Memory and Emotion

Daniel Reisberg
Paula Hertel,
Editors

OXFORD UNIVERSITY PRESS

Memory and Emotion

Series in Aective Science

Series Editors
Richard J. Davidson
Paul Ekman
Klaus Scherer
The Nature of Emotion: Fundamental
Questions
Edited by Paul Ekman and Richard J.
Davidson
Boo! Culture, Experience, and the Startle
Reex
By Ronald Simons
Emotions in Psychopathology: Theory
and Research
Edited by William F. Flack Jr. and
James D. Laird
What the Face Reveals: Basic and Applied
Studies of Spontaneous Expression
Using the Facial Action
Coding System (FACS)
Edited by Paul Ekman and Erika
Rosenberg
Shame: Interpersonal Behavior,
Psychopathology, and Culture
Edited by Paul Gilbert and Bernice
Andrews
Aective Neuroscience: The Foundations
of Human and Animal Emotions
By Jaak Panksepp
Extreme Fear, Shyness, and Social
Phobia: Origins, Biological Mechanisms, and Clinical Outcomes
Edited by Louis A. Schmidt and
Jay Schulkin
Cognitive Neuroscience of Emotion
Edited by Richard D. Lane and
Lynn Nadel

The Neuropsychology of Emotion

Edited by Joan C. Borod
Anxiety, Depression, and Emotion
Edited by Richard J. Davidson
Persons, Situations, and Emotions:
An Ecological Approach
Edited by Hermann Brandsttter and
Andrzej Eliasz
Emotion, Social Relationships, and
Health
Edited by Carol D. Ry and Burton
Singer
Appraisal Processes in Emotion:
Theory, Methods, Research
Edited by Klaus R. Scherer, Angela
Schorr, and Tom Johnstone
Music and Emotion: Theory and Research
Edited by Patrik N. Juslin and John A.
Sloboda
Handbook of Aective Sciences
Edited by Richard J. Davidson, Klaus
Scherer, and H. Hill Goldsmith
Nonverbal Behavior in Clinical Settings
Edited by Pierre Philippot, Erik J. Coats,
and Robert S. Feldman
Thinking about Feeling: Contemporary
Philosophers on Emotions
Edited by Robert C. Solomon
Memory and Emotion
Edited by Daniel Reisberg and
Paula Hertel

Memory and Emotion

Edited by
Daniel Reisberg and Paula Hertel

2004

Oxford New York

Auckland Bangkok Buenos Aires Cape Town Chennai
Dar es Salaam Delhi Hong Kong Istanbul Karachi Kolkata
Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi
Sao Paulo Shanghai Taipei Tokyo Toronto

Copyright 2004 by Oxford University Press, Inc.

Published by Oxford University Press, Inc.

198 Madison Avenue, New York, New York 10016

www.oup.com

Oxford is a registered trademark of Oxford University Press

All rights reserved. No part of this publication may be reproduced,

stored in a retrieval system, or transmitted, in any form or by any means,

electronic, mechanical, photocopying, recording, or otherwise,

without the prior permission of Oxford University Press.

Library of Congress Cataloging-in-Publication Data

Memory and emotion / edited by Daniel Reisberg and Paula Hertel.

p. cm.(Series in aective science)

Includes bibliographical references and index.

ISBN 0-19-515856-3

1. Autobiographical memory. 2. Emotions. 3. Psychophysiology.

4. Psychiatry. I. Reisberg, Daniel. II. Hertel, Paula. III. Series.
BF378.A87 M46 2003
152.4dc21
2003006595

2 4 6 8 9 7 5 3 1
Printed in the United States of America
on acid-free paper

he study of the interplay between emotion and memory inevitably involves a mixing of perspectives. After all, the study of emotion is an important and sophisticated research enterprise in its own right, and
so, too, is the study of memory. But, in addition, each of these topics in turn has
been studied from multiple perspectivesthe nature of emotion in normal populations and in pathological populations, in children, in adults, and in the elderly;
the nature of the biological changes, and then the psychological and subjective
changes, that accompany (and perhaps constitute) emotion; the range of emotions common in day-to-day life, and also the range of emotions one experiences
only in extreme circumstances.
It should be obvious that each of these perspectives on emotion has its own
value and asks questions that do not arise within these other contexts. But it
should also be obvious that common themesor points of dissonancemay
emerge when we start putting these various perspectives side by side. Unfortunately, though, those common themes often do not come into view: After all, in
our eld research tends to be published in specialized venues, and so there is less
cross-talk among related specialties than one might wish. It is precisely this situation that creates a need for volumes like this one.
In this volume, we have brought together some of the leading investigators
in the broad area of emotion and memory, deliberately seeking to span as best
we could the full range of approaches visible in the current scene. This juxtaposition should provide easy access to any reader, from any perspective, hoping to
enter this broad research literature. In addition (and as wed hoped), the juxtaposition highlighted a number of themes that cross research areas. For example,
many authors have proposed that emotion promotes memory, so that materials
one has experienced while emotional are better retained than materials experienced while one is calm. Indeed, some evidence suggests that emotional char-

acteristics of an event help to ameliorate memorial deciencies of one sort or

another (in elderly adults, for example, or in individuals diagnosed with schizophrenia). But a broader view of the literature makes it plain that the opposite
pattern also appears in some circumstances, with emotion serving to undermine
memoryeven memory for nonemotional events, particularly in the case of
anxiety and depression.
How should we think about this contrast? Is one of the eects artifactual? Or
(as seems likely) are both eects genuine, with other important variables guiding the direction of the event? And, if the latter, what are those variables? Is the
nature of the emotion crucial? The strength of the emotion? The nature of the
to-be-remembered materials? The way in which memory is assessed? Each of
these proposals arises in one or more chapters in this volume, and each of the
proposals has arguments in its favor. Our intention in this preface, therefore, is
surely not to settle the question of contrast, or even to suggest that the question
should be settled in any straightforward way. Instead, we raise this point simply
as a reminder that investigators in this broad domaindespite their very dierent methods, working assumptions, and forms of dataare often working on
closely related questions, so that all may benet from cross-talk among these
various approaches. It is that hope that provides the main impetus for this volume.
The volume begins with a contribution by Daniel Reisberg and Friderike
Heuer. Reisberg and Heuer celebrate the progress that has been made in our
understanding of how people remember emotional events, but also highlight the
substantial gaps in our knowledge. A prominent theme in their chapter is that,
in a wide range of circumstances, emotion promotes memory for an events central materials, but seems to have the opposite eect, undermining memory, for
details at an events periphery. Reisberg and Heuer argue, however, that this
latter eect may be produced not by emotion itself but by the presence of powerful attention magnets found within some (but by no means all) emotional
events. From this base, Reisberg and Heuer explore possible limits of this data
pattern, with an eye on how dierent types and dierent strengths of emotion
may inuence memory dierently.
The second chapter, by Tony Buchanan and Ralph Adolphs, explores the
neuroanatomy of emotional memory, arguing that the brain mechanisms that
encode, consolidate, and retrieve memories may operate dierentially in emotional and nonemotional contexts. Specically, the chapter emphasizes the role
of the amygdala in the enhancement of memory for emotional events, largely
through the amygdalas inuence over brain areas (including the hippocampus
and striatum) during the period of memory consolidation and also during the
retrieval of emotional memories. Data are drawn from human studies and from
studies of other species, from lesion evidence, and from neuroimaging. The specic eects reported, however, are not uniform, and here, as in other chapters,
we see evidence of emotions mixed eects on memory. Thus, for example, bilateral amygdala damage seems to eliminate both emotions tendency to promote

vi

memory for an events gist and also its tendency to undermine memory for an
events periphery. At the same time, the eects do seem general in other ways,
and the chapter argues that the amygdala eects (and, more broadly, emotions
memory eects) depend on arousal itself, not the emotions valence.
While these rst two chapters acknowledge emotions mixed eects on
memory, there is no question that these chapters emphasize the ways in which
emotional events seem to be remembered better than neutral (but otherwise
comparable) events. This emphasis is reversed, however, in the third chapter,
by Jessica Payne, Lynn Nadel, Willoughby Britton, and Jake Jacobs. They focus
on the biopsychological eects of trauma and how these inuence memory. They
argue that trauma (and, more precisely, the stress that usually accompanies
trauma) has identiable eects on the hippocampus, impairing both the neuronal
structure and the function of this brain region. As a direct consequence, stress
(especially uncontrollable stressors) impairs various forms of memory. However,
not all memories suer this eect, and the chapter seeks to explain this point by
arguing that the experience of stress (through its impact on the hippocampus)
causes stressful events to be recorded in a fragmented manner, with the elements of the event not woven into a coherent remembered episode. At the same
time, emotion works (via the amygdala) to promote memory for the gist of an
event, leading to well-encoded memories for the thematic content of an emotion
event, but, again, without the coherent spatio-temporal framework needed to
organize the memory (because this framework relies on hippocampal circuits
disrupted by stress). In this fashion, the authors seek to explain both the positive
eects of emotionality on memory and its negative eectswith an emphasis
on dierentiable neural structures and, with that, on dierent types of remembered information.
The fourth chapter, by Richard McNally, Susan Clancy, and Heidi Barrett, is
also on trauma and, like the Payne et al. chapter, seeks to understand the conditions under which traumatic events are remembered or forgotten. The chapters focus, however, is on the frequently espoused belief that all trauma will be
forgottena belief presumably based on the notion that too much emotion hurts
memory. In particular, McNally et al. review the evidence relevant to the debate
about the extent to which trauma victims typically repress and then later recover
memories of the traumatic event. In examining the reported evidencefrom
victims of childhood sexual abuse, concentration-camp survivors, war veterans,
and alien abducteesthe authors distinguish among phenomena that have
more often been collapsed in discussions of repressed and recovered memories.
They invite us to consider that the forgetting of traumatic events can arise from
quite normal phenomenasuch as the absence of rehearsal or the initial lack
of attention to some aspects of the eventinstead of special mechanisms of motivated repression. Another important feature of the chapter is their review of the
laboratory research on forgetting mechanisms employed by people who have
suered through various types of trauma. The pursuit of this line of research

vii

seems certain to shed light on the memory processes that distinguish those who
remember from those who forget and, in turn, from those who forget and then
remember.
Chapter 5, by Colin MacLeod and Andrew Mathews, considers how memory
might be inuenced by a variety of emotional states and conditions experienced
by people with anxiety disorders. This chapter represents a thorough review of
research performed with people who describe themselves as generally anxious
(without formal diagnosis), as well as with people who have been diagnosed
as experiencing generalized anxiety disorder, post-traumatic stress disorder,
phobias, obsessive-compulsive disorder, and panic disorder. In the context of research on mood congruent memory, one might expect that these individuals
will better remember stimuli that t with their anxious thoughts and beliefs;
this expectation, however, is clearly challenged by the results that MacLeod and
Mathews review. In some cases, anxious people do show evidence of anxietyrelated biases in memory, but the chapter argues that these probably result from
special instances of emotional interpretation of events with ambiguous meaning. Under conditions less prone to interpretive ambiguity, anxious people tend
not to remember in emotionally special ways.
Chapter 6, by Paula Hertel, addresses what is by far the largest research area
in the examination of memory in clinical groupsresearch conducted with
depressed or naturally unhappy (but possibly nondiagnosed) people. The
chapter theme is the connection between memory phenomena and habits of
thought. More so than most anxious people, depressed people ruminate about
their troubles. The practiced thought patterns of rumination facilitate memory
for emotionally consistent events and interfere with memory for other events.
Moreover, as Hertel wants us to understand, these habits take over under conditions of poor cognitive control, the main feature of cognition in depression.
Hertel argues that the negative consequences of habitual thinking can be overcome by external control or by the training of new habits of thought.
The literature on emotional memory in individuals diagnosed as schizophrenic is much less extensive than the literatures in depression and anxiety,
and it is often ignored in collections on cognition and emotion. Recently, however, it has been augmented by the research conducted by Jean-Marie Danion,
Caroline Huron, and their colleagues. The chapter here (by Danion, Huron,
Lydia Rizzo, and Pierre Vidailhet) rst reviews evidence that both memory disturbances and emotional disturbances characterize schizophrenia and then
raises questions about their interaction. Danion et al. argue that memory for
emotional material operates normally in schizophrenic states when emotional aspects of experience are genuinely noted. Of course, often these experiences are not interpreted by a person with schizophrenia in a fashion
that accurately reects the events emotionality, with a corresponding memory
decit. But in many cases the emotional characteristics of an event require little
controlled attention, and, under these conditions, we should expect to see intact

viii

emotional memory, even in this population disrupted in so many other ways.

Finally, in a concluding section of the chapter, Danion and his coauthors make
the provocative suggestion that intact emotional processing potentially plays a
role in perpetuating delusional cognitions, through constructive memory
processes.
The eighth chapter continues the emphasis on emotional remembering in
special populations, but the population at issue here is not pathological. Instead,
Robyn Fivush and Jessica Sales examine emotional memory in children. They
also focus on a theme that appears in many other chapters of the volumethe
fact that the functional role of emotional remembering often depends on attributes of memories other than their historical accuracy. In particular, Fivush
and Sales discuss the ways in which the structure and content of young childrens
emotional memories is co-constructed by the children and their parents, a process, they argue, that varies from culture to culture and also depends on the
childs gender. This co-construction is important for many reasons; among them,
it is one of the essential means through which children gain a foundation for
understanding themselves and their autobiographies. Also crucial here are the
ways in which parent-child reminiscing about stressful experiences can guide
the childs understanding of and coping with aversive events. Exploring this latter
point demands a comparison of how parents and children reminisce about both
aversive and pleasant events and also how they reminisce about both emotionally
mild events and emotionally extreme ones, themes that provide another important
point of contact between this chapter and other chapters in the volume.
In chapter 9, Mara Mather discusses memory at the other end of the lifespanin the elderly. She rst examines age-related changes in emotional
processingin mechanisms of emotional regulation and arousal, for example,
considered both psychologically and in terms of their neural mechanisms
and from this base oers an intriguing set of suggestions about how these
changes should inuence emotional memory. Among other issues, she considers whether the older persons improved ability to regulate emotion implies
that memories should become more emotionally gratifying, as well as whether
the emotional qualities of experience might actually protect an individual
against the age-related decline in memory. These suggestions are then evaluated in her subsequent review of the current literature on age dierences in
the eects of emotion on memory. In the end, Mather argues that emotional
memory does constitute a somewhat surprising island of maintained functioning in a sea of general decline in old age.
Chapter 10, by Robin Edelstein, Kristen Alexander, Gail Goodman, and
Jeremy Newton, covers topics that have arisen in many other chaptersthe
memory eects of trauma, emotional remembering by children, the long-term
durability of emotional memories, and so onbut covers these themes from the
perspective of the legal system, asking how eyewitnesses to crimes remember the
events they have observed (or, in many cases, the events in which they were vic-

ix

timized). Of special importance here is the question of whether laboratory ndings taken as characterizing emotional memory can be reasonably applied to
real-life crime situations. Also prominent in this chapter is the special case of
memories for childhood sexual abuse, including memories that are apparently
lost and then recovered. The chapter discusses the complexity that arises when
one tries to assess these memories and also factors (including a tendency toward
dissociation, or various forms of psychopathology) that play a role in determining when a traumatic event will be vividly remembered and when that event will
(apparently) be forgotten.
Finally, chapter 11, by Robert Kraft, focuses on a particularly important form
of emotional memory: memory of the Holocaust by its survivors. Throughout
this book, chapters have commented on the memory eects of traumatic or otherwise horric events; chapters have likewise commented on the process of sharing emotional memories with othersand perhaps sharing again and again. Just
as important is the fact of deliberately not sharing memories. How does this sharing (or not sharing) shape the recollection? Archives of Holocaust memories
speak powerfully to these issues, and Kraft provides a close qualitative analysis
of these memories in order to explore (among other topics) the general characteristics of traumatic memory, how emotion itself is recalled, and how the recall
of a memory can lead to the re-experiencing of emotion. Also of interest here is
the way in which these emotional memories can shape the survivors beliefs and
emotions throughout their lives. All of these considerations lead Kraft to proposals about multiple systems of memory that in some ways echo, and in other
ways may challenge, related multiple-systems proposals oered in several other
chapters.
We end with the happy task of thanking those who have helped us in assembling this volume. First, we are grateful to Catharine Carlin, our editor at
Oxford University Press, who persuaded us to launch this project, and who has
been enthusiastically encouraging ever since. We are also grateful to John
Rauschenberg, also at OUP, for his labors in putting the volume together. We
also thank the Series Editors, Richard Davidson, Klaus Scherer, and Paul Ekman,
for their support of this endeavor. Finally, we are grateful to Pat Ullmann, in
Trinity Universitys Instructional Media Service, for transcribing some gures.

Contributors,

xiii

1. Memory for Emotional Events,

3

Daniel Reisberg and Friderike Heuer

2. The Neuroanatomy of Emotional Memory in Humans,
Tony W. Buchanan and Ralph Adolphs

42

3. The Biopsychology of Trauma and Memory,

76

Jessica D. Payne, Lynn Nadel, Willoughby B. Britton,

and W. Jake Jacobs

4. Forgetting Trauma?
129

Richard J. McNally, Susan A. Clancy, and Heidi M. Barrett

5. Selective Memory Eects in Anxiety Disorders: An Overview of

Research Findings and Their Implications,

155

Colin MacLeod and Andrew Mathews

6. Memory for Emotional and Nonemotional Events in Depression:

A Question of Habit?
186

Paula Hertel
7. Emotion, Memory, and Conscious Awareness in

Schizophrenia,
217

Jean-Marie Danion, Caroline Huron, Lydia Rizzo,

and Pierre Vidailhet

8. Childrens Memories of Emotional Events,

Robyn Fivush and Jessica McDermott Sales

242

xi

9. Aging and Emotional Memory,

Mara Mather

272

10. Emotion and Eyewitness Memory,

308

Robin S. Edelstein, Kristen Weede Alexander, Gail S. Goodman,

and Jeremy W. Newton
11. Emotional Memory in Survivors of the Holocaust:

A Qualitative Study of Oral Testimony,

347

Robert N. Kraft
Index,

xii

391

Ralph Adolphs, Department of Neurology, University of Iowa, Iowa City, Iowa

Kristen Weede Alexander, University of California, Sacramento, California

Heidi M. Barrett, Department of Psychology, Harvard University, Cambridge,

Massachusetts

Willoughby B. Britton, Department of Psychology, University of Arizona, Tucson,

Arizona

Tony W. Buchanan, Department of Neurology, University of Iowa, Iowa City,

Iowa

Susan A. Clancy, Department of Psychology, Harvard University, Cambridge,

Massachusetts

Jean-Marie Danion, Unit INSERM 405, Hopital Universitaire, Strasbourg, France

Robin S. Edelstein, Department of Psychology, University of California, Davis,

California

Robyn Fivush, Department of Psychology, Emory University, Atlanta, Georgia

Gail S. Goodman, Department of Psychology, University of California, Davis,

California

Paula Hertel, Department of Psychology, Trinity University, San Antonio, Texas

Friderike Heuer, Lewis and Clark College, Portland, Oregon

Caroline Huron, INSERM, Paris, France

W. Jake Jacobs, Department of Psychology, University of Arizona, Tucson,

Arizona

xiii

Robert N. Kraft, Department of Psychology, Otterbein College, Westerville, Ohio

Colin MacLeod, Department of Psychology, University of Western Australia,

Perth, Australia

Mara Mather, Department of Psychology, University of California, Santa Cruz,

California

Andrew Mathews, MRC Cognition & Brain Sciences Unit, Cambridge, United

Kingdom

Richard J. McNally, Department of Psychology, Harvard University, Cambridge,

Massachusetts

Lynn Nadel, Department of Psychology, University of Arizona, Tucson, Arizona

Jeremy W. Newton, Department of Psychology, University of California, Davis,

California

Jessica D. Payne, Department of Psychology, University of Arizona, Tucson,

Arizona

Daniel Reisberg, Department of Psychology, Reed College, Portland, Oregon

Lydia Rizzo, Laboratoire Universitaire de Psychologie, Metz, France

Jessica McDermott Sales, Department of Psychology, Emory University, Atlanta,

Georgia

Pierre Vidailhet, Unit INSERM 405, Hopital Universitaire, Strasbourg, France

xiv

Memory and Emotion

This page intentionally left blank

ollege graduation. The death of a pet. A particularly romantic

evening. A moments reection will persuade most people that
events like theseevents that were emotional when they occurredprovide
many of their most vivid, most detailed, most compelling autobiographical
memories.
For many of us, these emotional memories are a basis for engaging and
sometimes enjoyable reminiscence. More important, though, these memories
also matter for us in important ways. After all, these emotional events are likely
to have been consequential for us in one way or another, and are, in most cases,
closely linked to issues, goals, or people that we care deeply about. It seems certain, therefore, that the recollection of these crucial events is likely to shape
our sense of who we are and also our broad perceptions of the world. This, in
turn, will inuence many aspects of our actions, perceptions, and beliefs. In
addition, many of these emotional events are consequential for other people,
and so we may be called on to report the event in one setting or another; when
that happens, it is of obvious interest to ask how full and accurate that report
will be.
These broad issues provide the motivation for our inquiry into the nature
of autobiographical recollection. As we will see in this chapter, research
in this domain has made considerable progress, but on some key issues, the
available data are still rather sparse. Hence, this chapter simultaneously celebrates our advances and highlights the points on which more work is urgently
needed.
3

The Accuracy

of Emotional Memories

Setting the Issue: Vividness Versus

Memory Accuracy

There is no question that emotional memories tend to be quite vivid, and, in one
early study, we reported a correlation of .71 between participants ratings of their
memories vividness and their ratings of how emotional the original event had
been (Reisberg, Heuer, McLean, & OShaughnessy, 1988). Interestingly, this
relationship was observed no matter what emotion was attached to the event,
so the correlation between vividness and emotionality ratings was .89 for sad
events, .68 for angry events, .90 for fearful events, and .71 for happy events
(Reisberg et al., 1988, Experiment 2). It also didnt matter whether the event
being recalled was a personal one (e.g., death of a parent, graduation from college) or a public one (e.g., rst moon landing). The personal events were more
vividly recalled and were more emotional than the public events, but the relationship between vividness and strength of aect was the same for both (.661
and .641, respectively).
Other studies conrm this pattern, observing extreme memory vividness both
for memories of traumatic events and memories for extremely positive events
(e.g., Pillemer, Goldsmith, Panter, & White, 1988; Rubin & Kozin, 1984). One
recent illustration of this nding is provided in Walker, Vogl, and Thompson
(1997), who found that the rated emotionality of an event was consistently a
strong predictor of whether participants believed they could remember the event
or not, and the same data pattern emerged when participants were recalling
extremely pleasant events and when they were recalling extremely unpleasant
ones. Likewise, Porter and Birt (2001) invited their participants to describe their
most traumatic autobiographical memory and also their most positive emotional
experience. In both cases, they found that strong emotion was reliably associated
with a high degree of memory vividness. Similarly, Berntsen (2001) examined
the highly vivid involuntary (spontaneously arising) memories often called ashbacks; the data showed little dierence between involuntary memories for
trauma and those for extremely happy events.
But are these emotional memories accurate? It would be easy to assume they
must be, since, after all, an emotional event is likely to be important to us, virtually guaranteeing that we will pay close attention as the event unfolds. In
addition, emotional events are often emotional precisely because they are related to issues we care about and have thought about in other contexts; this
will foster the sort of memory connections that we know promote retention and
recall. Moreover, we tend to mull over emotional events in the minutes (or

hours) following the event, and this is tantamount to memory rehearsal (cf.
Bower, 1992). Finally, we are likely to revisit emotional events periodically,
in our own thoughts or in conversations with others, and this too should promote retention.
Even with these considerations, however, it is certain that at least some of
our emotional memories do contain errors, and some may be wholly mistaken.
Plainly, therefore, the accuracy of emotional memories must be tested and cannot be taken for granted. Evidence relevant to this point comes from many
sources, but most forcefully from the debate over ashbulb memories, a debate prominent in the research literature in the 1980s and 1990s. The term
ashbulb memory was coined by Brown and Kulik (1977), and referred to
the exceptionally clear and detailed recollection people seem to have for singular, emotional, and consequential events they have experienced. Brown and
Kulik oered the example of peoples memory for John Kennedys assassination, an event that was still remembered as if it were yesterday many years
after the event; more current examples would include the memory that many
people have for Princess Dianas death (in 1997), the O. J. Simpson trial (in
1995), and the destruction of the World Trade Center (in 2001).
We will return to the topic of ashbulb memories later in this chapter; for present
purposes, we wish only to make a simple point about these memories. Flashbulb
recollections tend to be extraordinarily vivid and detailed, and these memories
are recalled with enormous condence that the memory is, in fact, correct. But
ashbulb memories can be shown in some circumstances to be wrong, making
it plain that we cannot equate memory vividness with memory accuracy or
memory condence with accuracy. One compelling illustration of this point
comes from Neisser and Harsch (1992), who interviewed people days after the
1986 space-shuttle disaster, to determine where they were when they heard the
news of the disaster, who brought them the news, and so on. Neisser and Harsch
then reinterviewed these same people roughly 3 years later (3234 months after
the initial data collection), to see how memories for the event had fared over this
time span. They found that people still reported detailed, high-condence memories in this 3-year follow-up, but many of these memories were simply wrong, sometimes in important ways, completely misrepresenting the original event.
We hasten to say that other studies have yielded rather dierent results,
with impressive accuracy in peoples ashbulb recollection (e.g., Conway et al.,
1994), and so, as we have noted, we will need to return to this topic before we
are done in order to address the question of why some ashbulb memories seem
accurate and long-lasting, while others do not. For now, though, we draw from
the Neisser and Harsch data only the moral that emotional memories can be
extremely vivid, extremely compelling, and yet completely out of step with the
historical facts. This demands that we test the accuracy of emotional memories and not assume it.

Studying Emotional Memory

in the Laboratory

Studies of ashbulb memory obviously provide insights into how people remember the genuinely emotional events that actually take place in their lives. The
disadvantage of these studies, though, is that we often have no way of knowing
exactly what happened within the target event and thus no means of assessing
the accuracy of memory. For this reason, the study of memory accuracy has often
been taken into the laboratory so that we now have full knowledge of, and full
control over, the to-be-remembered event.
These laboratory studies have employed a variety of to-be-remembered materials. Here our emphasis will be on memory for events, and so we will hold to
the side studies of peoples memory for emotional pictures (e.g., Canli, Zhao,
Brewer, Gabrieli, & Cahill, 2000), emotion-laden word lists (e.g., Dietrich et al.,
2001; Jones, OGorman, & Byrne, 1987), and memory for humor (e.g., Schmidt
& Williams, 2001). Even with this narrowed focus, however, the evidence derives from a diversity of studies. In some studies, participants have witnessed an
event presented via a series of slides, depicting successive moments within a story,
with an accompanying (tape-recorded) narrative telling the emotional tale. In
a smaller number of cases, the to-be-remembered event has been presented as a
video clip (or, in just a few studies, an animation). An even smaller number of
studies have employed live events, witnessed by the studys participants.
In virtually all cases, though, the studies compare participants memory for
emotional materials to their memory for neutral materials, with the assessment
of memory typically provided by a ne-grained test in four-alternative forcedchoice (4AFC) format. The experimental materials are usually claimed to be
emotional on three bases: the content itself, specically chosen (or designed) for
its emotional themes; self-reported emotionality from participants viewing the
content; and physiological measures of arousal (usually heart rate). The neutral
materials are then matched as well as possible to the experimental materials in
structure and content but, of course, are claimed not to be emotional (usually
on the same three grounds).
As an illustration, several dierent laboratories have used the doctor/mechanic stimulus set rst developed by Heuer (1987). In both versions of this
stimulus, participants see slides and hear a narration about a mother and son
going to visit father at work. In the emotional version of the stimulus, the father is a surgeon, and the son watches as his father completes a dicult surgical procedure. One slide for this sequence shows a scene of surgery with the
patients viscera in plain view; another shows the severed and reattached legs
of a child. In the neutral version of the stimulus, the father is an automobile
mechanic, and the son watches as his father completes a dicult repair. One
slide shows a scene of the repair with the engine in view; another shows the
supposedly damaged part.

The Easterbrook Hypothesis

Most of the laboratory studies of emotional remembering have been cast as tests
of two intertwined hypotheses. One hypothesis concerns participants memory
for the gist or central materials within the event. As we will see, the weight of
the evidence on this point is reasonably clear: in general, emotion seems to promote memory for an events gist or center.
A second hypothesis derives from claims oered many years ago by
Easterbrook (1959). Based on animal studies, Easterbrook proposed that arousal
causes a narrowing of attention, so that an aroused organism becomes less sensitive to information at the periphery of an event. As a consequence, the organism is likely to become more sensitive to information at the center of the event,
perhaps because of diminished distraction from the periphery, or perhaps because
the organisms attentional resources are more concentrated on the events center. In either case, the hypothesis implies that the memory advantage associated
with arousal will be observed only for the events center. Arousal should produce
a disadvantage in remembering the events periphery (relative to memory for otherwise comparable neutral materials).
Easterbrooks own studies manipulated arousal by depriving animals of food
for various lengths of time; his measures then assessed the animals sensitivity
to cues in its immediate environment. There is obviously some extrapolation
needed, therefore, to apply his hypotheses to the situation of a human, aroused
through anger or fear or joy, remembering a complex event. Even so, that extrapolation is invited by a pattern often observed among victims or witnesses to
crimes, a pattern known as the weapon focus eect. This term refers to the fact
that the witnesses to crimes often seem to lock their attention onto the
criminals weapon and seem oblivious to much else in the scene. As a result of
this attentional pattern, later on the witness will remember the perpetrators gun
or knife with great clarity but may remember little else about the crime, including such crucial details as what the perpetrator looked like!
Weapon focus has often been alleged by those in law enforcement and has
also been documented in a variety of laboratory studies (e.g., Loftus, Loftus, &
Messo, 1987; Stanny & Johnson, 2000; Steblay, 1992). In obvious ways, this
eect parallels the pattern observed by Easterbrook, with good memory for items
at the center of the crime (the gun or knife) but poor memory for items at the
periphery. Could it be, therefore, that Easterbrooks claim applies to emotional
remembering in general?

Studies of Memory Narrowing

Many studies have asked whether participants show the pattern of memory
narrowing for emotional events that we might expect by extension of the
weapon-focus data. Early evidence on this point was mixed, and, indeed, one can

easily nd in the literature claims that emotion promotes memory, claims that
emotion undermines memory, and also claims that emotion has diverse eects.
For example, and in keeping with the Easterbrook suggestion, Christianson and
Loftus (1991) found improved memory for the central materials within an emotional event but impoverished memory for the peripheral materials, in comparison to memory for a neutral (but otherwise similar) event. However, both Cliord
and Scott (1978) and Loftus and Burns (1982) reported evidence suggesting that
emotionality impaired memory, and, using a somewhat dierent paradigm, Heuer
and Reisberg (1990) found that emotion seemed uniformly to improve participants memory, a benet that emerged for both central and peripheral materials (in comparison to memory for a matched neutral event).
Although confusing when they were rst published, these contrasting data
patterns are in retrospect easily explained. If emotion does have opposite eects
on central and peripheral materials, it is essential that we dene these categories
with care. To the extent that items on the memory test are miscategorized, this
could easily obscure the results pattern. Similarly, our ability to make comparisons across studies is obviously compromised if dierent investigators use dierent categorization schemes for central and peripheral materials; that was, in fact,
a problem in this early literature. Christianson and Loftus (1991), for example,
dened their categories largely in spatial or perceptual terms; peripheral details
were those that were truly in the background. Heuer and Reisberg (1990), on the
other hand, used denitions that were more conceptual and counted as central
any bits of information directly relevant to the plot, or in any way important for
how the story unfolded. Peripheral information, in contrast, was information that
could be changed without in any material way changing the story.
Which of these categorization schemes carves nature at the joints? Burke,
Heuer, and Reisberg (1992) decided to treat this as an empirical question and
therefore used a very ne-grained categorization scheme in order to ask, in a
data-driven fashion, which of these distinctions were relevant to emotions
memory eects and which were not. In their data analysis, Burke et al. separated
items that were relevant to how the story unfolded from items that were not (thus
replicating Heuer and Reisbergs relevance-based distinction). The plot-relevant
items were then subdivided into gist items, items that essentially dened the
story, and basic level visual information, items that, in the broadest terms,
described what each of the slides in the sequence showed. Likewise, the plotirrelevant items were also subdivided, into those (irrelevant) details that happened to be spatially associated with plot-relevant actors or objections and those
truly in the background. Finally, all of these categories were subdivided once
again, but this time temporally, with questions divided according to whether they
probed memory for materials that happened before, during, or after the parts of
the story that were, in fact, arousing.
Figure 1.1 depicts the Burke et al. categorization scheme, and, in a rough fashion, describes their results. As can be seen, the dierence between plot-relevant

Figure 1.1. A schematic description of the Burke et al. design and findings.

and plot-irrelevant materials does matter, inasmuch as the pattern in the top half
of the gure is dierent from the pattern in the bottom half. Likewise, the spatial/perceptual distinction also matters, evident in the contrast between the
gures third and fourth rows. Finally, the temporal dimension also matters, as
shown in the contrast between the tables middle column and its two outer columns. (For related data, showing a similar eect of the temporal dimension, see,
for example, Bornstein, Liebel, & Scarberry, 1998.)
How to summarize this pattern? First, these data do conrm the memorynarrowing pattern, with emotion improving memory for materials central
to the event and hurting memory for more peripheral materials. Second, the
denition of the events center is complexwith materials favored by emotion
if they are in any fashion tied to the action in the storyeither conceptually or spatio-temporally.
Several other studies have since conrmed this pattern, albeit not at the same
ne grain. Safer, Christianson, Autry, and sterlund (1998), for example, showed
participants a sequence in which a woman was shown either gathering owers
in a park (neutral version) or stabbed in the throat and lying on the ground,
bleeding (emotional version). The participants memory was then tested with
various photographs diering only in how much of a close-up they were; par-

ticipants were asked to select the exact photo they had seen in the earlier sequence. We know from other studies (e.g., Intraub & Richardson, 1989) that
people often remember photographs as being less zoomed-in than they actually
were and correspondingly remember the photo as including more of the background than it actually did, a pattern known as boundary extension. Safer et al.,
however, found precisely the opposite pattern, with the emotional photos remembered as more zoomed in. Apparently, the participants memories excluded the
peripheral information and also excluded the fact that there even was peripheral information.
Similarly Wessel and Merckelbach (1997, 1998) invited spider phobics to the
laboratory and, in one procedure, showed them a large, live spider (contained
within a glass jar) and, in another procedure, showed them pictures of spiders
mounted on a bulletin board. In both cases, these were particularly arousing
stimuli for these participants but not for control subjects. In a subsequent
memory test, the more aroused (phobic) participants showed the expected pattern of narrowing, with better memory for the events center (the spider) and
worse memory for the events periphery than control participants.
It should be acknowledged, though, that not all studies conrm this broad
picture, and two studies, one by Libkuman, Nichols-Whitehead, Grith, and
Thomas (1999) and one by Wessel, van der Kooy, and Merckelbach (2000), have
failed to replicate the memory narrowing result. (We return to these nonreplications later.) Even so, the memory-narrowing pattern associated with
emotional events has been replicated often enough to be regarded as wellestablishedespecially when it is joined with the separate but substantial body
of research directly examining the closely related weapon eect.
Overall, then, where does this leave us? We began this section by asking
whether emotional events are remembered not just vividly but also accurately.
The answer depends on what aspects of the emotional event we are considering. If
we focus on central materials within an eventthe gist of the event and details
that are spatially central and also associated with the gistthe data seem reasonably uniform, with emotion seeming rather reliably to improve memory; we
will consider even more data on this point (and some complications) later in the
chapter. If we look instead at more peripheral materials (plot-irrelevant details
that are spatially removed from the action), the evidence is less clear, but, overall, the data seem to indicate that memory is impaired by emotion. Thus, emotional events seem to be remembered accurately but incompletely.

Could Memory Narrowing

Be Artifactual?

Emotional events are distinctive in many ways. Quite obviously, they are accompanied by the feelings and bodily changes that we call emotion and, for this
reason alone, may be remembered dierently from neutral events. But, as we

10

have already mentioned, several other factors may also be crucial here. Emotional events typically receive closer scrutiny than neutral events, and it may
be this scrutiny, not the emotion itself, that explains the data we have so far reviewed. Emotional events are also likely to be perceived as worth thinking about
after the fact and so probably are rehearsed to a greater extent than neutral events
are. Perhaps it is this that shapes how these events are remembered, rather than
the emotionality per se. In addition, emotional events are likely to be somewhat
unusual (if they were more familiar, it seems likely that they would lose their
emotional power), and this, too, may inuence how they are remembered.
Finally, emotional events are usually related to important themes or goals in our
lives; this is presumably what makes them emotional in the rst place. Perhaps
it is this relation to important themes that is crucial for memory and, again, not
the emotionality itself.
These factors are dicult to control in any study of emotional remembering,
but, even so, at least some evidence suggests that it is the emotionality that
matters, over and above the undeniable contribution of these other points. For
example, it is important that, in the Wessel and Merckelbach (1997, 1998) studies of spider phobics, it is the same event that is being remembered by the aroused
and less aroused (nonphobic) subjects. This allows us to set aside a range of concerns that focus on some aspect of the emotional events (their familiarity, their
coherence, their plausibility, their intelligibility) other than their emotionality.
Moreover, a number of studies have directly tackled some of these extraneous factors. Christianson and Loftus (1991), for example, compared participants
memory for three types of stimuli: a neutral stimulus, an emotional stimulus,
and a stimulus designed to be unusual and attention-grabbing but unemotional.
Specically, the neutral stimulus told a story about a woman riding her bicycle;
the arousal story told of a woman injured while riding her bicycle; the unusual
story showed the woman carrying her bicycle on her shoulder. The results
showed similar patterns for the emotional and unusual story with regard to the
peripheral aspects of the story (a car seen in the background), but, for central
information (the color of the womans clothing), performance was markedly
better for the emotional event than for the unusual event. In short, then, the
emotional and unusual stories led to dierent memory patterns, suggesting that
the eects of emotion must be dierent from those of sheer distinctiveness or
unfamiliarity.
In a similar spirit, Heuer and Reisberg (1990) compared memory under four
dierent circumstances. One group of subjects viewed an emotional sequence,
and one a neutral sequence. (The doctor/mechanic series, already described, was
used in this study.) A third group viewed the neutral sequence but was specically urged to memorize the story as best they could; this condition seemed likely
to elicit extra rehearsal of the story and so could illuminate rehearsals eects. A
fourth group was urged to scrutinize the (neutral) story closely and was told
specically that the stimulus they were about to see paralleled a recent event in

11

the news; their task was to discern what that event was. (In truth, there was no
such deliberate parallel, but participants had no way to know this.) This group
was included to provide a comparison between emotions eects and the eects
of close scrutiny (without emotion) and postevent contemplation.
The comparison of these conditions is important if we are to isolate and identify emotions memory eects. It is unfortunate, therefore, that this study, done
some years ago, relied on a relatively crude measure of memory, one resting on
a categorization of central and peripheral materials that the investigators subsequently abandoned (cf. Burke et al., 1992). Nonetheless, it is notable that the
memory data for the emotion group were easily distinguishable from the data
for any of the other three groups. Thus, though some caution is needed here, this
study does suggest that emotions memory eects cannot be reduced to these
other factors.
Roughly the same conclusion follows from a number of studies that have
sought to examine emotions biological impact on memory. For example, several
investigators have suggested that the amygdala plays a crucial role in emotional
memory, a claim supported, for example, by studies of patients with UrbachWeithe disease, a disease that damages the amygdala without harming other
brain structures. Adolphs, Cahill, Schull, and Babinsky (1997; also chapter 2
in this volume) presented two such patients with the stimuli developed by Heuer
and Reisberg (1990). The patients found the emotional slides initially arousing,
but showed no memory benet from this arousal, in comparison to their memory
for neutral materials.
These data are interesting for what they say about the biological mechanisms
underlying emotions memory eects but are also important for our purposes in
this section. Presumably, Urbach-Weithe disease makes these stimuli no less
unusual and no less coherent; if it is these factors that made the stimuli memorable, then the patients with this disease would remember the stimuli just as
ordinary participants do. The data, of course, suggest otherwise.
A parallel argument can be made for a study by Cahill, Prins, Weber, and
McGaugh (1994). They showed (their adaptation of) the doctor/mechanic
stimuli to participants but, prior to the presentation, injected half of the participants with propanolol, a beta-adrenergic blocker chosen to diminish
emotions bodily (arousal) eects, and injected the remainder with a placebo. If
the memory eects of the emotional stimulus truly depend on emotion (and, in
particular, on the arousal that accompanies emotion), then the beta-blocker
would be expected to reduce or eliminate these memory eects. If, on the other
hand, the memory eects depend on factors like rehearsal or the unusual nature
of the story, then it is not obvious why the beta-blocker should have an eect.
Cahill et al.s results indicate that beta-blockers do reduce (and may even
eliminate) the memory dierences between neutral and arousal stories, buttressing the claim that it is indeed emotion that matters for memory, and not some
other attributes of these stimuli. Once again, though, caution is needed here,

12

because Cahill et al.s study did not distinguish between central and peripheral
elements of the story and so may not provide a complete portrait of emotions
eects.
In short, the evidence on these points is strongly suggestive, but not conclusive. The weight of the evidence suggests that emotion itself has an impact on
memory, with the eect plausibly depending on emotions arousing eects (which
are specically diminished by the administration of beta-blockers). For reasons
we have mentioned, however, this is certainly a point in need of further research
scrutiny.

Does All Emotion Have

the Same Eects on Memory?

Does the Source of

the Emotion Matter?

We have so far painted a relatively straightforward picture of emotions eects.

A large number of studies indicate that emotion improves memory for an events
center; a somewhat smaller, but still substantial, group of studies indicate that
emotion disrupts memory for an events periphery. There are, however, two
conspicuous problems with this claim. One is the studies, already mentioned, that
have explicitly tried, but failed, to replicate the memory-narrowing pattern
(Libkuman et al., 1999; Wessel et al., 2000). A second, and perhaps broader,
problem lies with the generalizability of the studies cited so far.
Emotions capacity to promote memory for an events gist or center has been
demonstrated in many settings, both inside the laboratory and out (e.g.,
Bohannon, 1988; Brewer, 1988; Linton & Melin, 1982; Pillemer, 1984). To
mention just a few of the more recent studies, Bluck and Li (2001) showed that
the strength of participants emotional feelings about the O. J. Simpson verdict
was predictive of how fully the event was recalled 8 months later. Bornstein et
al. (1998) reported that the emotional center of an R-rated commercially produced movie was better recalled than a comparable but neutral episode. Peterson
and Whalen (2001) reported that high stress levels facilitated young childrens
recall of an emergency trip to the hospital 5 years earlier. Davidson, Luo, and
Burden (2001) reported that children are better able to remember emotional behaviors than unemotional ones. And so on.
But what of the narrowing eect and, specically, the tendency for emotion
to diminish memory for an events periphery? Here the diversity of evidence is
actually rather limited, because virtually all of the studies relevant to this eect
have induced emotion in roughly the same way: by showing participants something gruesome, or gory, or anxiety-provoking. Examples include pictures of
disgured faces, a picture of a boy shot in the eye, a picture of a woman with her

13

throat slit, a picture of the severed legs of a child, or the sight of a dangerouslooking spider. All of these manipulations produce strong reactions, evident both
in heart rates and self-report, but one might worry that these studies do not represent emotion as it naturally occurs outside the laboratory. In particular, one
might argue that the emotion experienced outside the laboratory is more often
induced, not by a particular visual stimulus but instead by involvement and
empathy with an unfolding event. We become emotional, in other words, when
we encounter issues and information that are pertinent to our lives, goals, and
values (or, perhaps, the lives and values of people we care about). We refer to
this more common kind of emotion as thematically induced, in contrast to the
visually induced reactions involved in most previous studies, produced by the presentation of a specically dened emotional visual stimulus.
This distinction between thematically and visually induced emotion raises
two concerns about the extant evidence. First, can we generalize from the prior
studies, or does the evidence instead allow us only to characterize one type of
emotional event, with a dierent prole needed for other events? Second, and
more troubling, this distinction raises the specic possibility that the narrowing of memory, observed in many studies, may be artifactual. Recall that the
Easterbrook claim attributes the pattern of memory narrowing to arousal, with
the implication that this pattern would not be observed if the arousal were somehow avoided or diminished. An alternative possibility, however, is that this narrowing has nothing to do with arousal but is observed simply because the
experimental stimuli provide a highly salient stimulus to focus on, a strong attention magnet that seizes participants concentration during the event and
therefore dominates their subsequent recollection of that event.
We note that essentially the same concerns have been raised by investigators examining the phenomenon of weapon focus, which we described in an
earlier section. Many authors have attributed this pattern to the emotional
arousal experienced by the eyewitness; on this view, weapon focus is just another manifestation of the eect described by Easterbrook. Other authors have
pointed out, however, that the eyewitness may focus on the weapon simply
because it is by far the most important and interesting aspect of the visual
input. After all, what could be more important to a crime witness than to know
whether he or she is in immediate danger or not, and, to this end, nothing in
the scene is more important than knowing whether the weapon is cocked and
pointed at him or her, or whether the criminals nger is on the trigger? On
this logic, even an entirely calm witness might still show the weapon focus
pattern, zooming in on the weapon because looking toward the weapon provides crucial information!
As it turns out, the available evidence suggests that both of these mechanismsone hinging on arousal, and one hinging on the weapons visual importancemay play a role in producing weapon focus. In several studies, for
example, weapon focus has been observed even in the absence of emotion, indi-

14

cating the importance of the weapons potency as an attention magnet (Kramer,

Buckhout, & Eugenio, 1990; Loftus et al., 1987; Maass & Khnken, 1989). Other
studies indicate that it may be the unusualness of the weapon, and not the threat,
that produces weapon focus (Pickel, 1998). Still other studies, however, suggest
that the strength of the weapon focus eect increases as arousal increases, indicating that arousal also plays a role (e.g., Peters, 1988).
To the best of our knowledge, though, this issue has rarely been raised in the
study of emotional memory (i.e., in the absence of a weapon). In the next section, therefore, we describe four studies designed to explore these points.

Thematically Induced Versus

Visually Induced Emotion

As a rst step toward addressing the issues just raised, it seems sensible to ask
about the nature of the emotional events that actually ll our lives. Is the emotion in these events typically visually induced (in which case the laboratory studies might be representative of emotionality in our day-to-day lives), or is the
emotion typically thematically induced?
In a pair of studies, Laney, Heuer, and Reisberg (in press) asked participants
simply to list a series of emotional events from their lives; in their rst study, the
participants were all college undergraduates at a prestigious institution; in their
second study, the participants were much more diverse in ages, professions, and
educational backgrounds. In both cases, the participants were then interviewed
about the events they had listed, and, based on these interviews, the events were
coded as either thematically or visually arousing.
These two studies yielded virtually identical data. Despite a coding scheme
set up to bias things toward counting events as visually induced, the huge
majority of events reported by participants were in fact thematically induced
82% in the study with undergraduates and 71% with the broader population.
These results strongly suggest that (as Laney et al. [in press] put it) human
emotional memory is not like a blockbuster movie with great special eects; it is
instead like a docudrama with complex characters and an emotionally engaging plot. Thus, most laboratory studies of emotional remembering are considering a form of emotion that is neither typical nor representative, and that of course
invites the next question. Can we explore in the laboratory how participants
remember thematically arousing events?
In two unpublished studies, Laney et al. pursued this issue, asking how participants remember events that are arousing for thematic, not visual, reasons.
The rst of these studies evoked arousal by appealing to an issue clearly emotional for the college-age participants, namely, date rape. As in previous research,
the stimuli involved a slide sequence plus tape-recorded narration. The sequence
showed a man and woman on their rst date, and the neutral and arousal sequences were visually identical except for one slide, late in the series. Hence, the

15

arousal manipulation was not contained within the slides themselves, but within
the narration, ensuring that the arousal was not induced by a specic visual
target. In the neutral version, subjects heard that the woman was relaxed and
happy about the date, and the man was polite and friendly. In the arousal version, subjects heard that the woman was growing increasingly apprehensive
as the date progressed, and these fears turn out to be well founded, as the man
attacks the woman late in the sequence and has to be forcibly pushed away.
The memory data in this study showed a robust eect of story, with participants having more complete and more accurate memories for the arousal story.
This conrms the positive eect of emotion on memory for central materials, even
with story materials that are thematically, not visually, arousing. On this point,
Laney et al.s study (unpublished) conrms the generality of prior ndings. However (and crucially), there was no hint of an impairment in memory for peripheral aspects of the emotional story; instead, these were remembered better than
peripheral aspects of the neutral story. Said dierently, emotion seemed in this
case to improve memory for all aspects of the story, and thus there was no indication at all, in these data, of memory narrowing.
Another study conrmed these ndings. In this case, participants viewed a series of 33 photographs conveying a story about a college student named Megan.
In the arousal version, subjects learned early on that Megan is doing badly in her
classes and may lose the nancial support shes been getting from her parents. In
addition, Megans boyfriend just dumped heron her birthday. As the story unfolds, Megan discusses the possibility of suicide and gets quite specic about it,
contemplating the combination of pills and alcohol that is, in fact, one of the most
common paths to suicide among college students. We emphasize, though, that all
of this upsetting information was conveyed in the narrative that accompanied
these slides; there was nothing in the visuals that was at all upsetting. Indeed,
neutral subjects saw the exact same visuals and heard a story with it that paralleled the arousal story, but with some essential dierences. They heard that Megan
was doing well in her classes and that she and her boyfriend were getting along
ne. They also heard about Megan reaching for a pill bottle, but, this time, in response to a hangover caused by her birthday celebration.
Again, the memory data showed a strong eect of story, with better memory
overall for the arousal story. But, as in the previous study, there was no indication in the data of memory narrowing; the positive eect that emotion had on
memory was as reliable for peripheral details as it was for central materials.
These two studies obviously suggest that memory narrowing is not inevitably produced by emotionality, in clear contrast to Easterbrooks suggestion, many
years ago, that it is arousal per se that leads to narrowed attention. Instead, these
studies suggest that we need separate hypotheses to account for emotions positive eects on memory for central materials and its negative eects on memory
for peripheral materials. The former is surely an eect of emotion and the arousal
that accompanies it; this is suggested by the beta-blocker results and other nd-

16

ings. But the latter eect seems not to be produced by emotion. Instead, it is
the result of an event containing a powerful attention magnet that summons
attention and, correspondingly, draws attention away from other aspects of
the event.
We hasten to say, however, that these claims about thematic arousal must
be tentative, and this is a point on which further data are surely needed. A number of studies have documented memory narrowing for emotional events, but
few studies have specically examined thematic arousal, and, until the data base
is larger, we urge caution on this theme. Indeed, one of our own early studies
indicated that thematically arousing events do produce memory narrowing
(Heuer, Reisberg, & Rios, 1997). We now believe that this early result, with its
relatively crude stimuli and test materials, should be set aside in favor of more
recent data collected with improved procedures. Even so, we must not pretend
the data are univocal on the memory eects of thematic arousal, and replication studies are plainly required.
Let us for the moment, however, assume that Laney et al.s (unpublished)
third and fourth studies can be taken at face value. Concretely, this is a claim
that emotion improves memory for all aspects of a story, not just the center, and
that the negative memory eects observed in prior studies are not a consequence
of emotion but are, instead, produced by powerful attention magnets embedded
within the emotional story. Does this mean investigators interested in emotion
should henceforth ignore the memory-narrowing pattern, realizing that it is not
a pattern produced by emotion?
We believe that this question should get a rm no for an answer, for three
reasons. First, it simply is the case that many emotional events do have a visual
focus, do have an attention magnet embedded within the event. Laney et al.s
data tell us that such events are atypical, far outnumbered by thematically arousing events, but even so at least some emotional events do contain visible wounds,
horrifying scenes, threatening weapons, and the like. If we are to understand how
these events are remembered, we must consider the role of these salient visual
targets in shaping memory.
Second, even if the narrowing pattern is not produced directly by emotion, it
may be potentiated by emotion. After all, what makes a visual stimulus a powerful magnet for our attention? Arguably, the key lies (at least in part) in the emotional importance of that stimuluswhether as a source of threat (e.g., a
weapon), a focus of horror (e.g., a gaping wound), or a source of joy (e.g., the
sight of a long-absent friends face). Thus, as emotion grows, the visual salience
of these stimuli grows (cf. hman, Flykt, & Esteves, 2001), and it is then the visual salience that produces memory narrowing.
Evidence consistent with this conjecture comes from a study to which we alluded earlier, by Peters (1988), examining adults memory for an occasion on
which they received a rubella injection. The data showed a weapon focus eect,
with the sight of the hypodermic needle apparently seizing the adults attention,

17

leading to an inability, later on, to identify the nurse who had given the injection. Importantly, participants ability to make this identication was negatively
correlated with arousal (measured by heart rate), suggesting that arousal served
to magnify the threatening quality of the weapon and thus to increase the
weapons power to seize attention, thus undermining memory for other aspects
of the event.
Third, and more immediately, this separation of mechanisms may allow us
to untangle some of the apparent contradictions in the empirical literature. We
noted earlier that two studies, by Libkuman et al. (1999) and Wessel et al.
(2000), have failed to obtain the standard memory-narrowing eect, despite
their use of stimuli and procedures that seemed fully appropriate. If we assume
that their stimuli were arousing, and if we assume that memory narrowing is
(as Easterbrook suggested) a direct consequence of arousal, then it is odd indeed
that these studies did not replicate the narrowing pattern.
However, the framework we are developing here oers a path toward explaining this nonreplication. Memory narrowing, we are proposing, is not a direct
consequence of arousal; instead, it is a consequence of how witnesses direct their
attention during an emotional event. This allocation of attention is to some extent under strategic control; it is likely to be inuenced by instructions and taskset; it is likely to dier somewhat from one subject population to the next. For
these reasons, our current hypothesis (unlike Easterbrooks claim) leads us to
expect some degree of unevenness in the data pattern, with some studies demonstrating the narrowing pattern and some not. To be sure, this line of argument
contains a substantial promissory note, because we have provided no details
about any of these factors guiding attention and so no rm predictions about
when narrowing will or will not be observed. Even so, an account cast in terms
of attention oers exibility that an account cast in terms of arousal does not,
and this by itself seems noteworthy, given the mixed pattern of evidence.

Emotion Intrinsic to the Event,

Emotion Extraneous to the Event

We have now argued that the source of emotion may matter in determining how
an emotional event is remembered. Does the arousal arise from the themes and
meaning inherent in the emotional event or from some salient visual stimulus
within the scene, such as the sight of a wound or a weapon? A related question
also concerns the source of the emotion. Sometimes we are emotional during an
event because the event itself is emotionaluplifting, perhaps, or frightening,
or anger-provoking. But sometimes we are emotional for reasons external to the
event we are participating in. We might still be emotionally aroused from some
earlier event, now done. Or we might be aroused for some reason other than
emotionexercise, perhaps, or an overlarge dose of caeine, or perhaps even an
experimenters injection. Does this contrast matter? Are the memory eects of

18

emotion intrinsic to an event the same as the eects of emotion that merely accompanies the event?
This issue is of interest for several reasons, including a question of how we
should conceptualize emotions impact on memory. Many accounts describe
this impact in terms of arousal mechanismsincreased norepinephrine levels in the bloodstream, for example, and increased serum glucose levels, and
their eect on brain mechanisms serving memory (cf. chapters 2 and 3). If these
are the keys in producing emotions memory eects, then the nature of the
arousal (intrinsic to the event or extraneous) should not matter. But if, on the
other hand, the content and meaning of an emotional event inuence memory
(as plausibly they would), then the contrast between intrinsic and extraneous
arousal may be crucial.
Over the last few years, a number of studies have pursued this issue. Christianson and Mjrndal (1985; also see Christianson, Nilsson, Mjrndal, Perris, &
Tjellden, 1986) asked how injections of adrenaline inuenced memory for pictures.
They reported no eects of the injection on memory for neutral pictures and
a dierence between how participants remembered emotional pictures (after a
saline injection) and how they remembered neutral pictures after the injection of
a stimulant. This result suggests that the source of the arousal does matter and
that only arousal intrinsic to the to-be-remembered materials has memory eects.
(For a related, but somewhat more complex study, see Clark, Milberg, & Ross,
1983.)
In contrast, though, we know that ingestion of glucose does improve memory
in some circumstances. Hall, Gonder-Frederick, Chewning, Silveira, and Gold
(1989) and Manning, Hall and Gold (1990) reported that ingestion of glucose
(dissolved in a lemon-avored drink) improved performance on a number of
memory measures, including recall of an earlier heard narrative. In this case,
mimicking one of arousals eects in a fashion entirely external to the to-beremembered material did improve memory.
It is not obvious what to make of these mixed results, although we suspect
that part of the problem lies with the arousal construct itself. There has been
considerable dispute over whether arousal can be understood as a single, unied construct or whether, instead, we need to distinguish various types and
multiple dimensions of arousal (for some classic statements on this issue, see
Anderson, 1990; Neiss, 1990). With this point unsettled, it is dicult to make
any comparisons across procedures.
In addition, and perhaps more important for our purposes, what about memory
for events, rather than memory for word lists, picture series, or verbal narrative?
To address this question, Libkuman et al. (1999) showed half of their participants an arousing story and half a neutral story (using Heuer and Reisbergs
doctor/mechanic stimulus, already described). Within each group, half of the
participants viewed the slides after a minute of sitting quietly, and half viewed
the slides after spending a minute energetically running in place, to produce a

19

baseline of elevated arousal. Their results showed little or no eect of exerciseproduced arousal on memory. A second study replicated this nding, using a
cued recall measure in place of a recognition test. A third study yielded similar
data but this time with the exercise maintained during the slide presentation.
(Participants pedaled on an exercise bicycle while viewing the slides.)
Where does all of this leave us? We know that arousal does play an important
role in mediating emotions memory eects. This is evident, for example, in Cahill
et al.s (1994, 2000) studies, showing that an interruption of arousal (via betablockers) disrupts emotions impact on memory. But, as we have now seen, at least
some evidence suggests that this arousal must come from the right source and
must somehow be tied to the to-be-remembered event; otherwise, the arousal
seems to have no eect on memory. How should we put these clues together? One
obvious conjecture is that physiological arousal may be necessary for promoting
event memory but is not by itself sucient for producing this memory eect. Also
necessary is an emotional content to the event, something that engages participants attention, processing, and subsequent rehearsal. In essence, this cognitive
work could provide the specic steps needed to establish an event memory and to
link this memory to other bits of knowledge (making the target memory accessible
after on). The physiological arousal could serve to facilitate these steps or to consolidate the memory once it is established through these steps.
This position is fully compatible with the claim, already in the literature (e.g.,
McGaugh, 2000), that physiological arousal has its eect on memory chiey
through its impact on postevent memory-consolidation processes. In this view,
arousal might play little role in establishing the content of emotional memories;
that would be determined by other factors (such as the nature and meaning of
the emotional episode itself). Arousal would instead play its role by cementing
the content (whatever it is) in place, so that the resulting memory would be more
complete and more long-lasting. From this perspective, neither the cognitive
work evoked by an emotional memory nor the arousal by itself would be sucient to produce emotions full memory eect; instead, both acting in concert
would produce the eects we have been describing.

Does the Valence of

the Emotion Matter?

We have now argued that the source of emotional arousal matters in shaping
emotions memory eects. Thematically induced emotion seems to produce dierent memory eects than visually induced, and emotional arousal intrinsic to the
to-be-remembered event seems in several studies to produce dierent eects than
extraneous arousal. But, of course, emotionality also varies in other ways, including the emotions valence. We saw earlier in the chapter that valence seems
largely irrelevant to the relationship between emotion and memory vividness, so
that sad memories are recalled just as vividly as happy memories, and traumatic

20

memories are no more vivid than memories for intensely pleasant events. But
what about memory accuracy or completeness?
Relatively few studies have pursued this issue. We do know that humor promotes memoryboth for verbally presented jokes and for cartoons (e.g., Schmidt
& Williams, 2001). But what about more complex materials? There is some suggestion in the literature that emotions memory eects are mediated by the release of stress hormones (Gold, 1989, 1992; McGaugh et al., 1993); one might
draw from this the claim that only aversive, stressful events will show the
memory benets of these hormones. But, of course, there are many similarities
between the biological eects of emotionally positive experiences and those of
aversive experiences (e.g., Hamann & Ely, 1999), so, on this view, it may be the
arousal itself that promotes memory, independent of the emotions valence.
A recent study from our laboratory (Moyer, 2002) was designed to explore
these issues, albeit using an unusual stimulus: the form of animation known as
anime. This choice of stimulus was motivated by the fact that it is often dicult
to evoke an emotional response in the laboratory, especially with a relatively brief
stimulus designed to experimental specications. It is helpful, therefore, to rely
on a genre and a medium that is already engaging to students, and anime, a style
of Japanese animation immensely popular with many college students, oered
these advantages. We created three stimuli, each an edited version of commercially available animes, one depicting an emotionally neutral story, one depicting an emotionally negative story, and one a positive story. The emotionality of
the stimuli was conrmed by pilot testing, as was the comparability of the stimuli
on several other dimensions (quality, complexity, duration, number of scene
changes). The positive event was drawn from the Magic Users Club (Takahashi
& Asari, 1996) and begins with four student members of the club, discussing ying on their brooms. One of the students attempts to launch her broom, and her
eorts are both comical and erotic. Later, the anime shows the friends ying
along, and one of the students zooms past them, unable to control her broom.
She is forced to let go of the broom, and her rescue (by another of the students)
is again both funny and sexy.
Participants reliably rated this video as making them amused and, perhaps
more impressive, many laughed out loud while viewing the video. How did they
remember the video? Questions in the memory test had been classied as central
or peripheral by a panel of judges, using criteria derived from the Burke et al. (1992)
data. Items were central if they were either relevant to how the story unfolded or
visually prominent within the anime (e.g., tied to a plot-relevant character or
object). Items were peripheral if they were both irrelevant to the story and distant,
in the judges assessment, from the attention centers of the video.
In the recognition test (Moyer, 2002), the participants remembered 65% of
the central details in the comical anime, compared to 51% in the emotionally
neutral stimulus. Thus, once again we see the broad advantage for emotional
materials relative to neutral ones. In addition, participants remembered 34%

21

of the details considered peripheral by the judges, compared to 46% for the
neutral stimulus. These data therefore replicate the memory-narrowing pattern
(with improved memory for central materials, impaired memory for peripheral
materials, in the emotional condition) but show this eect for the rst time, as
far as we know, with an emotionally positive stimulus.
Moyers (2002) study also included a negatively tinged anime, drawn from
Nosaka, Takahata, and Satos (1988) Graveyard of the Fireies. The clip begins
with a view of American ghter planes and a voice yelling air raid. The scene
then shifts to a child (about 15 years old) and his sister, leaving their house for
the air raid shelter. The anime shows them running among falling bombs and
debris and eventually shows them at a school that has been converted to a hospital. A neighbor takes the young boy to see his mother, who is covered in bandages and obviously dying.
Although we did not intend this, the negative anime lacks any specic visual
target that can be construed as a focus for viewers attention or as an attention magnet (to use the term we introduced earlier). Instead, the aect, demonstrated clearly in pilot testing, involves an overarching feeling of sadness over
these two children trying to survive and ultimately losing their mother. Therefore, inadvertently, this stimulus is akin to the thematically arousing stimuli
studied by Laney et al. (in press), and this is reected in the data: Once again,
the emotion improved memory, with 65% of the central details remembered
for this negative stimulus, in comparison to 51% for the neutral stimulus.
However, in keeping with Laney et al.s results, this (thematically arousing)
anime showed no evidence of memory narrowing, that is, no evidence of an
emotion-produced impairment for peripheral materials. Concretely, participants remembered 48% of the peripheral details of this anime, in comparison
to 46% for the neutral stimulus.
This study provides an initial hint, therefore, that emotionally positive events
are remembered in much the same way that emotionally negative events are.
Positive emotion conveys a broad memory advantage for the storys gist and
central materials, but positive emotion can also be accompanied by its own version of memory narrowing, its own version of weapon focus.
A recently published study, however, seems at rst pass to be inconsistent with
Moyers (2002) data. Berntsen (2002) asked participants to think of the most
traumatic experience they had ever experienced, and the happiest experience
they had ever experienced, and then to record as many details of these events as
they could. These details were then categorized as either central or peripheral.
(In one experiment, external judges did this categorization; in another, the participants themselves did this.) Berntsen concludes from this study that tunnel
memories (a pattern of memory narrowing) are due to the combined eects of
high arousal and negative valence, whereas high positive arousal . . . is not able
to create the eect (p. 1018, emphasis added). This seems to contradict Moyers
nding, which did show memory narrowing with positively valenced events.

22

However, this data contrast is dicult to interpret. As one concern, Berntsens

study (2002) does not assess memory accuracy, because there was no way to
conrm (or disconrm) her respondents recollections. Second, with participants able to select their own events for recollection, there is no way to know
if the positive and negative events in Bernstens study were truly comparable.
In particular, it is hard to know whether visually evoked and thematically
evoked events were equivalently represented in the positive and negative recalls, and, as we have discussed, this may well have shaped the pattern of
memory narrowing.
Third, and most important, Berntsens study (2002) does not include data for
neutral events, and this makes her results, as valuable as they are for her own
purposes, impossible to interpret for our purposes. Concretely, what her data
actually show is that the proportion of central details (relative to peripheral details) is greater for shocking memories than it is for positive memories. However, this leaves open the possibility that the positive memories may nonetheless
have contained a greater proportion of central details (and, more to the point, a
smaller number of correctly recalled peripheral details) than a set of neutral
memories would have. With no way to evaluate this possibility (and with the
other concerns just mentioned), there is no way to know from these data whether
the positive events produced a pattern of memory narrowing or not.
If we hold Bernstens study (2002) aside, then, Moyers (2002) data suggest
that it is emotionality per se that matters for memory, not the emotional valence.
A dierent line of data, however, raises questions about this proposal and suggests overall that we may need a more ne-grained account of how dierent
qualities of emotion inuence memory.
Levine and Burgess (1997) have argued that the valence of emotion does
matter for memory; they argue that the specic emotion will inuence not just
how much within an event is remembered but also what within the event is
remembered. In their view, emotions enhance memory for information that
is functionally relevant to the emotional state, and what information that
is will vary from one emotion to another. Thus, a broad categorization of an
episodes elements as central and peripheral may be too crude, because information central for one emotional state may be less so for another state.
To explore this claim, Levine and Burgess (1997) manipulated their participants initial mood by telling them they had received either an A or a D on
a surprise quiz. Participants then took part in what they believed to be an unrelated study during which they heard a narrative that they subsequently had
to recall. The results from this memory test were complex, with the content of
participants recall inuenced (as predicted) by the participants exact emotional
state. Thus, for example, participants who were angry tended to have better
recall for information that concerned goals of the individuals described in the
narrative; there was no relation between degree of anger and recall of other types
of information. Sadness, on the other hand, was associated with enhanced recall

23

of information about outcomes for individuals described in the narrative but not
with other aspects of recall.
In our view, Levine and Burgesss (1997) argument merits close consideration. It is not yet clear whether the details of their proposal (with regard to
goals and outcomes and so on) are warranted by data, and it is also worrisome from our point of view that the arousal at stake in their experiment was
produced by a source external to the to-be-remembered event. (That is, it was
the report of a grade that produced the arousal, but this grade had nothing to do
with the narrative the participants subsequently had to remember.) Nonetheless, Levine and Burgesss more general proposal seems plausible and is surely
consistent with themes we have developed in earlier sections of this chapter. For
example, we have suggested that neither arousal nor emotion is enough itself
to produce the pattern of memory narrowing; instead, this narrowing will be
observed only if the emotional event contains a salient stimulus, drawing participants attention. Moreover, we have suggested that in many cases a stimulus will gain its salience, that is, will gain its power to draw attention, from its
meaning within the emotional event. Let us now add that this emotional meaning may well depend on the particular emotion being experienced, so that a
stimulus salient for someone who is afraid may well be less salient for someone
who is angry. With this additional step, the position we have sketched blends
smoothly into that presented by Levine and Burgess.
With this said, however, the fact remains that the available data are rather
sparse for exploring the memory eects of dierent qualities of emotion. In our
view, this provides powerful reason for investigators to broaden the focus of their
work, and, until that is done, any claims regarding these issues need to be
couched with caution. In the meantime, though, we do have at least some indications that the memory pattern for emotionally positive materials will resemble
the pattern for emotionally negative materials, with an advantage overall (primarily deriving from information at the events center) but also with a disadvantage for peripheral materials if the to-be-remembered event contains a suitable
visual focus. Even so, we are mindful of Levine and Burgesss (1997) potentially
crucial point that dierent emotions will favor memory for dierent aspects of
the target event, and this is a proposal that cries out for further research.

Does the Intensity of

the Emotion Matter?

We have now considered dierence sources of emotion and dierent qualities

of emotion; what about dierent intensities of emotion? For years, the answer
to this question has been cast in terms of the Yerkes and Dodson (1908)
inverted-U curve, with the proposal that there is some optimal level of emotion for promoting memory and that poorer and poorer memory is observed
as one moves further away (in either direction) from this optimum. This pro-

24

posal seems intuitive enough and nds support in the fact that many studies
have shown a positive relationship between memory and level of arousal, while
other studies have shown a negative relationship. (Studies in the latter group
are common, for example, in the research literature on stress and are also reported in the literature on trauma; for discussion of both, see chapter 3, this
volume.) Plausibly, studies in the rst group are on the uphill side of the YerkesDodson function, so that increasing the level of arousal moves us toward the
optimum; studies in the second group could then be understood as on the
downhill side of the function, so that increasing the level of arousal moves us
away from this optimum.
However, it may well be time to retire the venerable Yerkes-Dodson function,
at least as applied to emotional memory. Other chapters in this volume will pursue the issue of how people remember extremely emotional events, but, for now,
we will simply note that these events should not be remembered on most construals of the Yerkes-Dodson function but that they often are rememberedfor
quite some time and in considerable detail (e.g., chapters 2, 4, and 11). To be
sure, there are some complications here (see chapters 3 and 8), but in any case
the data cannot be read as indicating the Yerkes-Dodson, no-memory-for-extreme-emotion pattern.
In addition, and more directly, Christianson (1992) argued persuasively that
there is very little direct evidence in favor of the Yerkes-Dodson pattern in emotional memory. Indeed, the Yerkes-Dodson claim may actually be untestable,
since it can explain virtually any result. (Note, for example, how readily this claim
accommodates both positive eects of emotion on memory and their opposite.)
With what should we replace the Yerkes-Dodson function? At the least, we
should acknowledge that multiple mechanisms are likely to contribute to emotions memory eects, and it is plausible that each will have its own optimal
operating circumstances (cf. chapter 3). If so, the function linking arousal level
and memory may not be unimodal, as the Yerkes-Dodson function is, but more
complex. Consistent with this suggestion, consider a result reported by White
(1991). Using an animal model, White examined the rate of learning as a function of the sugar dose delivered to the animal via a posttraining injection. His
data do not reveal anything like the inverted-U function suggested by Yerkes and
Dodson. Instead, his data plots have two peaks, with the clear suggestion that
multiple mechanisms are in play, each contributing to the overall performance
but with its own optimal level.
This result implies that the relation between arousal and memory will be more
complex than Yerkes and Dodson envisioned but still a relation that can be understood in terms of a single predictor variable. Figure 1.2 oers an appreciably more
complex model, initially oered as a proposal for describing sports performance
by Fazey and Hardy (1988) and then applied to memory by Deenbacher (1994).
This model implies that the relation between arousal and performance may depend
on other variables, such as the persons level of anxiety. At low anxiety levels, Fazey

25

Figure 1.2. The Yerkes-Dodson inverted-U function is often offered as a description

of how memory performance will change as a function of increasing arousal. This
figure, suggested by Fazey and Hardy (1988), provides a more complex and perhaps
more realistic model. Whether the details of this model turn out to be correct or not,
Fazey and Hardy remind us that the relationship between memory and arousal is
likely to interact with other variables. They suggest that a key third is cognitive
anxiety.

and Hardy forecast a pattern similar to the Yerkes-Dodson inverted-U. As anxiety increases, however, the pattern changes, with a sharp discontinuity in the
curve and the level of performance depending on whether arousal is gradually
being increased or decreased.
Not enough data are available to evaluate this model (or variants on it), but
it does remind us that the optimal arousal level for memory (if one can be dened) is very likely to depend on other factors, much as gure 1.2 suggests. If so,
the Yerkes-Dodson function is not just unproven and probably untestable, it may
be on the wrong track altogether, conceptualizing the data pattern in a (unidimensional) fashion that is far too simple. Once again, though, we call attention to the fact that this theme is plainly underresearched.

Lost Memories?
Before moving on, we need to address one further topic clearly related to the
themes of the last two sections. What happens when the emotion associated with
an event is extremely intense and also strongly negative in its valence, perhaps
with the valence marked by anxiety, perhaps by a sense of betrayal? Many authors have suggested that, under these circumstances, the relation between
emotion and memory changes dramatically. No longer does emotion promote

26

memory, but, instead, in these circumstances, emotion will be associated with

amnesia for the painful events.
This is not the place to delve deeply into this dicult and contentious issue,
but we do wish to touch a few points related to themes elsewhere in this chapter. (For more on this broad issue, see chapters 3, 4, and 11.) In doing so, we
believe it important to separate the empirical claims from the interpretations
often oered for the data; we begin, therefore, by considering the facts and (as
we view it) the considerable ambiguity attached to those facts.
Can one experience a highly emotional event but not remember it? The answer is yes, for several reasons. First, some emotional events discussed in the
literature have been experienced by very young children (e.g., Williams, 1995)
and so are open to the widespread pattern of childhood amnesia (cf. Shobe &
Kihlstrom, 1997), a pattern that applies to all events in the rst years of life,
emotional or not. Second, emotional memories, as strong as they tend to be,
are still vulnerable to the same sorts of forgetting mechanisms as other memories, including interference from other remembered events and retrieval failure. This, too, can lead to apparent amnesia for emotional events, and there is
at least some indication that the rate of forgetting is similar for emotional and
nonemotional memories (Read & Lindsay, 2000). Third, there is reason to believe that extreme levels of emotion can, in some circumstances, have a catastrophic eect on memory, causing relatively complete amnesia for horric
events that happened just hours or days earlier (Arrigo & Pezdek, 1997). It
seems likely in these cases that the amnesia is caused by the bodily changes
associated with extreme stress, changes that could disrupt the biological processes of memory consolidation needed to establish a memory in the rst place.
On this view, the high levels of stress have washed away the seeds of memory,
so to speak, before they had a chance to take root.
What about cases in which people report that they have long been amnesic
about an emotional event but have now recovered the relevant memories? There
is no doubt that these reports of memory discovery do occur, and here too there
are many mechanisms that might contribute to such reports. First, we believe
that some of these cases involve memories that have never been truly lost; the
recovery merely refers to a new-found willingness to talk about this always
remembered but never discussed episode. This willingness might emerge in the
safety of a therapy setting; it might be encouraged by a social climate in which
reports of previous traumas are supported and perhaps even encouraged. Second,
we believe that some of these cases involve recategorizations of the relevant
memory. I always knew that he did X, but until recently I didnt realize there
was something blameworthy in what he did, and its only now, when I realize
that the action is culpable, that Im talking about it. Third, there are some cases
in which, remarkably, someone is convinced he was amnesic about an event at
some prior time but, in truth, did remember the event all along. In other words,
this is an example of a memory error, but what is being misremembered is the

27

fact of amnesia itself! Several such cases have now been documented, cases that
are, in a sense, the inverse of someone falsely claiming, I knew it all along.
Hence, these cases are sometimes referred to as revealing the I forgot it all along
eect (Schooler, Bendriksen, & Amadar, 1997; also Padilla-Walker & Poole,
2002). Fourth, we have noted that at least some emotional memories will be
forgotten for routine reasons such as retrieval failure; these cases of forgetting
are likely to be reversed if suitable retrieval cues should present themselves. Fifth,
and nally, it does seem sadly plausible that at least some recovered memories
may turn out to be false memories, recalling events that never took place or
events that unfolded dierently from the way they are remembered.
In light of all these comments, we believe it is unsurprising that some emotional memories do seem genuinely to be forgotten, and also unsurprising that
there are many reports of emotional memories allegedly forgotten and then rediscovered. Moreover, we have suggested that many dierent causal sequences
could lead to these facts, and, notably, in most of these causal sequences, there
is no reason at all to challenge the veracity of the allegedly recovered memory.
Why, then, have recovered memories been so controversial? The debate, we
believe, does not revolve around the facts themselves; those, we have just suggested, are relatively uncontentious but also open to multiple interpretations.
Instead, the debate hinges on a particular claim about the facts, namely, that
emotional memories are lost because of a special mechanism of repression or
dissociation, a mechanism that involves mental processes markedly dierent
from those involved in ordinary emotional remembering. Our view is that the
facts do not warrant these latter claims, because, as just discussed, the available
facts are easily accommodated with no need for new and distinctive mechanisms.
We stress, though, that this is in no way intended as a statement of blanket skepticism about these lost or recovered memories. Instead, it is simply an assertion
that there are multiple non-exotic steps through which emotional memories
might well be apparently lost and then rediscovered.

Alternative Approaches
Beyond Accuracy:

Memory Detailedness and Coherence

We began this chapter by asking how accurately emotional events are remembered. The answer, it seems, is it depends. Overall, many studies show that
emotion promotes memory for an events center and that at least part of this
promotion depends on bodily arousal. But many studies also show that emotion
impedes accurate memory if the to-be-remembered event contains a salient visual
target, summoning attention; in this case, memory for the events center will be
enhanced but at the cost of poorer memory for the events periphery. In addi-

28

tion, we have at least some indications that this pattern can be observed for
emotionally positive events (e.g., an event that is comical and somewhat erotic)
as well as emotionally negative events. Beyond this, though, at least two complications are needed in this account. First, the pattern of what specic content
is remembered and what is not may depend on the nature of the emotion (after
Levine & Burgess, 1997), and this issue needs further exploration. Second, the
relation between memory and emotional intensity remains to be fully explicated,
and it does seem likely that there will be circumstances in which emotion (perhaps only extreme emotion) has none of the eects we are describing but, instead,
undercuts memory (see, for example, chapter 8).
It should be acknowledged, though, that this account views emotional
memory from a distinct perspectiveone highlighting a memorys accuracy and
completeness as its most important attributes. To be sure, this concern with
memory accuracy is easy to justify, especially in light of psychologys increasing concern over the accuracy of eyewitness memory in the courts. But other
perspectives on emotional memory should also be explored and have their own
value. As just one example, consider a study by Williams et al. (1996). Their
study begins with the fact that hopelessness is a key factor in leading a depressed
person to contemplate suicide, and so Williams et al. ask: Where does hopelessness come from? They explore the proposal that one becomes hopeless when one
cannot imagine future events in any detail, and this impairment in imagination,
in turn, becomes likely when one does not recall past events in detail. Their study
conrms these claims, with both a correlational design (using hospitalized patients who had, in fact, recently attempted suicide) and an experimental design
(to explore causal relations).
The Williams et al. article is interesting for several reasons, including the fact
that memory accuracy plays no role in their account. Instead, what matters is
memory detailedness and, arguably, memory vividness, and, more, these attributes matter for functions of considerable interestthe use of memory in problem solving; the use of memory in gauging the likelihood of future events; and,
prominently, the use of memory in generating hope for the future. Thus, this
article provides a compelling reminder that these other dimensions of memory
are well worth our scrutiny.
What do we know about emotions inuence on these other dimensions? As
we mentioned at the very outset, memory vividness is strongly correlated with
how emotional an event is recalled to have been (Reisberg et al., 1988), but there
are two concerns about this result (and related ndings in the literature). First,
note that the predictor variable here is retrospective assessment of an events
emotionality, and, as we will see in a moment, we may not be able to take this
retrospection at face value. Hence, we cannot be certain that there is a relation
between how emotional an event was at the time of its occurrence and subsequent memory vividness. Second, the Reisberg et al. data (and, again, more recent, related ndings) are in any case correlational, leaving questions about

29

causal relationships. Indeed, Reisberg et al. also reported a strong correlation

between memory vividness and how consequential the event was judged to have
been (and also a correlation between rated emotionality and rated consequentiality), and this obviously clouds causal interpretation.
If we can set these cautions aside, though, it does seem that emotional events
will be recalled more vividly than other events and, hence, following Williams
et al. (1996), will be more inuential and more compelling in gauging the future than other events. In addition, the content of emotional memories (whether
accurate or not) may be somewhat dierent from that for neutral memories, and
this, too, may matter for the function of these memories in problem solving, predicting the future, and the like. This dierence in content is suggested by a result reported by Heuer and Reisberg (1990). They found that the number of
intrusion errors made by participants was the same for emotional and neutral
events, but the type of errors was dierent. Participants who had viewed the
neutral stimulus (about the boy visiting his mechanic father) tended to recall plot
elements that had, in fact, not been mentioned. Participants who had viewed
the emotional stimulus (with the boy visiting his surgeon father) tended to recall psychological elementsabout how upset the boy was, for example, or how
angry the mother was that her son had seen such gruesome sights.
All of this leads us, then, to a by-now familiar message. We know a lot about
memory for emotional events, but it is not dicult to nd gaps in our knowledge, including our knowledge of the relationship between the emotionality
of an event and how detailed, vivid, compelling, or coherent the memory for
the event will be later on. In this setting, it is interesting to set psychologists
inquiries into memory (and autobiographical memory in particular) side by
side with the literary study of the genre of autobiography. In the study of this
genre, there is ample discussion of how a recollection shapes someones values or character and discussion of how the qualities of a remembered event can
compel someone to action, or fail to. Notably absent from this study, though,
is much concern about the historical correctness of these memories; it is not
the correctness that makes the memories formative or compelling. Perhaps
psychologists would be well-advised to learn from this traditioncertainly not
abandoning our research on memory accuracy but supplementing it with
a rich and (perhaps) equally important set of questions about how emotion
shapes memories in a way that can, in turn, mold who we are, how we will
act, and what we will believe in the future. (For an important psychological
foray into these issues, see Pillemer, 1998.)

Remembering the Emotion Itself

Our concern about alternative approaches to emotional memory also leads us
to another issue. Not only do we remember the content of an emotional event,
we also remember our assessment of, and reactions to, the event. In particular,

30

we remember how emotional the event was or was not for us, and this memory,
in turn, can also be more or less accurate.
For many people, introspection suggests that recall of prior feelings is immediate, long-lasting, and quite compelling. One may not remember much about
a particular speakers lecture, but one remembers that the talk was boring; one
may not recall the plot of a movie, but one remembers that it was quite funny;
and so on. Despite these intuitions, however, evidence suggests that retrospective reports of emotionality are often reconstructions, inuenced heavily by current assessments, and, in some cases, inaccurate.
Of course, some memory for past feelings is rather accurate. For example,
Safer, Bonanno, and Field (2001) surveyed participants whose spouse had died
6 months prior to the study. The participants reported their level of grief at the
time of the survey and, then, roughly 4.5 years later, tried to recall their level of
grief at the time of the survey. Safer et al. report that this recall was impressively
accurate, although participants whose grief diminished relatively little over time
did tend to overestimate their prior grief.
It is important, though, that participants contemporary assessment of prior
emotional events does change, so that the sting of past mishaps gradually fades,
as does the glow of past triumphs. This pattern was documented by Walker et al.
(1997), who asked participants to rate how they currently felt about previously
experienced emotional events; these ratings were collected 3 months, 1 year, and
4.5 years after the target event. Walker et al. report that all the ratings became
less extreme as time went on, and, interestingly, the unpleasantness of past
humiliations and defeats faded more quickly than the pleasantness of past joys
and celebrations.
How do these changes in current assessment of an event inuence the memory
for the event? Some insight is provided by Levines (1997) study of how supporters of Ross Perots presidential campaign (in 1992) recalled their feelings
upon hearing that Perot had withdrawn from the race and also their feelings
upon hearing about Perots subsequent decision to rejoin the race. Levine reports
that her participants showed frequent errors in how they recalled their past emotions but not (as some have suggested) a general tendency to overestimate past
emotion. Instead, she found systematic distortions in emotion recall, bringing
emotions-as-remembered into closer proximity to current appraisals. In short,
it appears that her participants were to some extent reconstructing what their
past emotions were likely to have been and were basing this reconstruction on
their current emotions.
A follow-up study found similar results in students recollection of their emotional reactions when they rst heard that O. J. Simpson had been acquitted of
accusations that he had murdered his wife (Levine, Prohaska, Burgess, Rise, &
Laulhere, 2001; also Levine & Safer, 2002). As time passed after the acquittal,
Levine et al. found systematic changes in how people remembered their own
emotional reactions, with these changes tending quite strongly toward bring-

31

ing past emotions (as remembered) into alignment with current assessments.
Thus, participants who had gradually grown angrier about the acquittal recalled
their initial reaction as being angrier than it was; participants who had grown
less upset recalled their initial reaction accordingly.
A dierent form of reconstruction of emotion, based on other cues, can also
be observed in a shorter time scale. In one laboratory study, participants viewed
photographs of faces while listening to sentences spoken in either a positive or
negative tone of voice (Ochsner, Schacter, & Edwards, 1997). The faces themselves conveyed either positive or negative affect, albeit in a subtle way. In a subsequent memory test, participants were shown the faces once again and asked
to recall the tone of voice in which the pictured person had spoken. The participants tended to recall the voice as having the same aect as the picturea
memory illusion in which remembered-aect is apparently reconstructed from
other information, both perceptually given and drawn from memory.
Other results show similar patterns, with aect associated with one aspect of
a stimulus altering (and in some cases distorting) how one remembers other
aspects of the stimulus. For example, in a study by Hertel and Narvaez (1986),
participants watched videotapes of conversations; memory was then tested (via
a recognition test in one procedure, recall in another) for the specic words
uttered in the conversation. Memory was clearly inuenced by the emotional
valence of the nonverbal gestures and facial expressions visible in the videotape
(with the actual script held constant) in a fashion that suggested that the (nonverbal) aect had inuenced how participants remembered the gist of the conversation. (For a related result, see Nygaard & Lunders, 2002.)
Surely, then, our memory for prior emotion (either experienced or observed)
is based to some extent on reconstruction, and our memory for prior emotion
can, in turn, shape how we recall other aspects of the event. These facts must be
part of our broader account of how emotional events are remembered, but note
that these facts also have methodological implications. We mentioned in the
previous section that the data linking emotion to memory vividness rest on retrospective assessments of emotion; we now see good reason to be cautious about
those assessments. Similarly, we will, in a later section, return to the topic of
ashbulb memories, and one of the issues there will be the role of emotion in
forming such memories; many of the relevant studies, however, have relied on
retrospective assessments of this emotion. For all of these reasons, we would be
well served by an improved understanding of how emotion itself is recalled.
Finally, one other complication should also be mentioned. Many events are
likely not to be uniformly emotional; instead, they will have a mix of intense
moments and less intense ones, moments of strong feelings interwoven with
moments of relative calm. How will this changing prole be reected in memory?
One might think that the subsequent memory will reect some sort of average of
the emotion felt over the course of an event, but several studies suggest that this is
not the case. Instead, people seem to retain just a few snapshots of the overall

32

event, and memory for the whole is dominated by the contents of these few snapshots. One snapshot seems to record the emotional peak of the target event
not surprisingly, since that peak is likely to be salient both in perception and in
memory. Another snapshot seems to record the emotional ending of the event
again, not surprisingly, given what we know about recency eects in memory
(e.g., Reisberg, 2001). These two snapshotsthe peak and the endare then
weighted heavily in subsequent recall of the events emotional meaning, so much
so that the subsequent evaluation can be predicted almost entirely from evaluations of these two moments (Fredrickson & Kahneman, 1993; Kahneman,
2000; Schreiber & Kahneman, 2000). Certainly, though, further work is needed
to explore how one derives a summary evaluation from these snapshots and then
how ones broader recollection of the target event is shaped by this summary
evaluation.

Working Backward

From an Emotional Memory

One last perspective on emotional memories invites our attention, but it is a perspective we have already touched in passing. In thinking about memory for
emotional events, it is tempting to begin with the event itself and ask: If someone
experiences an event that makes him angry, or happy, or afraid, how will he remember the event later on? But we could plausibly reverse this logic and begin
instead with a bit of emotional recall. In this case, we might ask: If someone remembers an emotional event, then what can we conclude from this recollection?
This shift in perspective is important for a simple reason. Some instances of
emotional remembering may be false, recording events that unfolded rather
dierently from the way they are recalled, or perhaps recording events that never
happened at all. In other words, it is true that, if an emotional event occurs, there
will likely be a memory corresponding to that event later on, and this memory
is likely to be detailed and reasonably accurate. (This is a consequence of the
positive eects of emotion on memory, already discussed.) But the converse is
not true. If an emotional event is recalled in vivid detail, it does not follow that
there is likely to have been a prior event corresponding to the memory. Or, to
put this succinctly, if there is an emotional event, then there will probably be an
emotional memory for it; but if there is an emotional memory, there may (or may
not) have been an emotional event.
As illustrations, we have already mentioned that the intensity of emotion in
an event may well be dierent from the intensity as it is recalled (Levine, 1997;
Levine et al., 2001). We have also mentioned the Neisser and Harsch (1992)
data, in which students very emotional recollections of the space-shuttle explosion turned out to be wrong not just in detail but in major elements as well.
Presumably, Neisser and Harschs participants were the victims of their own
errors in reconstructing the earlier episode, and, ironically, emotional recollec-

33

tion may be particularly vulnerable to this sort of error. To see why, imagine an
event that was not considered noteworthy when it occurred, but which was later
deemed consequential. One example might be a rst (and quite unremarkable)
encounter between two people that later grew into an intense romance. A dierent example might be a political event that seemed small at the time, but which
marked the beginning of a gradual slide (visible only later) toward some triumph
or some tragedy. In such cases, there is little reason to have encoded the initial
event with great care, no reason to have paid special attention, no reason to have
rehearsed the event, no arousal to promote memory consolidation. Yet, after the
fact, there is ample reason to want to recall the event, and this will spur extra
eort toward reconstruction. Of course, without much initial encoding, information actually in memory will provide slim support for this reconstruction,
leaving a large potential for schema-based inferences (which may or may not
be correct) and intrusions from other, related, episodes.
To explore these points, one would want a design in which two groups of
individuals are exposed to a minor and emotionally bland episode and then led
to recall the episode as best they can later on. Prior to the recall, however, one
group would be exposed to information persuading them that the episode was,
in fact, emotionally important in some fashion; this would tell us how, and how
fully, perceived emotion inuences memory reconstruction (its vividness, its
completeness, its accuracy). To the best of our knowledge, however, no study of
this sort has been conducted, despite the relevance of such a study to the ferocious debate over (potentially) false memories. The literature does contain hints
on this topicthe dierent types of intrusion errors for emotional and neutral
memories (e.g., Heuer & Reisberg, 1990) or the impact of dierent story outcomes
on how a story is recalled (e.g., Spiro, 1980)but, unmistakably, this is another
topic crying out for empirical pursuit.

Flashbulb Memories Revisited

We began this chapter with a mention of ashbulb memories, and it will be useful
to return to this topic in light of the wide range of issues we have now considered.
We know that these memories are vivid, detailed, and long-lasting, reasonably
accurate in many circumstances (e.g., Conway et al., 1994; McCloskey, Wible, &
Cohen, 1988) but also sometimes error-lled (e.g., Neisser & Harsch, 1992).
What produces this pattern? Why are some ashbulb events remembered
accurately, while others are not? As a number of authors have noted, several
factors are relevant (cf. Conway, 1995; Finkenauer et al., 1998). One is the
timing of the memory teststhat is, when the initial recording of the event
occurs and then when the follow-up testing takes place (e.g., Winningham,
Hyman, & Dinnel, 2000). It may also matter whether, how, and how often the
ashbulb event is discussedfor example, as part of the routine sharing of special memories that people often do (e.g., Philippot & Rime, 1998; Rime, Philippot,

34

Boca, & Mesquita, 1992). Another, and crucial, factor is the consequentiality of
the remembered event, with consequential ashbulb events (e.g., Margaret
Thatchers resignation, for many in England, or the 1989 Santa Clara earthquake, for those living near the epicenter) remembered more accurately than less
consequential events such as Thatchers resignation for those living in the United
States or the earthquake for those living far from the epicenter (Conway et al.,
1994; Neisser, 1996). This is perhaps an unsurprising result. Consequentiality
will motivate close attention and will spur rehearsal; consequentiality is also
likely to be associated with stronger emotion and thus will invite emotions overall positive contribution to remembering (see Finkenauer et al., 1998, for a more
extensive discussion of emotions role in ashbulb formation).
In short, then, ashbulb memorieslike other memoriestend to be accurate but are not always accurate. Flashbulb memories, like other memories, fade
with the passage of time, are promoted by rehearsal, are enhanced if the eventas-experienced is consequential. In these (and other) regards, ashbulb memories seem qualitatively similar to other emotional memories and so, despite early
claims to the contrary, are almost surely not in a class by themselves.

Conclusions
The main points of this chapter are relatively easy to describe. In general, emotion seems to have a positive eect on memory, increasing memory vividness,
accuracy, completeness, and longevity. But emotions eects are not uniformly
positive. Many emotional events contain a prominent visual stimulus, and, if so,
emotion seems to promote a focus on this stimulus in a fashion that impairs
memory for the events periphery. Emotion assigned to an event after the fact
may also spur memory reconstruction based on too little information, and this
may foster reconstructive error. And, nally, extremely intense emotion may
work against memory, perhaps by interrupting the biological processes needed
for memory consolidation.
This summary of the evidence reects considerable progress over the last
50 years, progress that has included the accumulation of a large quantity of data
and has also included questions that have arisen in the literature and then been
resolved to the satisfaction of many investigators (e.g., the special status of ashbulb memories). This progress has also included evidence that directly challenges
some long-held and often-quoted claims, including the Yerkes-Dodson function
and the Easterbrook notion that arousal directly leads to memory narrowing.
At the same time, though, the progress on these various fronts has helped
highlight the gaps in what we know about memory for emotional events. In this
chapter, we have considered the distinction between events that are visually
arousing and those that are thematically arousing, but in truth the data pertinent to this distinction are few. We have also considered how people remember

35

events that are emotionally positive or how the exact content of an emotional
memory might be inuenced by the particular emotion in place during an event.
On these points, too, we know far less than we might wish. Other issues are also
largely untouched, including the nature of intrusion errors in memory for emotional events, emotions inuence on our eorts toward reconstructing a poorly
remembered past event, and emotions role in shaping the sorts of memory
detailedness that seem to matter for problem solving and the generation of hope.
Finally, also in need of elaboration is the exact role that arousal plays in shaping emotional memories. We have considered evidence that arousal may be
necessary but not sucient for producing emotions memory eects, and the
mechanisms behind this pattern need to be specied.
We close, therefore, simultaneously celebrating what we know about emotional remembering and showcasing the work still to be done. We have come a
long way in the last 50 years, and the knowledge we have gained provides a rich
base for now tackling the questions currently before us.

References
Adolphs, R., Cahill, L., Schul, R., & Babinsky, R. (1997). Impaired declarative memory
for emotional material following bilateral amygdala damage in humans. Learning & Memory, 4, 291300.
Anderson, K. J. (1990). Arousal and the inverted-U hypothesis: A critique of Neisss
Reconceptualizing arousal. Psychological Bulletin, 107, 96100.
Arrigo, J. M., & Pezdek, K. (1997). Lessons from the study of psychogenic amnesia.
Current Directions in Psychological Science, 6, 148152.
Berntsen, D. (2002). Tunnel memories for autobiographical events: Central details
are remembered more frequently from shocking than happy experiences. Memory
& Cognition, 30, 10101020.
Bluck, S., & Li, K. Z. H. (2001). Predicting memory completeness and accuracy:
Emotion and exposure in repeated autobiographical recall. Applied Cognitive
Psychology, 15, 145158.
Bohannon, J. N. (1988). Flashbulb memories of the space shuttle disaster: A tale of
two theories. Cognition, 29, 179196.
Bornstein, B. H., Liebel, L. M., & Scarberry, N. C. (1998). Repeated testing in eyewitness memory: A means to improve recall of a negative emotional event.
Applied Cognitive Psychology, 12, 119132.
Bower, G. H. (1992). How might emotions aect learning? In S.-. Christianson
(Ed.), The handbook of emotion and memory: Research and theory (pp. 332).
Hillsdale, NJ: Erlbaum.
Brewer, W. F. (1988). Memory for randomly sampled autobiographical events. In
U. Neisser & E. W. Winograd (Eds.), Remembering reconsidered (pp. 2190). Cambridge: Cambridge University Press.
Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 7399.
Burke, A., Heuer, F., & Reisberg, D. (1992). Remembering emotional events.
Memory & Cognition, 20, 277290.

36

Cahill, L., Prins, B., Weber, M., & McGaugh, J. (1994). -Adrenergic activation and
memory for emotional events. Nature, 371, 702704.
Canli, T., Zhao, Z., Brewer, J., Gabrieli, J. D. E., & Cahill, L. (2000). Event-related
activation in the human amygdala associates with later memory for individual
emotional response. Journal of Neuroscience, 20, RC99.
Christianson, S.-. (1992). Emotional stress and eyewitness memory: A critical review. Psychological Bulletin, 112, 284309.
Christianson, S.-., & Loftus, E. (1991). Remembering emotional events: The fate
of detailed information. Cognition & Emotion, 5, 81108.
Christianson, S.-., & Mjrndal, T. (1985). Adrenalin, emotional arousal, and
memory. Scandinavian Journal of Psychology, 26, 237248.
Christianson, S.-., Nilsson, L.-G., Mjrndal, T., Perris, C., & Tjellden, G. (1986).
Psychological versus physiological determinants of emotional arousal and its
relation to laboratory induced amnesia. Scandinavian Journal of Psychology, 27,
300310.
Clark, M. S., Milberg, S., & Ross, J. (1983). Arousal cues arousal-related material in
memory: Implications for understanding eects of mood on memory. Journal
of Verbal Learning and Verbal Behavior, 11, 671684.
Cliord, B., & Scott, J. (1978). Individual and situational factors in eyewitness testimony. Journal of Applied Psychology, 63, 352359.
Conway, M., Anderson, S., Larsen, S., Donnelly, C., McDaniel, M., McClelland,
A. G. R., et al. (1994). The formation of ashbulb memories. Memory & Cognition, 22, 326343.
Conway, M. A. (1995). Flashbulb memories. Hillsdale, NJ: Erlbaum.
Davidson, D., Luo, Z., & Burden, M. J. (2001). Childrens recall of emotional behaviours, emotional labels, and nonemotional behaviours: Does emotion enhance memory? Cognition & Emotion, 15, 126.
Deenbacher, K. (1994). Eects of arousal on everyday memory. Human Performance, 7, 141161.
Dietrich, D. E., Waller, C., Johannes, S., Wieringa, B. M., Emrich, H. M., & Muente,
T. F. (2001). Dierential eects of emotional content on event-related potentials in word recognition memory. Neuropsychobiology, 43, 96101.
Easterbrook, J. A. (1959). The eect of emotion on cue utilization and the organization of behavior. Psychological Review, 66, 183201.
Fazey, J. A., & Hardy, L. (1988). The inverted-U hypothesis: A catastrophe for sport psychology. British Association for Sports Sciences Monograph No. 1. Leeds,
England: National Coaching Foundation.
Finkenauer, C., Luminet, O., Gisle, L., El-Ahmadi, A., van der Linden, M., & Philippot,
P. (1998). Flashbulb memories and the underlying mechanisms of their formation: Toward an emotional-integrative model. Memory & Cognition, 26, 516531.
Fredrickson, B., & Kahneman, D. (1993). Duration neglect in retrospective evaluations of aective episodes. Journal of Personality and Social Psychology, 65, 4555.
Gold, P. E. (1989). Neurobiological features common to memory modulation by
many treatments. Animal Learning & Behavior, 17, 94100.
Gold, P. E. (1992). A proposed neurological basis for regulating memory storage for
signicant events. In E. Winograd & U. Neisser (Eds.), Aect and accuracy in re-

37

call: Studies of ashbulb memories (pp. 141161). New York: Cambridge University Press.
Hall, J., Gonder-Frederick, L., Chewning, W., Silveira, J., & Gold, P. (1989). Glucose
enhancement of memory in young and aged humans. Neuropsychologia, 27,
11291138.
Hamann, S. B., & Ely, T. D. (1999). Amygdala activity related to enhanced memory
for pleasant and aversive stimuli. Nature Neuroscience, 2, 289293.
Hertel, P. T., & Narvaez, A. (1986). Confusing memories for verbal and nonverbal communication. Journal of Personality and Social Psychology, 50(3), 474
481.
Heuer, F. (1987). Remembering detail: The role of emotion in long-term memory. Unpublished doctoral dissertation, New School for Social Research.
Heuer, F., & Reisberg, D. (1990). Vivid memories of emotional events: The accuracy of remembered minutiae. Memory & Cognition, 18, 496506.
Heuer, F., Reisberg, D., & Rios, C. (1997). The memory eects of thematically induced emotion. In D. G. Payne & F. Conrad (Eds.), Intersections in basic and applied memory research (pp. 113132). Hillsdale, NJ: Erlbaum.
Intraub, H., & Richardson, M. (1989). Wide-angle memories of close-up scenes. Journal of Experimental Psychology: Learning, Memory & Cognition, 15, 179187.
Jones, E. B., OGorman, J. G., & Byrne, B. (1987). Forgetting of word associates as a
function of recall interval. British Journal of Psychology, 78, 7989.
Kahneman, D. (2000). Evaluation by moments: Past and future. In D. Kahneman
& A. Tversky, A. (Eds.), Choices, values and frames (pp. 693708). New York:
Cambridge University Press.
Kramer, T., Buckhout, R., & Eugenio, P. (1990). Weapon focus, arousal and eyewitness memory: Attention must be paid. Law and Human Behavior, 14, 167184.
Laney, C., Heuer, F., & Reisberg, D. (in press). Exploring the contrast between
thematically-induced and visually-induced emotion. Applied Cognitive Psychology.
Levine, L. J. (1997). Reconstructing memory for emotions. Journal of Experimental
Psychology: General, 126, 165177.
Levine, L. J., & Burgess, S. L. (1997). Beyond general arousal: Eects of specic emotions on memory. Social Cognition, 15, 157181.
Levine, L. J., Prohaska, V., Burgess, S. L., Rice, J. A., & Laulhere, T. M. (2001). Remembering past emotions: The role of current appraisals. Cognition & Emotion,
15(4), 393417.
Levine, L. J., & Safer, M. A. (2002). Sources of bias in memory for emotions. Current
Directions in Psychological Science, 11, 169173.
Libkuman, T. M., Nichols-Whitenead, P., Grith, J., & Thomas, R. (1999). Source
of arousal and memory for detail. Memory & Cognition, 27, 166190.
Linton, S. J., & Melin, L. (1982). The accuracy of remembering chronic pain. Pain,
13, 281285.
Loftus, E., & Burns, T. (1982). Mental shock can reproduce retrograde amnesia.
Memory & Cognition, 10, 318323.
Loftus, E., Loftus, G., & Messo, J. (1987). Some facts about weapon focus. Law and
Human Behavior, 11, 5562.

38

Maass, A., & Khnken, G. (1989). Eyewitness identication. Law and Human Behavior, 13, 397408.
Manning, C., Hall, J., & Gold, P. (1990). Glucose eects on memory and other neuropsychological tests in elderly humans. Psychological Science, 1, 307311.
McCloskey, M., Wible, C. G., & Cohen, N. J. (1988). Is there a special ashbulbmemory mechanism? Journal of Experimental Psychology: General, 117, 171
181.
McGaugh, J. L. (2000). MemoryA century of consolidation. Science, 287, 248251.
McGaugh, J. L., Intoini-Collision, I. B., Cahill, L. F., Castellenano, C., Dalmaz,
C., Parent, M. B., et al. (1993). Neuromodulatory systems and memory storage: Role of the amygdala. Behavioral Brain Research, 58, 8190.
Moyer, M. (2002). Polyanna revisited: Do we remember equally the good and the bad?
Unpublished B.A. thesis, Reed College, Portland, Oregon.
Neiss, R. (1990). Ending arousals reign of error: A reply to Anderson. Psychological Bulletin, 107, 101105.
Neisser, U. (1996). Remembering the earthquake: Direct experience vs. hearing the
news. Memory, 4, 337358.
Neisser, U., & Harsch, N. (1992). Phantom ashbulbs: False recollections of hearing the news about Challenger. In E. Winograd & U. Neisser (Eds.), Aect and
accuracy in recall: Studies of ashbulb memories (pp. 931). Cambridge: Cambridge University Press.
Nosaka, A. (Writer), Takahata, I. (Director), & Sato, R. (Producer). (1988). Graveyard of the reies [DVD]. Central Park Media Corporation, USA.
Nygaard, L. & Lunders, E. (2002). Resolution of lexical ambiguity by emotional tone
of voice. Memory & Cognition, 30, 583593.
Ochsner, K. N., Schacter, D. L., & Edwards, K. (1997). Illusory recall of vocal aect.
Memory, 5, 433455.
hman, A., Flykt, A., & Esteves, F. (2001). Emotion drives attention: Detecting the
snake in the grass. Journal of Experimental Psychology: General, 130, 466478.
Padilla-Walker, L. & Poole, D. (2002). Memory for previous recall: A comparison of
free and cued recall. Applied Cognitive Psychology, 16, 515524.
Peters, D. (1988). Eyewitness memory and arousal in a natural setting. In M. Gruneberg, P. Morris, & R. Sykes (Eds.), Practical aspects of memory: Current research
and issues (pp. 8994). New York: John Wiley.
Peterson, C., & Whalen, N. (2001). Five years later: Childrens memory for medical
emergencies. Applied Cognitive Psychology, 15(7), S7S24.
Philippot, P., & Rime, B. (1998). Social and cognitive processing in emotion: A heuristic for psychopathology. In W. F. Flack, Jr. & J. D. Laird (Eds.), Emotions in
psychopathology: Theory and research. Series in aective science (pp. 114129).
New York: Oxford University Press.
Pickel, K. L. (1998). Unusualness and threat as possible causes of weapon focus.
Memory, 6, 277295.
Pillemer, D. B. (1984). Flashbulb memories of the assassination attempt on President Reagan. Cognition, 16, 6380.
Pillemer, D. B. (1998). Momentous events, vivid memories. Cambridge, MA: Harvard
University Press.

39

Pillemer, D., Goldsmith, L. Panter, A., & White, S. (1988). Very long-term memories of the rst year in college. Journal of Experimental Psychology: Learning,
Memory and Cognition, 14, 709715.
Porter, S., & Birt, A. (2001). Is traumatic memory special? A comparison of traumatic memory characteristics with memory for other emotional life experiences. Applied Cognitive Psychology, 15, S101S117.
Read, J. D., & Lindsay, D. S. (2000). Amnesia for summer camps and high school
graduation: Memory work increases reports of prior periods of remembering
less. Journal of Traumatic Stress, 13, 129147.
Reisberg, D. (2001). Cognition: Exploring the science of the mind. 2nd ed. New York:
W. W. Norton.
Reisberg, D., Heuer, F., McLean, J., & OShaughnessy, M. (1988). The quantity, not
the quality, of aect predicts memory vividness. Bulletin of the Psychonomic
Society, 26, 100103.
Rime, B., Philippot, P., Boca, S., & Mesquita, B. (1992). Long-lasting cognitive and
social consequences of emotion: Social sharing and rumination. In W. Stroebe
& M. Hewstone (Eds.), European review of social psychology, Vol. 3 (pp. 225258).
New York: John Wiley.
Rubin, D. C., & Kozin, M. (1984). Vivid memories. Cognition, 16, 8195.
Safer, M. A., Bonanno, G. A., & Field, N. P. (2001). It was never that bad: Biased
recall of grief and long-term adjustment to the death of a spouse. Memory, 9,
195204.
Safer, M. A., Christianson, S.-., Autry, M. W., & sterlund, K. (1998). Tunnel
memory for traumatic events. Applied Cognitive Psychology, 12, 99118.
Schmidt, S. R., & Williams, A. R. (2001). Memory for humorous cartoons. Memory
& Cognition, 29, 305311.
Schooler, J., Bendriksen, M., & Amadar, Z. (1997). Recovered memories and false
memories. Oxford: Oxford University Press.
Schreiber, C., & Kahneman, D. (2000). Determinants of the remembered utility of
aversive sounds. Journal of Experimental Psychology: General, 129, 2742.
Shobe, K. K., & Kihlstrom, J. F. (1997). Is traumatic memory special? Current Directions in Psychological Science, 6, 7074.
Spiro, R. J. (1980). Accommodative reconstruction in prose recall. Journal of Verbal
Learning and Verbal Behavior, 19, 8495.
Stanny, C. J., & Johnson, T. C. (2000). Eects of stress induced by a simulated shooting on recall by police and citizen witnesses. American Journal of Psychology,
113, 359386.
Steblay, N. J. (1992). A meta-analytic review of the weapon focus eect. Law and
Human Behavior, 16, 413424.
Takahashi, H. (Director), & Asari, S. W. (Producer). (1996). Magic Users Club: Ill
follow you [DVD]. Bandai Visual Corporation, Inc.
Walker, W. R., Vogl, R. J., & Thompson, C. P. (1997). Autobiographical memory:
Unpleasantness fades faster than pleasantness. Applied Cognitive Psychology, 11,
399414.
Wessel, I., & Merckelbach, H. (1997). The impact of anxiety on memory for details
in spider phobics. Applied Cognitive Psychology, 11, 223232.

40

Wessel, I., & Merckelbach, H. (1998). Memory for threat-relevant and threatirrelevant cues in spider phobics. Cognition & Emotion, 12, 93104.
Wessel, I., van der Kooy, P., & Merckelbach, H. (2000). Dierential recall of central and peripheral details of emotional slides is not a stable phenomenon.
Memory, 8, 95109.
White, N. M. (1991). Peripheral and central memory-enhancing actions of glucose.
In R. C. A. Frederickson, J. L. McGaugh, & D. L. Felten (Eds.), Peripheral signalling of the brain: Neural, immune and cognitive function (pp. 421441). Toronto:
Hogrefe and Huber.
Williams, J. M. G., Ellis, N. C., Tyers, C., Healy, H., Rose, G., & MacLeod, A. (1996).
The specicity of autobiographical memory and imageability of the future.
Memory & Cognition, 24, 116125.
Williams, L. M. (1995). Recovered memories of abuse in women with documented
child sexual victimization histories. Journal of Traumatic Stress, 8, 649673.
Winningham, R. G., Hyman, I. E., Jr., & Dinnel, D. L. (2000). Flashbulb memories?
The eects of when the initial memory report was obtained. Memory, 8, 209
216.
Yerkes, R., & Dodson, J. (1908). The relation of strength of stimulus to rapidity of
habit-formation. Journal of Comparative Neurology of Psychology, 18, 459482.

41

re there special neural mechanisms to account for our memories

for emotional events? In the service of a species survival, has evolution equipped organisms with a specialized set of mechanisms that encode,
consolidate, and retrieve memories in a domain-specic manner, operating
dierentially in emotional and in nonemotional contexts? Indeed, evidence from
cognitive psychology and neuroscience suggests that such distinct emotional
memory mechanisms exist and depend on specic neural structures that we will
review.
The uncommon vividness and (presumed) accuracy of ashbulb memories
led Brown and Kulik (1977) to suggest that autobiographical memories surrounding the learning of emotionally arousing information relied on a unique
memory process. These authors speculated that encountering emotionally
salient events resulted in enhanced activity in the brainstems reticular formation, leading to a now-print mechanism that permanently xed memories of
these eventsin addition to autobiographical ephemera accompanying them
in memory (Brown & Kulick, 1977; Schooler & Eich, 2000). More recent research
has revised this proposal somewhat and indicates that while such memories
can indeed be vivid, their accuracy is modest at best and often distorted (see
Schmolck, Bualo, & Squire, 2000). Though ashbulb memories are no longer
held to be a qualitatively dierent kind of memory, the broader notion of emotional memories complements earlier ideas that such memories are typically
experienced with special vividness and personal meaning and that there is evidence for specialized processes that subserve them.
The current taxonomy of memory systems describes distinct neural mechanisms for declarative and nondeclarative memory processing (Squire & Kandel,
2000). The medial temporal lobe memory systemincluding the hippocampal
formation as well as the overlying cortexis known to be integral to the forma-

42

42

tion and consolidation of declarative memories (or relational memories; see

Eichenbaum & Cohen, 2001), whereas nondeclarative memories depend on regions outside the medial temporal lobe. This classic taxonomy distinguishes
memory systems on the basis of their mode of operationin what way they are
acquired and in what way they are accessible to guide behavior. Of the other ways
of distinguishing memory systems, two are important to operationalizing the
term emotional memory: (1) systems directly involved in memory versus those
having a modulatory role and (2) systems dierentially involved depending on
the nature of the material processed (domain-specic). Emotional memory concerns a specic domain of declarative memory, namely, memory for events or
stimuli that are themselves emotional or that occurred in an emotional context.
Emotional memory encompasses the enhancement of memory (assessed by, for
example, a recognition test) for stimuli that are emotionally arousing (e.g., pictures of mutilated faces) compared to memory for stimuli that are emotionally
neutral (e.g., pictures of neutral faces); it also includes the enhancement of
memory for neutral faces that were encoded in an emotionally arousing context
compared to a neutral context; and it even includes the possible repression of
memory for faces associated with extremely traumatic events (a controversial
possibility).
Our review focuses on the facilitative eects of emotion on declarative memory
rather than its possible suppressive eects. As such, memory for the mutilated
faces may be a particularly potent declarative memory, relying on the medial
temporal lobe system just as in memory for neutral faces. This eect, though well
described, needs to be distinguished from the equally well-described eects of
chronic stress and severely traumatic context on memory; in those cases, declarative memory is also modulated but in a qualitatively dierent way. We thus
envision a nonmonotonic relationship of emotional arousal to memory accuracy.
Mild to moderate, and transient, emotional arousal generally enhances memory;
severe or chronic emotional arousal instead suppresses and distorts memory.
Though emotional arousal can be mapped on a continuum, its eects on declarative memory are not linear [see chapters 1, 3, and 8 for more discussion of this
topicEds.].
Another important distinction concerns the eects of emotional arousal triggered by properties that are intrinsic to the stimuli being encoded (e.g., memory
for pictures of mutilations) versus eects provided by a context (e.g., memory for
neutral pictures encoded in a highly arousing context). Again, without going
into detail, our view is that the two engage equivalent mechanisms and lead to
similar eects when the emotional arousal relates to the stimulus. They may lead
to dierent eects when the emotional arousal is irrelevant to the stimulus (see
chapter 1).
As a nal note, and to avoid any confusion, we emphasize that we use emotional memory to refer to declarative memory for emotional stimuli (where
emotional stimuli means stimuli that induce an emotion in those who perceive

43

the stimulieither as a result of intrinsic properties of the stimuli themselves

or as a result of an emotional context related to the stimuli). This usage diers
from another usage of the term, especially in animal studies, as a form of nondeclarative memory (e.g., fear conditioning). Thus, classical fear conditioning
has been occasionally referred to as emotional memory, and in a sense, it is that.
But we do not use the term that way here.
What neural structures are involved in making emotionally arousing stimuli
more memorable than neutral stimuli? Are these structures in some way specialized to process emotional memory? And what mechanisms do such structures
implementin particular, do they modulate declarative memory at the level of
encoding, consolidation, or retrieval?
These questions will be the focus of this chapter. We will emphasize the role
of the human amygdala in the enhancement of memory for emotional events
through its inuence over medial temporal lobe structures and other neural
regions; we will also survey several other structures that may participate in emotional memory.

Emotional Memory in

Nonhuman Animals

Much of the human research presented in this chapter is predicated on animal

research, specically work in rats and nonhuman primates. In this section, we
review some of the studies in animals, but the breadth of this work is beyond the
scope of this chapter. We encourage the reader to look elsewhere for a comprehensive review of the literature on the topic (Baxter & Murray, 2000; Eichenbaum & Cohen, 2001; LeDoux, 2000; McGaugh, 2000).
By virtue of its location in the brain, for many years the amygdala has been the
focus of much interest as a possible site of learning and memory (Squire, 1987). It
was originally thought to play an integral role in the formation of declarative
memories, together with other proximal structures in the medial temporal lobe.
More careful study has revealed the role of the amygdala to be separate from the
traditional medial temporal lobe memory system that includes the hippocampus
and adjacent cortex (Murray, 1992; Zola-Morgan, Squire, Alvarez-Royo, & Clower,
1991; Zola-Morgan, Squire, & Amaral, 1986). Following the initial characterization of the memory decits of the famous patient H. M. (Scoville & Milner, 1957),
considerable work went into developing a nonhuman primate model of amnesia
through lesions of the medial temporal lobe (see Eichenbaum & Cohen, 2001, for
a review of these initial studies). Typically, these early studies used a surgical approach similar to that used in patient H. M. Namely, the entire medial temporal
lobe of the monkeys was removed, including the amygdala, hippocampus, and
surrounding cortical regions: the entorhinal, parahippocampal, and perirhinal
cortices. After the successful development of a nonhuman primate model of

44

amnesia, work began on delineating the role of the individual anatomical components of the medial temporal lobe in the amnesic animal. This work led to the nding that circumscribed lesions of the amygdala (sparing the surrounding cortical
areas) do not produce the amnesic eects produced by damage to other medial temporal structures, such as the hippocampus and entorhinal cortex (Zola-Morgan
et al., 1991; Zola-Morgan, Squire, & Amaral, 1989). This is not to say that these
lesions have no behavioral or mnemonic eects. Many of these animals displayed
altered emotional behavior reminiscent of a subset of the symptoms described
by Klver and Bucy in the 1930s (1937; see also Brown & Schafer, 1888, and
Weiskrantz, 1956), such as increased visual and oral inspection of potentially
threatening objects and increased tameness. These studies by Zola-Morgan and
colleagues showed that reduced declarative memory performance after damage
to the medial temporal lobe cannot be due to amygdala damage per se.
Whereas studies in nonhuman primates showed that the amygdala is not
necessary for the formation of new declarative memories, earlier work in rats had
begun to elucidate the modulatory role that the amygdala plays in memory
formation. For the purposes of this review, we will focus on the role of the
amygdala in the modulation of long-term declarative memories and refer
the reader to other reviews (Davis, 1997; LeDoux, 2000) for coverage of the
amygdalas role in nondeclarative forms of emotional memory, such as Pavlovian
fear conditioning. Perhaps the rst indication of the amygdalas modulatory role
in memory came from studies of the eects of electrical stimulation of the rodent
amygdala. Goddard (1964) demonstrated that amygdala stimulation following
aversive training resulted in reduced memory for the training period, pointing to
a role in the modulation of consolidation rather than a direct role in memory
formation. Subsequent work showed that amygdala stimulation could either
reduce or enhance the consolidation of previously learned materials depending on the intensity of the stimulation (Gold, Hankins, Edwards, Chester, &
McGaugh, 1975; McGaugh, 2000), possibly analogous to the nonmonotonic
relation between emotional arousal and human declarative memory that we
briey alluded to in the introduction. The neurobiology of this phenomenon has
subsequently been the focus of much research. Based on this work, we now know
that key components through which the amygdala modulates declarative
memory include the stress hormones epinephrine and corticosterone, as well as
noradrenergic, GABAergic, glutamatergic, and peptidergic neurotransmission
within the amygdala (McGaugh, 2000; Roozendaal, 2000; Tomaz et al., 1993).
Hormonal modulation of memory storage has been a major focus of study in
physiological psychology (McGaugh, 1983). Many of the behavioral paradigms
used to study learning and memory in the rat are suciently stressful to result
in the release of the adrenal hormones epinephrine and corticosterone. Gold and
van Buskirk (1975) showed that the administration of epinephrine after a learning episode enhanced memory for that episode. Numerous studies have replicated
this nding using both endogenous and exogenous manipulation of epinephrine.

45

Similarly, corticosterone manipulation has shown that this hormone plays a

modulatory role in memory (Roozendaal, 2000). The mechanism of action of
these hormones on memory function appears to be at the level of the amygdala
(McGaugh, 2000). Manipulations such as amygdala lesions or blockade of either
the -adrenergic receptor (Liang, Juler, & McGaugh, 1986) or the glucocorticoid receptor (Oitzl & de Kloet, 1992) in the amygdala block the memory-enhancing eects of emotional arousal. The nal common pathway of the eects of these
hormones on memory appears to be noradrenergic neurotransmission within
and originating from the amygdala (McGaugh, 2000). Blockade of -adrenergic
activity within the amygdala blocks the memory-enhancing eects of both
epinephrine (Liang et al., 1986) and corticosterone (Roozendaal, Williams, &
McGaugh, 1999). The relationships among epinephrine, the amygdala, and
memory are somewhat complicated, since peripheral epinephrine does not cross
into the brain, but instead exerts its eects via the release of glucose (Gold & van
Buskirk, 1975) or via neural transmission of body-state information through
the vagus nerve (Clark, Krahl, Smith, & Jensen, 1995; Clark, Naritoku, Smith,
Browning, & Jensen, 1999). The amygdala is not involved in all aspects of the
memory but only inuences other regions, such as the hippocampus and striatum (Packard & Teather, 1998), during emotional arousal and during a specic
time window in consolidation (Bianchin, Mello e Souza, Medina, & Izquierdo,
1999). Inactivation of the amygdala does not block the memory for an emotionally arousing event once established (Packard, Cahill, & McGaugh, 1994), suggesting that the memory trace per se does not depend on the amygdala.
As we said in the introduction, an important issue in relation to the eects of
emotion on long-term memory is the well-documented deleterious eects of stress
on memory. (A full account of the eects of stress on cognitive function is beyond
the scope of this chapter; a brief explanation follows.) We think of emotion as a
transient state (on the order of msec to seconds) during which physiological response systems are activated in preparation for action; stress, on the other hand,
is a more prolonged physiological perturbation due to a challenge to homeostasis
(see Lovallo, 1997; McEwen, 2000). The onset of a stressful experience likely involves the experience of negative emotion, often characterized by a lack of control
(Lovallo, 1997). The more prolonged nature of stress may result in a completely
dierent pattern of eects on physiology and behavior. One might think of this
relationship as analogous to the Yerkes-Dodson law (1908), which is characterized by an inverted-U-shaped eect wherein midrange levels of emotion are benecial to memory whereas higher, more prolonged emotional experience may
result in stress and subsequently impaired memory performance. Considerable
research has documented just such a relationship between stress and memory
performance in animals and humans (see Lupien & Lepage, 2001; McEwen &
Sapolsky, 1995). This inverted-U eect likely does not operate within the realm of
emotional memory that we refer to in this chapter (see chapter 1 for a discussion
of this topic); instead, the descending limb of the inverted-Uat which point

46

memory would be deleteriously aectedcorresponds to the onset of stress eects

on memory. The relationships among emotion, stress, and memory are quite complicated and the focus of much research. Throughout this chapter, we focus primarily on the eects of emotion on memory. We refer the reader to one of many
recent reviews on the topic of stress and memory for full coverage of this fascinating topic (see de Kloet, Oitzl, & Joels, 1999; Kim & Diamond, 2002; Lupien &
Lepage, 2001; see also chapter 3 in this volume).
Results from these studies suggest that a special neurobiological system
relying on the amygdala and stress hormonesis active during arousing learning situations to enhance memory for these events. The role of the amygdala,
then, is not in the storage of emotional memories but in the modulation of other
neural structures that directly implement such storage. During emotional
arousal, the amygdala increases its modulatory eect and, together with the
inuence of peripheral stress hormones, modulates other neural structures such
as the hippocampus and striatum to enhance the memory trace for emotional
events. This animal research has served as a basis for much of the work in the
neuropsychology of human emotional memory. Are the same neural mechanisms outlined in nonhuman primates and rodents applicable to the study of
human emotional memory in the laboratory? The answer to this question will
be the focus of the remainder of this chapter.

The Human Amygdala

and Emotional Memory

Lesion Studies
For most brain-behavior relationships, our initial source of information on the
role of the brain in human behavior, as well as on the relationship between the
amygdala and emotional declarative memories, has come from patients with
specic brain lesions. Though selective bilateral lesions of the amygdala are rare,
the few reported cases are illustrative. In addition to bilateral temporal lobectomieswhich are no longer performed due to the profound amnesia produced in
patient H. M. (Scoville & Milner, 1957)two disorders have been described that
result in the bilateral destruction of the amygdala: Urbach-Wiethe disease and
herpes simplex encephalitis. Urbach-Wiethe disease (also known as lipoid proteinosis) is a rare hereditary disorder characterized by the deposition of hyaline
material in the skin and mouth areas and is associated with bilateral mineralization of medial temporal lobe structures in about half of the cases, with specic
mineralization of the amygdala in some cases. Herpes simplex encephalitis is an
inammation of neural structures following viral infection. Even though this
disease may result in widespread pathology throughout the nervous system,
its earliest pathology almost invariably includes the amygdalae. Additionally,

47

unilateral temporal lobectomy is commonly employed in the surgical treatment

of intractable epilepsy. Here, we report on studies of long-term declarative
memory of emotional events from subjects with both bilateral and unilateral
amygdala damage.
Bilateral Amygdala Damage One well-characterized case of bilateral amygdala
damage was originally reported by Tranel and Hyman (1990). The patient
(SM046) had been diagnosed with Urbach-Wiethe disease. Computerized tomography (CT) and magnetic resonance imaging (MRI) conrmed the mineralization
of both amygdalae, as well as minimal damage to anterior entorhinal cortices,
but no other structural abnormality, an anatomical picture conrmed by functional imaging. Neuropsychological evaluation of this patient revealed normal
general intellect, language, and verbal memory function. These results are in
stark contrast to the profound declarative memory impairment following hippocampal damage (Squire, 1987).
Follow-up research with patient SM046 and others with amygdala damage
has highlighted the modulatory role in memory played by the amygdala.
Markowitsch and colleagues examined emotional memory formation in two
patients with Urbach-Wiethe disease, including bilateral mineralization of the
amygdala (Babinsky et al., 1993; Markowitsch et al., 1994). Neuropsychological evaluation revealed that though neither patient was amnesic, they showed
reduced performance on several standard neuropsychological tests of memory
(e.g., Auditory Verbal Learning Test). Memory for emotional material was similarly depressed, with one patient (C. P.) showing more impairment than the other
(B. P.). Specically, in a word stem completion task of previously presented emotional or neutral words, C. P. remembered the neutral words better than the
emotional words, whereas B. P. showed roughly equivalent performance for
neutral and emotional stimuli, as compared to control subjects who tended to
remember the emotional stimuli better. Similarly, in a recognition test of previously presented emotional pictures, C. P. recognized neutral pictures better than
emotional pictures in contrast to normal controls who showed enhanced recognition performance of the emotional materials.
Testing one of the same patients previously mentioned (B. P.), Cahill, Babinsky,
Markowitsch, and McGaugh (1995) showed that this patient did not show the
normal enhancement of memory for an emotionally arousing slide show. In
the task used in this study, originally described in Cahill and McGaugh (1995;
see also Heuer & Reisberg, 1990), all participants were exposed to a slide show
that included both neutral and emotionally arousing slides. Age-matched control subjects recalled the emotionally arousing stimuli much better than the
neutral stimuli, but patient B. P. showed equivalent memory for both types of
stimuli. Patient B. P. rated his emotional reaction to the slide show similarly
to control participants (B. P.s rating: 8, controls rating: 7.25 1.2, on a scale
of 0 to 10). Even though B. P. rated his emotional experience the same as

48

that of control subjects, this emotional experience did not translate into enhanced memory.
Using the same task, Adolphs, Cahil, Schul, and Babinsky (1997) also tested
the performance of an additional patient with bilateral amygdala damage
(SM046). This study featured analysis of memory for each individual slide for
both SM046 and for B. P. Whereas normal participants remembered emotionally
arousing slides signicantly better than the neutral slides, neither amygdaladamaged patient showed this pattern (see g. 2.1). Specically, the slide remembered best by the control participants (a slide showing the surgically reattached
legs of a car crash victim) was the one on which the two patients deviated most
from the controls scores (see g. 2.1A). Like B. P., SM046 also endorsed normal ratings of subjective emotional arousal for the story.
The pattern of impaired facilitation of memory for emotionally arousing material in the face of apparently normal memory performance for neutral material in
subjects with bilateral amygdala damage contrasts sharply with the performances
of amnesic subjects. Subjects with hippocampal or diencephalic amnesia are impaired in their overall memory performance, regardless of the nature of the material, but show a normal enhancement (albeit of smaller magnitude) when the
subject matter is emotionally arousing (Hamann, Cahill, & Squire, 1997).
Unilateral Amygdala Damage Using the same task, Adolphs, Tranel, and Denburg
(2000) examined the pattern of emotional memory performance following unilateral amygdala damage. Eight subjects with unilateral amygdala damage consequent to temporal lobectomy (6 left; 2 right), 9 brain-damaged controls with
no damage to the anterior temporal lobe, and 7 normal controls participated in
the study. In this experiment, each slide in the story was rated on scales of emotional valence, arousal, unusualness, and complexity. There were no group
dierences in slide ratings. As in the previous study, both normal controls and
brain-damaged controls showed enhanced memory for phase 2 of the slide/narrative story, specically for the most highly arousing slide. By contrast, the group
with left amygdala damage failed to show enhanced memory for this slide, showing the same pattern as previously reported for subjects with bilateral amygdala
damage. (The two subjects with right amygdala damage appeared to perform
normally, but these ndings are inconclusive due to the small sample size of this
group.) These ndings point to a role for the left amygdala in the consolidation
of declarative memory for emotionally arousing stimuli.
Further studies have addressed the role of unilateral amygdala damage in the
formation of memory for emotional words. Phelps, LaBar, and Spencer (1997)
examined emotional memory formation in 26 subjects following unilateral temporal lobectomy. In this study, subjects were presented with a list of 27 words
(9 positive, 9 negative, and 9 neutral) while skin conductance responses (SCRs)
were recorded. A surprise recall test was administered 1 minute after the presentation of the word list. Results illustrated that the left temporal lobectomy

49

Figure 2.1. A. Plots of standard deviations from normal of patient SM046 across
all 15 picture stimuli. B. Raw data for patient SM046 and normal controls. (Reprinted from Impaired declarative memory for emotional material following bilateral amygdala damage in humans, Adolphs, Cahill, Schul, & Babinsky, 1997,
Learning & Memory, 4, 291300. Copyright 1997 by Cold Spring Harbor Laboratory Press.)

group had the worst recall of the word list, but this eect was not statistically
signicant. Each group (controls and both temporal lobectomy groups) recalled
the negative and positive words better than the neutral words, but there was no
dierence among the groups in terms of the pattern of word recall. The authors
noted that the words used in the negative and positive categories (e.g., victim,
comedy) were perhaps not salient enough to produce the emotional arousal
necessary to show any group dierences in emotional memory performance. In

50

fact, psychophysiological data from SCR, a measure of autonomic arousal, illustrated that the neutral word category elicited a greater response than did either
emotion word category.
In light of these ndings, these investigators reported an additional study designed specically to address the role of the arousal dimension in memory consolidation (LaBar & Phelps, 1998). In this follow-up study, 22 temporal lobectomy
patients (10 left, 12 right) were presented with a list of 40 words (20 arousing,
20 neutral). The arousing words chosen in this study consisted of profanities,
sexually explicit words, and words depicting social taboos. These words were
successful in producing increased SCRs over neutral words and were rated as
signicantly more arousing than neutral words in controls, as well as in both
right and left temporal lobectomy groups. Free recall for the words was assessed
both immediately and at 1hour postencoding. The results were assessed in terms
of forgetting rates between the immediate and delayed free recall tests. Only the
control group showed a dierential forgetting rate for the arousing versus neutral words, showing increased recall for the arousing words at the delayed recall test. Both the right and the left temporal lobectomy groups showed decreased
memory for the arousing words at delayed recall. Consistent with the two previously reported studies, in this study the left temporal lobectomy group showed
the poorest levels of recall; their performance was signicantly worse than that
of both controls and the right temporal lobectomy group. This eect was not,
however, specic to emotionally arousing material, as the left temporal lobectomy group showed reduced overall performance but not a specic impairment
on memory for emotional stimuli. Although these results are consistent with
previous research showing a left-hemisphere dominance for verbal memory, they
do not address a specic role of the right or left amygdala in memory for verbal
or visual emotional stimuli.
A recent study in our lab has attempted to address this issue (Buchanan,
Denburg, Tranel, & Adolphs, 2001). Participants consisted of 20 subjects with
unilateral amygdala damage following temporal lobectomy for treatment of
epilepsy (11 left, 9 right) and 25 brain-damaged controls with unilateral lesions
outside the temporal lobe. Additionally, 35 normal control volunteers were recruited for participation. Participants were tested on 2 days; on the rst session
15 pictures diering in emotional salience (5 pleasant, 5 unpleasant, 5 neutral)
were presented along with a one-sentence verbal narrative description (e.g.,
accompanying a picture of two parents with their new twin babies was the narrative: After the babies were born, both parents were very happy, although a
bit exhausted). Subjects were told to watch attentively while the emotional
responses to the stimuli were recorded; no mention of a follow-up memory test
was made. Twenty-four hours after the rst session, subjects memory for the
slides was assessed with free recall, multiple choice, and four-alternative forcedchoice recognition tests. Within both free recall and multiple choice tests,
memory for narrative and picture information was assessed. The recognition

51

test assessed only recognition of visual detail. Results illustrated that the group
with left amygdala damage was specically impaired on memory for emotional
narratives relative to memory for neutral narratives. Interestingly, this group
was not impaired on memory for emotional picture information. The right
amygdala group, on the other hand, was impaired on visual recognition memory;
however, this impairment was not specic to emotional pictures (see g. 2.2).
These ndings support a material-specic role of the left amygdala in the
processing of verbal emotional stimuli and a role of the right amygdala in processing visual emotional stimuli, corroborating previous work describing the
separable language versus visuospatial processing roles for the left and right
hemispheres, respectively (Dobbins, Kroll, Tulving, Knight, & Gazzaniga, 1998).
These data replicate previous work illustrating a decit in verbal emotional
memory in individuals with left amygdala damage (LaBar & Phelps, 1998;
Phelps et al., 1997) while illustrating a lateralized pattern of the amygdalas
inuence on emotional memory.

The Human Amygdala and Memory

for Gist Versus Detail

Memory for gist as well as for peripheral details has been a major topic in the
study of the eects of emotion on memory (Christianson & Loftus, 1991; Heuer
& Reisberg, 1990). Despite discrepancies in the ndings, several studies have
shown that emotional arousal enhances memory for gist, but not memory for
detail (Burke, Heuer, & Reisberg, 1992; Reisberg & Heuer, 1992; see also chapter 1 here). More work is needed in operationalizing these constructs; however,
the basic idea is that gist or central information pertains to the most salient, relevant aspects of a stimulus (the aspects that one would focus on when describing
it to someone else and whose alteration would change the meaning of the stimulus), whereas detail or peripheral information is everything else. Clearly, these
categories do not form a strict dichotomy, but some features of a stimulus are
almost unanimously deemed gist, whereas others are described as detail.
Recent work in our lab has addressed the role of the amygdala in memory for
the gist/central information versus the peripheral details of emotional stimuli.
Using a task similar to that already described (Buchanan et al., 2001), Adolphs,
Denburg, and Tranel (2001) assessed memory for the gist and details of emotionally negative and neutral slides and narratives from patient SM046 as well
as from patients with unilateral temporal lobectomy (12 left, 8 right), braindamaged controls (n = 15), and age-matched controls (n = 47). Memory for gist
was operationalized as salient, general information of the stimulus sucient
to distinguish that particular stimulus from all the other stimuli and that did
not depend on remembering details of the scene; it was assessed with a fouralternative forced-choice written questionnaire. Detail memory was dened as

52

Figure 2.2. Free recall for narratives and pictures across valence categories.
A. Mean ( standard error) of correctly recalled narratives across all subjects from
each valence category. B. Mean ( standard error) of correctly recalled pictures
across all subjects from each valence category. (Reprinted from Verbal and nonverbal emotional memory following unilateral amygdala damage, Buchanan,
Denburg, Tranel, & Adolphs, 2001, Learning & Memory, 8, 326335. Copyright
2001 by Cold Spring Harbor Laboratory Press.)

53

information that could be accessed only from a detailed memory of the visual
image and was assessed using a four-alternative forced-choice visual recognition task showing the original visual stimulus and three computer-manipulated
foils that diered only in their details. The study found that all groups (including those with unilateral temporal lobectomies) showed enhanced memory for
the gist of negative rather than neutral stimuli, whereas patient SM046 showed
the opposite pattern, with greater memory for the gist of the neutral than the
negative stimuli. Performance on detail memory, on the other hand, showed
a dierent pattern. Both control groups (age-matched and brain-damaged
controls), as well as the left temporal lobectomy group, showed greater detail
memory for the neutral than the negative stimuli, whereas SM046 and the right
temporal lobectomy group remembered the details of both the negative and neutral stimuli equivalently. The general pattern of memory performance on this
task in the control subjects was that emotionally negative stimuli enhanced
gist memory but reduced detail memory, compared to neutral stimuli. Bilateral amygdala damage, on the other hand, interfered with both of these eects:
SM046 remembered gist for unpleasant stimuli relatively worse than controls,
and her detail memory performance for negative stimuli was better than that
of controls (see g. 2.3). This study had an inherent confound precluding an
unambiguous interpretation of the data; gist and detail memory were assessed
with dierent methods (questionnaire and visual recognition, respectively).
Current studies under way in our laboratory are providing data conrming that
the eects reported here can really be attributed to the amygdalas role in processing gist and detail information.
These ndings provide evidence that the human amygdala modulates declarative memory for emotionally arousing stimuli through dierential eects on
memory for gist and for visual detail. The data are consistent with the idea that
the amygdala acts as a lter in the encoding of relevant information from emotional stimuli. Whereas the healthy amygdalae are able to enhance the memory
processing of the gist of these stimuli and disregard the irrelevant details, damage to this structure impairs this ability. This pattern is consistent with recent
work showing that the amygdala is involved in the enhancement of perception
of emotionally salient events (Anderson & Phelps, 2001; hman & Mineka,
2001), suggesting that both perceptual and mnemonic processing of unpleasant events critically involve the amygdala.
Findings from these studies indicate that the emotional arousal, not the
pleasantness of the stimuli, most determines whether a stimulus will be remembered (see Bradley, Greenwald, Petry, & Lang, 1992). This arousal-mediated
memory enhancement is the feature that appears to be most aected following
amygdala damage (it is also arousal that is most predictive of amygdala activity
in functional imaging studies, as we discuss in the next section). These ndings
accord with recent neuroimaging work showing amygdala activity while viewing both negative and positive emotional stimuli, whether pictures (Hamann,

54

Figure 2.3. A. Memory for gist. Mean ( standard error) number of questions (out
of a maximum of two) answered correctly per stimulus, for the two emotion categories. All subjects except SM046 showed the same pattern: superior memory for
emotionally aversive stimuli compared with neutral stimuli. B. Memory for visual
detail. Subjects mean percentage correct on a four-alternative forced choice recognition memory task. The four stimuli consisted of the original stimulus and three
computer manipulated foils that were identical in gist but diered in detail. Whereas
controls remembered the details of aversive stimuli less well than those of positive
stimuli, subjects with amygdala damage remembered details about both types of
stimuli equally well. (Reprinted from The amygdalas role in long-term declarative memory for gist and detail, Adolphs, Denburg, & Tranel, 2001, Behavioral
Neuroscience, 15, 983992. Copyright 2001 by the American Psychological Association, Inc.)

55

Ely, Homan, & Kilts, 2002) or words (Hamann & Mao, 2002). Together, these
lines of work suggest that the primary variable that determines memory for emotional material is arousal and also that this variable is the most salient instigator of amygdala activity. The principal diculty here has been in constructing
stimuli of pleasant valence that can be as arousing as those of negative valence.

Neuroimaging of the Amygdala

and Emotional Memory

Functional neuroimaging has provided another tool with which to examine the
role of the human amygdala in the formation of emotional memories. Several
recent studies have been able to test specic hypotheses derived from both animal research and studies in humans with amygdala lesions using positron emission tomography (PET) and functional magnetic resonance imaging (fMRI).
The rst study to examine the role of the human amygdala in the formation
of emotional memories using functional neuroimaging was conducted by Cahill
et al. (1996). Healthy participants viewed emotionally arousing and neutral
videos during PET scanning. Three weeks later, participants were asked to recall all the information that they could remember from each video. As expected,
participants recalled signicantly more information from the emotionally arousing video than from the neutral video. Correlation analyses revealed a signicant
positive correlation between the glucose metabolic rate of the right amygdala
and the number of emotional lm clips recalled (r = 0.93; see section on gender
dierences for discussion of a lateralization of amygdala activity in a similar task).
Further analyses showed no such association with recall of neutral lm clips.
These ndings suggest that the amygdala is activated during the encoding of
emotionally arousing events and is involved in the translation of these events
into long-term memory. A follow-up study corroborated the nding that amygdala
activity is not involved in the formation of declarative memory for nonemotional
material (Alkire, Haier, Fallon, & Cahill, 1998). This study did, however, document an association between hippocampal activity and the formation of memory
for a nonemotional word list. These two studies further illustrate the dissociation of memory functions between the hippocampus and amygdala and highlight a specic role of the amygdala in the formation of emotional memories.
Subsequent studies have extended these ndings to show an association between bilateral amygdala activity during encoding and memory for both emotionally pleasant and unpleasant stimuli using both PET (Hamann, Ely, Grafton,
& Kilts, 1999) and fMRI (Canli, Zhao, Desmond, Glover, & Gabrieli, 1999). The
study by Hamann et al. additionally tested the emotional specicity of the
amygdalas inuence on memory by including a stimulus category of interesting and unusual pictures. These interesting pictures included a chrome rhinoceros and a scene from a surrealist painting. Presumably, if the amygdala is
involved in general memory enhancement, then amygdala activity should be

56

associated with memory for unusual yet nonemotional stimuli. Results from this
study illustrate that while these unusual pictures were better remembered than
neutral (e.g., a book) and pleasant pictures (e.g., opposite-sex nudes) 4 weeks
after encoding, this enhancement was unrelated to amygdala activity, which was
instead specically related only to memory for aversive and pleasant stimuli
(a nding also consistent with studies in monkeys, which have failed to nd
any eect of amygdala lesions on the von Restor eectenhanced memory for
especially unusual or distinctive stimuli in a set; Parker, Wilding, & Akerman,
1998). These results highlight the specic role of the amygdala in the enhancement of memories for emotionally signicant material regardless of valence (see
g. 2.4).
A recent event-related fMRI study has shown that stimuli rated as emotionally intense are associated with increased amygdala activity and increased
memory performance (Canli, Zhao, Brewer, Gabrieli, & Cahill, 2000). In this
study, 10 female volunteers were exposed to a selection of neutral and emotionally negative pictures (e.g., scenes of mutilation) while the fMRI response was
recorded for each picture. Immediately after viewing each picture, subjects rated
their emotional response on a scale from 0 (not emotionally intense at all) to 3
(extremely emotionally intense). Analysis of the fMRI response illustrated that
bilateral amygdala activity was correlated with increased ratings of emotional
intensity, such that the greater the emotional intensity, the greater the amygdala
response. Three weeks later, subjects returned for a surprise memory test in
which they were asked to report whether they were certain that they remembered a slide, whether the slide seemed familiar, or whether they did not remember the slide (in a remember/know recognition paradigm). Performance data
illustrated that those slides rated as extremely emotionally intense were remembered signicantly better than those rated as less intense. Additionally, the degree of left amygdala activation during picture encoding was correlated with
subsequent memory for the pictures (see section on gender dierences for discussion of lateralization of amygdala activity related to memory performance).
Those pictures that tended to produce the greatest response in the left amygdala
were also remembered most often.

Autobiographical Memories and

the Amygdala

In addition to the evidence reviewed, that the amygdala plays a role in the encoding and consolidation of emotional memories, there is evidence that the retrieval of autobiographical memories may depend, in part, on the amygdala.
Likely, the majority of our distant autobiographical memories are associated with
an emotional response, suggesting that perhaps the amygdala plays a role both
in the encoding and retrieval of these memories by virtue of their emotionally
arousing nature. Mori et al. (1997) assessed memory for autobiographical events

57

Figure 2.4. Top. Brain activity correlated with memory enhancement. Maps of pixels in which individual subject rCBF was signicantly correlated with individualsubject episodic memory enhancement superimposed on an axial MRI image. Left:
correlation map for pleasant stimuli at z = -10.5. Right: correlation map for aversive stimuli at z = -16.5. Note that in this gure the right hemisphere is shown on
the right. Bottom. Relationship between pleasant-picture memory and brain activity for individual subjects. Correlation scatterplots for the pleasant picture memory
with rCBF. a. Left amygdala. b. Right amygdala. c. Left hippocampus. d. Right hippocampus. (Reprinted with permission from Amygdala activity related to enhanced
memory for pleasant and aversive stimuli, Hamann, Ely, Grafton, & Kilts, 1999,
Nature Neuroscience, 2, 289294. Copyright 1999 by Nature America Inc.)

58

surrounding the 1995 Kobe earthquake in patients with probable Alzheimers

disease. Neuropathological studies have shown that the characteristic plaques
and neurobrillary tangles associated with Alzheimers disease are often located
in the amygdala, and atrophy of this structure has been reported in both autopsied patients (Scott, DeKosky, & Sche, 1991) and in vivo in patients with early
signs of the disease (Cuenod et al., 1993). Mori et al. examined the preexisting
individual dierences in the amygdala volume of a group of patients with
putative Alzheimers disease (diagnosed prior to the earthquake) to determine
whether changes in the size of this structure inuenced emotional autobiographical memory. Results from this study illustrate that the volume of the amygdala
was positively correlated with patients memories for their experiences during
and after the earthquake (see g. 2.5). Hippocampal volume was also correlated
with emotional memory but not as strongly as the relationship between amygdala
volume and emotional memory. This association between amygdala volume and
emotional memory was signicant even when controlling for factors such as age,
education, whole brain volume, and ratings of dementia. There was no relationship between the size of the amygdala and memory for general (nonemotional)
knowledge about the events surrounding the earthquake. These ndings extend
work from laboratory research focusing on the amygdala and emotional memory

Total Emotional Memory Score

Total Emotional Memory Score

0
600

1200
1600
2000
2400
Amygdalar Volume (mm3)

0
1600 2000 2400 2800 3200 3600
Hippocampal Volume (mm3)

Figure 2.5. Scatterplots of amygdala and hippocampal volumes with total emotional
memory scores of 36 patients with Alzheimers disease. Pearson correlation analysis showed signicant correlations between amygdala volume and total emotional
memory score and between hippocampal volume and total emotional memory score.
After controlling for the eects of age, sex, education, whole brain volume, and disease severity, the former correlation remained signicant, while the latter was no
longer signicant. (Reprinted with permission from Amygalar volume and emotional memory in Alzheimers disease, Mori, Ikeda, Hirono, Kitagaki, Imamura, &
Shimomura, 1999, American Journal of Psychiatry, 156, 216222. Copyright
1999 by the American Psychiatric Association, Inc.).

59

and suggest that memory for real-life emotional situations depends on the integrity of this structure.
A hypothesis has been proposed suggesting that the temporofrontal cortical
areas including the amygdala and uncinate fasciculus ber bundle connecting
these areas are specically involved in the retrieval of autobiographical memories (Kroll, Markowitsch, Knight, & von Cramon, 1997; Markowitsch et al., 2000).
More specically, a right-sided temporofrontal network purportedly retrieves
autobiographical memories, while a left-sided temporofrontal network encodes
this information into memory (Tulving, Kapur, Craik, Moscovitch, & Houle,
1994; Tulving et al., 1994). This hypothesis has found support from studies of
patients with lesions of these areas (Kroll et al., 1997; Levine et al., 1998), as
well as functional neuroimaging studies of healthy participants (Fink et al.,
1996; Markowitsch et al., 2000). Fink et al. showed that when subjects listened
to transcripts of their own autobiographical memories, activity in areas including the right amygdala, right hippocampus, and right prefrontal cortex increased
signicantly. The authors of several of these studies suggest that the role of these
areas in the retrieval of memory is, in part, due to the emotional nature of autobiographical memories (Markowitsch et al., 2000). Memories from our lives that
we are able to retrieve presumably have some aective character, thus making
these memories salient for recollection. These studies, in combination with previous work on emotional memory encoding, suggest that the amygdala may be
involved not only in encoding emotional information but also in its retrieval. One
note of caution regarding the studies reviewed in this last section is that they have
not pinpointed the observed eects solely to the amygdala but may reect structures in the immediate vicinity of the amygdala including the temporal pole.

Gender Dierences in

Amygdala Activity

Findings from several of these studies have suggested the possibility of gender
dierences related to lateralized amygdala activity in memory for emotional
visual stimuli. In studies on men, results have shown a predominantly right-sided
activation of the amygdala (Cahill et al., 1996; Hamann et al., 1999), whereas
two studies documented more left lateralized activation in women (Canli et al.,
1999, 2000). These gender dierences have recently been investigated directly
in a study including both men and women (Cahill et al., 2001) and using experimental conditions identical to those previously used with men only (described
in the section on neuroimaging; Cahill et al., 1996). Results from this study illustrated the same gender-specic lateralized pattern of activation previously documented across separate studies, with enhanced emotional memory performance
correlating with right amygdala activity in men but left amygdala activity in
women (Cahill et al., 2001). Another interesting facet of this work is the nding
that women show better memory for emotionally arousing material than do men

60

(Seidlitz & Diener, 1998). Canli, Desmond, Zhao, and Gabrieli (2002) examined
the possible neural mechanisms of womens enhanced memory for emotional
material. In a functional MRI study similar in methodology to previous experiments using emotional picture stimuli (Cahill et al., 2001; Canli et al., 2000),
the authors replicated the nding of womens enhanced memory for emotional
stimuli. Additionally, they found that women showed a greater network of neural
activity related to emotional memory (including the left amygdala, hippocampus, and frontal cortex) than men (whose activity was primarily located in the
right amygdala). Subjective ratings of emotional arousal were equivalent between men and women, so this nding suggests that gender dierences in emotional responsiveness alone are not responsible for this cognitive dierence. The
authors of these studies suggest that these gender dierences may reect dierent
cognitive strategies between men and women in the processing of these stimuli,
perhaps related to dierential attention to gist and to peripheral details, a cognitive phenomenon that also shows a gender dierence (see Cahill & van Stegeren,
2003). These intriguing ndings clearly warrant the inclusion of gender and
laterality as factors in future work on the relationship between the amygdala and
emotional memory. [For another perspective on gender dierences in emotional
memory, see chapter 8Eds.]

The Human Hippocampus

and Emotional Memory

The role of the hippocampus in the enhancement of memory by emotion has been
assessed in several studies. Two studies by Hamann, Cahill, McGaugh, and Squire
(1997) and Hamann, Cahill, and Squire (1997b) assessed memory for emotionally arousing materials in amnesic patients with damage to either the
hippocampus only, damage to the hippocampus and amygdala, or damage to
the diencephalon. Results from these studies illustrated that damage to the
hippocampus alone or damage to the diencephalon leaves the enhancement
of declarative memory by emotional arousal intact, presumably by relying on
the intact amygdala in the subjects included in these studies. Two patients with
combined damage to the hippocampus and amygdala (including overlying cortices) showed no discernible memory for the emotionally arousing or neutral
stimuli. A neuroimaging study (mentioned earlier in the section on the role of
the amygdala; Alkire et al., 1998) has demonstrated that activity of the
hippocampus, but not the amygdala, correlates with memory for nonemotional
materials. In another neuroimaging study, Hamann et al. (1999) addressed
the relationship between the amygdala and hippocampus in the enhancement
of memory by emotion. Bilateral activity in both the amygdala and hippocampus was associated with memory enhancement for both pleasant and aversive
stimuli in this study. Results from the studies with amnesics illustrate that the

61

hippocampus, while necessary for the formation of new declarative memories, is

not itself responsible for the modulation of memory by emotional arousal. The
neuroimaging work shows that correlated activity between the amygdala and
hippocampus may be a necessary mechanism for the enhancement of memory for
emotionally arousing events. This work further demonstrates the modulatory role
of the amygdala, which in combination with the hippocampus lays down a stronger memory trace for emotional situations than for nonemotional situations.

The Prefrontal Cortex

and Emotional Memory

The prefrontal cortex, notably its dorsolateral regions, is known to be involved

in both the encoding and retrieval of memories (Buckner & Wheeler, 2001;
Tulving, Kapur, Craik, et al., 1994) and is thought to provide executive strategies for recollecting, comparing, and ordering memories. Work from numerous
investigators has also suggested a role of the orbitofrontal region in emotional processing (Bechara, Damasio, & Damasio, 2000; Rolls, 2000b). As dorsolateral and
orbital sectors of the prefrontal cortex are intimately connected (Rolls, 2000a), and
as orbitofrontal cortex is also massively connected with the amygdala, it is natural to hypothesize that networks within the prefrontal cortex participate in structuring emotional memories as well.
Whether and how the prefrontal cortex may be involved in the modulation of
memory by emotion has been less well studied than the case of the amygdala, but
several studies have begun to provide some hints. Numerous studies have documented the eects of orbitofrontal damage on decision making (Bechara et al.,
2000; Bechara, Damasio, Damasio, & Lee, 1999; Bechara, Damasio, Tranel, &
Damasio, 1997; Bechara, Tranel, Damasio, & Damasio, 1996). This work has
outlined a model whereby the orbitofrontal cortex processes emotional signals from
the body that aect decision-making processes (Damasio, 1994). One study has
sought to distinguish this decision-making bias from the enhancement of memory
by emotion (Bechara et al., 2000). Patients with orbitofrontal damage were presented with a series of neutral and emotionally arousing picture stimuli in four sets.
One set of pictures was presented only once, another set was presented twice, a
third set was presented three times, and a fourth set was presented eight times.
Memory for these stimuli was assessed immediately after all stimuli had been presented. Results from this study illustrated that while the orbitofrontal patients
showed a reduced overall level of memory, these patients still showed enhanced
memory for the emotionally arousing pictures rather than for the neutral pictures,
and both groups showed improved memory for those stimuli that were repeated
most often. This study suggests that the biasing of decision making by emotion that
depends on the orbitofrontal cortex is separable from the inuence of emotion on
memory, which appears to rely less on the orbitofrontal cortex.

62

Recent neuroimaging studies have documented a role of the prefrontal cortex in the memory for neutral material encoded in an emotionally arousing context (Maratos, Dolan, Morris, Henson, & Rugg, 2001; Maratos & Rugg, 2001). For
example, these studies give the example of the word corn used in a negative
context: The farmer was shredded when he fell into the corn grinder, and used
in a positive context: The farmer was overjoyed with his bountiful crop of corn.
Using electrophysiological recordings, the authors in these two studies showed
that the areas that are normally activated during the retrieval of old neutral
words (including the right prefrontal cortex) exhibited enhanced activity when
these words were presented in a negative context (Maratos & Rugg, 2001). In
an event-related fMRI study using a similar experimental design (with the addition of a positive emotional context), Maratos et al. (2001) demonstrated that
the re-presentation of neutral words that had been presented in a positive emotional context resulted in heightened activity in the orbitofrontal cortex, the
dorsolateral prefrontal cortex, and the medial temporal lobe areas such as the
amygdala and hippocampus. Results from these studies suggest (1) that activity within areas normally involved in the retrieval of neutral information (including the prefrontal cortex) is enhanced by an emotionally arousing context
and (2) that the same areas activated during encoding of emotionally arousing
materials (including the amygdala) are active during the retrieval of these
materials.
These studies from both lesion and neuroimaging paradigms illustrate the
complex relationship between the prefrontal cortex and the enhanced memory
for emotional material. Though the role of the PFC in emotional memory has
not been well documented in lesion patients, the neuroimaging data suggest a
subtle role for this region in the encoding and recognition of emotional stimuli.
It is worth emphasizing the close connectivity between the amygdala, orbitofrontal cortex, basal forebrain, and the dorsal prefrontal cortex. Studies in monkeys have demonstrated that the amygdala and orbitofrontal cortex are two
essential components of a system for processing the emotional properties of
stimuli; disconnection of the two structures can result in impairments on certain tasks of associative emotional memory that are as severe as lesions in either
structure (Gaan, Murray, & Fabre-Thorpe, 1993). It seems plausible to suggest
the same systems-level architecture for declarative emotional memory.

Neurodegeneration

and Emotional Memory

A nal section to our review concerns the insights gleaned not from focal lesion or
activation studies but from the more diuse types of damage incurred by neurodegenerative diseases (see chapter 9 here for a discussion of these conditions in
relation to aging). Alzheimers disease (AD) is initially characterized by memory

63

complaints, often concomitant with alteration in emotional behavior. The plaques

and neurobrillary tanglesthe neuropathological hallmarks of ADare often
rst noted in medial temporal lobe structures including the amygdala, entorhinal
cortex, and hippocampus (Hyman, Van Hoesen, & Damasio, 1990; Van Hoesen,
Augustinack, & Redman, 1999). These ndings have led several investigators recently to examine emotional memory in AD. Results from two studies have shown
impairments in the enhancement of memory by emotion in patients with probable AD (Abrisqueta-Gomez, Bueno, Oliveira, & Bertolucci, 2002; Hamann, Monarch, & Goldstein, 2000), yet two other studies have shown no such impairment
(Kazui et al., 2000; Moayeri, Cahill, Jin, & Potkin, 2000). The major dierence
between the studies showing impairment and those showing no impairment is that
the latter studies used a slide show paradigm similar to that described in Cahill,
Prins, Weber, and McGaugh (1994), whereas the former studies showing an
impairment used individual emotional pictures devoid of a narrative. This dierence in methodology deserves attention in future research in emotional memory.
Perhaps a cohesive narrative structure is necessary to keep such patients attention, whereas the presentation of unrelated emotional pictures may not be salient
enough (see discussion of the distinction between visually arousing and thematically arousing inucences on memory in chapter 1 for a fuller explication of this
possibility). The aforementioned study by Mori et al. (1999) that demonstrated
an association between amygdala volume and emotional memories for the Kobe
earthquake in AD patients suggests that, at least in some of these patients, the
volumetric loss of the amygdala plays a role in the ability to remember emotional
events (see previous section on autobiographical memories and the amygdala for
fuller description of this study).
The same neural structures (amygdala and hippocampus) that have been
described as playing a role in emotional memory in neurodegenerative diseases
are also known to be aected in normal aging. Specically, the volumes of both
the amygdala (Jack et al., 1997; Mu, Xie, Wen, Weng, & Shuyan, 1999) and hippocampus (Golomb et al., 1993; Jernigan et al., 2001) diminish with age. In line
with this reduction in brain volume are the well-documented memory impairments that accompany aging (Grady & Craik, 2000). The eect of aging on
memory for emotional material has not been well studied. A recent study from
our laboratory has sought to examine this issue (Denburg, Buchanan, Tranel,
& Adolphs, in press). In the study, 80 healthy adults between the ages of 35 and
85 years were presented with a series of pictures that ranged from pleasant to
neutral to unpleasant scenes (the same ones mentioned in connection with the
studies by Buchanan et al. [2001] and Adolphs et al. [2001] on pages 51 and
52). Memory for these pictures was then tested 24 hours later. Results from this
study showed that regardless of the age of the subject, memory for the emotionally arousing pictures was better than that for neutral pictures. Though there
was an overall decline in memory function with age, there was only a small dierence in the degree of memory modulation by emotion across the age range. These

64

ndings indicate a sparing of the inuence of emotion on memory with age,

perhaps due to a more pronounced age-related degeneration of the hippocampus than for the amygdala. Studies designed to assess the relationship between
age-related declines in brain volume and emotional memory function are necessary to test this hypothesis.

Summary and Integration

The studies reviewed in this chapter illustrate the complex interrelationships
among brain structures involved in the modulation of memory by emotion. Numerous brain structuresincluding the amygdala, hypothalamus, hippocampus,
cingulate, insular and orbitofrontal cortices, among othersare involved in both
the evaluation of and response to emotional stimuli. Findings from lesion studies
and neuroimaging point to the amygdala as the most critical structure for the
enhanced memory of stimuli encoded during emotional arousal.
Our focus on the amygdalas (and other structures) role in modulation of declarative memory for stimuli does not exclude its role in other aspects of emotional
memory, more broadly construed. In our view, the amygdala (together with structures like the ventral striatum and orbitofrontal cortex) helps to mount a coordinated, multisystem response to an emotional challenge. Such a response consists
of concerted changes in many parameters of an organisms body (autonomic,
endocrine, and other changes) and brain. The changes in brain functioning are
also multicomponent, and include, among others, changes in attention, alertness,
andthe focus of this chaptermemory encoding and consolidation. The known
neuroanatomical connectivity of the amygdala bears out this sketch; it receives
rudimentary (LeDoux, 2000) as well as highly processed sensory information
(Amaral, Price, Pitkanen, & Carmichael, 1992), which allows for quick and specic
activation of response systems that inuence behavior and subsequent memory
formation. The specic role of the amygdala appears to be that of instigator, the
catalyst that initially imbues incoming stimuli with emotional signicance and
triggers other neural regions to react accordingly. The detection and experience
of emotion increases the processing of the emotional stimuli by focusing attention
and enacting response systems (Frijda, 1987; Lang, Bradley, & Cuthbert, 1990).
The amygdala plays an integral role in this enhanced processing through inuence
over numerous brain structures, including the hippocampus, hypothalamus, brain
stem arousal centers, basal forebrain, and bidirectional connections with neocortex (McGaugh, 2000; Packard & Teather, 1998; Pitknen, 2000).
Along with the integrative, multisystem role played by structures like the
amygdala in coordinating emotional responses, its role in memory is equally
diverse. We can think of the ways in which emotion inuences memory as
changes in memory for (1) the emotion-inducing sensory properties of stimuli
themselves (e.g., remembering that a picture showed mutilated bodies), (2) the

65

intrinsically neutral sensory properties of stimuli associated with a related emotional state (e.g., remembering that one saw a picture of landscape as part of
story about a terrible disaster), (3) the emotional state itself in the perceiver
(e.g., remembering that one felt upset), or (4) higher-order information about
the emotion and the stimuli (e.g., remembering simply that there were some
emotional stimuli or that one felt an emotion). None of these is mutually exclusive, and there is evidence to suggest that the amygdala may contribute to
all of them. Interesting questions for the future remain about how these aspects of information are related and how they might depend dierentially on
certain brain structures.
The ndings we have reviewed place the amygdala in a modulatory position
of the traditional medial temporal lobe memory system, its activity inuencing
memory formation only under conditions of emotional arousal. In response to
emotion, the same brain structures are recruited in the service of both response
output and memory formation. The nding that enhanced activity of response
systemsautonomic, endocrine, and skeletal motor activityaccompanies
increased mnemonic representation illustrates the coordinated activity of these
brain structures in the service of survival. The lesion studies described suggest
that the amygdala is necessary for the enhancement of memory for emotionally
arousing stimuli. The dysfunction of one or both amygdalae (because of disease
or surgery) results in reduced memory for emotionally arousing materials. The
precise nature of such dysfunctionwhether it is hyper- or hypoactivation
can lead either to compromised emotional memories or to the pathological exaggeration of such memories, as seen in traumatic and phobic memories that
we have not reviewed here (Davis, Walker, & Lee, 1997; hman & Mineka,
2001). The nature of these studies has made it dicult for the assessment of
consolidation or retrieval eects but clearly implicate the amygdala in the encoding of emotional material into memory. Similarly, functional neuroimaging
studies have most consistently pointed to encoding as the memory stage at which
the amygdala exerts its eects (but see Dolan, Lane, Chua, & Fletcher, 2000).
The most consistent nding from the discussed studies is an association between
amygdala activity during encoding and the enhancement of subsequent memory.
The activation of the amygdala would seem to be a potent predictor of memory
performance.
Another pathway through which the amygdala may inuence memory performance is via hormonal output (see chapter 3 for a more thorough discussion
of these issues). The previously described pharmacological studies in animals
have illustrated the eect of manipulations of both catecholamines and glucocorticoids on emotional memory performance. Direct connections between the
amygdala and the hypothalamus mediate the release of both epinephrine and
glucocorticoids during emotionally arousing situations (Davis, 1997). Following their release, these hormones exert actions throughout the central nervous
system and specically at the amygdala (Honkaniemi et al., 1992; Shepard,

66

Barron, & Myers, 2000). A great deal of animal research has focused on the bidirectional nature of the relationship between the amygdala and stress hormones
(McGaugh & Izquierdo, 2000; Roozendaal, 2000). Enhancement of either
adrenergic or glucocorticoid activity improves, but blockade of these hormones
reduces, memory performance. These ndings have been documented in both
animals (Roozendaal, 2000) and humans (Buchanan & Lovallo, 2001; Cahill
et al., 1994). The nal common pathway of these hormones eects on memory
performance appears to be at the amygdalaspecically through noradrenergic neurotransmission in the lateral/basolateral nuclei (Quirarte, Roozendaal,
& McGaugh, 1997). As previously mentioned, manipulations of this activity
within these nuclei inuence hippocampal and cortical function (Escobar,
Chao, & Bermudez-Rattoni, 1998; Ikegaya, Nakanishi, Saito, & Abe, 1997),
providing a potential mechanism whereby the actions of these hormones at the
amygdala could inuence the formation of emotional memories. Even though
human research has yet to show a relationship among stress hormones,
amygdala activity, and emotional memory, animal research suggests such a
relationship; future work will no doubt focus on elucidating this topic.
Consistent with the nding that systemic hormones inuence subsequent
amygdala activity, one mechanism through which an emotional response could
inuence memory is via perception of the physiological response in the body.
Vagal stimulation in both animals (Clark et al., 1995) and humans (Clark et al.,
1999) results in increased memory performance. In fact, the sensory role of
the vagus nerve is proposed as one mechanism whereby peripherally released
epinephrinewhich does not readily pass the blood-brain barrierinuences
amygdala function in the formation of emotional memories (McGaugh, 2000).
These results illustrate that the inuence of the amygdala on memory could
occur at multiple stages: (1) through rapid initial responses inuencing neural
information processing with a short latency after the occurrence of a stimulus,
or (2) indirectly, through rst triggering an emotional response in the body and
subsequent central eects of the bodys physiological state (of which some of the
latter may then also be mediated by the amygdala, albeit at a later time than
mechanism 1). Most current studies have not permitted a separation of the emotion depicted in the stimulus, from the emotional reactions and feelings of the
subject, and an important goal for the future will be to disentangle these issues.
In all likelihood, a variety of structures participate in modulating the encoding, consolidation, and retrieval of our memories for emotional events, with respect to diverse components of information and at multiple temporal scales.
Structures such as the amygdala and the ventral striatum perhaps provide an
initial, rapid, automatic evaluation of the emotional signicance of stimuli, or
of the context within which they occur, and can then consequently modulate
our attention to those stimuli (e.g., via basal forebrain), the strength of their
memory consolidation (e.g., via hippocampus), or the nature of their contextual
links to autobiographical information and to other items stored in memory (e.g.,

67

via prefrontal regions). Perhaps, at a later point, many of the same structures
can modulate memory through less direct mechanisms. Given the enormous and
varied amount of declarative knowledge that humans store over their lifetime,
it seems prudent to acknowledge that multiple, complex neural mechanisms
would exist to encode, order, and eciently store such knowledge.

Note
We thank Antonio and Hanna Damasio, Daniel Tranel, and Natalie Denburg for
their participation in some of the studies reviewed. Supported in part by an NRSA
grant from the National Institute on Aging to T.W.B. and from the National Institute of Mental Health to R.A.

References
Abrisqueta-Gomez, J., Bueno, O. F., Oliveira, M. G., & Bertolucci, P. H. (2002). Recognition memory for emotional pictures in Alzheimers patients. Acta Neurologica
Scandinavica, 105, 5154.
Adolphs, R., Cahill, L., Schul, R., & Babinsky, R. (1997). Impaired declarative memory
for emotional material following bilateral amygdala damage in humans. Learning & Memory, 4, 291300.
Adolphs, R., Denburg, N. L., & Tranel, D. (2001). The amygdalas role in longterm declarative memory for gist and detail. Behavioral Neuroscience, 115, 983
992.
Adolphs, R., Tranel, D., & Denburg, N. (2000). Impaired emotional declarative memory
following unilateral amygdala damage. Learning & Memory, 7, 180186.
Alkire, M. T., Haier, R. J., Fallon, J. H., & Cahill, L. (1998). Hippocampal, but not
amygdala, activity at encoding correlates with long-term, free recall of nonemotional information. Proceedings of the National Academy of Science, USA, 95,
1450614510.
Amaral, D. G., Price, J. L., Pitkanen, A., & Carmichael, S. T. (1992). Anatomical
organization of the primate amygdaloid complex. In J. P. Aggleton (Ed.), The
amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction
(pp. 166). New York: Wiley-Liss.
Anderson, A. K., & Phelps, E. A. (2001). Lesions of the human amygdala impair
enhanced perception of emotionally salient events. Nature, 411, 305309.
Babinsky, R., Calabrese, P., Durwen, H. F., Markowitsch, H. J., Brechtelsbauer, D.,
Heuser, L., & Gehlen, W. (1993). The possible contribution of the amygdala to
memory. Behavioural Neurology, 6, 167170.
Baxter, M. G., & Murray, E. A. (2000). Reinterpreting the behavioral eects of
amygdala lesions in non-human primates. In J. Aggleton (Ed.), The amygdala:
A functional analysis (Vol. 2, pp. 545568). Oxford: Oxford University Press.
Bechara, A., Damasio, H., & Damasio, A. R. (2000). Emotion, decision making and
the orbitofrontal cortex. Cerebral Cortex, 10, 295307.
Bechara, A., Damasio, H., Damasio, A. R., & Lee, G. P. (1999). Dierent contributions of the human amygdala and ventromedial prefrontal cortex to decisionmaking. Journal of Neuroscience, 19, 54735481.

68

Bechara, A., Damasio, H., Tranel, D., & Damasio, A. R. (1997). Deciding advantageously before knowing the advantageous strategy. Science, 275, 1293
1295.
Bechara, A., Tranel, D., Damasio, H., & Damasio, A. R. (1996). Failure to respond
autonomically to anticipated future outcomes following damage to prefrontal
cortex. Cerebral Cortex, 6, 215225.
Bianchin, M., Mello e Souza, T., Medina, J. H., & Izquierdo, I. (1999). The amygdala
is involved in the modulation of long-term memory, but not in working or shortterm memory. Neurobiology of Learning and Memory, 71, 127131.
Bradley, M. M., Greenwald, M. K., Petry, M. C., & Lang, P. J. (1992). Remembering
pictures: Pleasure and arousal in memory. Journal of Experimental Psychology:
Learning Memory and Cognition, 18, 379390.
Brown, R., & Kulick, J. (1977). Flashbulb memories. Cognition, 5, 7399.
Brown, S., & Schafer, A. (1888). An investigation in the functions of the occipital
and temporal lobes of the monkeys brain. Philosophical Transactions of the Royal
Society of London, 179, 303327.
Buchanan, T. W., Denburg, N. L., Tranel, D., & Adolphs, R. (2001). Verbal and nonverbal emotional memory following unilateral amygdala damage. Learning &
Memory, 8, 326335.
Buchanan, T. W., & Lovallo, W. R. (2001). Enhanced memory for emotional material
following stress-level cortisol treatment in humans. Psychoneuroendocrinology,
26, 307317.
Buckner, R. L., & Wheeler, M. E. (2001). The cognitive neuroscience of remembering. Nature Reviews Neuroscience, 2, 624634.
Burke, A., Heuer, F., & Reisberg, D. (1992). Remembering emotional events.
Memory & Cognition, 20, 277290.
Cahill, L., Babinsky, R., Markowitsch, H. J., & McGaugh, J. L. (1995). The amygdala
and emotional memory. Nature, 377, 295296.
Cahill, L., Haier, R. J., Fallon, J., Alkire, M. T., Tang, C., Keator, D., Wu, J., & McGaugh,
J. L. (1996). Amygdala activity at encoding correlated with long-term, free recall of emotional information. Proceedings of the National Academy of Science,
USA, 93, 80168021.
Cahill, L., Haier, R. J., White, N. S., Fallon, J., Kilpatrick, L., Lawrence, C., et al.
(2001). Sex-related dierence in amygdala activity during emotionally inuenced memory storage. Neurobiology of Learning and Memory, 75, 19.
Cahill, L., & McGaugh, J. L. (1995). A novel demonstration of enhanced memory
associated with emotional arousal. Consciousness and Cognition, 4, 410
421.
Cahill, L., Prins, B., Weber, M., & McGaugh, J. L. (1994). Beta-adrenergic activation and memory for emotional events. Nature, 371, 702704.
Cahill, L., & van Stegeren, A. (2003). Sex-related impairment of memory for emotional events with beta-adrenergic blockade. Neurobiology of Learning and
Memory, 79, 8188.
Canli, T., Desmond, J. E., Zhao, Z., & Gabrieli, J. D. (2002). Sex dierences in the
neural basis of emotional memories. Proceedings of the National Academy of
Science, USA, 99, 1078910794.

69

Canli, T., Zhao, Z., Brewer, J., Gabrieli, J. D., & Cahill, L. (2000). Event-related activation in the human amygdala associates with later memory for individual
emotional experience. Journal of Neuroscience, 20, RC99.
Canli, T., Zhao, Z., Desmond, J., Glover, G., & Gabrieli, J. D. (1999). fMRI identies a
network of structures correlated with retention of positive and negative emotional memory. Psychobiology, 27, 441452.
Christianson, S.-., & Loftus, E. F. (1991). Remembering emotional events: The fate
of detailed information. Cognition and Emotion, 5, 81108.
Clark, K. B., Krahl, S. E., Smith, D. C., & Jensen, R. A. (1995). Post-training unilateral vagal stimulation enhances retention performance in the rat. Neurobiology
of Learning and Memory, 63, 213216.
Clark, K. B., Naritoku, D. K., Smith, D. C., Browning, R. A., & Jensen, R. A. (1999).
Enhanced recognition memory following vagus nerve stimulation in human
subjects. Nature Neuroscience, 2, 9498.
Cuenod, C. A., Denys, A., Michot, J. L., Jehenson, P., Forette, F., Kaplan, D., et al.
(1993). Amygdala atrophy in Alzheimers disease. An in vivo magnetic resonance imaging study. Archives of Neurology, 50, 941945.
Damasio, A. R. (1994). Descartes error and the future of human life. Scientic American, 271, 144.
Davis, M. (1997). Neurobiology of fear responses: The role of the amygdala. Journal
of Neuropsychiatry and Clinical Neuroscience, 9, 382402.
Davis, M., Walker, D. L., & Lee, Y. (1997). Roles of the amygdala and bed nucleus
of the stria terminalis in fear and anxiety measured with the acoustic startle
reex. Possible relevance to PTSD. Annals of the New York Academy of Science,
821, 305331.
de Kloet, E. R., Oitzl, M. S., & Joels, M. (1999). Stress and cognition: Are corticosteroids good or bad guys? Trends in Neuroscience, 22, 422426.
Denburg, N. L., Buchanan, T. W., Tranel, D., & Adolphs, R. (in press). Evidence for
preserved emotional memory in normal elderly persons. Emotion.
Dobbins, I. G., Kroll, N. E., Tulving, E., Knight, R. T., & Gazzaniga, M. S. (1998).
Unilateral medial temporal lobe memory impairment, type decit, function decit, or both? Neuropsychologia, 36, 115127.
Dolan, R. J., Lane, R., Chua, P., & Fletcher, P. (2000). Dissociable temporal lobe
activations during emotional episodic memory retrieval. Neuroimage, 11,
203209.
Eichenbaum, H., & Cohen, N. J. (2001). From conditioning to conscious recollection,
memory systems of the brain. Oxford: Oxford University Press.
Escobar, M. L., Chao, V., & Bermudez-Rattoni, F. (1998). In vivo long-term potentiation in the insular cortex, NMDA receptor dependence. Brain Research, 779,
314319.
Fink, G. R., Markowitsch, H. J., Reinkemeier, M., Bruckbauer, T., Kessler, J., &
Heiss, W. D. (1996). Cerebral representation of ones own past, neural networks involved in autobiographical memory. Journal of Neuroscience, 16, 4275
4282.
Frijda, N. H. (1987). Emotion, cognitive structure, and action tendency. Cognition
and Emotion, 1, 115143.

70

Gaan, D., Murray, E. A., & Fabre-Thorpe, M. (1993). Interaction of the amygdala
with the frontal lobe in reward memory. European Journal of Neuroscience, 5,
968975.
Goddard, G. V. (1964). Amygdaloid stimulation and learning in the rat. Journal of
Comparative and Physiological Psychology, 58, 2330.
Gold, P. E., Hankins, L., Edwards, R. M., Chester, J., & McGaugh, J. L. (1975). Memory
interference and facilitation with posttrial amygdala stimulation, eect on
memory varies with footshock level. Brain Research, 86, 509513.
Gold, P. E., & van Buskirk, R. (1975). Facilitation of time-dependent memory
processes with posttrial epinephrine injections. Behavioral Biology, 23, 509
520.
Golomb, J., de Leon, M. J., Kluger, A., George, A. E., Tarshish, C., & Ferris, S. H.
(1993). Hippocampal atrophy in normal aging. An association with recent
memory impairment. Archives of Neurology, 50, 967973.
Grady, C. L., & Craik, F. I. (2000). Changes in memory processing with age. Current
Opinion in Neurobiology, 10, 224231.
Hamann, S., & Mao, H. (2002). Positive and negative emotional verbal stimuli elicit
activity in the left amygdala. Neuroreport, 13, 1519.
Hamann, S. B., Cahill, L., McGaugh, J. L., & Squire, L. R. (1997). Intact enhancement of declarative memory for emotional material in amnesia. Learning &
Memory, 4, 301309.
Hamann, S. B., Cahill, L., & Squire, L. R. (1997). Emotional perception and memory
in amnesia. Neuropsychology, 11, 104113.
Hamann, S. B., Ely, T. D., Grafton, S. T., & Kilts, C. D. (1999). Amygdala activity
related to enhanced memory for pleasant and aversive stimuli. Nature Neuroscience, 2, 289293.
Hamann, S. B., Ely, T. D., Homan, J. M., & Kilts, C. D. (2002). Ecstasy and agony,
activation of the human amygdala in positive and negative emotion. Psychological Science, 13, 135141.
Hamann, S. B., Monarch, E. S., & Goldstein, F. C. (2000). Memory enhancement
for emotional stimuli is impaired in early Alzheimers disease. Neuropsychology,
14, 8292.
Heuer, F., & Reisberg, D. (1990). Vivid memories of emotional events, the accuracy
of remembered minutiae. Memory & Cognition, 18, 496506.
Honkaniemi, J., Pelto-Huikko, M., Rechardt, L., Isola, J., Lammi, A., Fuxe, K., et al.
(1992). Colocalization of peptide and glucocorticoid receptor immunoreactivities in rat central amygdaloid nucleus. Neuroendocrinology, 55, 451459.
Hyman, B. T., Van Hoesen, G. W., & Damasio, A. R. (1990). Memory-related neural systems in Alzheimers disease, an anatomic study. Neurology, 40, 1721
1730.
Ikegaya, Y., Nakanishi, K., Saito, H., & Abe, K. (1997). Amygdala betanoradrenergic inuence on hippocampal long-term potentiation in vivo.
Neuroreport, 8, 31433146.
Jack, C. R., Jr., Petersen, R. C., Xu, Y. C., Waring, S. C., OBrien, P. C., Tangalos,
E. G., et al. (1997). Medial temporal atrophy on MRI in normal aging and very
mild Alzheimers disease. Neurology, 49, 786794.

71

Jernigan, T. L., Archibald, S. L., Fennema-Notestine, C., Gamst, A. C., Stout, J. C.,
Bonner, J., et al. (2001). Eects of age on tissues and regions of the cerebrum
and cerebellum. Neurobiology of Aging, 22, 581594.
Kazui, H., Mori, E., Hashimoto, M., Hirono, N., Imamura, T., Tanimukai, S., et al.
(2000). Impact of emotion on memory, controlled study of the inuence of
emotionally charged material on declarative memory in Alzheimers disease.
British Journal of Psychiatry, 177, 343347.
Kim, J. J., & Diamond, D. M. (2002). The stressed hippocampus, synaptic plasticity
and lost memories. Nature Reviews Neuroscience, 3, 453462.
Klver, H., & Bucy, P. C. (1938). Psychic blindness and other symptoms following bilateral temporal lobectomy in rhesus monkeys. American Journal of Physiology, 5, 3354.
Kroll, N. E., Markowitsch, H. J., Knight, R. T., & von Cramon, D. Y. (1997). Retrieval
of old memories, the temporofrontal hypothesis. Brain, 120, 13771399.
LaBar, K. S., & Phelps, E. A. (1998). Arousal-mediated memory consolidation: Role
of the medial temporal lobe in humans. Psychological Science, 9, 490493.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1990). Emotion, attention, and the
startle reex. Psychological Review, 97, 377395.
LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review of Neuroscience,
23, 155184.
Levine, B., Black, S. E., Cabeza, R., Sinden, M., McIntosh, A. R., Toth, J. P., et al.
(1998). Episodic memory and the self in a case of isolated retrograde amnesia.
Brain, 121, 19511973.
Liang, K. C., Juler, R. G., & McGaugh, J. L. (1986). Modulating eects of posttraining
epinephrine on memory, involvement of the amygdala noradrenergic system.
Brain Research, 368, 125133.
Lovallo, W. R. (1997). Stress and health: Biological and psychological interactions.
Thousand Oaks, CA: Sage.
Lupien, S. J., & Lepage, M. (2001). Stress, memory, and the hippocampus, cant live
with it, cant live without it. Behavioral Brain Research, 127, 137158.
Maratos, E. J., Dolan, R. J., Morris, J. S., Henson, R. N. A., & Rugg, M. D. (2001).
Neural activity associated with episodic memory for emotional context. Neuropsychologia, 39, 910920.
Maratos, E. J., & Rugg, M. D. (2001). Electrophysiological correlates of the retrieval
of emotional and non-emotional context. Journal of Cognitive Neuroscience, 13,
877891.
Markowitsch, H. J., Calabrese, P., Wurker, M., Durwen, H. F., Kessler, J., Babinsky,
R., Brechtelsbauer, D., Heuser, L., & Gehlen, W. (1994). The amygdalas contribution to memoryA study on two patients with Urbach-Wiethe disease.
Neuroreport, 5, 13491352.
Markowitsch, H. J., Thiel, A., Reinkemeier, M., Kessler, J., Koyuncu, A., & Heiss,
W. D. (2000). Right amygdalar and temporofrontal activation during autobiographic, but not during ctitious memory retrieval. Behavioural Neurology,
12, 181190.
McEwen, B. S. (2000). The neurobiology of stress, from serendipity to clinical relevance. Brain Research, 886, 172189.

72

McEwen, B. S., & Sapolsky, R. M. (1995). Stress and cognitive function. Current
Opinion in Neurobiology, 5, 205216.
McGaugh, J. L. (1983). Hormonal inuences on memory. Annual Review of Psychology, 34, 297323.
McGaugh, J. L. (2000). MemoryA century of consolidation. Science, 287, 248251.
McGaugh, J. L., & Izquierdo, I. (2000). The contribution of pharmacology to research
on the mechanisms of memory formation. Trends in Pharmacological Science, 21,
208210.
Moayeri, S. E., Cahill, L., Jin, Y., & Potkin, S. G. (2000). Relative sparing of emotionally inuenced memory in Alzheimers disease. Neuroreport, 11, 653655.
Mori, E., Yoneda, Y., Yamashita, H., Hirono, N., Ikeda, M., & Yamadori, A. (1997).
Medial temporal structures relate to memory impairment in Alzheimers disease, an MRI volumetric study. Journal of Neurology, Neurosurgery, and Psychiatry, 63, 214221.
Mu, Q., Xie, J., Wen, Z., Weng, Y., & Shuyan Z. (1999). A quantitative MR study
of the hippocampal formation, the amygdala, and the temporal horn of the
lateral ventricle in healthy subjects 40 to 90 years of age. American Journal of
Neuroradiology, 20, 207211.
Murray, E. A. (1992). Medial temporal lobe structures contributing to recognition
memory: The amygdaloid complex versus the rhinal cortex. In J. P. Aggleton
(Ed.), The amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction (pp. 453470). New York: Wiley-Liss.
hman, A., & Mineka, S. (2001). Fears, phobias, and preparedness, toward an
evolved module of fear and fear learning. Psychological Review, 108, 483522.
Oitzl, M. S., & de Kloet, E. R. (1992). Selective corticosteroid antagonists modulate specic aspects of spatial orientation learning. Behavioral Neuroscience,
106, 6271.
Packard, M. G., Cahill, L., & McGaugh, J. L. (1994). Amygdala modulation of
hippocampal-dependent and caudate nucleus-dependent memory processes.
Proceedings of the National Academy of Science, USA, 91, 84778481.
Packard, M. G., & Teather, L. A. (1998). Amygdala modulation of multiple memory
systems, hippocampus and caudate-putamen. Neurobiology of Learning and
Memory, 69, 163203.
Parker, A., Wilding, E., & Akerman, C. (1998). The Von Restor eect in visual object recognition memory in humans and monkeys. The role of frontal/perirhinal interaction. Journal of Cognitive Neuroscience, 10, 691703.
Phelps, E. A., LaBar, K. S., & Spencer, D. D. (1997). Memory for emotional words
following unilateral temporal lobectomy. Brain and Cognition, 35, 85109.
Pitknen, A. (2000). Connectivity of the rat amygdaloid complex. In J. P. Aggleton
(Ed.), The amygdala: A functional analysis. Oxford: Oxford University Press.
Quirarte, G. L., Roozendaal, B., & McGaugh, J. L. (1997). Glucocorticoid enhancement
of memory storage involves noradrenergic activation in the basolateral amygdala. Proceedings of the National Academy of Science, USA, 94, 1404814053.
Reisberg, D., & Heuer, F. (1992). Remembering the details of emotional events. In
E. Winograd & U. Neisser (Eds.), Aect and accuracy in recall: Studies of ashbulb memories (pp. 163190). Cambridge, MA: Harvard University Press.

73

Rolls, E. T. (2000a). Memory systems in the brain. Annual Review of Psychology, 51,
599630.
Rolls, E. T. (2000b). The orbitofrontal cortex and reward. Cerebral Cortex, 10, 284
294.
Roozendaal, B. (2000). Glucocorticoids and the regulation of memory consolidation. Psychoneuroendocrinology, 25, 213238.
Roozendaal, B., Williams, C. L., & McGaugh, J. L. (1999). Glucocorticoid receptor
activation in the rat nucleus of the solitary tract facilitates memory consolidation, involvement of the basolateral amygdala. European Journal of Neuroscience,
11, 13171323.
Schmolck, H., Bualo, E. A., & Squire, L. R. (2000). Memory distortions develop over
time: Recollections of the O. J. Simpson trial verdict after 15 and 32 months.
Psychological Science, 11, 3945.
Schooler, J. W., & Eich, E. (2000). Memory for emotional events. In E. Tulving &
F. I. M. Craik (Eds.), The Oxford handbook of memory (pp. 379392). Oxford:
Oxford University Press.
Scott, S. A., DeKosky, S. T., & Sche, S. W. (1991). Volumetric atrophy of the amygdala in Alzheimers disease, quantitative serial reconstruction. Neurology, 41,
351356.
Scoville, W. B., & Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. Journal of Neurology, Neurosurgery, and Psychiatry, 20, 1121.
Seidlitz, L., & Diener, E. (1998). Sex dierences in the recall of aective experiences.
Journal of Personality and Social Psychology, 74, 262271.
Shepard, J. D., Barron, K. W., & Myers, D. A. (2000). Corticosterone delivery to the
amygdala increases corticotropin-releasing factor mRNA in the central
amygdaloid nucleus and anxiety-like behavior. Brain Research, 861, 288295.
Squire, L. (1987). Memory and brain. New York: Oxford University Press.
Squire, L. R., & Kandel, E. R. (2000). Memory: From mind to molecules. New York:
Scientic American Library.
Tomaz, C., Dickinson-Anson, H., McGaugh, J. L., Souza-Silva, M. A., Viana, M. B.,
& Grae, F. G. (1993). Localization in the amygdala of the amnestic action of
diazepam on emotional memory. Behavioral Brain Research, 58, 99105.
Tranel, D. & Hyman, B. (1990). Neuropsychological correlates of bilateral amygdala
damage. Archives of Neurology, 47, 349355
Tulving, E., Kapur, S., Craik, F. I., Moscovitch, M., & Houle, S. (1994). Hemispheric
encoding/retrieval asymmetry in episodic memory, positron emission tomography ndings. Proceedings of the National Academy of Science, USA, 91, 20162020.
Tulving, E., Kapur, S., Markowitsch, H. J., Craik, F. I., Habib, R., & Houle, S. (1994).
Neuroanatomical correlates of retrieval in episodic memory, auditory sentence
recognition. Proceedings of the National Academy of Science, USA, 91, 20122015.
Van Hoesen, G. W., Augustinack, J. C., & Redman, S. J. (1999). Ventromedial temporal lobe pathology in dementia, brain trauma, and schizophrenia. Annals of
the New York Academy of Science, 877, 575594.
Weiskrantz, L. (1956). Behavioral changes associated with ablation of the amygdaloid complex in monkeys. Journal of Comparative and Physiological Psychology,
49, 381391.

74

Yerkes, R. M., & Dodson, J. D. (1908). The relation of strength of stimulus to rapidity of habit formation. Journal of Comparative Neurology and Psychology, 18, 31
39.
Zola-Morgan, S., Squire, L. R., Alvarez-Royo, P., & Clower, R. P. (1991). Independence of memory functions and emotional behavior, separate contributions of
the hippocampal formation and the amygdala. Hippocampus, 1, 207220.
Zola-Morgan, S., Squire, L. R., & Amaral, D. G. (1986). Human amnesia and the
medial temporal region, enduring memory impairment following a bilateral
lesion limited to eld CA1 of the hippocampus. Journal of Neuroscience, 6, 2950
2967.
Zola-Morgan, S., Squire, L. R., & Amaral, D. G. (1989). Lesions of the amygdala that
spare adjacent cortical regions do not impair memory or exacerbate the impairment following lesions of the hippocampal formation. Journal of Neuroscience,
9, 19221936.

75

. , , . ,
.

rauma aects memory. We have known that for millennia, but only
recently have we been able to ask and answer pointed questions
about why and how. Only by elucidating the neurological consequences of
trauma, which until recently had been deemed a purely psychological event,
can we unravel its impact on memory. In this chapter we (1) briey discuss the
nature and organization of memory; (2) attempt to dene the notion of trauma
and the related phenomenon of stress; (3) consider the impact of trauma and
stress on the brain and, by extension, on the memory functions subserved by the
brain; (4) oer an integrated account connecting these data with several clinical manifestations of trauma, stress, and memory; and (5) use this account to
resolve an apparent tension between two parallel literatures: one describing
a trauma- or stress-based impairment of memory and another describing an
emotion-based enhancement of memory.

The Nature of Memory:

A Taxonomy

One of the key insights of recent cognitive- and neuroscience-based analyses of

memory is that multiple systems underlie this function. Unlike earlier perspectives, which viewed memory as a single system, perhaps subserved by a restricted
part of the brain, the modern consensus view is that several types of memory,
each obeying rather dierent rules of operation, and each subserved by discrete
neural systems, interact to produce our subjective sense of remembering. This
insight is critical if we are to understand the impact of any experience, normal
or traumatic, on memory, because it raises the possibility that dierent kinds of
memory will face dierent fates in the presence of trauma.

76

76

We neednt go into much detail about the nature of these multiple memory
systems or the types of information they process. Various views suggest how to
classify and arrange these systems (cf. Schacter & Tulving, 1994); most agree
on a distinction between two broad classes of memory. First, there are memories of the events in our lives and the knowledge of the world that we obtain from
those events. Typically, this class of memory, referred to as explicit memory, can
be explicitly retrieved on demand. Second, there are memories for the skills, procedures, and habits we acquire through experience. These implicit memories are
not so readily made explicit; indeed, in most instances these memories cannot
be expressed explicitly and are evident only in behavior.
A distinction between explicit and implicit memory is important but not sucient for understanding the impact of trauma on remembering. Within explicit
memory, there is a further crucial subdivision between episodic and semantic
memories. Episodic memory incorporates the specic context of an experienced
event, including the time and place of its occurrence. Semantic memory, on the
other hand, is concerned with the knowledge one acquires during events but is
itself separated from the specic event in question. Thus, our knowledge about
the meaning of words, and facts about the world, though acquired in the context of some specic experience, appears to be stored in a form that is not bound
to the originating context. The sensory features of experiences, their positive and
negative attributes, other objects with which they might have been associated
knowledge of this kind constitutes a part of semantic memory. The exact relation between episodic and semantic memory remains to be unraveled; however,
there is considerable agreement that they are separable in the brain. Evidence
suggests a third, emotional memory, system that mediates the encoding and
storage of emotionally charged events. This system pertains to learning about
fearful and unpleasant stimuli, although some evidence suggests it plays a role
in memory for pleasant information as well (see chapter 2, this volume; Hamann,
Ely, Grafton, & Kilts, 1999).
Apparently, each of these memory systems is subserved by a dierent brain
area, and, consequently, each is open to dierential impact by exposure to stressors and to trauma. Episodic memory seems to be the most inuenced by trauma
and stress and is neuroanatomically linked to the hippocampal formation in the
medial temporal lobe. But before embarking on a description of the hippocampal system and detailing its response to stress, we rst dene stress and trauma.

What Are Stress and Trauma?

Despite the widespread use of stress and trauma, these terms lack precise denitions. Although stressor is unambiguous according to the Oxford English
Dictionary (OED), which gives it a single denition, a single condition or agent
that constitutes a stress for an organism, the term stress is utterly ambiguous.

77

The OED provides at least nine distinct senses of the word when it is used as a
noun. These range from stress dened as emphasis, to stress dened as physical hardship, strain or pressure. Furthermore, the OED provides at least ve distinct meanings for stress used as a verb.
Trauma is similarly ambiguous. Although the OED considers it only as a
noun, it gives ve distinct senses for its use, including two that are germane
here:
1. Pathology. A wound, or external bodily injury in general; also the condition caused by this.
2. Psychoanalysis and Psychiatry. A psychic injury, especially one caused
by emotional shock the memory of which is repressed and remains unhealed; an internal injury, esp. to the brain, which may result in a behavioral disorder of organic origin. Also, the state or condition so caused.

In the scientic literature, both trauma and stress are often treated as static
things, as if they were well-dened entities characterized by psychophysiological measures that identify and quantify them (Kim & Diamond, 2002). Such is
clearly not the case, at least not at this stage in our understanding. Both stress
and trauma are events, that is, dynamic processes associated with various physical and psychological responses that can dier in magnitude and expression.
A common working denition of stress is any event that seriously disturbs the
physiological or psychological homeostasis of an organism. Cannon (1929) was
the rst to use homeostasis to characterize processes that are essential to the
maintenance of life and to highlight the bodys ability to control its internal
environment within narrow limits. Selye (1956) initiated the rst major study
of the eects of stress on systemic regulation, dening stress as any severe threat
to homeostasis that could result in death. More recently, however, McEwen
(2000, 2001) has suggested that because homeostasis applies to a limited number of systems (such as pH, body temperature) that are essential for survival,
another term, allostasis, may better describe the processes disturbed during
stress. Allostasis refers to the systems that actively maintain homeostasis even
when life and death are not at stake.
Kim and Diamond (2002) argue that a good denition of stress must address
the following two facts. First, stress is determined not by a given environmental
situation but by how an organism perceives and reacts to a situation. This denition permits individual dierences in response to stressors, as well as variations
in response to a xed stressor over time and experience within an individual. The
need to take experience, personal biases, and temperament into account should
be obvious to anyone who has learned to engage in public speaking; at rst a
nerve-racking experience, for most, public speaking eventually becomes easier
and less anxiety-provoking. Second, there is no physiological state that always
denes stress. Although most investigators rely on elevations of stress hormones

78

(e.g., glucocorticoids) as an indicator of stress, such hormones can also increase

in response to physical exertion and daily changes in the sleep/wake cycle (e.g.,
Kanaley, Weltman, Pieper, Weltman, & Hartman, 2001; Plihal & Born, 1999;
Van Cauter & Turek, 2001).
With these limitations in mind, Kim and Diamond (2002) presented a threepart denition of stress, which they claim is broadly applicable across species and
paradigms. They argue that, rst, stress is always accompanied by high levels of
physiological arousal, as measured by behavior (motor behaviors) or neurochemistry (increased stress hormones). Second, stress must be perceived as
threatening and something that would be avoided if possible. For example,
giving a speech is considered threatening, and many people would avoid it if
possible. When the speech cannot be avoided, however, the task is accompanied
by increased physiological excitability (heart rate, sweating, glucocorticoid increases, etc.). Finally, stress depends on whether an organism perceives that it
has control over the stressful experience. This component reects evidence from
animal and human studies demonstrating that perceived control can profoundly
mitigate the experience of stress (e.g., Fox & Dwyer, 2000; Maier & Watkins,
1998; Prince & Anisman, 1990).
We agree with the rst component of the denition but take issue with the
second and third components. We suggest that stressor should not always
mean something negative. Instead, it should simply designate a dynamic strain
or pressure exerted on an individual. This strain or pressure could be pleasant
(e.g., skydiving, sexual activity) or unpleasant (e.g., a loss of social status), but
in both cases it disturbs the individuals allostatic balance. Further, the data
suggest that controllability can sometimes mitigate the action of a stressor but
at other times has no impact at all.
People often have control over events they experience as stressful; giving
speeches and willingly jumping out of airplanes are two good examples. Thus,
although controllability is certainly important and may, in some cases, both
mitigate the perceived negativity of stress and determine its magnitude, we assert that stress reactions, even intense stress reactions, occur in the presence of
perceived control.
For our purposes, then, we use stress to designate activation of a dynamic
physiological statea system that, according to current knowledge, consists of the
hypothalamic-pituitary-adrenal (HPA) axis. We use the activation of this system
as a proxy for stress, bearing in mind the noted qualication that the mere presence of this physiological state does not indicate inevitable stress. Further details
of the physiology of stress will be provided in the section that follows. Moreover,
we use the term stressor to designate the environmental conditions that elicit
or inuence this physiological system, recognizing the need to specify the formal
relations between given environmental stressors and the stress response.
We do not consider psychological trauma the equivalent of intense stress.
Although stress coupled with environmental variables may produce a traumatic

79

psychological state, and a traumatic psychological state may induce stress (i.e.,
serve as a stressor), a traumatic state is, by our reckoning, a quasi-independent
mental state that systematically interacts with stress and stressors.1

Eects of Stress: Brain Physiology

and Anatomy

As mentioned, stressors activate the HPA axis. Although the stress response
system is complex, to date much of the research focus has been on the HPA axis
and the hippocampus. Stress leads to the release of corticotropin-releasing factor
(CRF) from the hypothalamus and subsequently triggers adrenocorticotropic
hormone (ACTH) release from the pituitary gland. In response to ACTH, stress
hormones called glucocorticoids are secreted and released from the adrenal
glands. Circulating glucocorticoids then feed back onto the pituitary and hypothalamus to inhibit the secretion of CRF and ACTH. The hippocampus (and
certain cortical areas, particularly the frontal cortex) plays an important role in
this negative feedback system as well, helping the pituitary and hypothalamus
inhibit HPA activity (Feldman & Conforti, 1980, 1985; Jacobson & Sapolsky,
1991; Sapolsky & Meaney, 1986).
Hippocampal involvement in the regulation of the stress response turns out
to be critically important to any discussion of stress and memory, because, as
we noted, this structure has long been implicated in explicit memory. In particular, the hippocampal formation has been linked to episodic memory, to context,
and to spatial maps (e.g., OKeefe & Nadel, 1978). Debate rages over aspects of
this story, but the link between the hippocampal system and episodic memory
seems rm. The seminal studies of Scoville and Milner (1957), for example, made
it clear that bilateral ablation of the hippocampal complex resulted in profound
loss of memory. We now know that damage to the hippocampus results in a
specic form of memory dysfunction, in particular that involving spatial/contextual and episodic memories (Hirsch, 1974; Nadel, 1991, 1994; Nadel &
Willner, 1980; Rosenbaum et al., 2000), and, further, that stress appears to
impair precisely these types of memory (see Lupien & McEwen, 1997). Together
with the discovery of dense concentrations of stress hormone receptors in the
hippocampus (e.g., McEwen, Weiss, & Schwartz, 1968), these ndings provide
a basis for the claim that stress hormone modulation of hippocampal activity
underlies the damaging eects of stress in animals and humans (see Bremner,
1999; McEwen, 2000, for reviews).
Thus, increases in stress hormones not only trigger inhibition of HPA activity via hippocampal mediation but at high enough levels also impair the neuronal structure and function of the hippocampus. Glucocorticoids at abnormally
high or abnormally low concentrations can, by changing the physiological characteristics of large populations of neurons, disrupt hippocampal function. Thus,

80

a moderate level of circulating glucocorticoids appears to enhance memory,

whereas maximal or minimal levels appear to disrupt it. The stress response typically involves a dramatic increase in the level of circulating glucocorticoids,
which can produce changes in hippocampal neuronal physiology and anatomy.
In short, stress appears to alter hippocampal morphology, disrupt neurogenesis,
and block synaptic plasticity thought to underlie memory formation (see McEwen,
2000, 2001, for reviews).
For example, both glucocorticoid exposure and restraint stress over 21 days
have been shown to result in reduced dendritic branching in pyramidal neurons
of the CA3 region of the rat hippocampus (Magarinos & McEwen, 1995; Watanabe,
Gould, & McEwen, 1992; Woolley, Gould, & McEwen, 1990). Psychosocial stress
leads to atrophy of dendrites as well, in the tree shrew (Magarinos, McEwen, Flugge,
& Fuchs, 1996) and in rats (McKittrick et al., 2000). In addition, stress and glucocorticoid elevation can produce alterations in synaptic terminal structure in CA3
(Magarinos, Verdugo, & McEwen, 1997).
With even longer periods of exposure to stress and glucocorticoids, initial
evidence suggested that overt loss of hippocampal neurons occurred (e.g.,
Sapolsky, Krey, & McEwen, 1985; Uno, Ross, Else, Suleman, & Sapolsky, 1989).
Such ndings led to assumptions that neuronal death associated with glucocorticoid release might underlie the severe memory impairments seen in posttraumatic stress disorder (see later discussion). More recent work, however, raises
questions about whether glucocorticoids are themselves toxic or simply create
conditions for other factors to exert neurotoxic eects (Sapolsky, 1996). Nonetheless, chronic exposure to stressors is undoubtedly associated with hippocampal impairment. For example, post-mortem analysis of chronically stressed
monkeys showed extensive damage in the CA3 region of the hippocampus; subsequent research suggested that high levels of the glucocorticoid cortisol were
responsible for the hippocampal damage, as cortisol-secreting pellets implanted
in the monkey hippocampus resulted in the same pattern of damage.
In addition to having neurotoxic eects on hippocampal neurons, chronic exposure to stressors may also inhibit the growth of new neurons (granule cells)
in the dentate gyrus, thereby changing its structure. A stress-induced decrease
in neurogenesis has been demonstrated in the dentate gyrus of rats, tree shrews,
and primates (Gould, Tanapat, McEwen, Flugge, & Fuchs, 1998; Gould &
Tanapat, 1999; Tanapat, Hastings, Rydel, Galea, & Gould, 2001) and has been
hypothesized as the primary mechanism of stress-related hippocampal impairment (Gould & Gross, 2002; Gould & Tanapat, 1999). Glucocorticoids appear
to mediate these stress-induced changes in neurogenesis. For example, normalizing glucocorticoid levels was shown to prevent predator odor stress from inhibiting granule cell production (Tanapat et al., 2001). In addition, corticosterone
treatment has been shown to decrease dentate cell proliferation, whereas adrenalectomy increases the production of new granule cells (Cameron & Gould, 1994;
Cameron & McKay, 1999).

81

In addition to these structural eects, stress and stress hormones also modulate the excitability of hippocampal neurons. Acute stress and glucocorticoid
release can impair hippocampal plasticity (Diamond, Fleshner, & Rose, 1994;
Foy, Stanton, Levine, & Thompson, 1987). Long-term potentiation (LTP) and
the related phenomenon of primed burst potentiation (PBP) are two processes
by which hippocampal neuronal responsiveness changes in response to experience (McEwen, 2001). Thus, these forms of potentiation are often conceptualized as physiological models of memory. Studies show that hippocampal LTP can
be impaired by the administration of glucocorticoids (e.g., Diamond & Rose,
1994; Filipini, Gijsbers, Birmingham, & Dubrovsky, 1991). Diamond, Bennet,
Fleshner, and Rose (1992) reported an inverted U-shaped relationship between
the level of circulating glucocorticoids and the extent of LTP, demonstrating that
glucocorticoids facilitated LTP at low levels but disrupted it at high levels. Stressinduced disruption of LTP targets CA1 and the dentate gyrus and has been shown
to last up to 48 hours in rats and 24 hours in mice (e.g., Garcia, Musleh, Tocco,
Thompson, & Baudry, 1997; Shors, Gallegos, & Breindl, 1997). These studies
show that LTP can be disrupted by stress and that changes can last for days
after exposure to a single stressor; however, these changes are typically not permanent. More recent studies suggest that PBP may be even more sensitive to
psychological stress (e.g., predator exposure) than is LTP (Mesches, Fleshner,
Heman, Rose, & Diamond, 1999).

Stress and Brain Anatomy:

Evidence for Hippocampal

Volume Reduction

Given the evidence of hippocampal impairment in response to stress, one might

expect to see gross structural changes in the hippocampi of certain patient populations. Indeed, the human hippocampus does appear to show signs of atrophy as a
result of severe stress or persistently elevated glucocorticoids. Reductions have been
noted in disorders associated with prolonged glucocorticoid elevations, such as
Cushings syndrome, recurrent major depression, schizophrenia, posttraumatic
stress disorder, and in some cases of normal aging (Starkman et al., 1999; Sheline
et al., 1996; Bremner et al., 1993; Bremner, Randall, Capelli, et al., 1995; Gurvits
et al., 1996; Fukuzako et al., 1996; Convit et al., 1995). For example, Cushings
syndrome arises from adrenocorticotropic hormone or corticotropin-releasing
hormone secreting tumors (Sapolsky, 2000). Those with the disorder thus
have hypercortisolemia and often exhibit cognitive dysfunction (Sapolsky, 2000;
Starkman et al., 1999). Using magnetic resonance imaging (MRI), researchers
have reported selective hippocampal atrophy in individuals suering from
Cushings syndrome for approximately 1 to 4 years (Starkman, Gebarski, Berent,
& Schteingart, 1992). Although a trauma control group was not tested in this
study, hippocampal reduction was strongly correlated with cortisol levels.

82

MRI studies of PTSD have revealed a pattern of evidence many consider

indicative of hippocampal atrophy. When Vietnam veterans with combat-related
PTSD were compared to healthy control subjects, the PTSD patients were found
to have an 8% reduction in right hippocampal volume (Bremner, Randall, Scott,
et al., 1995). This dierence in volume was associated with short-term memory
impairments as assessed by the WMS-Logical memory test. Similar results were
obtained in childhood abuse survivors (physical and sexual) with PTSD (Bremner
et al., 1997). In this sample, however, left hippocampal volume was 12% smaller
than that of controls. Although right hippocampi were 3.8% smaller than those
of controls, this dierence did not reach statistical signicance. In both studies,
control subjects were matched for sex, age, race, height, weight, socioeconomic
status, education, and alcohol abuse. Other studies nding similar dierences in
hippocampal volume include Gurvits et al. (1996) (26% smaller bilateral volume in Vietnam combat veterans with PTSD compared to non-PTSD combat veterans) and Stein, Koverola, Hanna, Torchia, & McClarty (1997) (5% smaller left
hippocampal volume in women who reported sexual abuse than in controls).
Concerns about the reliability and interpretability of these studies, which
center on the co-occurrence of hippocampal atrophy with other disorders (e.g.,
depression, substance abuse), led Bremner and other researchers to revisit their
dataset and statistically exclude those subjects who had abused alcohol or other
substances during the 6 months prior to their MRIs. Despite these changes, results remained suggestive of a relationship between PTSD and reduced hippocampal volume not explainable by alcohol or drug abuse (see Bremner, 2001b).
Though in general MRI studies are suggestive of stress-related hippocampal
atrophy, interpretation of the results is compromised by cross-sectional designs
and the use only of subjects with chronic PTSD. Consequently, one cannot use
these results to infer causality. Reduced hippocampal volume could either precede the traumatic event, making individuals vulnerable to developing PTSD,
or could develop in the aftermath of the trauma as a consequence of prolonged
exposure to stress hormones or other uncontrolled factors.
To correct for these confounds, Bonne et al. (2001) conducted a prospective
MRI study examining the hippocampal volume of individuals exposed to acute
traumas (traumas experienced by subjects admitted to a hospital emergency
room). The hippocampal volume of those who developed PTSD 6 months later
did not dier from those who had not developed PTSD at the 6-month followup. Although these results are inconsistent with earlier studies of hippocampal
volume reductions and PTSD, they may be explained by dierences in the nature
or duration of trauma. The earlier studies tested individuals exposed to chronic
traumatization, such as child abuse or war experience, whereas Bonne et al.
tested individuals exposed to acute stressors, perhaps of lesser severity.2 Another
possibility is that the earlier studies failed to screen tightly for confounding factors such as alcoholism and depression, although Bremner (2001b) attempted
to do so. In any case, the conservative conclusion is that tentative evidence

83

exists for reduced hippocampal volume in PTSD. Further studies that better control for comorbidity, type and duration of trauma, and duration of PTSD are
needed. Nonetheless, stress-induced decits in memory may exist in the absence
of gross neuroanatomical changes, reecting either functional or subtle structural impairments.

Eects of Stress on Memory: Data

We have shown that exposure to trauma may result in damage to brain regions
important for memory function. Now we turn to evidence supporting a relationship between stressors and memory function itself, a relationship with a long
history. Stress-related memory impairments have been reported in a variety of
clinical settings. As we mentioned, patient populations with chronically elevated
levels of stress hormones, such as Cushings syndrome, schizophrenia, some
forms of recurrent depression, PTSD,3 dementia of the Alzheimers type, and
asthmatic patients treated with the glucocorticoid prednisone (Keenan, Jacobson, Soleymani, & Newcomer, 1995; Mauri et al., 1993; Rasmusson et al., 2001;
Sapolsky, 2000; Sheline, Sanghavi, Mintun, & Gado, 1999; Starkman et al.,
1999), are characterized by impaired memory function. However, despite consistent correlations between chronic elevations in stress hormones and memory
impairment, these data are limited by the problems endemic to all naturalistic
studies. Because nonrepresentative sampling, possible noncausal associations,
and confounding variables such as extraneous disease factors may be responsible for memory impairments (e.g., Sapolsky, 2000), these observations must
be treated with caution. The correlations do, however, serve as a starting point
and encourage us to take a closer look at the nature of memory and its interactions with stress and trauma.
A growing body of randomized, placebo-controlled studies, using reasonably
precise denitions of stress, has demonstrated that exposure to uncontrollable
stressors can impair memory. This research has proceeded along two separate
but related paths, one dedicated to the study of relations between episodic
memory and long-term/chronic stress and the other to relations between episodic memory and short-term/acute stress.
The ethics inherent in human research make experimental studies of relations
between memory and chronic stressors dicult if not impossible, leaving the eld
to naturalistic research designs. Experimental studies of relations between
memory and acute stressors, such as stressful speeches or participation in highrisk tasks (e.g., skydiving), though dicult ethically, are at least possible. Finally,
the ethics governing animal research permit the induction of short-term (acute)
and long-term or repeated (chronic) stress using stressors such as tail-shock,
restraint, swimming in cold water, and more naturalistic stressors such as

84

predator exposure. We consider each of these literatures briey in the following sections.
The administration of stress-related hormones permits well-controlled studies of relations between memory and acute stress in humans. Acute doses of stress
hormones such as glucocorticoids (e.g., cortisol) can be safely administered to
humans over relatively brief periods. In some studies, agonists (corticocortisone)
or antagonists (e.g., dexamethasone) of naturally occurring glucocorticoids are
used instead of cortisol. Glucocorticoid antagonists reduce the level of circulating
glucocorticoids, and both abnormally high and abnormally low levels of glucocorticoids have been shown to produce memory decits (Conrad & Roy, 1996;
Luine, Villegas, Martinez, & McEwen, 1994; Vaher, Luine, Gould, & McEwen,
1994).

Acute Stressors: Memory in Humans

Inducing acute stress in the laboratory typically involves exposing the subject,
whether animal or human, to a brief, one-time stressor (e.g., a speech task, a
predators odor) or administering a single dose of glucocorticoids. In both cases,
glucocorticoid levels increase, usually for no longer than several hours in
humans. However, the cognitive eects of acute stress or glucocorticoid administration are often apparent for several hours to as long as a day; moreover, they
are usually reversible and are often specic to the task or experimental situation
(McEwen, 2001). These eects are typically specic to memory function, with
spatial and verbal-episodic memory often showing the greatest impairment.
Wolkowitz, Reus, and Weingartner (1990) and Wolkowitz et al. (1993), for
example, examined recall of previously learned verbal-episodic material after the
administration of dexamethasone or prednisone and found that these glucocorticoids interfered with word recall. In the dexamethasone study, participants
learned a list of 12 semantically associated words, which they were rst asked
to recall following a 90-second distractor task and then asked to recognize in a
larger list of related distractor words. Participants were tested 1 week before and
1 day after administration of 1 mg of dexamethasone. Although this single dose
of dexamethasone did not change recall, it did increase errors of commission, or
false recognition of related words not presented on the study list.
The prednisone study used a similar test of memory, but here both recall and
recognition of previously learned words were required 24 hours after presentation of the list, a manipulation that created a test of longer-term episodic memory.
Participants were given 80 mg of prednisone daily for 5 days. Word recognition
and recall were tested once during an initial 5-day placebo period, again after
4 days of prednisone administration, and once again 7 days after discontinuation of the treatment. Again, participants receiving prednisone showed no dierences in free recall but falsely recognized more semantically related words than

85

did controls. Thus, administration of dexamethasone and prednisone resulted

in diculties recognizing words presented in standard word lists (see also Bender,
Lerner, & Poland, 1991), a diculty characterized by an impaired ability to discriminate previously studied information from new, but similar, information.
Newcomer, Craft, Hershey, Askins, and Bardgett (1994) also investigated the
eects of dexamethasone on measures of memory, as well as attention and perceptual function. Dexamethasone treatment administered over 4 days (in successive doses of 0.5, 1, 1, and 1 mg) was associated with impaired recall of verbal
material originally presented in a paragraph (Wechsler, 1945). This impairment
was strongest on the 4th day of treatment and was specic to memory for verbal material. The presence of dexamethasone did not aect measures of attention or perception.
Unlike Wolkowitz et al. (1990, 1993), Newcomer et al. (1994) found that
dexamethasone treatment resulted in errors of omission (information in the
paragraph was forgotten) but not errors of commission (related but nonpresented
information was added to the material presented in the paragraph). The dierences between the Wolkowitz et al. (1990) and Newcomer et al. studies are likely
due to dierences in the strategies required by the two tasks. The recall task
Newcomer et al. used required the verbatim recall of story-like information presented in paragraph format. Given these instructions, normal subjects usually
make more omission than commission errors, as thematically related intrusions
errors rarely occur in this type of task. By comparison, semantically associated
word lists, like the ones used in Wolkowitz et al. (1990), are known to produce
commission errors at high rates (see Roediger & McDermott, 1995). We have
recently shown that psychologically induced stress increases commission errors
on semantically associated word lists (Payne, Nadel, Allen, Thomas, & Jacobs,
2002), a result we will discuss in more detail.
Results of the rst few experimental studies indicate that pharmacological
doses of the glucocorticoids dexamethasone and prednisone impair verbal episodic memory. However, caution should be used when thinking of this memory
impairment as causally related to stress-induced increases in cortisol. Although
changes in recall and recognition could result from increased glucocorticoid
receptor binding in a manner similar to the direct administration of cortisol
(the traditional explanation), an alternative explanation is possible. Dexamethasone suppresses cortisol, and because very high or very low levels of
cortisol can aect the biological structures on which memory depends, the
observed changes could result from dexamethasone-induced decreases in cortisol secretionand, hence, a failure to bind to glucocorticoid receptors. This
uncertainty, along with dierences in the binding of dexamethasone and cortisol to type I and type II glucocorticoid receptors in the hippocampus, limits
our condence in the generality of conclusions reached when dexamethasone
and similar ligands are used to examine the eects of cortisol or stress in humans (Newcomer et al., 1999).

86

Similar concerns motivated a second study by Newcomer et al. (1999), which

examined interactions between hydrocortisone (cortisol) and explicit memory
for words, again using a paragraph recall task. Two xed oral doses of hydrocortisone (160 mg/d or 40 mg/d) were administered over 4 days. The lower-dose
treatment was selected to approximate physiological levels of cortisol associated
with mild to moderate stress, and the higher-dose treatment was selected to
approximate physiological levels of cortisol associated with severe stress. Sampling of plasma cortisol and memory testing were conducted at baseline (day 0),
on days 1 and 4, and after a 6-day washout period (day 10). Impaired paragraph
recall was reported in the presence of the higher, but not the lower, dose of hydrocortisone. As in their previous study (Newcomer et al., 1994), this eect was
maximal on the 4th day of treatment. Newcomer et al. (1999) did not observe
the immediate eects of glucocorticoids seen in some studies (e.g., Wolkowitz
et al., 1990) but instead showed that several days of cortisol exposure at doses
and plasma concentrations associated with high stress could impair the recall
of verbal material presented in a paragraph format.
An earlier study by Kirschbaum, Wolf, Wippich, and Hellhammer (1996),
however, reported that a single, low dose of hydrocortisone (10 mg) led to a decit
in verbal episodic memory. In this study, subjects who received hydrocortisone
recalled fewer words (via a cued recall test) from a previously learned word list
than did control subjects when recall occurred 60 minutes after receiving the
drug. Thus, hydrocortisone apparently can produce a decit in at least one type
of verbal memory (cued word recall) at smaller doses and with shorter delays
between administration and test than suggested by the results of the Newcomer
et al. (1999) experiment. Again, the dierences in these two experiments might
depend on the nature of the material recalled, the delay between learning and
test, or dierent task demands. Although the details vary from experiment to
experiment, the administration of drugs simulating acute stress clearly changes
the recall pattern of words presented in word lists or paragraphs, both of which
measure aspects of explicit verbal memory.
Many of the glucocorticoid-administration studies are limited in that
the physiological eects of cortisol are longer-lived than those of the standard
memory experiment (which typically lasts no longer than an hour or two). Consider the designs used by Kirschbaum et al. (1996) and Newcomer et al. (1999)
already described. In each case cortisol levels were probably elevated continuously throughout the study, making it dicult to determine what aspect of
memory was impaired. In an attempt to sort this out, de Quervain, Roozendaal,
Nitsch, McGaugh, and Hock (2000) administered 25 mg doses of cortisone
(which is quickly absorbed and transformed into cortisol) at dierent stages of
a verbal delayed free-recall experiment. In this way they could evaluate the
impact of elevated cortisol on each of three distinct memory processes: acquisition, storage, and retrieval. Participants learned word lists composed of unrelated
words and received cortisone either (1) before list learning, (2) immediately after

87

list learning, or (3) 24 hours later, but immediately before recall. Impairment in
recall was found only when cortisone was given immediately before recall, thus
implicating retrieval as the memory stage most aected by a low dose of cortisol.
Although a 25 mg dose of hydrocortisone administered at encoding did not
aect recall of verbal material presented in paragraphs (Newcomer et al., 1999),
the same low dose of cortisone did aect the retrieval of words presented in lists
(de Quervain et al., 2000). This discrepancy could have arisen from dierences
in the kind of information recalled (stories vs. word lists), or it might reect a
greater impact of stress on retrieval than on encoding or consolidation.
In addition to explicit memory for words, spatial memory also appears to suer
in the presence of stress. To our knowledge, spatial memory in humans has not
yet been assessed with the direct administration of glucocorticoids. However,
Laurance, Hardt, Nadel, and Jacobs (2001) found that exposure to a social stressor (public speaking) slows the rate at which spatial information is acquired in
a virtual environment. The impact of stress in this study was specic to spatial
map-based performance, whereas performance based on individual cues, and
therefore dependent on dierent neural circuits, remained intact. Public speaking tasks have also resulted in verbal episodic decits, as measured by listlearning tasks, in healthy elderly subjects (Lupien et al., 1997) and in younger
adults (Kirschbaum et al., 1996).

Acute and Chronic Stressors:

Memory in Animals

Animal studies of stress and memory yield results that are generally consistent
with the human studies; exposure to glucocorticoids, acute or long-term, is often
associated with deleterious eects on memory tasks dependent on hippocampal
integrity (e.g., Bodno et al., 1995; Luine et al., 1994; Ohl & Fuchs, 1999; see
Kim & Diamond, 2002; Lupien & Lepage, 2001, for reviews). As in humans, however, acute stress eects are often short-term and reversible, whereas repeated
or long-term stress eects can lead to frank cognitive dysfunction (McEwen &
Sapolsky, 1995) that may result from the changes in hippocampal morphology
already described.
Administering corticosterone to rats facilitates the extinction of forms of
conditioned shock-avoidance that depend on the hippocampal formation (e.g.,
Bohus & Lissak, 1968; Greidanus, 1970; Roozendaal & McGaugh, 1996).
Spatial memory, also known to depend on the hippocampus, is impaired in rats
exposed to stressors or given corticosterone and in transgenic mice with elevated
corticosterone levels. For example, 21 days of restraint was correlated with impairments in spatial memory as measured by performance in the radial arm maze
task (Luine et al., 1994). Exposure to stressors has also been shown to disrupt
hippocampal-dependent object recognition (e.g., Clark, Zola, & Squire, 2000) and

88

working memory (Diamond, Fleshner, Ingersoll, & Rose, 1996; Diamond, Park,
Heman, & Rose, 1999). Likewise, long-term subcutaneous implants of glucocorticoids, which mimic the eects of chronic stress, impair new learning and
memory for maze escape behaviors in rats. In one study, the severity of such
decits was correlated with the number of damaged cells in the CA3 region of
the hippocampus (Arbel, Kadar, Silberman, & Levy, 1994). A thorough review
of the eects of stress in animals is beyond the scope of this chapter, but the reader
is directed to excellent reviews (de Kloet, Oitzl, & Joels, 1999; Kim & Diamond,
2002; Lupien & Lepage, 2001; Lupien & McEwen, 1997; McEwen, 2000).
Collectively, ndings from these studies converge on the idea that elevated
levels of glucocorticoids released during stress disrupt normal hippocampal
physiology, which, in turn, leads to changes in specic forms of memory. These
eects do not, however, occur alone. Instead, they occur in the context of a larger
stress response. Other neurochemical systems interact with glucocorticoids to
mediate the eects of stress on memory and hippocampal function, including
norepinephrine, serotonin, dopamine, amino acids (e.g., GABA, glutamate), and
immune compounds. Because many reviews focus exclusively on the glucocorticoids, a brief mention of the other systems will be appropriate here.
Exposure to stress causes the release of norepinephrine from the locus
coeruleus (a nucleus in the pontine region of the brainstem), and this is often
thought of as the rst wave response to stress. The norepinephrine system
responds almost immediately to a stressor, whereas the HPA axis, or second
wave response, reacts more slowly. Norepinephrine binds to receptors throughout the brain and facilitates motor responsiveness, thus governing the ight or
ght response (e.g., Stoddard, Bergdall, Townsend, & Levin, 1986). Moreover,
norepinephrine has potentiating eects on memory for emotional experiences,
a nding that may be mediated by the numerous norepinephrine receptors
located in the amygdala (e.g., McGaugh, 2000). Norepinephrine can reach very
high levels and still facilitate memory for emotion, even while similarly high
levels of glucocorticoids impair contextual and verbal episodic memory.
Exposure to stress also produces serotonin release in the hippocampus, medial
prefrontal cortex (PFC), and amygdala, among other brain regions (e.g., Inoue,
Tsuchiya, & Koyama, 1994; Kaehler, Singewald, Sinner, Thurnher, & Philippu,
2000), and stress-induced alterations in serotonin impact memory function. For
example, substances that reduce extracellular serotonin levels appear to block
the eects of stress on memory and hippocampal function, suggesting that serotonin release during stress might contribute to hippocampal damage (McEwen
et al., 1997; McEwen, 2000). Tianeptine, an antidepressant that works by
enhancing serotonin reuptake (and therefore reducing extracellular serotonin
levels), blocks both stress- and corticosterone-induced dendritic atrophy of neurons in the CA3 region of the hippocampus (Watanabe, Gould, Daniels, Cameron,
& McEwen, 1992).

89

Inhibition of hippocampal neurons and associated memory decits may also

be mediated by a stress-induced increase in proinammatory cytokines. Specically, interleukin-1 (IL-1), which is released by macrophages in response to
stress, may mediate stress-induced memory impairments by stimulating serotonin release in the hippocampus (Linthorst, Flachskamm, Holsboer, & Reul,
1994) or directly interfering with LTP (Bellinger, Madamba, & Siggins, 1993)
by disrupting calcium inux mechanisms (Cunningham, Murray, ONeill, &
OConnor, 1996). In any case, IL-1 may play a key role in mediating hippocampusspecic memory impairments. For example, IL-1 induction has been found to
impair performance on spatial (i.e., hippocampal-dependent) versions of the
Morris water maze task but to leave performance on the nonspatial versions
unaected (Gibertini, Newton, Friedman, & Klein, 1995). IL-1 has also been
shown to interfere with fear conditioning to contextual but not discrete cues
(Aubert, Vega, Dantzer, & Goodall, 1995).
The stress response also involves the dopaminergic system, primarily the system that innervates the medial PFC. This system appears to be vulnerable to mild
and brief stress, and some studies suggest that stress has a negative eect on
dopaminergic function. This is important because this region, which appears
essential for the normal functions associated with working memory, shares
responsibility with the hippocampus for processing the source of information
(source is a type of contextual information specifying where, or from whom,
something was learned) and plays a role in episodic binding. Dopamine turnover
is necessary to ensure proper PFC function; in line with this observation, recent
evidence suggests that working memory is impaired by exposure to extreme
or chronic stressors (Arnsten, 2000). Dopamine may play a role in emotional
memory as well, perhaps mediating the cognitive modulation of fear responses
(see Pani, Porcella, & Gessa, 2000).
Benzodiazepines play a role in stress and memory by potentiating the eects
of the inhibitory neurotransmitter GABA. Blocking benzodiazepine receptors
increases the experience of anxiety, while Valium, a benzodiazepine receptor
agonist, decreases it. Results from several studies suggest that chronic stress
reduces benzodiazepine receptor binding in the frontal cortex, not only increasing the subjective experience of stress and anxiety but also perhaps impairing
hippocampal function (McEwen, 2000). Moreover, recent experiments in young
animals show that inducing stress through maternal deprivation also reduces
benzodiazepine receptor binding in the frontal lobes, as well as the amygdala and
locus coeruleus (Caldji, Francis, Sharma, Plotsky, & Meaney, 2000; cf. Bremner
& Vermetten, 2001).
Finally, excitatory amino acids and endogenous opiate peptides may be involved in the stress-induced impairment of hippocampal function (Moghaddam,
Bolinao, Stein-Behrens, & Sapolsky, 1994). In brief, excitatory amino acids (e.g.,
glutamate) may abet glucocorticoids in causing pyramidal cell loss in the hippocampus (e.g., Virgin et al., 1991). Stress- and glucocorticoid-induced atrophy

90

of dendrites can be blocked by the anti-epileptic drug phenytoin (Dilantin), which

works primarily by blocking the release of glutamate. Some evidence suggests
that glucocorticoids and stress elevate levels of glutamate in the hippocampus
and other brain areas (Stein-Behrens, Lin, & Sapolksy, 1994) and that a high
concentration of glutamate is toxic to neurons. In addition, excitatory amino
acids aid in the damage of hippocampal neurons in culture and appear to exacerbate kainic acid-induced damage, as well as ischemic damage, to the hippocampus (Stein-Behrens et al., 1992). NMDA receptors are thought to be
involved, because blockade of such receptors is another eective means of preventing stress-induced atrophy of dendrites (McEwen et al., 1995).
Clearly, then, glucocorticoids are accompanied by, and in some cases modulated or mediated by, other neurochemicals that respond to stressors. This point
is important because it suggests that any eects of stress on memory may reect
complex cascades; thus, understanding stress- or trauma-induced memory impairments in humans, to which we now turn our attention, will not be a simple
matter.

Chronic Stress: Memory in Humans

As in animals, the memory-impairing eects of acute stress or glucocorticoid
administration are generally reversible in humans (Lupien & McEwen, 1997;
McEwen, 2000).4 Whether the eects of chronic stress are reversible, however,
remains an open question. Research generally suggests that chronic glucocorticoid elevation is associated with memory dysfunction in humans (e.g.,
Mauri et al., 1993). Although some studies examining the long-lasting eects
of Cushings disease suggest that moderate recovery from cognitive decit after
cortisol reduction may be possible (Starkman et al., 1999), other studies suggest
more permanent damage (Sheline et al., 1996; see also Bremner, 2001b). Memory
decits associated with aging in humans appear related to progressive dysregulation of HPA activity with increasing age. For instance, Lupien et al. (1994)
reported that elderly individuals whose cortisol levels increased over a 4-year
period performed signicantly poorer on tests of explicit memory than individuals whose cortisol levels declined over the same period (see also Lupien et al.,
1998). These observations led to the clinically relevant idea that some individuals
who endure life-threatening experiences with associated increases in stressrelated hormones might experience a dysregulation of their memory systems,
perhaps for specic traumatic events and for autobiographical events in general.
Traumatic Stress and Memory in Humans: The Special Case of Posttraumatic Stress
Disorder For every 100 people who endure horrifying experiences (e.g., Holocaust survivors, victims of physical or sexual abuse or rape, war veterans, etc.),
about 15 will develop posttraumatic stress disorder (American Psychiatric Association, 1994 [DSM-IV]), which characteristically involves a collection of memory

91

problems (see Bremner, 1999; Lemieux & Coe, 1995; Yehuda et al., 1995). Diagnostic criteria for PTSD include previous exposure to an event or events that involved actual or threatened death or serious injury, or a threat to the physical
integrity of self or others. In addition, the persons response must have involved
intense fear, helplessness, or horror (DSM-IV). PTSD is characterized by intrusive
recollections of the traumatic experience, avoidance of the event and reminders
of the event, hyperarousal, dissociation, sleep disturbance, emotional numbing,
and, paradoxically, increased emotional responsivity and anxiety (Emilien et al.,
2000; Golier, Yehuda, & Southwick, 1997; Saigh & Bremner, 1999).
Epidemiological studies suggest that the lifetime prevalence of PTSD in the
general population ranges from 8 to 9 per 100 and that roughly 60% of these
cases become chronic (e.g., Breslau, Davis, Andreski, & Peterson, 1991;
Davidson, Roth, & Newman, 1991; Kessler, Sonnega, Bromet, Hughes, & Nelson,
1995). Prevalence rates may be even higher in children and adolescents and are
twice as frequent in women as in men (e.g., Costello, Erkanli, Fairbank, & Angold,
2002; Kessler et al., 1995; see Fairbank, Ebert, & Caddell, 2001, for a more comprehensive review).
Clinicians have long reported that memory diculties are also associated with
PTSD, noting that war veterans often forget appointments and what they had
for breakfast yet describe the central aspects of their combat experiences in excruciating detail, as if those experiences happened yesterday (Bremner, 2001b).
Although forgetting appointments was once understood as a psychoanalytical
defense mechanism and treated as resistance on the part of the patient, a growing body of empirical research supports a more biological explanation for the
memory problems seen in PTSD (Bremner, 2001b). Both short-term memory and
episodic memory appear to suer in those diagnosed with PTSD (Bremner et al.,
1993; Bremner, 2001a, b; Torrie, 1944). Severe memory decits have been measured in PTSD using portions of the Wechsler Adult Intelligence Scale (WAIS),
the Wechsler Memory Scale (WMS), and the Verbal Selective Reminding Test
(vSRT). For example, veterans with PTSD performed more poorly than healthy
controls on the logical memory component of the WMS (about 2 of 5 show impaired immediate recall and 1 of 2 show impaired delayed recall; Bremner et al.,
1993). More recently, Vasterling et al. (2002), using the Rey Auditory Verbal
Learning Test (AVLT), showed that Vietnam veterans with PTSD manifested
greater decits on this working memory task than controls (Vietnam Veterans
without diagnosed psychopathology). These authors demonstrated that the cognitive impairments associated with PTSD are independent of estimated verbal
general intelligence (see, however, McNally & Shin, 1995).
Similar results were obtained from a sample of adults reporting childhood
sexual abuse (Bremner, Randall, Capelli, et al., 1995). These individuals performed more poorly than controls on the WMS Logical component for immediate and delayed recall, as well as percent retention. They also had diculty with

92

the immediate recall of words as measured with the vSRT. These impairments
cannot be explained by dierences in general intelligence as reected by IQ,
which did not dier in either study. Vietnam veterans, Desert Storm veterans,
and victims of rape diagnosed with PTSD all have impaired verbal episodic
memory in comparison to control subjects (Jenkins, Langlais, Delis, & Cohen,
1998; Uddo, Vasterling, Brailey, & Sutker, 1993; Vasterling, Brailey, Constans,
Borges, & Sutker, 1997; Vasterling, Brailey, Constans, & Sutker, 1998; Yehuda
et al., 1995, but see Stein et al., 1997).
Although these studies show that PTSD and memory decits are related,
design limitations prevent unequivocal causal conclusions. The rst limitation
is associated with the well-understood problems of quasi-experimental designs.
The second concerns a failure to examine memory performance in the context
of other neuropsychological functions, which leaves open the question of
whether PTSD patients exhibit selective memory impairments or more general
cognitive impairments. The third concerns lack of control over the potential relations among dierent cognitive functions, leaving us asking, for example, if
memory changes in PTSD can be explained more generally by changes in attention, concentration, visuospatial abilities, or other cognitive functions related
to memory. The fourth concerns what accounts for changes in memory observed
under these conditions; memory decits may be due to development of PTSD
symptoms per se, to the result of trauma exposure generally, or to some unknown
set of factors. A fth limitation concerns comorbidity, that is, whether other disorders associated with changes in memory (e.g., depression, alcoholism) might
account for many of the memory eects observed in PTSD.
Horner and Hamner (2002) reviewed 19 studies examining cognitive function in PTSD and concluded that converging evidence for memory decits in
PTSD does exist. They caution, however, that methodological problems, particularly comorbidity (e.g., substance abuse, mood disorders, and other anxiety
disorders), were a problem in most studies. Sixteen of the 19 studies reported
evidence of attention or memory diculties, or both, in PTSD. However, 15 of
these studies included patients with other comorbid disorders, reecting the diculty recruiting cases of pure PTSDand raising the question of whether such
a pure sample exists.
An important study by Gilbertson, Gurvits, Lasko, Orr, and Pitman (2001),
which was not discussed in the Horner and Hamner (2002) review, examined
performance on a range of neuropsychological tests including memory, attention, visuo-spatial skills, and executive functioning. These authors reported that,
although those meeting criteria for PTSD were impaired on all these tasks compared to controls, performance on the attention and memory tasks uniquely
predicted a PTSD diagnosis. In addition, verbal memory emerged as distinguishable from generalized attentional dierences, as well as from the severity of the
triggering trauma, general intelligence, depression, alcohol use, or a history of

93

developmental learning problems (Gilbertson et al.). This study provides some

of the rst evidence for a specic association between changes in episodic memory
and the presence of PTSD.
Autobiographical memory also suers in PTSD. Individuals diagnosed with
PTSD often have overly general (i.e., nonspecic, lacking context and detail)
memories of the past (McNally, Lasko, Macklin, & Pitman, 1995) and show
selective decits when retrieving specic personal information in response to cue
words having positive (e.g., kindness), neutral (e.g., appearance), or negative (e.g., panic) emotional valence (Kaspi, McNally, & Amir, 1995; McNally,
Litz, Prassas, Shin, & West, 1994; McNally, 1998). When individuals with PTSD
were asked to retrieve a personal event memory triggered by a cue, they tended
to retrieve generic memories without reference to a specic event. That is, rather
than reporting contextually specic memories with specic source (time, place)
information, they appear to report schema- or knowledge-based narratives.
Moreover, they report diculty retrieving specic positively valenced personal
memories, suggesting that trauma-related thoughts hamper access to positive
or neutral memories.
As these results clearly demonstrate, memory problems are a hallmark of
PTSD. In addition to general memory eects, those with the disorder tend to
either vividly re-experience (i.e., ashbacks), or fail to remember, information
related to their traumas. PTSD can thus be characterized as a disorder that paradoxically includes excessive and uncontrolled memory retrieval (e.g., intrusive
memories, nightmares, ashbacks) and memory disruption (e.g., traumatic
amnesia, fragmented memory).

Stress and Memory: What Is

Forgotten and Why

Although the foregoing discussion makes it clear that stress can impair memory,
the data leave one wondering precisely what aspect(s) of memory are impacted
by stress. All aspects, and all systems of memory, obviously are not entirely disrupted; if they were, researchers would be left with nothing to measure. So what,
exactly, does stress cause us to fail to learn, fail to store, or fail to retrieve?
In this section, we suggest that (1) contextual memories (i.e., memories of the
spatial and environmental aspects of experience), (2) detailed memories (i.e.,
those details that allow episodic memories to be distinguished from one another),
and (3) coherent, explicit autobiographical memories (i.e., the conscious, contextually, or episodically grounded recall of material with details that t together
in sensible, meaningful ways) suer in the presence of high levels of stress.5
What evidence is there that memory for stressful experiences lacks coherence,
context, and episodic detail? Initial support came primarily from anecdotal clinical reports, which suggested that experiences encoded during high levels of stress

94

are often fragmented (Bremner, 1999; Golier et al., 1997; Gray & Lombardo,
2001; van der Kolk, Hopper, & Osterman, 2001; van der Kolk, 1994). The bits
and pieces of experience are not related to one another as a whole. That is, these
pieces and patches are not bound together in a way that produces a good
gestalt, or a coherent episode. Fragmentation is an important feature of PTSD.
Patients sometimes describe gaps in recalled experiences, not only of trauma
but also of other personal experiences (Bremner, 1999). Some laboratory evidence suggests that stress-induced fragmentation can emerge in experiments assessing memory for verbal material (e.g., Payne et al., 2002, described later); it
is, however, most commonly observed in response to trauma.
Rather than simply retrieving fragments and reporting them as such, some
individuals who have experienced trauma appear to make educated guesses
about memory in a process called narrative smoothing. Jacobs and Nadel (1998)
oered a neurobiological theory of reconstructed memory taking this phenomenon into account. Starting with the observation that multiple memory systems
underlie the storage of episodic information, they suggested that retrieving an
autobiographical memory entails accessing and integrating fragments of information stored in these dierent systems. Many dierent types of information go
into the retrieval of a specic experience. There are the elements composing the
episode: what happened; who engaged in which actions; what it all looked like,
sounded like, smelled like; etc. In addition to this episodic content, there is also
spatial or contextual information. This is a particularly important type of information because all episodes occur someplace and thus are dened partly by the
space or context in which they occur. It can be dicult if not impossible to divorce
memories of episodes from memories of context, a notion that led us to suppose
that context serves as an organizing frame to which the elements of an episodic
memory trace are attached (Jacobs & Nadel, 1998; Nadel & Payne, 2002; Nadel,
Payne, & Jacobs, 2002).
Consider a hypothetical traumatic war experience; a memory of this experience might include smells (the jungle, unwashed bodies), sounds (gunre and
cursing), sights (the ash of a snipers weapon, the sight of a wounded combatant), tactile feelings (the humidity, thorns slashing skin), actions (diving for
cover, returning re), and emotions (anger and fear). Each of these independent
features is stored in the relevant part of the brain, typically, but not exclusively,
in the neocortex.
The hippocampus and adjacent medial temporal regions appear to be critical
to a process by which these disparate fragments of information from multiple
brain regions are bound into a unied memory trace at the time of retrieval
(Mitchell, Johnson, Raye, Mather, & DEsposito, 2000; Mitchell, Johnson, Raye,
& DEsposito, 2000). At the same time, the hippocampus appears to be critical
for placing an autobiographical memory in time, place, and context. Hence, the
disruption of hippocampal function by high levels of stress and glucocorticoids
could disrupt the storage of information about place, time, and context during a

95

traumatic experience, as well as the binding process that connects features of

the experience to the context and to one another (Bremner, Krystal, Charney, &
Southwick, 1996; Jacobs & Nadel, 1998).
The idea that memories are disaggregated during storage and then reaggregated during retrieval has several implications in explaining interactions
between extreme stress and memory (Jacobs & Nadel, 1998). By disrupting the
hippocampally based binding function, stress may leave the pieces, patches, and
context of a memory trace disconnected. So memories formed in the presence of
stressors, and hence under the inuence of high levels of glucocorticoids, will be
disjointed and fragmented, lacking spatial or temporal context, and thus creating decits in episodic memory.
Although traumatic stress appears to disrupt hippocampally based features
of memory (e.g., context, episodic detail), quite the contrary appears to be the
case for emotional features of memory based in other brain regions.6 The amygdala, a brain region adjacent to the hippocampus, is known to process, and perhaps store, emotional memories (Fanselow & LeDoux, 1999; LeDoux, 2000). The
amygdala is active during emotionally charged conditions, and the emotional
memories dependent on this region are highly resistant to forgetting (LeDoux,
1992). This stamping in of emotional memories is at least partly mediated by
the stress hormone norepinephrine, which, as noted earlier, also reaches high
levels during stressful experiences (see McGaugh, 2000, for a review).
Hence, two prominent stress hormones, both released during intense stress,
have opposite eects on dierent structures in the brain. High levels of cortisol
disrupt hippocampal function, impairing episodic (e.g., de Quervain et al., 2000;
Payne et al., 2002) and spatial memory (Laurance et al., 2001). In contrast, high
levels of both norepinephrine and cortisol facilitate amygdalar functions and
the emotional memories dependent on them (see Buchanan & Lovallo, 2001;
Metcalfe & Jacobs, 1996, 1998, 2000).
Jacobs and Nadel (1998) suggested that in the absence of an intact
hippocampus-based memory system, the amygdala-based system stores emotional information unbound to the spatio-temporal context of the relevant
events. This process results in a pool of emotional memories encoded without
a coherent spatio-temporal frame to organize them. This pool is, essentially, a
population of sensory and perceptual fragments acquired during the traumatic
event (Jacobs, Laurance, Thomas, Luzcak, & Nadel, 1996). By this model, traumatic events do not typically lead to a complete eradication of memory but rather
to the storage of fragments lacking an organizing framework. Devoid of this
framework, traumatized individuals resort to narrative smoothing; retrieved
memories are composed of re-aggregated representations cobbled together by a
narrative based on gist, inference, and educated guesswork (Jacobs & Nadel,
1998; Nadel & Jacobs, 1998; see also Shimamura, 1997).
By this model, a fragmentation of normally coherent elements of an experience resulting from disruption of hippocampally based contextual encoding,

96

coupled with the sparing or even enhancement of amygdala-based emotional

encoding, leads to a theoretically interesting clinical picture of trauma and
memory. Van der Kolk and Fisler (1995) presented evidence that traumatic
events are remembered initially as disconnected images and waves of disjointed
sensations and emotions. The emotions can emerge as punctate waves (van
der Kolk & Fisler, 1995) or as persistent anxiety without an apparent source
(e.g., Freud, 1926). Clinicians sometimes suspect that clients presenting with
such generalized (or freeoating) anxietyan emotional disturbance without obvious anchors in autobiographical memoryhave suered past traumas. Claims about generalized anxiety are supported by some empirical
evidence, which suggests that memory disturbance in response to trauma
may be associated with dysregulation of the autonomic nervous system,
heightened arousal or vigilance, and increased startle reex (Joseph, 1996).
Indeed, emotional dysregulation is also characteristic of PTSD, and while
patients often suer from extreme anxiety, they may also exhibit isolation of
aect (described in the following) and describe emotional blunting and feelings
of numbness.
Cortisone therapy (a common treatment for rheumatoid arthritis) has been
associated with both anxiety and panic, on the one hand, and with aective dissociation, depersonalization, and emotional atness, on the other (Clark, Bauer,
& Cobb, 1952).
These observations are reminiscent of a defense mechanism proposed late in
Freuds career: defensive undoing (Freud, 1926). Freud thought of anxiety as
a marker that emerges when the mind cannot cope with the threatening information assailing it. Thus, it pushes autobiographical information associated with
the trauma out of awareness, and anxiety emerges instead. Like much of Freuds
work, defensive undoing is an intuitively appealing idea but dicult to test. In
contrast, the biological response to stress oers a testable and more rigorous
explanation of this phenomenon. If high levels of cortisol disrupt normal neuronal function of the structure responsible for the storage of critical contextual
information (i.e., the hippocampus), then coherent memories of trauma will be
rendered inaccessible.7 Moreover, if physiological processes impair cognitive but
not emotional aspects of memory, as suggested by the opposite actions of catecholamines on amygdala function and cortisol on hippocampal function, then
feelings of anxiety disconnected from the corresponding factual or autobiographical memory of a traumatic event might be expected. In other words,
emotional memory for the traumatic event might continuously enter awareness
in the form of generalized feelings of anger, fear, or uneasiness (referred to as
body memories by van der Kolk, 1994) unbound to the originating episodic
context. Without a contextual framework to relate emotional memories to associated traumatic events, these emotional memories surface as vague disconnected feelings of unease, anger, or fear that are troublesome. Anxiety is thus
experienced as generalized and freeoating.

97

Isolation of Aect in PTSD:

A Paradoxical Memory Decit

Perhaps the most common pattern of decit seen in PTSD involves what we described: a disruption of episodic memory, along with the experience of extreme
anxiety apparently devoid of a source. Clinicians also describe cases exhibiting
the opposite pattern of forgetting, however, when a person retains episodic
information concerning a traumatic event but loses the corresponding emotional
response (see chapter 11). Modern clinicians use the terms emotional blunting
or emotional numbing to describe this pattern (e.g., Golier et al,, 1997; chapter 11). However, this observation is reminiscent of Freuds concept of isolation of aect, when people may recount traumatic experiences in the absence
of the emotion that must have initially accompanied them; these individuals are,
in fact, often overtly blas about the upsetting information they recount (also
called splitting o of aect; Freud, 1926). Thus, isolation of aect involves a
lack of emotion and may contribute to experiential aspects of PTSD (Harvey &
Bryant, 1998; van der Kolk, 1991). Freud speculated that individuals experiencing emotional numbness do not repress traumatic memories, thereby holding disturbing information out of awareness. Rather, unlike the failure to
remember in undoing, individuals experiencing emotional numbness have no
trouble recollecting the episodic aspects of the trauma. They do, however, have
diculty consciously accessing the emotion that originally belonged to it. By this
conjecture, these individuals remember only the cognitive aspects of trauma
(Mitchell & Black, 1995). To bring memory of the emotion into awareness would
be too threatening, unduely straining the ego.
Although there is little or no epidemiological data on which to rest the case,
isolation of aect (emotional blunting) lends itself to a biological explanation.
One possibility (we speculate on it in the next section) is that stress disconnects
normal chains of communication among normally interacting hippocampal,
amygdala, and cortical networks.
A Speculative Neurobiological Account of Isolation of Aect Metcalfe and Jacobs
(1996, 1998, 2000), working within the theoretical framework just outlined,
explored interactions between two learning/memory systems, (1) a cool system, based in interactions between hippocampal and frontal regions, that is
slow, emotionally neutral, exible, integrative, episodic, strategic, and highly
contextualized and (2) a hot system, based in the amygdala, that is the basis of
emotionality, fears, and passions, that is fast and impulsive and controlled by
innate and acquired releasing stimuli. Stress has a dierent impact on the cool
hippocampal-based system and the hot amygdala-based system. Exposure to
intense stressors or trauma disrupts the hippocampus, leaving disaggregated
memories ruled by the cool system. Exposure to these same stressors, however,

98

enhances the hot system, encoding context-free emotional memories rapidly

and eciently.
This approach seems to account for many features of recovered memory,
specic phobia, panic attacks, and PTSD (Jacobs & Nadel, 1985, 1998; Nadel &
Jacobs, 1996; Metcalfe & Jacobs, 1996), with one glaring exception. As we have
described, modern clinicians often describe an emotional blunting associated
with the presence of PTSD; some even describe this as a core dening feature
of PTSD (e.g., Litz & Gray, 2002). To oer an account of emotional blunting (i.e.,
isolation of aect), we must consider two possibilities, neither well supported
empirically.
The rst possibility is that chronic exposure to glucocorticoids damages neurons in the amygdala, disrupting its processing of emotional information. The
most serious weakness of this proposal is that chronic stress does not appear to
damage the amygdala. Indeed, there is some evidence that, in the rat, predictable exposure to a stressor simultaneously disrupts dendritic arborization in the
hippocampus but enhances this same process in the amygdala (Vyas, Mitra, Rao,
& Chattarji, 2002). Under these conditions, one expects to observe disruption of
hippocampal-mediated memory (e.g., contextualization) and enhancement of
amygdala-mediated memory (e.g., emotional memory). In contrast, chronic
exposure to unpredictable stressors leaves the dendritic arborization process relatively intact in the hippocampus but induces dendritic atrophy in the basolateral
nucleus of the amygdala (Vyas et al., 2002). Under these conditions, one expects
to observe highly contextualized episodic and emotional memories but greatly
disrupted processing of emotional memories. That is, one expects broad emotional attening.
In the presence of a damaged amygdala, however, one would expect a global
disruption of emotional processing, yet this apparently does not happen to individuals who experience isolation of aect. Instead, these individuals appear to
be unaware of their emotional reactions while simultaneously showing enhanced behavioral and physiological reactivity (see, for example, Jacobs, Nadel,
& Hayden, 1992).
This nding recalls work reported by Peter Lang and his colleagues and
introduces a second possible explanation of isolation of aect. Lang and Cuthbert
(1984) conceptualized anxiety as including three loosely coupled response systems (constellations) of overt behavior, verbal report (cognition), and physiological activation. Within this framework, those who experience isolation of aect
behave as if they do not experience subjective emotion (i.e., feelings) but do experience heightened arousal at the level of both behavioral and physiological
constellations (Jacobs et al., 1992). It is as if the physiological concomitants of
emotion do not enter awareness. The physiological changes occur but without
corresponding emotional valence. More formally, aect is not expressed in the
cognitive constellation, while heightened emotion (e.g., hypervigilance, in-

99

creased startle, etc.) remains expressed in both behavioral and physiological

constellations.
We account for this pattern of responding in isolation of aect by making
several potentially defensible assumptions about interactions among stress, stressors, and the functions of hippocampus, amygdala, and PFC.

Stress and the Frontal Cortex:

Stress-Induced Forgetting Beyond

the Hippocampus

We have focused almost exclusively on the hippocampus and amygdala. The

frontal lobes, however, have received increasing attention in the stress literature. Recent experiments using functional magnetic resonance imaging (fMRI)
(Henson, Shallice, & Dolan, 1999) and electrophysiological techniques (Otten &
Rugg, 2002) have demonstrated an important role for the PFC in memory (Raye,
Johnson, Mitchell, Nolde, & DEsposito, 2000; Buckner, Logan, Donaldson, &
Wheeler, 2000). Traditionally, the PFC was thought to contribute mainly to
encoding processes through its presumed role in working memory, which
allows information to be maintained temporarily in an active state (Baddeley,
1986). New evidence from PET and fMRI studies, however, shows that the PFC is
also involved in memory retrieval (see Henson et al., 1999; Nyberg & Cabeza,
2000; Rugg, Flethcer, Chua, & Dolan, 1999) and that the PFC may share, with
the hippocampus, responsibilities for binding episodic material (Mitchell,
Johnson, Raye, & DEsposito, 2000).
The frontal cortex is also involved in the modulation of emotion and inhibitory responses (e.g., Arnsten, 2001; Rosenkranz & Grace, 2002). Moreover, damage to this area, particularly if it is conned to the prefrontal region, leads to
perseveration, diculty choosing correct responses, and distortions and confabulations of memory. For example, patients with frontal lobe damage tend to
make an abnormally large number of condent false recognition errors, in spite
of good recall and recognition generally (Curran, Schacter, Norman, & Galluccio, 1997; Delbecq-Derouesne, Beauvois, & Shallice, 1990; Melo, Winocur, &
Moscovitch, 1999; Parkin, Bindschaedler, Harsent, & Metzler, 1996). With the
existence of glucocorticoid receptors in the frontal lobes (reviewed later), stressrelated increases in commission errors likely are not caused solely by a disrupted
hippocampus.
An important review by Lupien and Lepage (2001) underscores the contribution of frontal regions to stress-related eects on episodic and emotional
memory. These authors argue that the hippocampus has been studied at the
expense of other brain regions likely involved in glucocorticoid action and that
attempts to explain stress impairments on memory solely on the basis of the
impact of stress on the hippocampus may be incomplete.

100

Lupien and Lepage (2001) note that both Type I (mineralocorticoid, MR) and
Type II (glucocorticoid, GR) stress receptors are present in cortical regions, with
a preferential and dense distribution in the PFC. Hence, stress may exert eects
not only on hippocampal neurons but also on neurons critical for the normal
functioning of the PFC. If one takes the inuence of stress and corticosteroids on
human memory to reect the action of GRs, then one must acknowledge that
both hippocampal and PFC function are likely involved.
Dierential MR and GR action in the hippocampus and PFC may underlie
some of the dierences in the magnitude and nature of memory impairments
reviewed in this chapter. The hippocampus, mainly through MR, but also moderate GR, receptor binding, may be a primary site of short-term, reversible eects
that occur in response to stress (e.g., de Quervain et al., 2000; Newcomer et al.,
1994, 1999; Payne et al., 2002; Wolkowitz et al., 1990, 1993). MRs must be
totally saturated before GRs become activated, and signicant reductions in LTP
do not take place until both MRs and GR receptors are occupied. Thus, experiments using humans (for obvious ethical reasons) likely rarely attain the levels
of stress needed to produce lasting changes in memory.
On the other hand, exposure to trauma may lead to the widespread and sustained saturation of both MR and GR receptors in the PFC and hippocampus.
These conditions cannot be ethically reproduced in the laboratory and thus may
emerge only in the context of naturally occurring chronic stress in the patient
populations described earlier (e.g., Cushings syndrome, PTSD, major recurrent
depression, etc.). In such cases, extensive binding of cortisol to both receptor types
in the hippocampus and PFC may lead to long-lasting and harmful changes in
neuronal functioning in the entire memory circuit, resulting in dramatic changes
in how environmental information is encoded, stored, and retrieved. Evidence
that excessively high levels of norepinephrine (e.g., during exposure to uncontrollable stress) impair cognitive functions of the PFC (Arnsten, 1998) makes this
concept all the more feasible.
Consequently, exposure to intense stress appears to aect the PFC, as well as
the hippocampus and amygdala. Models of stress and memory that focus solely
on the hippocampus and amygdala cannot adequately account for isolation of
aect, but perhaps the addition of the PFC can explain this phenomenon. Does
the PFC play a role in the modulation of emotions? What happens to that role
under conditions of extreme stress? And, if we can specify that role adequately,
how does this help us understand isolation of aect?
We start by assuming that the PFC plays two roles in emotional processing
and regulation: (1) inhibitory control of emotional responses generated by an
amygdala-centered system and (2) subjective awareness of that emotion (e.g.,
Amaral, Price, Pitkanen, & Carmichael, 1992; Reiman, Lane, Ahern, Schwartz,
& Daivdson, 2000; Rosenkranz & Grace, 2002; Anderson, Bechara, Damasio,
Tranel, & Damasio, 1999). Under the model outlined earlier, these assumptions
have two important implications. First, disruption of the PFC should disinhibit

101

the amygdala and thereby amplify the behavioral and physiological constellations of emotions controlled by it. Second, disruption of the PFC should disrupt
the subjective experience of emotionality. Under these conditions, we expect
individuals to experience high levels of emotional arousal in the behavioral and
physiological constellations but not in awareness. This process would then account for narrow (or cognitive) emotional attening with enhanced reactivity
in the behavioral and physiological constellations.
Although the model remains to be rounded out and tested, the welldocumented processes that simultaneously impair PFC- and hippocampalbased memories and facilitate amygdala-based memories provide empirical
physiological support for defensive undoing and emotional blunting, or the isolation of aect.

Trauma and Memory:

What Is Remembered?

Intense stress is associated with memory disruption, either partial or complete,

relating to the trauma itself or to episodic memory in general. Yet, paradoxically,
other cases of trauma provoke an inability to forget (hypermnesia) and a desire
to talk about the experiences constantly (e.g., Terr, 1991, 1993). Along similar
lines, some individuals report spontaneous ashbacks and hallucinations, as well
as severe emotional upset in relation to the trauma. Sometimes these ashbacks
appear to be triggered by environmental stimuli (e.g., the sound of a recracker
or an engine backring) and are reportedly so real that the person feels he or she
is re-experiencing, rather than remembering, the event. The fact that people
report ashbacks suggests that some representation of traumatic events has been
stored but is not always accessible. Interestingly, ashbacks reportedly plague
trauma survivors regardless of whether they report amnesia for traumatic events
or experience hypermnesia and talk about them obsessively.8 Further, whether
they experience ashbacks may be unrelated to general level of memory function (e.g., Joseph, 1996).
A thorough investigation of ashbacks is beyond the scope of this chapter.
Nonetheless, we briey note that an understanding of the neurobiological processes involved may help elucidate ashback phenomena. Although fear is tied
to specic stimuli in a persons immediate environment, anxiety appears to be a
secondary response (LeDoux, 1992). Pathways of fear involve interactions between cortical sensory processing areas and the amygdala, where emotionally
neutral sensory signals are transmitted to receive their aective connotation.
Depending on whether the emotional processing mechanisms of the amygdala
are activated directly by incoming sensory information or indirectly by cognitive
processes organized elsewhere, a person will either experience fear, a primary
response, or anxiety, a secondary response (Davis, 1992; Davis & Shui, 1999;
Gray, 2000). The former pathway may initiate ashbacks, which elicit fear di-

102

rectly and are experienced in the here and now. The secondary pathway may
characterize normal reactivation of trauma memories, which should elicit
negative aect and anxiety that are correctly referenced to the past. This latter
pathway appears to be disrupted during the trauma, leading to decontextualized
and fragmented memory for the experience.
Acknowledging the diering contributions of these two pathways, we might
expect the content of ashbacks to be highly emotional, as fear circuits become
activated. The fact that amygdala function is preserved during intense stress ts
this suggestion. Moreover, due to stress-induced disruption of the structures
underlying context and detail memory (i.e., hippocampus and PFC) during the
trauma, we might expect ashbacks to lack both contextual information and
details outside the emotional center of the traumatic event. In the absence of
contextual detail, ashbacks should consist mainly of the central content or
theme of the traumatic experience.9 In other words, we expect the content of
ashbacks to be rarely, if ever, about the peripheral details of a traumatic experience. Rather, we expect the images that emerge during a ashback to be the
most horrible, emotionally salient, and central features of the traumatic experience. These considerations force us to examine another literature, one that seemingly unfolded parallel to the trauma and memory literature.

The Emotion and Memory Literature

We have focused almost exclusively on memory impairment that results as a
consequence of stress and trauma. Many readers, however, particularly those
with a background in emotion, may be wondering how these data t with evidence seemingly pointing in the opposite directionnamely, that memory often
improves in response to emotion, even intense emotion (see also chapter 2 here).
A key nding in the emotion and memory literature is that memory for emotional events is superior to memory for similar, neutral events (e.g., Heuer &
Reisberg, 1990; chapter 1 here). Most people have ready examples of emotional
events that are remembered vividly as if they happened yesterday, and a substantial literature conrms that some emotional experiences are unforgettable
(see chapter 1). People also report vivid recollections for intensely negative experiences accompanied by strong negative emotion (e.g., Porter & Birt, 2001). So we
now have one literature demonstrating that high levels of stress can disrupt
memory for stressful experiences and another literature demonstrating that intense negative emotion can improve memory for negative experiences. The key
to this apparent conict, we think, is embedded in an as yet unacknowledged
dierence between strong negative emotion and stress. Although stressful experiences are usually negative, and although substantial overlap may connect
negative emotion and stress, we claim that they are not the same.
In the following pages, we try to address the tension between these two literatures directly. Although at rst blush data from the two research traditions seem

103

contradictory, we suggest that a solution resides in the dierential contributions

of hippocampus-based circuits and amygdala-based circuits. Can intensely negative memories be vivid and intact in some cases and disrupted and fragmented
in others? Can traumatic memories be both vivid and fragmented? The short
answer to both questions, we think, is yes. The key to resolving this apparent
paradox likely resides in (1) the magnitude of the stress, (2) the corresponding
amount of stress hormones released during emotional experiences of dierent
intensities (i.e., the degree of stress inherent in a given emotional experience),
and (3) the degree to which changes in these stress hormones aect dierent
memory structures.
The thrust of our argument resides in the (possibly nonlinear) relations between experienced emotion and the magnitude or intensity of the stress response.
However, we briey point out two additional diculties that may help account
for the contradictory literatures. The rst diculty is denitional. It concerns the
often-overlooked fact that there are multiple memory systems. Memory is not a
single entity that goes up or down in response to emotional arousal; rather,
memory is an aggregate of many systems of which emotional memory is just one.
If this is true, emotional arousal (even in the absence of stress) may facilitate some
attributes of memory yet disrupt others. The second concerns how emotion
is conceptualized and where and how it comes into play in memory experiments.
In the stress/trauma literature, emotion (emotional memory) is a dependent variable, one of several types of memory dierentially aected by stress, the independent variable. Conversely, in the emotion literature, emotion (or emotional
arousal) is often used as an independent variable, which inuences dierent types
of memory (usually enhancing memory for emotional content: words, pictures,
videos, experiences, etc.). Subtle dierences between emotion and stress need
untangling, but distinguishing between emotional content and emotional
arousal that may be associated with that content is critical. We consider each of
these problems.
Something about emotionality leads to particularly vivid and long-lasting
memories. Attempts to isolate mechanisms underlying this eect have led to
discoveries anchored in research techniques derived from cognitive theory (e.g.,
chapter 1) and in research techniques derived from learning/neuroscientic
theory (e.g., McGaugh, 2000, 2002).
The Learning/Neuroscientic Approach to Emotion and Memory Research A crucial contribution of the neuroscience approach is the unshakable notion that
specic neural and hormonal mechanisms underlie the enhanced consolidation
of memory for emotional materials. One group of researchers (Cahill, Babinsky,
Markowitsch, & McGaugh, 1995; Cahill & McGaugh, 1998), for example, has
demonstrated that stress hormones interact with the amygdala and, through
that interaction, enhance the storage of explicit memories. This group has shown
that adrenergic activity enhances (e.g., OCarroll, Drysdale, Cahill, Shajahan,

104

& Ebmeier, 1999; Southwick et al., 2002) and blockade of adrenergic activity
reduces (e.g, Cahill, Prins, Weber, & McGaugh, 1994) the consolidation of explicit
memory for emotional materials in humans. The activation of beta-adrenergic
receptors, located on neurons in the basolateral nucleus of the amygdala, appears critical for this noradrenergic mediation of emotional memory consolidation (McGaugh, 2002; McGaugh & Roozendaal, 2002). In addition to adrenergic
inuences on memory, McGaugh, Cahill and their associates point out that low
doses of glucocorticoids can enhance memory as well, suggesting that both epinephrine (and norepinephrine) and cortisol enhance memory consolidation
(Cahill & McGaugh, 1998). Indeed, the enhancing eects of cortisol on memory
for emotional (i.e., nonhippocampal-dependent) information have been demonstrated in animal studies of aversive conditioning and in human studies that use
emotional memory tasks (e.g., Buchanan & Lovallo, 2001; Roozendaal, 2000).
For example, Buchanan and Lovallo showed that 20 mg of hydrocortisone facilitated memory for emotional pictures, but not neutral pictures, relative to a
placebo control. Again, the basolateral amygdala (BLA) is the specic nucleus
thought to underlie these eects. The BLA is not only a locus of interaction
between norepinephrine and cortisol (McEwen, 2000) but also a major hub of
communication between the amygdala and hippocampus. As such, the BLA
is thought to be a critical locus regulating norepinephrine and glucocorticoid inuences on memory consolidation enhancement eects (Roozendaal &
McGaugh, 1997).
These ndings appear to contradict the stress data we reviewed. Yet a close
examination of the designs, the data, and the measures to obtain them leads to
a proposal that could resolve this apparent contradiction, given a clear understanding of features composing emotional memory.
The design of a typical study examining emotion and memory goes something like this. Human subjects watch a series of slides depicting successive
moments in a story that is either emotionally arousing or emotionally neutral.
A tape-recorded narrative accompanies the slides, helping subjects interpret the
scenes and synthesize them into a story. Many studies have used variations
on the doctor/mechanic slide sequences originally developed by Heuer and
Reisberg (1990). In both stories, a mother and young boy visit the father at work.
In the emotional series, the father is a surgeon performing explicit surgeries
(where viscera or severed legs are in plain view); in the neutral series, the father
is a mechanic working on a broken car, and disconnected car parts are in plain
view. Studies in this tradition focus mainly on consolidation processes, so
after waiting for a time (usually minutes after viewing the slides; although see
Cahill et al., 1994), participants receive a substance that either enhances or
blocks adrenergic activity. The participants receive a single test of memory,
usually after a long delay. The typical result shows enhanced memory for the
emotional slide sequence with adrenergic facilitation and decreased memory
for the emotional slide sequence under adrenergic blockade. Neutral slides,

105

on the other hand, remain unaected by adrenergic alterations (e.g., Cahill et al.,
1994).
These studies have made critical contributions to our understanding of the
neural and hormonal mechanisms underlying emotional memory enhancement.
Nevertheless, they tempt us to believe, perhaps prematurely, that emotional
arousal improves emotional memory overall. Despite having borrowed emotional
memory materials and designs from cognitive researchers, these researchers
measure memory dierently. Those from the learning-based tradition often obtain high-level aggregate measures of memory, measures that may not be sensitive to subtle nuances obtained when measuring the components of memory
directly (although see Adolphs, Denburg, & Tranel, 2001). For example, Cahill
and colleagues assess memory using traditional recall and recognition measures,
focusing on the overall amount of memory preserved posttreatment. In contrast, cognitive researchers often examine qualitative dierences within emotional memoryfor example, dierentiating memory for central or gist-like
information from memory for peripheral detail (see chapter 1), often showing
that only the former is enhanced by emotion. Hence, although arousal may lead
to broad enhancement of emotional memory, we do not yet have the necessary
data to draw such a conclusion. Until the eects of emotion on gist-like information and peripheral detail are directly compared, it may be safest to assume
that memory enhancement maps onto a fairly robust pattern of ndings in the
cognitive literaturethat emotional arousal promotes memory for an events
gist or center, often at the expense of other (contextual) details (Adolphs et al.,
2001; Metcalfe & Jacobs, 1998, 2000; chapter 2 here).
The Cognitive Approach to Emotion and Memory Research: Central Versus Peripheral Information Results anchored in the cognitive tradition are generally
consistent with the notion that increased levels of norepineprhine (and arousal)
enhance recall of memory for emotional events. An early nding in the emotion
and memory literature was that individuals initially remember highly unpleasant events more poorly than neutral events but that, over time, the pattern
reverses, with enhanced recall for arousing events after a delay (e.g., Goodman,
Hirschman, Hepps, & Rudy, 1991; Osborne, 1972). Clark, Milberg, and Ross
(1983) and Kleinsmith and Kaplan (1963, 1964) demonstrated that highly arousing events are initially less well remembered than low-arousal or neutral events,
but a recovery eect for the high-arousal events occurs after a delay. The eventual enhancement of memory by arousal is not surprising because such memories may be essential for survival. For instance, remembering to avoid a certain
poisonous food could save ones life (Bremner, Southwick, & Charney, 1997, 1999;
McEwen, 2000). It seems safe to assume that negative emotional events lead to
overall enhanced memory for these events, particularly after a delay.
Recent research, however, suggests that the story isnt quite so straightforward. First, newer studies show that emotional arousal can promote imme-

106

diate memory enhancement, implicating an eect of arousal on encoding, in addition to consolidation, processes (see Hamann, 2001, for a review). Second, and
more important, many studies suggest that emotion enhances episodic memory
but in a way often limited to the central and thematic features of such experiences (Burke, Heuer, & Reisberg, 1992; Christianson & Loftus, 1987, 1991;
Loftus, Loftus, & Messo, 1987). People tend to remember the gist or aective
signicance of a stimulus or event at the expense of peripheral details
(Christianson, 1992, Goodman et al., 1991; Ochsner & Schacter, 2000). This
tendency may appear in the weapon focus phenomenon, when witnesses to
crimes recall information about the weapon used during an assault at the expense of other important details (Stanny & Johnson, 2000). Loftus and Burns
(1982), for example, reported that subjects who watched an emotionally arousing videotape of a violent bank robbery (in which a little boy is shot in the face)
did not recall the events immediately preceding the attack as well as control
subjects who watched a less emotional videotape (in which bank robbers simply ran past the little boy, leaving him unharmed). Similarly, Goodman et al.
(1991) showed that receiving an unpleasant inoculation leads to enhanced
memory for central information related to the procedure; likewise, Christianson
and Loftus (1991) found enhanced memory for central information but impaired
memory for peripheral information within an emotional event, as compared to
a similar, neutral event.
Burke, Heuer, and Reisberg (1992) demonstrated the same emotion-driven
enhancement of central or gist-like information but added to our understanding of centrality by showing that memory improved both for visually central
information and information associated with an events central theme or plot.
As in other studies, however, memory suered for details that were not associated with central events (see also Christianson & Loftus, 1991). Christianson and
Loftus (1990) found that people describing their most traumatic memories reported more central than peripheral detail information. Study after study has
shown that memory for peripheral detail suers while memory for gist or thematic content improves with heightened emotionality (see chapter 1 for a comprehensive review and more detailed scrutiny of these data).
There are several possible explanations for this phenomenon (see Heuer &
Reisberg, 1992, also chapter 1 here). For example, central information may be
better remembered simply because people attend to it longer. Increased memory
for emotional information, however, can occur even when viewing time for
emotional and nonemotional information is held constant (Christianson et al.,
1991; Ochsner & Schacter, 2000). Another possibility is that the nature of
attention may dier for emotional and nonemotional events, leading to better
memory. Emotional events could also be associated with more elaborate processing and more rehearsal of the information to which people attend (Ochsner &
Schacter, 2000). For example, witnesses to emotional attention-grabbing
episodes (e.g., accident, crime, assault), compared with a neutral attention-

107

grabbing episode, may be more concerned with what they have just seen and
think about (rehearse) it more often. According to Christianson and Safer (1995),
this process of elaboration implies a psychological focusing on salient information that is the source of emotional stress, while at the same time limiting access
to the information in the mental periphery.
Although most discussion of these ndings focuses on memory dierences for
central detail versus peripheral detail, there is another way to conceptualize the
ndings. As Christianson (1992) points out, central detail represents the source
of the emotional arousal; it includes the most relevant information for extracting
the emotional signicance of a stimulus or event. So there may be an alternative denition of what centrality is and what it means for traumatic memory.
First, central information may represent a concentration of experience, where
disproportionate emphasis is placed on emotion. Whatever is emotional about
an experience will therefore be well remembered, whereas more neutral information will not fare as well. Thus, emotion, likely through the activation of the
adrenergic system, may be responsible for the memory enhancement of central
experiences, because centrality and the source of emotionality often overlap.
Second, central information may also represent a schematization of experience,
where peripheral details are not well remembered but the gist or essence of an
event is particularly well preserved. The process of schematization occurs when
preserved emotional details are woven into a meaningful narrative memory
on the basis of schematic inference. The result is a gist-based memory that may
be biased toward central and emotional features of experience. By this view,
schematized memories are likely thematically consistent, but they may not be
accurate. Our notion of schematized memories dovetails with evidence suggesting that emotional memories are prone to errors and can even be entirely
false (see chapter 1). Due to the way they are formed, schematically consistent
memories are not always accurate, especially if one believes that the truth is
in the details. Centrality and gist are quite compatible within this interpretation and likely represent the same information. Gist memories, we say, are
composed of preserved bits of central information, linked together by schemabased inferences.
Glucocorticoid binding at hippocampal receptor sites may shed light on this
process, particularly if one thinks of peripheral details as contextually important
features (i.e., specifying source information, spatial and temporal relationships)
that rely on a functional hippocampus and central information as gist-like and
relatively noncontextual, capturing the essence of experience but missing many
of the episodic details that would produce complete and veridical memories.
By this view, then, central information represents not only the emotional tone
but also the overall gist or theme of a stressful experience. Central or gist-like
information may survive, whereas detail memory may suer, in response to
emotional stress.

108

This claim brings us to the inevitable place where the words emotion and
stress must be pulled apart. Despite obvious overlap between the two,
emotional arousal is not always suciently intense to elicit a stress response, and
occasionally a stress response is elicited but is relatively mild (and thus may
enhance memory; see Cahill & McGaugh, 1998). Thus, while the pattern of
enhanced gist and impaired detail memory may be inuenced by denitional
problems with these terms and limits in memory measurement (i.e., in the
neuroscientic experiments previously described), the most compelling reason
for such a pattern concerns the magnitude of negative emotion and, critically,
the degree of HPA-axis activity associated with negative emotion. Stress is distinct from ordinary emotion insofar as HPA-axis activity and cortisol reach high
enough levels for memory impairment. Whether emotion promotes memory for
gist and centrality while disrupting memory for noncentral details may depend
on the level of stress elicited by an emotional experience.10 Along these lines, we
suspect that HPA-axis activity may be associated with the viewing of emotional
slides and videotapes used in most emotional memory experiments. This
hypothesis is easy to test if cortisol measures are taken during these procedures.
A study of this nature is currently under way in our laboratory, with the goal of
testing our predictions about how gist and detail memory fare in response to
stress.
Our suggestion nds support in other studies as well. There is good evidence
in the memory literature that gist and detail memory are dierent (Neisser, 1981;
Reyna & Kiernan, 1994; see also chapter 11 here, which distinguishes between
core memory and narrative memory). The fuzzy trace theory of Brainerd and
Reyna (1998), for example, posits two types of processing that eventually lead
to two dierent representations of experience. During encoding, verbatim and
gist traces are formed in parallel, creating a hierarchy of independent representations at varying levels of precision. In many cases, gist representations are more
easily retrieved than detail representations, and we suggest that stress is one such
case. Interestingly, fuzzy trace theory has gained attention as a theoretical framework that could explain a variety of false memory errors, and we note here that
preservation of gist at the expense of detail may inspire lling in, or narrative
smoothing, setting the stage for emotional memories that are true to gist yet
inaccurate (see Golier et al., 1997; Heuer & Reisberg, 1992).
Studies demonstrating increased memory for central, thematic, and appraisalrelevant information (e.g., Burke et al., 1992) are consistent with the results
of stress studies by Payne et al. (2002) and Lindsay and Jennings (2000). Both
studies sought to determine the eect of stress on false memory as assessed
by the Deese-Roediger-McDermott (DRM) paradigm (Deese, 1959; Roediger &
McDermott, 1995). In this work, participants studied a number of lists of semantically related nouns (e.g., candy, sour, sugar, bitter, chocolate, cake, etc.), followed by a four-item recognition task consisting of three types of words: words

109

presented in the original list (e.g., candy), words not presented in the original
list and not related to the theme of the list (e.g., hat), and critical lure words
not presented in the original list but highly related to the them of the list (e.g.,
sweet). Payne et al. showed that experiencing an acute stressor followed by a brief
waiting period dramatically increased false recognition of the critical lures. In
particular, thematically related or gist-based false alarms (incorrect identication of critical lures as present in the original list, or commission errors) were
amplied by the psychologically induced stressor. This increase in false alarms
using the DRM paradigm has also been observed in chronically anxious participants (Lindsay & Jennings, 2000), in traumatized patients with and without the
diagnosis of PTSD (Zoellner, Foa, Brigidi, & Przeworski, 2000), and in sexually
abused women with PTSD (Bremner, Shobe, & Kihlstrom, 2000).
Both Payne et al. (2002) and Lindsay and Jennings (2000) concluded that
elevated stress hormones increase the production of false memories in the DRM
paradigm. One must, however, ask false memory for what? Only on very rare
occasions did stressed subjects falsely recognize unrelated distractor words, and
their rate of false recognition of these distractors did not dier from that of nonstressed control subjects. Stressed subjects exhibited false memory of a particular type; they falsely remembered more words related to the thematic structure
of the word list and therefore made more gist-based errors than their nonstressed
counterparts. We suggest that this may be due to impaired contextual and detail memory and, as a result, an overreliance on gist memory (see Payne et al.,
2002).
These ndings converge on the idea that stress disrupts memory for spatial
and contextual details of experience, leading to an overreliance on thematic information.11 A closer look at the studies described earlier supports this idea. Recall that in Wolkowitz et al. (1990, 1993; see also Deptula, 1983), dexamethasone
treatment was associated with signicantly greater rates of commission errors,
or intrusions (self-generated words) into free recall, than control (placebo) treatment. It was also associated with greater rates of false recognition of semantically related distractor words. There was, however, no dierence between the
treatment and control groups on correct free recall or total hits on the recognition task. The same pattern emerged for people given prednisone, though the
tendency to false alarm to related distractors disappeared 7 days after discontinuation of treatment. Although not conceptualized as false memory experiments,
the work of Deptula (1983) and Wolkowitz et al. (1990, 1993) found the same
false remembering of related but nonpresented information as did Lindsay and
Jennings (2000) and Payne et al. (2002). Excessive reliance on gist information
implies some lling in of missing context and details and leads to errors logically consistent with the theme of an experience but nonetheless incorrect. Indeed, innocent confabulation and narrative smoothing during traumatic
memory retrieval may stem from an overreliance on gist information. When
people face decontextualized fragments, they may infer thematic material based

110

on gist to make sense of their experience. Thus, many memories formed under
stress may in some sense be false, and many false memories may reect gist-based
narrative smoothing. Although the context and details of experience may be
incorrect, the gist may well be accurate, an important consideration for the
validity of memories for traumatic experience.
The relation between emotional stress and memory for central or gist-like
information clearly warrants further exploration. Nonetheless, the research
reviewed here tempts one to conclude that focusing on the central or thematic
features of an emotionally arousing event, perhaps at the expense of memory for
context and detail, may be a characteristic reaction to stress.

Putting It All Together

To summarize, stress, on the whole, appears to disrupt types of memory that
depend on a functional hippocampus and PFC. Disruption of these structures is
associated with impairment in episodic memory. Episodic memory, however,
appears to be composed of dierent parts: memory for context (spatial memory),
memory for events that occur within the specic context, and thematic (gist)
information, derived by inference. Trauma appears to disrupt memory for the
context (e.g., Jacobs & Nadel, 1985, 1998) and the details of experienced events
(e.g., Bartlett, 1932/1995). The thematic content of the memory, on the other
hand, appears to survive intact, perhaps because it is facilitated by an amygdalabased modulation of emotional aspects of experience.
We suggest, speculatively, that this same pattern of impairment and sparing
might characterize how memory is disrupted in PTSD. We have also pointed out
that traumatic memories can apparently suer two rather dierent fates: In one
case, episodic information is disrupted while emotion is preserved; for example,
emotion might be remembered in the form of intrusive fear and recurrent images of the trauma (ashbacks) or as persistent anxiety (free-oating anxiety).
All episodic information associated with the trauma is not necessarily forgotten.
Rather, the memory survives in a fragmented form that renders those fragments
susceptible to narrative smoothing. Narrative smoothing does not reect intentional fabrication but an automatic and inferential use of preserved emotional
and gist information that lls in gaps and smooths over inconsistencies. In the
second case, described in detail in the previous section, memory for traumatic
events is preserved, but the associated emotion is cut o (isolation of aect).
Based on the central idea that trauma results in abnormal functioning in
various brain regions, with consequences for the memory roles these structures
play, we have described patterns of disruption that account for the abnormal
intrusion of traumatic memories into consciousness (i.e., ashbacks), the lack
of recall for traumatic memory (amnesia), and the fragmentation and subsequent
lling in of traumatic memories, depending on the specic circumstances. We

111

hope these ideas stimulate debate and experimentation and, with proper development, eective intervention.

Notes
The writing of this chapter was supported by grants to L. N. from the Flinn Foundation and the McDonnell-Pew Cognitive Neuroscience Program. Address correspondence to any of the authors at: Psychology Department, University of Arizona,
Tucson, AZ 89719 (jdpayne@u.arizona.edu, nadel@u.arizona.edu, or wjj@u.
arizona.edu).
1. But what is trauma? Trauma is an event that often has a specied role in the
etiology of a mental disorder. But what is a mental disorder? Building on the work
of Klein (1978) and Wakeeld (1992), we use the term mental disorder to designate an authentic dysregulation of adaptive emotional, cognitive, or behavioral
systems designed through natural or sexual selection to solve specic adaptive problems. A disorder becomes noticeable when a dysregulated system becomes maladaptive (dysfunctional), that is, consistently fails to solve the adaptive problem for
which it was designed. Accordingly, the presence of a dysregulated system (a mental disorder) may be inferred from measurable maladaptive (dysfunctional) characteristics at an emotional, cognitive, behavioral, physiological, or anatomical level.
Note that this denition of mental disorder diers from that oered by Klein (1978)
and Wakeeld (1992) in two ways. First, the term harmful has been removed from
the phrase harmful dysfunction. The term has been a source of both controversy
and confusion (see McNally [2001] for a brief summary). Second, the notion of
sexual selection has been added to natural selection as a gold standard against
which one might determine normality (see Miller [2000] for a delightful introduction to the notion of sexual selection and the scope of its inuence). It is far beyond
the reach of this chapter to discuss the implications of these moves. Suce it to say
that these distinctions provide theoretically useful predictions about distinctions
between extant judgments of disorder and nondisorder.
2. The nature of traumatic events leading to the emergency room visits was
not described in the Bonne et al. (2001) article. However, subjects were not included in the study if they suered head trauma or physical injury that required
hospitalization.
3. Although PTSD has been associated with persistently low levels of cortisol
(hypocortisolemia) (Yehuda, 1997; Yehuda, Southwick, Nussbaum, Giller, & Mason,
1991), some cases, particularly in the early stages of the disorder, show elevated
levels of stress hormones (e.g., Pitman & Orr, 1990; see Bremner, 2001b).
4. Reversible here implies that memory function eventually returns to baseline
in acute stress studies.
5. These three types of memory are not mutually exclusive. For example, context can be a part of detail memory, helping to distinguish one episode from another, and coherence is a direct outcome of contextual information and details
being combined properly to yield a veridical representation of experience.
6. It behooves the reader to distinguish between the eects of emotion on memories of various forms and the eects of traumatic stress on emotional memories. The

112

data we discuss here apply to the latter. In the section entitled The Emotion and
Memory Literature, we discuss eects of emotion on memory (traditionally conceptualized) and provide a brief critique of the methods and interpretations oered
by that literature.
7. This idea nds some support in the controversial evidence pointing to the
existence of psychogenic amnesia, where trauma memories, or even memory for
ones entire identity, can be temporarily lost (Emilien et al., 2000; Markowitsch,
1999; but see Kihlstrom, 2001).
8. We do not know if these two reactions to trauma reect dierent underlying
brain functions, individual dierences wherein some talk about traumatic events
and others attempt to forget them, the level of stress experienced, the nature of the
stressor (events that inspire guilt, shame, or severe physical threat may be less likely
to be discussed), or some combination of these factors.
9. See Nadel et al. (2002) and Nadel and Payne (2002) for denitions of context
and detail and an explanation of how gist-based memories result from a lack of both.
10. An important distinction in the emotional memory literature is that some of
the material we remember is itself emotional, while other material is emotionally
neutral but encountered while one is emotionally aroused by some other experience.
In our view, the source of emotional arousal is not critical for the pattern of enhanced
gist/impaired detail memory to emerge because elevated stress and HPA activity determine the pattern. Given sucient elevation, regardless of its source, this HPA activity will dierentially inuence hippocampus-based and amygdala-based memory
function, provided that the stress response occurs within a specic timeline.
11. This information on gist and thematic information brings to mind work
on schemas and scriptsrepresentations that describe knowledge people can abstract from common, frequently occurring events (e.g., Abelson, 1981; Mandler,
1984)and recalls the classic studies of Bartlett (1932/1995), who demonstrated
that schemas can inuence retrieval. His classic War of the Ghosts study showed
that people reconstruct their memories based on schemas as they try to ll in missing details and make sense out of incomprehensible information.

References
Abelson, R. P. (1981). Psychological status of the script concept. American Psychologist, 36, 715729.
Adolphs, R., Denburg, N. L., & Tranel, D. (2001). The amygdalas role in long-term
declarative memory for gist and detail. Behavioral Neuroscience, 115, 983992.
Amaral, D. G., Price, J. L., Pitkanen, A., & Carmichael, S. T. (1992). Anatomical
organization of the primate amygdaloid complex. In J. P. Aggleton (Ed.), The
amygdala: Neurobiological aspects of emotion, memory and mental dysfunction
(pp. 166). New York: Wiley-Liss.
Anderson, S. W., Bechara, A., Damasio, H., Tranel, D., & Damasio, A. R. (1999).
Impairment of social and moral behavior related to early damage in human
prefrontal cortex. Nature Neuroscience, 2, 10321037.
Arbel, I., Kadar, T., Silberman, M., & Levy, A. (1994). The eects of long-term corticosterone administration on hippocampal morphology and cognitive performance of middle-aged rats. Brain Research, 657, 227235.

113

Arnsten, A. (1998). Catecholamine modulation of prefrontal cortical cognitive function. Trends in Cognitive Sciences, 2, 436447.
Arnsten, A. (2000). Stress impairs prefrontal cortical function in rats and monkeys:
Role of dopamine D1 and norepinephrine alpha-1 receptor mechanisms. Progress in Brain Research, 126, 183192.
Arnsten, A. F. T. (2001). Dopaminergic and noradrenergic inuences on cognitive
functions mediated by prefrontal cortex. In M. V. Solanto, A. F. T. Arnsten, &
F. X. Castellanos (Eds.), Stimulant drugs and ADHD: Basic and clinical neuroscience
(pp. 185208). London: Oxford University Press.
Aubert, A., Vega, C., Dantzer, R., & Goodall, G. (1995). Pyrogens specically disrupt the acquisition of a task involving cognitive processing in the rat. Brain,
Behavior and Immunity, 9, 129148.
Baddeley, A. (1986). Working memory. Oxford: Clarendon.
Bartlett, F. C. (1932/1995). Remembering: A study in experimental and social psychology. New York: Cambridge University Press.
Bellinger, F. P., Madamba, S., & Siggins, G. R. (1993). Interleukin 1 b inhibits synaptic strength and long-term potentiation in the rat CA1 hippocampus. Brain
Research, 628, 227234.
Bender, B. G., Lerner, J. A., & Poland, J. E. (1991). Association between corticosteroids and psychologic change in hospitalized asthmatic children. Annals of
Allergy, 66, 414419.
Bodno, S. R., Humphreys, A. G., Lehman, J. C., Diamond, D. M., Rose, G. M., &
Meaney, M. J. (1995). Enduring eects of chronic corticosterone treatment on
spatial learning, synaptic plasticity, and hippocampal neuropathology in young
and mid-aged rats. Journal of Neuroscience, 15, 6169.
Bohus, B., & Lissak, K. (1968). Adrenocortical hormones and avoidance behaviour
of rats. International Journal of Neuropharmacology, 7, 301306.
Bonne, O., Brandes, D., Gilboa, A., Gomori, J. M., Shenton, M. E., Pitman, R. K., et al.
(2001). Longitudinal MRI study of hippocampal volume in trauma survivors
with PTSD. American Journal of Psychiatry, 158, 12481251.
Brainerd, C. J., & Reyna, V. F. (1998). Fuzzy-trace theory and childrens false memories. Journal of Experimental Child Psychology, 71, 81129.
Bremner, J. D. (1999). Does stress damage the brain? Biological Psychiatry, 45, 797
805.
Bremner, J. D.. (2001a). A biological model for delayed recall of childhood abuse.
In J. Freyd & A. Deprince (Eds.), Trauma and cognitive science: A meeting of
minds, science, and human experience (pp. 165183). The Haworth Maltreatment & Trauma Press.
Bremner, J. D. (2001b). Hypotheses and controversies related to eects of stress
on the hippocampus: An argument for stress-induced damage to the hippocampus in patients with posttraumatic stress disorder. Hippocampus, 11,
7581.
Bremner, J. D., Krystal, J. H., Charney, D. S., & Southwich, S. M. (1996). Neural
mechanisms in dissociative amnesia for childhood abuse: Relevance to the
current controversy surrounding the False Memory Syndrome. America Journal of Psychiatry, 153, FS7182.

114

Bremner, J. D., Licinio, J., Darnell, A., Krystal, J. H., Owens, M., Southwick, S. M.,
et al. (1997). Elevated CSF corticotropin-releasing factor concentrations in
posttraumatic stress disorder. American Journal of Psychiatry, 154, 624629.
Bremner, J. D., Randall, P. R., Capelli, S., Scott, T., McCarthy, G., & Charney, D. S.
(1995). Decits in short-term memory in adult survivors of childhood abuse.
Psychiatry Research, 59, 97107.
Bremner, J. D., Randall, P. R., Scott, T. M., Bronen, R. A., Delaney, R. C., Seibyl, J. P.,
et al. (1995). MRI-based measurement of hippocampal volume in posttraumatic stress disorder. American Journal of Psychiatry, 152, 973981.
Bremner, J. D., Randall, P., Vermetten, E., Staib, L., Bronen, R. A., Mazure, C. M.,
et al. (1997). MRI-based measurement of hippocampal volume in posttraumatic stress disorder related to childhood physical and sexual abuse: A preliminary report. Biological Psychiatry, 41, 2332.
Bremner, J. D., Scott, T. M., Delaney, R. C., Southwick, S. M., Mason, J. W., Johnson,
D. R., et al. (1993). Decits in short-term memory in post-traumatic stress disorder. American Journal of Psychiatry, 150, 10151019.
Bremner, J. D., Shobe, K. K., & Kihlstrom, J. F. (2000). False memories in women
with self-reported childhood sexual abuse: An empirical study. Psychological
Science, 11, 333337.
Bremner, J. D., Southwick, S. M., & Charney, D. S. (1997). Neuroanatomical correlates of the eects of stress on memory: Relevance to the validity of memories
of childhood abuse. In P. S. Appelbaum, L. A. Uyehara, & M. R. Elin (Eds.),
Trauma and memory (pp. 6192). London: Oxford.
Bremner, J. D., Southwick, S. M., & Charney, D. S. (1999). The neurobiology of PTSD:
An integration of animal and human research. In J. D. Bremner & P. Saigh
(Eds.), Posttraumatic stress disorder: A comprehensive text (pp 103143). New
York: Allyn and Bacon.
Bremner, J. D. Southwick, S. M., Johnson, D. R., Yehuda, R., & Charney, D. S. (1993).
Childhood physical abuse in combat-related posttraumatic stress disorder.
American Journal of Psychiatry, 150, 235239.
Bremner, J. D., & Vermetten, E. (2001). Stress and development: Behavioral and
biological consequences. Development and Psychopathology, 13, 473489.
Breslau, N., Davis, G. C., Andreski, P., & Peterson, E. (1991). Traumatic events and
post-traumatic stress disorder in an urban population of young adults. Archives
in General Psychiatry, 48, 216222.
Buchanan, T. W., & Lovallo, W. R. (2001). Enhanced memory for emotional material following stress-level cortisol treatment in humans. Psychoneuroendocrinology, 26, 307317.
Buckner, R. L., Logan, J., Donaldson, D. I., & Wheeler, M. E. (2000). Cognitive neuroscience of episodic memory encoding. Acta Psychologica, 105, 127139.
Burke, A., Heuer, F., & Reisberg, D. (1992). Remembering emotional events.
Memory & Cognition, 20, 277290.
Cahill, L., Babinsky, R., Markowitsch, H. J., & McGaugh, J. L. (1995). The amygdala
and emotional memory. Nature, 377, 295296.
Cahill, L., & McGaugh, J. L. (1998). Mechanisms of emotional arousal and lasting
declarative memory. Trends in Neuroscience, 21, 294299.

115

Cahill, L., Prins, B., Weber, M., & McGaugh, J. L. (1994). Beta adrenergic activation and memory for emotional events. Nature, 371, 702704.
Caldji, C., Francis, D., Sharma, S., Plotsky, P. M., & Meaney, M. J. (2000). The eects
of early rearing environment on the development of GABA-A and central benzodiazepine receptor levels and novelty-induced fearfulness in the rat. Neuropsychopharmacology, 22, 219229.
Cameron, H. A., & Gould, E. (1994). Adult neurogenesis is regulated by adrenal steroids in the dentate gyrus. Neuroscience, 61, 203209.
Cameron, H. A., & McKay, R. (1999). Restoring production of hippocampal neurons in old age. Nature Neuroscience, 2, 894897.
Cannon, W. B. (1929). Bodily changes in pain, hunger, fear, and rage. 2nd ed. New
York: Appleton.
Christianson, S.-. (1992). Emotional stress and eyewitness memory: A critical review. Psychological Bulletin, 112, 284309.
Christianson, S.-., & Loftus, E. F. (1987). Memory for traumatic events. Applied
Cognitive Psychology, 1, 225239.
Christianson, S.-A, & Loftus, E. F. (1990). Some characteristics of peoples traumatic memories. Bulletin of the Psychonomic Society, 28, 195198.
Christianson, S.-., & Loftus, E. F. (1991). Remembering emotional events: The fate
of detailed information. Cognition & Emotion, 5, 81108.
Christianson, S.-., Loftus, E. F., Homan, H., & Loftus, G. R. (1991). Eye xations
and memory for emotional events. Journal of Experimental Psychology: Learning, Memory, and Cognition, 17, 693701.
Christianson, S.-., & Safer, M. A. (1995). Emotional events and emotions in autobiographical memories. In D. C. Rubin (Ed.), Remembering our past: Studies in autobiographical memory (pp. 218243). Cambridge: Cambridge University Press.
Clark, L. D., Bauer, W., & Cobb, S. (1952). Preliminary observations on mental disturbances occurring in patients under therapy with cortisone and ACTH. New
England Journal of Medicine, 246, 205216.
Clark, M. S., Milberg, S., & Ross, J. (1983). Arousal cues arousal-related materials
in memory: Implications for understanding eects of mood on memory. Journal of Verbal Learning and Verbal Behavior, 22, 633649.
Clark, R. E., Zola, S. M., & Squire, L. R. (2000). Impaired recognition memory in rats
after damage to the hippocampus. Journal of Neuroscience, 20, 88538860.
Conrad, C. D., & Roy, E. J. (1996). Dentate gyrus destruction and spatial learning
impairment after corticosteroid removal in young and middle-aged rats. Hippocampus, 5, 115.
Convit, A., DeLeon, M. J., Tarshish, C., De Santi, S., Kluger, A., Rusinek, H., et al.
(1995). Hippocampal volume losses in minimally impaired elderly. Lancet, 345,
266266.
Costello, E. J., Erkanli, A., Fairbank, J. A., & Angold, A. (2002). The prevalence of
potentially traumatic events in childhood and adolescence. Journal of Traumatic
Stress, 15, 99112.
Cunningham, A. J., Murray, C. A., ONeill, L. A., & OConnor, J .J. (1996). Interleukin
1-b (IL-1b) and tumor necrosis factor (TNF) inhibit long-term potentiation in
the rat dentate gyrus in vitro. Neuroscience Letters, 203, 1720.

116

Curran, T., Schacter, D. L., Norman, K. A., & Galluccio, L. (1997). False recognition after a right frontal lobe infarction: Memory for general and specic information. Neuropsychologia, 35, 10351049.
Davidson, J. R. T., Roth, S., & Newman, E. (1991). Treatment of posttraumatic stress
disorder with uxetine. Journal of Traumatic Stress, 4, 419423.
Davis, M. (1992). The role of the amygdala in fear and anxiety. Annual Review of
Neuroscience, 15, 353375.
Davis, M., & Shui, C. (1999). The extended amygdala: Are the central nucleus of
the amygdala and the bed nucleus of the stria terminalis dierentially involved
in fear versus anxiety? In: J. McGinty (Ed.), Advancing from the ventral striatum
to the extended amygdala: Implications for neuropsychiatry and drug use: In honor
of Lennart Heimer (pp. 281291). New York: New York Academy of Sciences.
Delbecq-Derouesne, J., Beauvois, M. F., & Shallice, T. (1990). Preserved recall versus impaired recognition. Brain, 113, 10451074.
de Quervain, D. J.-F., Roozendaal, B., Nitsch, R. M., McGaugh, J. L., & Hock, C.
(2000). Acute cortisone administration impairs retrieval of long-term declarative memory in humans. Nature Neuroscience, 3, 313314.
Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in
immediate recall. Journal of Experimental Psychology, 58, 1722.
De Kloet, E. R., Oitzl, M. S., & Joels, M. (1999). Stress and cognition: Are corticosteroids good or bad guys? Trends in Neuroscience, 22, 422426.
Deptula, D. J. (1983). Lateralized brain dysfunction in depression: An analysis of
memory. Dissertation Abstracts International.
Diamond, D. M., Bennet, M. C., Fleshner, M., & Rose, G. M. (1992). Inverted U relationship between the level of peripheral corticosterone and the magnitude of
hippocampal primed burst potentiation. Hippocampus, 2, 421430.
Diamond, D. M., Fleshner, M., Ingersoll, N. & Rose, G. M. (1996). Psychological stress
impaires spatial working memory: Relevance to electrophysiological studies of
hippocampal function. Behavioral Neuroscience, 110, 661672.
Diamond, D. M., Fleshner, M., & Rose, G. M. (1994). Psychological stress repeatedly blocks hippocampal primed burst potentiation in behaving rats.
Behavioural Brain Research, 62, 19.
Diamond, D. M., Park, C. R., Heman, K. L., & Rose, G. M. (1999). Exposing rats to a
predator impairs spatial working memory in the radial arm water maze. Hippocampus, 9, 542552.
Diamond, D. M., & Rose, G. M. (1994). Stress impairs LTP and hippocampaldependent memory. Annals of New York Academy of Sciences, 746, 411414.
Emilien, G., Penasse, C., Charles, G., Martin, D., Lasseaux, L., & Waltregny, A.
(2000). Post-traumatic stress disorder: Hypotheses from clinical neuropsychology and psychopharmacology research. International Journal of Psychiatry in
Clinical Practice, 4, 318.
Fairbank, J. A., Ebert, L., & Caddell, J. M. (2001). Posttraumatic stress disorder. In
H. E. Adans & P. B. Sutker, (Eds.), (pp. 183209). Comprehensive handbook of
psychopathology (3rd ed., pp. 183209). New York: Plenum.
Fanselow, M. S., & LeDoux, J. E. (1999). Why we think plasticity underlying Pavlovian fear conditioning occurs in the basolateral amygdala. Neuron, 23, 229232.

117

Feldman, S., & Conforti, N. (1980). Participation of the dorsal hippocampus in the
glucocorticoid negative-feedback eect on adrenocortical activity. Neuroendocrinology, 30, 5255.
Feldman, S., & Conforti, N. (1985). Modications of adrenocortical responses following frontal cortex stimulation in rats with hypothalamic dierentiations and
medial forebrain bundle lesions. Neuroscience, 15, 10451047.
Filipini, D., Gijsbers, K., Birminghan, M. K., & Dubrovsky, B. (1991). Eects of adrenal steroids and their reduced metabolites on hippocampal long-term potentiation. Journal of Steroid Biochemistry and Molecular Biology, 40, 8792.
Fox, M. L., & Dwyer, D. J. (2000). Reducing stress by increasing control. Clinical
Leadership and Management Reviews, 14, 114117.
Foy, M. R., Stanton, M. E., Levine, S., & Thompson, R. F. (1987). Behavioral stress
impairs long-term potentiation in rodent hippocampus. Behavioral Neural
Biology, 48, 138149.
Freud, S. (1926). Inhibitions, symptoms, and anxiety. London: Hogarth Press and the
Institute of Psycho-analysis.
Fukuzako, H., Fukuzako, T., Hashiguchi, T., Hokazono, Y., Takeuchi, K., Hirakawa,
K., et al. (1996). Reduction in hippocampal formation volume is caused mainly
by its shortening in chronic schizophrenia: Assessment by MRI. Biological Psychiatry, 39, 938945.
Garcia, R., Musleh, W., Tocco, G., Thompson, R. F., & Baudry, M. (1997). Timedependent blockade of STP and LTP in hippocampal slices following acute stress
in mice. Neuroscience Letters, 233, 4144.
Gibertini, M., Newton, C., Friedman, H., & Klein, T. W. (1995). Spatial learning
impairment in mice infected with Legionella pneumophila or administered
exogenous interleukin-1b. Brain, Behavior and Immunity, 9, 113128.
Gilbertson, M. W., Gurvits. T. V., Lasko, N. B., Orr, S. P., & Pitman, R. K. (2001).
Multivariate assessment of explicit memory function in combat veterans with
posttraumatic stress disorder. Journal of Traumatic Stress, 14, 413432.
Golier, J. A., Yehuda, R., & Southwick, S. M. (1997). Memory and posttraumatic
stress disorder. In P. S. Appelbaum, L. A. Uyehara, & M. R. Elin (Eds.), Trauma
and memory (pp. 225242). London: Oxford.
Goodman, G. G., Hirschman, J. E., Hepps, D. H., & Rudy, L. (1991). Childrens
memory for stressful events. Merrill-Palmer Quarterly, 37, 109158.
Gould, E., & Gross, C. G. (2002). Neurogenesis in adult mammals: Some progress
and problems. Journal of Neuroscience, 22, 619623.
Gould, E., & Tanapat, P. (1999). Stress and hippocampal neurogenesis. Biological
Psychiatry, 46, 14721479.
Gould, E., Tanapat, P., McEwen, B. S., Flugge, G., & Fuchs, E. (1998). Proliferation
of granule cell precursors in the dentate gyrus of adult monkeys is diminished
by stress. Proceedings of the National Academy of Science USA, 95, 31683171.
Gray, J. A. (2000). The neuropsychology of anxiety: An enquiry into the function of the
septo-hippocampal system. New York: Oxford.
Gray, M. J., & Lombardo, T. W. (2001). Complexity of trauma narratives as an index
of fragmented memory in PTSD: A critical analysis. [Special issue]. Applied
Cognitive Psychology, 15, S171S186.

118

Greidanus, T. J. B. (1970). Eects of steroids on extinction of an avoidance response

in rats. A structure-activity relationship study. Progress in Brain Research, 32,
185191.
Gurvits, T. G., Shenton, M. E., Hokama, H., Ohta, H., Lasko, N.B., Gilbertson, M.W.,
et al. (1996). Magnetic resonance imaging study of hippocampal volume in
chronic, combat-related posttraumatic stress disorder. Biological Psychiatry, 40,
10911099.
Hamann, S. (2001). Cognitive and neural mechanisms of emotional memory. Trends
in Cognitive Sciences, 5, 394400.
Hamann, S. B., Ely, T. D., Grafton, S. T., & Kilts, C. D. (1999). Amygdala activity
related to enhanced memory for pleasant and aversive stimuli. Nature Neuroscience, 2, 289293.
Harvey, A. G., & Bryant, R. A. (1998). Relationship of acute stress disorder and posttraumatic stress disorder following motor vehicle accidents. Journal of Consulting and Clinical Psychology, 66, 507512.
Henson, R. N. A., Shallice, T., & Dolan, R. J. (1999). Right prefrontal cortex and
episodic memory retrieval: A functional MRI test of the monitoring hypothesis.
Brain, 122, 13671381.
Heuer, F., & Reisberg, D. (1990). Vivid memories of emotional events: The accuracy of remembered minutiae. Memory & Cognition, 18, 496506.
Heuer, F., & Reisberg, D. (1992). Emotion, arousal, and memory for detail. In
S. Christianson (Ed.), The handbook of emotion and memory: Research and theory
(pp. 151180). Hillsdale, NJ: Erlbaum.
Hirsh, R. (1974). The hippocampus and contextual retrieval of information from
memory: A theory. Behavioral Biology, 12, 421444.
Horner, M. D., & Hamner, M. B. (2002). Neurocognitive functioning in posttraumatic stress disorder. Neuropsychology Review, 12, 1530.
Inoue, T., Tsuchiya, K., & Koyama, T. (1994). Regional changes in dopamine and
serotonin activation with various intensity of physical and psychological stress
in the rat brain. Pharmacology, Biochemistry, & Behavior, 49, 911920.
Jacobs, W. J., & Nadel, L. (1985). Stress-induced recovery of fears and phobias. Psychological Review, 92, 512531.
Jacobs, W. J., & Nadel, L. (1998). Neurobiology of reconstructed memory. Psychology, Public Policy and Law, 4, 1110.
Jacobs, W. J., Nadel, L., & Hayden, V. (1992). Anxiety disorders. In D. J. Stein &
J. Young (Eds.), Cognitive science and clinical disorders. (pp. 211233) New York:
Academic Press.
Jacobs, W. J., Laurance, H. E., Thomas, K. G. F., Luzcak, S. E., & Nadel, L. (1996).
On the veracity and variability of recovered traumatic memory. Traumatology,
2. Available http://www.rdz.stjohns.edu: 80/trauma/traumaj.html
Jacobson, L., & Sapolsky, R. M. (1991). The role of the hippocampus in feedback
regulation of the hypothalamic-pituitary-adrenal axis. Endocrinology Reviews,
12, 118134.
Jenkins, M. A., Langlais, P. J., Delis, D., & Cohen, R. (1998). Learning and memory
in rape victims with posttraumatic stress disorder. American Journal of Psychiatry, 155, 278.

119

Joseph, R. (1996). Neuropsychiatry, neuropsychology, and clinical neuroscience. Baltimore: Williams & Wilkins.
Kaehler, S. T., Singewald, N., Sinner, C., Thurnher, C., & Philippu, A. (2000). Conditioned fear and inescapable shock modify the release of serotonin in the locus
coeruleus. Brain Research, 859, 249254.
Kanaley, J. A., Weltman, J. Y., Pieper, K. S., Weltman, A., & Hartman, M. L. (2001).
Cortisol and growth hormone responses to exercise at dierent times of day.
Journal of Clinical Endocrinology and Metabolism, 86, 28812889.
Kaspi, S. P., McNally, R. J., & Amir, N. (1995). Cognitive processing of emotional
information in posttraumatic stress disorder. Cognitive Therapy and Research,
19, 319330.
Keenan, P. A., Jacobson, M. W., Soleymani, R. M., & Newcomer, J. W. (1995). Commonly used therapeutic doses of glucocorticoids impair explicit memory. Annals
of the New York Academy of Sciences, 761, 400402.
Kessler, R. C., Sonnega, A., Bromet, E., Hughes, M., & Nelson, C. B. (1995). Posttraumatic stress disorder in the national comorbidity survey. Archives of General Psychiatry, 52, 10481060.
Kihlstrom, J. F. (2001). Dissociative disorders. In P. B. Sutker & H. E. Adams (Eds.),
Comprehensive handbook of psychopathology (3rd ed.). New York: Kluwer Academic/Plenum Publishers.
Kim, J. J., & Diamond, D. M. (2002). The stressed hippocampus, synaptic plasticity
and lost memories. Nature Reviews Neuroscience, 3, 453462.
Kirschbaum, C., Wolf, O. T., May, M., Wippich, W., & Hellhammer, D. H. (1996).
Stress- and treatment-induced elevations of cortisol levels associated with impaired declarative memory in healthy adults. Life Sciences, 58, 14751483.
Klein, D. F. (1978). A proposed denition of mental illness. In R. L. Spitzer & D. F.
Klein(Eds.), Critical issues in psychiatric diagnosis (pp. 4171). New York: Raven
Press.
Kleinsmith, L. J., & Kaplan, S. (1963). Paired-associate learning as a function of arousal
and interpolated interval. Journal of Experimental Psychology, 65, 190193.
Kleinsmith, L. J., & Kaplan, S. (1964). The interaction of arousal and recall interval
in nonsense syllable paired-associate learning. Journal of Experimental Psychology, 67, 124126.
Lang, P. J., & Cuthbert, B. N. (1984). Aective information processing and the assessment of anxiety. Journal of Behavioral Assessment, 6, 369395.
Laurance, H. E., Hardt, O. M., Nadel, L., & Jacobs, W. J. (2001). Stress aects spatiotemporal processing: Support for a model of traumatic memory. Poster session
presented at the 109th meeting of the American Psychological Association.
LeDoux, J. E. (1992). Emotion and the amygdala. In J. P. Aggleton (Ed.), The
amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction
(pp. 339351). New York: Wiley-Liss.
LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review of Neuroscience,
23, 155184.
Lemieux, A. M., & Coe, C. L. (1995). Abuse-related posttraumatic stress disorder:
Evidence for chronic neuroendocrine activation in women. Psychosomatic Medicine, 57, 105115.

120

Lindsay, A. L., & Jennings, J. M. (2000). Chronic anxiety and false memory: Potential
implications for medial temporal lobe functioning. Poster session presented at the
2000 meeting of the Society for Neuroscience.
Linthorst, A. C., Flachskamm, C., Holsboer, F., & Reul, J. M. (1994). Local administration of recombinant human interleukin-1b in the rat hippocampus increases
serotonergic neurotransmission, hypothalamic-pituitary-adrenocortical axis
activity and body temperature. Endocrinology, 135, 520532.
Litz, B. T., & Gray, M. J. (2002). Emotional numbing in posttraumatic stress disorder: Current and future research directions. Australian and New Zealand Journal of Psychiatry, 36, 198204.
Loftus, E. F., & Burns, T. (1982). Mental shock can produce retrograde amnesia.
Memory & Cognition, 10, 318323.
Loftus, E. F., Loftus, G., & Messo, J. (1987). Some facts about weapon focus. Law
and Human Behavior, 11, 5562.
Luine, V., Villegas, M., Martinez, C., & McEwen, B. S. (1994). Repeated stress causes
reversible impairments of spatial memory performance. Brain Research, 639,
167170.
Lupien, S. J., de Leon, M., de Santi, S., Convit, A., Tarshish, C., Nair, N.P.V., et al.
(1998). Cortisol levels during human aging predict hippocampal atrophy and
memory decits. Nature Neuroscience, 1, 6973.
Lupien, S. J., Gaudreau, S., Tchiteya, B. M., Maheu, F., Sharma, S., Nair, N. P. V.,
et al. (1997). Stress-induced declarative memory impairments in healthy
elderly subjects: Relationship to cortisol reactivity. Journal of Clinical Endocrinology and Metabolism, 82, 20702075.
Lupien, S. J., Lecours, A. R., Lussier, I., Schwartz, G., Nair, N. P. V., & Meaney, M. J.
(1994). Basal cortisol levels and cognitive decits in human aging. Journal of
Neuroscience, 14, 28932903.
Lupien, S. J., & Lepage, M. (2001). Stress, memory, and the hippocampus: Cant live
with it, cant live without it. Behavioural Brain Research, 127, 137158.
Lupien, S. J., & McEwen, B. S. (1997). The acute eects of corticosteroids on cognition: Integration of animal and human model studies. Brain Research Reviews,
24, 127.
Magarinos, A. M., & McEwen, B. S. (1995). Stress-induced atrophy of apical dendrites of hippocampal CA3c neurons: Involvement of glucocorticoid secretion
and excitatory amino acid receptors. Neuroscience, 69, 8998.
Magarinos, A. M., McEwen, B. S., Flugge, G., & Fuchs, E. (1996). Chronic psychosocial stress causes apical dendritic atrophy of hippocampal CA3 pyramidal neurons in subordinate tree shrews. Journal of Neuroscience, 16, 3534
3540.
Magarinos, A. M., Verdugo, J. M., & McEwen, B. S. (1997). Chronic restraint stress
alters synaptic terminal structure in hippocampus. Proceedings of the National
Academy of Sciences, 94, 1400214008.
Maier, S. F., & Watkins, L. R. (1998). Stressor controllability, anxiety, and serotonin.
Cognitive Therapy Research, 22, 595613.
Mandler, J. M. (1984). Stories, scripts, and scenes: Aspects of schema theory. Hillsdale,
NJ: Erlbaum.

121

Markowitsch, H. J. (1999). Functional neuroimaging correlates of functional amnesia. Memory, 5/6, 561583.
Mauri, M., Sinforiani, E., Bono, G., Vignati, F., Berselli, M. E., Attanasio, R., et al.
(1993). Memory impairment in Cushings disease. Acta Neurologica Scandinavica,
87, 5255.
McEwen, B. S. (2000). The neurobiology of stress: From serendipity to clinical relevance. Brain Research, 886, 172189.
McEwen, B. S. (2001). Neurobiology of interpreting and responding to stressful
events: Paradigmatic role of the hippocampus. In B. S. McEwen (Ed.). Handbook
of physiology. Section 7: The endocrine system. Volume 4: Coping with the environment: Neural and endocrine mechanisms. New York: Oxford University Press.
McEwen, B. S., Albeck, D., Cameron, H., Chao, H. M., Gould, E., Hastings, N., et al.
(1995b). Stress and the brain: A paradoxical role for adrenal steroids. In G. D.
Litwack (Ed.), Vitamins and hormones (pp. 371402). New York: Academic Press.
McEwen, B. S., Conrad, C. D., Kuroda, Y., Frankfurt, M., Magarinos, A. M., &
McKittrick, C. (1997). Prevention of stress-induced morphological and cognitive consequences. European Neuropsychopharmacology, 7(suppl.), 322328.
McEwen, B. S., & Sapolsky, R. M. (1995). Stress and cognitive function. Current
Opinion in Neurobiology, 5, 205216.
McEwen, B. S., Weiss, J. M., & Schwartz, L. S. (1968). Selective retention of corticosterone by limbic structure in rat brain. Nature, 220, 911912.
McGaugh, J. L. (2000). MemoryA century of consolidation. Science, 287, 248251.
McGaugh, J. L. (2002). The amygdala regulates memory consolidation. In L. R.
Squire & D. L. Schacter (Eds.), Neuropsychology of memory. 3rd ed. New York:
Guilford Press.
McGaugh, J. L., & Roozendaal, B. (2002). Role of adrenal stress hormones in forming
lasting memories in the brain. Current Opinion in Neurobiology, 12, 205210.
McKittrick, C.nR., Magarinos, A. M., Blanchard, D. C., Blanchard, R. J., McEwen,
B. S., & Sakai, R. R. (2000). Chronic social stress reduces dendritic arbors in
CA3 of hippocampus and decreases binding to serotonin transporter sites. Synapse, 36, 8594.
McNally, R. J. (1998). Experimental approaches to cognitive abnormality in posttraumatic stress disorder. Clinical Psychology Review, 18, 971982.
McNally, R. J. (2001). On Wakeelds harmful dysfunction analysis of mental disorder. Behaviour Research & Therapy 39, 309314.
McNally, R. J., Lasko, N. B., Macklin, M. L., & Pitman, R. K. (1995). Autobiographical memory disturbance in combat-related posttraumatic stress disorder.
Behaviour Research and Therapy, 33, 619630.
McNally, R .J., Litz, B. T., Prassas, A., Shin, L. M., & West, F. W. (1994). Emotional
priming of autobiographical memory in post-traumatic stress disorder. Cognition & Emotion, 8, 351367.
McNally, R. J., & Shin, L. M. (1995). Association of intelligence with severity of posttraumatic stress disorder symptoms in Vietnam combat veterans. American
Journal of Psychiatry, 152, 936938.
Melo, B., Winocur, G., & Moscovitch, M. (1999). False recall and false recognition:
An examination of the eects of selective and combined lesions to the medial

122

temporal lobe/diencephalon and frontal lobe structures. Cognitive Neuropsychology, 16, 343359.
Mesches, M. H., Fleshner, M., Heman, K. L., Rose, G. M., & Diamond, D. M. (1999).
Exposing rats to a predator blocks primed burst potentiation in the hippocampus in vitro. Journal of Neuroscience, 19, RC18(15).
Metcalfe, J., & Jacobs, W. J. (1996). An interactive hot system/cool system view of
memory under stress. PTSD Research Quarterly, 7, 16.
Metcalfe, J., & Jacobs W. J. (1998). Emotional memory: The eects of stress on cool
and hot memory systems. In D. L. Medin (Ed.), The psychology of learning and
motivation: Vol. 38. Advances in research and theory (pp. 187222). San Diego,
CA: Academic Press.
Metcalfe, J., & Jacobs, W. J. (2000). Can hot human emotions be captured by cool
mathematical models? In E. Tulving (Ed.), Memory, consciousness, and the brain:
The Tallinn Conference (pp. 228242). Philadelphia: Psychology Press.
Miller, G. F. (2000). The mating mind: How sexual choice shaped the evolution of
human nature. New York: Doubleday.
Mitchell, K. J., Johnson, M. K., Raye, C. L., & DEsposito, M. (2000). fMRI evidence
of age-related hippocampal dysfunction in feature binding in working memory.
Cognitive Brain Research, 10, 197206.
Mitchell, K. J., Johnson, M. K., Raye, C. L., Mather, M., & DEsposito, M. (2000).
Aging and reective processes of working memory: Binding and test load
decits. Psychology and Aging, 15, 527541.
Mitchell, S. A., & Black, M. J. (1995). Freud and beyond: A history of modern psychoanalytic thought. New York: BasicBooks.
Moghaddam, B., Bolinao, M. L., Stein-Behrens, B., & Sapolsky, R. (1994). Glucocorticoids mediate the stress-induced extracellular accumulation of glutamate.
Brain Research, 655, 251254.
Nadel, L. (1991). The hippocampus and space revisited. Hippocampus, 1, 221
229.
Nadel, L. (1994). Multiple memory systems: What and why. An update. In D. L.
Schacter & E. Tulving (Eds.), Memory systems 1994 (pp. 3963). Cambridge,
MA: MIT Press.
Nadel, L., & Jacobs, W. J. (1996). The role of the hippocampus in PTSD, panic, and
phobia. In N. Kato (Ed.), The hippocampus: Functions and clinical relevance (pp. 455
463). Amsterdam: Elsevier Science.
Nadel, L., & Jacobs, W. J. (1998). Traumatic memory is special. Current Directions
in Psychological Science, 7, 154157.
Nadel, L., & Payne, J. D. (2002). The hippocampus, waynding, and episodic
memory. In P. Sharp (Ed.), The neural basis of navigation: Evidence from single cell
recording (pp. 235247). Boston: Kluwer Academic Publishers.
Nadel, L., Payne, J. D., & Jacobs, W. J. (2002). The relationship between episodic
memory and context: Clues from memory errors made while under stress.
Physiological Research, 51, S3S11.
Nadel, L., & Willner, J. (1980). Context and conditioning: A place for space. Physiological Psychology, 8, 218228.
Neisser, U. (1981). John Deans memory: A case study. Cognition, 9, 122.

123

Newcomer, J. W., Craft, S., Hershey, T., Askins, K., & Bardgett, M. E. (1994). Glucocorticoid-induced impairment in declarative memory performance in adult
humans. Journal of Neuroscience, 14, 20472053.
Newcomer, J. W., Selke, G., Melson, A. K., Hershey, T., Craft, S., Richards, K., et al.
(1999). Decreased memory performance in healthy humans induced by stresslevel cortisol treatment. Archives of General Psychiatry, 56, 527533.
Nyberg, L., & Cabeza, R. (2000). Brain imaging of memory. In E. Tulving & F. I. M.
Craik (Eds.), The Oxford handbook of memory (pp. 501519). Oxford: Oxford University Press.
OCarroll, R. E., Drysdale, E., Cahill, L., Shajahan, P., & Ebmeier, K. P. (1999). Stimulation of the noradrenergic system enhances and blockage reduces memory for
emotional material in man. Psychological Medicine, 29, 10831088.
Ochsner, K. N., & Schacter, D. L. (2000). A social cognitive neuroscience approach
to emotion and memory. In J. C. Borod (Ed.), The neuropsychology of emotion
(pp. 163193). New York: Oxford University Press.
Ohl, F., & Fuchs, E. (1999). Dierential eects of chronic stress on memory processes
in the tree shrew. Cognitive Brain Research, 7, 379387.
OKeefe, J., & Nadel, L. (1978). The hippocampus as a cognitive map. Oxford: Clarendon
Press.
Osborne, J. (1972). Short- and long-term memory as a function of individual dierences in arousal. Perceptual and Motor Skills, 34, 587593.
Otten, L. J., & Rugg, M. D. (2002). The birth of memory. Trends in Neurosciences, 25,
279281.
Pani, L., Porcella, A., & Gessa, G. L. (2000). The role of stress in the pathophysiology of the dopaminergic system. Molecular Psychiatry, 5, 1421.
Parkin, A. J., Bindschaedler, C., Harsent, L., & Metzler, C. (1996). Pathological false
alarm rates following damage to the left frontal cortex. Brain and Cognition, 32,
1427.
Payne, J. D., Nadel, L., Allen, J. J. B., Thomas, K. G. F., & Jacobs, W. J. (2002).
The eects of experimentally-induced stress on false recognition. Memory, 10,
16.
Pitman, R., & Orr, S. (1990). Twenty-four hour urinary cortisol and catecholamine
excretion in combat-related posttraumatic stress disorder. Biological Psychiatry, 27, 245247.
Plihal, W., & Born, J. (1999). Memory consolidation in human sleep depends on
inhibition of glucocorticoid release. Neuroreport, 10, 27412747.
Porter, S., & Birt, A. (2001). Is traumatic memory special? A comparison of traumatic memory characteristics with memory for other emotional life experiences. Applied Cognitive psychology, 15, S101S117.
Prince, C. R., & Anisman, H. (1990). Situation specic eects of stressor controllability on plasma corticosterone changes in mice. Pharmacology, Biochemistry,
and Behavior, 37, 613621.
Rasmusson, A. M., Lipschitz, D. S., Wang, S., Hu, S., Vojvoda, D., Bremner, J. D.,
et al. (2001). Increased pituitary and adrenal reactivity in premenopausal
women with posttraumatic stress disorder. Biological Psychiatry, 50, 965977.

124

Raye, C. L., Johnson, M. K., Mitchell, K. J., Nolde, S. F., & DEsposito, M. (2000). fMRI
investigations of left and right PFC contributions to episodic remembering.
[Special issue]. Psychobiology, 28, 197206.
Reiman, E. M., Lane, R. D., Ahern, G. L., Schwartz, G. E., & Davidson, R. J. (2000).
Positron emission tomography in the study of emotion, anxiety, and anxiety
disorders. In R. D. Lane & L. Nadel (Eds.), Cognitive neuroscience of emotion
(pp. 389406). Oxford: Oxford University Press.
Reyna, V. F., & Kiernan, B. (1994). Development of gist versus verbatim memory in
sentence recognition: Eects of lexical familiarity, semantic content, encoding
instructions, and retention interval. Developmental Psychology, 30, 178191.
Roediger, H. L., & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning,
Memory & Cognition, 21, 803814.
Roozendaal, B. (2000). Glucocorticoids and the regulation of memory consolidation. Psychoneuroendocrinology, 25, 213238.
Roozendaal, B., & McGaugh, J. L. (1996). Amygdaloid nuclei lesions dierentially
aect glucocorticoid-induced memory enhancement in an inhibitory avoidance
task. Neurobiology of Learning and Memory, 65, 18.
Roozendaal, B., & McGaugh, J. L. (1997). Glucocorticoid receptor agonist and antagonist administration into the basolateral but not central amygdala modulates memory storage. Neurobiology of Learning and Memory, 67, 176179.
Rosenbaum, R. S., Priselac, S., Kohler, S., Black, S. E., Gao, F, Nadel, L., et al. (2000).
Remote spatial memory in an amnesic person with extensive bilateral hippocampal lesions. Nature Neuroscience, 3, 10441048.
Rosenkranz, J. A., & Grace, A. A. (2002). Cellular mechanisms of infralimbic
and prelimbic prefrontal cortical inhibition and dopaminergic modulation
of basolateral amygdala neurons in vivo. Journal of Neuroscience, 22, 324
337.
Rugg, M. D., Fletcher, P. C., Chua, P. M.-L., & Dolan, R. J. (1999). The role of the
prefrontal cortex in recognition memory and memory for source: An fMRI
study. NeuroImage, 10, 520529.
Saigh, P. A., & Bremner, J. D. (1999). The history of posttraumatic stress disorder.
In P. A. Saigh & J. D. Bremner (Eds.), Posttraumatic stress disorder: A comprehensive text. Needham Heights, MA: Allen & Bacon.
Sapolsky, R. M. (1996). Why stress is bad for your brain. Science, 273, 749750.
Sapolsky, R. M. (2000). Glucocorticoids and hippocampal atrophy in neuropsychiatric disorders. Archives of General Psychiatry, 57, 925935.
Sapolsky, R. M., Krey, L. C., & McEwen, B. S. (1985). Prolonged glucocorticoid exposure reduces hippocampal neuron number: implications for aging. Journal
of Neuroscience, 5, 12221227.
Sapolsky, R. M., & Meaney, M. J. (1986). Maturation of the adrenocortical stress
response: Neuroendocrine control mechanisms and the stress hyporesponsive
period. Brain Research, 396, 6476.
Schacter, D. L., & Tulving, E. (Eds.), (1994). Memory systems 1994. Cambridge, MA:
MIT Press.

125

Scoville, W. B., & Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. Journal of Neurology, Neurosurgery, and Psychiatry, 20, 1121.
Selye, H. (1956). The stress of life. New York: McGraw-Hill.
Sheline, Y. I., Sanghavi, M., Mintun, M. A., & Gado, M. H. (1999). Depression duration but not age predicts hippocampal volume loss in medically healthy women
with recurrent major depression. Journal of Neuroscience, 19, 50345043.
Sheline, Y. I., Wang, P. W., Gado, M. H., Csernansky, J. C., & Vannier, M. W (1996).
Hippocampal atrophy in recurrent major depression. Proceedings from the National Academy of Sciences USA, 93, 39083913.
Shimamura, A. P. (1997). Recollection: Perspectives on reinstated memory and
child trauma. In J. D. Read & D. S. Lindsay (Eds.), Recollections of trauma
(pp. 253266). New York: Plenum Press.
Shors, T. J., Gallegos, R. A., & Breindl, A. (1997). Transient and persistent consequences of acute stress on long-term potentiation (LTP), synaptic ecacy, theta
rhythms and bursts in area CA1. Synapse, 26, 209217.
Southwick, S. M., Davis, M., Horner, B., Cahill, L., Morgan, C. A., Golc, P. E., et al.
(2002). Relationship of enhanced norepinephrine activity during memory consolidation to enhanced long-term memory in humans. American Journal of Psychiatry, 159, 14201422.
Stanny, C. J., & Johnson, T. C. (2000). Eects of stress induced by a simulated shooting on recall by police and citizen witnesses. American Journal of Psychology,
113, 359386.
Starkman, M. N., Gebarski, S. S., Berent, S., & Schteingart, D. E. (1992). Hippocampal formation volume, memory dysfunction, and cortisol levels in patients with
Cushings syndrome. Biological Psychiatry, 32, 756765.
Starkman, M. N., Giordani, B, Gebrski, S. S., Berent, S., Schork, M. A., & Schteingart,
D. E. (1999). Decrease in cortisol reverses human hippocampal atrophy following treatment of Cushings disease. Biological Psychiatry, 46, 15951602.
Stein-Behrens, B. A., Elliott, E. M., Miller, C. A., Schilling, J. W., Newcombe, R., &
Sapolsky, R.M. (1992). Glucocorticoids exacerbate kainic acid-induced extracellular accumulation of excitatory amino acids in the rat hippocampus. Journal of Neurochemistry, 58, 17301735.
Stein-Behrens, B. A., Lin, W. J., & Sapolsky, R. M. (1994). Physiological elevations
of glucocorticoids potentiate glutamate accumulation in the hippocampus.
Journal of Neurochemistry, 63, 596602.
Stein, M. B., Koverola, C., Hanna, C., Torchia, M. G., & McClarty, B. (1997). Hippocampal volume in women victimized by childhood sexual abuse. Psychological
Medicine, 27, 951959.
Stoddard, S. L., Bergdall, V. K., Townsend, D. W., & Levin, B. E. (1986). Plasma catecholamines associated with hypothalamically-elicited ight behavior. Physiology & Behavior, 37, 709715.
Tanapat, P., Hastings, N. B., Rydel, T. A., Galea, L. A., & Gould, E. (2001). Exposure
to fox odor inhibits cell proliferation in the hippocampus of adult rats via an
adrenal hormone-dependent mechanism. Journal of Comparative Neurology,
437, 496504.

126

Terr, L. (1991). Childhood traumas: An outline and overview. American Journal of

Psychiatry, 148, 1020.
Terr, L. (1993). Unchained memories. New York: Basic Books.
Torrie, A. (1944). Psychosomatic casualties in the Middle East. Lancet, 29, 139
143.
Uddo, M., Vasterling, J. T., Brailey, K., & Sutker, P. B. (1993). Memory and attention in posttraumatic stress disorder. Journal of Psychopathology and Behavioral
Assessment, 15, 4352.
Uno, H., Ross, T., Else, J., Suleman, M., & Sapolsky, R. (1989). Hippocampal damage associated with prolonged and fatal stress in primates. Journal of Neuroscience, 9, 17051711.
Vaher, P., Luine, V., Gould, E., & McEwen, B. (1994). Eects of adrenalectomy on
spatial memory performance and dentate gyrus morphology. Brain Research,
656, 7176.
van Cauter, E., & Turek, F. W. (2001). Roles of sleep-wake and dark-light cycles in
the control of endocrine, metabolic, cardiovascular, and cognitive function. In
B. S. McEwen & H. M. Goodman (Eds.), Handbook of physiology: Section 7, The
endocrine system. New York: Oxford University Press.
van der Kolk, B. A. (1991). The biological response to psychic trauma: Mechanisms
and treatment of intrusion and numbing. Anxiety Research, 4, 199212.
van der Kolk, B. A. (1994). The body keeps the score: Memory and the evolving
psychobiology of posttraumatic stress. Harvard Reviews Psychiatry, 1, 253265.
van der Kolk, B. A., & Fisler, R. (1995). Dissociation and the fragmentary nature of
traumatic memories: Review and experimental conrmation. Journal of Traumatic Stress, 8, 505525.
van der Kolk, B. A., Hopper, J. W., & Osterman, J. E. (2001). Exploring the nature of
traumatic memory: Combining clinical knowledge with laboratory methods.
Journal of Aggression, Maltreatment & Trauma, 4, 931.
Vasterling, J. J., Brailey, K., Allain, A. N., Duke, L. M., Constans, J. I., & Sutker, P. B.
(2002). Attention, learning, and memory performances and intellectual resources in Vietnam veterans: PTSD and no disorder comparisons. Neuropsychology, 16, 514.
Vasterling, J .J., Brailey, K., Constans, J. I., Borges, A., & Sutker, P. B. (1997). Assessment of intellectual resources in Gulf War veterans: Relationship to PTSD.
Assessment, 4, 5159.
Vasterling, J. J., Brailey, K., Constans, J. I., & Sutker, P. B. (1998). Attention and
memory dysfunction in posttraumatic stress disorder. Neuropsychology, 12,
125133.
Virgin, C. E., Taryn, P. T. H., Packan, D. R., Tombaugh, G. C., Yang, S. H., Horner,
H. C., et al. (1991). Glucocorticoids inhibit glucose transport and glutamate
uptake in hippocampal astrocytes: Implications for glucocorticoid neurotoxicity. Journal of Neurochemistry, 57, 14221428.
Vyas, A., Mitra, R., Rao, S. B. S., & Chatterji, S. (2002). Chronic stress induces contrasting patterns of dendritic remodeling in hippocampal and amygdaloid neurons. Journal of Neuroscience, 22, 68106818.

127

Wakeeld, J. C. (1992). Disorder as harmful dysfunction: A conceptual critique of

DSM-III-Rs denition of mental disorder. Psychological Review, 99, 232247.
Watanabe, Y., Gould, E., Daniels, D. C., Cameron, H., & McEwen, B. S. (1992).
Tianetine attenuates stress-induced morphological changes in the hippocampus. European Journal of Pharmacology, 222, 157162.
Watanabe, Y., Gould, E., & McEwen, B. S. (1992). Stress induces atrophy of apical
dendrites of hippocampal CA3 pyramidal neurons. Brain Research, 588, 341
345.
Wechsler, D. (1945). A standardized memory scale for clinical use. Journal of Psychology, 19, 8795.
Wolkowitz, O. M., Reus, V. I., & Weingartner, H. (1990). Cognitive eects of corticosteroids. American Journal of Psychiatry, 147, 12971303.
Wolkowitz, O. M., Weingartner, H., Rubinow, D. R., Jimerson, D., Kling, M.,
Berretini, W., et al. (1993). Steroid modulation of human memory: Biochemical correlates. Biological Psychiatry, 33, 744746.
Wooley, C. S., Gould, E., & McEwen, B. S. (1990). Exposure to excess glucocorticoids
alters dendritic morphology of adult hippocampal pyramidal neurons. Brain
Research, 531, 225231.
Yehuda, R. (1997). Stress and glucocorticoid. Science, 275, 16621663.
Yehuda, R., Keefe, R. S. E., Harvey, P. D., Levengood, R. A., Gerber, D. K., Geni, J.,
et al. (1995). Learning and memory in combat veterans with posttraumatic
stress disorder. American Journal of Psychiatry, 152, 137139.
Yehuda, R., Southwick, S. M., Nussbaum, E. L., Giller, E. L., & Mason, J. W. (1991).
Low urinary cortisol in PTSD. Journal of Nervous and Mental Disorders, 178,
366369.
Zoellner, L. A., Foa, E. B., Brigidi, B. D., & Przeworski, A. (2000). Are trauma victims susceptible to false memories? Journal of Abnormal Psychology, 109, 517
524.

128

. , . ,

ost experts agree that people exposed to traumatic events remember them all too well. (For reviews, see McNally, 2003c; Pope,
Oliva, & Hudson, 1999.) Indeed, survivors of trauma often report intrusive
memories despite their eorts to forget. But debate persists regarding whether
a substantial subset of survivors repress, dissociate, or otherwise forget having been traumatizedand then remember it all later (McNally, 2003a).
Consider the views of Brown, Schein, and Hammond (1998), who proclaim
overwhelming scientic support for the existence of repressed or dissociated
memory (pp. 538539). Summarizing across studies, they conclude that
generally speaking, approximately a third of sexually abused victims report
some period of their lives where they did not remember anything about the
abuse and later recovered the memory of the abuse (p. 196). Yet after considering much of the same evidence, Loftus and Ketcham (1994) published their
dierent conclusions in The Myth of Repressed Memory. How can scholars scrutinize the same evidence yet arrive at such diametrically opposed conclusions?
To address this question, we examine the evidence adduced in support of repressed and recovered memory for traumatic experiences. To provide a fresh
perspective on the debate, we also summarize our recent laboratory research on
cognitive functioning in people who report repressed, recovered, or continuous
memories of trauma.

Common Assumptions About

Forgetting Trauma

Some theorists distinguish between repression and dissociation, dening the

former as inhibition of taboo urges and the latter as inhibition of disturbing
129

memories. But in the recovered memory debate, this is a distinction without a

dierence. In fact, as Brown et al. indicate, repressed memory and dissociated
memory are used interchangeably. Regardless of whether the theorists favored
mechanism is repression or dissociation, he or she usually endorses the following assumptions about forgetting and trauma. First, people are motivated to forget unpleasant experiences. Second, they are especially likely to forget atrocious
trauma: The ordinary response to atrocities is to banish them from consciousness (Herman, 1992, p. 1). Third, because sexual abuse, in particular, is so traumatic, if someone fails to think about the abuse for many years, then he or she
must have repressed or dissociated it from awareness. That is, an active inhibitory force must be keeping it out of consciousness. Otherwise, why would someone not think about it for years? Fourth, the more frequently a person (especially
a child) is traumatized, the more dicult it will be for him or her to remember
having been traumatized (Terr, 1991). That is, people exposed to repetitive
trauma develop dissociative skills to cope with inescapable situations, making
it dicult to recall these experiences many years later. Fifth, children are more
likely to forget having been traumatized if they were abused by their parents than
if they were abused by strangers (Freyd, 1996). Sixth, forgotten trauma is neither inert nor benign; it is the silent source of diverse psychological problems
(Blume, 1990). Accordingly, remembering forgotten trauma, emotionally processing it, and integrating it into ones autobiography are important steps toward healing (Courtois, 1992). All assumptions except the rst have been
ashpoints for intense controversy.

Evidence Adduced for

Forgetting of Trauma

The trauma therapists Brown et al. (1998) have assembled evidence in support of the six assumptions. Many of the studies they cite were modeled on an
inuential survey conducted by Briere and Conte (1993). Recruiting subjects
through a network of therapists specializing in the treatment of sexual abuse,
these authors obtained questionnaire data on 450 patients who reported having been sexually abused as children. In response to the question, Was there
ever a time when you could not remember the forced sexual experience, 59%
of the subjects responded in the armative. Many psychologists, such as Brown
et al., interpreted these ndings as evidence that many sexual abuse survivors
experience amnesia for their traumatic experiences, only to remember them
later in life.
Scholars soon drew attention to the methodological limitations of this survey and others like it. First, the subjects were patients potentially exposed to
therapeutic techniques likely to foster illusory memories of abuse (Poole, Lindsay, Memon, & Bull, 1995). As in many abuse recollections, external corrobo-

130

ration was apparently unavailable, and the duration of amnesia for trauma
was unspecied (a week? decades?).
The most important issue concerned the key survey question about forgetting. Interpreted literally, this question seems to make little sense. An armative response implies that the subject had repeatedly attempted to remember
the abuse but failed to do so (i.e., could not remember). But if subjects were
unaware of having been abused, on what basis would they attempt to recall it
in the rst place? Indeed, the most sensible way to regard armative responses
to this question is to assume that subjects interpreted it to mean, Was there
ever a time when you did not think about your abuse? A person who answered
yes to this question might have experienced sexual abuse as a child, managed not to think about it for many years, but was reminded of the abuse in
adulthood. But the long time when the person did not think about his or her
abuse cannot be equated with amnesia, or an inability to recall something
from memory when furnished with adequate retrieval cues. Just because a
person did not remember (did not think about it) does not mean that the person could not remember.
Moreover, as Schooler and others have shown, some people who believe they
have not thought about their (documented) traumatic experiences for many
years are surprised to learn that they had, in fact, discussed these events with
family members when they thought the memories never came to mind (Schooler,
Bendiksen, & Ambadar, 1997). That is, a person can forget having remembered
(thought about) a traumatic event, producing an illusion of amnesia. A person might be especially likely to forget a previous instance of recalling abuse if
the recollection was not accompanied by strong emotion. For example, in a conversation with her husband, a woman might mention that she was once molested by a distant relative. If this conversation occurs during an especially
positive period in her life, she might not experience much emotion while recounting her abuse and therefore not recall the conversation with her husband.
A key point in examining the literature on forgetting trauma is that one cannot equate amnesia for an event with merely not having thought about the event
for a time. To conrm amnesia, one must show that the person had encoded the
event in the rst place and is now incapable of recalling it despite the adequate
retrieval cues (e.g., such as being asked whether one had ever been abused).
Failure to think about something for a time must not be confused with an inability to remember it.
Moreover, as we have learned in our ongoing studies, many individuals report that the abuse was not experienced as traumatic when it occurred. Many
of our subjects have told us that they felt it must be wrongthat something
was not right about it. At the time, they often felt confused or embarrassed by
the experience. Only years later did they grasp the full signicance of what had
happened (i.e., that they were sexually abused). In fact, many of them rate the
abuse as more traumatic now than it was at the time.

131

These points also apply to the growing number of case reports concerning people
who recall long-forgottenand often corroboratedepisodes of abuse (for reviews, see Cheit, 1998, 1999; McNally, 2003c, chapter 7). Some children are exposed to one (or sometimes more) episodes of molestation, often nonpenetrative
(e.g., fondling) but abusive nevertheless. They nd the experiences confusing or
upsetting but not necessarily terrifying. They avoid dwelling on the experience and
may succeed in forgetting it, especially if reminders are no longer present (e.g., the
perpetrator moves away). Then, many years later, they encounter retrieval cues
(e.g., someone mentions the perpetrator, or they hear news reports on television
regarding sexual abuse), and the memory suddenly pops back into awareness. Such
cases clearly qualify as recovered memories of sexual abuse, but they do not count
as amnesia (i.e., an inability to recall when provided adequate retrieval cues). In
fact, sudden recollections of seemingly forgotten experiences are fairly common
in the general population and are not conned to adverse events like sexual abuse
(Read, 1997). Most important, a failure to think about something for many years
does not mean that active forces (repression, dissociation) are preventing these
memories from entering awareness.
An important study by Williams (1994) further illustrates these interpretive
pitfalls. Her research team interviewed 129 women who had been assessed for
suspected sexual abuse approximately 17 years earlier, when they were children.
Medical evidence conrmed the abuse in many cases. The main purpose of the
study was to survey women about their health and experiences with the health
care system. Embedded within the interview were questions about sexual abuse.
Women who acknowledged having been abused were questioned about the
event. Their narratives were compared with the hospital records that described
the index event. Of the 129 women, 49 of them did not mention the index event
(although 33 of them did describe other incidents of sexual abuse). Strikingly,
16 of the 49 women denied ever having been sexually abused.
Some authors have interpreted these data as showing that victims can repress
or dissociate all memory of their abuse rather than merely forget it (i.e., not think
about it for a long time), but other interpretations are possible. Many of the subjects were younger than 5 when assessed for abuse. Because few memories from
very early childhood survive into adulthood, memories of the index event may
have undergone ordinary forgetting. Moreover, the younger the child was when
abused, the more likely she might have failed to understand what the perpetrator was doing, thereby undermining encoding and later retrieval of the memory.
Other subjects may have remembered the index event but may have been too
embarrassed to disclose such personal matters to the interviewer.
A study by Goodman et al. (2003) provides further data on failure to disclose
abuse. They assessed 175 subjects (81% female) who had been involved in legal
proceedings concerning sexual abuse. The proceedings occurred when the subjects were approximately 9 years old (range: 3 years to 16 years). Questions about
sexual abuse were embedded in a longer survey concerning attitudes about the

132

law and experiences with the legal system. The survey was administered 13 years
after the subjects had been involved in the legal proceedings. The results revealed
that 88% of the subjects mentioned the index event that led to the legal proceedings. Like some subjects in Williamss study, 17 subjects denied ever having been
abused. However, further analysis of the data undermines many popular explanations for why these nondisclosing individuals may have forgotten their
abuse. Contrary to the notion that people prone to dissociation are most likely
to forget their abuse, the higher a subjects score on the Dissociative Experiences
Scale (DES; Bernstein & Putnam, 1986), the more likely the subject was to disclose the abuse. Contrary to Freyds (1996) conjecture that children exposed to
parental abuse are those most likely to be unable to remember it, there was no
correlation between failure to disclose and relationship to the perpetrator.
The studies by Williams (1994) and Goodman et al. (2003) do not provide
clear support for any special repression or dissociation mechanism that actively
inhibits recollection of the index event. The ndings from these studies can be
most parsimoniously interpreted as showing that a minority of adults with childhood abuse histories will deny having been abused when questioned by a survey interviewer. However, in neither study did researchers conduct subsequent
clarication interviews to ascertain the basis for the discrepancy between documents concerning the abuse and the subjects denial of it. In a study on young
adults with documented histories of physical abuse, Femina, Yeager, and Lewis
(1990) found that all subjects who had denied (or minimized) their abuse during an interview later acknowledged that they had not forgotten it when originally questioned by the interviewer. Subjects who had earlier denied their
documented abuse later said that they did not wish to discuss such upsetting
experiences or did not like the interviewer. None had repressed or forgotten the
abuse. The ndings of this study suggest that some of the nondisclosers in the
studies of Williams and Goodman et al. may have been disinclined to discuss
abuse they had, in fact, remembered.
In any event, survey studies have not been the only source of evidence adduced in support of repressed memories of trauma (Brown et al., 1998; van der
Kolk & Fisler, 1995). For example, Brown et al. wrote that Dollinger (1985)
found that two of the 38 children studied after watching lightning strike and kill
a playmate had no memory of the event (pp. 609610). Although this trauma
was well documented, the amnesia experienced by these two children for this
horror has nothing to do with dissociation or repression. Brown et al. forgot to
mention that both children had themselves been struck by side ashes from
the main lightning bolt, had been knocked unconscious, and had nearly died
(Dollinger, 1985). Given the severe eects on the central nervous system of a
lightning strike (Kotagal, Rawlings, Chen, Burris, & Nouri, 1982), it is little
wonder that these two children had amnesia for the episode. Obviously, their
amnesia was entirely organic. Indeed, none of the other children who had not
been struck by lightning forgot this horric event.

133

Interpreting research on memory for trauma among former inmates of concentration Camp Erika (Wagenaar & Groeneweg, 1990), Brown et al. (1998)
claimed that amnesia for Nazi Holocaust camp experiences has also been reported (p. 156). Wagenaar and Groeneweg assessed the memories of camp
survivors 40 years after the survivors had provided testimony about their internment shortly after their release. Impressed by the strikingly accurate recollections of the former inmates, Wagenaar and Groeneweg emphasized that
there is no doubt that almost all witnesses remember Camp Erika in great detail, even after 40 years (p. 84). Why, then, did Brown et al. cite this study as
providing evidence for traumatic amnesia? As it turns out, several subjects
claimed to have forgotten the name of an especially brutal guard or failed to
mention several violent incidents that they had mentioned 40 years earlier.
But even these isolated incidents of forgetting do not indicate amnesia. With
one exception, subjects immediately recollected the forgotten details once they
examined their original testimony. Given the remarkable degree of remembering (Wagenaar & Groeneweg, 1990, p. 80) exhibited by the former inmates, Brown et al.s claim about amnesia regarding the Holocaust seems
far-fetched. Because memory does not operate like a video recorder, it is hardly
surprising that not every detail mentioned was instantly recalled 40 years later.
[See chapter 11Eds.]
One of the most common blunders made by theorists of traumatic amnesia
is to confuse general memory impairment in the wake of trauma with an inability to remember that one has been traumatized. Listing studies that supposedly support the notion of traumatic amnesia, Brown et al. (1998) mention
Wilkinsons (1983) study of individuals who had witnessed the horric collapse
of the Hyatt Regency skywalks in Kansas City. Of those assessed, 27% endorsed
the DSM-III symptom of memory diculties while arming their intensely
vivid recollections of the disaster itself. Indeed, intrusive memories of the disaster may have contributed to problems with forgetfulness in everyday life.
In another often misinterpreted study (Brown et al., 1998; van der Kolk &
Fisler, 1995), Archibald and Tuddenham (1965) reported that 65% of their
World War II combat fatigue cases complained of diculty in memory versus
only 5% of psychiatrically healthy combat veterans. However, 60% of veterans
who were not in combat and suered from other psychiatric disorders endorsed
the same complaint. Moreover, the authors themselves emphasized how unforgettable combat was for the traumatized patients, noting that these veterans
cannot blot out their painful memories (p. 480). Obviously, the memory problems reported by these men concern everyday forgetfulness, not amnesia for
their combat trauma.
However, other studies on combat veterans have uncovered evidence of apparent amnesia for combat itself among groups of men referred for psychiatric
evaluation and treatment. Eder (1917) reported that 2% of his psychiatric battle

134

casualties could not recall certain combat events occurring in World War I.
According to Sargant and Slater (1941), 14% of psychiatric casualties who had
escaped from Dunkirk reported amnesia for events occurring during the chaotic
retreat. World War II psychiatrists reported rates of amnesia for combat in 5%
of patients in the Pacic Theater (Henderson & Moore, 1944) and 8.6% in North
Africa (Torrie, 1944). But as Pope et al. (1999) pointed out, these psychiatrists
were often unable to rule out organic causes of amnesia (e.g., exhaustion, head
injury) or malingering.
Other more recent cases of amnesia for traumatic events provide an interesting contrast with the typical case of repressed and recovered memory of childhood sexual abuse. Swihart, Yuille, and Porter (1999) have described cases of
red-out in which a person murders a loved one in a t of rage but experiences
amnesia for the murder itself. Swihart et al. suspect that these cases may be genuine, not malingered. Indeed, the murderer often phones the police and immediately admits to committing the crime despite claiming not to have remembered
the actual attack.
There have been other cases of psychogenic amnesia in which an event,
often shocking, triggers retrograde amnesia (see McNally, 2003c, chapter 7, for
a review of this case study literature). Reviewing this literature, Arrigo and
Pezdek (1997) connected it to the furor over recovered memories of sexual abuse.
However, the phenomenon of psychogenic amnesia is drastically dierent from
the controversial cases in the child abuse eld. In cases of psychogenic amnesia,
say, after the death of a loved one, the person develops complete retrograde
amnesia shortly after the shocking event, accompanied by loss of personal identity. Psychogenic amnesia seldom lasts for more than a few hours to a few weeks.
Typically, memory and personal identity return abruptly without any psychotherapy. In contrast, cases of traumatic amnesia, say, of repressed and recovered
memories of childhood sexual abuse, have no clear onset, never involve identity loss, may last for years or decades, and often gradually return piece by piece
during the course of psychotherapy. Moreover, the amnesia selectively blocks
retrieval of trauma; it does not blot out the entire persons life and identity. Accordingly, classic psychogenic amnesia has little to do with the controversy over
repressed and recovered memories of childhood abuse.
Finally, some theorists note that inability to recall an important aspect of
the trauma (psychogenic amnesia) has been a diagnostic criterion of PTSD ever
since DSM-III-R (American Psychiatric Association [APA], 1987, p. 250) and
adduce this fact to support the notion that trauma survivors can repress or dissociate their memory for trauma. Unfortunately, this diagnostic criterion is fatally ambiguous because it does not distinguish between an inability to recall
an aspect of the trauma that was never encoded into memory in the rst place
and an inability to recall an aspect of the trauma because an inhibitory mechanism (e.g., repression) blocks its access to awareness. Because the mind is not a

135

video recorder, there will inevitably be aspects of the traumatic event that never
make it into memory. Encoding failure must not be confused with amnesia, let
alone repression.
In summary, extant data do not support the claim that unconscious repression or dissociation mechanisms expel traumatic memories from awareness and
prevent them from entering consciousness. People often try not to think about
their sexual abuse or other unpleasant experiences, and sometimes they succeed.
But not thinking about something for a long time does not imply an inability to
remember it (i.e., amnesia).

A View From the Cognitive

Psychology Laboratory

Cognitive psychologists have conducted considerable research on mechanisms

of inhibition in memory retrieval (e.g., Anderson & Neely, 1996). And they have
drawn on basic research to speculate about how false memories of trauma might
be inadvertently fostered in clinical settings (e.g., Ceci & Loftus, 1994; Lindsay
& Read, 1994). But this work is swiftly dismissed by repressed memory theorists,
who assert that generalizations from laboratory studies with college students
have questionable relevance for how abuse survivors might remember or forget
their trauma (e.g., Brown et al., 1998, p. 98). In fact, one of the most striking
features of the recovered memory debate has been the near absence of any research on cognitive functioning in people reporting repressed or recovered
memories of abuse. This situation may have arisen because few clinicians possess expertise in laboratory research, and few cognitive psychologists have access to trauma populations. In any event, during the past few years our research
group has been conducting laboratory studies designed to test hypotheses relevant to (1) mechanisms implicated in either the ability to repress and recover
traumatic memories or (2) mechanisms relevant to proneness to forming false
memories of trauma (McNally, 2001, 2003b).
This line of research began quite by accident. The rst author was conducting
psychiatric diagnostic interviews for a positron emission tomography (PET) study
designed to map the functional neuroanatomy of autobiographical recollection of
traumatic memories in women who had been sexually abused as children (Shin
et al., 1999). We had posted advertisements seeking female adult survivors of childhood sexual abuse, and we planned to distinguish those with and without PTSD
prior to the study. During the diagnostic interview, three women who expressed
interest in the project turned out not to have any explicit autobiographical memories of molestation, despite considering themselves survivors of abuse. Further
questioning revealed that they assumed that they harbored repressed memories
of abuse because they were inexplicably tense in the presence of certain relatives,
experienced diverse symptoms (e.g., depressed mood), and so forth; lacking a suit-

136

able explanation, they assumed that their problems must be attributable to repressed memories of early abuse. Because they lacked autobiographical memories
of trauma, these women did not qualify for the neuroimaging study. But their reports of repressed memories inspired us to recruit and study individuals who reported diverse memory manifestations of trauma.
In our rst series of studies, we recruited four groups of women from the community. The repressed memory group included subjects who suspected they had
been sexually abused as children but had no autobiographical memories of abuse.
These subjects inferred their abuse history from a diverse range of symptoms (e.g.,
nightmares, depressed mood, bulimia). About a third of this group mentioned
having acquired the belief in their repressed memories of abuse during psychotherapy. The recovered memory group included subjects who reported recollecting abuse after long periods of not having thought about it. Nearly half of them
remembered their abuse during the course of psychotherapy, but only one mentioned recovering a memory in session. The continuous memory group reported
never having forgotten their abuse. The control group included individuals who
denied ever having been sexually abused. We were unable to corroborate reports
of the abuse (although we are actively attempting to do so in our current set of
studies). Accordingly, we do not know whether the recovered (or continuous)
memories of abuse are genuine or whether those who suspected they harbored
repressed memories of abuse were really abused. Of course, the absence of corroboration does not mean that the memories are false.
To characterize our groups, we conducted a study on personality proles and
clinical symptoms (McNally, Clancy, Schacter, & Pitman, 2000b). Our subjects
completed the Multidimensional Personality Questionnaire (MPQ; Tellegen,
1982), an inventory designed to characterize normal personality variation that
includes an Absorption scale (Tellegen & Atkinson, 1974) that taps fantasy
proneness. In addition to these personality measures, they also completed three
psychiatric questionnaires. The Beck Depression Inventory (BDI; Beck & Steer,
1987) measured symptoms of depression, the Dissociative Experiences Scale
(DES; Bernstein & Putnam, 1986) measured alterations in consciousness (e.g.,
memory lapses, depersonalization, episodes of spacing out), and the civilian
version (Vreven, Gudanowski, King, & King, 1995) of the Mississippi Scale for
Combat-Related Posttraumatic Stress Disorder (CMISS; Keane, Caddell, & Taylor, 1988) measured PTSD symptoms, as well as other problems (e.g., occupational diculties, suicidal ideation) sometimes linked to PTSD.
The MPQ uncovered striking similarities and dierences in personality prole among the groups. The continuous memory group and the control group
were indistinguishable on every measure of personality. The same was true for
the repressed and recovered memory groups, who did not dier on any scale.
Moreover, the four groups did not dier on positive aectivity, or proneness to
experience joy and enthusiasm. In contrast, the groups did dier in negative
aectivity, or proneness to experience sadness, anxiety, anger, and guilt. Re-

137

pressed memory subjects scored higher than did either the continuous memory
or control groups, whereas recovered memory subjects scored midway between
the repressed subjects and subjects in the other two groups.
Consistent with patterns of negative aectivity, repressed memory subjects
had more depressive, dissociative, and PTSD symptoms than did continuous
memory and control subjects. Repressed memory subjects also had more depressive and PTSD symptoms than recovered memory subjects, who, in turn, had
more dissociative and PTSD symptoms than did control subjects. Strikingly,
continuous memory subjects were indistinguishable from control subjects in
PTSD, dissociative, and depressive symptoms.
Finally, the repressed and recovered memory groups did not dier on the absorption scale, but both groups scored higher than the control group. The repressed memory group also scored signicantly higher than the continuous
memory group. Table 4.1 summarizes these ndings.
These data indicate that people who believe they harbor repressed memories
of sexual abuse are more psychologically distressed than those who have never
forgotten their abuse. There are at least two explanations for this pattern. Heightened levels of distress among repressed memory subjects may reect the psychic
toll of blocking out memories of abuse. On the other hand, their distress may have
arisen from diverse and poorly understood causes, motivating an eort after
meaning that led them to attribute their problems to repressed memories of abuse.
As Bass and Davis (1988) emphasized in The Courage to Heal, When you rst remember your abuse or acknowledge its eects, you may feel tremendous relief.
Finally there is a reason for your problems. There is someone, and something to
blame (p. 173). When confronted with a choice between being miserable and not
knowing why and being miserable and knowing why, many people will choose
the latter, especially if they can also blame others for their diculties.
To test Bass and Daviss (1988) hypothesis, we recontacted 11 of our recovered memory subjects and asked them to complete a Child Abuse Survivor Questionnaire (Clancy, 2000). In this pilot study, subjects wrote brief responses to

table 4.1 Summary of Psychometric Results

Variable
Positive affectivity
Negative affectivity
Depression symptoms
PTSD symptoms
Dissociative symptoms
Absorption

Repress
a
a
a
a
a
a

Group
Recover
Continuous
a
ab
b
b
ab
ab

a
b
b
bc
bc
bc

Control
a
b
b
c
c
c

Note. Repress = repressed memory group; recover = recovered memory group.

Means of groups sharing a letter do not differ significantly (p > .05). A mean value
represented by a is greater than a mean value represented by b, which, in turn, is
greater than a mean value represented by c (i.e., a > b > c).

138

each question and then rated their strength of endorsement of each item on a
9-point Likert scale. The questions and anchors for each scale follow.
1. Since you recovered the memory, do you feel your self-esteem has changed?
The anchors were 1 (I feel much worse about myself ) and 8 (I feel much
better about myself).
2. Since you recovered the memory, do you nd that other people are more
or less supportive of you? The anchors were 1 (I have many less social
supports since I recovered the memory) and 8 (I have many more social
supports since I recovered the memory).
3. Since you recovered the memory, has your level of happiness changed? The
anchors were 1 (I feel much less happy) and 8 (I feel much happier).
4. Overall, since you recovered the memory, have things been better or worse
for you? The anchors were 1 (Things have been much worse) and
8 (Things have been much better).
5. If you had to do it all over again, would you choose NOT to remember that
you were abused? The anchors were 1 (I would denitely choose not to
remember) and 8 (I would denitely choose to remember).
6. Do you feel that recovering the memory changed how you understand
yourself? The anchors were 1 (I understand myself much less now) and
8 (I understand myself much better now).

The results revealed that 100% of the subjects reported at least some benets
from recovering their memories of abuse, and 73% reported benets in response
to all questions. After recovering their memories, 100% of the subjects reported
increased self-esteem (M = 6.7), 82% reported increased social support (M = 5.4),
91% reported increased happiness (M = 5.8), 82% reported overall life improvement (M = 6.2), 91% said they would choose to remember their abuse if they
could do it all over again (M = 7.2), and 100% reported increased self-understanding (M = 7.5).
These pilot data suggest that Bass and Davis (1988) may be correct. Identifying oneself as an abuse survivor by recovering memories of abuse may yield
psychological benets to already distressed individuals even if the eects of the
abuse itself have been clearly harmful.

Forgetting Trauma-Related Material

Some trauma theorists assert that sexually abused children develop an avoidant
(or dissociative) encoding style that enables them to disengage attention from terrifying stimuli during abuse episodes and redirect it elsewhere (e.g., Terr, 1991).
Unable to escape physically from the perpetrator, abused children may endeavor
to escape mentally. Although adaptive in the short run, a dissociative style of coping with threats presumably fosters long-term psychological problems.

139

This hypothesized cognitive style ought to be evident in the laboratory and

detectable through directed forgetting methods. There are two main directedforgetting paradigms, and each engages dierent mechanisms (Golding, in press;
Johnson, 1994). In the list method, the experimenter presents the subject with
a list of words, halfway through the procedure, he or she tells the subject that
the rst set of words was only for practice and hence can be forgotten. After
this instruction to forget, the experimenter resumes presentation of the words
to be remembered. In the item method, an instruction to remember or to forget
follows the presentation of each word. For both paradigms, the experimenter asks
the subject to recall as many words as possible, regardless of original instructions.
Generally, subjects tend to recall more remember-words than forget-words, but
the mechanisms underlying this eect dier for each paradigm. In the list paradigm, forget-words appear to have been encoded but areinhibited at retrieval.
In the item paradigm, subjects appear to attenuate their encoding of a word as
soon as the forget instruction occurs; hence, later recall failures seem chiey
attributable to poor encoding of the forget-words.
In our rst experiment, we tested adult female survivors of childhood sexual
abuse with PTSD, psychiatrically healthy survivors of childhood sexual abuse,
and nonabused control subjects (McNally, Metzger, Lasko, Clancy, & Pitman,
1998). Subjects had experienced penetrative abuse, and most had always remembered it. The subject viewed a series of words on a computer screen that were
either trauma-related (e.g., incest), positive (e.g., cheerful), or neutral (e.g., cupboard). After each words presentation, instructions appeared that told the subject either to remember or to forget the immediately preceding word. Immediately
after this encoding phase, we asked subjects to write down all the words they
could remember, regardless of whether a word had been followed by remember
or forget instructions. The standard directed forgetting eect emerging from basic
research indicates that subjects abort encoding of forget-words and, accordingly,
exhibit better recall for remember-words than for forget-words (Johnson, 1994).
However, if psychologically traumatized survivors are characterized by a superior ability to disengage attention from trauma cues, then the PTSD group, compared to healthy survivors and controls, should exhibit superior forgetting of
trauma words relative to other words.
The results contradicted this hypothesis: childhood sexual abuse survivors
with PTSD exhibited memory decits for positive and neutral words they had
been instructed to remember, and they exhibited excellent memory for trauma
words, including words they had been instructed to forget. Exhibiting the standard directed forgetting eect, healthy survivors and control subjects recalled
remember-words more often than forget-words, regardless of the words valence.
Contrary to our original hypothesis, survivors with PTSD seemed to exhibit an
impairment in the ability to banish trauma words from awareness.
One limitation of the previous study was that our sexual abuse survivors were
suering from intrusive thoughts as part of their current PTSD. Perhaps indi-

140

viduals reporting repressed or recovered memories of abuse would be most likely

to exhibit superior forgetting of trauma words in the laboratory. To test this
hypothesis, we used the same directed-forgetting protocol to compare the relative ability to forget trauma words in recovered memory subjects, repressed
memory subjects, and nonabused control subjects (McNally, Clancy, & Schacter,
2001). If anyone should exhibit heightened ability to forget trauma words, it
should be individuals who once repressed (or still have repressed) their memories of abuse.
Contrary to this hypothesis, the memory performances of the repressed and
recovered memory groups were entirely normal. Both groups recalled remember-words better than forget-words, regardless of valence. (Control subjects,
however, exhibited diculty forgetting trauma words.)
The directed-forgetting paradigm involves the forgetting of trauma words,
mere pale proxies for autobiographical memories of abuse. If these individuals
developed a skill for repressing or dissociating genuine memories of trauma, it
should easily be brought into play in the laboratory when they are asked to forget words linked to abuse. Yet we found no evidence of any superior ability to
forget trauma words among our childhood sexual abuse groups, and those with
PTSD appear to experience a breakdown in the ability to banish trauma words
from awareness. The very fact that abuse survivors with PTSD have diculty
forgetting is fully consistent with the cardinal symptom of the disorder: intrusive recollection of traumatic memories.
On the other hand, preoccupation with the possibility that one might have
been sexually abused might heighten the salience of trauma words, making it
dicult for subjects to disengage attention from them and forget them in this
paradigm.

Intrusion of Traumatic Material

PTSD is characterized by intrusive recollection of traumatic memories. Assessment of this symptom has traditionally relied on introspective self-reports disclosed in interviews or questionnaires. The emotional Stroop color-naming
paradigm provides a quantitative measure of intrusive cognition that does not
rely on introspection (for reviews, see McNally, 1998; Williams, Mathews, &
MacLeod, 1996). In this laboratory paradigm, subjects view words that vary in
emotional signicance and are asked to name the colors of the words while ignoring their meanings. When the meaning of a word captures the subjects attention, the subject exhibits a delay in naming its color. The intrusiveness of
meaning is reected in delayed color naming (Stroop interference). Psychiatrically disturbed trauma survivors who qualify for PTSD take longer to name the
colors of words related to their trauma than to name the colors of neutral words
or emotional words unrelated to their trauma (for a review, see McNally, 1998).

141

Trauma survivors who do not suer from PTSD seldom show much Stroop interference. Psychologically traumatized survivors of combat (e.g., McNally,
Kaspi, Riemann, & Zeitlin, 1990), rape (e.g., Foa, Feske, Murdock, Kozak, &
McCarthy, 1991), and sexual abuse (Dubner & Motta, 1999) exhibit Stroop interference for words related to their traumatic memories. Indeed, color naming
of trauma words is more strongly associated with intrusive symptoms than with
avoidance/numbing symptoms (Cassiday, McNally, & Zeitlin, 1992).
We used the emotional Stroop paradigm to test whether subjects reporting
either continuous, repressed, or recovered memories of sexual abuse, compared
to nonabused control subjects, would exhibit interference for trauma words
(McNally, Clancy, Schacter, & Pitman, 2000a). Although we did not conduct
diagnostic interviews to assess for PTSD, the repressed memory group exhibited
elevations on the CMISS. Moreover, if severity of trauma is what causes PTSD
and what motivates repression of traumatic memories, subjects who cannot
recall their presumably repressed memories may nevertheless exhibit interference on the emotional Stroop taska measure of automatic (obligatory) emotional processing (McNally, 1995).
Subjects named the colors of a series of trauma-related (e.g., molested), positive (e.g., elation), and neutral (e.g., carpet) words on a computer screen as
quickly as possible. Unlike patients with PTSD, none of the groups exhibited
delayed color naming of trauma words relative to neutral or positive ones. These
data suggest that interference eects for trauma-related material may be conned to abuse survivors who qualify for PTSD (Dubner & Motta, 1999). Believing that one harbors repressed memories of abuse is not associated with patterns
of interference that characterize survivors with PTSD.

Memory Distortion
Many psychologists have warned that certain therapeutic techniques may inadvertently foster false memories of abuse in distressed patients (e.g., Poole
et al., 1995). One such technique is guided imagery. Misguided therapists who
believe that patients with certain symptom patterns harbor repressed memories
of abuse will ask these patients to visualize abuse scenarios that might have
happened, hoping that the content of guided imagery might trigger recollection
of the blocked trauma. Unfortunately, repeated visualization of imagined events
may increase condence that the events actually occurred (Garry, Manning,
Loftus, & Sherman, 1996; Heaps & Nash, 1999).
Using an imagination ination paradigm developed by Garry et al. (1996), we
tested whether recovered memory subjects are more susceptible than control subjects to this form of memory distortion (Clancy, McNally, & Schacter, 1999). Subjects rated their condence regarding whether they had experienced unusual, but

142

nontraumatic, childhood events (e.g., getting stuck in a tree, nding a $10 bill in
parking lot). During a subsequent visit to our laboratory, they performed a brief
guided imagery task requiring them to visualize certain events but not others.
Immediately thereafter, they rated their condence in whether any of these events
had actually occurred during their childhoods. Perhaps because of modest statistical power, the increase in condence that imagined events had in fact happened
fell short of signicance. However, the eect size for this imagination ination
phenomenon was more than twice as large in the control group than in the recovered memory group. Several subjects in the recovered memory group later said
that they thought the experiment was about creating false memories in the laboratory, perhaps indicating that the procedure was transparent.
Continuing to explore whether recovered memory subjects are especially
vulnerable to false memory eects in the laboratory, we next applied a variant
of the Deese/Roediger-McDermott (DRM) paradigm (Deese, 1959; Roediger &
McDermott, 1995). This paradigm is less transparent than the guided imagery
one. During the encoding phase, subjects hear a series of word lists, each consisting of semantically related items (e.g., sour, bitter, candy, sugar) that converge on a nonpresented wordthe false targetthat captures the theme of
the list (e.g., sweet). The false memory eect occurs when subjects remember
having heard the false target on subsequent recognition tests. Research suggests
that false memory eects in this paradigm occur when people rely on their
memory for the general semantic aspects (or gist) of the items they studied.
We tested subjects who reported either repressed, recovered, or continuous
memories of childhood sexual abuse and nonabused control subjects (Clancy,
Schacter, McNally, & Pitman, 2000). None of the lists was trauma-related. The
results revealed that the recovered memory group was more prone to remember false targets than were the other groups. This group exhibited no general
memory impairment; material that had been presented was remembered as well
by the recovered memory subjects as by the other subjects. Moreover, the higher
a subjects DES score, the more likely she was to exhibit the false memory eect.
(The DES also predicts this eect among college students in the DRM paradigm;
Winograd, Peluso, & Glover, 1998.) Thus, self-reported dissociation in everyday
life was linked to remembering words never presented. This tendency for false
memory formation notwithstanding, we do not know whether the eect would
be more or less pronounced for material directly related to abuse.
What implications do these ndings have for false memories of trauma? The
fact that recovered memory subjects apparently relied on gist rather than
memory for specic words suggests a processing style that might contribute to
memory confusion later in life. Arming that something occurred based on
general resemblance to events that did occur might lead one to suspect they were
sexually abused if they did, in fact, experience other similar adverse events in
childhood (e.g., emotional abuse, neglect).

143

False Memory Eects in

Space Alien Abductees

The DRM experiment suggested that individuals who report recovered memories of sexual abuse are prone to exhibit false memory eects in the laboratory.
But this does not mean that their recovered memories of abuse are false. We were
unable to obtain independent evidence bearing on the veracity of their abuse
reports, but external corroboration of abuse is often dicult to obtain.
In our next experiment, we extended our work to a group of subjects whose
memories of trauma were likely inaccurate: people who report recovering memories of abduction by space aliens. Applying Robinson and Roedigers (1997) DRM
paradigm, we tested the false recognition and false recall propensities in three
groups of subjects (Clancy, McNally, Schacter, Lenzenweger, & Pitman, 2002).
One group reported recollections of alien abduction. A second group, similar to
our repressed memory subjects, included individuals who believed they had been
abducted but had no explicit memories of the trauma. They inferred a history of
abduction from sources such as unexplained marks on their bodies, interpreted
as evidence of alien medical probes; a passion for reading science ction books;
panic attacks triggered by seeing drawings of aliens on book covers; and so forth.
They assumed that the aliens had control of their memories or that the abduction occurred in another dimension. A third (control) group included individuals who denied a history of alien abduction.
Like subjects who report having recovered memories of sexual abuse, those
reporting abduction by space aliens exhibited greater false memory eects in the
DRM paradigm than did either the control group or the group that had suspected
(but had no memory) they had been abducted. The group that believed they had
been abducted, but who had no conscious memories of abduction, exhibited
greater false memory eects than the control group but not as great eects as
the group that had recovered memories of abduction. Therefore, subjects who
had recovered memories of abduction relied most on gist memory in the DRM
paradigm, a propensity that may explain why they recall these experiences
after having undergone quasi-hypnotic regression therapies, read books about
abduction, and so forth. Accordingly, this pattern may explain why the recovered memory group went on to develop full-blown false memories of abduction,
whereas others did not.
People who have recovered memories of sexual abuse often experience intense
emotion while remembering these long-forgotten events, tempting some therapists to credit the veracity of the accounts. Surely the memories must be genuine, some therapists believe; otherwise, how could these recollections trigger such
extreme aect?
In fact, considerable research indicates that when people with PTSD recollect their traumatic experiencesmore specically, when they listen to

144

audiotaped descriptions of themthey exhibit heightened psychophysiologic

reactivity (for a review, see Orr & Roth, 2000). Indeed, Orr et al. (1998) found
not only that female survivors of sexual abuse with PTSD exhibited greater physiological responding while listening to autobiographical trauma scripts than did
abuse survivors without PTSD but also that those with PTSD who had recently
recovered their memories tended to be more physiologically responsive than were
those who had always remembered their abuse.
Using this script-driven imagery procedure, we tested whether people reporting abduction by space aliens exhibit heightened psychophysiologic responding
while listening to scripts describing their purported abduction experiences
(McNally et al., 2002). If they did, then this nding would imply that belief that
one has been traumatized can generate a physiologic prole similar to that of
PTSD patients who have endured traumatic conditioning experiences.
Each of our 10 abductees traced contact with space aliens to an episode of
apparent sleep paralysis accompanied by hyponopompic (upon-awakening)
hallucinations. These episodes occur as the person is awakening from rapid eye
movement (REM) sleep (Cheyne, Rueer, & Newby-Clark, 1999). Dreams occur
during REM sleep, and this stage of sleep is associated with fullbody paralysis.
However, during an episode of sleep paralysis, the cognitive and motoric aspects
of REM become desynchronized for several seconds to several minutes. The
sleeper awakens before paralysis has waned and becomes aware of an inability
to move. No more pathological than a hiccup, sleep paralysis occurs at least once
to approximately 30% of the general population (Cheyne, Newby-Clark, &
Rueer, 1999). About 5% of the population, however, has also experienced hallucinations during these episodes, including ashing lights, buzzing sounds, feelings of levitating o the bed, electrical tingling sensations shooting through ones
body, and seeing threatening gures hovering near ones bed. Within a few seconds or minutes, the person is entirely awake, can move again, and the hallucinations cease. None of our subjects realized that he or she was experiencing sleep
paralysis, and all eventually interpreted the episode as an encounter with alien
beings. Frightened by such experiences, some sought hypnosis to recover presumably repressed aspects of these episodes. During these sessions, they recalled
additional memories of having been transported into spacecraft, where they
endured medical and sexual experimentation at the hands of alien beings.
For the script-driven imagery protocol, each abductee furnished material for
ve individualized, autobiographical scripts: two scripts related to abduction
trauma; a script related to a dierent, highly stressful event; a script related to a
positive experience; and one related to an emotionally neutral experience. Our
control group consisted of individuals who denied a history of alien abduction.
Each control subject heard the scripts of one of the abductees.
Compared with control subjects, abductees exhibited greater heart rate and
skin conductance responses during imagery of abduction scripts than to imagery of the other scripts (ps < .06; eect size rs = .40). Although none of the

145

abductees ever met full criteria for PTSD, the magnitude of their heart rate and
skin conductance responses to their abduction scripts matched or exceeded the
reactivity exhibited by PTSD subjects to their trauma scripts in previous research.
For example, abductees had a mean skin conductance response of 1.82 S,
whereas Vietnam combat veterans with PTSD in Keane et al.s (1998) denitive
study had a mean response of 0.8 S. Likewise, the mean heart rate response of
abductees was 7.8 bpm, whereas for Vietnam veterans with PTSD it was 3.2 bpm
(Keane et al., 1998). Believing that one has been traumatized by space aliens
produces heightened physiologic responding similar to that of people exposed to
combat and other traumatic events. Therefore, emotional responding during
recollection provides no guarantee that the memory is veridical.

Are Memories of Trauma Special?

Some clinical theorists believe that memories of traumatic events dier qualitatively from memories of other emotionally intense experiences. Consider the
claim of Brown et al. (1998):
Thus, across all studies, a robust nding is that narrative memory for the gist
or central action of a negative emotionally arousing event is extremely well
retained. The more involved the person is, the more likely the memory will
be retained. Traumatic events are a signicant exception to this general rule.
A signicant portion of traumatized individuals suer from traumatic amnesia; even if their narrative memory for the traumatic event is retained, it may
not be readily accessible. (p. 368)

Intensity of emotion, Brown et al. believe, is a key variable. [W]hen emotional

material reaches the point of being traumatic in intensitysomething that cannot be replicated in articial laboratoriesin a certain subpopulation of individuals, material that is too intense may not be able to be consciously processed and
so may become unconscious and amnesic (p. 97). Others have made similar
assertions (e.g., Joseph, 1999; van der Kolk, 1994), suggesting that memory for
traumatic experience follows an inverted-U function (Yerkes & Dodson, 1908).
That is, emotional arousal enhances memory, but only up to a point. After that
point is reached, these theorists believe that intense emotion somehow undermines hippocampal function and impairs explicit memory for the trauma.
This theory is riddled by many conceptual and empirical problems (see McNally,
2003c, chapter 6, for detailed critique). First, intense arousal does not block the
formation of explicit narrative memory for traumatic experiences. As Langers
(1991) work has shown, survivors of the Nazi concentration camps are haunted
by detailed, explicit, vivid memories of the horrors they experienced. Indeed, the
evidence overwhelmingly shows that traumatic experiences are remembered all
too well (Pope et al., 1999). People tend not to forget the central aspects of trau-

146

matic events, even if they fail to encode or remember details of these experiences.
[as with nontraumatic emotional memories; see chapter 1Eds.].
Second, abundant laboratory research shows that amygdala activation promotes encoding and hippocampus-mediated explicit memory for emotionally intense experiences (for a review, see McGaugh, Ferry, Vazdarjanova, & Roozendaal,
2000). To be sure, release of stress hormones can impair performance on certain explicit memory tasks mediated by the hippocampus (for a review, see Lupien
& McEwen, 1997). But stress does not impair memory for the stress-producing
experience itself. For example, Kirschbaum, Wolf, May, Wippich, & Hellhammer
(1996) exposed subjects to a social stressor (giving a speech prior to trying to
solve math problems out loud) before having them learn and recall a list of words.
The higher the levels of stress hormone (cortisol) triggered by this stressor, the
fewer words subjects recalled. Hence, stress can impair memory for an incidental activity, such as memorizing words. But it does not abolish memory for the
stressor itself.
Third, as Christiansons (1992) review shows, studies on emotional stress and
memory fail to support a Yerkes-Dodsonlike (1908) notion that moderately
high stress enhances memory and extreme stress impairs it. (Moreover, as some
trauma theorists seem to have forgotten, Yerkes and Dodson studied visual discrimination learning in mice, not memory for trauma. Hence, the work of these
two scientists has scant relevance for the recovered memory debate.) As
Easterbrooks (1959) theory implies, extreme stress enhances memory for the
central aspects of the traumatic experience, sometimes at the expense of the
peripheral details (McNally, 2003c, chapter 5).
Fourth, some theorists have suggested that extreme stress establishes implicit memories of the traumatic experience while sometimes undermining
explicit memories of it (e.g., van der Kolk, 1994). As Brown et al. (1998) asserted, Much of the memory for trauma is retained as an implicit rather than
explicit memory (p. 483). In contrast to memory for ordinary events, dissociated trauma memory supposedly remains frozen and relatively impervious to
distortion during the passage of time. These allegedly dissociated implicit
memories of trauma are nevertheless manifested in the form of body memories [conditioned emotional responses?], ashbacks, fragments, sudden intense
feelings, avoidant behaviors, images, sensory processes, and dreams (p. 187).
There are several problems with this formulation (McNally, 2003c, chapter 6).
In the absence of explicit memory, one cannot reason backward and infer a
repressed or dissociated memory from dreams, sudden feelings, and so forth.
These alleged implicit markers of buried memories can arise from many causes.
Moreover, implicit memories are just as subject to change and distortion as
explicit ones are (Lustig & Hasher, 2001). Finally, the notion that people can
undergo trauma yet retain only implicit traces (e.g., conditioned emotional
responses) contradicts recent critiques of the literature on conditioning without awareness (Lovibond & Shanks, 2001; Shanks & Lovibond, 2001). The

147

clinical implications are clear: if people are suciently traumatized to acquire

conditioned emotional responses and other implicit memories of the experience, they will retain explicit memories as well.

Dispelling Semantic Confusion

The debate about repressed and recovered memories has been partly fueled by
confusion concerning the meaning of basic concepts such as remembering and
forgetting. At times, the debate has been inamed by proponents on each side
who use the same words for dierent issues. Consider the verb forget. In accordance with everyday parlance, many of our research subjects say, I was
abused as a child, but then I forgot about it. Further questioning claries what
they mean. Many say they deliberately tried to push these experiences out of
their minds after they occurred. Engaged by other matters, they often succeeded
in their attempts not to think about their abuse, sometimes for long periods.
Prior to their enrolling in our research program, they encountered cues that
reminded them of their forgotten abuseabuse that had not entered their
mind for some time.
Other subjects say, I never forgot that I was abused as a child. Obviously,
memories of their abuse were often absent from awareness for hours, days, or
perhaps even weeks at a time, only to pop into mind either spontaneously or
when triggered by some reminder.
Some traumatic amnesia theorists seem to believe that memories of sexual
abuse ought to come to mind with reasonable frequency, and if someone does not
think about the abuse for months or years at a time, some mechanism must be
actively inhibiting its retrieval during these periods of forgetting. Because these
theorists assume that sexual abuse memories are highly emotionally charged, they
assume that a failure of these memories to enter the mind during long periods must
be attributable to an active force keeping them out of awareness.
The problem with this assumption is that the evidence for this postulated
mechanismrepression or dissociationis inferred from the very phenomena
it is designed to explain. In these cases, there is no independent evidence of inhibitory mechanisms other than the mere absence of certain memories from
awareness for periods of time. To be sure, cognitive psychologists have devised
ingenious methods for conrming active inhibitory processes in memory, at least
for neutral words (e.g., Anderson & Green, 2001). But to defend a claim of active inhibition, one should show that people are unable to recall their abuse in
the presence of adequate retrieval cues, such as the question: Were you ever
sexually abused as child? An inability to recall abuse when it should be readily
accessible provides a stronger basis for inferring inhibition than does the mere
failure to think about it for periods of time (everyday forgetting), which does not
require postulation of any inhibitory mechanism.

148

Thus, merely because someone has not thought about something for a long
timehas forgotten itdoes not mean that the person has been unable to
remember it. One need not postulate any special mechanisms to explain why
someone tried not to think about something unpleasant and managed not to
think about it for long stretches of time. Only an inability to remember when
exposed to adequate retrieval cues can provide a reasonable basis for labeling the
phenomenon as amnesia (assuming, of course, that the event was encoded in
the rst place). Until such evidence is adduced, the most parsimonious explanation for not thinking about trauma for long periods of timefor forgetting it
must lie with the conventional memory and forgetting mechanisms of cognitive
psychology, not repression or traumatic dissociation.
Finally, some recovered memories may not correspond to genuine events.
Some recovered memories are likely false. In our research, we have noted two
strikingly dierent types of recovered memory experience. In one type, subjects
are suddenly reminded of events that they had not thought about in many years.
They are surprised at their recollection but not at the content of the memory per
se. For example, in one of our cases, the subject suddenly recalled her own abuse
on learning that her daughter had just been molested. These individuals often
remark, My God! I cant believe I forgot that.
In the other type, subjects realize that they are abuse survivors, sometimes
gradually recalling new memories over a time. For example, one of our cases
recovered memories of being involved in cult abuse and cannibalism. This kind
of recollection diers from those in which the person is surprised not by the content of the recollection but by not having thought about it for so long. The second type of experience, we suspect, may be more likely to involve false memory
than the rst type.

Note
Preparation of this chapter was supported in part by NIMH grant MH61268 awarded
to the first author.

References
American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders. 3rd ed., rev. Washington, DC: APA.
Anderson, M. C., & Green, C. (2001). Suppressing unwanted memories by executive control. Nature, 410, 366369.
Anderson, M. C., & Neely, J. H. (1996). Interference and inhibition in memory retrieval. In E. L. Bjork & R. A. Bjork (Eds.), Memory (pp. 237313). San Diego,
CA: Academic Press.
Archibald, H. C., & Tuddenham, R. D. (1965). Persistent stress reaction after combat: A 20-year follow-up. Archives of General Psychiatry, 12, 475481.
Arrigo, J. M., & Pezdek, K. (1997). Lessons from the study of psychogenic amnesia.
Current Directions in Psychological Science, 6, 148152.

149

Bass, E., & Davis, L. (1988). The courage to heal: A guide for women survivors of child
sexual abuse. New York: Harper and Row.
Beck, A. T., & Steer, R. A. (1987). Beck Depression Inventory manual. San Antonio,
TX: The Psychological Corporation.
Bernstein, E. M., & Putnam, F. W. (1986). Development, reliability, and validity of
a dissociation scale. Journal of Nervous and Mental Disease, 174, 727735.
Blume, E. S. (1990). Secret survivors: Uncovering incest and its aftereects in women.
New York: Wiley.
Briere, J., & Conte, J. (1993). Self-reported amnesia for abuse in adults molested as
children. Journal of Traumatic Stress, 6, 2131.
Brown, D., Schein, A. W., & Hammond, D. C. (1998). Memory, trauma treatment,
and the law. New York: Norton.
Cassiday, K. L., McNally, R. J., & Zeitlin, S. B. (1992). Cognitive processing of trauma
cues in rape victims with post-traumatic stress disorder. Cognitive Therapy and
Research, 16, 282295.
Ceci, S. J., & Loftus, E. F. (1994). Memory work: A royal road to false memories?
Applied Cognitive Psychology, 8, 351364.
Cheit, R. E. (1998). Consider this, skeptics of recovered memory. Ethics and Behavior, 8, 141160.
Cheit, R. E. (1999). Junk skepticism and recovered memory: A reply to Piper. Ethics
and Behavior, 9, 295318.
Cheyne, J. A., Newby-Clark, I. R., & Rueer, S. D. (1999). Relations among hypnagogic and hypnopompic experiences associated with sleep paralysis. Journal of
Sleep Research, 8, 313317.
Cheyne, J. A., Rueer, S. D., & Newby-Clark, I. R. (1999). Hypnagogic and hypnopompic hallucinations during sleep paralysis: Neurological and cultural construction of the night-mare. Consciousness and Cognition, 8, 319337.
Christianson, S.-. (1992). Emotional stress and eyewitness memory: A critical review. Psychological Bulletin, 112, 284309.
Clancy, S. A. (2000). Child Abuse Survivor Questionnaire. Unpublished questionnaire,
Department of Psychology, Harvard University, Cambridge, MA.
Clancy, S. A., McNally, R. J., & Schacter, D. L. (1999). Eects of guided imagery on
memory distortion in women reporting recovered memories of childhood sexual
abuse. Journal of Traumatic Stress, 12, 559569.
Clancy, S. A., McNally, R. J., Schacter, D. L., Lenzenweger, M. F., & Pitman, R. K.
(2002). Memory distortion in people reporting abduction by aliens. Journal of
Abnormal Psychology, 111, 455461.
Clancy, S. A., Schacter, D. L., McNally, R. J., & Pitman, R. K. (2000). False recognition in women reporting recovered memories of sexual abuse. Psychological
Science, 11, 2631.
Courtois, C. A. (1992). The memory retrieval process in incest survivor therapy.
Journal of Child Sexual Abuse, 1, 1531.
Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in
immediate recall. Journal of Experimental Psychology, 58, 1722.
Dollinger, S. J. (1985). Lightning-strike disaster among children. British Journal of
Medical Psychology, 58, 375383.

150

Dubner, A. E., & Motta, R. W. (1999). Sexually and physically abused foster care
children and posttraumatic stress disorder. Journal of Consulting and Clinical
Psychology, 67, 367373.
Easterbrook, J. A. (1959). The eect of emotion on cue utilization and the organization of behavior. Psychological Review, 66, 183201.
Eder, M. D. (1917). War-shock. London: William Heinemann.
Femina, D. D., Yeager, C. A., & Lewis, D. O. (1990). Child abuse: Adolescent records
vs. adult recall. Child Abuse and Neglect, 14, 227231.
Foa, E. B., Feske, U., Murdock, T. B., Kozak, M. J., & McCarthy, P. R. (1991). Processing of threat-related information in rape victims. Journal of Abnormal Psychology, 100, 156162.
Freyd, J. J. (1996). Betrayal trauma: The logic of forgetting childhood abuse. Cambridge,
MA: Harvard University Press.
Garry, M., Manning, C. G., Loftus, E. F., & Sherman, S. J. (1996). Imagination ination: Imagining a childhood event inates condence that it occurred.
Psychonomic Bulletin & Review, 3, 208214.
Golding, J. M. (in press). Understanding directed forgetting. In A. Wenzel & D. C.
Rubin (Eds.), Cognitive methods in psychopathology research. Washington, DC:
American Psychological Association Press.
Goodman, G. S., Ghetti, S., Quas, J. A., Edelstein, R. S., Alexander, K. W., Redlich,
A. D., et al. (2003). A prospective study of memory for child sexual abuse: New
ndings relevant to the repressed-memory controversy. Psychological Science,
14, 113118.
Heaps, C., & Nash, M. (1999). Individual dierences in imagination ination.
Psychonomic Bulletin and Review, 6, 313318.
Henderson, J. L., & Moore, M. (1944). The psychoneuroses of war. New England Journal of Medicine, 230, 273278.
Herman, J. L. (1992). Trauma and recovery. New York: Basic Books.
Johnson, H. M. (1994). Processes of successful intentional forgetting. Psychological
Bulletin, 116, 274292.
Joseph, R. (1999). The neurology of traumatic dissociative amnesia: Commentary and literature review. Child Abuse & Neglect, 23, 715727.
Keane, T. M., Caddell, J. M., & Taylor, K. L. (1988). Mississippi Scale for CombatRelated Posttraumatic Stress Disorder: Three studies in reliability and validity.
Journal of Consulting and Clinical Psychology, 56, 8590.
Keane, T. M., Kolb, L. C., Kaloupek, D. G., Orr, S. P., Blanchard, E. B., Thomas,
R. G., et al. (1998). Utility of psychophysiological measurement in the diagnosis of posttraumatic stress disorder: Results from a Department of Veterans Aairs Cooperative Study. Journal of Consulting and Clinical Psychology,
66, 914923.
Kirschbaum, C., Wolf, O. T., May, M., Wippich, W., & Hellhammer, D. H. (1996).
Stress- and treatment-induced elevations of cortisol levels associated with impaired declarative memory in healthy adults. Life Sciences, 58, 14751483.
Kotagal, S., Rawlings, C. A., Chen, S.-C., Burris, G., & Nouri, S. (1982). Neurologic,
psychiatric, and cardiovascular complications in children struck by lightning.
Pediatrics, 70, 190192.

151

Langer, L. L. (1991). Holocaust testimonies: The ruins of memory. New Haven, CT: Yale
University Press.
Lindsay, D. S., & Read, J. D. (1994). Psychotherapy and memories of childhood sexual

abuse: A cognitive perspective. Applied Cognitive Psychology, 8, 281338.

Loftus, E., & Ketcham, K. (1994). The myth of repressed memory. New York: St. Martins.

Lovibond, P. F., & Shanks, D. R. (2002). The role of awareness in Pavlovian conditioning: Empirical evidence and theoretical implications. Journal of Experimental Psychology: Animal Behavior Processes, 28, 326.
Lupien, S. J., & McEwen, B. S. (1997). The acute eects of corticosteroids on cognition: Integration of animal and human model studies. Brain Research Reviews,
24, 127.
Lustig, C., & Hasher, L. (2001). Implicit memory is not immune to interference.
Psychological Bulletin, 127, 618628.
McGaugh, J. L., Ferry, B., Vazdarjanova, A., & Roozendaal, B. (2000). Amygdala:
Role in modulation of memory storage. In J. P. Aggleton (Ed.), The amygdala: A
functional analysis (2nd ed., pp. 391423). Oxford: Oxford University Press.
McNally, R. J. (1995). Automaticity and the anxiety disorders. Behaviour Research
and Therapy, 33, 747754.
McNally, R. J. (1998). Experimental approaches to cognitive abnormality in posttraumatic stress disorder. Clinical Psychology Review, 18, 971982.
McNally, R. J. (2001). The cognitive psychology of repressed and recovered memories
of childhood sexual abuse: Clinical implications. Psychiatric Annals, 31, 509514.
McNally, R. J. (2003a). Progress and controversy in the study of posttraumatic stress
disorder. Annual Review of Psychology, 54, 229252.
McNally, R. J. (2003b). Recovering memories of trauma: A view from the laboratory. Current Directions in Psychological Science, 12, 3235.
McNally, R. J. (2003c). Remembering trauma. Cambridge, MA: Belknap Press/
Harvard University Press.
McNally, R. J., Clancy, S. A., & Schacter, D. L. (2001). Directed forgetting of trauma
cues in adults reporting repressed or recovered memories of childhood sexual
abuse. Journal of Abnormal Psychology, 110, 151156.
McNally, R. J., Clancy, S. A., Schacter, D. L., & Pitman, R. K. (2000a). Cognitive
processing of trauma cues in adults reporting repressed, recovered, or continuous memories of childhood sexual abuse. Journal of Abnormal Psychology, 109,
355359.
McNally, R. J., Clancy, S. A., Schacter, D. L., & Pitman, R. K. (2000b). Personality
proles, dissociation, and absorption in women reporting repressed, recovered,
or continuous memories of childhood sexual abuse. Journal of Consulting and
Clinical Psychology, 68, 10331037.
McNally, R. J., Kaspi, S. P., Riemann, B. C., & Zeitlin, S. B. (1990). Selective processing of threat cues in posttraumatic stress disorder. Journal of Abnormal Psychology, 99, 398402.
McNally, R. J., Lasko, N. B., Clancy, S. A., Macklin, M. L., Pitman, R. K., & Orr, S. P.
(2002). Psychophysiologic responding during script-driven imagery in people reporting abduction by space aliens. Manuscript submitted for publication.
McNally, R. J., Metzger, L. J., Lasko, N. B., Clancy, S. A., & Pitman, R. K. (1998).

152

Directed forgetting of trauma cues in adult survivors of childhood sexual abuse

with and without posttraumatic stress disorder. Journal of Abnormal Psychology, 107, 596601.
Orr, S. P., Lasko, N. B., Metzger, L. J., Berry, N. J., Ahern, C. E., & Pitman, R. K.
(1998). Psychophysiologic assessment of women with posttraumatic stress
disorder resulting from childhood sexual abuse. Journal of Consulting and Clinical Psychology, 66, 90613.
Orr, S. P., & Roth, W. T. (2000). Psychophysiological assessment: Clinical applications for PTSD. Journal of Aective Disorders, 61, 225240.
Poole, D. A., Lindsay, D. S., Memon, A., & Bull, R. (1995). Psychotherapy and the
recovery of memories of childhood sexual abuse: U.S. and British practitioners
opinions, practices, and experiences. Journal of Consulting and Clinical Psychology, 63, 426437.
Pope, H. G., Jr., Oliva, P. S., & Hudson, J. I. (1999). Repressed memories: The scientic status. In D. L. Faigman, D. H. Kaye, M. J. Saks, & J. Sanders (Eds.), Modern
scientic evidence: The law and science of expert testimony (Vol. 1), Pocket part
(pp. 115155). St. Paul, MN: West.
Read, J. D. (1997). Memory issues in the diagnosis of unreported trauma. In J. D.
Read & D. S. Lindsay (Eds.), Recollections of trauma: Scientic evidence and clinical practice (pp. 79108). New York: Plenum.
Robinson, K. J., & Roediger, H. L., III. (1997). Associative processes in false recall
and false recognition. Psychological Science, 8, 231237.
Roediger, H. L., III, & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 803814.
Sargant, W., & Slater, E. (1941). Amnesic syndromes in war. Proceedings of the Royal
Society of Medicine, 34, 757764.
Schooler, J. W., Bendiksen, M., & Ambadar, Z. (1997). Taking the middle line: Can
we accommodate both fabricated and recovered memories of sexual abuse? In
M. A. Conway (Ed.), Recovered memories and false memories (pp. 251292). Oxford: Oxford University Press.
Shanks, D. R., & Lovibond, P. F. (2002). Autonomic and eyeblink conditioning are
closely related to contingency awareness: Reply to Wiens and hman (2002)
and Manns et al. (2002). Journal of ExperimentalPsychology: Animal Behavior
Processes, 28, 3842.
Shin, L. M., McNally, R. J., Kosslyn, S. M., Thompson, W. L., Rauch, S. L., Alpert,
N. M., et al. (1999). Regional cerebral blood ow during script-driven imagery
in childhood sexual abuse-related PTSD: A PET investigation. American Journal of Psychiatry, 156, 575584.
Swihart, G., Yuille, J., & Porter, S. (1999). The role of state-dependent memory in
red-outs. International Journal of Law and Psychiatry, 22, 199212.
Tellegen, A. (1982). Brief manual for the Dierential Personality Questionnaire. Minneapolis: Department of Psychology, University of Minnesota.
Tellegen, A., & Atkinson, G. (1974). Openness to absorbing and self-altering experiences (absorption), a trait related to hypnotic susceptibility. Journal of Abnormal Psychology, 83, 268277.

153

Terr, L. C. (1991). Childhood traumas: An outline and overview. American Journal

of Psychiatry, 148, 1020.
Torrie, A. (1944, January 29). Psychosomatic casualties in the Middle East. Lancet,
246, 139143.
van der Kolk, B. A. (1994). The body keeps the score: Memory and the evolving psychobiology of posttraumatic stress. Harvard Review of Psychiatry, 1, 253265.
van der Kolk, B. A., & Fisler, R. (1995). Dissociation and the fragmentary nature of
traumatic memories: Overview and exploratory study. Journal of Traumatic
Stress, 8, 505525.
Vreven, D. L., Gudanowski, D. M., King, L. A., & King, D. W. (1995). The civilian
version of the Mississippi PTSD Scale: A psychometric evaluation. Journal of
Traumatic Stress, 8, 91109.
Wagenaar, W. A., & Groeneweg, J. (1990). The memory of concentration camp
survivors. Applied Cognitive Psychology, 4, 7787.
Wilkinson, C. B. (1983). Aftermath of a disaster: The collapse of the Hyatt Regency
Hotel skywalks. American Journal of Psychiatry, 140, 11341139.
Williams, J. M. G., Mathews, A., & MacLeod, C. (1996). The emotional Stroop task
and psychopathology. Psychological Bulletin, 120, 324.
Williams, L. M. (1994). Recall of childhood trauma: A prospective study of womens
memories of child sexual abuse. Journal of Consulting and Clinical Psychology,
62, 11671176.
Winograd, E., Peluso, J. P., & Glover, T. A. (1998). Individual dierences in susceptibility to memory illusions. Applied Cognitive Psychology, 12, S5S27.
Yerkes, R. M., & Dodson, J. D. (1908). The relation of strength of stimulus to rapidity of habit-formation. Journal of Comparative Neurology and Psychology, 18,
459482.

154

An Overview of Research Findings
and Their Implications

xperimental support for the idea that anxious individuals may display anomalous performance on cognitive tasks extends back
many decades (Calvin, Koons, Bingham, & Fink, 1953). Early research focused
primarily on investigating the patterns of cognitive impairment associated with
anxiety. Now convincing evidence shows that anxious participants indeed demonstrate performance decrements across measures of cognitive functioning, especially dicult and demanding memory tests (cf. Eysenck, 1982). However,
contemporary accounts of such performance decits attribute them, in part, to
anxious individuals tendency to process task-irrelevant information associated
with their worries. Recently, many researchers have directly examined whether
the emotional valence of stimulus materials can moderate anxiety-linked patterns of performance on cognitive tasks. One hypothesis suggests that anxious
individuals may display memory biases that selectively favor the retrieval of emotionally threatening information. Following a brief review of research showing
that anxiety can be associated with memory decits, we provide an overview of
the extensive experimental work testing whether anxious participants preferentially retrieve threat-related memories. Support for this position is by no means
compelling, though favorable evidence has been more apparent in certain populations of anxious individuals and for certain types of memory tasks.

Memory Decits in Anxiety

Like other dysphoric states such as depression, anxiety is associated with cognitive performance decits, including those involving memory (Eysenck, 1982;
Humphreys & Revelle, 1984; Mueller, 1992). In a highly inuential account,
155

Eysenck and Calvo (1992) proposed that anxiety aects information processing
in two quite dierent ways. First, by eliciting task-irrelevant cognitive activity
such as worry, anxiety consumes processing resources within working memory
(Baddeley, 1986). At the same time, however, their concern to avoid poor performance motivates anxious individuals to invest greater eort in task performance to compensate for their capacity limitations. If the task is not too
demanding, this additional eort may lead anxious individuals to perform as well
as, or even better than, nonanxious individuals. In contrast, if a task places high
demands on temporary storage capacity, performance of anxious participants
likely suers.
In general, research ndings have supported predictions derived from Eysenck
and Calvos (1992) account. For example, Darke (1988b) demonstrated that
highly anxious participants display a lower digit span than do less anxious individuals and show poorer ability to intentionally recall words presented at the end
of sentences. In another study, Darke (1988a) observed that, though highly
anxious readers demonstrate a normal ability to draw the elementary inferences
necessary to comprehend text, their performance is impaired on inference tasks
that plausibly demand greater processing resources. Consistent with the hypothesis that anxiety is associated with restricted cognitive capacity, MacLeod and
Donellan (1993) found that increases in mental load disproportionately impaired
reasoning task performance in highly anxious individuals.
One important but unresolved issue concerns whether performance decits
in anxious individuals result from mood state at the time of testing or represent
an enduring characteristic of individuals with high trait vulnerability to anxiety. Eysenck and Calvos (1992) hypothesis causally implicates anxious mood
state in such cognitive impairment. Under many circumstances, the high correlation between trait and state anxiety can make dissociating their respective
roles dicult. However, state anxiety elevated by quite specic worries still causes
impaired cognitive performance. For example, anxiety about mathematics is
associated with temporary impairment of the ability to hold numbers in memory
and with lesser ability to avoid distracting thoughts when one performs calculations (Ashcroft & Kirk, 2001). Although little direct evidence is available concerning longer-term memory consequences, Ashcroft and Kirk suggest that these
temporary problems likely contribute to impaired long-term knowledge about
mathematical operations.
Such performance decits clearly need to be a part of discussion on memory
in anxiety disorders. However, the bulk of research and theory to be reviewed
here has focused on whether anxiety is associated with the relative enhancement
of memory for threatening material. There are several reasons why this pattern
of memory selectivity might be expected. One is that most people tend to remember the gist of emotional events, such as the content of threatening pictures,
better over time (cf. chapter 1 here). Furthermore, people remember emotional
events better than matched unemotional events perhaps because of individual

156

dierences in emotional arousal (see Andrews, cited in Reisberg & Heuer, 1992).
Various explanations for this eect have been proposed (Christianson, 1992;
Wessel, van der Kooy, & Merckelbach, 2000). Selective attention to the emotional
focus of a scene may lead to better encoding of information in that location, rather
than in other less wellattended locations. Also, people may elaborate disproportionately on the meaning of emotional scenes, leading to more extensive
memorial representations and consequently to improved retrieval. Additionally,
emotional arousal at the time of processing the scene may directly enhance the
durability of the memory trace (Revelle & Loftus, 1990; Cahill & McGaugh, 1995).
Though these are not mutually exclusive possibilities, particularly strong
evidence supports the rst suggestion, that the elevation of anxiety may be associated with an attentional focus on threat during encoding. People with heightened levels of anxiety are more likely than nonanxious individuals to let
threatening words or pictures hold their attention (Fox, Russo, Bowles, & Dutton,
2001; Mathews & MacLeod, 1994; Yiend & Mathews, 2001). Even if anxiety exerts no direct eect on the retrieval process itself, greater attention to threatening aspects of stimuli during encoding likely enhances later memory for such
information. However, whether more attention to a threatening stimulus event
results in better or worse memory for that event also may depend on the type of
encoding, as well as the task used to assess memory. If attention to threat during encoding is accompanied by subsequent processing of the distinctive meaning of this emotional information, then its later recall is likely enhanced. On the
other hand, if attention to threat during encoding results, for example, principally in registration only of the aversive quality of the feelings evoked, then there
may be fewer distinctive features within the memory representation to help in
its later recall. A memory test requiring retrieval of certain aspects of an event
could thus reveal either good or poor recall, despite the greater allocation of attention resources to this event at the time of presentation, depending on how
the event has been encoded. Therefore, an encoding bias for threatening information may result in preferential memory for such material, even when the retrieval process itself is not biased to favor this type of information. Conversely, if
there is a memory advantage for threatening information or not, such information nevertheless may have recruited disproportionate attentional resources
during encoding.
The evidence we review here suggests that increased selective attention to
threatening stimuli, demonstrated so reliably by highly anxious individuals, is
not consistently associated with better subsequent memory for this information.
In contrast to research on depressed participants, who demonstrate recall advantages for negative information that are easy to replicate, the bulk of research
on memory bias in anxious individuals has proven surprisingly negative. Anxious individuals commonly display no greater tendency to recall emotionally
threatening material than less anxious individuals exposed to the same material do. Exceptions to this general rule will be highlighted during our review and

157

will be discussed at the end of this chapter. To anticipate, despite variable results
across studies, greater evidence of a recall bias favoring threatening information
has been obtained in studies examining panic disorder patients than in studies
of other anxiety disorders or of nonclinical individuals with elevated levels of trait
or state anxiety. Also, measures of implicit memory, especially those involving
perceptually driven processes, may detect an anxiety-linked memory advantage
for threatening information more often than measures of explicit memory, such
as recall and recognition.
In drawing conclusions, we distinguish between eects that could arise solely
from selective encoding and those attributable to biases aecting subsequent
access to, or retrieval of, emotionally valenced information. We suggest that an
anxiety-linked recall advantage for threatening material is more likely when
participants are free to initially encode emotional material without restrictions.
Under such free encoding conditions, highly anxious groups are disproportionately likely to encode personally relevant meanings of threatening stimulus
items, and this facilitates memory performance for such information, even without anxiety-linked bias in the retrieval process itself.
The review that follows is organized into sections that separately describe
research investigating the association between anxiety and tests of autobiographical memory, recall, recognition, and implicit memory. Within these main
sections, we review studies carried out on nonclinical populations selected according to their scores on anxiety questionnaires, as well as studies that have
assessed patients suering from anxiety disorders, including generalized anxiety disorder (GAD), phobias, obsessive compulsive disorder (OCD), posttraumatic
stress disorder, and panic disorder.

Memory for Emotional Information

in Anxiety Disorders

Recall of Autobiographical Memories

A number of researchers have sought to establish whether autobiographical
memory functioning is biased within anxious individuals. For example, Richards
and Whittaker (1990) asked high and low trait-anxious individuals to recall a
specic personal memory evoked by either happy cue words (e.g., pleased, lucky)
or anxiety-related cue words (e.g., anger, pain). High trait-anxious participants,
unlike low trait anxious individuals, were signicantly faster in retrieving autobiographical memories elicited by the anxiety-related cue words, suggesting
that anxiety-related memories may be disproportionately accessible for such
participants. However, because Richards and Whittaker did not examine the
emotional content of the memories produced, their data instead could reect high
trait-anxious participants producing more memories with objectively neutral

158

content in response to the anxiety-related cues. That is, high trait-anxious individuals may simply have associated more of their memories with the anxietyrelated cues, regardless of the emotional tone of the memories themselves.
Subsequent studies by Burke and Mathews (1992), carried out on generalized
anxiety disorder (GAD) patients and nonanxious control participants, employed
independent raters to assess the emotional tone of memory content. GAD patients
and controls were presented with emotionally neutral cue words and were instructed to recall the rst personal memory brought to mind by each item. Independent blind ratings revealed a nonsignicant trend for the GAD patients to
report memories that were more threatening than those reported by controls.
In a second study, GAD patients, unlike controls, produced signicantly more
memories when directly instructed to recall anxious autobiographical memories to neutral cues. However, judges classied only 80% of these memories to
be truly anxious in GAD patients, in comparison to 92% in the case of control
participants. Thus, GAD patients may indeed use more liberal criteria when dening recollected events as anxietyrelevant.
More important, it has proven dicult to replicate the early nding that anxiety is associated with biased autobiographical memory performance. Levy and
Mineka (1998) obtained independent ratings of the emotional tone of memories
elicited in response to neutral, positive, and threatening cue words in high and low
trait-anxious participants. There were no group dierences in the relative number of memories produced in response to each type of cue, in the emotional tone of
these memories, or in the pattern of recall latencies across these three cue conditions. Rapee, McCallum, Melville, Ravenscroft, and Rodney (1994) presented social phobics and nonanxious participants with social-threat and neutral cue words
and instructed them to retrieve associated memories of events either they or a close
friend or relative experienced. Across all conditions, there was no evidence that
the social phobics diered from the nonanxious participants in their tendency to
recall negative events. Wenzel, Jackson, and Holt (2002) also presented social
phobics and nonanxious controls with socialthreat and neutral cue words,
instructing them to report the rst specic personal memory evoked by each cue.
The social phobics showed no enhanced ability to recall memories in response to
the threatening retrieval cues.
Using a similar autobiographical memory task, with both happy and anxious
cue words, Wilhelm, McNally, Baer, and Florin (1997) found an equivalent pattern of retrieval latencies across obsessive-compulsive disorder patients and control participants. Even in posttraumatic stress disorder, where patients might be
expected to show better recall of trauma-related events, the experimental evidence provides little support for such an autobiographical memory bias. McNally,
Litz, Prassas, Shin, and Weathers (1994) presented Vietnam combat veterans,
some who suered from PTSD, with negative, positive, and neutral words as
retrieval cues, again instructing them to recollect personal memories elicited by
each cue word. All participants were faster to retrieve memories in response to

159

the negative cues than in response to the neutral or positive cues, but this eect
did not dier across groups with or without PTSD. [For evidence concerning the
forgetting of traumatic memories, see chapter 4Eds.]
Overall, the balance of evidence provides little support for the idea that anxiety disorders are associated with the facilitated recall of negative autobiographical material. In many ways, this nding is rather surprising, given the strong
possibility that individuals who develop anxiety pathology may actually have
experienced a disproportionate number of threatening events (Finlay-Jones &
Brown, 1981). Indeed, the fact that one cannot assume equivalence of negative
event frequency across anxious and control populations compromises interpretation of those few early studies that claimed to demonstrate evidence of an autobiographical memory bias. To overcome this methodological problem, many
researchers have turned away from autobiographical memory tasks and instead
have tested for memory bias following the presentation of to-be-recalled stimuli
under controlled conditions. We now turn to these studies.

Recall of Experimentally

Presented Stimuli

The emotional stimuli presented in such controlled experiments usually have

been single emotional words. Participants have either read these words or used
them in an imagery or judgment task, before an unexpected recall test for the
items. Following both reading and imagery tasks, Richards and French (1991)
assessed recall of emotionally threatening stimulus words (e.g., trapped, humiliated) and neutral words in high and low trait-anxious participants. Regardless
of whether these words had simply been read or had been employed to evoke
imagined scenes, no dierences in relative recall separated high and low traitanxious participants. Using a rather dierent encoding task, Nugent and Mineka
(1994) directed high and low trait-anxious participants to rate how much they
liked or disliked presented threatening and nonthreatening words, then unexpectedly tested ability to recall these words. Although there was some evidence
from an initial study to suggest relatively better recall of threat words in high
trait-anxious participants, this eect failed to emerge in a second experiment,
leading these researchers to conclude that trait anxiety is not reliably associated
with a recall advantage for threatening information.
In contrast, three articles have reported nding that the recall of negative
words is enhanced for high trait-anxious participants (Eysenck & Byrne, 1994;
Reidy & Richards, 1997; Russo, Fox, Bellinger, & Nguyen-Van-Tam, 2001).
Under certain combinations of encoding and recall tasks, Eysenck and Byrne
observed relatively facilitated recall of negative stimulus words in their high traitanxious participants, though correlational analysis failed to disentangle the
potential mediating roles of anxiety and depression, both elevated within this
sample. Similarly, following a self-descriptive or other-descriptive encoding task,

160

Reidy and Richards found that recall of self-encoded negative words was disproportionately good for high trait-anxious participants. However, this eect proved
to be correlated with depression rather than with anxiety level.
Russo et al. (2001) observed no dierences in the patterns of recall shown by
high and low trait-anxious individuals following a semantic encoding task that
required participants to rate the pleasantness of emotionally toned stimulus
words. They suggest that this nding could reect a ceiling eect, given that the
high level of overall recall left little scope for stimulus emotionality to inuence
memory performance. Consistent with this account, Russo et al. found greater
evidence of an anxiety-linked recall bias following an alternative encoding task
that required participants to make structural judgments about the stimulus
words (i.e., counting the number of syllables). Although recall was poorer following this structural encoding task than in the semantic encoding task, threat
words now were recalled disproportionately by the high trait participants. However, as the authors themselves acknowledge, this eect might be attributed to
dierential encoding and so does not require the conclusion that anxiety is associated with a biased retrieval process. The structural encoding task instructed
participants to attend to the syllabic format of each word while ignoring its semantic content. Previous research convincingly demonstrates that high traitanxious individuals display an impaired ability to direct attention away from the
semantic content of threatening stimuli, even when explicitly instructed to do
so (e.g., Williams, Mathews, & MacLeod, 1996; Wood, Mathews, & Dalgleish,
2001). Certainly, despite instructions to process syllabic structure, Russo et al.s
(2001) simple task permitted ample time for the processing of semantic content,
as each word was exposed for 4 seconds. Under these circumstances, one would
predict on the basis of established research ndings that high trait-anxious participants would attend disproportionately to threat word content. Thus, although
the net result was one of better memory for threat words in anxious participants,
compared to nonanxious individuals, it seems more parsimonious to regard this
result as further evidence of selective encoding, rather than of an anxiety-linked
bias in the memory retrieval process.
Those investigators who have examined memory for emotional stimulus
words in clinical populations have not obtained consistent evidence of a recall
advantage for threatening material in patients suering from anxiety disorders.
In an early study, Mogg, Mathews, and Weinman (1987) required patients
with GAD, and nonanxious control participants, to judge whether positive and
negative words were either self-descriptive or other-descriptive. In a subsequent
unexpected memory test, there was no group dierence in relative ability to
recall these two classes of stimulus words. Indeed, a subsidiary analysis carried out on the subset of negative words judged to be most threatening revealed
that they actually were recalled disproportionately poorly by the anxiety patients. In a virtually identical experiment, reported by Mogg and Mathews
(1990), GAD patients did display a nonsignicant trend toward relatively better

161

recall of negative words, compared to nonanxious control participants, across

both encoding conditions. However, Mogg and Mathews noted that their GAD
patients also generated a disproportionate number of negative words as intrusion errors while performing the recall task, suggesting that the observed eects
may result from a negative response bias, rather than from any enhanced ability to retrieve the negative material from memory. Becker, Roth, Andrich, and
Margraf (1999) had GAD participants and nonanxious controls rate the ease
with which they could imagine scenes evoked by emotionally valenced words,
some closely related to the usual concerns of GAD patients, before giving them
an unexpected recall test for these words. No group dierences in pattern of
recall performance were observed, leading Becker et al. to conclude that GAD
is not associated with any memory advantage for threatening words, even
when they are relevant to disorder-related concerns. In a study of cued recall
in GAD patients and nonanxious controls, Mathews, Mogg, May, and Eysenck
(1989) used a similar encoding method but then provided three-letter word
stems that participants were directed to complete to yield one of the previously
exposed words. There were no dierences between groups in the observed pattern of cued recall accuracy. Otto, McNally, Pollack, Chen, and Rosenbaum
(1994) replicated this null result.
As noted, Russo et al. (2001) suggested that the failure to nd an anxietylinked memory bias in certain studies may reect ceiling eects, which render
the task insensitive to group dierences in the ability to recall valenced stimuli.
However, in a study comparing patients suering from GAD and those with clinical depression, Bradley, Mogg, and Williams (1995) also failed to obtain evidence
of any anxiety-linked recall bias, while nevertheless conrming their tasks sensitivity to individual dierences in the selective recall of emotional materials.
Participants rst rated depression-relevant words (e.g., discouraged, despair),
anxiety-relevant words (e.g., embarrassed, assaulted), positive words (e.g., paradise, bliss), and neutral words (e.g., bench, cereal) for their relevance to personal
concerns. In an unexpected subsequent memory test, the depressed patients
showed evidence of a recall bias favoring depression-related words. However,
once again there was no dierence between the recall proles of GAD patients
and nonanxious controls. In contrast, Friedman, Thayer, and Borkovec (2000)
did observe better recall of threat words in GAD patients than controls when they
used a less directive approach to initial stimulus processing. Specically, threat
and neutral words were presented for 8 seconds each, during which time the
participant was required simply to read each item aloud. When memory for these
words then was tested, GAD patients recalled more threat than nonthreat words
and more threat words than did control participants. The initial word processing task used by Friedman et al. was similar, in potentially important ways, to
that employed by Russo et al. This task permitted, but did not require, the processing of semantic content, and the exposure time was long. These conditions
likely allowed participants ample time and freedom to encode the words in quite

162

dierent ways, so an anxiety-linked tendency to selectively encode the aversive

semantic content of the threat words may underlie the enhanced recall of such
materials shown by the anxious patients.
Studies that have investigated the memorial characteristics of social anxiety
have failed to yield evidence that it is associated with any recall advantage for
threatening information. Sanz (1996) found no dierences between socially anxious students and nonanxious controls in their recall of neutral, positive, and negative trait words previously encountered in a self-descriptiveness rating task, even
when these words were closely related to social anxiety. When comparing memory
for words appraised earlier by participants for emotional tone, Rapee et al. (1994)
likewise found no dierence between clinical social phobics and nonanxious controls in the relative recall of neutral words, positive words, and negative words
related either to general threat or specically to social threat. In a second study,
Rapee et al. modied their encoding task, now requiring socially phobic patients
and control participants to imagine a scene involving themselves and the referent
of each stimulus word and then to rate this scene for pleasantness. Afterward,
Rapee et al. tested cued recall, rather than free recall, by instructing participants
to complete three-letter word stems to produce previously encountered words.
Once again, however, no group dierence in the relative recall of neutral, positive,
and negative words emerged. Finally, in a third experiment designed to increase
ecological validity, Rapee et al. provided positive and negative feedback concerning participant performance on an initial task (e.g., your material was poorly organized) and then unexpectedly tested recall of this emotional feedback. Compared
to nonanxious control participants, social phobic patients displayed no facilitated
recall of the negative information. Indeed, these patients actually recalled a disproportionately small number of the negative feedback statements.
Cloitre, Cancienne, Heimberg, Holt, and Liebowitz (1995) and Lundh and Ost
(1997) also examined the ability of socially phobic patients to recall emotional
stimuli, using free and cued recall procedures, respectively, and found no evidence of any anxiety-linked memory bias. Becker et al.s (1999) study, cited earlier, included a group of social phobic as well as GAD patients. These social phobic
patients, like the GAD patients, did not dier from nonanxious control participants in terms of their relative ability to recall threat words associated with social concerns, general threat words, neutral words, and positive words. Amir,
Coles, Brigidi, and Foa (2001) investigated whether practicing the recall of neutral, positive, or social threat words would reveal dierences between socially
phobic and nonanxious participants. Words initially were exposed together with
a preceding category label, such as FruitMango or JobStress. Subsequently,
half of the participants were given cued retrieval practice, which involved presenting the category labels together with the initial fragment of the associated
test word (e.g., FruitMa- -), which had to be completed to produce the original item. Finally, all participants were given a cued recall test, within which they
were shown parent category words and told to recall the words paired with them.

163

Without retrieval practice, the groups did not dier in their relative ability to
recall the emotional categories of stimulus words. Retrieval practice did not elicit
any relative recall advantage for negative words in the socially phobic patients.
Indeed, the socially phobic patients gained less than controls from retrieval practice with socially threatening words. Collectively, therefore, these experimental
studies provide no empirical support for the proposal that the recall of emotional
information is enhanced in social phobia.
Fewer experimental studies have investigated whether patients suering from
OCD or PTSD demonstrate preferential recall of threatening information. One
problem that arises quite commonly in these studies is that supposedly emotional
material may be neutral for control participants. For example, Radomsky and
Rachman (1999) found better recall of contaminated than clean objects in
OCD patients, relative to nonanxious control participants. However, contamination was manipulated by varying whether objects were touched by the experimenter, a distinction unlikely to aect the control group emotionally.
Within a directed forgetting paradigm, Wilhelm et al. (1996) exposed OCD
patients and control participants to neutral words (e.g., spoon, curtain), positive words, (e.g., laugh, condent), and negative words often associated with
obsessive concerns (e.g., danger, disease). In this experimental approach, after
each word has been presented, a signal directs the participant to either remember or forget that item. Wilhelm et al. found that when participants were instructed to remember words, the two groups demonstrated equivalent
subsequent recall across all three classes of stimuli. However, for those items
they were directed to forget, the OCD patients subsequent recalled the negative words more often than the controls did. Using a similar paradigm, McNally,
Metzger, Lakso, Clancy, and Pitman (1998), exposed survivors of childhood
abuse, with or without PTSD, to neutral words, positive words, and negative
words related to their traumatic experience (e.g., molested, abuse). Their results were similar to those obtained by Wilhelm et al. using OCD patients. Participants did not dier in their patterns of recall for the words they had been
instructed to remember, but for those they had been directed to forget, the PTSD
patients recalled a disproportionate number of the negative items. Anxiety is
associated with a reduced capacity to inhibit attention to threatening meanings during the encoding of stimulus words (Wood et al., 2001). Directing
participants to forget a word, at the time of its initial exposure, means that they
must try not to process the meaning of that stimulus, which is disproportionately dicult for anxious individuals when this meaning is emotionally threatening. Thus, the relative diculty PTSD and OCD patients appear to experience,
when trying not to process threat words under such forget instructions, may
reect this well-established encoding eect, rather than any anxiety-linked bias
in the retrieval process itself.
A related issue arises in a study by Vrana, Roodman, and Beckham (1995),
who exposed Vietnam veterans with and without PTSD to neutral and threat-

164

ening words (e.g., reght, death), within an emotional Stroop task (Williams
et al., 1996). Consistent with much previous research, PTSD participants were
disproportionately slow to color-name the threatening words during their initial exposure, an eect commonly attributed to anxious individuals diculty in
suppressing the processing of these words negative semantic content. Vrana et
al. also found that the PTSD patients subsequently recalled a disproportionate
number of the threatening words. However, because of the observed pattern of
color-naming interference, these researchers quite reasonably ascribe this group
dierence to the PTSD patients tendency to selectively encode threat word meaning during the color-naming task, despite instructions to ignore word content.
Thus, the few available studies of OCD or PTSD patients are either rendered problematic by the use of stimuli unlikely to be threatening for control participants
or they invite explanation in terms of an anxiety-linked failure to inhibit the
encoding of negative meanings, not in terms of an anxiety-linked bias in the
retrieval of information from memory.
In contrast, there is much more empirical support for the existence of a
memory bias that favors the retrieval of threatening information in panic disorder patients. Of course, some of the relevant studies suer from the methodological limitations already described, and null results sometimes have been obtained
even in sound experimental designs. For example, in an early study of agoraphobia (now considered to be closely associated with panic disorder), Nunn,
Stevenson, and Whalan (1984) assessed recall of words and short passages.
Agoraphobic patients recalled more words and propositions from threatening
passages than did nonanxious control participants. However, the threatening
material described activities such as shopping and included words such as street,
travel and cinema, which control participants probably would not consider
emotionally negative. Furthermore, Pickles and van den Broek (1988) failed to
replicate Nunn et al.s original ndings, despite employing a very similar experimental design. Otto et al. (1994) found no dierences between the patterns of
recall shown by panic disorder patients and nonanxious control participants for
neutral, positive, general threat, and panic-related words that earlier had been
rated for personal emotional signicance. Likewise, Rapee (1994) did not observe
any dierence in the patterns of recall shown by panic disorder patients and
nonclinical controls for diering classes of emotional words previously exposed
within a free association task.
However, these few negative reports are outweighed by a greater number of
studies that have found that panic disorder patients demonstrate enhanced recall of threatening information. McNally, Foa, and Donnell (1989) required
panic disorder patients and nonanxious control participants to rate anxietyrelated words (e.g., apprehensive, uptight) and other words unrelated to anxiety
(e.g., charming, outgoing) for self-descriptiveness. In an unexpected subsequent
memory test, control participants recalled fewer anxiety-related words than
anxiety-unrelated words, whereas this pattern reversed in the panic disorder

165

patients. Contrary to the ndings of Mogg and Mathews (1990), the emotional
valence of intrusion errors did not dier between patients and controls, suggesting that these results reect a genuine group dierence in recall, rather than in
response bias. Similar ndings have been reported in studies that have employed
dierent encoding tasks. For example, Lundh, Czyzykow, and Ost (1997) required
participants to imagine scenes evoked by each of their stimulus words and then
tested memory using a cued recall test that required three-letter word stems to
be completed to make previously exposed words. Control participants recalled
similar numbers of threat and neutral words, whereas panic disorder patients
recalled a disproportionate number of physically threatening words (but not
socially threatening words), with the magnitude of this recall advantage for
threat predicted by a measure of anxiety sensitivity.
Using the same encoding task with panic disorder patients and control group
participants, Becker, Rinck, and Margraf (1994) observed no group dierence
in the subsequent recall of positive and generally negative words (e.g., lonely,
brutal) but found that panic disorder patients did recall more of the negative
words specically related to panic (e.g., fainting, madness). Once again, such
words did not appear more commonly among the intrusion errors made by panic
patients, mitigating against a response bias explanation of the eect. In their
subsequent research, Becker et al. (1999) included three types of panic-related
words: somatic (e.g., palpitations, sweating), cognitive (e.g., helplessness, dying),
and situational (e.g., tunnel, airplane). Panic disorder patients demonstrated
enhanced recall of somatic and situational threat words but not of cognitive
threat words, without showing increased representation of such items among
their intrusion errors.
Cloitre and Liebowitz (1991) employed two quite dierent encoding tasks to
expose panic disorder patients and nonclinical controls to neutral, positive, and
threat words. One task required participants to decide whether each item described a feeling state, whereas the other task required them to decide only
whether the item was a legitimate English word. The nonanxious group subsequently recalled a similar number of words from all three emotional categories,
whereas the panic disorder patients recalled more threat words than either positive or neutral words, regardless of encoding task. Again, the patterns of intrusion errors were equivalent across groups, suggesting that the observed recall
eects could not be attributed to a group dierence in response bias. Cloitre,
Shear, Cancienne, and Zeitlin (1994) used another encoding task variant, in
which panic disorder and control participants assessed the semantic association
between members of word pairs, which could be neutral (e.g., gauge-metric),
positive (e.g., smiles-elated), or panic-related (e.g., dizzy-faint). An unexpected
cued recall test then was given, with participants shown the rst word of each
pair followed by the rst three letters of the second word, which they were required to complete to yield the original item. Control participants displayed

166

equivalent recall of the classes of emotional words, but panic disorder patients
recalled a disproportionate number of the panic-related threat words.
In summary, then, although recall studies have yielded little compelling experimental evidence to support a negative retrieval bias among individuals who report high levels of trait anxiety, or in patients suering from most clinical anxiety
disorders, they have provided rmer grounds for believing that such a retrieval bias
may characterize panic disorder. The few studies that have demonstrated a recall
advantage for threat stimuli among other anxious populations typically have tested
memory following encoding tasks that have either instructed participants to ignore word content or permitted individuals considerable freedom to choose how
they process stimulus content during encoding. When, instead, participants have
been directed to process the semantic content of stimuli in a specied manner
during encoding, typically no anxiety-linked bias in recall performance has been
observed. This pattern of ndings suggests that the occasionally observed superior recall of threatening words by such anxious participants likely results from
their established tendency to selectively process threat stimulus content during
encoding rather than from any bias in the retrieval process itself. The same is not
true, however, for panic disorder patients, for whom encoding instructions appear
to have little impact on observed patterns of selective recall. Across a wide range
of studies, varying in both encoding and recall methodology, panic disorder patients repeatedly have displayed a recall bias favoring threatening stimuli, suggesting that the retrieval of such information may indeed be facilitated within this
particular anxious population. We now will turn to the consideration of experiments that have employed measures of recognition, rather than recall, to test for
selective memory bias in anxious participants. Such studies continue to provide
little evidence that the retrieval of threatening information is generally enhanced
across diering populations of anxious participants.

Recognition Memory
Recognition memory diers from recall; for example, it does not need to engage
the search component of memory retrieval that characterizes recall, which may
make recognition measures less sensitive to emotional biases that exert their
primary inuence on this search process. However, recognition tasks have the
advantage of permitting assessment of memory for types of stimuli that do not
readily t within recall tasks, such as pictures of emotional faces. Furthermore,
recognition data can be subjected to signal detection analysis, allowing a distinction between measures of memory sensitivity (d) and response bias (B).
Across two experiments, Nugent and Mineka (1994) assessed high and low
trait-anxious participants recognition memory for threat and nonthreat words,
which previously had been rated for likeability, and subjected their data to signal detection analysis. In one of these studies, high trait individuals were found

167

to demonstrate a response bias favoring threat words. However, in neither of the

experiments was any dierence observed between the groups in their relative
memory sensitivity for the two classes of stimulus words. Similarly, Dalgleish
(1994) found no evidence from signal detection analysis to indicate that high
and low trait-anxious participants diered in relative ability to accurately recognize neutral, positive, or anxiety-related words that earlier had been encountered as solutions to an anagram task.
A number of researchers have compared the ability of GAD patients and
nonanxious controls to recognize threat and nonthreat words previously presented within an emotional Stroop task. Using signal detection analysis, both
Mathews and MacLeod (1985) and Mogg, Mathews, and Weinman (1989) failed
to nd any group dierence in the sensitivity of recognition memory for these
two word types. However, recognition performance was generally poor in these
studies, probably because the encoding task invited little processing of word
content. To counteract this limitation, MacLeod and McLaughlin (1995) modied the initial emotional Stroop task, by requiring participants to read each word
aloud after it had been color-named. This did indeed improve recognition
memory for the stimulus words, but still GAD patients and nonanxious controls
did not dier in relative recognition memory for the threat and nonthreat words.
This failure to observe an anxiety-linked recognition memory advantage for
threat words extends to other studies that have employed quite dierent encoding tasks. Mogg et al. (1987) examined recognition memory for negative and
positive trait adjectives that previously had been rated by participants for selfdescriptiveness or other-descriptiveness. Patients suering from GAD showed no
relative enhancement of recognition memory for threat words, compared to
nonanxious control participants. Indeed, the anxiety patients tended to recognize fewer threat words, and more positive words, than the nonanxious participants did, although this group dierence in recognition sensitivity was not
statistically signicant. Rapee et al. (1994) included recognition measures to
assess memory for words that participants previously had rated for likeability
but observed no dierences between social phobics and controls in ability to
recognize threat and nonthreat words. Cloitre et al. (1995) also found equivalent patterns of recognition memory for emotional words in social phobics and
nonanxious control participants.
Lundh and Ost (1996a) assessed recognition memory for images of faces
rather than word stimuli. These researchers required socially phobic patients and
nonanxious controls initially to classify faces according to whether they would
anticipate a high- or low-quality interaction with the person shown. Subsequently, both groups of participants demonstrated an equivalent ability to recognize the faces, regardless of the classication imposed. However, in a second
study, Lundh and Ost (1996b) had participants rate the faces as either critical or
accepting. In a later memory task, the social phobics, unlike the control participants, recognized a disproportionate number of those faces they earlier had rated

168

as critical. Although this may suggest that social phobics display a recognition
memory advantage for emotionally negative faces, there are other possible interpretations. Because the experimenters did not directly manipulate the emotional valence of the faces, their ndings may reect anxiety-linked dierences
in rating task performance, rather than in recognition memory performance. For
example, social phobics may have been disproportionately inclined to judge the
more perceptually distinctive faces as critical. Even if all participants then nd
these distinctive faces easier to subsequently recognize, to the same degree, this
rating bias would lead to the social phobics recognizing a disproportionate number of the faces they deemed critical, though to construe this dierence as a group
dierence in selective memory would be misleading. This problem was circumvented by Mansell, Clark, Ehlers, and Chen (1999), who presented social phobics
and nonanxious individuals with equivalent arrays of diering faces, some displaying neutral, some happy, and some negative emotional expressions (anger,
disgust, fear, or sadness). Signal detection analysis revealed no subsequent group
dierences in relative recognition memory sensitivity for faces with diering
emotional expressions. Thus, to conclude that social phobics display a recognition memory advantage for negative faces would be premature.
Similarly, measures of recognition memory for neutral and negative words
do not reveal dierences between nonanxious controls and participants suering from OCD, PTSD, or phobic states. Using intentional memory instructions,
Wilhelm et al. (1996) found that OCD patients and controls display an equivalent pattern of recognition memory for neutral, positive, and negative stimulus
words. Even when textual stimuli have been employed, tailored to describe events
closely related to typical OCD concerns, these anxiety patients do not demonstrate
enhanced recognition memory for such emotionally negative information. Foa,
Amir, Gershuny, Molnar, and Kozak (1997) presented OCD patients and
nonanxious individuals with neutral sentences and others related to contamination concerns (they soiled their clothes with blood). Both groups subsequently displayed an equivalent pattern of recognition memory, with all
participants recognizing the neutral sentences better than the contamination
sentences. Vrana et al. (1995) exposed Vietnam veterans with and without PTSD
to neutral, negative, and trauma-related words within an emotional Stroop task.
Subsequent recognition memory was better for the negative than for the neutral words, but the observed pattern did not dier between the anxiety patients
and the control participants. Even when anxious participants are characterized
by specic fears, they still do not generally show enhanced recognition memory
for stimuli associated with these fears. Watts, Trezise, and Sharrock (1986) had
spider-phobics and nonanxious controls look at a variety of dead spiders and then
tested their capacity to subsequently recognize them. The phobics showed no
overall superiority in recognition memory performance. Indeed, for the larger
spider, which presumably represented the more highly feared stimuli, the phobics
actually demonstrated poorer recognition memory than control participants did.

169

We concluded earlier that good evidence indicates that panic disorder patients
display relatively enhanced recall of emotionally negative material, yet such
patients appear to show no parallel recognition memory advantage for negative
information. Ehlers, Margraf, Davies, and Roth (1988) gave panic disorder patients and control participants a recognition memory test for words that had been
presented within an earlier emotional Stroop task. Signal detection analysis revealed that threat words were recognized more accurately than neutral words,
but this eect was equivalent across both patient and control groups. Beck,
Stanley, Averill, Baldwin, and Deagle (1992) assessed recognition memory, in
panic disorder patients and control participants, for previously exposed neutral
and emotional words. Overall, they found greater recognition memory sensitivity for socially threatening words than for neutral words, but once again this
pattern did not distinguish the two groups.
Across two experiments, Lundh, Thulin, Czyzykow, and Ost (1998) assessed
recognition memory in panic disorder patients and nonanxious controls, using
the face recognition methodology previously employed by Lundh and Ost
(1996b) to test social phobics. There were no group dierences in relative recognition memory for faces participants previously had classied as accepting or
critical or for faces they had classied as high or low in degree of desired contact. Only when participants initially had classied faces according to whether
they would consider the displayed person to be safe or unsafe were dierent patterns of subsequent recognition memory observed between the patients and the
controls. However, the nature of this eect was that the panic disorder patients
displayed disproportionately poor recognition memory for those faces that they
previously rated unsafe. Clearly, therefore, these experiments provide no support
for the proposal that recognition memory for negative information is enhanced
in panic disorder patients.
Only one study has found evidence to support the existence of such a bias,
and it employed a rather unusual measure of recognition memory. Cloitre and
Liebowitz (1991) rst exposed neutral, positive, or threatening words within a
task that required panic disorder patients and control participants to judge
whether the item described a feeling state or simply whether it was a legitimate
English word. Participants subsequently were given a high-speed recognition
memory task adapted from Jacoby and Dallas (1981). Previously seen and new
words were displayed for only 35 ms each, and participants made a forcedchoice
decision concerning whether the word had been encountered earlier in the experiment. Although such a brief exposure likely restricted awareness of these test
stimuli, Cloitre and Liebowitz nevertheless contend that their task assessed explicit memory, because participants were directed to identify which words they
had previously seen. There was a trend toward greater recognition memory sensitivity for threat words in the panic disorder group alone, but this eect reached
signicance only when performance on negative and positive words was directly
compared, after removing neutral word data from the analysis.

170

Overall, research on recognition memory has provided little evidence that

anxiety is associated with enhanced recognition memory for negative information. Most well-designed experiments have yielded null results. Only the unusual
recognition task Cloitre and Liebowitz (1991) used has oered limited support
for the hypothesis, though the eect was signicant only when an experimental
condition was excluded within a post-hoc subsidiary analysis, so whether the
nding will prove replicable remains uncertain. Nevertheless, the best evidence
for an anxiety-linked recognition memory advantage for negative information
comes from a task that imposed a substantial restriction on awareness. We turn
now to the consideration of experiments designed specically to assess implicit
memory performance. The results of such studies have been very variable, and
although they have yielded somewhat greater evidence for such an anxietylinked bias than explicit memory tasks, the phenomenon nevertheless remains
elusive.

Implicit Memory
Whereas recall and recognition tests require participants to consciously recollect previously encountered information, implicit tests infer memory indirectly,
by examining how prior exposure to stimulus items inuences later task performance. For example, when asked to complete word stems with the rst items that
come to mind, people are disproportionately likely to generate words they have
seen recently (Richardson-Klavehn & Bjork, 1988). Similarly, people are more
likely to accurately identify a briey exposed word if this word has been recently
encountered (Jacoby & Dallas, 1981). Evidence that measures of implicit
memory, yielded by such indirect tests, can usefully be distinguished from the
more common recall and recognition measures of explicit memory has come from
work on amnesic patients, who typically show impaired performance on explicit
memory tasks without demonstrating corresponding decits on implicit memory
tasks (Jacoby & Witherspoon, 1982).
In a study that compared the performance of high and low trait-anxious participants on several implicit tests, Russo, Fox, and Bowles (1999) failed to nd
evidence that anxiety is associated with enhanced implicit memory for threatening words. No group dierences were observed in the impact exerted by previously exposed emotional words (which had been simply read or used within
an imagery task) on either subsequent word fragment completion or on the perceptual identication of threat and nonthreat words. However, though null results of this type have not been uncommon, the pattern of ndings has been
highly inconsistent across studies, and a number of researchers also have presented evidence to support anxiety-linked bias on implicit memory tasks.1
Nugent and Mineka (1994) initially presented high and low trait-anxious
participants with emotional words they were to rate for likeability and then assessed implicit memory for this information using a word stem completion task.

171

As evidence of implicit memory, more of these stems were completed to yield

previously exposed (primed) words than alternative (nonprimed) words. However, the relative magnitude of this eect, across threat and nonthreat words,
did not dier between the groups. In contrast, using this same word stem completion task, Richards and French (1991) did nd a dierent pattern of implicit
memory for threat words and nonthreat words in high and low trait-anxious
individuals, though only when these words initially had been processed in a selfreferential manner during their initial exposure. For high trait-anxious participants, but not for low trait-anxious individuals, the priming measure of implicit
memory was signicantly greater for threat words than for non-threat words.
Eysenck and Byrne (1994) also used a stem word completion task to compare
high and low trait-anxious participants implicit memory for emotional words,
encountered earlier in an encoding task that required either simple reading or
the generation of the words from their denitions. Regardless of encoding task,
Eysenck and Byrne observed that the priming index of implicit memory was
greater for threat than for neutral words in high trait-anxious individuals but
not in either medium for low trait-anxious participants.
Mathews et al. (1989) found evidence that GAD patients also display disproportionately good implicit memory for threatening words. They presented neutral, positive, and threatening words to these anxiety patients, and to nonanxious
controls, within an initial encoding task that required participants to generate
a self-referent image using each stimulus item. Following this, implicit memory
was assessed by measuring the magnitude of the priming eects exerted by these
stimuli on a subsequent word stem completion task. Control participants evidenced signicantly smaller priming eects for the threat words than for the
nonthreatening words, whereas GAD patients showed slightly greater priming
eects for the threat words than for the nonthreatening words. Although this
pattern of ndings is consistent with the presence of an anxiety-linked implicit
memory advantage for threatening information, the eect has not replicated
consistently across other studies of GAD patients, possibly because of variations
in the encoding tasks employed. Mathews, Mogg, Kentish, and Eysenck (1995)
exposed GAD patients and controls to word sets similar to those used by Mathews
et al. (1989), but instead of instructing participants to generate images from these
words, they asked them to count occurrences of the letter e in each word. Under
these encoding conditions, there was no tendency for the GAD and control participants to dier in the relative degrees to which the threatening and nonthreatening words primed a later word stem completion, suggesting no dierential
pattern of implicit memory between the groups. Bradley et al. (1995) also employed fairly supercial encoding tasks, requiring participants to judge personal
usage frequency of words or simply exposing them very briey (14 ms), before
inferring implicit memory for these word from the magnitude of repetition priming eects observed on a subsequent lexical decision task. These researchers, too,
failed to observe any dierence between GAD patients and control participants

172

in relative implicit memory for neutral, positive, and anxiety- or depressionrelevant words. Nevertheless, MacLeod and McLaughlin (1995) did nd evidence
of an implicit memory advantage for threat words in GAD patients, without using
a self-referential encoding task. GAD patients and controls were presented with
threat and nonthreat words in an initial emotional Stroop task and were instructed rst to name the ink color and then to pronounce each word aloud.
Subsequently, these words, together with matched new words, were exposed for
a duration calibrated to permit around 70% accurate identication, and implicit
memory was inferred from participants increased ability to identify the old words
relative to the new words. GAD patients, unlike nonanxious control participants,
showed evidence of greater implicit memory for the threat words than for the
neutral words. MacLeod and McLaughlin concluded that enhanced implicit
memory for negative information is a replicable, even if not entirely reliable,
characteristic of GAD patients.
The results of studies designed to investigate implicit memory bias in social
phobics also have been inconsistent. Using a word stem completion task to assess implicit memory for neutral and social threat words, previously encountered
in a self-referential imagery task, Rapee et al. (1994) found no dierences between the patterns of performance by social phobics and by nonclinical controls.
However, using this same word stem completion procedure to assess implicit
memory for emotionally toned stimuli following a similar encoding task, Lundh
and Ost (1997) obtained more encouraging ndings. Despite the absence of a
signicant interaction of group by word type, hypothesis-driven tests revealed a
trend toward disproportionately strong priming eects for social threat words
in the social phobics, compared to the control participants, consistent with an
implicit memory advantage for such material. Amir, Foa, and Coles (2000) used
a dierent procedure to assess implicit memory, based on Jacoby, Allan, Collins,
and Larwills (1988) nding that background noise is judged to be quieter when
familiar material is embedded within it. Amir et al. rst had social phobics and
nonanxious controls listen to, and repeat aloud, emotionally toned sentences.
Later, the participants heard sentences, some previously presented and some
new, embedded within white noise. Implicit memory was revealed by their tendency to rate the white noise intensity as lower when this embedded sentence
had been exposed earlier. For social phobics, but not for control participants, the
perceived intensity of this white noise was disproportionately low when it accompanied previously presented sentences with threatening emotional content.
This nding was taken as evidence that implicit memory for such threatening
sentences was enhanced in these anxious patients.
An insucient number of studies have examined implicit memory for emotional information in either OCD or PTSD patients to permit clear conclusions.
Only one such experiment has been carried out on each of these populations, and
each used the white noise paradigm to assess implicit memory. Foa et al. (1997)
exposed contamination-fearful OCD patients to contamination-related and neu-

173

tral sentences before presenting these same sentences and matched new sentences embedded in white noise. Perceived noise intensity was lower when the
noise accompanied old sentences, conrming implicit memory for this material.
However, the relative magnitude of this eect across both emotional categories
of sentences was equivalent for OCD patients and nonanxious individuals. In
contrast, when Amir, McNally, and Wiegartz (1996) used the same approach
to assess implicit memory for neutral and for combat-related sentences in Vietnam veterans with and without PTSD, evidence of a group dierence was obtained, though only when white noise volume was high. Veterans with PTSD,
unlike those without, judged subjective noise intensity as lower when it accompanied the old combat-related sentences. Thus, though only a few studies have
addressed the issue, no evidence of an implicit memory advantage for negative
information yet has been found in OCD patients, though some evidence indicates
that such a bias may characterize PTSD patients.
Amir, McNally, Riemann, and Clements (1996) used this same white-noise methodology to assess implicit memory for emotional sentences in panic disorder patients
and control participants. Old and new sentences were embedded in a low level
of white noise, and panic disorder patients demonstrated disproportionate
attenuation of subjective noise intensity when it accompanied old sentences describing panic-related threat scenarios. Although this nding is consistent with
an implicit memory advantage for threatening information in panic disorder, some
studies using other measures of implicit memory to assess this same population of
clinical patients have failed to support this conclusion. Rapee (1994) used the word
stem completion task to measure implicit memory for neutral and threat words,
initially exposed within a word association task, and found no dierence in the
patterns of performance shown by panic disorder patients and nonanxious control participants. Similarly, Lundh et al. (1997) failed to nd any dierence between
panic disorder patients and controls on a word stem completion measure of implicit memory for such emotional words, initially encoded in a self-referential imagery task. In a later experiment, Lundh, Wikstrom, Westerlund, and Ost (1999)
employed a perceptual-identication measure to assess implicit memory for these
emotional words, following their initial exposure in an emotional Stroop task.
Perceptual identication was facilitated by prior exposure, conrming the presence
of implicit memory, but the magnitude of this eect was equivalent across all three
classes of words for both panic disorder patients and control participants. Nevertheless, Amir et al.s claim that implicit memory for threat is enhanced in panic
disorder patients gains some support from a study by Cloitre et al. (1994). In this
experiment, emotional words rst were presented in pairs, and participants rated the
strength of semantic association between items. They later were shown the rst word
from each pair, together with the rst three letters of the second word, and were
instructed to complete these stems to make the rst words that came to mind. A
measure of memory was provided by degree to which the previously exposed words,

174

rather than alternatives, appeared in these completions. Cloitre et al. observed that,
among panic disorder patients, this priming eect was disproportionately great for
threat words and took this result as evidence of an implicit memory advantage for
threatening information in patients with this anxiety disorder.
As shown in this review, the overall pattern of implicit memory ndings is
not yet suciently consistent to sustain rm conclusions. There are many possible reasons for this inconsistency, including the potential impurity of the
memory measures. Though implicit memory tasks are designed to be sensitive
to the inuence of implicit memory, performance on most implicit memory tasks
also can be inuenced by explicit memory (Perruchet & Baveux, 1989), and the
degree to which observed eects reect the impact of each type of memory may
depend on subtle aspects of the experimental procedure. For example, the fact
that word stems in Cloitre et al.s (1994) study were accompanied by another
word, which earlier had been paired with an item that could complete this stem,
might encourage the use of explicit memory to identify this candidate completion. Thus, one might plausibly contend that Cloitre et al.s nding could reect
the inuence of enhanced explicit memory, rather than enhanced implicit
memory, for threatening information in panic disorder. However, with the exception of panic disorder, evidence of an anxiety-linked explicit memory advantage for threat is conspicuously lacking. Therefore, it would seem unreasonable
to argue that the more common appearance of such an anxiety-linked bias on
implicit memory tasks should be attributed to the contaminating inuence of
explicit memory. Indeed, contamination from explicit memory might reduce the
capacity of these tasks to detect an anxiety-linked bias in implicit memory. If so,
perhaps task renements that increase the purity of the resulting implicit
memory measures may result in more consistent evidence of such a bias. For the
moment, however, we must concede that, whereas a greater proportion of implicit memory tasks than explicit memory have supported the idea that memory
for threatening information may be enhanced in anxious individuals, neither
approach has provided suciently reliable evidence to conrm the validity of this
hypothesis.

Theoretical Implications
The most obvious conclusion we draw from this review is that, from the wide
range of studies that now have examined memory performance in high traitanxious groups and in patients suering from most anxiety disorders, not enough
evidence has been obtained to support the existence of a memory bias that
consistently favors the retrieval of threatening information. Despite the wellestablished fact that both clinically anxious patients and high trait-anxious members of the normal population selectively direct attentional resources toward such

175

information during encoding (cf. Mathews & MacLeod, 1994), they do not reliably display an enhanced ability to subsequently access this type of material from
memory. Such ndings contrast markedly with those observed in depressed participants, who do indeed commonly display disproportionately good recall of
negative self-related information (e.g., Bradley et al., 1995). [See chapter 6
Eds.] Within the anxiety disorders, such a pattern of selective recall appears only
in panic disorder. For this condition alone, the majority of experimental studies
have supported the existence of an explicit recall bias favoring the retrieval of
threatening or panic-related information.

Why Is There So Little Evidence

of a General Anxiety-Related

Memory Bias?

Elsewhere, we have suggested two possible reasons why anxiety-related memory

advantages for threatening information typically are not observed (e.g., Mathews
& MacLeod, 1994). First, although anxious individuals selectively attend to threat
cues during encoding, this information may not be processed in a manner that
produces representations that can readily be accessed during memory search.
The success of such a search process likely depends highly on the degree to which
stimulus information has undergone elaborative processing, thereby establishing
associations between the new to-be-remembered information and prior knowledge
(Graf & Mandler, 1984). If the aversive nature of threatening stimuli motivates
highly anxious individuals to rapidly detect them, but also to limit their further
processing, such elaboration will be restricted, countering any subsequent retrieval
advantage that otherwise would result from increased attention (Eysenck & Mogg,
1992; Williams, Watts, MacLeod, & Mathews, 1997). Although many studies
have instructed participants to process stimulus information during the encoding
task, in ways that might counter such avoidance (such as relating word meanings to the self), highly anxious individuals may fail to comply fully with such
instructions, perhaps to avoid a sustained aversive experience.
There is little direct evidence to support the idea that anxious individuals avoid
elaboratively processing threatening stimuli. Indeed, the widely recognized tendency for anxious individualsparticularly those suering from GADto worry
excessively suggests that extensive elaborative processing of threat occurs in
some anxiety disorders, especially in the form of verbal propositions (Borkovec
& Inz, 1990). One could argue that this type of elaboration would aid the recall
of such information. However, elaborative processing in the form of worry may
exert no benecial impact on recall performance, particularly if the worry involves stereotypical themes. Instead, if dierent threat cues elicit similar worry
content in anxious individuals, the distinctiveness of the representations resulting from this type of elaboration could be reduced, compromising retrieval of the
original cues.

176

The robust tendency to display relatively facilitated recall of negative information in depression often has been attributed to processing that establishes
novel linkages with other self-related negative information, already represented
within depressives memory. Depressive individuals typically demonstrate such
a negative memory bias only for words encoded in tasks that encourage selfreferential processing (Bradley & Mathews, 1983). If depression alone is associated with particular negative ideation about the self, the apparent dierence in
the patterns of memory selectivity demonstrated by depressed and by anxious
participants might be explained. For depressed individuals, the elaborative processing of new negative information links it to existing negative self-schema,
which then functions as a retrieval aid during memory search. In contrast, although anxious individuals worry about future events, there is little evidence
to suggest that they habitually ruminate about their own negative qualities (as
do depressives), so there is no compelling reason to expect anxious individuals
to display a negative recall bias following such commonly used self-referential
encoding procedures (e.g., Mogg et al., 1989).
Another reason that could explain the lack of an anxiety-linked recall bias is
that anxious individuals may tend to represent threatening information perceptually rather than in a conceptual form based on verbal propositions. This contention is consistent with the supposition that threat can be processed within a
primitive fear system (or module) that evolved prior to the development of language (hman & Mineka, 2001). If anxious participants are particularly inclined
to represent threat using perceptual codes, perhaps because this primitive fear
system is overly active, then according to the transfer-appropriate processing
accounts of memory (Roediger & Blaxton, 1987), retrieving this information
when performing memory tasks that encourage conceptual processing, such as
free recall, would prove dicult for them. In contrast, memory tasks that encourage perceptual processing should more eectively tap anxious participants representations of previously encountered threat stimuli. This may explain why
conventional implicit memory tasks, such as rapid word identication, which
characteristically encourage perceptual rather than conceptual processing, have
more often revealed evidence of an anxiety-related memory advantage for threat.
According to this account, the widespread tendency to use self-referential or
other semantic encoding procedures may reduce the likelihood of nding such
biases on this type of implicit memory task. Certainly, as we have shown, the
evidence in support of an anxiety-linked implicit memory advantage for threat
is very mixed, but it may not be fully explained by variations in encoding tasks.
Several studies that have used structural encoding methods have failed to reveal
any implicit memory advantage for threat in anxious participants, though others have yielded more supportive ndings (e.g., MacLeod & McLaughlin, 1995).
Perhaps the prevailing use of words as stimuli also may undermine the likelihood of nding enhanced perceptual memory for threat in anxious individuals,
as the perceptual encoding of such stimuli seems less likely than would be ex-

177

pected with pictorial materials. Though such a proposal is speculative at this

stage, both the restricted elaboration hypothesis and the perceptual encoding
hypothesis remain viable explanations for the absence of any general memory
advantage for threat stimuli in anxious individuals, despite their tendency to
selectively allocate attention to such information. Neither account, however, has
been directly supported by experimental evidence; indeed, few experiments have
been designed to test these hypothetical accounts.

Why Is There a Memory Bias

in Panic Disorder?

Despite the sparse evidence that anxiety is associated with a memory bias favoring threat, panic disorder patients tend to recall more panic-related threat words
than nonanxious controls do. We cannot attribute this eect to anxiety per se,
given the absence of parallel eects in other anxiety disorders, so it seems necessary to explain why panic disorder is an exception to the rule. One obvious possibility, already noted in our comments on the study by Nunn et al., (1984), is
that the supposedly negative words in these studies often may be threatening
only for the panic patients. For example, words describing common cues for the
panic attacks experienced by such patients (e.g., street, crowd) frequently have
been used in these experiments, even though such words are unlikely to be emotionally charged for nonanxious control participants. As we reported in the introduction, emotional information is more likely to be recalled than neutral
information, regardless of anxiety level, and the nding that panic patients recall these words disproportionately well may reect only this general eect. The
confound between clinical status and stimulus valence makes it dicult to determine whether the panic disorder patients and the controls would have recalled
dierently stimulus materials that both groups had considered threatening.
This problem is less evident within studies that have used stimulus materials
likely to be emotionally threatening for all participants. Words such as torture
and death presumably would have similar threat values for all readers, so any
general recall advantage for emotional words would not produce group dierences in recall. Nevertheless, as we have pointed out, likely group dierences in
the perceived threat value of stimulus materials have compromised at least one
study of OCD patients (Radomsky & Rachman, 1999). In this study, discrepant
patterns of memory performance arose perhaps because a contamination manipulation increased stimulus threat value for patients, but not for nonanxious
controls. The more common nding is that OCD patients do not dier from
nonanxious controls in memory for threatening stimuli, even when these are
related to their concerns, although they do often report less condence in their
memory (e.g., Tolin et al., 2001). Thus, the anomalous memory advantage for
threat-related words shown by panic disorder patients may reect the fact that
panic-related word cues possess more emotionally negative meaning for panic

178

patients than for nonanxious controls. The use of stimulus words that have special emotional signicance only for one group of participants prevents the attribution of observed group dierences in the recall of such material to a tendency
to selectively remember emotional information.

Are Discrepancies in Memory Eects

Due to Diering Encoding Tasks?

We have suggested that some of the memory eects shown by anxious participants can readily be attributed to anxiety-linked patterns of selective encoding,
without the need to implicate a bias in the retrieval process itself. For example,
in the experiments carried out by Friedman et al. (2000) and by Russo et al.
(2001), words were exposed for durations that far exceeded the minimum time
required to register their identity. Anxious individuals characteristically direct
attention toward, and nd it disproportionately dicult to disengage attention
from, emotionally threatening words during such encoding procedures. Consequently, their enhanced subsequent ability to recall the threatening words can
be attributed to the superior encoding of the threatening information, resulting
from this reduced ability to inhibit attention to such aective meanings (Williams et al., 1996; Wood et al., 2001). Further evidence to support this account
comes from a series of experiments reported by Pury and Mineka (2001). In a
typical study, participants were required to judge, as quickly as possible, whether
a presented word (e.g., bleeding, condence) described something dangerous or
safe or whether it described a physical or psychological state. When participants
were not informed which decision was required until immediately prior to word
presentation, highly anxious participants were slower than low anxious controls
to make the physical versus psychological judgment but not to make the dangerous versus safe judgment. This nding, together with the results of other
experiments from this same series, provides evidence that high anxious individuals are more likely than their low anxious counterparts to encode stimuli in terms
of their aective meaning.
If anxious individuals infrequently observed threat-related bias in recall results
from their heightened inclination to encode the emotionally threatening content
of aective stimulus words, then this recall bias should be particularly likely following encoding tasks that permit ample opportunities for participants to select
how to process these stimuli. Under such encoding conditions, we suppose that
increased attention paid to threatening aspects of meaning may help high anxious individuals to retain such words better. In contrastagain assuming that the
enhanced memory for threat occasionally demonstrated by anxious participants
reects an anxiety-linked tendency to process the aective meaning of such stimuli
during encodingthis eect should not be observed following encoding tasks that
require all participants to process stimuli in terms of their aective meaning. In
general, as this review has indicated, anxious individuals more often show en-

179

hanced relative recall of threatening information following encoding tasks that

permit but do not require the encoding of aective meaning.
Technically, of course, the term memory bias is appropriate to describe the
recall advantage for threat sometimes shown by anxious participants following
free encoding conditions that employ long exposure durations (e.g., Friedman
et al., 2000; Russo et al., 2001). However, we contend that it is not parsimonious to ascribe the eect to a bias in the process of retrieving information from
memory, when it can be explained by the well-established tendency for highly
anxious individuals to attend to and selectively encode threatening meanings.
We believe that the eect may better be construed as an indirect consequence of
the bias that favors the selective encoding of threatening stimulus content in
anxiety.
Nevertheless, the possibility that the patterns of selective attention and encoding associated with vulnerability to anxiety may, under certain circumstances, give rise to long-term memory eects is clearly important. These patterns
of encoding selectivity may aect not only how old events are recalled but also
how new events are understood. For example, our work on the experimental
manipulation of such biases has shown that people trained to attend selectively
to threatening meanings of stimulus information become more likely to store
negatively disambiguated representations of newly presented ambiguous events
in memory (Mathews & MacLeod, 2002). Distinctions between attention, interpretation, and memory thus can become blurred when life events are processed,
given that initial encoding biases so readily can translate into dierential patterns of interpretation and recall.
In closing, however, we emphasize that, despite this potential capacity for
encoding biases to inuence memory performance, most experimental studies
have failed to demonstrate that anxious individuals (with the possible exception
of panic disorder patients) consistently display enhanced recall or recognition
memory for emotionally threatening information. Though an implicit memory
bias favoring threat stimuli sometimes has been observed in anxious participants,
at best it is a fragile eect that often has proven dicult to replicate. It remains
to be seen whether the future renement of experimental methodologies will
identify particular conditions under which anxious participants will more reliably demonstrate a memory advantage for negative information. For the moment, the most appropriate conclusion is that such a memory bias does not
represent a robust general characteristic of anxiety.

Note
1. Whereas methodological rigor has varied across these studies, and it sometimes is possible to criticize experimental decisions such as choice of appropriate
baselines, the pattern of observed inconsistences cannot readily be explained in
terms of such design variations.

180

References
Amir, N., Coles, M. E., Brigidi, B., & Foa, E. B. (2001). The eect of practice on recall
of emotional information in individuals with generalized social phobia. Journal of Abnormal Psychology, 110, 7682.
Amir, N., Foa, E.B., & Coles, M. (2000). Implicit memory bias for threat-relevant
information in individuals with generalized social phobia. Journal of Abnormal
Psychology, 109, 713720.
Amir, N., McNally, R. J., Riemann, B. C., & Clements, C. (1996). Implicit memory
bias for threat in panic disorder: Application of the white noise paradigm.
Behaviour Research and Therapy, 34, 157162.
Amir, N., McNally, R. J., & Wiegartz, P. S. (1996b). Implicit memory bias for threat
in posttraumatic stress disorder. Cognitive Therapy and Research, 20, 625635.
Ashcraft, M. H., & Kirk, E. P. (2001). The relationship among working memory,
math anxiety, and performance. Journal of Experimental Psychology: General,
130, 224237.
Baddeley, A. D. (1986). Working memory. Oxford: Oxford University Press.
Beck, J. G., Stanley, M. A., Averill, P. M., Baldwin, L. E., & Deagle, E. A. (1992). Attention and memory for threat in panic disorder. Behaviour Research and
Therapy, 30, 619629.
Becker, E., Rinck, M., & Margraf, J. (1994). Memory bias in panic disorder. Journal
of Abnormal Psychology, 103, 396399.
Becker, E. S., Roth, W. T., Andrich, M., & Margraf, J. (1999). Explicit memory in
anxiety disorders. Journal of Abnormal Psychology, 108, 153163.
Borkovec, T. B., & Inz, J. (1990). The nature of worry in generalized anxiety disorder: A predominance of thought activity. Behaviour Research and Therapy, 28,
153158.
Bradley, B. P., & Mathews, A. (1983). Negative self-schemata in clinical depression.
British Journal of Clinical Psychology, 22, 173181.
Bradley, B. P., Mogg, K., & Williams, R. (1994). Implicit and explicit memory for
emotional information in non-clinical subjects. Behaviour Research and Therapy,
32, 6578.
Bradley, B. P., Mogg, K., & Williams, R. (1995). Implicit and explicit memory for
emotion-congruent information in clinical depression and anxiety. Behaviour
Research and Therapy, 33, 755770.
Burke, M., & Mathews, A. (1992). Autobiographical memory and clinical anxiety.
Cognition & Emotion, 6, 2335.
Cahill, L., & McGaugh, J. L. (1995). A novel demonstration of enhanced memory
associated with emotional arousal. Consciousness and Cognition, 4, 410421.
Calvin, A. D., Koons, P. B., Bingam, J. L., & Fink, H. H. (1953). A further investigation of the relationship between manifest anxiety and intelligence. Journal of
Consulting Psychology, 19, 280282.
Christianson, S-. (1992). Emotional stress and eyewitness memory: A critical review. Psychological Bulletin, 112, 284309.
Cloitre, M., Cancienne, J., Heimberg, R. G., Holt, C. S., & Liebowitz, M. R. (1995).
Case histories and shorter communications: Memory bias does not generalize
across anxiety disorders. Behaviour Research and Therapy, 33, 305307.

181

Cloitre, M., & Liebowitz, M. R. (1991). Memory bias in panic disorder: An investigation of the cognitive avoidance hypothesis. Cognitive Therapy and Research,
15, 371386.
Cloitre, M., Shear, M. K., Cancienne, J., & Zeitlin, S. B. (1994). Implicit and explicit
memory for catastrophic associations to bodily sensation words in panic disorder. Cognitive Therapy and Research, 18, 225240.
Dalgleish, T. (1994). The relationship between anxiety and memory biases for material that has been selectively processed in a prior task. Behaviour Research and
Therapy, 32, 227231.
Darke, S. (1988a). Anxiety and working memory capacity. Cognition & Emotion, 2,
145154.
Darke, S. (1988b). Eects of anxiety on inferential reasoning task performance. Journal of Personality and Social Psychology, 55, 499505.
Ehlers, A., Margraf, J., Davies, S., & Roth, W. T. (1988). Selective processing of threat
cues in subjects with panic attacks. Cognition & Emotion, 2, 201220.
Eysenck, M. W. (1982). Attention and arousal: Cognition and performance. Berlin:
Springer.
Eysenck, M. W., & Byrne, A. (1994). Implicit memory bias, explicit memory bias,
and anxiety. Cognition & Emotion, 8, 415431.
Eysenck, M. W., & Calvo, M. G. (1992). Anxiety and performance: The processing
eciency theory. Cognition & Emotion, 6, 409434.
Eysenck, M. W., & Mogg, K. (1992). Clinical anxiety, trait anxiety, and memory bias.
In S-. Christianson (Ed.), Handbook of emotion and memory (pp. 429450).
Hillsdale, NJ: Erlbaum.
Finlay-Jones, R., & Brown, G. W. (1981). Types of stressful life event and the onset
of anxiety and depressive disorders. Psychological Medicine, 11, 803815.
Foa, E. B., Amir, N., Gershuny, B., Molnar, C., & Kozak, M. J. (1997). Implicit and
explicit memory in obsessive-compulsive disorder. Journal of Anxiety Disorders,
11, 119129.
Fox, E., Russo, R., Bowles, R., & Dutton, K. (2001). Do threatening stimuli draw or
hold visual attention in subclinical anxiety? Journal of Experimental Psychology:
General, 130, 681700.
Friedman, B. H., Thayer, J. F., & Borkovec, T. D. (2000). Explicit memory bias for
threat words in generalized anxiety disorder. Behavior Therapy, 31, 745756.
Graf, P., & Mandler, G. (1984). Activation makes words more accessible but not
necessarily more retrievable. Journal of Verbal Learning and Verbal Behaviour,
23, 553568.
Humphreys, M. S., & Revelle, W. (1984). Personality, motivation, and performance:
A theory of the relationship between individual dierences and information
processing. Psychological Review, 91, 153184.
Jacoby, L. L., Allan, L. G., Collins, J. C., & Larwill, L. K. (1988). Memory inuences
subjective experience: Noise judgments. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 14, 240247.
Jacoby, L. L., & Dallas, M. (1981). On the relationship between autobiographical
memory and perceptual learning. Journal of Experimental Psychology: General,
110, 303340.

182

Jacoby, L. L., & Witherspoon, D. (1982). Remembering without awareness. Canadian Journal of Psychology, 36, 300324.
Levy, E. A., & Mineka, S. (1998). Anxiety and mood-congruent autobiographical
memory: A conceptual failure to replicate. Cognition & Emotion, 12, 625634.
Lundh, L.-G., Czyzykow, S., & Ost, L.-G. (1997). Explicit and implicit memory bias
in panic disorder with agoraphobia. Behaviour Research and Therapy, 35, 1003
1014.
Lundh, L.-G., & Ost, L.-G. (1996a). Face recognition in patients with social phobia.
Scandinavian Journal of Behaviour Therapy, 25, 139148.
Lundh, L.-G., & Ost, L.-G. (1996b). Recognition bias for critical faces in social
phobics. Behaviour Research and Therapy, 34, 787794.
Lundh, L.-G., & Ost, L.-G. (1997). Explicit and implicit memory bias in social phobia.
The role of subdiagnostic type. Behaviour Research and Therapy, 35, 305317.
Lundh, L.-G., Thulin, U., Czyzykow, S., & Ost, L.-G. (1998). Recognition bias for safe
faces in panic disorder with agoraphobia. Behaviour Research and Therapy, 36,
323337.
Lundh, L.-G., Wikstrom, J., Westerlund, J., & Ost, L.-G. (1999). Preattentive bias for
emotional information in panic disorder with agoraphobia. Journal of Abnormal Psychology, 108, 222232.
MacLeod, C., & Donellan, A. M. (1993). Individual dierences in anxiety and the
restriction of working memory capacity. Personality and Individual Dierences,
15, 163173.
MacLeod, C., & McLaughlin, K. (1995). Implicit and explicit memory bias in anxiety: A conceptual replication. Behaviour Research and Therapy, 33, 114.
Mansell, W., Clark, D. M., Ehlers, A., & Chen, Y.-P. (1999). Social anxiety and attention away from emotional faces. Cognition & Emotion, 13, 673690.
Mathews, A., & MacLeod, C. (1985). Selective processing of threat cues in anxiety
states. Behaviour Research and Therapy, 23, 563569.
Mathews, A., & MacLeod, C. (1994). Cognitive approaches to emotion and emotional
disorders. Annual Review of Psychology, 45, 2550.
Mathews A., & MacLeod, C. (2002). Induced processing biases have causal eects
on anxiety. Cognition & Emotion, 16, 310315.
Mathews, A., Mogg, K., Kentish, J., & Eysenck, M. (1995). Eect of psychological
treatment on cognitive bias in generalized anxiety disorder. Behaviour Research
and Therapy, 33, 293303.
Mathews, A., Mogg, K., May, J., & Eysenck, M. (1989). Implicit and explicit memory
bias in anxiety. Journal of Abnormal Psychology, 98, 236240.
McNally, R. J., Foa, E. B., & Donnell, C. D. (1989). Memory bias for anxiety information in patients with panic disorder. Cognition & Emotion, 3, 2744.
McNally, R. J., Litz, B. T., Prassas, A., Shin, L. M., & Weathers, F. W. (1994). Emotional priming of autobiographical memory in post-traumatic stress disorder.
Cognition & Emotion, 8, 351367.
McNally, R. J., Metzger, L. J., Lasko, N. B., Clancy, S. A., & Pitman, R. K. (1998).
Directed forgetting of trauma cues in adult survivors of childhood sexual abuse
with and without posttraumatic stress disorder. Journal of Abnormal Psychology, 107, 596601.

183

Mogg, K., & Mathews, A. (1990). Is there a self-referent mood-congruent recall bias
in anxiety? Behaviour Research and Therapy, 28, 9192.
Mogg, K., Mathews, A., & Weinman, J. (1987). Memory bias in clinical anxiety. Journal of Abnormal Psychology, 96, 9498.
Mogg, K., Mathews, A., & Weinman, J. (1989). Selective processing of threat cues
in anxiety states: A replication. Behaviour Research and Therapy, 27, 317323.
Mueller, J. H. (1992). Anxiety and performance. In A. P. Smith & D. M. Jones (Eds.),
Factors aecting human performance (Vol. 3, pp. 127160). London: Academic
Press.
Nugent, K., & Mineka, S. (1994). The eect of high and low trait anxiety on implicit
and explicit memory tasks. Cognition & Emotion, 8, 147163.
Nunn, J. D., Stevenson, R., & Whalan, G. (1984). Selective memory eects in agoraphobic patients. British Journal of Clinical Psychology, 23, 195201.
hman, A., & Mineka, S. (2001). Fears, phobias, and preparedness: Toward an
evolved module of fear and fear learning. Psychological Review, 108, 483522.
Otto, M. W., McNally, R. J., Pollack, M. H., Chen, E., & Rosenbaum, J. F. (1994).
Hemispheric laterality and memory bias for threat in anxiety disorder. Journal
of Abnormal Psychology, 103, 828831.
Perruchet, P., & Baveux, P. (1989). Correlational analyses of explicit and implicit
memory performance. Memory & Cognition, 17, 7786.
Pickles, A. J., & van den Broek, M. D. (1988). Failure to replicate evidence for phobic schemata in agoraphobic patients. British Journal of Clinical Psychology, 27,
271272.
Pury, C. L. S., & Mineka, S. (2001). Dierential encoding of aective and nonaective
content information in trait anxiety. Cognition & Emotion, 15, 659693.
Radomsky, A. S., & Rachman, S. (1999). Memory bias in obsessive-compulsive disorder (OCD). Behaviour Research and Therapy, 37, 605618.
Rapee, R. M. (1994). Failure to replicate a memory bias in panic disorder. Journal of
Anxiety Disorders, 8, 291300.
Rapee, R. M., McCallum, S. L., Melville, L. F., Ravenscroft, H., & Rodney, J. M. (1994).
Memory bias in social phobia. Behaviour Research and Therapy, 32, 8999.
Reidy, J., & Richards, A. (1997). Anxiety and memory: A recall bias for threatening
words in high anxiety. Behaviour Research and Therapy, 35, 531542.
Reisberg, D., & Heuer, F. (1992). Remembering the details of emotional scenes. In
E. Wingrad & U. Neisser (Eds.), Aect and accuracy in recall: Studies of ashbulb
memory (pp. 162190). New York: Cambridge University Press.
Revelle, W., & Loftus, D. A. (1990). Individual dierences and arousal: Implications
for the study of mood and memory. Cognition & Emotion, 3, 209237.
Richards, A., & French, C. C. (1991). Eects of encoding and anxiety on implicit and
explicit memory performance. Personality and Individual Dierences, 12, 131139.
Richards, A., & Whittaker, T. M. (1990). Eects of anxiety and mood manipulation in
autobiographical memory. British Journal of Clinical Psychology, 29, 145153.
Richardson-Klavehn, A., & Bjork, R. A. (1988). Measures of memory. In M. R.
Rosenzweig & L. W. Porter (Eds.), Annual Review of Psychology, 9, 475543.
Roediger, H. L., & Blaxton, T. A. (1987). Retrieval modes produce dissociations in
memory for surface information. In D. Gorfein & R. R. Homan (Eds.), Memory

184

and cognitive processes: The Ebbinghaus Centennial Conference. Hillsdale, NJ:

Erlbaum.
Russo, R., Fox, E., Bellinger, L., & Nguyen-Van-Tam, D. P. (2001). Mood-congruent free recall bias in anxiety. Cognition & Emotion, 15, 419433.
Russo, R., Fox, E., & Bowles, R. J. (1999). On the status of implicit memory bias in
anxiety. Cognition & Emotion, 13, 435456.
Sanz, J. (1996). Memory biases in social anxiety and depression. Cognition & Emotion, 10, 87105.
Tolin, D. F., Abramowitz, J. S., Brigidi, B. D., Amir, N., Street, G. P., & Foa, E. B.
(2001). Memory and memory condence in obsessive-compulsive disorder.
Behaviour Research and Therapy, 39, 913927.
Vrana, S. R., Roodman, A., & Beckham, J. C. (1995). Selective processing of traumarelevant words in posttraumatic stress disorder. Journal of Anxiety Disorders, 9,
515530.
Watts, F. N., Trezise, L., & Sharrock, R. (1986). Processing of phobic stimuli. British Journal of Clinical Psychology, 25, 253261.
Wenzel, A., Jackson, L. C., & Holt, C. S. (2002). Social phobia and the recall of autobiographical memories. Depression & Anxiety, 15, 186189.
Wessel, I., van der Kooy, P., & Merckelbach, H. (2000). Dierential recall of central
and peripheral details of emotional slides is not a stable phenomenon. Memory,
8, 95109.
Wilhelm, S., McNally, R. J., Baer, L., & Florin, I. (1996). Directed forgetting in obsessive-compulsive disorder. Behaviour Research and Therapy, 34, 633641.
Wilhelm, S., McNally, R. J., Baer, L., & Florin, I. (1997). Autobiographical memory
in obsessive-compulsive disorder. British Journal of Clinical Psychology, 36, 21
31.
Williams, J. M. G., Mathews, A., & MacLeod, C. (1996). The emotional Stroop task
and psychopathology. Psychological Bulletin, 120, 324.
Williams, J. M. G., Watts, F. N., MacLeod, C., & Mathews, A. (1997). Cognitive psychology and emotional disorders. 2nd ed. Chichester, UK: Wiley.
Wood, J., Mathews, A., & Dalgleish, T. (2001). Anxiety and cognitive inhibition.
Emotion, 1, 166181.
Yiend, J., & Mathews, A. (2001). Anxiety and attention to threatening pictures.
Quarterly Journal of Experimental Psychology, 54, 665681.

185

A Question of Habit?

he truest claim that cognitive science can make might also be the
least sophisticated: the mind tends to do what it has done before.
In previous centuries philosophers and psychologists invented constructs such
as associations, habit strength, and connectivity to formalize the truism, but
others have known about it, too. In small towns in the Ozarks, for example,
grandmothers have been overheard doling out warnings such as, Dont think
those ugly thoughts; your mind will freeze that way. Depressed persons, like
most of us, usually dont heed this advice. The thoughts frozen in their minds
might not be ugly, but they often reect disappointments, losses, failures, other
unhappy events, and a generally negative interpretive stance toward ongoing
experience. By considering these habits of thinking, we should better understand
the nature of memory in depressed states. Deliberate attempts to remember are
either impaired or facilitated in ways that appear related to habits of thinking.
Even more commonly, memory is expressed indirectly through the content of
current thoughts and interpretations.
This chapter examines the relations among habitual thoughtsoften called
ruminationsand memory phenomena in depression. I use depression imprecisely, to refer to both diagnostic categories and the self-reported state of dysphoria. Dysphoria is used to denote undiagnosed negative aect, of the sort that
produces moderate-to-high scores on the Beck Depression Inventory (BDI: Beck,
Ward, Mendelson, Mock, & Erbaugh, 1961). Although the patterns of performance associated with self-reported measures of depressed moodsuch as BDI
scoresare often similar to those obtained from diagnosed samples, occasional
evidence for dierences should be noted (see Burt, Zembar, & Niederehe, 1995).
We should note the imprecision associated with diagnosis as well (e.g., the
boundary problems for mood disorders and generalized anxiety disorders; see

186

186

Brown, Di Nardo, Lehman, & Campbell, 2001). On the side of simplicity, however, many phenomena appear similar in form, if not extent, across studies measuring dysphoria and depression. (For a short review, see Wenzla, Meier, &
Salas, 2002.)
Further caution about the use of depression is appropriate when considering the many studies performed with experimental mood inductions (e.g.,
through the use of sad or happy music) in an eort to make causal statements
about mood and memory. In some of these studies, the language of depression
is used to describe the ndings from negatively valenced inductions. (See Parrott
& Hertel, 2000, for a description of these and other methodological issues.) Moodinduction studies sometimes reveal patterns similar to those in studies based on
self-reports or diagnoses. But because habits of ruminating do not likely characterize students randomly assigned to listen to sad music, this chapter rarely addresses ndings from mood inductions.

Habits of Thought
Negative thinking has been such a prevalent feature of depression, as observed
clinically, that Beck (1967) used schema to capture its habitual and interrelated qualities. These stable cognitive structures were proposed as one way of
organizing and describing thought patterns in depression. Rumination is used
as a process-oriented companion to the structural construct of a schema. Depressed people tend to ruminate. They ponder the episodes associated with sad
feelings and imagine similar future occurrences. They focus on their feelings and
wonder if they will ever change (see Ingram, 1990; Nolen-Hoeksema, 1991).
Rumination can be an intensely attention-demanding process, yet the initiation of a ruminative episode is often thoughtless or automatic, even to the extent that the person can become so engaged without awareness. The draw
toward rumination has a lot in common with the tendency to attend to external
stimuli that are emotionally congruent with depressive concepts. For example,
dysphoric students are slower to name the color of ink when it spells a word related to depression (e.g., Gotlib & McCann, 1984). On similar tasks nondepressed
participants preferentially attend to positive members of word pairs and avoid
the negative members, but clinically depressed and dysphoric participants lack
this positive bias (McCabe & Gotlib, 1995; McCabe & Toman, 2000).
Mood-related experience also encourages the tendency to nd negative meaning in ambiguous information, relatively automatically. Lawson, MacLeod, and
Hammond (2002) devised a clever indirect measure of such interpretive biases
by measuring the blink reex to noise occurring as participants imaged situations evoked by auditorially presented ambiguous and nonambiguous words.
Participants who had scored high on the BDI produced particularly amplied
blink reexes during the imaging of ambiguous words that lend themselves to

187

negative interpretations, in a pattern similar to that for clearly negative words.

Using a quite dierent paradigm, Wenzla and Bates (1998) presented strings
of words that could be unscrambled to form either negative or non-negative sentences. Dysphoric students formed negative versions more often than others did.
Even students who were formerly dysphoric displayed this bias under dual-task
conditions. Although direct reports of beliefs and thoughts tend toward less negativity as depression lessens (Haaga, Dyck, & Ernst, 1991), ndings such as those
by Wenzla and Bates (1998) reveal the enduring nature of depressive habits of
thought (also see Wenzla, Rude, Taylor, Stultz, & Sweatt, 2001).
Depressive habits of thought seem, moreover, to exacerbate sad moods and
predict future depressive episodes. In an extensive program of research by NolenHoeksema and her colleagues (e.g., Lyubomirsky, Caldwell, & Nolen-Hoeksema,
1998; Nolen-Hoeksema & Morrow, 1993), depressed and nondepressed participants have been asked to concentrate on either self-focused or distracting
phrases. The self-focused phrases themselves do not suggest sad or depressed
mood; in fact, the nondepressed participants typically report feeling no sadder
after concentrating on them than they feel after entertaining thoughts about
distracting phrases (e.g., geographical locations). But in study after study, the
sad moods of depressed participants have increased following the self-focused
statements and decreased following the distracting phrases. Similarly, Fennell,
Teasdale, Jones, and Damle (1987) reported improvements in mood after depressed participants focused on distracting images of outdoor scenes. Therefore,
habits of thought can aect mood, and recent discoveries have shown that they
can also predict future mood impairment (e.g., Alloy, Abramson, & Francis,
1999; Nolen-Hoeksema, 2000; Rude, Wenzla, Gibbs, Vane, & Whitney, 2002).
Measures of negative-thinking styles, demonstrated either through self-reports
or by performance on laboratory tasks, predicted episodes of dysphoria and depression in weeks to come. So, just as depressed mood might establish habits of
negative thinking, such habits at least portend, and perhaps help establish, not
only temporary changes in mood but also future depressive episodes. Indeed,
substantial empirical evidence now supports the concept of a vicious cycle of
rumination and mood, described years ago by John Teasdale (1983).

Habits of Memory
Separating habits of memory from habits of thinking is a somewhat arbitrary
exercise. Obviously, ruminative episodes often include autobiographical memories that come to mind habitually. Less obviously, perhaps, even the acts of interpreting current experience in habitually negative ways are themselves
instances of using memory implicitly or without intention. Jacoby and Kelley
(1987, citing Polyani, 1958) distinguished the use of memory as an object for
examination from the use of memory as a tool for perception and interpretation,

188

a helpful distinction in the context of considering cognitive habits. Both uses of

memoryobject and toolcan become biased out of habit.
To refer to memory as an object (explicit memory) is to communicate awareness that the content of ones thoughts derives from a past event. Episodes from
ones personal past can come to mind habitually, sometimes uninvited (but with
full awareness of their temporal status) and sometimes deliberately and repeatedly sought. At the end of this section, I examine the well-documented tendency
for depressed persons to remember negative episodes from their biographical past.
Depressed persons habitually remember the bad.
Habits of remembering negative events also show up in the laboratory. On
tests of explicit memory, the tendency to think negatively can facilitate the deliberate remembering of experimental materials from the same conceptual neighborhood. On tests ostensibly unrelated to memory (implicit or indirect tests of
memory), performance can also reect habits of prior preferential processing. The
experimental goal in both cases is to control the initial experience intended to
operate as tool or object on the test (i.e., the experience to be remembered);
however, the autobiographical method obviously lacks such control over the
initial experience. In experimental studies, at least we know the characteristics
of the immediate experience to be remembered. Yet it is important to keep in mind
that the power of any mood-congruent eectexperimental or autobiographicalno doubt lies in uncontrolled, pre-experimental habits of thought.

Memory-as-Object in the Laboratory

If we present positively and negatively aective words in a rst experimental task
and later ask the depressed and nondepressed participants to recall those words,
chances are good that recall will in some way be congruent with mood. (See the
meta-analysis by Matt, Vazquez, & Campbell, 1992.) Maybe the dierences will
be lopsided, especially in the case of dysphoria, with dysphoric participants recalling fewer positive words but the same number of negative words as
nondysphoric participants do. Likely, the dierences will obtain only for words
encountered conceptually in the rst task. In fact, some evidence suggests further restrictions to words considered in relation to the self (e.g., Bradley &
Mathews, 1983; Derry & Kuiper, 1981; Dobson & Shaw, 1987). These limitations support the claim that habits of thought play a role in producing moodcongruent recall.
Habits likely play a large role during the test itself. An illustration is the experiments performed by Murray, Whitehouse, and Alloy (1999). They found
evidence of mood-congruent recall following a self-referential task, but only
when participants were not required to guess in order to produce a criterion
number of words. When recall was forced, the mood-related dierences vanished. These results should encourage researchers to use a forcing procedure
more often. A reasonable hypothesis is that mood-congruent biases inuence the

189

uency or ease of remembering, but that materials with other meaning can be
brought to mind with sucient persistence or external aid. Such an outcome,
however, does not reduce our interest in the ubiquitous ndings of moodcongruent recall and their interpretation. Most acts of recall are probably uencydriven, without prolonged pursuit of additional material.
Several theoretical frameworks have emphasized the greater degree of conceptual processing that depressed persons devote to negative materials. Some
approaches propose that elaborative conceptual processing facilitates recall by
establishing richer and more diverse retrieval routes (see Williams, Watts,
MacLeod, & Mathews, 1988, and the network account by Bower, 1981). According to the framework of transfer-appropriate processing (Morris, Bransford, &
Franks, 1977; see Roediger & McDermott, 1992), mood-congruent recall is established by the match between conceptual elaboration during the initial task
and the conceptual basis of the attempt to remember. This account can be extended to include a third type of occasion: prior habits of thinking. Predated
by such habits, negative self-referential material comes to mind with greater uency, either during the initial episode to be remembered, during the test itself,
or on both occasions.
Probably more often than we realize, however, evidence for mood-congruent recall of conceptually processed material is not obtained. Can a consideration
of cognitive habits suggest boundary conditions for the mood-congruent eect,
at least on a post-hoc basis? Among others, Parrott and Spackman (2000) have
written about the tendency for some people in negative moods to attempt mood
repair by deliberately thinking positive thoughts. Indeed, evidence for moodincongruent recall might result from these attempts, as reconrmed recently by
Rusting and DeHart (2000). They instructed some participants to keep focusing on negative events (imagined vignettes for specied words or autobiographical episodes) used to induce negative mood, much like what one does in
rumination. Other participants were instructed to engage in positive reappraisals of those events. Subsequent recall was mood-congruent for the former
participants and mood-incongruent for the latter. These participants were not
depressed; nevertheless, the results provide a model for what might happen naturally when depressed persons develop either set of cognitive habits.1
The literature on depression and mood-congruent recall is substantial. Many
studies now conducted to examine depression-related biases in other types of
cognitive tasks often include a demonstration of mood-congruent recall for
comparison (e.g., Watkins, Mathews, Williamson, & Fuller, 1992). There are
also published reports of mood-congruent recognition in depression (e.g.,
Wenzla et al., 2002), even one that reports mood-related dierences in eventrelated potentials during the initial task of rating pleasantness, as well as during the recognition test (Deldin, Keller, Gergen, & Miller, 2001). Dierent levels
of brain activity should indeed inform our understanding of mood-congruent
memory, but they do not necessarily reveal the causal mechanisms underly-

190

ing the dispositions for negative interpretations. Habits might serve as instigators, outcomes, or both.

Memory as Tool
In the last decade of the 20th century, a number of experiments on implicit
memory in depression were published (e.g., Bazin, Perruchet, De Bonis, & Fline,
1994; Danion et al., 1991; Denny & Hunt, 1992; Elliott & Greene, 1992; Hertel,
1994; Hertel & Hardin, 1990; Watkins et al., 1992). Some of these experiments
were not designed to address mood-congruence. But others varied the emotional
valence of the materials in attempts to determine whether mood-congruent
memory would be revealed on indirect tests, in situations in which people
presumably are not trying to remember. In the same experiments that produced
mood-congruent recall, there was at rst a notable failure to nd moodcongruent dierences on indirect tests. Some researchers anticipated this outcome, predicting that prior conceptual processing, so advantageous for strategic
retrieval tasks, is unimportant for nonstrategic indirect tests (e.g., Williams
et al., 1988). If mood-congruent memory relies on dierences during initial conceptual processing, mood congruencelike other conceptual manipulations, in
this line of reasoningshould not characterize performance on indirect tests
(also see Denny & Hunt, 1992).
An exception to the initial rule was reported by Ruiz-Caballero and Gonzlez
(1994), who found evidence of mood congruence on a stem-completion task. In
this indirect test, beginning letters of both previously read words and new words
are provided, along with instruction to complete the stems with the rst words
that come to mind. Such a test can easily be turned into an explicit test of
memory-as-object if participants begin to use the stems as cues for deliberate
recall, even on occasional trials (see Watkins et al., 1992). Responding to these
concerns about the process purity of the test, Ruiz-Caballero and Gonzlez manipulated intention to learn in a second experiment. This manipulation aected
levels of free recall (which followed stem completion), but not levels of priming on stem completion, although performance on both tests showed evidence
of mood congruence. If the participants engaged in deliberate recall on the stemcompletion test, the authors argued, performance on that test should have shown
eects of intention to learn. But, as Chapman and Chapman (1973) alerted us,
the two memory tests might be dierentially sensitive to the manipulation of
intention to learn, while showing similar dierences according to emotional
valence. The provision of stems might compensate for the lack of intentionality
in the unintentional condition and thereby overwhelm the eect of intentional
learning, while nevertheless cuing recall in mood-congruent ways. Therefore,
doubts about mood-congruent uses of memory as tool remained.
Evidence for mood-congruent memory from tests of stem completion is surprising, particularly to a reader in a transfer-appropriate-processing frame of

191

mind. Among others, Roediger and McDermott (1992) discussed the importance of the match between types of processing at study and test. Moodcongruent thoughts during initial exposure should facilitate similar thoughts,
but not perceptions, at the time of testing, regardless of whether memory is used
as tool or object. Word-stem cues should function more perceptually than conceptually. Therefore, mood-congruent performance on stem completion should
not occur. In hindsight, it now seems clear that prior habits of reading moodcongruent words can sometimes facilitate performance on so-called perceptually driven tests (e.g., unprimed word identication, as used by Von Hippel,
Hawkins, & Narayan, 1994, and stem completion of both old and new items,
as reported by Bazin et al., 1994). Nevertheless, it was rare to see or to expect
to see evidence of mood-congruent memory on tests lacking a substantial reliance on conceptual processing.
Taking Roediger and McDermotts (1992) suggestion to use conceptual indirect tests, Watkins and his associates (Watkins, Martin, & Stern, 2000;
Watkins, Vache, Verney, Muller, & Mathews, 1996) have found mixed evidence
of mood-congruent performance by clinically depressed participants. Watkins
et al. (1996) found it on a test of free association, but there was no accompanying assurance against contamination by explicit remembering. Using both perceptual and conceptual indirect tests, Watkins et al. (2000) found evidence of
mood congruence following a conceptual orienting task, but only on one conceptually driven test, the test they called word retrieval (in which one produces
words when cued by dictionary denitions). Again, we cant be sure that participants were unaware of the memorial nature of the task or that they did not
attempt to remember deliberately if they were aware.
Contamination on indirect tests by deliberate remembering should not undermine interest in the many ways that habits of thought can inuence memory.
No doubt, all tests that reect prior experience do so through a mixture of automatic and recollective processes (see Jacoby, 1991). Even so, it is important to
know whether instances in which memory is used primarily as a tool for understanding current experience are aected by habits of negative thinking in depression, and in this regard more evidence is needed. Obtaining that evidence in the
laboratory is likely made more dicult by problems in controlling the thoughts
that should come to mind on indirect tests. Finding conceptual tests that do not
invite deliberate uses of memory as object and, at the same time, produce eects
that override other sources of uency from the past has been dicult. Thinking
of words from denitions might work because other words cannot be used, but
tests of free associations may invite too many extra-experimental responses. In
other words, memory-as-tool is hard to control experimentally.
A general-purpose or prototypical tool might work better in experimental
demonstrations of implicit bias. Past habits of thinking inuence performance
on current tasks not only because the exact thought keeps returning (the realworld analog for the indirect test of memory as tool) but also because past-

192

related thoughts guide current thoughts. Recent experiments in my lab illustrate

this point (Hertel, Mathews, Peterson, & Kintner, in press). An initial training
phase was designed to encourage either negative or nonnegative interpretations
of homographs as participants judged semantic relatedness (e.g., pursue vs. celery, in relation to stalk). The subsequent task directed participants to form images of individual words, none of which had been used during training and many
of which were homographs with both negative and nonnegative interpretations;
performance in this task reected the bias established during training. The training phase in this line of research is intended to model what happens naturally
when prior habits of thought inuence current interpretations in emotionally
biased ways. The inuence itself is an example of memory, put to use as tool,
although we tend not to think about it that way. Our typical impressions about
memorial inuences more often occur in conjunction with being reminded about
our personal past.

Memory as Autobiographical Experience

Habits of thought and rumination often include memories of events from the
personal past. Some of the earliest evidence for mood-congruent memory in depression was obtained in autobiographical studies (e.g., Clark & Teasdale, 1982),
with accompanying interpretations that related mood congruence to habits of
thinking. Now, direct evidence of this linkage is available. Using the same set of
ruminative or distracting procedures previously used to aect the temporary
mood of dysphoric students, Lyubomirsky et al. (1998) produced dierentially
negative biases in autobiographical recall.
In experiment 1, Lyubormirsky et al. (1998) randomly assigned dysphoric
and nondysphoric students to engage in 8 minutes of rumination or distraction
and then requested free recall of events from their lives. Following recall, the students rated the hedonic tone of the memories they produced. Dysphoric students
who had ruminated prior to recall rated their memories as less positive and more
negative than students in the other three conditions did. The same pattern of
results was obtained in experiment 2, in which the students were prompted to
recall two happy and two unhappy events, and in experiment 3, in which the
students judged frequency of occurrence in their lives for 10 positive and 10
negative events listed by the experimenter. In experiment 4, students were instructed to think aloud during the rumination or distraction phase, and their
thoughts were recorded. The rst six autobiographical memories mentioned in
the tapes were rated for negativity and unhappiness by independent judges. The
same pattern of mood congruence occurred: dysphoric students who ruminated
produced more negative memories, compared to those who were distracted and
compared to nondysphoric students in both conditions. All experiments also
produced the pattern of eects on mood ratings described previously in this chapter: only the dysphoric students felt more or less sad and depressed as a function

193

of rumination or distraction, respectively. In summary, ruminative thoughts not

only aected mood; they also invoked or contained more negatively toned memories from the dysphoric students personal pasts.
The experiments by Lyubomirsky et al. (1998) could not address the precise
nature of the mood-congruent nding. The authors acknowledge diculties in
concluding that the eects in the rst three experiments pertained to memory
instead of merely to ratings exaggerated by rumination. And, of course, in any
autobiographical study, we cannot be sure that the memories are accurate (although accuracy issues do not trump all other interests in bias). So, with these
qualications, it is reasonable to conclude that ruminative habits encourage
negative thoughts about the past. Especially those students with a reason for
focusing on negative past events were led to do so by a brief ruminative episode.
Making a similar point, Rothkopf and Blaney (1991) found clearer evidence of
mood-congruent autobiographical recall when students lled out the BDI prior
to the recall test rather than at other times. One interpretation of this type of
result points to demand established by the statements on the BDI. Another interpretation, however, posits that a depressed person must become aware of her
depressed state for the ruminative memories to be invoked. The self-schema must
be activated (Beck, 1967). Caught unaware of self, he might sidestep habits of
remembering the negative, and then other characteristics of the situation could
more powerfully guide performance.
The notion that a schema or prototype guides attempts to remember is consistent with a second major characteristic of autobiographical memory in depression, that is, its tendency to be overly general, as if attempts to remember events
from ones life stop short at the categorical level (see the review by Healy & Williams, 1999). Williams, Teasdale, Segal, and Soulsby (2000) provide the example
of the cue word kindness eliciting the following memory: My grandmother
was always kind to me. She used to take me out when my father got cross
(p. 150). The response referred to a category of events instead of to a (requested)
specic memory. Both positive and negative autobiographical memories tend to
be described at this categorical level by depressed people (as well as by those who
suer from posttraumatic stress disorder; see chapter 4). This tendency might
reect an unconscious habit or a conscious strategy to avoid emotion associated
with specic details (Williams et al., 2000).2 As with so many aspects of depressive cognition, the direction of cause is not at all clear. An overly categorical
pattern of thinking and remembering might very well contribute to onset of
depressive episodes, as suggested by Becks (1967) theory and by hopelessness
formulations of depression (Abramson, Metalsky, & Alloy, 1989). Whatever the
origin, the style is not so habitual that it cannot be modied (Williams et al.,
2000). There is even evidence that a focus on specic experience (instead of abstract analysis) can at least temporarily reduce the categorical extent of autobiographical memory (Watkins & Teasdale, 2001). Building new habits to recruit
specic instances can deect a habitual dark-cloud schema and reveal a dier-

194

ent self to the rememberer. Of course, this assumption has long been incorporated into the techniques of cognitive-behavioral therapy.

Summary
When past events serve as the objects of current thought, these thoughts are both
more negative and more abstract for depressed people than for others. Habits
role in explicit memory bias is depicted in results from manipulations of selfreferential processing and rumination. When memory is being used as a tool for
other purposes, however, convincing evidence of similar bias has not yet been
established for specic events, even though common sense tells us that these
unintended eects are ubiquitous. After all, these cognitive habits are nothing
other than the nondeliberate inuences of past conceptual experiences during
current, similar conceptual acts. Interpretive biases, surely established primarily by past experience, are well documented in depression.
A primary question about the inuence of cognitive habits on both forms of
memory concerns the extent to which the inuence relies on awareness of ones
mood state. Tasks that occur prior to tests for mood-congruent recall include selfreferential judgments of trait words, self-referential images, ruminative training, and a variety of mood-related forms and inventories. These tasks are
guaranteed to make one consider ones mood state. The distracting phrases used
by Lyubomirsky et al. (1998) might sidestep mood congruence not merely because mood is temporarily improved but also because other habits of thinking
have not been recently exercised. A continuing task for future research is to ask
whether negative biases can be invoked without awareness of self.
Interpretive and memory biases are not always symmetrical with mood. Depending on the type of depression and on the nature of the materials, the depressed participants do not always recall more negative words than positive
words, but they do fail to show the positive bias in the nondepressed group. Sometimes they recall negative materials at the same level as the nondepressed group
(but fewer positive words). In part, these asymmetries reect other dierences
associated with mood and memorydierences in the degree to which controlled, strategic thoughts are initiated and used. This is the other side of the
habit coin.

Impaired Control

in Nonhabitual Tasks

Decient cognitive control sets the stage for habits to emerge. At the same time,
thoughts that habitually occupy attention leave little mental room for thoughts
about anything else. These assertions of a reciprocal relation between habits of
thought and controlled attention characterize theoretical approaches to memory

195

in depressed states, particularly when the topic pertains to impaired performance.

Memory impairment is one of the frequent complaints of people in depressed
states (Beck, 1967).
Control refers to the operation of cognitive procedures at the opposite end
of the continuum from habit (see Jacoby, Jennings, & Hay, 1996). Controlled
procedures are initiated with awareness or intent, and their components are not
well integrated. From a phenomenological perspective, an individual must decide what to do or think nextquite a strain in depression. No wonder that eort
metaphors have been used so often in descriptions of depressed participants
diculties in tasks that require or reect control (see Hasher & Zacks, 1979;
Weingartner, Cohen, Murphy, Martello, & Gerdt, 1981).

Eortful Construction

of Memory as Object

The idea that depressed people have trouble carrying out eortful mental procedures emerged rst in early studies of intellectual functioning. (See the review
by Hartlage, Alloy, Vazquez, & Dykman, 1993.) Only after attention theorists
(e.g., Posner & Snyder, 1975) began to write about the control of attention, however, did researchers begin to connect eortful or controlled processing to performance on memory tests (Hasher & Zacks, 1979; Tyler, Hertel, McCallum, &
Ellis, 1979). In a nutshell, the idea was that the more eortful or attentiondemanding the process, the more likely that the product would be remembered.
Like the levels-of-processing framework popular among memory researchers at
that time (Craik & Lockhart, 1972), the connection of cognitive eort to memory
was based on assumptions about the strength of memory traces or links in an
associative network. In the case of eort, however, strength was thought to directly reect the amount of attention or eort expended during initial processing, instead of the type or level of processing. Orienting tasks vary in the degree
of eortful, attention-demanding processes required to complete them; those
requiring more eort supposedly produce a stronger memory. This idea was not
unlike the idea of diculty, but eort referred to a characteristic of processing,
not the task itself. Eort-inducing tasks inspired or required more focused concentration, as revealed by longer latencies to perform a simple secondary task
(e.g., Tyler et al., 1979).
The applicability of the eort-memory connection to depression was a central feature of Hasher and Zackss (1979) often-cited work. Since then, literature reviews and a meta-analysis have supported the idea that depression-related
decits in memory occur primarily when the initial orientation task requires
controlled concentration (Burt et al., 1995; Hartlage et al., 1993). Hartlage
et al. also described the variety of theoretical accounts of eorts role in depression-related impairment. Condensed and simplied, one account connects the
concept of reduced attentional control to abnormal frontal function. (See

196

Davidson, 2000, for a more recent review.) Another claims that eortful processing is limited by the allocation of resources to personal concerns, presumably for the duration of the depressive episode (e.g., Ellis & Ashbrook, 1988).
Implicit or explicit in all eort accounts is the assumption that greater eort produces a memory trace that is more easily found when needed. This assumption,
rarely stated quite so simply, is clearly too simple in any form.
The ability to focus attention and concentrateeither because the neurotransmitters are fully supplied to the appropriate areas of the brain or because
other matters are not more compellinglikely has benecial consequences for
memory, but only when those particular procedures are subject to replication
on the test. The intensity of concentration is important to memory only in the
sense that it must be sucient to carry out the procedures that will later, when
replicated, benet performance on a memory test. According to a transferappropriate-processing perspective, research should examine that suciency, as
well as the match in attentional focus across the occasions of initial exposure and
memory test. One obvious way to evaluate the transfer-appropriate claim is to
examine the literature on memory as tool-versus-object in depressed states, because both the need for controlled attention and its focus varies across types of
tests. If depressed people do not focus on aspects that will later guide performance
on memory tests, and if they do not concentrate suciently for later replication,
performance should suer, regardless of the intentionality of the test or its
diculty.

Appropriate Use of Memory

as Tool or Object

Indirect tests of memory are designed to elicit automatic or habitual use of prior
experience and often do not require concentrated eort during their performance.
Spelling homophones is an example of such a test. It takes little concentration
to spell common words read aloud by the experimenter at a fast pace (e.g., week/
weak). Similarly, the orienting tasks prior to the spelling test often require little
concentration on one of the two alternative meanings (e.g., while listening to
the question: Name the days of the week). Simple spelling and comprehension
use habitual procedures. As expected, the dysphoric students spelling performance matched that of controls in eects of question-biased meaning; yet in the
same experiment, the subsequent use of recognition strategies was impaired in
the dysphoric group (Hertel & Hardin, 1990).
On the other hand, memory can operate as a tool in performing somewhat
more attention-demanding tasks. Even stem- and fragment-completion tests,
when not performed habitually, might function as puzzles that require at least
momentary eort and concentration, yet the available reports show no evidence
of depression-related impairment (in the same experiments that revealed impaired free or cued recall of nonnegative words; e.g., Bazin et al., 1994; Denny

197

& Hunt, 1992; Ruiz-Caballero & Gonzlez, 1994; Watkins et al. 1992).3 Why
should we not see occasional evidence of impairment? One possible reason is that
it is sucient to have focused attention on the look or sound of the words in the
orienting task in order that memory operates as a tool during perceptual indirect teststests that might well demand concentration on seeing or hearing them
again. Sustained processing of meaning does not boost performance, but any task
that discourages attention to words as units might decrease it. For example, consider an experiment that used a word-identication test, in which previously presented and nonpresented words were displayed very briey and back-masked, and
the task was merely to try to name the words. Anyone who has tried these tests
knows that they require focused attention to see the words. To what extent
should identication benet from prior exposure (memory as tool)? In the orienting task of one such experiment (Hertel, 1994), I asked participants to evaluate
some words according to the degree of roundness in their letters. The words evaluated in this task were identied more frequently on the subsequent perceptual test
than were words not previously exposed, but this eect was smaller for the depressed participants! A reasonable postdiction is that during the orienting task the
depressed participants attended less often to the words as integrated units and so
were disadvantaged in naming them on the test. Word identication benets from
prior lexical processing, even as it requires concentration or eort.
Conceptual processing facilitates use of the same concepts later on. Conceptual implicit tests require prior attention to meaning and therefore should reect
the extent to which meaning has been fully attended. For example, Jenkins and
McDowall (2001) found a depression-related impairment on both free recall and
the indirect test of category association, but only for the words that had been
generated from antonyms (vs. merely read) during the orienting phase. No clear
evidence of impaired fragment completion was found for either the previously
generated or previously read words. These results are not as unambiguous as I
describe them; for example, baseline performance on the category-association
test was low for the depressed group. Nevertheless, the pattern illustrates the
importance of transfer-appropriate conceptual processing. And the results also
suggest that the depressed participants conceptual processing during antonymgeneration was insucient to support the later production of those concepts
during recall or category association. Whether category association functioned
purely as a test of memory as a tool cannot be assured, of course. Like the freeassociation task used by Watkins et al. (1996), it might have been contaminated
by attempts to remember, particularly in the nondepressed group. Because other
category associates are available, the depressed participants could perform the
test without concentrating on the past and, therefore, without showing the benet of prior generation of category members, and a mood-related impairment
would thereby result.
Depressed people are less likely to devote eort to thinking about the past when
the current task can be performed without doing so. This claim receives indirect

198

support from a series of problem-solving experiments in which memory for similar past problems operated as tool or object in facilitating solutions for current
problems (Hertel & Knoedler, 1996). The logic-based word problems were themselves attentiondemanding and quite dicult without prior experience in learning how to solve similar problems. During training, problems were presented for
solution attempts, and each (almost always unsuccessful) attempt was followed
by experimenter-provided instruction in how to solve that problem. Following
the training phase, a series of structurally identical but supercially dissimilar
target problems was presented, always without such instruction. Because we
were curious about how memory operates as a tool in problem solving, we included a condition to measure spontaneous transfer; in this condition, the training problems were not mentioned during the target series. We also included a
condition in which we provided very specic cues and instructions for using
memory for the structurally similar training problem as object just prior to
presenting each target. In two experiments, these hints for recalling the prior
analogy actually disrupted performance by the nondysphoric students. Focusing attention on the details of the past problem while solving a new one seemed
to be a transfer-inappropriate use of prior experience; in comparison, more targets were solved when memory for the analogy was allowed to operate as a tool.
The dysphoric students performed similarly well, with or without hints to think
back, which led us to suspect that they did not sustain attention to the relevant
past problem and therefore were not led astray in transfer-inappropriate ways.
Attending to important features of the target problem itself beneted from earlier attention to those features of the analog.
In predicting performance in the problem-solving experiments, Knoedler and
I initially reasoned that if the nondepressed students solution rates were similar with and without hints, they might have been using memory as object on
their own initiative in the no-hint condition. If this pattern were obtained, we
expected to nd depressive impairments. When the hints turned out to be detrimental for nondepressed students, we inferred that memory for training problems in the no-hint condition operated spontaneously as a tool. In reasoning this
way, we did not truly believe that either use was all or none. Similarly, most
researchers understand that memory usually operates as both tool and object
in a variety of memory tasks and that memory tasks are not process-pure
(Jacoby, 1991). Yet when no attempt is made to determine the degree to which
one use of memory contaminates the other use, reasoning about the role of controlled attention in depressive memory is made more dicult. Luckily, Jacobys
process-dissociation procedure directly addressed the problem of how to examine the separate inuences of tool and object memory.
The central method used by Jacobyand others, subsequentlyentails two
conditions of instructions: one in which automatic or habitual uses of memory
operate in concert with controlled attention to the past and one in which they
operate in opposition to each other in producing a response. In recognition tasks,

199

for example, participants are instructed to endorse items that either feel familiar or are recollected from a specic prior task on inclusion trials, whereas on
exclusion trials, they are instructed to reject items from that specic task while
endorsing other old items. In the inclusion case, recognition hits are produced
by both controlled memory for specic prior occurrences and habitual procedures
of reading or thinking words that were earlier read or thought. The uses of
memory operate in concert. In the exclusion case, however, items from the specic prior task are (erroneously) endorsed only if they produce a feeling of familiarity in the absence of memory for source as object. Habit and control operate
in opposition to each other. When independence of the two classes of inuences
can be assumed, their estimates can be computed from the proportion of positive recognition responses under the two instructional conditions. The relevance
of these procedures to depression-related impairment should be obvious: Unless
habits are disrupted during orientation, habitual use of memory as tool should
remain intact, and only the estimates of controlled uses of memory as object
should reveal the impairment. This indeed is what we found on recognition tests
following a conceptual orienting task (Hertel & Milan, 1994). The dissociation
was replicated through similar procedures designed for conceptually cued fragment-completion tests (e.g., building-s_o_e) following the reading of meaningful cue-target pairs in the orienting task (e.g., building-stone), (Hertel, 1998; also
see Hertel & Meiser, 2000, for a multinomial model applied to those results).
Again, the importance of transfer-appropriate attention across orienting task and
test is illustrated. To the extent that participants attended to the connection
between the cue and target initially, they would be able to attend to it as an object
in the past, including the target or excluding it when required by instruction. A
less careful focus on the pair initially would make it harder to both exclude and
include deliberately on the test. The component of memory that represents the
past as the object of attention is the one that shows a decit associated with depression. However, the mere reading of the word pair during the orienting task
makes the target available later, as a completion of the fragment, in the presence or absence of controlled recollection, and this habit-oriented component
of memory was unimpaired. Thus, attentional control is indeed a factor in producing memory impairment in depression, when performance relies on the replication of focus across the two occasions.4

Habit as Detriment to Control

Poor performance often reects absence of habit in performing a cognitive act.
This is why controlled attention is requiredto do the nonhabitual and think
about things one does not usually think about. Seemingly, all of the reported
cases of depression-related impairment involved nonnegative materialsmaterials unrelated to the personal concerns thought to occupy attention habitually
in depression. Habits of negative thinking benet the controlled use of memory

200

for related stimuli, as is the case on tests of mood-congruent recall, but otherwise they can detract. Therefore, attentional control in depression should depend
on the suppression of habitual thoughts in the service of focusing elsewhere. A
line of research begun by Hasher and Zacks (1988) has shown evidence of inhibition diculties associated with aging that might also characterize depression.
Like eort-related diculties, inhibitory diculties in depression could logically
arise from fundamental frontal dysfunction, from specic concerns associated
with the depressed state, or from both (see Hertel, 1997).
Reliance on an inhibition construct is not the only way to represent the problem of impaired control in the context of habitual thinking. We might slightly
shift perspective to propose that control is made more dicult when other habits are strong. Habits of attending (e.g., to negatively toned events) might make
the act of disengaging attention more dicult than it otherwise might be in the
course of turning attention to a dierent event. At this stage of research, however, the aptness of the metaphor (inhibition vs. disengagement) is less important than gathering evidence for depression-related diculties in turning away
from habit-related stimuli in order to do something else.
The something else in experimental studies has often been the task to respond to a dot or color patch in a screen location that diers from the one occupied by a mood-relevant word (e.g., Bradley, Mogg, & Lee, 1997; McCabe &
Gotlib, 1995). Bradley et al. found dysphoria-related biases under conditions of
suprathreshold but not subthreshold exposures to the words. They argued that
the problem was therefore one of disengaging attention, because a bias in allocating attention to relevant stimuli would pertain for both exposure conditions.
Recently, investigators have used the negative-priming paradigm to examine depression-related diculties in disengagement or inhibition. In the standard
version of this paradigm, items to be attended and ignored are presented concurrently on each trial, with the attended item on trial X sometimes having
served as the ignored item on trial X-1. When the ignored is subsequently attended, or even when the same category of the ignored is now the category of
the attended, the judgment for trial X is slowed, compared to control conditions.
This slowing (called negative priming) is attributed sometimes to inhibition of
the ignored item or category and sometimes to transfer-inappropriate attention
(Neill & Mathis, 1998). If depressed people have trouble disengaging from moodrelated thoughts, negative-priming eects should be reduced or absent. Some
suggestion of this diculty has been demonstrated in a simple letter-identication
task after students were induced by music to feel sad (von Hecker, Conway,
Meiser, & Holm, 2002). However, Joormann (in press) has used a more conceptually relevant task. She asked nondysphoric and mildly dysphoric students to
judge the self-relevance of negative and positive targets while ignoring distractors
of the same or opposite valence. All participants took longer to judge the selfrelevance of positive words when positive words had been ignored on the previous trial, compared to previous trials of ignored negative words. This nding

201

roughly corresponds to the typical negative-priming eect. The same pattern of

negative priming was obtained for nondysphoric participants in judging negative items after having ignored negative (vs. positive) items, but it was not pertain obtained for mildly dysphoric participants or for those who reported prior
episodes of depressed mood. The dierence in these patterns suggests that the
participants who tended currently to be in more negative mood states (or reported
feeling depressed in the past) ignored the negative adjectives less successfully.5
Diculty in disengaging attention from negative material is more clearly
documented in connection with anxiety disorders or high levels of trait anxiety
(see Fox, Russo, Bowles, & Dutton, 2001). Again, when researchers use scores
on inventories like the BDI or even when they use diagnostic labels to form groups
for experimental purposes, their dysphoric or depressed participants are likely
to be anxious. In the other direction, ruminative habits more typically attributed
to depression also typify anxious states or traits (Nolen-Hoeksema, 2000). Regardless of their lack of specicity to depression, however, habits of dwelling on
the negative have consequences for memory beyond those reported in relation
to evidence for mood-congruent memory, and these consequences are not always
desirable.
One consequence might be trouble in forgetting. Not only do depressed people
have trouble ignoring mood-congruent material when they focus on other material presented concomitantly (as in Joormanns negative-priming studies), they
also less successfully suppress mood-congruent material in more sustained attempts. Evidence for diculties in sustained thought suppression has been documented by Wegner, Wenzla, and their colleagues (for review, see Wenzla &
Bates, 1998). Thought suppression, whether accomplished directly or indirectly
in the service of thinking about something else, plays an important role in forgetting (see Anderson & Neely, 1996). And forgetting can be a valuable outcome,
particularly for the depressed person who might prefer not to ruminate about
past events beyond ones control.
Intentional forgetting has been studied in the lab by directing participants to
forget items or lists of items.6 Clearly, instructions to forget lists can have their
intended eect, particularly when the participants are given something else to
think about. To document eects of direct suppression on individual items,
Anderson and Green (2001) developed the think/no-think procedure. After
learning pairs of unrelated words to a xed criterion, the participants were given
a varied number of trials (0, 1, 8, or 16) to practice the retrieval of the second
member of some pairs and to practice the suppression of the second member of
other pairs. The rst member of each pair was presented as a cue for either retrieval or suppression. This practice phase was followed by a nal test of cued
recall, in which participants were instructed to recall all of the response words,
regardless of previous instruction. In several experiments, the main nding was
increased forgetting as a function of practice in suppression, both when the original studied cues were used at test and when new cues related categorically to

202

the response words were used at test. The simple story is that people can successfully suppress to the point of incurring forgetting. Can depressed people do
this as well as others? Melissa Gerstle and I have recently conducted an experiment to answer that question (Hertel & Gerstle, in press).
To investigate the possibility of mood-congruent suppression, we replicated
the procedures developed by Anderson and Green (2001), but with one major
change: The cues we used throughout the experiment were adjectives related to
the target nouns. On half of the pairs, the adjectives gave the essentially neutral
nouns a negative meaning, and on the other half, the meaning was positive. (The
materials were fully counterbalanced, e.g., gloomy cottage vs. splendid cottage.)
To encourage initial attention to the emotional valence of the materials, in the
learning task we instructed participants to construct a self-referential image for
each pair and to rate the meaningfulness of the image. These materials and instructions had been used earlier in a recall experiment without suppression instructions, together with other materials in which a third set of adjectives
produced neutral concepts when paired with the same nouns (e.g., plain cottage;
Hertel & Parks, 2002). In that experiment, dysphoric and nondysphoric participants alike recalled the nouns when they had been given emotional meanings
(both positive and negative) more often than when they had been given neutral
meanings. In the suppression experiment, Gerstle and I also found that recall was
not signicantly associated with emotional valence for the dysphoric participants
(although the controls showed more forgetting of nouns from positive pairs).
More important, however, was our nding that the eect of practicing suppression on later recall was signicantly reduced in the dysphoric group. The
dysphoric participants recalled as many of the rehearsed items as did the nondepressed participant and signicantly more of the targets that they had practiced suppressing, regardless of the cues valence (see g. 6.1).
Similarly, Power, Dalgleish, Claudio, Tata, and Kentish (2000) found evidence for depression-related diculties in one of their three experiments on directed forgetting. Perhaps because their participants were clinically depressed
(experiment 3), however, this dierence was restricted to negative adjectives.
Recall of positive adjectivesto be remembered or to be forgottenwas no better or worse than that of anxious participants and controls.
At the same time that depressed people might have trouble deliberately forgetting self-referential thoughts, they also have trouble remembering what they
had been thinking about prior to an episode of self-referential rumination. At
least this is true for the controlled component of memory, as estimated through
process-dissociation procedures on a test of cued fragment completion (Hertel,
1998). In the experiment previously mentioned, we inserted what might be seen
as an interference phase between the initial orienting task of reading word pairs
and the nal memory phase. We assigned a third of the dysphoric and control
participants to merely sit and wait for 7 minutes, another third to ruminate, and
the nal third to think about other matters, such as geographical locations. The

203

Nondys Respond

Nondys Suppress

Dysphoric Respond

Dysphoric Suppress

Baseline

Practice

Mean % Recalled

100

90

80

70

Intervening Task
Figure 6.1. The mean percentage of targets recalled on the final test, after depressed
and nondepressed participants had practiced retrieving or suppressing them in response to cues. (Adapted from Depressive Deficits in Forgetting, by P. T. Hertel
and M. Gerstle, in press, Psychological Science. Copyright 2003 by the American
Psychological Society.)

latter two groups actually underwent the rumination or distraction procedures

used by Nolen-Hoeksema and Morrow (1993) that I described elsewhere in this
chapter. Following the period of self-focus in the rumination condition, the dysphoric participants showed impaired control on the memory test (relative to the
other participants), much like those who simply waited during the 7-minute
interval. However, the participants who thought about objects and locations
showed no depression-related impairment in control (see g. 6.2). In addition
to improving mood under some conditions (e.g., Lyubormirsky et al., 1998), distraction also seems to improve controlled attention to the past when the
rememberer is depressed. And because the distracted group showed no impairment, it is tempting to infer that even those who merely waited for 7 minutes
focused on suciently compelling matters that thinking back during the memory
test was made more dicult.
Inherent in those results (Hertel, 1998) are the nal two points of this chapter: situations that permit mind wandering invite habitual thinking and impair memory for other matters. Decient performance is not, therefore, an
inevitable outcome of depression, because habits can be opposed in ways that
benet control.

204

Mean Estimated Recollection

0.5

Nondysphoric
Dysphoric

0.4
0.3
0.2
0.1
0
Wait

Ruminate

Distract

Intervening Task
Figure 6.2. The mean estimate of controlled retrieval by dysphoric and nondysphoric participants, following an intervening period of waiting for 7 minutes, entertaining ruminative thoughts, or entertaining distracting thoughts. Adapted
from The Relationship between Rumination and Impaired Memory in Dysphoric Moods, by P. T. Hertel, 1998, Journal of Abnormal Psychology, 107, p. 170.
Copyright 1998 by the American Psychological Association.

Opposing Habit Through Control

Here is the prototype of the depressed rememberer as constructed by the ndings described thus far: she (well, yes, she) is someone who habitually ruminates
about personal concerns and other negative events and whose interpretations
and perceptions reect those habits; someone who tends to remember negatively
aective events more often or remembers events as more negatively toned; and
someone whose memory for neutral or positive events is occasionally impaired,
at least when attention is allowed to stray to habitual patterns. This person seems
to have little cognitive self-control. (One can almost hear her grandmothers
admonishments.)
I use self-control to imply that the most important aspect of impaired control in depression is control initiated by the self. Controlled attention to any stimulus can be facilitated by the environment; the structure of the task, stimulus
salience, and its interest value are some of the relevant dimensions. Cognitive
self-control is merely a convenient term for referring to the focusing, switching,
and sustaining of attention under conditions where stimulus and task dimensions do not play obvious roles. In structural terms, self-control is loosely analogous to the central executive component of working memory (Baddeley & Hitch,
1974). Its physical components are frontal functions; indeed, much of the evidence for depression-related dierences in brain activity implicates activity in the
prefrontal cortex (see Davidson, 2000; Heller & Nitschke, 1998). Similar

205

behavioral evidence is supplied by Channon (1996), who found impairments on

a neuropsychological test of frontal functionthe Wisconsin Card Sorting Test
(WCST)even in a sample of dysphoric college students. Furthermore, by statistically controlling for correlations with BDI scores, Davis and Nolen-Hoeksema
(2000) found WCST dierences associated with the tendency to ruminate. Selfreported ruminators made more perseverative errors on the WCST and failed to
maintain set more often than did nonruminators. In a basic sense, then, cognitive self-control is a problem in depression, as well as a problem for people who
tend to ruminate.
A good simple example of the memory problems associated with depressive
decits in self-control or initiation is impaired prospective memory in the absence
of reminding cues (Rude, Hertel, Jarrold, Covich, & Hedlund, 1999). Depressed
and control participants were instructed at the start of the session to press a function key every 5 minutes during their performance of an ongoing test of general
knowledge. The depressed participants checked the time less often and thereby
made more errors on the prospective task.
Reduced self-control aects performance in memory experiments under conditions of poor external control. For example, dysphoric students performance
yielded lowerestimates of controlled recollection in my cued-fragment-completion study (1998) after a period of lax external control. Doing nothing for 7 minutes presented ample opportunity for mind wandering, because the dysphoric
students had more trouble subsequently turning attention to the task of remembering. Although this depression-related impairment in controlled recollection
was closely mimicked by data from participants in the rumination condition, it
is not certain that the waiting students also ruminated during that period. For
some, the break from the ongoing experiment might have invited mind blanking. As a learned response to stress, the habit to tune out might be just as reective of poor self-control as the habit to ruminate. Watts, MacLeod, and Morris
(1988) found that blanking tended to occur in tasks that involve planning (tapping frontal functions). Although we know too little about blanking in comparison to mind wandering or rumination, either type of habit should take over
during periods of lax external control of the procedures to be performed.
As I have argued previously (1997, 2000), lax external control often typies
real-world thinking, and it sometimes characterizes the trials that constitute the
orienting task in laboratory experiments. The orienting task is, according to eort
metaphors and other processing frameworks, the critical period in determining
whether the material will be recalled later. Therefore, in allowing sucient time
on each trial for the elaborative processing that benets recall, memory researchers also encourage habits of mind wandering or blanking. Years ago, Stephanie
Rude and I used a simple manipulation of either allowing an orienting decision
to be made at any time during an 8-second exposure or requiring the decision at
the end of the trial, along with a repetition of the target word. (The decision concerned whether a target word t sensibly into a sentence frame.) We called the

206

two conditions unfocused and focused, respectively (Hertel & Rude, 1991a). Inthe
focused condition, the materials did not stay on the screen for the 8-second period, as they did in the unfocused condition. To perform each trial in the focused
condition, the participants had to hold the word and decision in mind for the
duration of the trial, but their minds could wander in the unfocused condition
because decisions could be reported early on during the 8 seconds. On the subsequent test of free recall, depressed participants from the unfocused condition
performed poorly, compared to their nondepressed counterparts. This decit,
however, was eliminated for participants who had been required to focus during the orienting task.
The depressed participants ability to perform the task could not ever have been
at issue. The task was not dicult in either condition (even though it had been
used previously as an example of an eortful orienting task by Tyler et al., 1979,
and Ellis, Thomas, & Rodriguez, 1984). Indeed, the more dicult condition clearly
was the focused one, because it required sustained attention. Therefore, instead
of making the task easier to accomplish, the focusing instruction structured the
task transfer-appropriately. Participants were required to attend to the recent past
(a few seconds earlier) in the same way that they were later required to attend to
the past during the recall test. Moreover, external control compensated for poor
self-control by not allowing habits of thought (or no-thought) to carry the day.
There were a number of variations of this attention-focusing procedure (Hertel
& Rude, 1991b). We also tried to stien the requirements at the time of the test
by using a forced-recall procedure. Although forcing the participants to write
a certain number of words on the recall test merely increased levels of recall for
everyone in our experiments, others have found that this procedure eliminated
the depression-related decit (e.g., Murray et al., 1999). The forcing procedure
counteracts a possibly conservative response tendency, but it does so by asking
the participant to refocus attention on the past events, which might be at least
as important as overcoming hesitancy or conservatism.
Other methods of improving performance by providing external support have
not always been successful. For example, in the recognition experiment with
process-dissociation procedures, Milan and I reinstated the original context word
in some test conditions. That procedure boosted estimates of controlled recollection for both dysphoric and nondysphoric participants (Hertel & Milan, 1994).
Clearly, the nondysphoric students had done something that we had failed to
capture by the manipulation. This was also true when recognition was much
more strategically dependent (Hertel & Hardin, 1990). Hardin and I noticed that
the nondysphoric students seemed to be using their performance on the prior
indirect test of homophone-spelling to decide whether the recognition item had
occurred in the original set of questions (e.g., Week/Weak? Did I just spell that
word? How did I spell it? And was that word in a question? Oh, name the days of
the week? Yes.). In a subsequent experiment, we asked participants to answer a
similar set of questions before making each recognition judgment. By doing so,

207

we found a pattern of dependency in the dysphoric group that mimicked the

pattern of nondysphoric judgments under unassisted conditions, but the
nondysphoric participants who got the set of question cues outwitted us. Their
performance improved even for those items that they had spelled in the opposite
way (e.g., Week/Weak? Did I just spell that word? Weak? Is there another way
to spell it?). These and similar dierences illustrate the diculty in predicting
exactly what people do under conditions of lax external control, but the ndings
also show that performance in depressed states can be improved by at least some
version of external support. And the main point is that this support works because it forestalls the emergence of cognitive habits in depressionhabits that
interfere under poorly controlled conditions and make it dicult for the depressed
person to take control and show cognitive initiative (see Hertel, 2000).
Habits of thought are dicult to oppose. The historically important construct
of habit strength in Hulls theory of learning (1943) was opposed by the construct
of inhibition, but Hulls rats experienced immediate consequencesmissing
rewardthat led to the orderly development of inhibitory strength. The consequences of thought are often much farther removed, so suppressing a turn of the
mind is perhaps much harder to learn than is suppressing a turn in the maze.
Modern approaches to thought suppression acknowledge this diculty. Wegners
ironic-processes theory (1994), for example, includes a construct called a monitor that checks whether thoughts are being successfully suppressed. According to
this point of view, the depressed persons need to suppress self-concerns, particularly during tasks that require self-control and initiative, presumably activates
the monitor and ironically brings those concerns back to mind. At the very least,
any attention devoted to the task of suppression likely disrupts procedures that
require sustained attention or planning. Instructions to suppress are therefore
much harder to follow when not accompanied by a task that directs attention
through external means.
The short version of that argument is that simple self-controlled thought
suppression likely does not work well for depressed people, both in the sense of
potentiating later forgetting (Hertel & Gerstle, in press) or freeing attention to
focus in ways that later benet remembering. In the latter case, what works best
is environmental support, but the world is not often arranged to be helpful. What
also seems to work in the sense of improving both memory and momentary mood
is a good distraction procedure (e.g., Hertel, 1998; Lyubormirsky et al., 1998).
Distraction at least interrupts ongoing habitual thoughts so that the depressed
person begins the recall task on even footing. There is even some reason to wonder if reappraisals of intrusive thoughts, under some circumstances, might aid
recall of nonnegative events through temporary mood improvement or altered
perspective (in the manner found by Rusting & DeHart, 2000; also see Parrott &
Spackman, 2000). However, are these measures enough?
According to Teasdale et al. (2000) and others, approximately 80% of people
diagnosed with major depressive disorder experience recurrent depressive epi-

208

sodes. Practical considerations therefore encourage us to consider not merely the

opposition of habitual thinking by controlled procedures but the training of new
habits through practice in control. Virtually the only line of research with that
general aim is the research on mindfulness training, adapted from the meditation work of Kabat-Zinn by Teasdale, Williams, and Segal. The training consists
of a sustained focus on some simple act (typically breathing). When attention
wanders, the individual acknowledges the new thought or feeling but then redirects attention to the simple act. It diers from simple thought suppression in
two main respects: acknowledgment of the thought and practice in redirection.
Do the benets of such training extend to long-term habits of control? So far,
Williams et al. (2000) have found that mindfulness training boosts the eects
of cognitive-behavioral therapy in reducing overly general autobiographical
memory by formerly depressed participants. Moreover, for people with at
least three prior episodes of depression, the addition of mindfulness training to
cognitive-behavioral therapy signicantly reduced recurrence during the subsequent year (Teasdale et al., 2000). These studies, of course, are merely a beginning.
In short, the best antidote to maladaptive habits is a new set of habitsnot
the opposite sort of habits recommended by Pollyanna and the teachings of
Norman Vincent Peale (1956) and not the habits of suppression as recommended
by certain grandmothers but the habits of thought control. Cognitive control,
as conceived by cognitive psychologists, is a concept that opposes habit, in that
habitual procedures require little attention. But if the necessary control proceduresawareness of the wandering stream of thought and subsequent disengagement, for exampleare practiced religiously, they too can become
habitual. In the absence of mind wandering and rumination, memory biases are
less extreme, and memory impairment is repaired.

Notes
1. Induction procedures are particularly useful when it is dicult to determine
the content of natural thought, although some procedures do not produce memorial eects that mimic patterns in natural depression. For a case in which they fail
to show mood-congruent recall in the same study in which it was obtained for dysphoric students, see Kwiatkowski and Parkinson (1994; also see Eich & McCaulay,
2000; Perrig & Perrig, 1988).
Inductions have also been used to study mood-dependent memory: better
memory when mood at initial encounter matches mood at test. Mood-dependent
memory has also been studied with participants diagnosed with bipolar disorder.
When the mood during initial encounter (mania or depression) matched the
mood at test, higher levels of recall or recognition were found (Eich, Macaulay, &
Lam, 1997; Weingartner, Miller, & Murphy, 1977). See Blaney (1986, p. 237)
for a discussion of how mood-dependent memory can be seen as a special case of
mood-congruent memory.
2. Ironically, overgeneral tendencies have also been correlated with spontaneous

209

intrusions of stressful memories (Brewin, Reynolds, & Tata, 1999; Wessel, Merckelbach, & Dekkers, 2002).
3. In one exception to the early rule, Elliott and Greene (1992) did show impaired
word-stem completion and homophone spelling but omitted unprimed items on the
test, which might have encouraged participants awareness of the memorial goal
and a concomitant use of memory as object. (See Bazin et al., 1994, and Roediger &
McDermott, 1992, for full discussions of their design.)
4. Of course, attention might well be impaired during initial exposure. Consider
a report by Rokke, Arnell, Koch, and Andrews (2002), who investigated the size of
attentional blinks in a rapid-serial-visual-presentation paradigm (RSVP) administered to students with varying degrees of dysphoria. Moderately to severely dysphoric
participants performed relatively well when the only task was to detect the presence
of a single letter in the stream of rapidly presented letters. However, their performance suered more, compared to that of nondysphoric or mildly dysphoric participants, when the task also involved attention to a letter occurring earlier in the
stream (a deeper attentional blink). This outcome implicates fundamental decits
in redirecting attention.
5. Some of the results reported in this section should be viewed cautiously, given
either unexpected dierences on baseline trials (e.g., Joormann, in press) or lack of
information about baseline (e.g., Bradley et al., 1997).
6. Studies using this directed forgetting paradigm have revealed interesting
dierences associated with clinical disorders. For example, Sonntag et al. (2003)
have shown that schizophrenics fail to show eects of directed forgetting on measures of recollection (see chapter 7). McNally, Metzger, Lasko, Clancy, and Pitman
(1998) found poor directed forgetting of trauma words by women with posttraumatic stress disorder, as well as decits in remembering neutral and positive words
on the lists to be remembered (see chapter 4).

References
Abramson, L. Y., Metalsky, G. I., & Alloy, L. B. (1989). Hopelessness depression: A
theory-based subtype of depression. Psychological Review, 96, 358372.
Alloy, L. B., Abramson, L. Y., & Francis, E. L. (1999). Do negative cognitive styles
confer vulnerability to depression? Current Directions in Psychological Science,
8, 128132.
Anderson, M. C., & Green, C. (2001). Suppressing unwanted memories by executive control. Nature, 410, 366369.
Anderson, M. C., & Neely, J. H. (1996). Interference and inhibition in memory retrieval. In E. L. Bjork & R. A. Bjork (Eds.), Memory (pp. 237313). New York:
Academic Press.
Baddeley, A., & Hitch, G. J. (1974). Working memory. In G. Bower (Ed.), The psychology of learning and motivation (Vol. 7, pp. 4789). New York: Academic
Press.
Bazin, N., Perruchet, P., De Bonis, M., & Fline, A. (1994). The dissociation of explicit and implicit memory in depressed patients. Psychological Medicine, 24,
239245.

210

Beck, A. T. (1967). Depression: Clinical, experimental and theoretical aspects. New York:
Harper & Row.
Beck, A. T., Ward, C., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory
for measuring depression. Archives of General Psychiatry, 4, 561571.
Blaney, P. H. (1986). Aect and memory: A review. Psychological Bulletin, 99, 229
246.
Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129148.
Bradley, B., & Mathews, A. (1983). Negative self-schemata in clinical depression.
British Journal of Clinical Psychology, 22, 173181.
Bradley, B. P., Mogg, K., & Lee, S. C. (1997). Attentional biases for negative information in induced and naturally occurring dysphoria. Behaviour Research and
Therapy, 35, 911927.
Brewin, C. R., Reynolds, M., & Tata, P. (1999). Autobiographical memory processes
and the course of depression. Journal of Abnormal Psychology, 108, 511517.
Brown, T. A., Di Nardo, P. A., Lehman, C. L., & Campbell, L. A. (2001). Reliability
of DSM-IV anxiety and mood disorders: Implications for the classication of
emotional disorders. Journal of Abnormal Psychology, 110, 4958.
Burt, D. B., Zembar, M. J., & Niederehe, G. (1995). Depression and memory impairment: A meta-analysis of the association, its pattern, and specicity. Psychological Bulletin, 117, 285305.
Channon, S. (1996). Executive dysfunction in depression: The Wisconsin Card Sorting Test. Journal of Aective Disorders, 39, 107114.
Chapman, L. J., & Chapman, J. P. (1973). Problems in the measurement of cognitive decits. Psychological Bulletin, 79, 380385.
Clark, D. M., & Teasdale, J. D. (1982). Diurnal variation in clinical depression and
accessibility of memories of positive and negative experiences. Journal of Abnormal Psychology, 91, 8795.
Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for
memory research. Journal of Verbal Learning and Verbal Behavior, 11, 671684.
Danion, J.-M., Willard-Schroeder, D., Zimmermann, M.-A., Grange, D., Schlienger,
J.-L., & Singer, L. (1991). Explicit memory and repetition priming in depression.
Archives of General Psychiatry, 48, 707711.
Davidson, R. J. (2000). Aective style, mood, and anxiety disorders: An aective
neuroscience approach. In R. J. Davidson (Ed.), Anxiety, depression, and emotion (pp. 88108). New York: Oxford University Press.
Davis, R. N., & Nolen-Hoeksema, S. (2000). Cognitive inexibility among ruminators and nonruminators. Cognitive Therapy and Research, 24, 699711.
Deldin, P. J., Keller, J., Gergen, J. A., & Miller, G. A. (2001). Cognitive bias and emotion
in neuropsychological models of depression. Cognition & Emotion, 15, 787802.
Denny, E. B., & Hunt, R. R. (1992). Aective valence and meory in depression: Dissociation of recall and fragment completion. Journal of Abnormal Psychology,
101, 575580.
Derry, P. A., & Kuiper, N. A. (1981). Schematic processing and self-reference in clinical depression. Journal of Abnormal Psychology, 90, 286297.
Dobson, K. S., & Shaw, B. F. (1987). Specicity and stability of self-referent encoding in clinical depression. Journal of Abnormal Psychology, 96, 3640.

211

Eich, E., & Macaulay, D. (2000). Are real moods required to reveal mood-congruent and mood-dependent memory? Psychological Science, 11, 244248
Eich, E., Macaulay, D., & Lam, R. W. (1997). Mania, depression, and mood dependent memory. Cognition & Emotion, 11, 607618.
Elliott, C. L., & Greene, R. L. (1992). Clinical depression and implicit memory. Journal of Abnormal Psychology, 101, 572574.
Ellis, H. C., & Ashbrook, P. W. (1988). Resource allocation model of the eects of
depressed mood states on memory. In K. Fiedler & J. Forgas (Eds.), Aect, cognition and social behavior (pp. 2543). Toronto: Hogrefe.
Ellis, H. C., Thomas, R. L., & Rodriguez, I. A. (1984). Emotional mood states and
memory: Elaborative encoding, semantic processing, and cognitive eort.
Journal of Experimental Psychology: Learning, Memory, and Cognition, 10, 470
482.
Fennell, M. J. V., Teasdale, J. D., Jones, S., & Damle, A. (1987). Distraction in neurotic and endogenous depression: An investigation of negative thinking in
major depressive disorder. Psychological Medicine, 17, 441452.
Fox, E., Russo, R., Bowles, R., & Dutton, K. (2001). Do threatening stimuli draw or
hold attention in subclinical anxiety? Journal of Experimental Psychology: General, 130, 681700.
Gotlib, I. H., & McCann, C. D. (1984). Construct accessibility and depression: An
examination of cognitive and aective factors. Journal of Personality and Social
Psychology, 47, 427439.
Haaga, D. A., Dyck, M. J., & Ernst, D. (1991). Empirical status of the cognitive theory
of depression. Psychological Bulletin, 110, 215236.
Hartlage, S., Alloy, L. B., Vazquez, C., & Dykman, B. (1993). Automatic and eortful
processing in depression. Psychological Bulletin, 113, 247278.
Hasher, L., & Zacks, R. T. (1979). Automatic and eortful processes in memory. Journal of Experimental Psychology: General, 108, 356388.
Hasher, L., & Zacks, R. T. (1988). Working memory, comprehension, and aging: A
review and a new view. In G. H. Bower (Ed.), The psychology of learning and
motivation (Vol. 22, pp. 193225). San Diego, CA: Academic Press.
Healy, H., & Williams, J. M. G. (1999). Autobiographical memory. In T. Dalgleish &
M. Power (Eds.), The handbook of cognition and emotion (pp. 229242).
Chichester: Wiley.
Heller, W., & Nitschke, J. B. (1998). The puzzle of regional brain activity in depression and anxiety: The importance of subtypes and comorbidity. Cognition &
Emotion, 12, 421447.
Hertel, P. T. (1994). Depressive decits in word identication and recall. Cognition
& Emotion, 8, 313327.
Hertel, P. T. (1997). On the contributions of decient cognitive control to memory
impairments in depression. Cognition & Emotion, 11, 569584.
Hertel, P. T. (1998). The relationship between rumination and impaired memory
in dysphoric moods. Journal of Abnormal Psychology, 107, 166172.
Hertel, P. T. (2000). The cognitive-initiative account of depression-related impairments in memory. In D. Medin (Ed.), The psychology of learning and motivation
(Vol. 39, pp. 4771). New York: Academic Press.

212

Hertel, P. T., & Gerstle, M. (in press). Depression-related decits in forgetting. Psychological Science.
Hertel, P. T., & Hardin, T. S. (1990). Remembering with and without awareness in
a depressed mood: Evidence of decits in initiative. Journal of Experimental Psychology: General, 119, 4559.
Hertel, P. T., & Knoedler, A. J. (1996). Solving problems by analogy: The benets
and detriments of hints and depressed moods. Memory & Cognition, 24, 1625.
Hertel, P. T., Mathews, A., Peterson, S., & Kintner, K. (in press). Transfer of training emotionally biases interpretations. Applied Cognitive Psychology.
Hertel, P. T., & Meiser, T. (2000). Capacity and procedural accounts of impaired
memory in depression. In U. von Hecker, S. Dutke, & G. Sedek (Eds.), Generative mental processes and cognitive resources: Integrative research on adaptation
and control (pp. 283308). Dordrecht, The Netherlands: Kluwer Academic
Publishers.
Hertel, P. T., & Milan, S. (1994). Depressive decits in recognition: Dissociation of
recollection and familiarity. Journal of Abnormal Psychology, 103, 736742.
Hertel, P. T., & Parks, C. (2002). Emotional episodes facilitate word recall. Cognition & Emotion, 16, 685694.
Hertel, P. T., & Rude, S. S. (1991a). Depressive decits in memory: Focusing attention improves subsequent recall. Journal of Experimental Psychology: General,
120, 301309.
Hertel, P. T., & Rude, S. S. (1991b). Recalling in a state of natural or induced depression. Cognitive Therapy and Research, 15, 103127.
Hull, C. L. (1943). Principles of behavior. New York: Appleton-Century-Crofts.
Ingram, R. E. (1990). Self-focused attention in clinical disorders: Review and a conceptual model. Psychological Bulletin, 107, 156176.
Jacoby, L. L. (1991). A process dissociation framework: Separating automatic from
intentional uses of memory. Journal of Memory and Language, 30, 513541.
Jacoby, L. L., Jennings, J. M., & Hay, J. F. (1996). Dissociating automatic and consciously controlled processes: Implications for diagnosis and rehabilitation of
memory decits. In D. Herrmann, C. McEvoy, C. Herzog, P. T. Hertel, & M. K.
Johnson (Eds.), Basic and applied memory research: Theory in context (pp. 161
194). Mahwah, NJ: Erlbaum.
Jacoby, L. L., & Kelley, C. M. (1987). Unconscious inuences of memory for a prior
event. Personality and Social Psychology Bulletin, 13, 314336.
Jenkins, W., & McDowall, J. (2001). Implicit memory and depression: An analysis
of perceptual and conceptual processes. Cognition & Emotion, 15, 803812.
Joormann, J. (in press). Attentional bias in dysphoria: The role of inhibitory processes. Cognition & Emotion.
Kwiatkowski, S. J., & Parkinson, S. R. (1994). Depression, elaboration, and mood
congruence: Dierences between natural and induced mood. Memory & Cognition, 22, 225233.
Lawson, C., MacLeod, C., & Hammond, G. (2002). Interpretation revealed in the
blink of an eye: Depressive bias in the resolution of ambiguity. Journal of Abnormal Psychology, 111, 321328.
Lyubomirsky, S., Caldwell, N. D., & Nolen-Hoeksema, S. (1998). Eects of rumina-

213

tive and distracting responses to depressed mood on retrieval of autobiographical memories. Journal of Personality and Social Psychology, 75, 166177.
Matt, G. E., Vazquez, C., & Campbell, W. K. (1992). Mood-congruent recall of
aectively toned stimuli: A meta-analytic review. Clinical Psychology Review,
12, 227255.
McCabe, S. B., & Gotlib, I. H. (1995). Selective attention and clinical depression:
Performance on a deployment-of-attention task. Journal of Abnormal Psychology, 104, 241245.
McCabe, S. B., & Toman, P. E. (2000). Stimulus exposure duration in a deploymentof-attention task: Eects on dysphoric, recently dysphoric, and nondysphoric
individuals. Cognition & Emotion, 14, 125142.
McNally, R. J., Metzger, L. J., Lasko, N. B., Clancy, S. A., & Pitman, R. K. (1998).
Directed forgetting of trauma cues in adult survivors of childhood sexual abuse
with and without posttraumatic stress disorder. Journal of Abnormal Psychology, 107, 596601.
Morris, C. D., Bransford, J. P., & Franks, J. J. (1977). Levels of processing versus transfer-appropriate processing. Journal of Verbal Learning and Verbal Behavior, 16,
519533.
Murray, L. A., Whitehouse, W. G., & Alloy, L. B. (1999). Mood congruence and depressive decits in memory: A forced-recall analysis. Memory, 7, 175196.
Neill, W. T., & Mathis, K. M. (1998). Transfer-inappropriate processing: Negative
priming and related phenomena. In D. Medin (Ed.), The psychology of learning
and motivation (Vol. 38, pp. 144). New York: Academic Press.
Nolen-Hoeksema, S. (1991). Responses to depression and their eects on the duration of depressive episodes. Journal of Abnormal Psychology, 100, 569582.
Nolen-Hoeksema, S. (2000). The role of rumination in depressive disorders and mixed
anxiety/depressive symptoms. Journal of Abnormal Psychology, 109, 504511.
Nolen-Hoeksema, S., & Morrow, J. (1993). The eects of rumination and distraction on naturally occurring depressed moods. Cognition & Emotion, 7, 561570.
Parrott, W. G., & Hertel, P. T. (1999). Research methods in cognition and emotion.
In T. Dalgleish & M. Power (Eds.), The handbook of cognition and emotion (pp. 61
81). Chichester: Wiley.
Parrott, W. G., & Spackman, M. P. (2000). Emotion and memory. In M. Lewis &
J. Haviland-Jones (Eds.), Handbook of emotions (2nd ed., pp. 476490). New
York: Guilford.
Peale, N. V. (1956). The power of positive thinking. Englewood Clis, NJ.
Perrig, W. J., & Perrig, P. (1988). Mood and memory: Mood-congruity eects in
absence of mood. Memory & Cognition, 16, 102109.
Polyani, M. (1958). Personal knowledge: Toward a post-critical philosophy. Chicago:
University of Chicago Press.
Posner, M. I., & Snyder, C. R. R. (1975). Attention and cognitive control. In R. L.
Solso (Ed.), Information processing and cognition: The Loyola Symposium (pp. 55
85). Hillsdale, NJ: Erlbaum.
Powers, M. J., Dalgleish, T., Claudio, V., Tata, P., & Kentish, J. (2000). The directed
forgetting task: Application to emotionally valent material. Journal of Aective
Disorders, 57, 147157.

214

Roediger, H. L., & McDermott, K. B. (1992). Depression and implicit memory: A commentary. Journal of Abnormal Psychology, 101, 587591.
Rokke, P. D., Arnell, K. M., Koch, M. D., & Andrews, J. T. (2002). Dual-task attention decits in dysphoric mood. Journal of Abnormal Psychology, 111, 370379,
Rothkopf, J. S., & Blaney, P. H. (1991). Mood congruent memory: The role of aective focus and gender. Cognition & Emotion, 5, 5364.
Rude, S. S., Hertel, P. T., Jarrold, W., Covich, J., & Hedlund, S. (1999). Remembering to remember: A diculty for depressed individuals? Cognition & Emotion,
13, 267276.
Rude, S. S., Wenzla, R. M., Gibbs, B., Vane, J., & Whitney, T. (2002). Negative processing biases predict subsequent depressive symptoms. Cognition & Emotion,
16, 423440.
Ruiz-Caballero, J. A., & Gonzlez, P. (1994). Implicit and explicit memory bias in
depressed and nondepressed subjects. Cognition & Emotion, 8, 555569.
Rusting, C. L., & DeHart, T. (2000). Retrieving positive memories to regulate negative mood: Consequences for mood-congruent memory. Journal of Personality
and Social Psychology, 78, 737752.
Sonntag, P., Gokalsing, E., Olivier, C., Robert, P., Burglen, F., Kaumann-Muller,
F., et al. (2003). Impaired strategic regulation of contents of conscious awareness in schizophrenia. Consciousness and Cognition, 12, 190200.
Teasdale, J. D. (1983). Negative thinking in depression: Cause, eect, or reciprocal
relationship? Advances in Behavior Research and Therapy, 5, 326.
Teasdale, J. D., Segal, Z. V., Williams, J. M. G., Ridgeway, V. A., Soulsby, J. M., & Lau,
M. A. (2000). Prevention of relapse/recurrence in major depression by mindfulness-based cognitive therapy. Journal of Consulting and Clinical Psychology,
68, 615623.
Tyler, S. W., Hertel, P. T., McCallum, M. C., & Ellis, H. C. (1979). Cognitive eort
and memory. Journal of Experimental Psychology: Human Learning and Memory,
5, 607617.
Von Hecker, U., Conway, M., Meiser, T., & Holm, S. (2002). Inhibition decits in dysphoric mood. Manuscript submitted for publication.
Von Hippel, W., Hawkins, C., & Narayan, S. (1994). Personality and perceptual
expertise: Individual dierences in perceptual identication. Psychological Science, 5, 401406.
Watkins, E., & Teasdale, J. D. (2001). Rumination and overgeneral memory in depression: Eects of self-focus and analytic thinking. Journal of Abnormal Psychology, 110, 333357.
Watkins, P. C., Martin, C. K., & Stern, L. D. (2000). Unconscious memory bias in
depression: Perceptual and conceptual processes. Journal of Abnormal Psychology, 109, 282289.
Watkins, P. C., Mathews, A., Williamson, D. A., & Fuller, R. D. (1992). Mood-congruent memory in depression: Emotional priming or elaboration? Journal of
Abnormal Psychology, 101, 581586.
Watkins, P. C., Vache, K., Verney, S. P., Muller, S., & Mathews, A. (1996). Unconscious mood-congruent memory bias in depression. Journal of Abnormal Psychology, 105, 3441.

215

Watts, F. N., MacLeod, A. K., & Morris, L. (1988). Association between phenomenal and objective measures of concentration problems in depressed patients.
British Journal of Psychology, 79, 241250.
Weingartner, H., Cohen, R. M., Murphy, D. L., Martello, J., & Gerdt, C. (1981). Cognitive processes in depression. Archives of General Psychiatry, 38, 4247.
Weingartner, H., Miller, H., & Murphy, D. L. (1977). Mood-state-dependent retrieval
of verbal associations. Journal of Abnormal Psychology, 86, 276284.
Wegner, D. M. (1994). Ironic processes of mental control. Psychological Review, 101,
3452.
Wenzla, R. M., & Bates, D. E. (1998). Unmasking a cognitive vulnerability to depression: How lapses in mental control reveal depressive thinking. Journal of
Personality and Social Psychology, 75, 15591571.
Wenzla, R. M., Meier, J., & Salas D. M. (2002). Thought suppression and memory
biases during and after depressive moods. Cognition & Emotion, 16, 403422.
Wenzla, R. M., Rude, S. S., Taylor, C. J., Stultz, C. H., & Sweatt, R. A. (2001). Beneath the veil of thought suppression: Attentional bias and depression risk.
Cognition & Emotion, 15, 435452.
Wessel, I., Merckelbach, H., & Dekkers, T. (2002). Autobiographical memory specificity, intrusive memory, and general memory skills in Dutch-Indonesian survivors of the World War II era. Journal of Traumatic Stress, 15, 227234.
Williams, J. M. G., Teasdale, J. D., Segal, Z. V., & Soulsby, J. (2000). Mindfulnessbased cognitive therapy reduces overgeneral autobiographical memory in formerly depressed patients. Journal of Abnormal Psychology, 109, 150155.
Williams, J. M. G., Watts, F. N., MacLeod, C., & Mathews, A. (1988). Cognitive psychology and emotional disorders. New York: Wiley.

216

, ,

- , , ,

he conceptualization of schizophrenia has changed radically over

the past decade. Cognitive decits are now viewed as core symptoms and major disabilities of the disease (Green, 1996); virtually all cognitive functions are considered defective (Aleman, Hijman, de Haan, & Kahn, 1999;
Heinrichs & Zakzanis, 1998). Among these decits, memory impairments likely
play a crucial role. Not all cognitive abilities are equally aected; several studies have shown that memory functions are disproportionately impaired (e.g.,
Bilder et al., 2000; Gold, Randolph, Carpenter, Goldberg, & Weinberger, 1992;
Heinrichs & Zakzanis, 1998; Palmer et al., 1997; Saykin et al., 1991, 1994).
Moreover, memory impairments are strongly correlated with disturbed functioning in everyday life (Green, 1996), and a large part of clinical work (e.g.,
collecting historical information; conducting psychodynamic, behavioral, or
cognitive therapies; ensuring medication compliance) relies on the ability of
the patient to learn and remember.
In normal subjects, the evidence that memory is strongly inuenced
by emotions is overwhelming (e.g., Blaney, 1986; Bower, 1981; Conway &
Pleydell-Pearce, 2000). Memory for emotional events is better than memory for
neutral events, a phenomenon known as the emotionality eect (e.g., Dutta &
Kanungo, 1975; Rubin & Friendly, 1986). Moreover, memory for positive events
is often better than memory for negative eventsa dierence attributed to the
Pollyanna tendency (Boucher & Osgood, 1969), a disposition for normal subjects to construct an aectively positive world view. Finally, normal subjects
exhibit better memory for events whose emotional valence is congruent with
their current mood (mood-congruent memory). How do memory and emotion
interact in patients with schizophrenia? Do these patients exhibit a particular
diculty in remembering emotional events? Do they still show the Pollyanna
tendency? Is there any impairment of the subjective experience of retrieving
217

emotional memories? Because emotional disturbances, which interfere dramatically with behavior control and with social interaction, are another major aspect of schizophrenia (Taylor & Liberzon, 1999), the answers to these questions
are crucial for the development of cognitive models of schizophrenia that take
emotions into account. They may help us identify the mechanisms underlying
the formation and the persistence of poorly understood symptoms comprising a
strong emotional component, such as delusional experience. In this chapter, we
begin by reviewing the studies of memory impairment and emotion disturbances
in schizophrenia. Then we discuss experimental investigations of how memory
and emotion interact in schizophrenia. Because conscious awareness, which
plays a crucial role both in memory and emotion, may represent the fundamental impairment in schizophrenia (Andreasen, 1999; Danion, Huron, & Robert,
2001; Frith, 1992), we emphasize not only objective accuracy of memory but
also states of awareness associated with emotional memories.

Memory Impairment in

Schizophrenia

The Dissociation of Performance in

Explicit and Implicit Memory Tasks

The long-term memory impairment associated with schizophrenia is selective.

Whether patients with schizophrenia display a decit of performance depends
on the task. Several studies have shown that patients display normal or quasinormal performance in implicit memory tasksthat is, when participants are
not asked to retrieve material consciously. This pattern has been demonstrated
in tasks exploring perceptual priming (Clare, McKenna, Mortimer, & Baddeley,
1993; Doniger, Silipo, Rabinowicz, Snodgrass, & Javitt, 2001; Gras-Vincendon
et al., 1994), conceptual priming (Schwartz & Winstead, 1985), and procedural
memory (Clare et al., 1993; Dominey & Georgie, 1997; Michel, Danion, Grang,
& Sandner, 1998; Schwartz, Rosse, Veazey, & Deutsch, 1996). It has also
been demonstrated in implicit learning tasks (Danion, Meulemans, KaumannMuller, & Vermaat, 2001). In accordance with these ndings, Kazs et al. (1999),
who used the process dissociation procedure (Jacoby, 1991) to obtain uncontaminated estimates of memory processes, found that automatic inuences
of memory, but not consciously controlled uses of memory, were intact in
schizophrenia.
In contrast to normal performance in implicit memory tasks, impaired performance is typically observed in explicit or episodic memory tasks. This impairment
has been consistently demonstrated in free-recall (Koh & Peterson, 1978), cuedrecall (Schwartz, Rosse, & Deutsch, 1993), and, to a lesser degree, recognition tasks
(Aleman et al., 1999; Calev, 1984a, 1984b). Episodic memory impairment

218

occurs for verbal material, such as words, sentences (Koh & Peterson, 1978), and
stories (Abbruzzese & Scarone, 1993); visual material, such as faces (Addington
& Addington, 1998); and contextual information, such as space and time (Rizzo,
Danion, Van der Linden, & Grang, 1996; Rizzo, Danion, Van Der Linden,
Grang, & Rohmer, 1996). Autobiographical memories (memories for personal
events and facts from ones life) are also aected. Patients with schizophrenia
display an abnormal pattern of memories when theirs are compared to those of
healthy individuals: the most severe impairment occurs for early adulthood
memories, childhood memories are quasi-normal, and recent past memories are
decient to a lesser extent (Feinstein, Goldberg, Nowlin, & Weinberger, 1998).
Finally, despite diculties with explicit memory, patients with schizophrenia
remain able to learn (Goldberg, Weinberger, Pliskin, Berman, & Podd, 1989;
Hawkins, 1999), although their learning curve is lower than normal (Aleman
et al., 1999).

Recognition Memory

and Associated Subjective States

of Conscious Awareness

Most of the memory investigations carried out in normal controls and patients with
schizophrenia primarily concern the objective accuracy of recall and recognition
of stimuli. However, what often matters most is not the objective fact that we can
say an event occurred but what it feels like and means to us to remember what
took place. Recently, in recognition of the importance of subjective experience,
many researchers have shifted their focus from questions about the absolute accuracy of memories and toward questions about the states of awareness accompanying memories. The notion of conscious awareness is pivotal to the distinction
between explicit and implicit tasks. Therefore, we can infer from the pattern of
learning and memory dysfunction observed in schizophrenia that memory is impaired when conscious awareness is required at encoding or at retrieval and is
spared when conscious awareness is not required. In other words, conscious awareness might be the critical determinant of memory impairment both at encoding
and at retrieval. However, this inference can be questioned because explicit and
implicit memory tasks assess conscious awareness only indirectly, from an objective, or third-person, perspective. To obtain direct evidence of impaired conscious
awareness, we should investigate the experiences of an individual from a subjective, or rst-person, perspective.
Tulving (1985) was the rst to propose a rst-person approach to conscious
awareness associated with recognition memory. According to Tulving, conscious
awareness associated with recognition memory is not a unitary phenomenon. It
includes at least two distinct subjective states that may be investigated experimentally using the Remember/Know procedure (R/K; Gardiner, Java, & RichardsonKlavehn, 1996; Tulving, 1985). In a recognition task, the participants are asked

, ,

219

to report their subjective state of awareness at the time they recognize each individual item. They are instructed to make a Remember response if they consciously recollect something they experienced when they learned the item, that
is, if they relive mentally the learning episode. This qualitatively rich mental
experience includes perceptual, spatial, temporal, semantic, and emotional details that are attributed to a past event (Johnson, Hashtroudi, & Lindsay, 1993;
Johnson & Raye, 1981). The participants are instructed to make a Know response
if recognition is accompanied by feelings of familiarity yet no specic memories
for the learning episode. Many experiments carried out with normal controls and
the R/K procedure have shown functional dissociations between Remember and
Know responses; these ndings indicate that the two responses index two
distinct states of conscious awareness (Gardiner et al., 1996; Gardiner & RichardsonKlavehn, 2000; Rajaram & Roediger, 1997; Yonelinas, Kroll, Dobbins, Lazzara, &
Knight, 1998). Remember responses are inuenced by variables that increase
the salience of the events, or make them more distinctive: For instance, a deeper
level of processing enhances the frequency of Remember, but not Know, responses (Gardiner, 1988). Know responses, on the other hand, depend on the
ease with which an item can be processed either perceptually or conceptually:
Repetition test priming increases Know responses, without inuencing Remember responses (Rajaram, 1993). Moreover, neuropsychological data from
brain-lesioned patients and brain imaging studies carried out in normal controls
indicate that Remember and Know responses rely on dierent neural correlates
(Duzel, Yonelinas, Mangun, Heinze, & Tulving, 1997; Henson, Rugg, Shallice,
Josephs, & Dolan, 1999; Smith, 1993). Using the R/K procedure, Huron et al.
(1995; also see Huron, Danion, Rizzo, Killofer, & Damiens, in press) and Danion,
Rizzo, and Bruant (1999) studied patients with schizophrenia. They showed that
recognition performance of patients was associated with low levels of Remember, but not Know, responses, as compared to performance of normal controls.
Impairment of conscious recollection was attributed to a failure of strategic processes at encoding, at retrieval, or at both points.
Episodic memories are not a literal reproduction of the past but instead depend on constructive and reconstructive processes that are sometimes prone to
errors, distortions, and illusions (e.g., Schacter, Norman, & Koutstaal, 1998).
Recently developed paradigms for investigating false memories (memories for
events that never happened) have been used to address issues in constructive
memory. Using both the R/K procedure and the procedure initially introduced
by Deese (1959) and subsequently modied by Roediger and McDermott (1995)
to investigate false memories, Huron and Danion (2002) showed that the frequency of false memories was lower in patients with schizophrenia than in normal controls. These low levels of false memories were associated with a reduced
frequency of Remember, but not Know, responses. Danion, Gokalsing, Robert,
Massin-Krauss, and Bacon (2001) also showed that schizophrenia impairs the
relationship between subjective experience and behavior, suggesting that pa-

220

tients behavior is less determined by subjective experience, and less controlled

by intentions, than that of normal controls. This is reminiscent of the traditional
view that schizophrenia is characterized by a dissociation between thought and
action (Bleuler, 1911). Because conscious recollection provides a basis for decision and action and a foundation for social relationships, these ndings may have
important clinical implications. Reduced conscious recollection leads to poor
ability to apprehend subjective experiences in all their richness and may account
for many behavioral disturbances exhibited by patients with schizophrenia
(Danion, Huron, et al., 2001).

Emotional Disturbances

in Schizophrenia

According to a number of authors (e.g., Damasio, 1995; Ekman, 1993; Lazarus,

1991; Mandler, 1984; Zajonc, 1984), emotions are best conceptualized as a concatenation of cognitive evaluation with changes in body and brain state, arising in response to particular events or mental images. As the consequences of
how people construe situations, emotions dier according to the complexity of
the cognitive evaluation from which they stem. In some situations, this evaluation may be rudimentary, whereas in others the full complex emotion can develop only with substantial reasoning about the situation and its implications.
Body state changes correspond to physiological modications, many perceptible
to an external observer; examples include changes in skin color, facial expression, or skin conductance. Brain changes correspond to the activation of specic
cerebral regions and produce cognitive processing modications; some of these
modicationssuch as the rate at which images are produced or the uency with
which aective stimuli are processedare perceptible to the individual in whom
they are enacted. In addition to these evaluative and physiological components,
emotions may involve consciousness when the subject becomes aware of the link
between emotion and the causal event. The implication of conscious awareness
and its experiential dimension explains why, in humans, emotions are not only
closely related to adaptive behavior at an automated level but also oer the
exibility of response based on memory for emotional events from the personal
past. A primary function of emotions is to make thought, action, and social interaction more eective (Lazarus, 1991).

Inappropriate or Flat Aect

Emotional disturbances such as inappropriate or at aect are a cardinal symptom of schizophrenia. An inappropriate aect is one that is discordant with the
content of the patients speech or ideation; sudden and unpredictable changes
in aect may also occur. In at aect, there is little or no aective expression in

, ,

221

general and a reduced range of facial expression in particular. The mechanism

underlying this symptom has long been controversial. On the one hand, at aect
has been taken as evidence that the primary emotional disturbance in schizophrenia is an inability to experience pleasure, that is, anhedonia (e.g., Rado,
1962). According to this view, the absence of pleasure is a central feature of
schizophrenia; hence, the diminished expression of positive emotions is secondary to anhedonia. On the other hand, classical authors such as Bleuler (1911)
noticed that, although patients often report that they experience strong emotions, observers do not always conrm their reports. He believed that patients
do experience emotions but that expression is blunted. Experimental evidence
relevant to this debate has been obtained from laboratory studies. A disjunction
between expressed and subjectively experienced emotions has been shown using emotion-eliciting lms: patients with schizophrenia experience as many
positive and negative emotions as normal controls, but they are less facially expressive of both positive and negative emotions (Kring, Kerr, Smith, & Neale,
1993; Kring & Neale, 1996).
Myin-Germeys, Delespaul, and deVries (2000) recently assessed emotional
experience of patients with schizophrenia and normal controls in daily life. They
used a structured time-sampling technique to collect the data. Participants received a digital wristwatch and a set of assessment forms. Ten times a day on
6 consecutive days, the watch emitted a beep at unpredictable moments between
7:30 AM and 10:30 PM. After every beep, reports of thoughts, current activity,
moods, and severity of symptoms were collected. The assessment forms contained
10 Likert-type items about positive and negative emotions. Patients were divided
into blunted and nonblunted subgroups on the basis of Brief Psychiatric Rating
Scale behavioral ratings of at aect. Patients with schizophrenia experienced
more intense and more variable negative emotions than normal controls did. In
contrast, they experienced less intense and less variable positive emotions. This
nding, at variance with evidence from laboratory studies that patients experience the same intensities of positive emotions as normal controls do, may be
explained by the absence of social context in laboratory studies. Patients with
schizophrenia may indeed be capable of experiencing the same positive emotions
as normal controls but, in the naturalistic context of daily life, engage less often
in situations likely to elicit positive experiences. These patterns of emotion were
observed whether the patients were emotionally blunted or not. In another study
that used the same structured time-sampling technique, Myin-Germeys, van Os,
Schwartz, Stone, and Delespaul (2001) provided additional evidence that patients with psychotic illness reacted with more intense negative emotions to subjective appraisals of stress in daily life than did controls. In summary, these results
bear out the view that patients with schizophrenia are more emotionally active
and responsive than what is usually assumed on the basis of behavioral observations and facial modications perceptible to an external observer. In some circumstances, patients may experience more intense negative emotions and,

222

possibly, less intense positive emotions than controls do. Therefore, the identication of the circumstances in which patients with schizophrenia do experience
emotions like controls and those in which they do not is a prerequisite for the
investigation of how memory and emotion interact.

Cognitive Evaluation

of Emotional Events

A large body of experimental evidence shows that patients with schizophrenia

fail to recognize emotion in others (see the review by Bryson, Bell, & Lysaker,
1997; Walker, Marwit, & Emory, 1980). Whether this decit reects specic disturbances or is part of a more generalized impairment, such as, for instance,
attentional, executive or evaluative functions, is a matter of debate. Kerr and Neal
(1993) have shown that patients with schizophrenia do not exhibit greater impairment in aect recognition than in nonemotional perception tasks. However,
evidence from two studies that varied the emotional valence of words in working memory tasks suggests that patients with schizophrenia may still be sensitive to the emotional salience of stimuli. Koh, Szoc, and Peterson (1977) used
pleasant, unpleasant, and neutral words in a Sternberg task.1 Despite their overall poorer performance, patients, like normal controls, were faster for neutral
words than for pleasant and unpleasant words. Moreover, in both groups, there
was no dierence between pleasant and unpleasant words. Kay (1982) used a
modication of the Brown-Peterson task (Wickens, 1970) to investigate
whether patients with schizophrenia were sensitive to the emotional salience
of to-be-learned words. Participants were presented with three trials involving word triads (e.g., love, excellent, rest) drawn from a single aective valence
(e.g., positive) and then switched to the opposite aective valence on the fourth
trial (e.g., negative). After each trial, participants performed a distractor task
over a short retention interval before recalling the word triads. Patients with
schizophrenia performed progressively worse across trials but returned to the
level of the rst trial for the fourth trial, namely, when the aective valence
shifted. This increase in performance induced by the shift of aective valence
indicates a release of proactive interference, suggesting that patients with
schizophrenia were still sensitive to the aective salience of the to-be-learned
words. This nding is consistent with evidence from studies conducted in normal controls showing that words elicit an emotional evaluation automatically
or with minimal strategic processes and conscious cognitive eort (review in
Ochsner, 2000). In summary, these ndings suggest that patients with schizophrenia exhibit defective emotional processing when a highly complex or very
specic cognitive evaluation of the emotional stimuli is required, such as recognizing, discriminating, and interpreting emotions in others. In contrast,
patients remain sensitive to the emotional valence of stimuli when cognitive
evaluation is more automatic.

, ,

223

Body State Changes

Body state changes arising in response to emotional events have been widely
investigated in schizophrenia with measurements of skin conductance. Forty to
50% of patients with schizophrenia exhibit a reduced phasic response of skin
conductance to orienting stimuli (review in Bernstein, 1987). But patients may
also exhibit a tonic hyperactivity of skin conductance (Dawson, Nuechterlein,
Schell, Gitlin, & Ventura, 1994). Finally, in the study using emotion-eliciting
lms, Kring and Neale (1996) showed that, compared to normal controls, patients with schizophrenia evidenced greater reactivity of skin conductance to
positive and negative lms. Yet they also exhibited greater reactivity to neutral
lms than controls did. Thus, while they were reporting comparable levels of
subjectively experienced emotions, the patients were more electrodermally
aroused. The pattern of ndings provided by these studies is complex: the level
of physiological responses to emotional events is sometimes lower and sometimes
higher in patients with schizophrenia than in normal controls. It suggests that
schizophrenia is associated with a dysregulation of body responses, rather than
a reduced reactivity to emotional events (Taylor & Liberzon, 1999).

Brain State Changes

Advances in functional brain imaging have recently made it possible to investigate the functional brain correlates of mood changes in schizophrenia. Schneider
et al. (1998) used functional magnetic resonance imaging to examine brain activity in patients with schizophrenia and normal controls during happy and sad
mood induction. To induce happy and sad mood, researchers presented the participants with stimuli that consisted of happy and sad facial expressions. They
were required to look at each face and use it to help them to feel happy or sad. In
line with previous ndings, results show brain activity in the amygdala of normal controls during negative aect. Unlike controls, patients with schizophrenia did not demonstrate amygdala activation during sadness, despite similar
subjective ratings of negative aect. Using morphological procedures, researchers have reported structural abnormalities of the amygdala in schizophrenia
(Bogerts, 1993). The absence of amygdala activation is consistent with the hypothesis that the inappropriate behavior displayed by patients with schizophrenia in an emotional context is related to a dysregulation of dopamine systems
resulting from functional impairment of the amygdala (Louilot & Besson, 2000).
Interestingly, three independent studies of patients with schizophrenia have
recently shown that the response of dopamine terminals to systemic administration of amphetamine, a dopaminergic agonist, is enhanced (Abi-Dargham
et al., 1998; Breier et al., 1997; Laruelle et al., 1996). However, to the best of
our knowledge, there is no published study in which brain imaging techniques
were used to investigate in patients with schizophrenia the functional brain cor-

224

relates of emotions arising specically in responses to discrete emotional events.

In particular, in view of the dopamine hypothesis about schizophrenia and of the
role putatively played by dopamine in response to emotional stimuli, investigating the inuence of emotional stimuli on dopamine responses in patients with
schizophrenia should be revealing (Taylor & Liberzon, 1999). Moreover, because
most patients are treated with neuroleptic drugs, that is, dopamine blockers,
evaluating the inuence of treatment on the dierent components of emotion
should be worthwhile.

Implications for Research

Both clinical and experimental ndings show that schizophrenia is associated
with an impairment of dierent components of emotions, including evaluative
processes and body and brain correlates. They also show that patients
with schizophrenia often exhibit a disjunctive relationship among expressive,
subjective, and biological dimensions of emotions, indicating that the dierent
components of emotion may not be defective altogether. In particular, the impairment of the expressive and biological dimensions does not necessarily imply
a disturbance of the emotions subjectively experienced by the patient, which
seem to be preserved in some situations. Moreover, patients are still sensitive
to the emotional salience of events, provided that the cognitive evaluation of
emotional valence is relatively automatic. Whether the subjective experience of
emotion is intact depends on the type of information processing each emotional
event requires.
How emotion and memory interact in patients with schizophrenia should
depend primarily on whether the sensitivity to the emotional salience of stimuli
is intact. We can therefore conclude from these studies that emotional words,
which elicit a cognitive evaluation of the emotional valence that can take place
automatically, may represent appropriate stimuli for investigating the relationship between emotion and memory. Moreover, full-blown investigations of this
relationship should include not only memory assessments but also measurements of body and brain changes induced by emotional memories.

Interaction of Emotion and

Memory in Schizophrenia

In comparison to the numerous studies that have addressed the issue of emotional disturbances, experimental evidence pertaining to the interaction of emotion and memory in schizophrenia is rare. With the exception of one study (Koh,
Grinker, Marusarz, & Forman, 1981) that used photographs as to-be-memorized
stimuli, all others used emotional words. None of them investigated the body and
brain correlates of emotional memories. A series of studies (Kayton & Koh, 1975;

, ,

225

Koh et al., 1981; Koh, Kayton, & Peterson, 1976) compared the impact of the
emotional valence of to-be-learned words on recall performance in patients with
schizophrenia and normal controls. As a collection, these studies consistently
showed that normal controls recalled signicantly more pleasant words than
unpleasant words, therefore exhibiting a Pollyanna tendency. In contrast, the
studies did not provide consistent results in schizophrenia. Kayton and Koh used
a free-recall task involving pleasant, unpleasant, and neutral words in 17 patients with schizophrenia and 12 normal controls. The results showed an absence
of the Pollyanna tendency in schizophrenia: patients recalled pleasant words to
the same extent as unpleasant words. Moreover, in comparison with the performance of normal controls, their recall performance for pleasant words, but not
for unpleasant and neutral words, was impaired. In a second study, Koh, Kayton,
and Peterson used a set of recall tasks to investigate the recall performance of
patients with schizophrenia when they were asked to rate the pleasantness of
words during the encoding phase. This orienting task was used to make sure that
all participants encoded the aective valence of words to be recalled. The set of
recall tasks included an incidental free-recall task in which participants assessed
the pleasantness of the words but were not aware that they would have to recall
them; an intentional free-recall task, in which participants were instructed to
learn the words that they rated; and a cued recall task in which recall was cued
by dierent emotional categories (e.g., pleasant or unpleasant). For all these recall tasks, the results showed that patients, like normal controls, tended to recall pleasant words more often than unpleasant words. However, this Pollyanna
tendency was not statistically signicant in incidental free recall by the group
of patients with schizophrenia. Interestingly, across all these tasks, the recall
performance of patients with schizophrenia was not defective: they recalled as
many words as did the controls. This pattern of results led the authors to conclude that both the recall decit and the absence of a Pollyanna tendency observed in their previous study of patients with schizophrenia was remediable
when patients encoded the words according to their emotional features.
Unfortunately, numerous methodological aws marred the study by Koh
et al. (1976). For instance, the words used in the incidental free-recall task were
the same words participants had to recall in the subsequent cued-recall task.
Similarly, the words that they had to recall in the intentional free-recall task had
been used previously as distractors in a recognition task. These methodological
concerns limit the conclusions that can be drawn from these ndings. Moreover,
another study (Koh et al., 1981) using the same orienting task produced inconsistent results: Although overall recall performance was comparable in patients
with schizophrenia and normal controls, the Pollyanna tendency was absent in
patients. In a second experiment, Koh et al. investigated the interaction between
memory and emotion in schizophrenia using items characterized by a separation of target (photographs of faces) and emotional information (favorable or
unfavorable personality-trait words). They conrmed the absence of a Pollyanna

226

tendency in schizophrenia: controls recognized signicantly more favorably than

unfavorably encoded faces, but this asymmetry was absent in patients.
Finally, in a more recent study, Calev and Edelist (1993) investigated the eect
of the emotional valence of words on the rate of forgetting. They compared immediate and delayed recall of neutral, positive, and negative words in patients
with schizophrenia and normal controls. The results showed that patients, like
normal controls, forgot neutral words more rapidly than emotional words but
forgot negative words less rapidly than positive words, whereas normal controls
exhibited equivalent forgetting rates for positive and negative words. Moreover,
patients with schizophrenia recalled fewer words than normal controls in both
immediate and delayed recall tasks.
These ndings do not provide a clear picture of the impact of emotion on
memory in schizophrenia. Indeed, some studies (Kayton & Koh, 1975; Koh et al.,
1981) that revealed a defective Pollyanna tendency in schizophrenia suggest that
the memory performance of patients with schizophrenia might be less inuenced
by the emotional valence of stimuli than that of normal controls. In contrast,
Koh et al. (1976) reported an intact Pollyanna tendency that suggests that the
memory performance of patients remains inuenced by the emotional valence of
words.2 Similarly, the more rapid forgetting of positive than negative words in
schizophrenia reported by Calev and Edelist (1993) suggests that the emotional
valence of the words to-be-learned still has an inuence on patients memory.

An Investigation of the Inuence

of Emotional Words on Recall

Performance in Schizophrenia

In view of these inconsistent results, we investigated the emotionality eect in

schizophrenia using words as emotional stimuli. Because patients with schizophrenia are sensitive to the emotional salience of words (Kay, 1982; Koh et al.,
1977), we predicted that they should recall more emotional words than neutral
words, even though they might recall fewer words overall in comparison to controls. However, to investigate whether a putative impairment of the emotionality eect could be remediable, we asked each subject to complete two sessions.
In the rst session, participants were asked to remember the words, no mention
being made of their emotional component. The second session was identical to
the rst one, with the exception that it included an additional orienting task in
which we explicitly asked participants to assess the emotional dimension of words
(Koh et al., 1981; Koh et al., 1976).
Method Twenty-four clinically stabilized patients (10 men, 14 women) fullling DSM-III-R criteria for schizophrenia took part in the study. Twenty-two patients were receiving a long-term neuroleptic treatment (M = 561 mg of
chlorpromazine or equivalent, SD = 410 mg), whereas 2 patients were drug free.

, ,

227

Global psychiatric symptomatology (36.4, SD = 24.5) was assessed using the

Brief Psychiatric Rating Scale (BPRS). Positive (24.5, SD = 27.6) and negative
(38.4, SD = 31.1) symptoms were assessed using the Scale for the Assessment
of Positive Symptoms (SAPS) and the Scale for the Assessment of Negative Symptoms (SANS), respectively (Andreasen, 1982). The normal control group comprised 24 participants (10 men, 14 women). The groups did not dier
signicantly in age (31.3 years, SD = 24.5, and 31.8 years, SD = 8.0, respectively) or education (11.0 years, SD = 4.8, and 11.4 years, SD = 3.0).
Words were selected on the basis of their emotional valence (details of word
selection procedure can be found in Danion, Kaumann-Muller, Grang,
Zimmermann, & Greth, 1995; Grang, Greth, & Danion, 1991). There were two
lists of 30 words, each with 10 positive, 10 negative, and 10 neutral words, randomly selected from each of the three valence categories. Positive, negative, and
neutral words were mixed in each list. The participants completed two sessions,
separated by at least 24 hours. In the study phase of each session, the participants were exposed to the words of one list. The lists were counterbalanced across
sessions within each group.
The two sessions diered in study instructions. In the rst session, we asked
the participants to read the words aloud and remember them, no mention being
made of the emotional component of words. We attempted to equalize recall
performance for neutral words in both groups by giving more opportunity to
patients than controls to learn the words and by shortening the delay between
the study and test phases of the recall task. Normal controls were given 2 seconds per word and patients 5 seconds. In the second session, they were asked to
read the words aloud, to remember them, and to perform an orienting task that
required them to rate the words on a 100 mm visual analogue scale according
to their subjective feelings of pleasantness and unpleasantness. Normal controls
were given 5 seconds and patients 10 seconds per word. The test phases were
conducted 10 minutes after the study phases in the control group and 3 minutes after in the patients group. In each of the two test phases, the participants
were asked to write in 3 minutes as many words from the study list as possible.
Finally, at the end of the second test phase, they were asked to rate the emotional
valence of the words of the list presented in the rst study phase.
Results and Discussion For each subject, we computed the mean ratings of positive, negative, and neutral words. Patients and normal controls discriminated
among positive, negative, and neutral words to the same extent. In normal controls, the mean ratings for positive words (70.6, SD = 10.8) were higher than
those for neutral words (52.8, SD = 7.2), which were higher than those for negative words (28.3, SD = 14.3). A similar prole was observed in patients (67.5,
SD = 14.4; 56.9, SD = 10.1; 32.2, SD = 17.7, respectively). An analysis of variance (ANOVA) carried out on these ratings yielded a signicant word type eect,

228

Mean Percent Recalled

F(2, 88) = 91.2, p < .0001, but neither the group eect nor the interaction between group and word type was signicant.
The mean proportions of words recalled were subjected to an ANOVA, with
word type (positive vs. negative vs. neutral) and encoding condition as withinsubject factors and group as a between-subject factor. Figure 7.1 displays the
results pooled across the two sessions; recall performance was not aected by
the orienting task, as indicated by a lack of a signicant eect of encoding conditions or interactions (all Fs < 1). Patients recalled fewer words than controls,
F(1, 46) = 9.80, p < .005. This result indicated that the attempt to equalize
recall performance in the two groups failed, probably as a consequence of the
severity of the memory impairment that characterized the patients. There was
a signicant word-type eect, F(2, 92) = 5.98, p < .005; the mean proportion
of recalled words was signicantly higher for positive words compared to neutral words, t(47) = 3.35, p < .002, and for negative words compared to neutral words, t(47) = 2.81, p = .007. The dierence between positive and negative
words was not signicant, nor was the interaction between group and word
type. Therefore, as we predicted, while patients recall performance was poor,
both groups recalled emotional words better than neutral words. This pattern
of performance was observed under conditions in which patients and normal
controls discriminated between positive, negative, and neutral words to the
same degree. It was independent of whether emotional encoding was induced
by an orienting task forcing the participants to process the emotional valence
of words. Although the hypothesis that participants deliberately noticed valence in the rst task, regardless of instructions, cannot be excluded, this nding is consistent with the assumption that the encoding of emotional valence

Controls
Patients

25
20
15
10
5
0

Positive

Neutral

Negative

Materials
Figure 7.1. Free-recall performance as a function of affective valence of words in
patients with schizophrenia and normal participants.

, ,

229

of words does not require any controlled, strategic processes but involves more
automatic processes (e.g., Hasher & Zacks, 1979). It is also consistent with
evidence that automatic memory processes are intact in schizophrenia (GrasVincendon et al., 1994; Kazs et al., 1999).
To investigate whether aective attening (limited outward expression of
positive and negative emotions in face, voice, and gestures) blunted the emotionality eect on memory performance, we carried out a complementary ANOVA
of memory performance on two subgroups of 12 patients. We categorized these
patients according to the score (below or above the median of the group) of the
aective attening item of the SADS, a clinical scale measuring the intensity
of negative symptoms of schizophrenia. There was neither a signicant group
eect nor an interaction between group and word type. Therefore, our experiment suggests that the enhancement of memory performance for emotional
words is not attenuated by aective attening; it provides evidence of a disjunction between the expressed dimension of emotion, which is reduced, and emotions experienced at encoding and re-experienced at retrieval, which are
preserved. Discrepancies between our results and those from previous studies
suggesting that the emotionality eect on memory is disturbed in schizophrenia may be linked to their methodological dierences: dierences in the emotional
stimuli used and dierences in samples of patients. Some former studies did not
use explicit diagnostic criteria of schizophrenia.

Interaction of Emotion and

Conscious Awareness Associated

With Recognition Memory in

Normal Controls

Most of the previous investigations of the interaction of emotion and memory

carried out in normal controls and in patients with schizophrenia primarily concerned the objective accuracy of memory for emotional stimuli. However, as we
said, what often matters most is not the objective fact that an event occurred but
what it feels like and how it is re-experienced. This is particularly important for
emotional memories experienced by patients suering from psychiatric diseases,
such as schizophrenia, that are accompanied by major disturbances of emotion.
However, most of the studies carried out in normal controls with the R/K procedure used neutral material (review in Gardiner & Java, 1993). On the other hand,
studies that have addressed the issue of the subjective experience associated with
emotional memories in normal controls shared a common approach: The participants were simply asked to rate the vividness of their memories at retrieval.
These studies found that memories for emotional events are more vivid than
memories for neutral events (Ochsner, 2000).
Recently, Ochsner (2000) conducted three experiments using the R/K procedure with student participants to investigate the subjective states of aware-

230

ness accompanying recognition of emotional events. He reasoned that emotional

events possess unique attributes that make them more distinctive than neutral
events. They elicit evaluative and physiological responses not generated by neutral events, and they induce semantic processing and emotional subjective states
not evoked by neutral events. Because the distinctiveness with which stimuli are
encoded enhances the frequency of conscious recollection (Dewhurst & Conway,
1994; Huron et al., in press; Rajaram, 1993), Ochsner postulated that emotional
events should be consciously recollected more often than neutral events. Because
familiarity is thought to rely on how easily a stimulus is processed, Ochsner also
predicted that Know responses (which reect familiarity) would be more frequent
for emotional, and especially negative, stimuli than for neutral stimuli. Several
lines of evidence suggest that without prior study, emotional stimuli are processed more uently than neutral stimuli. Speeded perceptual or conceptual
access for emotionally arousing stimuli has been shown in various tasks (review
in Ochsner, 2000). Increased uency seems to be more reliably found for negative stimuli than for positive stimuli. However, by considering that nonstudied
aective stimuli seen at test should be processed more uently as well, Ochsner
reasoned that increases in familiarity for negative stimuli should be expected only
if encoding signicantly boosts uency above an already elevated baseline uency. Photographs of scenes and objects selected from the International Aective Picture System (IAPS; Lang, Greenwald, Bradley, & Hamm, 1993) were used
as emotional stimuli. In experiment 1, the participants were presented with these
photos and asked to rate each one according to its valence, arousal, and visual
complexity. Experiment 2 was a replication of experiment 1, but the participants
were not asked explicitly to attend and to rate the stimuli along these dimensions.
In experiment 3, the participants were asked to encode stimuli by judging their
subjective brightness.
Ochsner conducted two series of analyses. Considering that the subjective
states of awareness as assessed by Remember and Know responses are mutually
exclusive, he rst computed the proportions of Remember and Know responses
for old and new items as a function of word type. Computed values are measures
of discrete states of awareness. However, whether these values reect the underlying processes is a matter of debate. Initially, it was suggested that the states of
awareness captured by the R/K procedure map onto the underlying processes of
conscious recollection and familiarity (Gardiner & Java, 1993). According to this
view, an item either could be recollected, or familiar, but not both. However,
several authors have argued against such a direct mapping (e.g., Jacoby, 1991;
Yonelinas et al., 1998). They have suggested that the underlying processes of
conscious recollection and familiarity are independent, rather than mutually
exclusive. According to their view, some items can be exclusively recollected,
some can be exclusively familiar, and some can be both recollected and familiar. In other words, whether a given stimulus is consciously recollected does not
constrain whether it is familiar. Therefore, in a second series of analyses, Ochsner

, ,

231

used the independence model recently proposed by Yonelinas et al. to estimate

the respective contributions of conscious recollection and familiarity to recognition memory. Conscious recollection is an estimate of how often words were
correctly recollected relative to the number of opportunities participants had to
do so. Familiarity is derived from the participants tendencies to give a correct
Know response to old words and an incorrect Know response to new words.
The results of the three experiments conrmed most of Ochsners predictions.
Remember responses were more frequent for emotional than for neutral stimuli
and for negative than for positive stimuli; estimates of conscious recollection
followed the same pattern. However, contrary to Ochsners prediction, dierences
in familiarity were slight and nonsignicant. These ndings are consistent with
the notion that emotional events are better recalled or recognized than neutral
events, but they go further by demonstrating that greater accuracy is due primarily to better conscious recollection of emotional stimuli. The results also
support the hypothesis that emotion leads to increases in the distinctiveness with
which stimuli are encoded and re-experienced in memory later on. Moreover,
the results indicated that the emotionality eect obtained even when participants
did not explicitly evaluate the emotional valence of stimuli (Experiment 2) or
evaluated the subjective brightness of stimuli (Experiment 3). They thereby reconrmed the hypothesis that the emotionality eect does not depend on being
instructed explicitly to attend to emotional aspects of words.
Finally, the pattern of dierences observed in this study with positive and
negative photographs is opposite to the Pollyanna tendency usually observed
with positive and negative words. In keeping with Conway and Dewhurst
(1995), Ochsner suggested that a more general factor underlying the emotionality eect on conscious recollection might be the personal signicance of a stimulus and not its absolute emotional valence. This interpretation is also supported
by an investigation of autobiographical memories in normal controls that
Conway, Collins, Gathercole, and Anderson (1996) conducted. This experiment
showed that the personal signicance attributed to autobiographical events interacts with emotion to form a major determinant of conscious recollection for
autobiographical memories.

An Investigation of the Inuence of

Emotional Words on Recognition

Memory and Associated Conscious

Awareness in Schizophrenia

Ochsners results suggest that the R/K procedure is a more sensitive tool than
traditional recall and recognition tasks for investigating the relationship between
emotion and memory. The fact that emotional stimuli dierentially inuenced
conscious recollection and familiarity may explain why some previous studies
failed to nd a clear relationship between emotion and memory in normal con-

232

trols. This may also account for some of the inconsistent results obtained by investigations of the interaction of memory and emotion in schizophrenia. However, the studies of conscious awareness in patients with schizophrenia were
conducted using neutral stimuli, leaving unexplored the states of awareness that
accompany memory for emotional events. Danion, Kazs, Huron, and Karchouni
(in press) addressed this issue recently. They used the R/K procedure and the
recognition model of Yonelinas et al. (1998) to investigate the inuence of emotional words both on the subjective states of awareness accompanying recognition memory and the processes of conscious recollection and familiarity that
might underlie Remember and Know responses. During the study phase of a
recognition task, 24 patients with schizophrenia and 24 normal controls were
presented with positive, negative, and neutral words. They were asked to remember the words and to perform an orienting task that required them to rate the
words according to their subjective feelings of pleasantness and unpleasantness.
In the test phase, the participants were presented with a list of studied and
nonstudied words, and they were asked to recognize words from the study list.
In addition, they were told to make a R/K judgment. Because patients with
schizophrenia remain sensitive to the emotional salience of words, the researchers predicted that remember responses and conscious recollection should be more
frequent for emotional than for neutral events. Results conrmed this prediction.
During the study phase of the task, patients evaluated the emotional valence of
words just as normal controls did. At test, patients produced fewer Remember
responses and exhibited lower levels of conscious recollection than normal controls did. They gave more Know responses than controls but familiarity, as estimated using the model of Yonelinas et al., was slightly reduced. Their Remember
responses were more frequent for emotional than for neutral words and for positive than for negative words, with parallel variations in conscious recollection.
This pattern of responses was similar to the one observed in normal controls. The
levels of Know responses and familiarity for emotional and neutral words were
similar. Therefore, provided that the cognitive evaluation of emotional valence
is intact, patients, like controls, consciously recollected emotional words better
than neutral words.

Concluding Comments
This review of the investigations of conscious awareness in schizophrenia shows
that, whereas several determinants of conscious recollection such as word frequency (Huron et al., 1995), picture superiority (Huron & Danion, 2002), and
directed-forgetting eects (Sonntag et al., 2003) are impaired in patients with
schizophrenia, the emotionality eect is preserved when words are used. However, this review leaves numerous questions unanswered. For instance, are the
body correlates and the brain correlates of this emotionality eect similar in

, ,

233

patients with schizophrenia and in normal controls? Studies of the emotionality

eect of words in schizophrenia using psychophysiological and brain imaging
techniques would be worth carrying out to answer this question. Are patients
with schizophrenia still able to re-experience at retrieval the specic attributes
that make an event emotional at encoding, such as particular emotional details
or aspects of the event, components of the cognitive evaluation of the emotional
valence, and physiological changes? One way to investigate this issue might be
to use the R/K procedure to evaluate the subjective states of awareness that characterize the emotional component of the event. Another, complementary way
to investigate this issue might be to compare the qualitative characteristics of
the subjective experiences associated with emotional memories in patients with
schizophrenia and in normal controls. The Memory Characteristics Questionnaire (MCQ), designed by Johnson, Foley, Suengas, and Raye (1988) to assess
attributes of memories, could be used to achieve this goal. The MCQ asks the
participants to rate the specic qualitative characteristics of their subjective experiences, such as visual details, mental images, and emotional reactions. These
approaches to experiential aspects of emotions should provide information about
not only the inuence of emotional events on the frequency of conscious recollection in patients with schizophrenia but also the ability of patients to relive
what makes this event distinctive, its emotional dimension.
It is plausible that the relationship between emotion, memory, and conscious awareness depends on the type of information processing required by
the emotional items at encoding and at retrieval. For instance, Koh et al. (1981)
failed to observe a Pollyanna tendency, as noted earlier; the experiment used
items characterized by a separation of target (photographs of faces) and emotional information (favorable or unfavorable personality-trait words). This
failure may be related to a defective binding of target and emotional information. If so, whether this defective binding is specic to emotional information
is not clear. Instead, it could simply be a special case of a generalized defect in
binding separate dimensions of items. Such a decit has been described in
schizophrenia for items that comprise several spatially, temporally, or
cognitively separated components (Danion et al., 1999; Rizzo, Danion, Van der
Linden, & Grang, 1996; Rizzo, Danion, Van Der Linden, Grang, & Rohmer,
1996). These results raise the issue of the inuence of emotion on conscious
recollection in schizophrenia when emotional stimuli other than words are
usedsuch as stimuli that induce stronger emotional responses; stimuli that
are associated with primary, innate emotions (instead of secondary, acquired
emotions; e.g., Damasio, 1995); or stimuli that require a more controlled, strategic evaluation of emotional valence. Finally, patients with schizophrenia
often suer from depressive symptoms, and Brebion, Gorman, Malaspina,
Sharif, and Amador (2001) have suggested that the decit of verbal memory
observed in schizophrenia may be related to these symptoms. It would therefore be worthwhile to investigate whether schizophrenic patients suering from

234

depressive symptoms exhibit a mood congruency eect, that is, better memory
performance for negative than for positive stimuli [see chapter 6Eds.].
All the investigations of emotion and memory in schizophrenia reviewed in
this chapter shared a primary concern with episodic memory, as assessed in the
laboratory using experimental procedures far removed from real-life situations.
A crucial issue, albeit virtually unexplored in schizophrenia, is the relationship
between emotion and autobiographical memories, that is, memories for personal
events and facts from ones life. There is converging evidence that emotional
experience at the time of an event can inuence memory and associated conscious recollection for that event (Conway et al., 1996). Because autobiographical memories are emotional in essence and act as organizers of autobiographical
knowledge, a crucial function of the interaction of emotion and autobiographical memory is the development of the self and of personal identity (Conway &
Pleydell-Pearce, 2000). Investigating the interaction of emotion and autobiographical memories may thus provide insights into the current abnormalities
of personal identity, as well as the formation of the abnormal personal identity
that characterizes schizophrenia. Abnormalities in personal identity emerge at
the onset of schizophrenia, which occurs in adolescence and early adulthood,
that is, when autobiographical knowledge is being acquired and organized to
form personal identity.
What are the clinical implications of the investigations of emotion and
memory in schizophrenia? Evidence that the emotionality eect on memory and
subjective experience may be preserved in patients with schizophrenia has positive clinical implications. Emotions are closely related to adaptive behavior at
an automated level. But because emotions enhance conscious recollection, they
also oer the exibility of response based on memory for emotional events from
the subjects personal past. Evidence of a preserved emotionality eect therefore
suggests that patients may still benet from such exibility, at least in some situations. Other clinical implications may be negative. As emotion is the consequence of how people construe situations, some neutral events or experiences
may become emotional by virtue of false beliefs or delusions. Because these events
will be more easily recollectable later and, hence, more richly and vividly experienced in memory, they could contribute to the persistence and the enrichment
of delusional experiences. According to this view, symptoms of schizophrenia
such as delusions may arise as a consequence of a combination of impaired and
spared aspects of emotions and cognition.

Notes
1. In a typical Sternberg task, subjects are presented with a short list of items
(memory set) to remember. After a short delay (retention interval), they are presented with a probe item and instructed to decide as quickly as possible whether the
probe belongs to the previously presented memory set. Performance is measured by
the time to respond and the number of errors.

, ,

235

2. This conclusion implies that the nonaective characteristics (e.g., frequency,

imagery) of words to-be-learned can be considered equivalent.

References
Abbruzzese, M., & Scarone, S. (1993). Memory and attention dysfunctions in story
recall in schizophrenia: Evidence of a possible frontal malfunctioning. Biological Psychology, 35, 5158.
Abi-Dargham, A., Gil, R., Krystal, J., Baldwin, R. M., Seibyl, J. P., Bowers, M., et al.
(1998). Increased striatal dopamine transmission in schizophrenia: Conrmation in a second cohort. American Journal of Psychiatry, 155, 761767.
Addington, J., & Addington, D. (1998). Facial aect recognition and information
processing in schizophrenia and bipolar disorder. Schizophrenia Research, 32,
171181.
Aleman, A., Hijman, R., de Haan, E. H., & Kahn, R. S. (1999). Memory impairment in
schizophrenia: A meta-analysis. American Journal of Psychiatry, 156, 13581366.
Andreasen, N. C. (1982). Negative symptoms in schizophrenia. Denition and reliability. Archives of General Psychiatry, 39, 784788.
Andreasen, N. C. (1999). A unitary model of schizophrenia. Archives of General Psychiatry, 56, 781787.
Bernstein, A. S. (1987). Orienting response research in schizophrenia: Where we
have come and where we might go. Schizophrenia Bulletin, 13, 623641.
Bilder, R. M., Goldman, R. S., Robinson, D., Reiter, G., Bell, L., Bates, J. A., et al.
(2000). Neuropsychology of rst-episode schizophrenia: Initial characterization and clinical correlates. American Journal of Psychiatry, 157, 549559.
Blaney, P. H. (1986). Aect and memory: A review. Psychological Bulletin, 99, 229
246.
Bleuler, E. (1911). Dementia Praecox ou groupe des schizophrnies. Paris: E.P.E.L.
Bogerts, B. (1993). Recent advances in the neuropathology of schizophrenia. Schizophrenia Bulletin, 19, 431445.
Boucher, J., & Osgood, C. E. (1969). The Pollyanna hypothesis. Journal of Verbal
Learning and Verbal Behavior, 8, 18.
Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129148.
Brebion, G., Gorman, J. M., Malaspina, D., Sharif, Z., & Amador, X. (2001). Clinical
and cognitive factors associated with verbal memory task performance in patients with schizophrenia. American Journal of Psychiatry, 158, 758764.
Breier, A., Su, T. P., Saunders, R., Carson, R. E., Kolachana, B. S., de Bartolomeis,
A., et al. (1997). Schizophrenia is associated with elevated amphetamine-induced synaptic dopamine concentrations: Evidence from a novel positron emission tomography method. Proceedings of the National Academy of Sciences of the
USA, 94, 25692574.
Bryson, G., Bell, M., & Lysaker, P. (1997). Aect recognition in schizophrenia: A
function of global impairment or a specic cognitive decit. Psychiatry Research,
71, 105113.
Calev, A. (1984a). Recall and recognition in chronic nondemented schizophrenics:
Use of matched tasks. Journal of Abnormal Psychology, 93, 172177.

236

Calev, A. (1984b). Recall and recognition in mildly disturbed schizophrenics: The

use of matched tasks. Psychological Medicine, 14, 425429.
Calev, A., & Edelist, S. (1993). Aect and memory in schizophrenia: Negative emotion words are forgotten less rapidly than other words by long-hospitalized
schizophrenics. Psychopathology, 26(56), 229235.
Clare, L., McKenna, P. J., Mortimer, A. M., & Baddeley, A. D. (1993). Memory in
schizophrenia: What is impaired and what is preserved? Neuropsychologia,
31(11), 12251241.
Conway, M. A., Collins, A. F., Gathercole, S. E., & Anderson, S. J. (1996). Recollections of true and false autobiographical memories. Journal of Experimental Psychology: General, 125, 6995.
Conway, M. A., & Dewhurst, S. A. (1995). Remembering, familiarity, and source
monitoring. Quarterly Journal of Experimental Psychology, 48, 125140.
Conway, M. A., & Pleydell-Pearce, C. W. (2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107, 261288.
Damasio, A. R. (1995). Toward a neurobiology of emotion and feeling: Operational
concepts and hypotheses. Neuroscientist, 1, 1925.
Danion, J., Gokalsing, E., Robert, P., Massin-Krauss, M., & Bacon, E. (2001). Defective relationship between subjective experience and behavior in schizophrenia.
American Journal of Psychiatry, 158(12), 20642066.
Danion, J. M., Huron, C., & Robert, P. (2001). Schizophrenia and autonoetic awareness. European Neuropsychopharmacology, S133S134.
Danion, J. M., Kaumann-Muller, F., Grang, D., Zimmermann, M. A., & Greth, P.
(1995). Aective valence of words, explicit and implicit memory in clinical depression. Journal of Aective Disorder, 34, 227234.
Danion, J. M., Kazs, M., Huron, C., & Karchouni, N. (in press). Do patients with
schizophrenia consciously recollect emotional events better than neutral
events? American Journal of Psychiatry.
Danion, J. M., Meulemans, T., Kaumann-Muller, F., & Vermaat, H. (2001). Intact
implicit learning in schizophrenia. American Journal of Psychiatry, 158, 944948.
Danion, J. M., Rizzo, L., & Bruant, A. (1999). Functional mechanisms underlying
impaired recognition memory and conscious awareness in patients with schizophrenia. Archives of General Psychiatry, 56, 639644.
Dawson, M. E., Nuechterlein, K. H., Schell, A. M., Gitlin, M., & Ventura, J. (1994).
Autonomic abnormalities in schizophrenia. State or trait indicators? Archives
of General Psychiatry, 51(10), 813824.
Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in
immediate recall. Journal of Experimental Psychology, 58, 1722.
Dewhurst, S. A., & Conway, M. A. (1994). Pictures, images, and recollective experience. Journal of Experimental Psychology: Learning, Memory, and Cognition, 20,
10881098.
Dominey, P. F., & Georgie, N. (1997). Schizophrenics learn surface but not abstract
structure in a serial reaction time task. Neuroreport, 8(13), 28772882.
Doniger, G. M., Silipo, G., Rabinowicz, E. F., Snodgrass, J. G., & Javitt, D. C. (2001).
Impaired sensory processing as a basis for object-recognition decits in schizophrenia. American Journal of Psychiatry, 158, 18181826.

, ,

237

Dutta, S., & Kanungo, R. N. (1975). Aect and memory: A reformulation. New York:
Pergamon Press.
Duzel, E., Yonelinas, A. P., Mangun, G. R., Heinze, H. J., & Tulving, E. (1997). Eventrelated brain potential correlates of two states of conscious awareness in memory.
Proceedings of the National Academy of Sciences of the USA, 94(11), 59735978.
Ekman, P. (1993). Facial expression and emotion. American Psychologist, 48, 384
392.
Feinstein, A., Goldberg, T. E., Nowlin, B., & Weinberger, D. R. (1998). Types and
characteristics of remote memory impairment in schizophrenia. Schizophrenia
Research, 30, 155163.
Frith, C. D. (1992). The cognitive neuropsychology of schizophrenia. East Sussex, England: Erlbaum.
Gardiner, J. M. (1988). Functional aspects of recollective experience. Memory &
Cognition, 16, 309313.
Gardiner, J. M., & Java, R. I. (1993). Recognising and remembering. In A. Collins,
M. A. Conway, S. E. Gathercole, & P. Morris (Eds.), Theories of memory (pp. 163
188). Hillsdale, NJ: Erlbaum.
Gardiner, J. M., Java, R. I., & Richardson-Klavehn, A. (1996). How level of processing really inuences awareness in recognition memory. Canadian Journal of
Experimental Psychology, 50, 114122.
Gardiner, J. M., & Richardson-Klavehn, A. (2000). Remembering and Knowing. In
E. Tulving & F. I. M. Craik (Eds.), The Oxford handbook of memory (pp. 229244).
New York: Oxford University Press.
Gold, J. M., Randolph, C., Carpenter, C. J., Goldberg, T. E., & Weinberger, D. R.
(1992). Forms of memory failure in schizophrenia. Journal of Abnormal Psychology, 101, 487494.
Goldberg, T. E., Weinberger, D. R., Pliskin, N. H., Berman, K. F., & Podd, M. H.
(1989). Recall memory decit in schizophrenia. A possible manifestation of
prefrontal dysfunction. Schizophrenia Research, 2, 251257.
Grang, D., Greth, P., & Danion, J. M. (1991). Application des mthodes de classication la construction dun test de mmoire. La Revue de Modulad, 8, 3143.
Gras-Vincendon, A., Danion, J. M., Grang, D., Bilik, M., Willard-Schroeder, D.,
Sichel, J. P., et al. (1994). Explicit memory, repetition priming and cognitive
skill learning in schizophrenia. Schizophrenia Research, 13, 117126.
Green, M. F. (1996). What are the functional consequences of neurocognitive decits in schizophrenia? American Journal of Psychiatry, 153, 321330.
Hasher, L., & Zacks, R. T. (1979). Automatic and eortful processes in memory. Journal of Experimental Psychology: General, 108, 356388.
Hawkins, K. A. (1999). Memory decits in patients with schizophrenia: Preliminary
data from the Wechsler Memory Scale-Third Edition support earlier ndings.
Journal of Psychiatry and Neuroscience, 24, 341347.
Heinrichs, R. W., & Zakzanis, K. K. (1998). Neurocognitive decit in schizophrenia: A quantitative review of the evidence. Neuropsychology, 12, 426445.
Henson, R. N., Rugg, M. D., Shallice, T., Josephs, O., & Dolan, R. J. (1999). Recollection and familiarity in recognition memory: An event-related functional magnetic resonance imaging study. Journal of Neurosciences, 19(10), 39623972.

238

Huron, C., & Danion, J. M. (2002). Impairment of constructive memory in schizophrenia. International Clinical Psychopharmacology, 17, 127133.
Huron, C., Danion, J. M., Giacomoni, F., Grang, D., Robert, P., & Rizzo, L.
(1995). Impairment of recognition memory with, but not without, conscious
recollection in schizophrenia. American Journal of Psychiatry, 152(12), 1737
1742.
Huron, C., Danion, J. M., Rizzo, L., Killofer, V., & Damiens, A. (in press). Subjective
qualities of memories with the picture superiority eect in schizophrenia. Journal of Abormal Psychology.
Jacoby, L. L. (1991). A process dissociation framework: Separating automatic from
intentional uses of memory. Journal of Memory and Language, 110, 306340.
Johnson, M. K., Foley, M. A., Suengas, A. G., & Raye, C. L. (1988). Phenomenal
characteristics of memories for perceived and imagined autobiographical
events. Journal of Experimental Psychology: General, 117, 371376.
Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitoring. Psychological Bulletin, 114, 328.
Johnson, M. K., & Raye, C. L. (1981). Reality monitoring. Psychological Review, 88,
6785.
Kay, S. R. (1982). Conceptual disorder in schizophrenia as a function of encoding
orientation. Journal of Nervous and Mental Diseases, 170, 154163.
Kayton, L., & Koh, S. D. (1975). Hypohedonia in schizophrenia. Journal of Nervous
and Mental Diseases, 161, 412420.
Kazs, M., Berthet, L., Danion, J. M., Amado, I., Willard, D., Robert, P., et al. (1999).
Impairment of consciously controlled use of memory in schizophrenia. Neuropsychology, 13, 5461.
Kerr, S. L., & Neale, J. M. (1993). Emotion perception in schizophrenia: Specic decit
or further evidence of generalized poor performance? Journal of Abnormal Psychology, 102, 312318.
Koh, S. D., Grinker, R. R., Marusarz, T. Z., & Forman, P. L. (1981). Aective memory
and schizophrenia anhedonia. Schizophrenia Bulletin, 7, 292307.
Koh, S. D., Kayton, L., & Peterson, R. A. (1976). Aective encoding and consequent
remembering in schizophrenic young adults. Journal of Abnormal Psychology,
85, 156166.
Koh, S. D., & Peterson, R. A. (1978). Encoding orientation and the remembering of
schizophrenic young adults. Journal of Abnormal Psychology, 87, 303313.
Koh, S. D., Szoc, R., & Peterson, R. A. (1977). Short-term memory scanning in
schizophrenic young adults. Journal of Abnormal Psychology, 86, 451460.
Kring, A. M., Kerr, S. L, Smith, D. A. & Neale, J. M. (1993). Flat aect in schizophrenia does not reect diminished subjective experience of emotion. Journal of
Abnormal Psychology, 102, 507517.
Kring, A. M., & Neale, J. M. (1996). Do schizophrenic patients show a disjunctive
relationship among expressive, experiential, and psychophysiological components of emotion? Journal of Abnormal Psychology, 105, 249257.
Lang, P. J., Greenwald, M. K., Bradley, M. M., & Hamm, A. O. (1993). Looking at
pictures: Aective, visceral, and behavioral reactions. Psychophysiology, 30,
261273.

, ,

239

Laruelle, M., Abi-Dargham, A., Van Dyck, C. H., Gil, R., DSouza, C. D., & Erdos, J.
(1996). Single photon emisssion computerized tomography imaging of amphetamine-induced dopamine release in drug-free schizophrenic subjects. Proceedings of the National Academy of Sciences of the USA, 93, 92359240.
Lazarus, R. S. (1991). Cognition and motivation in emotion. American Psychologist,
46, 352367.
Louilot, A., & Besson, C. (2000). Specicity of amygdalostriatal interactions in the
involvement of mesencephalic dopaminergic neurons in aective perception.
Neuroscience, 96, 7382.
Mandler, G. (1984). Mind and body: Psychology of emotion and stress. New York:
Norton.
Michel, L., Danion, J. M., Grang, D., & Sandner, G. (1998). Cognitive skill learning
and schizophrenia: Implications for cognitive remediation. Neuropsychology,
12, 590599.
Myin-Germeys, I., Delespaul, P. A., & deVries, M. W. (2000). Schizophrenia patients
are more emotionally active than is assumed based on their behavior. Schizophrenia Bulletin, 26, 847854.
Myin-Germeys, I., van Os, J., Schwartz, J. E., Stone, A. A., & Delespaul, P. A. (2001).
Emotional reactivity to daily life stress in psychosis. Archives of General Psychiatry, 58(12), 11371144.
Ochsner, K. N. (2000). Are aective events richly recollected or simply familiar? The
experience and process of recognizing feelings past. Journal of Experimental
Psychology: General, 129, 242261.
Palmer, B. W., Heaton, R. K., Paulsen, J. S., Kuck, J., Bra, D., Harris, M. J., et al.
(1997). Is it possible to be schizophrenic yet neuropsychologically normal?
Neuropsychology, 11, 437446.
Rado, S. (1962). Psychoanalysis of behavior. Vol. 2. New York: Grune & Stratton.
Rajaram, S. (1993). Remembering and knowing: Two means of access to the personal past. Memory & Cognition, 21, 89102.
Rajaram, S., & Roediger, H. L. III. (1997). Remembering and knowing as states of
consciousness during retrieval. In J. D. Cohen & J. W. Schooler (Eds.), Scientic
approaches to consciousness: Carnegie Mellon symposia on cognition (pp. 213
240). Mahwah, NJ: Erlbaum.
Rizzo, L., Danion, J. M., van der Linden, M., & Grang, D. (1996). Patients with
schizophrenia remember that an event has occurred, but not when. British Journal of Psychiatry, 168, 427431.
Rizzo, L., Danion, J.-M., Van Der Linden, M., Grang, D., & Rohmer, J. G. (1996).
Impairment of memory for spatial context in schizophrenia. Neuropsychology,
10, 376384.
Roediger, H. L., & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 21, 803814.
Rubin, D. C., & Friendly, M. (1986). Predicting which words get recalled: Measures
of free recall, availability, goodness, emotionality, and pronunciability for 925
nouns. Memory & Cognition, 14, 7994.
Saykin, A. J., Gur, R. C., Gur, R. E., Mozley, P. D., Mozley, L. H., Resnick, S. M., et al.

240

(1991). Neuropsychological function in schizophrenia. Selective impairment

in memory and learning. Archives of General Psychiatry, 48, 618624.
Saykin, A. J., Shtasel, D. L., Gur, R. E., Kester, D. B., Mozley, L. H., Staniak, P., et al.
(1994). Neuropsychological decits in neuroleptic naive patients with rstepisode schizophrenia. Archives of General Psychiatry, 51, 124131.
Schacter, D. L., Norman K. A., & Koutstaal, W. (1998). The cognitive neurosciences
of constructive memory. Annual Review of Psychology, 49, 289318.
Schneider, F., Weiss, U., Kessler, C., Salloum, J. B., Posse, S., Grodd, W., et al. (1998).
Dierential amygdala activation in schizophrenia during sadness. Schizophrenia Research, 34, 133142.
Schwartz, B. D., & Winstead, D. K. (1985). Icon formation in chronic schizophrenics.
Biological Psychiatry, 20, 10151018.
Schwartz, B. L., Rosse, R. B., & Deutsch, S. I. (1993). Limits of the processing view
in accounting for dissociations among memory measures in a clinical population. Memory & Cognition, 21, 6372.
Schwartz, B. L., Rosse, R. B., Veazey, C., & Deutsch, S. I. (1996). Impaired motor
skill learning in schizophrenia: Implications for corticostriatal dysfunction. Biological Psychiatry, 39, 241248.
Smith, M. E. (1993). Neurophysiological manifestations of recollective experience
during recognition memory judgments. Journal of Cognitive Neuroscience, 5, 1
13.
Sonntag, P., Gokalsing, E., Olivier, C., Robert, P., Burglen, F., Kaumann-Muller,
F., et al. (2003). Impaired strategic regulation of contents of conscious awareness in schizophrenia. Consciousness and Cognition, 12, 190200.
Taylor, S. F., & Liberzon, I. (1999). Paying attention to emotion in schizophrenia.
British Journal of Psychiatry, 174, 68.
Tulving, E. (1985). Memory and consciousness. Canadian Psychology, 26, 112.
Walker, E., Marwit, S. J., & Emory, E. (1980). A cross-sectional study of emotion
recognition in schizophrenics. Journal of Abnormal Psychology, 89, 428436.
Wickens, D. D. (1970). Encoding categories of words: An empirical approach to
meaning. Psychological Review, 77, 115.
Yonelinas, A. P., Kroll, N. E. A., Dobbins, I., Lazzara, M., & Knight, R. T. (1998).
Recollection and familiarity decits in amnesia: Convergence of rememberknow, process dissociation, and receiver operating characteristic data. Neuropsychology, 12, 323339.
Zajonc, R. B. (1984). On the primacy of aect. American Journal of Psychology, 39,
117123.

, ,

241

he events of our lives are imbued with emotion. In the everyday

ups and downs of daily life, the cyclical joys and sorrows, or the
heart-stopping ecstasy, grief, or fear that comes just a few times in a lifetime,
emotion is woven deep into the fabric of our personal past. Moreover, we dont
simply experience emotional events; we share them with others, both as they
are occurring and in retrospect. How we experience and reminisce with others
about the meaningful and emotional events of our lives plays a critical role in
how we come to understand and evaluate our past and ourselves. For young
children who are just beginning to construct an autobiographical memory, the
way in which emotion and memory are intertwined provides a foundation for
understanding of self and other (Fivush & Buckner, 2000; Fivush, Berlin, et al.,
2003). Most important, early autobiographical memory is co-constructed in the
context of parent-child reminiscing; parents inuence both the structure and the
content of young childrens emerging autobiography (Fivush, Haden, & Reese,
1996; Haden, Didow, Ornstein, & Eckerman, 2001; Nelson, 1993; Nelson &
Fivush, 2002).
In this chapter, we discuss two aspects of childrens developing memories of
emotional events. First, we examine parent-child reminiscing about everyday
emotional experiences. We argue that through participating in parent-guided
reminiscing, children develop an emotional self-concept that simultaneously
inuences the way in which the past is remembered and forms the basis for
understanding self in the present. We then turn to an examination of childrens
memories of highly emotional and stressful experiences and how these memories may dier from memories of highly positive experiences. Given our perspective on the role of parent guided reminiscing in childrens developing autobiographical memories, we also discuss the role of parent-child reminiscing about
these kinds of stressful experiences and how parents might help or hinder young

242

242

children in understanding and coping with aversive events. Throughout, we take

a functional perspective on autobiographical memory (Bluck & Alea, 2002;
Fivush, 1988). Our focus is on how autobiographical memories are used to link
the past with current understanding of self through the creation of an autobiographical narrative, help dene self in relation to others through time, and
contribute to understanding, regulating, and coping with aversive experiences.
Thus, we emphasize meaning making rather than accuracy; how do individuals
make sense of their past experiences and how does this process emerge developmentally within joint reminiscing? To place this research in perspective, we rst
briey review the development of childrens autobiographical memory and emotional understanding more broadly.

Autobiographical Memory

Development

Although memory functions from birth, autobiographical memorydened as

consciously available and socially shareable memories of specic events experienced by the self at a specied time and placeemerges slowly across the
preschool years (Nelson & Fivush, 2002; Pillemer & White, 1989). Children
begin to converse about specic past experiences at about 16 to 18 months of
age (Eisenberg, 1985; Harley & Reese, 1999). However, at this early point, adults
provide much of the content and structure of what happened; children participate by conrming, denying, or repeating what the adult says. Very quickly,
children become more competent participants in autobiographical reminiscing,
and by 3 years of age, children are able to give accurate detailed accounts of their
personal past (Fivush, Gray, & Fromho, 1987), although they still rely on adults
to provide much of the coherence, or narrative structure (Hamond & Fivush,
1990; Ornstein, 1995). By the end of the preschool years, children can provide
relatively clear and organized narratives of their past (Fivush, Haden, & Adam,
1995), and these skills continue to develop throughout childhood (Hudson &
Shapiro, 1991; McCabe & Peterson, 1991).
Extensive research conrms that children are learning the forms and functions for talking about the past through participating in adult-guided reminiscing. In accord with Vygotskys (1978) social-cultural theory of development,
adults provide the structure, or scaold, for childrens early reminiscing; over
time, children begin to internalize the forms used by adults and begin to provide
the structure for themselves. More specically, children learn the canonical
narrative forms and the culturally appropriate content to include when recalling the past (see Fivush, 1994, and Fivush & Haden, 2003, for a review). Importantly, there are clear and enduring individual dierences in both parents
and childrens autobiographical reminiscing (Fivush & Fromho, 1988; Hudson,
1990; McCabe & Peterson, 1991). Some parents display a highly elaborative

243

reminiscing style, talking about the past in great detail and encouraging their
childrens participation. In contrast, some parents display a less elaborative reminiscing style, taking about the past in sparse detail, often simply asking the same
questions over and over. Not only is parental reminiscing style consistent over
time but highly elaborative parents facilitate their childrens developing autobiographical memory skills. Children of more elaborative parents eventually
come to tell more richly embellished narratives of their own past experience than
children of less elaborative parents (Harley & Reese, 1999; McCabe & Peterson,
1992; Reese, Haden, & Fivush, 1993).
In addition to individual dierences within cultures, there are also substantial dierences among cultures in parent-child reminiscing. Within Asian cultures, in which self is conceptualized as interdependent and integrated into a
social and moral community (Oyserman & Markus, 1993), there is less focus
overall on the past; parents in these cultures are less elaborative than Caucasian
parents (Leichtman, Wang, & Pillemer, 2003). Provocatively, there are also
dierences within middle-class American culture between boys and girls. Parents
are more elaborative and more evaluative overall when reminiscing with daughters than with sons (Reese, Haden, & Fivush, 1996). By middle childhood, the
eects of these dierences in early parent-child reminiscing become apparent.
Caucasian children tell longer and more elaborated narratives of their personal
past than do Asian children (Han, Leichtman, & Wang, 1998), and middle-class
Caucasian girls tell longer and more elaborated personal narratives than boys
(Buckner & Fivush, 1998). And as adults, Caucasians and women provide longer
and more detailed narratives of their personal past than do Asians and adult men
(see Pillemer, 1998, for a review). The emerging eects of culture and gender
dierences indicate that parental reminiscing style is an important avenue for
the socialization of self and other, as well as for the construction of a culturally
appropriate autobiography (Fivush & Haden, 2003).

Emotional Development
Children begin talking about their emotions as well as their past early in development. By about 13 months of age, children begin referencing their emotional states,
and by 2 years of age, children are integrating information about their own
and others emotions into their everyday conversations (Bretherton, Fritz, ZahnWaxler, & Ridgeway, 1986). Further, and again similar to the development of
autobiographical memory, the way in which parents incorporate emotion into
conversations plays a critical role in childrens developing understanding of
emotion. Mothers who talk more about emotion early in development have children who talk more about emotion later in development (Denham, Zoller,
& Couchoud, 1994; Dunn, Bretherton, & Munn, 1987). More broadly, families that
discuss emotional experiences in more open and integrative ways have children

244

who develop better prosocial skills, have more positive peer relations, and show
better psychological adjustment (see Halberstadt, Denham, & Dunsmore, 2001,
for a review). And just as there are gender dierences in parent-child reminiscing, there are also gender dierences in emotion socialization. Mothers talk more
about emotions overall with girls than with boys, and girls begin to talk more
about emotion than boys do as early as the preschool years (Dunn et al., 1987;
Zahn-Waxler, Cole, & Barrett, 1991).
To summarize, early socialization patterns mirror adult patterns of emotional
expression and understanding. Adult women express more emotion than men
and report experiencing emotions more frequently and intensely than do men
(see Fischer, 2000, for a review). Specic to autobiography, adult women include
more emotion when narrating their past than do adult men (Bauer, Stennes, &
Haight, in press; Davis, 1990).

Parent-Child Reminiscing About

Emotional Events

The developments in autobiographical memory and emotional development

suggest that children may be socialized in both domains in parallel ways and
along similar timelines. Not surprisingly, when reminiscing about the past,
parents and children include information about the emotional aspects of their
shared experiences. Reminiscing about past emotional experiences diers in
important ways from talking about emotions experienced in the present (Dunn,
Brown, & Beardsall, 1991; Fivush, 1993). First, parents and children are not
in the immediate heat of the emotional moment when reminiscing about past
emotions and therefore may be better able to reect on and interpret emotional
experience. Second, in reminiscing, parents can choose to talk about particular emotions rather than others; for example, they can emphasize times when
their children were sad but downplay times when their children were angry.
Finally, parents can choose to talk about various aspects of emotional experience; they can focus on the emotion itself and the way it is expressed; they can
focus on the causes of emotional experience, or on how negative emotion is
resolved; and they can talk about both the childs and others emotional experiences and explain similarities or dierences. Thus, in the context of reminiscing about emotional experiences, parents can provide a framework for childrens
developing understanding of how and why they and others experience emotion.
Through parent-guided interpretation and evaluation of emotional experiences,
children begin to develop an emotional self-concept (Fivush & Buckner, 2003;
Fivush, Berlin, et al., 2003), composed of three functions: self-dening (the kind
of emotional person I am), self-in-relation (how I express and share my emotions with others), and self-regulation (how I cope with and resolve negative
emotion).

245

In line with gender dierences in many other aspects of emotional processing (Fischer, 2000), several studies have documented gender dierences in the
emotional content of mother-child reminiscing (Adams, Kuebli, Boyle, & Fivush,
1995; Fivush, 1989, 1991; Kuebli, Butler, & Fivush, 1995; Kuebli & Fivush,
1992). In these studies, mothers and their preschool children are visited in their
homes and asked to talk about several past experiences they have shared. No
further instructions are given. Tape-recorded conversation are transcribed and
coded for various dimensions. Across studies, mothers talk more about emotion
with their preschool daughters than sons, especially about sadness. Mothers also
use a wider variety of emotion words with girls than with boys, for example, talking about being sad and upset and crabby rather than simply sad. As children grow older, mothers talk increasingly of other peoples emotions as well as
the childs, especially with girls. Mothers also place emotions in a more social and
relational context with girls than with boys. In reminiscing with daughters, mothers talk about how emotions emerge from, and are modulated by, interactions with
other people, whereas with sons, they are more likely to discuss emotions as internal and autonomous experiences. These patterns suggest that mothers are
providing a more embellished and dierentiated understanding of emotion with
daughter than with sons. Although there are no dierences in how girls and boys
talk about past emotions early in the preschool years, by the end of the preschool
years, girls are taking signicantly more about their past emotional experience
than are boys (see Fivush & Buckner, 2000, for a review).
These ndings, along with previous ndings of dierences in adults autobiographical narratives, raise the question of possible dierences between
mothers and fathers reminiscing with daughters and sons. To explore this issue,
we asked mothers and fathers to independently discuss four specic past experiences with their 4year-old children when the child experienced happiness, sadness, anger, and fear (Fivush, Brotman, Buckner, & Goodman, 2000). Overall,
mothers talked more about emotion (M = 5.23 emotion words per event discussed) than fathers (M = 3.99 emotion words), but both mothers and fathers
talked about sadness, especially the causes of sadness, more with daughters
(M = 4.05 mentions) than with sons (1.95 mentions). In this study, girls talked
more about being scared than did boys, but there were no dierences in conversations about feeling happy, sad, or angry. Finally, as in previous research, both
mothers and fathers placed emotions in a more social and relational context with
daughters than with sons.
Thus, overall, this body of research indicates that emotional experiences are
discussed dierently depending on both the gender of the parent and the child.
Parents, especially mothers, talk about emotional experiences more, use a wider
variety of emotional language, and place past emotions in a more social and
relational context with their preschool daughters than with their sons. In turn,
by the end of the preschool years, girls are talking more about the emotional

246

aspects of their past experiences and using a wider variety of emotion terms than
boys. These patterns suggest that children are being socialized into genderappropriate ways of understanding their emotional experience. Parents facilitate the development of a more complex and more dierentiated emotional sense
of self in the past with daughters than with sons (Fivush, 1998b).
Whereas previous research on the development of autobiographical memory
has demonstrated distinct parental reminiscing styles that vary in elaborativeness, research on reminiscing about emotional events focuses on dierences in
the emotional content of parent-child reminiscing. To integrate these research
ndings, we examined whether level of maternal elaboration is related to the type
of emotional event discussed (Fivush, Berlin, et al., 2003). We were particularly
interested in how mothers reminisce about emotionally negative events, because
negative emotions are often dicult to cope with, and children must learn how
to express and resolve negative aect in culturally appropriate ways. Are mothers
more elaborative when reminiscing about fear rather than sadness or anger?
Further, does content vary depending on the type of emotion discussed? Do
mothers dier in their focus on the emotion itself, the causes of emotional experience, or resolution of negative aect depending on the type of emotional event
under discussion or the gender of the child?
Mothers reminisced with their 4-year-old children about three specic events,
a time the child was scared, angry, and sad. With daughters, mothers were more
elaborative overall, talking in more embellished detail about emotional experiences than with sons. Mothers were also more evaluative with daughters than
with sons, providing more feedback and more conrmation of their daughters
contributions to the conversations than their sons. Similarly, girls were more
elaborative than boys.
But there were also dierences depending on the type of emotion discussed.
Across all three emotions, mothers discussed the causes of the emotional experiences to a greater extent than the emotion itself or its resolution, suggesting
that mothers are helping their children to understand how emotional experiences
arise. However, when discussing fearful events, mothers elaborated more, and
they talked more about the facts concerning the event itself than when discussing sadness or anger. Mothers also focused on resolving fear more so than anger
(but not sadness). Children also seemed to be more concerned with fearful experiences than with sadness or anger. They talked more about fear overall, and they
provided more evaluations and resolutions when discussing fear than when discussing sadness or anger. In contrast, in conversations about anger, mothers did
not elaborate, they did not discuss the facts surrounding the event itself, and they
did not resolve anger as much as they did in conversations about fear. Rather,
when discussing anger, mothers focused on emotional attributions. Children also
did not talk as much about anger, they did not evaluate information about anger,
and they did not discuss emotional resolutions of anger to the same extent as

247

resolutions of fear or sadness. Finally, sadness falls between these two extremes.
Conversations about sadness were not as elaborative overall as conversations
about fear, nor did mothers focus on the event itself, as they did for fear, or on
emotional attributions, as they did for anger. Rather, mothers focused on evaluating and resolving sad feelings. Children also focused on resolving sadness.
These conversational patterns are illustrated in table 8.1, which presents
excerpts from conversations between one mother-daughter dyad and one
mother-son dyad, each reminiscing about three emotional events. Mothers focus
on the causes of their childrens emotions is apparent, as are the gender dierences. Across all three emotional events, the mother and daughter talk more
about the emotional event, and the mother uses more nuanced emotion language and talks in more detail about the daughters experience and expression
of emotion than in the mother-son dyad. These patterns conrm the previous
suggestion that girls are socialized into a more elaborated and dierentiated
understanding of their past emotions.
Perhaps most important, research on parent-child reminiscing underscores
that even when recalling the everyday events of our lives, emotion is an integral aspect of what we remember. Beginning very early in development, parents
and children are co-constructing personal narratives replete with emotional
experience; through these narratives, parents are helping their children to interpret and evaluate their experience. Mothers seem to reminisce more about emotions than do fathers, and both mothers and fathers reminisce about emotions
more with daughters than with sons. Through participating in richly emotional
reminiscing, girls may be constructing a more emotionally laden and emotionally nuanced sense of self in the past. Further, by placing emotions in a more
interpersonal context with daughters than with sons, parents may be teaching
girls to integrate emotions into relationships with others to a greater extent than
are boys. Thus, girls may be developing a more embellished, more dierentiated, and more relational emotional self-concept than are boys. However, there
are also important dierences depending on the type of emotion discussed.
These patterns suggest that parents are inuencing the way in which their
children are learning to understand, express, and perhaps even experience
specic emotions.
An important limitation of this research is the focus on white middle-class
American families. Emotional expression is clearly culturally mediated (Lutz &
White, 1986); therefore, it is not surprising that there are cultural dierences
in parent-child reminiscing about emotional events (see Leichtman et al., 2003,
for a review). For example, Wang (2002) asked American and Chinese mothers
to reminisce about times when their preschool child was happy, scared, angry,
and sad. American mothers used what Wang labeled an emotion-explaining
style, a pattern similar to what we have just described, whereas Chinese mothers
used an emotion-criticizing style. An emotion-explaining style helps create
shared emotional experiences that dene one in relation to others; an emotion-

248

 . Excerpts From Mother-Child Conversations About Everyday Emotional

Events (. . . indicates some missing dialogue)
Mother-daughter dyad

Mother-son dyad

Conversations about fear

M: And what happens when theres a big
thunder and lightening storm?
C: Id want to be with my mom and dad
[whispering].
M: Yes, you want to be with your mom and
dad. And what happens if youre
sleeping, and sleeping and theres a big
bunch of thunder in the middle of the
night? What happens sometimes?
C: Scared
M: Scared, do you tremble like that? Do
you shake? Huh? Then what do you
do?
C: I get up and go to my mom and dad
[whispering].
M: You get up and go to your mom and
dad. And what do we do?
C: Say dont worry.
M: Say dont worry. Do we hold you? Yes,
does holding you help you when youre
scared? Yes.

M: . . . why did you come in bed with

mommy and daddy? Cause how did you
feel, how did that storm make you feel?
C: Umm, I feel really scared.
M: A little bit scared?
C: Really scared.
M: You were really scared. What scares
you about the thunderstorm honey?
C: . . . brokes the TV.
M: Yeah, that scared you because the TV
went o.

Conversations about anger

M: Well, I seem to remember that
yesterday you and Patrick were
swinging on the swings and he was
doing something that really annoyed
you. . . . Do you remember?
C: He wouldnt give me a swing.
M: Oh! Thats right . . . and did you wait
patiently for a long time, hmm?
C: Yes.
M: And is that when you started to get
really angry? Uh-huh, how do you feel
inside when you get angry?
C: . . . Mad.
M: Mad! And did your mouth turn down at
the corners? And do you feel like going
grrrr?
Conversations about sadness
...
M: Well, one thing that made you really
sad was when your best friend Shena
moved away, right? Yeah, and did we
watch all her things go on the moving
truck? Uh-huh, and do you remember
why she had to move away?
C: . . . Because Shenas Dad had to work.

M: Now how do you feel when Mommy

makes you take you vitamins in the
morning and you dont want to take
them?
C: I feel mad.
M: You do! And what do you do?
C: I hide.
M: . . . Uh-huh, then what happens to
Mommy? How does Mommy feel when
you didnt take your vitamins and your
calcium?
C: Mad.
M: Then what happens to [you]?
C: Get in trouble.
M: Yeah.

M: How did you feel when Mandy wouldnt

let you play the other day, when she
was playing with Jess and she told you
boys to go away? How did that make
you feel?
C: Umm, sad.
M: And why does Mandy do that?
(continued)

249

 . (continued)
Mother-daughter dyad

Mother-son dyad

Conversations about sadness

M: Shenas Daddy was going to start
working at a new job . . . And do you
still miss Shena when you think about
it? Yes?
C: Yes.
M: It makes you sad, doesnt it? But is she
still your friend even far away? Yes!
What can you do even though shes far
away?
C: Give her a happy letter with a [drawing]
on it.
M: Give her a happy letter, right, and we
have a drawing, dont we?

C: Because shes mean.

M: [laughing] Is she mean? Is she mean all
the time? When she doesnt let you
play, youre not very happy, are you?
C: Umm.

criticizing style uses emotion to instill proper behavior and values. Wang argues
that these dierent styles emerge from larger cultural constructions of self as
independent or interdependent, as discussed earlier. Independent selves own and
express their emotional experience, whereas interdependent selves use emotions
to regulate appropriate social behavior (see also Mullin & Yi, 1995). Obviously,
these are not orthogonal functions, and mothers and children from both cultures
likely show elements of both styles, but Wangs research highlights that children may be developing dierent understandings of self and emotion depending
on the larger social and cultural context within which reminiscing is embedded.
Thus far, we have discussed how parent-child reminiscing modulates childrens developing understanding of their everyday emotional experiences. But
what of more stressful and traumatic events? How do extreme levels of stress
aect childrens memories, and how might parent-guided reminiscing play a role
in helping children to understand and cope with aversive experiences?

Memories of Stressful

and Traumatic Events

Intertwined with everyday emotional experiences are the unfortunate accidents, disasters, and other stressful experiences that occur over the course
of a lifetime. The concept of stress is familiar to both laypersons and professionals. In fact, it is not uncommon to hear the words stress or trauma in everyday conversations. However, as Mason (1975) notes, The single most remarkable historical fact concerning the term stress is its persistent, widespread
usage in biology and medicine in spite of almost chaotic disagreement over

250

its denition (p. 9). Despite the lack of a universal denition, the topic of
stress and memory has been of interest to researchers from many areas of
study.
Because much of the developmental research stems from forensic issues
involving children as witnesses or victims of violence, the focus has been on how
stress aects the amount, accuracy, and suggestibility of childrens recall of past
experiences (see Ceci & Bruck, 1993, and chapter 10, this volume). Although
not the focus in this chapter, the consensus on accuracy is clear (see Westcott,
Davies, & Bull, 2002, for an overview). Even quite young children can respond
to open-ended free-recall questions (e.g., What was the man wearing?) with
accurate details of their past experiences. Moreover, repeatedly interviewing
children about the same event does not compromise accuracy in the absence
of suggestive and misleading questions (Fivush, Peterson, & Schwartzmueller,
2002). The problem is that preschoolers, though accurate, provide little information in response to these open-ended questions, and errors increase dramatically as questions become more close-ended.
Accuracy is compromised in young childrens memory reports in two ways.
First, responses to specic close-ended questions (e.g., What was the man wearing on his head?), especially yes/no questions (e.g., Was the man wearing a
hat on his head?), are vulnerable to error during the preschool years (Fivush
et al., 2002). Second, preschoolers are substantially more susceptible to misleading and suggestive questions (e.g., What color was the mans jacket? when no
jacket was worn) than older children and adults (Ceci & Bruck, 1993). And suggestive questions asked over repeated interviews can induce many errors into
young childrens memory reports (Leichtman & Ceci, 1995). What is still in question is the extent to which these kinds of misleading questions change the actual memory representation or whether children are responding to the implicit
social demands of the interviewers questions. Related to this, although children
change their responses to specic questions, it is not clear that these errors are
incorporated into childrens subsequent free recall (Cassal & Bjorkland, 1995;
Fivush et al., 2002). Moreover, the extent to which stress interacts with accuracy of recall is still controversial.
First and foremost, this research domain is plagued by the problem of operationalizing stress. Several methods have been devised to assess childrens stress
levels, including self-report, global ratings of stress level by parents or doctors,
objective stress ratings based on specic behavioral indices, and physiological
measures such as heart rate, skin conductance, and hormonal levels. When directly compared, the dierent measures of stress do not correlate highly with one
another (Eisen, Goodman, Ghetti, & Qin, 1999; Ornstein, 1995; Parker, Bahrick,
Merrit, Lundy, & Fivush, 1998); thus, it is not surprising that they do not correlate systematically with memory measures. In addition, dierent measures of
memory have been used across studies, some studies assessing free recall and
other studies assessing cued recall or recognition. Also, dierent aspects of the

251

event are targeted for recall, such as central or peripheral information. Moreover, studies have examined events ranging from mildly stressful, such as getting an inoculation, to highly traumatic, such as witnessing the murder of a
parent. Finally, stress may inuence memory dierently at dierent developmental points, further complicating comparisons across studies.
Several clinical case studies describe childrens memories of truly horrendous
experiences, such as being kidnapped and buried alive or witnessing the murder of a parent (Malmquist, 1986; Terr, 1979, 1988). Terr (1979) reports that
school-age children who were kidnapped from a school bus and essentially buried alive overnight were able to recall the horric event in vivid detail up to
4 years later. Similarly, in describing long-term memories of children traumatized during the preschool years, Terr (1988) reports that children 3 years of age
or older at time of traumatization continue to report vivid and elaborated memories of events such as abuse, kidnapping, and other violent acts. However, children under the age of 3 at the time of the event do not develop the ability to recall
the event verbally as they grow older, although aspects of the traumatic experiences may be present in their behavior. These ndings suggest that real-life traumatic events are remembered in great detail over long periods of time, at least in
children 3 or older at the time of the event.
Although clinical studies allow rich description of memories of real-life
trauma, they do not allow for systematic comparisons across children or type of
event. To examine this question, developmental researchers have taken advantage of necessary but painful medical procedures that many children must endure. Studies use one of two approaches. One approach compares one group of
childrens memories of stressful events, such as inoculations, to another group
of childrens memories of more mundane events such as a routine doctor examination. Children in the stressful situation generally show enhanced memory for
the event in comparison to the participants experiencing a more neutral event,
suggesting that stressful events are better recalled than emotionally neutral
events (see Pezdek & Taylor, 2001, for a review). The other methodological approach has been to select a single stress-producing event and examine individual
dierences in stress level. For instance, Merritt, Ornstein, and Spicker (1994)
examined childrens recall of a stressful medical procedure, a voiding cystourethrogram. They found that children who experienced more stress during this
procedure, as measured by behavioral ratings, recalled less accurate information than children who displayed less stress during the procedure. Similarly,
Vandermass, Hess, and Baker-Ward (1993) found that high anxiety or stress had
a debilitative eect on older childrens recall of a dental operation but not on
younger childrens, suggesting developmental dierences in the stress-memory
relation. More specically, preschool children may experience and recall stressful events dierently than older children.
In a systematic, longitudinal research program examining childrens memories of a stressful event, Peterson and others have examined childrens memories

252

of an injury resulting in emergency room treatment (Peterson, 1999; Peterson &

Bell, 1996; Peterson & Whalen, 2001). Injuries were relatively minor, requiring
only outpatient procedures, and included mostly lacerations requiring stitches,
severe burns, and broken bones. Children recalled their injury and treatment
within 2 weeks of the event and again at 6 months, 1 year, 2 years, and 5 years
after the event. Although there are several nuances to the results, the basic nding
is that children 3 years of age or older at the time of the experience recalled this
event extremely well over the 5year period. In fact, Peterson and Whalen report
virtually no forgetting of the injury itself and little forgetting of the hospital treatment. However, there were no relations between stress, as assessed by maternal
and child report on a Likert scale, and either amount or accuracy of recall. These
results conrm that children recall stressful events extremely well over long time
periods, although they also contribute to the mixed ndings in the literature over
possible relations between stress and memory.
Importantly, children under the age of 3 at the time of experience showed a
somewhat dierent pattern (Peterson & Rideout, 1998). One- and 2-year-old
children who were not able to verbally recall the event when it occurred did not
show accurate recall as they grew older. Although many of these children did
respond to the interviewers questions about what happened at later interviews,
about 50% of what they reported was wrong. Thus, similar to the clinical descriptions provided by Terr (1988), more systematically controlled research conrms that children under the age of 3 may have particular diculty verbally
recalling events even as they grow older and develop the required language skills.
Most important, Peterson and Rideout present evidence suggesting that it is not
age per se that is critical, but rather childrens ability to verbally recall the event
when it occurred. Those 2-year-olds who were able to give a coherent, although
obviously still limited, verbal report of what happened when rst interviewed remained able to recall the event accurately as they grew older. Thus, it seems that
language or narrative skill at the time of experience is a critical factor in predicting whether children will remain able to recall events accurately as they grow
older (see Fivush, 1998a, and Fivush, Pipe, Murachver, & Reese, 1997, for review and further discussion of this issue).
One of the limitations of this body of research is the focus on memories of
medical procedures. Whereas many of these procedures are quite painful and
stressful for young children, it is not clear that they approach the level of stress
experienced during traumatic events. Further, children often take their emotional cues from the adults around them as to the meaning and severity of the
event; in the case of medical procedures, parents may be stressed, but they are
not in fear of either their own or their childrens lives (at least for the events that
have been studied thus far). One of the prevalent theories about the relation
between memory and stress comes from the Yerkes-Dodson law that predicts an
inverted-U function (see Christianson, 1992, for an overview). Basically, it is
postulated that as stress increases to moderate levels, memory will be enhanced

253

due to increased attention and processing, but as stress increases to extreme

levels, memory will be hindered because the system will be overwhelmed. Although this theory is widely cited, little evidence shows that this function holds.
One problem in assessing this function is that control for level of stress while
holding the event to be recalled constant is dicult. And, of course, creating
extremely high levels of stress in the laboratory is not ethical.
Only one study has systematically examined relations between childrens
levels of stress and memory for a naturally occurring life-threatening traumatic
event. Bahrick, Parker, Fivush, and Levitt (1998) examined how stress aected
preschool childrens memories for Hurricane Andrew, a devastating storm that
hit the Florida coast in 1992. Stress was dened by the amount of property damage to the familys home. Children were placed into either a low, moderate, or
high damage group. Families who prepared for the hurricane, but sustained little
property damage, were placed in the low damage group. If they experienced
substantial damage to the perimeter of their home, such as the yard, windows,
or roof, but the house itself remained intact, they were placed into the moderate
damage group. Families who experienced a severe storm that penetrated the
home, often resulting in partial collapse of the roof and multiple broken windows
or doors, were placed into the high damage group.
Three- and 4-year-old children were interviewed about their experiences
within 2 to 6 months of the storm. Interviews began with an open-ended question about the hurricane, followed by nondirective prompts until the child did
not recall any additional information. Then the experimenter asked a series of
more specic questions about preparing for the storm, the storm itself, and the
aftermath of the storm. Childrens recall was related to the amount of stress
experienced, such that the children exposed to a moderate amount of damage
recalled more than children who experienced high or low damage did. This pattern would seem to support the proposed inverted-U function between memory
and stress. However, 6 years later, when the children were 9 and 10 years old,
we interviewed them again about their hurricane experiences (Fivush, Sales,
Goldberg, Bahrick, & Parker, in press). Figure 8.1 displays the mean amount of
information children recalled at each interview by stress group. All children were
able to recall the event in vivid detail 6 years later, and there were no dierences
in overall amount recalled among children in the low, moderate, or high damage groups. Thus, there is no longer an inverted-U function. In fact, children now
reported more than twice as much information about the experiences as they
had at the initial interview. We do not want to argue that no forgetting occurs;
we argue that children are now able to verbally report much more of what they
remember than they could when they were younger.
In this situation, children were able to recall the event when it occurred, and,
as their language skills increased, their verbal report of the event increased as
well (see Peterson & Whalen, 2001, for similar ndings and arguments). This
conrms our earlier discussion of Petersons research (Peterson & Rideout,

254

Mean propositions recalled

140
120
100
80

Low
Moderate
High

60
40
20
0
Immediate Recall

6 year follow-up Recall

Figure 8.1. Mean number of propositions recalled by children in each stress group
over time.

1998). In that research, children who were unable to recall the event at the time
of occurrence were not able to subsequently recall the event with any accuracy,
but like that research, we also found that children who could recall the core
aspects of the event when it occurred remained able to recall the event in vivid
detail over extended time, indicating that memory of stressful events is quite long
lived. Importantly, amount of recall was not related to either retention interval
or amount of rehearsal at either interview time.
However, there were still eects of stress on memory 6 years later. Children
in the high damage group needed more prompts and cues to recall as much information as children in the moderate and low damage groups, as shown in gure 8.2. This changes the interpretation of the original ndings. Obviously, if
children in the high stress group are now recalling as much information as children in the other two groups, arguing that stress interferes with memory per se
becomes dicult. Rather, stress may interfere with either the ability or the willingness to retrieve information for recall. If we interpret the lower recall in the
high damage group at the rst interview as a form of avoidance, in that highly
stressed children simply did not want to talk about their experiences, then 6 years
later children in the high damage group may express avoidance as being less
willing to respond freely to the experimenter because they found it stressful to
recall their experiences openly. However, when directly questioned and prompted,
these children demonstrated that they do indeed recall as much as the children
who experienced lesser damage. Thus, stress may interfere with retrieval in
dierent ways at dierent developmental ages or as time since the stressful experience increases.

255

Mean Amount Recalled

120
100
80
60

Free
Recall
Cued
Recall

40
20
0
Low Stress

Moderate
Stress

High Stress

Figure 8.2. Mean number of propositions recalled by children in each stress group
in free and cued recall.

Although stress aected recall at both interviews, children in all three damage
groups recalled detailed information about their hurricane experience. Though
we were unable to compare amount of recall about this event directly to other
events recalled by these same children, comparing across studies of childrens
memory, we found that children recalled substantially more about the hurricane
event than their age-peers recall about more mundane or even mildly stressful
events (see Fivush, 1998a, for a review). Further, although only a subset of the
childrens recall was judged for accuracy by mothers, ratings of accuracy were
extremely high, above 90%. Thus, it appears that highly stressful events are well
recalled overall, although as stress increases, children may be less willing to think
about and talk about their experiences.
To illustrate the richness and detail of childrens recall, table 8.2 presents
excerpts from a child in the high damage group and a child in the moderate damage group at each interview. These are childrens responses to the rst openended question, What do you remember about Hurricane Andrew? and the
nondirective prompts that followed this question. As shown, both children recalled the event in great detail at each interview. Even 6 years later, children were
able to narrate a detailed, coherent account of the storm and its consequences.
These excerpts also illustrate how emotionally compelling the memories are,
especially for children who experienced extensive damage to their homes. Children seem to remain highly aected by their experience even years later. As we
already mentioned, most of the research on stress and memory has focused on
accuracy and overall amount of information children recall. However, relations
between memory and stress are important not only for forensic reasons but also

256

table 8.2 Excerpts of Childrens Recall of Hurricane Andrew

Immediate interview

Follow-up interview

Child who experienced high damage

E: Can you think really hard about the
hurricane and tell me everything you
remember about it?
C: [Child nods yes]
E: Yeah? What do you remember rst?
C: Um, I remember we had friends over.
E: You had friends over, yeah, what else?
C: Um, I was, um, stued in a corner.
E: You were stued in a corner, yeah, what
else can you remember?
C: I was in my bedroom.
E: In your bedroom, yeah.
C: And um . . . and um, during the
hurricane, there was a window in my
bedroom, and um, well we, I was stued
in the corner and then we, then we, um,
we had the Farmigans over and they,
um, stayed at our house for a little bit
longer after the hurricane.
E: Yeah, what else can remember about
that?
C: Um, that they were sitting on the
furniture and I wasnt.
E: And you werent. O.K. Can you
remember anything else about the
hurricane?
C: [Child shakes head no]

E: Can you tell me everything you can

remember about Hurricane Andrew?
C: Okay. Um, I remember the hurricane
was the day after my fourth birthday
party. And we were scared. My grand,
grandparents came over. And our
neighbors Sue and Terry. And it was
me, my mom and my dad. And we were
in my room and my grandparents were
sleeping in my mom and dads room.
And I remember I had on, I sat, I wanted
to go close to the window because we
had a mattress over our window,
because theres only one window in my
room. And I sat and then this big
lightening bolt came I was scared a lot
and I remember half way through the
night, part, a corner of our roof blew o.
And our grandparents came and told us
that it was o and we put two or three
trashcans under it to catch it. And
Terry, he had out little weather radio
and he said he was sorry because there
were no cartoons on, there was only
weather about the hurricane. And I
remember, I had my bed is in the corner
of the room and about ve feet before
my closet and they made me, my mom
and my dad made me stay there the
whole night. Cause they were afraid if
the window blew o, the mattress would
hit me [she goes on like this for several
more sentences]. . . . And I was really
scared because I had never been
through anything like that, and um, I
dont remember a lot.

Child who experienced moderate damage

E: So, can you think really hard and tell
me what you remember about
Hurricane Andrew?
C: Ah. . . . You know, in my neighbors
house.
E: Ah.
C: All the screens cause she has a column
of screens . . . when . . . when the screen
was knocked over, they had to get new
screen, and there was all trees. My next
door neighbor had to go all [makes
noise] like that.

E: Can you think really hard about the

hurricane and tell me everything that
you can remember about it. We want to
hear anything and everything that you
can think of.
C: Well, I remember we had a coconut
palm in our backyard, and my dad had
to cut all the coconuts o to make sure
none of the windows broke or anything.
E: Uh-huh.
C: And, well, and that coconut palm broke
in the Hurricane Andrew so now thats
(continued)

257

table 8.2 (continued)

Immediate interview

Follow-up interview

Child who experienced moderate damage

E: Anything else you remember about
that?
C: Only this . . . only my little screen fell o
and nothing else. And all the trees fell
o, a lot a lot, but not all.
E: Right.
C: Any you know what, and even when it
was almost the hurricane we were
cutting coconuts, and then, then we got
back inside, and then, and my dad, and
then he is all sweaty and he has ants all
over. Thats it.
E: Ok. Anything else that you remember
about the hurricane
C: Um, I think nothing else.

not there anymore. But we didnt, we

were pretty lucky and we didnt have
much happen. All it was was, like,
fences and stu.
E: Uh-huh.
C: And our pool was very dirty.
E: Really?
C: Yeah. We, uh, stayed in our, what used
to be our play room, which is now our
computer room, and we had little
shutters that we put over the glass so no
glass would break there so we were very
safe there. And we had a, we had a
patio by the pool.
E: Uh-huh.
C: And we have some sort of electrical
shutters to close everything up in the
patio, so we can put lots of our stu in
there. And I dont remember much of
the hurricane because I dont know
how, but I fell asleep in the middle of it.
E: Really?
C: Uh-huh. So thats about it.

Note. C stands for Child, E stands for Experimenter

because how we remember and incorporate the stressful occurrences of our lives
has implications for future coping and physical and emotional well-being
(Pennebaker, 1997). What we remember and how we come to understand the
events of our lives may be more important for our everyday functioning than
whether our memories are perfect replicas of the actual event (Fivush et al., in
press). As the excerpts illustrate, children do not simply recall what happened
during the hurricane; they are actively trying to understand and process what
occurred, as indicated by the many references to their thoughts and emotions.
Thus, we were interested in examining the content of childrens recall of Hurricane Andrew in relation to stress in more detail (Sales, Fivush, Parker, & Bahrick,
2002). In particular, based on research on expressive writing about negative
events with adults (see Pennebaker, 1997, for a review), we focused on the inclusion of cognition and emotion words that have been related to higher levels
of stress and coping with aversive events. Further, given the mixed ndings in
previous research, we were interested in examining relations among multiple
measures of stress and memory.
We included two dierent types of stress measures at the initial interview: Likert
scale ratings by mother and child assessing global assessment of experienced stress
and more concrete, behaviorally based measures of stress, including the Frederick

258

Posttraumatic Stress Disorder Reaction Index (Frederick, 1985), which asks mothers and children to rate 20 specic behaviors associated with stress (e.g., Do
thoughts about the hurricane make you feel afraid or upset? and Have you had
bad dreams since the hurricane?); a child well-being score, which assessed 35
specic behaviors associated with stress (e.g., Wants to sleep with adults and
Laughs easily); and the damage groupings, as previously dened. Only the PTSD
Reaction Index was completed at the follow-up interview. As shown seen in table
8.3, the damage assessment, the child well-being survey and the PTSD Reaction
Index, which all rate specic physical events or specic child behaviors, were highly
correlated with one another, but not related to the more global and subjective Likert
scale stress measures. Thus, the most reliable measures of stress seem to be those
that decrease the amount of subjective interpretation of the rater and depend instead on more observable events and behaviors. Furthermore, only the objective
ratings of stress were related to the content of childrens recall.
As shown in table 8.4, which displays correlations among the stress and
memory measures, highly stressed children initially included less positive emotion, fewer cognitive processing words, and less information overall than less
stressed children. The inclusion of less language indicative of cognitive and
emotional processing of the event, as well as shorter narratives, suggests that
the highly stressed children had diculty processing the event immediately following the storm. However, 6 years later they included more negative emotion
words and more cognitive words in their recall than less stressed children, indicating that these highly stressed children are still trying to process and understand the event they experienced when they were only 3 and 4 years old to a
greater extent than children who were less stressed at the time. In contrast, children who exhibited more emotional and cognitive processing indicative of better
coping during their initial interview had better psychological outcomes 6 years
later, as indicated by lower PTSD scores. These patterns accord with our previous interpretation that children who were highly stressed by Hurricane Andrew
actively tried to avoid thinking about and processing the event. However, this
table 8.3 Correlations Between Stress Variables
Dmg
MSR
CSR
CWB
PTSD1
PTSD2

.43*

.03
-.36*
.43**
.13

MSR

CSR

CWB

.09

-.25
.19
.16

.16

-.07
.21

-.79**

-.13

PTSD1

.24

Note: Dmg = damage rating; MSR = mothers rating of childs

stress; CSR = childs rating of own stress; CWB = child wellbeing score; PTSD1 = Time 1 total PTSD score; PTSD2 = Time
2 total PTSD score.
**p<.05.
**p<.01.

259

table 8.4 Correlations Between Stress and Content Variables

Damage rating
Moms stress rating
Childs stress rating
Child well-being
PTSD Total Time 1
PTSD Total Time 2

Time 1 content variables

Time 2 content variables

Pos
Neg
Cog Total
words words words recall

Pos
Neg
Cog Total
words words words recall

-.37*
-.24
.07
.27
-.17
-.37*

-.06
-.12
.02
.51**
-.30
-.11

-.53**
-.19
-.09
.08
-.09
-.01

-.37*
-.25
-.18
.27
-.38*
-.37*

.21
.03
.06
-.06
.11
.14

.46**
.01
.04
-.16
.31
.26

.45**
.06
-.18
-.16
.29
.26

.02
-.29
-.26
-.11
.11
.14

*p<.05
**p <.01

early avoidance led to later diculty, and 6 years later, these children still seem
to be struggling with trying to process their experience. In contrast, those children who were able to talk about their hurricane experience in more cognitively
and emotionally integrative ways showed fewer stress symptoms related to the
hurricane 6 years later.
Consistent with these ndings, Wolitzky, Fivush, Zimand, Hodges, and Rothbaum (2001) found that children who included more internal state language indicative of cognitive and emotional processing in their narratives about a
cancer-related medical procedure had lower anxiety levels. Thus, across two types
of stressful experiences the same pattern emerges between stress levels and content of childrens recall. Children who are less anxious and stressed immediately
following an aversive event are able to express more about their thoughts and
emotions concerning the event than children who are more anxious and stressed,
and these children show fewer symptoms of stress over time. However, over time,
more highly stressed children seem to include more of their thoughts and emotions in their recall. More detailed longitudinal research is needed to elucidate temporal patterns, as well as to establish causal relations. Still, the ndings thus far
indicate that it is not just amount of recall that may be aected by stress but also
the content of what children focus on when they report traumatic experiences.
A major limitation of this research is the focus on negative events. Researchers
tend to use the terms arousal and stress interchangeably and ignore emotional
valence. It is possible that memory is aected by higher arousal in similar ways
for both positive and negative experiences, or there may be dierences depending on valence of the event. Few studies systematically compare memory of
events that are equally arousing but dier in valence. The few experimental
studies that have manipulated both of these dimensions found that arousal, not
valence, accounts for better memory in adults, at least for static images of events
presented in a laboratory setting (Hamann, Ely, Grafton, & Kilts, 1999).
In the rst study to compare childrens memories of positive experiences
directly to negative stressful life events, we examined narratives produced by

260

5- to 12-year-old children growing up in a violent community (Fivush, Hazzard,

Sales, Sarfati, & Brown, 2003).Because of the nature of these childrens living
environments, many of the negative events narrated were highly traumatic and
involved interpersonal violence, including witnessing a gun ght between relatives, seeing ones mother arrested and taken away by the police, and seeing an
assault victim. However, other children narrated negative events that fall well
within the range of normal childhood stressors, including serious illnesses and
invasive medical procedures. Most of the positive events were family and school
excursions to parks and beaches and church retreats. Because we allowed families to nominate the positive and negative events of their lives, it is not clear that
the positive and negative events selected were equally arousing.
We were specically interested in the type of information children recalled,
and we coded all unique units of information into one of seven categories:
mention of persons, places, objects, actions, descriptions, and internal states
including thoughts and emotions. As shown in gure 8.3, children recalled a

40
35

Mean Units

30
25
20
15

Positive
Negative

10
5
0
l
t
e
n
n
n
on ora
ec
at
io
io
io
s
j
t
t
t
t
r
c
b
a
S
p
ip
A
O
Pe
m
oc
al
cr
e
n
s
L
r
e

T
D
te
In
Codes

Figure 8.3. Mean amount of information recalled in each category for positive and
negative events.

261

great deal of information about both types of events. Overall amount of recall
did not dier between positive and negative events, but the content diered.
When recounting negative experiences, children included more information
about their thoughts and emotions than when they recounted positive experiences. In contrast, when recalling positive experiences, they recalled more
information about people, actions, and descriptions. Intriguingly, childrens
narratives of the negative events were also more coherent than the positive
event narratives. Coherence may emerge as children try to think through and
process aversive experiences.There were no dierences in parental reports of
how often the positive or negative events had been talked about, nor was time
since occurrence a factor in amount or type of recall. This pattern suggests that,
even though children report similar amounts of information quantitatively,
their reports of negative events dier in content from their reports of positive
events.
The inclusion of thoughts and emotions in childrens narratives likely is related to how aversive the event is and may reect the way in which children are
trying to process, understand, and possibly cope with these negative experiences.
Given that parent-guided conversations about everyday emotional experiences
may help children in understanding and coping with negative emotion, we ask
how parents may discuss highly stressful events with their young children and
how this may aect how children remember and cope with stress.

Parent-Child Reminiscing About

Stressful Events

To date, only two studies have examined how parents and children discuss highly
stressful experiences. Ackil, Waters, Dropnik, Dunisch, and Bauer (1999) compared mother-child conversations about a devastating tornado that leveled their
town to conversations about a nontraumatic event. Tornado conversations
included more mention of negative emotion and information about the causes
and consequences of the event than did the conversations about the nontraumatic events, suggesting that mothers are using the negative conversations
as an opportunity to facilitate their childrens understanding of this overwhelming event. However, this study focused on the content of mothers and childrens
conversations and did not assess parental reminiscing style. As we discussed
earlier, elaborative parents facilitate the development of more detailed and embellished autobiographical narratives in their children. If parental reminiscing
style remains consistent across positive and negative events, children of elaborative parents should produce more detailed narratives about negative past events
as well. Findings in the adult literature indicate that detailed and coherent narratives of stressful events are related to higher levels of coping and well-being
(Foa, Molnar, & Cashman, 1995; Pennebaker, 1997). As with adults, a more

262

detailed memory of negative events may facilitate childrens coping with such
events. Thus, an elaborative parental reminiscing style may help children cope
with stressful life experiences.
We examined both reminiscing style and content of parent-child conversations about two emotionally laden events: an injury requiring emergency room
treatment and an individually nominated positively valenced experience (Sales,
Fivush, & Peterson, 2003). As in previous research, we dened parental style as
extent of elaboration. Parental reminiscing style was consistent across the two
emotional conversations, such that parents who were more highly elaborative
when reminiscing about the positive experience were also more elaborative
when reminiscing about the negative experience. Further, parents with a highly
elaborative style had children who reported more new information during the
conversations.
However, there were also dierences in parental reminiscing style for positive and negative events. When reminiscing about stressful events, parents asked
a higher proportion of open-ended memory questions to their children, which
require children to provide information in response, whereas when reminiscing about the positive experiences, parents asked a higher proportion of yes/
no questions, which require children only to conrm or negate the questions.
Parents also focused more on emotion when discussing positive experiences
with their children and more on causal explanations when discussing the stressful experiences.
These dierences suggest that parents may have dierent underlying goals
in these two dierent emotional contexts. Reminiscing about shared positive experiences serves to create, maintain, and strengthen emotional bonds,
as well as to create a shared history, which is a basis for family identity (Fivush
et al., 1996). Thus, in emotionally positive conversations, parents engage in coconstructing the experience with their children, with each conversational partner contributing information about the shared event. Furthermore, a greater
focus on emotion, and especially positive emotion, may highlight how the event
was meaningful to the parent-child relationship.
In contrast, discussing stressful experiences may serve more of a didactic function. Parents are concerned with teaching their children how to cope with lifes
stressful experiences and to avoid similar situations in the future. To facilitate
their childrens thinking about what happened and why, parents may try harder
to elicit their childrens recall of details of the event, rather than providing this
information themselves and asking for conrmation. Moreover, as in Ackil et al.
(1999), parents focused on the causes of negative events more so than of positive events. Interestingly, parents focus on causes of emotion even when reminiscing about everyday negative events, as we noted earlier. Focusing on causal
information helps children understand how and why stressful events occurred,
thereby making the event more comprehensible and perhaps providing a lesson
on how to cope with similar events. Thus, the functions of reminiscing about

263

negative events, whether they are the everyday ups and downs or more serious
and even traumatic experiences, seem to dier from the functions of reminiscing about more positive experiences.
Most intriguing, we are now exploring whether parental reminiscing style
predicts childrens independent recall of stressful events in interviews with an
unfamiliar adult (Peterson, Sales, & Fivush, 2002). Regression analyses on the
amount and accuracy of childrens recall of their injury and hospital treatment
indicates that parental education level, parental report of amount of family discussion of the event, parental report of the childrens stress level, and childrens
language skills do not predict childrens recall, but parental reminiscing style
does. Children of highly elaborative parents recall the hospital treatment more
accurately and in more detail than children of less elaborative parents. Recall of
the injury is more variable, with parental reminiscing style predicting level of
detail but not accuracy. These ndings indicate that parents who discuss emotionally dicult events with their young children in elaborative ways may help
facilitate their childrens understanding and memory of these events. Future
analyses will allow us to examine whether parental style continues to facilitate
childrens memories for stressful events over time.

Narratives and Coping

Perhaps most intriguing, our research points to the close connection between
memory and coping. The ways in which we recall the negative events of our lives
is critical for subsequent physical and psychological well-being (Pennebaker,
1997). Adults who narrate stressful events more coherently, and include more
emotion words and more words indicative of causal and cognitive processing
(e.g., realize, understand, because) subsequently show lower levels of anxiety, a
higher sense of well-being, and better immune system functioning than adults
who narrate stressful events less coherently, and use fewer emotion and causal/
cognitive words.
Our results suggest that, for young children, who are not yet able to create a
coherent account of a stressful event for themselves, the ways in which adults,
and particularly parents, help them to organize and remember these events may
have important implications for childrens coping. Highly elaborative parents,
who help their children construct a coherent and emotionally detailed narrative and help provide a causal-explanatory framework for the event, may facilitate their childrens coping eorts. In turn, children of highly elaborative parents
come to be able to construct coherent and emotionally detailed narratives of past
experiences for themselves and thus may also come to develop better coping
strategies to handle aversive events. Thus, we posit a dialectical developmental
relation between parental reminiscing style and the development of childrens
coping strategies (see Bretherton, 1996, for related arguments).

264

Gender dierences in parental reminiscing may also have implications for

coping. Because parents are more elaborative and more emotional when reminiscing with their young daughters than with their sons, girls seem to develop
more detailed and emotionally dense narratives of their personal past than do
boys. This, in turn, suggests that girls may also develop better coping strategies
than boys. Little is known about the development of coping strategies in children in general, and little attention has been paid to possible gender dierences.
We are currently examining these issues in our laboratory.

Remembering Interpersonal

Violence

Finally, we emphasize that virtually all of the research on childrens memories

for stressful events has focused on medical procedures and natural disasters, and
this limitation may aect conclusions. First, these events do not involve interpersonal violence. Even though children may not understand the necessity of
painful and invasive medical procedures, these events may not approach the
interpersonal betrayal involved when one individual intentionally harms another. Interpersonal violence shatters basic assumptions about human interaction (Jano-Bulman, 1992). Second, and equally important, children too often
experience violence at the hands of trusted adults. Physical abuse and sexual
abuse are not simply stressful experiences; they fundamentally undermine a
childs belief in a safe world. In this way, interpersonal violence, specically abusive experiences, may be remembered very dierently from other kinds of stressful experiences (Fivush, 1998a; Freyd, 1996).
Moreover, given the role of parent-child reminiscing in helping children to
make sense of stressful experiences, and the posited relation to coping, children
who experience violence and abuse may not have the opportunity to talk about
these experiences with adults who can help them resolve and cope with the aversive aect. Often violence and abuse occurs within silenced communities, where
children are implicitly or explicitly not allowed to speak of their experiences.
Thus, children are left on their own to make sense out of what is essentially senseless. It is not clear from the research ndings thus far what might happen to these
kinds of memories (see Fivush, 1998a, for a full discussion of this issue). Certainly, this is a critical topic for future research.

Conclusions and Implications

We began this chapter with the assertion that emotion and memory are inseparable; the events of our lives are imbued with emotion, both at the time of experience and in retrospect. In the process of reminiscing with others, our life

265

experiences take on deep emotional and evaluative meaning that aects our
sense of our self and our autobiographical history. The research we reviewed
supports this assertion. Beginning very early in development, parents and children are already reminiscing in emotionally rich ways about their shared experiences. Parents who reminisce with their preschool children in elaborative and
emotional ways have children who begin to tell the stories of their own lives in
greater and more nuanced emotional detail. Further, these patterns are related
to gender. Both mothers and fathers reminisce in more emotionally embellished
ways with their preschool daughters than with their sons, and by the end of the
preschool years, girls are including more emotion and more dierentiated emotion in their own autobiographical narratives. This research demonstrates that
even when recalling and reminiscing about the everyday events of our lives,
emotion is an integral part of our experience. We do not simply remember what
happened; we remember what these events meant to us in the past and continue
to mean to us in the present.
Moreover, our review points to the importance of examining how emotional
events are recalled, examining issues that go beyond accuracy. Although stress
is a dicult concept to dene and measure, the research indicates that highly
stressful events are well recalled even over long periods. Our emphasis on content provides important new information about how and why stressful events
are remembered. Stress and memory are intricately interrelated, such that level
of experienced stress inuences the emotional and cognitive content of childrens
recall, and the content of childrens recall is related to continuing levels of stress
as time since the experience passes. What children are remembering about the
stressful events of their lives is important not just for understanding memory but
for understanding coping and emotional resilience. The way in which children
make sense of their stressful experiences, often through participating in joint
reminiscing with their parents, helps them to place these stressful experiences
in an appropriate context, in which children develop a sense of control over their
life experiences through the ability to understand and regulate their emotional
reactions. A functional approach to memory for emotional events expands the
focus of research from addressing accuracy and retention to examining relations
between parent-guided reminiscing, childrens developing narratives, and coping with aversive events.

References
Ackil, J., Waters, J., Dropnik, P., Dunisch, D., & Bauer, P. (1999). From the eyes of
the storm: Mother-child conversations about a devastating tornado. Poster session
presented at the biennial meetings of the Society for Research in Child Development, Albuquerque, NM.
Adams, S., Kuebli, J., Boyle, P., & Fivush, R. (1995). Gender dierences in parentchild conversations about past emotions: A longitudinal investigation. Sex
Roles, 33, 309323.

266

Bahrick, L. E., Parker, J. F., Fivush, R., & Levitt, M. (1998). The eects of stress on
young childrens memory for a natural disaster. Journal of Experimental Psychology: Applied, 4, 308331.
Bauer, P. J., Stennes, L., & Haight, J. (in press). Representation of the inner self in
autobiography: Womens and mens use of internal state language in personal
narratives. Memory.
Bluck, S., & Alea, N. (2002). Exploring the functions of autobiographical memory.
In J. D. Webster & B. K. Haight (Eds.), Critical advances in reminiscing work: From
theory to application (pp. 6175). New York: Springer.
Bretherton, I. (1996). Internal working models of attachment relationships as related to resilient coping. In G. G. Noam & K. W. Fischer (Eds.), Development and
vulnerability in close relationships (pp. 327). Mahwah, NJ: Erlbaum.
Bretherton, I., Fritz, J., Zahn-Waxler, C., & Ridgeway, D. (1986). Learning to talk
about emotions: A functionalist perspective. Child Development, 57, 529548.
Buckner, J. P., & Fivush, R. (1998). Gender and self in childrens autobiographical
narratives. Applied Cognitive Psychology, 12, 407429.
Cassel, W. S., & Bjorkland, D. F. (1995). Developmental patterns of eyewitness memory
and suggestibility: An ecologically based short-term longitudinal study. Law and
Human Behavior, 19, 507532.
Ceci, S. J., & Bruck, M. (1993). Suggestibility of the child witness: A historical review and synthesis. Psychological Bulletin, 113, 403439.
Christianson, S.-. (1992). Emotional stress and eyewitness memory: A critical review. Psychological Bulletin, 112, 284309.
Davis, P. J. (1990). Gender dierences in autobiographical memories for childhood
emotional experiences. Journal of Personality and Social Psychology, 76, 498510.
Denham, S. A., Zoller, D., & Couchoud, E. A. (1994). Socialization of preschoolers
emotion understanding. Developmental Psychology, 30, 928936.
Dunn, J., Bretherton, I., & Munn, P. (1987). Conversations about feeling states
between mothers and their young children. Developmental Psychology, 23, 132
139.
Dunn, J., Brown, J., & Beardsall, L. (1991). Family talk about feeling states and
childrens later understanding of others emotions. Developmental Psychology,
27, 448455.
Eisen, M., Goodman, G. S., Ghetti, S., & Qin, J. (1999, July). An examination of abuse
disclosures in maltreated children. Paper presented at the meetings of the Society for Applied Research in Memory and Cognition, Boulder, Colorado.
Eisenberg, A. (1985). Learning to describe past experience in conversation. Discourse
Processes, 8, 177204.
Fischer, A. H. (2000). Gender and emotion: Social psychological perspectives. New York:
Cambridge University Press.
Fivush, R. (1988). The functions of event memory: Some comments on Nelson and
Barsalou. In U. Neisser & E. Winograd (Eds.), Remembering reconsidered: Ecological and traditional approaches to memory (pp. 277282). New York: Cambridge University Press.
Fivush, R. (1989). Exploring sex dierences in the emotional content of mother-child
talk about the past. Sex Roles, 20, 675691.

267

Fivush, R. (1991). Gender and emotion in mother-child conversations about the

past. Journal of Narrative and Life History, 1, 325341.
Fivush, R. (1993). Emotional content of parent-child conversations about the past.
In C. A. Nelson (Ed.), The Minnesota Symposium on Child Psychology: Memory
and aect in development (pp. 3977). Hillsdale, NJ: Erlbaum.
Fivush, R. (1994). Constructing narrative, emotion and self in parent-child conversations about the past. In U. Neisser & R. Fivush (Eds.), The remembering self:
Accuracy and construction in the life narrative (pp. 136157). New York: Cambridge University Press.
Fivush, R. (1998a). Childrens memories for traumatic and non-traumatic events.
Development and Psychopathology, 10, 699716.
Fivush, R. (1998b). Gendered narratives: Elaboration, structure and emotion in
parent-child reminiscing across the preschool years. In C. P. Thompson, D. J.
Herrmann, D. Bruce, J. D. Read, D. G. Payne, & M. P. Toglia (Eds.), Autobiographical
memory: Theoretical and applied perspectives (pp. 79104). Hillsdale, NJ: Erlbaum.
Fivush, R., Berlin, L., Sales, J. M., Mennuti-Washburn, J., & Casidy, J. (2003). Functions of parent-child reminiscing about negative events. Memory, 11, 179192.
Fivush, R., Brotman, M., Buckner, J. P., & Goodman, S. (2000). Gender dierences
in parent-child emotion narratives. Sex Roles, 42, 233254.
Fivush, R., & Buckner, J. P. (2000). Gender, sadness and depression: Developmental and socio-cultural perspectives. In A. H. Fischer (Ed.). Gender and emotion:
Social psychological perspectives (pp. 232253). Cambridge: Cambridge University Press.
Fivush, R., & Buckner, J. P. (2003). Constructing gender and identity through autobiographical narratives. In R. Fivush & C. Haden (Eds.), Autobiographical
memory and the construction of a narrative self: Developmental and cultural perspectives (pp. 149168). Hillsdale, NJ: Erlbaum.
Fivush, R., & Fromho, F. (1988). Style and structure in mother-child conversations
about the past. Discourse Processes, 11, 337355.
Fivush, R., Gray, J. T., & Fromho, F. A. (1987). Two year olds talk about the past.
Cognitive Development, 2, 393409.
Fivush, R., & Haden, C. (1997). Narrating and representing experience: Preschoolers developing autobiographical recounts. In P. van den Broek, P. A.
Bauer, & T. Bourg (Eds.), Developmental spans in event comprehension and representation: Bridging ctional and actual events. Hillsdale, NJ: Erlbaum.
Fivush, R., & Haden, C. (2003). Autobiographical memory and the development of a
narrative self: Developmental and cultural perspectives. Hillsdale, NJ: Erlbaum.
Fivush, R., Haden, C., & Adam, S. (1995). Structure and coherence of preschoolers
personal narratives over time: Implications for childhood amnesia. Journal of
Experimental Child Psychology, 60, 3256.
Fivush, R., Haden, C., & Reese, E. (1996). Remembering, recounting and reminiscing: The development of autobiographical memory in social context. In D. Rubin
(Ed.), Reconstructing our past: An overview of autobiographical memory (pp. 341
359). New York: Cambridge University Press.
Fivush, R., Hazzard, A., Sales, J., Sarfati, D., & Brown, T. (2003). Creating coher-

268

ence out of chaos? Childrens narratives of emotionally negative and positive

events. Applied Cognitive Psychology, 17, 119.
Fivush, R., Peterson, C., & Schwarzmueller, A. (2002). Questions and answers: The
credibility of child testimony in the context of specic questions. In M. Eisen,
G. S. Goodman, & J. Quas (Eds.), Memory and suggestibility in the forensic interview (pp. 331354). Hillsdale, NJ: Erlbaum.
Fivush, R., Pipe, M.-E., Murachver, T., & Reese, E. (1997). Events spoken and unspoken: Implications of language and memory development for the recovered
memory debate. In M. Conway (Ed.), True and false memories (pp. 3462).
Oxford: Oxford University Press.
Fivush, R., Sales, J. M., Goldberg, A., Bahrick, L., & Parker, J. (in press). Weathering
the storm: Childrens long-term recall of Hurricane Andrew. Memory.
Foa, E. B., Molnar, C., & Cashman, L. (1995). Change in rape narratives during exposure therapy for posttraumatic stress disorder. Journal of Traumatic Stress, 8,
675690.
Frederick, C. J. (1985). Children traumatized by catastrophic situations. In S. Eth &
D.F. Ross (Eds.), Childrens eyewitness memory (pp. 123). New York: Springer.
Freyd, J. (1996). Betrayal trauma: The logic of forgetting childhood abuse. Cambridge,
MA: Harvard University Press.
Haden, C. A., Didow, S. M., Ornstein, P. A., & Eckerman C. O. (2001). Mother-child
talk about the here-and now: Linkages to subsequent remembering. Child Development, 72, 10161031.
Halberstadt, A. G., Denham, S. A., & Dunsmore, J. (2001). Aective social competence. Social Development, 10, 79119.
Hamann, S. B., Ely, T. D., Grafton, S. T., & Kilts, C. D. (1999). Amydala activity related to enhanced memory for pleasant and aversive stimuli. Nature Neuroscience, 2, 289293.
Hamond, N. R., & Fivush, R. (1991). Memories of Mickey Mouse: Young children
recount their trip to Disney World. Cognitive Development, 6, 433448.
Han, J. J., Leichtman, M. D., & Wang, Q. (1998). Autobiographical memory in Korean,
Chinese, and American children. Developmental Psychology, 34, 701713.
Harley, K., & Reese, E. (1999). Origins of autobiographical memory. Developmental
Psychology, 35, 13381348.
Hudson, J. A. (1990). The emergence of autobiographic memory in mother-child
conversation. In R. Fivush, & J. A. Hudson (Eds.), Knowing and remembering in
young children (pp. 166196). New York: Cambridge University Press.
Hudson, J. A., & Shapiro, L. (1991). Eects of task and topic on childrens narratives. In A. McCabe & C. Peterson (Eds.), New directions in developing narrative
structure (pp. 89136). Hillsdale, NJ: Erlbaum.
Jano-Bulman, R. (1992). Shattered assumptions. New York: Free Press.
Kuebli, J., Butler, S., & Fivush, R. (1995). Mother-child talk about past events: Relations of maternal language and child gender over time. Cognition & Emotion,
9, 265293.
Kuebli, J., & Fivush, R. (1992). Gender dierences in parent-child conversations
about past emotions. Sex Roles, 12, 683698.

269

Leichtman, M. D., & Ceci, S. J. (1995). The eects of stereotypes and suggestions on
preschoolers reports. Developmental Psychology, 31, 558578.
Leichtman, M., Wang, Q., & Pillemer, D. P. (2003). Cultural variation in interdependence and autobiographical memory. In R. Fivush & C. Haden (Eds.),
Autobiographical memory and the construction of a narrative self: Developmental
and cultural perspectives (pp. 7398). Hillsdale, NJ: Erlbaum.
Lutz, C. A., & White, G. M. (1986). The anthropology of emotion. Annual Review of
Anthropology, 15, 405436.
Malmquist, C. P. (1986). Children who witness parental murder: Post-traumatic
aspects. Journal of the American Academy of Child Psychiatry, 25, 320325.
Mason, J. W. (1975). A historical view of the stress eld. Part 1. Journal of Human
Stress, 1, 612.
McCabe, A., & Peterson, C. (1991). Getting the story: A longitudinal study of parental styles in eliciting narratives and developing narrative skill. In A. McCabe
& C. Peterson (Eds.), Developing narrative structure (pp. 217253). Hillsdale, NJ:
Erlbaum.
Merritt, K. A., Ornstein, P. A., & Spicker, B. (1994). Childrens memory for a salient
medical procedure: Implications for testimony. Pediatrics, 94, 1723.
Mullen, M., & Yi, S. (1995). The cultural context of talk about the past: Implications
for the development of autobiographical memory. Cognitive Development, 10,
407419.
Nelson, K. (1993). The psychological and social origins of autobiographical
memory. Psychological Science, 1, 18.
Nelson, K., & Fivush, R. (2002). The emergence of autobiographical memory. Manuscript submitted for publication.
Ornstein, P.A. (1995). Childrens long-term retention of salient personal experiences. Journal of Traumatic Stress, 8, 581606.
Oyserman, D., & Markus, H. (1993). The sociocultural self. In J. Suls (Ed.), Psychological perspectives on the self: The self in social perspective (Vol. 4, pp. 187220).
Hillsdale, NJ: Erlbaum.
Parker, J., Bahrick, L., Merrit. K., Lundy, B., & Fivush, R. (1998). Eects of stress on
childrens memory for a natural disaster. In C. P. Thompson, D. J. Herrmann,
J. D. Read, D. Bruce, D. G. Payne, & M. P. Toglia (Eds.), Eyewitness memory:
Theoretical and applied perspectives (pp. 3154). Hillsdale, NJ: Erlbaum.
Pennebaker, J. W. (1997). Opening up. New York: Guilford.
Peterson, C. (1999). Childrens memories for medical emergencies: 2 years later.
Developmental Psychology, 35, 14931506
Peterson, C., & Bell, M. (1996). Childrens memory for traumatic injury. Child Development, 67, 30453070.
Peterson, C., & McCabe, A. (1982). Developmental psycholinguistics: Three ways of
looking at a narrative. New York: Plenum.
Peterson, C. & Rideout, R. (1998). One- and two-year-olds remember medical emergencies. Developmental Psychology, 34, 10591072.
Peterson, C., Sales, J. M., & Fivush, R. (2002). Predicting childrens recall of emotionally negative events. Paper submitted to the biennial meetings of the Society for
Research in Child Development, Tampa, FL.

270

Peterson, C., & Whalen, N. (2001). Five years later: Childrens memories for medical emergencies. Applied Cognitive Psychology, 15, 118.
Pezdek, K., & Taylor, J. (2001). Memories of traumatic events. In M. Eisen, G. S.
Goodman, & J. S. Quas (Eds.), Memory and suggestibility in the forensic interview
(pp. 165184). Hillsdale, NJ: Erlbaum.
Pillemer, D. (1998). Momentous events, vivid memories. Cambridge, MA: Harvard
University Press.
Pillemer, D., & White, S.H. (1989). Childhood events recalled by children and adults.
In H. W. Reese (Ed.), Advances in child development and behavior, Vol. 22. New
York: Academic Press.
Reese, E., Haden, C. A., & Fivush, R. (1993). Mother-child conversations about the
past: Relationships of style and memory over time. Cognitive Development, 8,
403430.
Reese, E., Haden, C. A., & Fivush, R. (1996). Mothers, fathers, daughters and sons:
Gender dierences in reminiscing. Research on Language and Social Interactions,
29, 2756.
Sales, J. M., Fivush, R., Parker, J., & Bahrick, L. (2002). Stressing memory: Long term
relations among childrens stress, recall and psychological outcome following
Hurricane Andrew. Manuscript submitted for publication.
Sales, J. M., Fivush, R., & Peterson, C. (2003). Parental reminiscing about positive
and negative events. Journal of Cognition and Development, 4, 185209.
Terr, L. C. (1979). Children of Chowchilla: A study of psychic trauma. Psychoanalytic
Study of the Child, 34, 547623.
Terr, L. C. (1988). What happens to early memories of trauma? A study of twenty
children under age ve at the time of documented traumatic events. Journal of
the American Academy of Child and Adolescent Psychiatry, 27, 96104.
Vandermass, M. O., Hess, T. M., & Baker-Ward, L. (1993). Does anxiety aect childrens reports of memory for a stressful event? Journal of Applied Psychology, 7,
109128.
Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Cambridge, MA: Harvard University Press.
Wang, Q. (2002). Did you have fun?: American and Chinese mother-child conversations about shared emotional experiences. Cognitive Development, 16, 693715.
Westcott, H., Davies, G., & Bull, R. (Eds.). (2002). Childrens testimony in context. New
York: Wiley.
Wolitzky, K. Fivush, R., Zimand, E., Hodges, L., & Rothbaum, B. O. (2001). Coping
with cancer: Relations among parent and child coping, anxiety and distress, and
eectivenes of virtual reality during invasive medical procedures. Manuscript submitted for publication.
Zahn-Waxler, C., Cole, P., & Barrett, K. (1991). Guilt and empathy: Sex dierences
and implications for the development of depression. In J. Garber & K. Dodge
(Eds.), The development of emotion regulation and deregulation: Cambridge studies
in social and emotional development (pp. 243272). New York: Cambridge University Press.

271

ost studies investigating the relationship between emotion

and memory in adults have used college students as participants. However, the impact of emotion on memory may
change across the adult life span. Compared with younger adults, older adults
exhibit decits in many types of memory tasks (for reviews, see Light, 1996; Prull,
Gabrieli, & Bunge, 2000; Zacks, Hasher, & Li, 2000). In contrast, older adults
actually improve rather than decline in emotional well-being and regulation
(e.g., Carstensen & Charles, 1998). Emotions gain centrality in everyday information processing with age (Blanchard-Fields, 1998; Labouvie-Vief, 1998), and
there is evidence that brain regions associated with emotional processes deteriorate less with age than many other regions. As outlined in the rst part of this
chapter, this research on emotion and aging predicts that the relationship between emotion and memory should change as people age. In the second part of
the chapter, I review studies examining age dierences in memory for emotional
information. These studies reveal that emotion becomes more salient in memory
and that emotional goals have more of an inuence as one grows older.

Age-Related Changes in Emotional

Processes and Their Implications

for Memory

Emotional Well-Being Improves

Across the Adult Life Span

Emotional well-being remains stable or even is enhanced with age. Younger

adults are at higher risk for depression than older adults (Lawton, Kleban, &
Dean, 1993; Weissman, Leaf, Bruce, & Florio, 1988). Even among patients di-

272

272

agnosed with dysthymia (a milder, chronic form of depression), older adults have
higher ratings for emotional well-being and seem to be in better mental health
(Oxman, Barrett, Sengupta, & Williams, 2000). Almost all mental disorders are
more prevalent among younger than older adults (Regier et al., 1988). A longitudinal study found that clinician-rated psychological health shows a steady
improvement from age 30 to 62 (Jones & Meredith, 2000), indicating that these
mental health dierences do not simply reect a cohort eect.
Many studies have also found that normal everyday subjective emotional
experience either remains steady or actually improves across the life spana
quite remarkable nding considering that physical health tends to decline and
social networks tend to contract in old age. On questionnaires, older adults respond that they experience fewer negative emotions than younger adults (Gross
et al., 1997). Life satisfaction remains constant (Diener & Suh, 1998) or increases
(Lawton et al., 1993) with age. When asked to look back and rate how satised
they were at dierent points in their lives, people tend to report that they are most
satised at the current moment (Field, 1997), suggesting that people also have
a subjective sense that their own satisfaction increases across the life span. In
one experience sampling study including participants from 18 to 94 years old,
the frequency and duration of negative emotions experienced in daily life decreased with age (Carstensen, Pasupathi, Mayr, & Nesselroade, 2000). Positive
aect appears to remain mostly constant (Carstensen et al., 2000) or increase
(Mroczek, 2001) across the life span. In very old age (after the age of 85), crosssectional data indicate that positive aect may decrease again (Smith, Fleeson,
Geiselmann, Settersten, & Kunzmann, 1999), although a longitudinal study
found that positive outlook increased from ages 85 to 92 (Agren, 1998). Another
longitudinal study spanning 23 years revealed that negative aect decreased over
time in multiple age cohorts, whereas positive aect remained stable until about
age 60, when it decreased slightly (Charles, Reynolds, & Gatz, 2001).
Among younger adults, negative mood increases the likelihood of remembering negative information (Blaney, 1986; Bower, 1981). Depression also increases the likelihood of remembering negative information (Bradley, Mogg, &
Williams, 1995; Gilboa, Roberts, & Gotlib, 1997; Johnson & Magaro, 1987;
Mineka & Nugent, 1995; Watkins, Mathews, Williamson, & Fuller, 1992). [Also
see chapter 6Eds.] Older adults lower levels of negative aect suggest that they
may be less likely to recall negative information than younger adults. Thus,
across the life span, the likelihood of remembering negative information may
decrease relative to the likelihood of remembering positive or neutral information, as a result of decreasing negative mood.

Emotion Regulation Eectiveness Increases

Emotions often seem to be beyond our control, coming and going whether we
choose to experience them or not. Nevertheless, we actually have considerable

273

control over our emotions (e.g., Gross, 2001). Furthermore, there is evidence that
the ability to regulate emotions improves with age. Older adults are more likely
than younger adults to maintain positive emotional states and to maintain the
absence of negative states (Carstensen et al., 2000). They are more likely than
younger or middle-aged adults to endorse statements suggesting increased emotional control, such as I try hard to stay in a neutral state and to avoid emotional situations and Detachment or cool judgment is my best way to meet life
situations (Lawton, Kleban, Rajagopal, & Dean, 1992). They are also less likely
to ruminate about emotionally upsetting events (McConatha & Huba, 1999).
In addition, their coping and defense strategies tend to reect more impulse control (Diehl, Coyle, & Labouvie-Vief, 1996). Across an ethnically and culturally
diverse set of samples, older adults consistently reported being better able to control their emotions than younger adults (Gross et al., 1997). Ratings of emotional
control and emotional experience were correlated; higher ratings of emotional
control were associated with higher frequencies of happiness and lower frequencies of sadness and fear. This pattern suggests that changes in emotion regulation processes contribute to the improvement in emotional experience across the
life span.
People use a variety of strategies to regulate emotion. For example, reappraisal
and suppression are both eective ways to regulate emotion, but they have quite
distinct mechanisms and side eects (e.g., Gross, 2001). In general, these strategies can be classied as either response-focused or antecedent-focused (Gross,
2001, 2002). Response-focused strategies attempt to manage emotion after it
is already under way, whereas antecedent-focused strategies attempt to inuence
emotion before it occurs. Initial research suggests that older adults increased
ability to regulate emotion relies more on antecedent-focused strategies, such
as reappraising an event to alter its emotional impact, than on response-focused
strategies, such as suppressing the expression of an emotion (Carstensen, Gross,
& Fung, 1998). For example, compared with younger adults, older adults report
using less confrontative coping, greater distancing, and more positive reappraisal
(Folkman, Lazarus, Pimley, & Novacek, 1987). In general, older adults seem to
be better able to reappraise events cognitively (Diehl et al., 1996; Labouvie-Vief,
DeVoe, & Bulka, 1989).
Laura Carstensens socioemotional selectivity theory posits that the increasing eectiveness of emotion regulation results from the increasing salience of
emotional goals as people approach the end of life (Carstensen, 1992, 1995;
Carstensen, Isaacowitz, & Charles, 1999). According to her theory, perceived
limitations on time direct attention toward emotional goals. Thus, older adults
prefer to spend time with emotionally meaningful social partners (Fredrickson
& Carstensen, 1990; Fung, Carstensen, & Lutz, 1999; Fung, Lai, & Ng, 2001),
and younger people mimic this preference under experimental conditions where
time is limited (Fung et al., 1999). Older adults emphasize emotional dimensions
more than other personal dimensions in their mental representations of social

274

partners (Fredrickson & Carstensen, 1990), a pattern also evident in younger

adults facing the end of their lives because of terminal illness (Carstensen &
Fredrickson, 1998). Older adults are also more likely to adopt emotion-focused
strategies when attempting to solve interpersonal problems (Blanchard-Fields
& Camp, 1990; Blanchard-Fields, Camp, & Casper Jahnke, 1995). Thus, it appears that aging, because it brings people continually closer to the ultimate end,
shifts goal priorities so that emotional goals gain importance and have more of
an impact on behavior.
The research demonstrating improved emotion regulation and increasing
importance of emotional goals with age suggests several age dierences in
memory for emotional information. First, the increased importance of emotional
goals may increase the salience of all emotionally relevant information relative
to neutral information for older adults. Second, because one way to regulate
emotion is to avoid attending to or remembering negative informationfocusing instead on positive informationolder adults may have a bias against negative information in their memories. Thus, older adults memories may be more
emotionally gratifying. Third, insofar as older adults are more likely to use reappraisal strategies, they may remember originally negative events in a more
positive light. I review studies addressing these possibilities later in the chapter.

Decreasing Physiological Arousal

Associated With Emotion

Physiological reactions to emotion-eliciting events rely on the cardiovascular

system, which changes with age (Cacioppo, Berntsen, Klein, & Poehlmann,
1998). The muscle mass of the heart decreases, leading to a decline in the hearts
stroke volume. In addition, both blood ow resistance and resting blood pressure
increase. Thus, it is perhaps not surprising that cardiovascular reactivity (heart
rate and pulse transmission time) associated with emotional responding decreases with age (Levenson, Carstensen, & Gottman, 1994; Levenson, Friesen,
Ekman, & Carstensen, 1991; Tsai, Levenson, & Carstensen, 2000). This age
dierence in physiological reactivity appears when participants are asked to generate and relive past emotional episodes from their own lives (Levenson et al.,
1991), discuss marital problems with their spouses (Levenson et al., 1994), or
watch emotional lm clips (Tsai et al., 2000). These physiological dierences
occur despite similar subjective emotional experience and expressive behavior
among younger and adult participants.
According to the Easterbrook hypothesis (1959), physiological arousal leads
to a narrowing of attention. Many studies have examined this claim in the context of the relationship between emotional arousal and memory in younger
adults, but the results are mixed. Typically, these studies nd a positive eect of
arousal when the level of arousal is moderate and a negative eect when the it
is relatively high (Baddeley, 1998). [Also see chapter 1Eds.] Using these stud-

275

ies to make predictions about the implications of age-related changes in physiological arousal is further complicated because older adults lower levels of physiological reactivity are not associated with lower levels of emotional experience
relative to younger adults. Thus, the age dierences in physiological reactivity
do not lead to any clear predictions about how the relationship of emotion and
memory might change across the life span.

Relatively Well-Maintained Neural

Mechanisms of Emotion

Amygdala The emotional systems in the brain consist of interacting circuits

involving the prefrontal cortex, amygdala, and anterior cingulate. The amygdala
seems to play the most critical role in terms of the interaction between emotion
and memoryat least for younger adults. [For a general review of emotion,
memory, and the brain, also see chapter 2Eds.] Neural and hormonal mechanisms mediated by the amygdala enhance memory for emotional stimuli
(Hamann, 2001). It is highly interconnected with both cortical and subcortical
regions (Young, Scannell, Burns, & Blakemore, 1994) and so is in a position to
integrate disparate sources of information and aect a wide range of functions.
In particular, with its numerous connections to the hippocampus and other
medial temporal regions, the amygdala is ideally situated to inuence memory
processes.
Nevertheless, an intact amygdala is not needed to remember emotionally
neutral material (e.g., Bechara, Tranel, Damasio, & Adolphs, 1995; Scoville &
Milner, 1957). Instead, the amygdala plays a modulatory role, enhancing
memory associated with emotionally arousing material, especially during the
consolidation period after the emotional event (McGaugh, 2000). Animal studies indicate that adrenal stress hormones and the amygdala interact to modulate consolidation of recently acquired information (McGaugh, Ferry, &
Vazdarjanova, 2000). Case studies of patients with bilateral lesions in the
amygdala have found that they do not show the normal enhancement in
memory for emotional material, although they remember nonemotional information as well as normal controls (Adolphs, Cahill, Schul, & Babinsky, 1997;
Cahill, Babinsky, Markowitsch, & McGaugh, 1995). Conversely, amnesic patients with intact amygdalae show an enhancement in memory for emotional
material despite their overall memory impairment (Hamann, Cahill, McGaugh,
& Squire, 1997).
Brain imaging studies also provide evidence that the amygdala plays a critical role in enhanced long-term memory for emotionally arousing events (Cahill
et al., 1996; Canli, Zhao, Brewer, Gabrieli, & Cahill, 2000; Canli, Zhao, Desmond,
Glover, & Gabrieli, 1999). For example, in one study, participants received a PET

276

scan while watching either emotionally negative or emotionally neutral lms.

Greater amygdala activity while watching the negative lms predicted better
memory for these lms later, whereas there was no such correlation for neutral
lms (Cahill et al., 1996). Recent ndings indicate that the amygdala is also associated with enhanced memory for emotionally positive information (Hamann,
Ely, Grafton, & Kilts, 1999).
With age, brain volume decreases (for reviews, see Kemper, 1994; Raz, 2000).
This decrease, however, does not occur at an equal rate across all regions of the
brain. Aging has relatively little impact on the amygdala volume, especially
when compared with the atrophy in other regions such as the prefrontal cortex.
Although MRI studies focusing on the hippocampus or amygdala often nd decreases in volume with age (e.g., Bigler, Anderson, & Blatter, 2002; Jack,
Petersen, OBrien, & Tangalos, 1992; Jack et al., 1997; cf. Lim, Zipursky,
Murphy, & Pfeerbaum, 1990; Mu, Xie, Wen, Weng, & Shuyun, 1999; Smith,
Malcein, et al., 1999), studies that include a broader selection of brain regions
nd that the amygdala-hippocampus complex shows little or no change relative
to other regions of the brain (e.g., Coey et al., 1992; Good et al., 2001; Ohnishi,
Matsuda, Tabira, Asada, & Uno, 2001; Raz et al., 1997; Sullivan, Marsh,
Mathalon, Lim, & Pfeerbaum, 1995) or decreases at a rate no greater than that
of the volume of the whole brain (Tisserand, Visser, van Boxtel, & Jolles, 2000).
Furthermore, postmortem studies using stereologically based cell-counting techniques nd almost no changes with normal aging in the hippocampus (Raz,
2000; West, Coleman, Flood, & Troncoso, 1994) or in the entorhinal cortex
(Gomez-Isla et al., 1996), which is closely connected to the hippocampus and
limbic system. This pattern suggests that the enhancement in memory for emotional material due to the interaction of the amygdala and hippocampus should
remain relatively intact in older adults.
Prefrontal Cortex Researchers investigating memory and aging have become
interested in the role of the prefrontal cortex (PFC). The PFC shows a disproportionately large decrease in volume with age, compared with other regions involved in memory processes, such as medial temporal regions (Coey et al., 1992;
Cowell et al., 1994; DeCarli et al., 1994; Raz, 2000; Raz et al., 1997; West,
1996). Patients with lesions in the frontal lobes show decits in strategic memory
tasks that are striking in contrast with their intact performance on nonstrategic memory tasks (Shimamura, 1995). Older adults show a similar, although
less pronounced, disproportionate decit for strategic memory tasks (for reviews,
see Johnson & Raye, 2000; Light, 2000; Moscovitch & Winocur, 1995; Prull
et al., 2000). Furthermore, older adults performance on such tasks is correlated
with their performance on neuropsychological tests usually associated with frontal lobe damage (Craik, Morris, Morris, & Loewen, 1990; Fabiani & Friedman,
1997; Glisky, Polster, & Routhieaux, 1995; Henkel, Johnson, & De Leonardis,

277

1998; Mather, Johnson, & De Leonardis, 1999; McDaniel, Glisky, Rubin, Guynn,
& Routhieaux, 1999; Schacter, Kazniak, Kihlstrom, & Valdiserri, 1991).
The PFC also plays an important role in aective processing (for a review, see
Davidson & Irwin, 1999). Brain imaging studies reveal activity in the PFC during the normal experience of positive and negative emotions (George et al., 1995;
Kimbrell et al., 1999; Lane, Reiman, Ahern, Schwartz, & Davidson, 1997; Lane,
Reiman, Bradley, et al., 1997). Patient data going back to the famous case of
Phineas Gage indicate that dysfunction of the prefrontal cortex disrupts the regulation of emotion and social conduct (Damasio, Grabowski, Frank, Galaburda,
& Damasio, 1994; Dimitrov, Phipps, Zahn, & Grafman, 1999). An extreme example of a failure of social and emotional regulation is the act of killing someonein particular, when this murder is not an act of self-defense. A positron
emission tomography (PET) study of 41 murderers who had pleaded not guilty
by reason of insanity found that, compared with normal controls, the murderers showed reduced glucose metabolism in the PFC (Raine, Buschbaum, &
LaCasse, 1997).
On the face of it, this pattern seems paradoxical. Normal aging leads to disproportionate declines in the PFC yet is not associated with the types of problems regulating emotions and social behavior that are the hallmark of many
patients with frontal lobe damage. In fact, in addition to being better able to regulate their own emotional experience, older adults are also better able to control
their expressions of negative emotions. For example, they are less likely than
younger adults to report having a tendency to lose their temper quickly
(McConatha & Huba, 1999). This pattern suggests that some prefrontal regions
are more critical for the regulation of emotion and that prefrontal regions specialized for emotional processing are less subject to age-related decline than other
regions of the PFC.
Considering that the frontal lobes occupy about a third of the human cortex,
it is not surprising that thinking about the PFC as a monolithic unit can be misleading. In terms of the role of the PFC in aging, emotion, and memory, it seems
especially important to consider the dorsolateral PFC and the orbital PFC as discrete regions (Phillips & Della Sala, 1998). These two regions have separate anatomical pathways to communicate with subcortical regions and seem to mediate
very dierent processes (Masterman & Cummings, 1997). The orbital PFC seems
to be essential for regulating emotion (see Davidson, Putnam, & Larson, 2000,
for a review). Patients with lesions in orbitofrontal regions have diculty reversing responses previously rewarded but no longer rewarded (Rolls, 2000) and
show disinhibited behavior (Bechara, Damasio, Tranel, & Anderson, 1998). They
also show emotion regulation decits in gambling tasks, as they are driven by
short-term emotional payos, ignoring more signicant future emotional consequences (Bechara, Damasio, Damasio, & Anderson, 1994; Bechara, Damasio,
& Damasio, 2000). On the other hand, the dorsolateral PFC seems particularly
important for strategic memory processes. Declines in the volume and activity

278

of the dorsolateral PFC in it have been associated with older adults decit in strategic memory tasks (for a review, see Raz, 2000).
Unfortunately, there is little information about which areas of the PFC are
most aected by normal aging. Most studies examining age-related changes in
the frontal lobes have not considered subdivisions within the frontal lobes. Recent evidence, however, indicates that the orbital PFC may be selectively spared
in comparison to other areas of the PFC (Salat, Kaye, & Janowsky, 2001). In
contrast, the study by Salat et al. (2001) suggests that regions in the medial and
lateral PFC and anterior cingulate are more likely to be aected by aging. Thus,
preliminary evidence suggests that at least one region in the PFC that is important for regulating emotion is comparatively spared by aging.
Hemispheric Asymmetry Indirect techniques for examining laterality and emotion suggest that the right hemisphere plays a larger role than the left hemisphere
in processing emotions (for a review, see Kolb & Whishaw, 1990). For example,
people tend to make eye movements toward the left during emotional tasks and
to the right during nonemotional tasks (Schwartz, Davidson, & Maer, 1975), and
the emotional aspect of facial expressions tend to be more pronounced on the
left side of the face (Moscovitch & Olds, 1982). There is also evidence of laterality
eects for emotional valence, as people have a more negative response to lms
shown to the right than left hemisphere (Dimond, Farrington, & Johnson, 1976).
Data from case studies of patients with lesions suggest that this laterality eect
occurs in the frontal lobes, with the left frontal lobe normally mediating positive
emotions and the right frontal cortex mediating negative emotions (Grafman,
Vance, Weingartner, Salazar, & Amin, 1986). Furthermore, neuroimaging and
electrophysiological studies with normal younger adults reveal right PFC activation during negative emotional states (Davidson & Irwin, 1999).
Some researchers have proposed that certain age-related cognitive decits
occur because in aging the right hemisphere deteriorates faster than the left hemisphere (Ellis & Oscar-Berman, 1989; Evert & Oscar-Berman, 2001; Goldstein &
Shelly, 1981; Johnson et al., 1979; Nebes, 1990; Schaie & Schaie, 1977). Researchers suspect that the right hemisphere is more aected by aging because older
adults score as well as younger adults on verbal subtests of intelligence tests but
are impaired on subtests measuring visuospatial skills. This dissociation is consistent with the lateralization of verbal and spatial functions to the left and right
hemispheres, respectively. Furthermore, the performance of normal elderly persons on intelligence tests is similar to that of patients with damage restricted to
the right hemisphere (Schaie & Schaie, 1977).
A major problem with this hypothesis, however, is that the tests sensitive to
right-hemisphere damage tend to require more complex problem solving than
those sensitive to left-hemisphere damage. It may be the complexity of the tests
that leads to older adults pattern of performance, rather than their reliance on
the right hemisphere. Thus, a number of investigators have attempted direct tests

279

of the right-hemisphere hypothesis by comparing younger and older adults

performance on tasks in which the stimuli are tachistoscopically presented to
either the right or left visual eld. Most visual half-eld experiments have not
found age dierences in hemispheric asymmetry (e.g., Hutner & Oscar-Berman,
1996; Mittenberg, Seidenberg, Oleary, & Digiulio, 1989; Nebes, 1990; Nebes,
Madden, & Berg, 1983; Obler, Woodward, & Albert, 1984). In addition, studies
specically examining hemispheric asymmetry for the perception of emotional
stimuli have not found age dierences in how dominant the right hemisphere is
for processing emotions (Cherry, Hellige, & McDowd, 1995; Hutner & OscarBerman, 1996; Moreno, Borod, Welkowitz, & Alpert, 1990; Reminger, Kaszniak,
& Dalby, 2000). Finally, measurements of the volume of the right and left hemispheres have failed to nd evidence for greater age-related decreases in the right
hemisphere (e.g., Good et al., 2001). Thus, there is little evidence to show that
older adults have a selective decline in the right hemisphere that aects emotional processing.
In contrast, there is substantial evidence showing that prefrontal activity
associated with memory and other cognitive processes becomes less lateralized
with age (Cabeza, 2002). For example, brain activity associated with episodic
retrieval is right-lateralized among younger adults (Nyberg, Cabeza, & Tulving,
1996) but becomes more bilateral with age. This decreased lateralization may
reect dedierentiation of brain regions, changes in strategy, or compensatory
mechanisms (in which analogous regions on the contralateral side are recruited
to assist processing). In any case, this reduction in lateralization for memory
processes does not lead to any strong predictions about changes in the impact of
emotion on memory. One possibility, however, is that it interacts with the lateralization of positive and negative aect also seen in the PFC (Davidson & Irwin,
1999). The decrease seen with age in the right lateralization of memory processes
may lead to decreases in the degree to which memories are negative. Because
we do not know whether laterality eects for valence also decrease with age, this
possibility is only speculative.
Overview of Emotion and the Aging Brain In summary, studies investigating
age-related changes in the brain suggest that brain regions typically associated with emotional processes are relatively well preserved in older adults. The
amygdala shows little deterioration, and there does not appear to be greater
right than left hemisphere decline with aging. There is also some indication
that orbitofrontal regions associated with emotional control in the PFC are less
aected by aging than other parts of the PFC. Altogether, these ndings suggest that the impact of emotion on memory will still occur among older adults.
In fact, the impact of emotion might even be amplied, as emotional processes
may gain more inuence than other processes supported by brain regions that
deteriorate faster.

280

Summary of Predictions for

Memory Based on Research

on Emotion and Aging

The behavioral studies reviewed in the preceding sections indicate that emotional
processes gain centrality and eectiveness with age. In addition, the neurological
studies indicate that the brain regions supporting emotional processing are relatively spared the eects of age. The one major exception to this pattern of maintenance is that physiological intensity of emotion in the peripheral nervous system,
as measured by heart rate, declines with age. Nevertheless, the reduced physiological responsiveness in older adults does not seem to aect their subjective emotional
reactions or experience (Levenson et al., 1991, 1994; Tsai et al., 2000).
The ndings regarding emotional functioning in these dierent domains lead
to two major hypotheses about how the impact of emotion may dier for older and
younger adults. First is the emotional compensation hypothesis: well-maintained
emotional processes can help older adults remember information they otherwise
would have forgotten. Second is the goal-directed emotional memory hypothesis:
in general, memory should become more emotionally gratifying, as older adults
focus on regulating emotion more than younger adults. In the following section,
I review ndings from studies examining age dierences in the eects of
emotion on memory to see if either or both of these predictions hold.

Evidence for Changes in the Role

of Emotion in Remembering

as We Age

Memory for Emotional Versus

Neutral Information

Our most vivid and powerful memories involve emotional events (e.g., Brown &
Kulik, 1977; Reisberg, Heuer, McLean, & Oshaughnessy, 1988; Rubin & Kozin,
1984). Several studies that look at memory for emotional information (but do
not distinguish positive from negative information) suggest that memory accuracy for emotional information does not decrease with age as much as memory
for neutral information. In one study, older and younger participants participated
in scripted situations (such as packing a picnic basket) and imagined participating in others (Hashtroudi, Johnson, & Chrosniak, 1990). After intervening tasks,
they rated their memories for half the situations they had experienced, using a
memory characteristics questionnaire that assessed the amount of perceptual
and contextual detail and thoughts and feelings associated with their memories.
The following day, they rated their memories of each situation and then recalled

281

as much as they could about them. After this 1-day delay, older adults rated
thoughts and feelings as more memorable than younger adults did. In contrast,
this age dierence did not appear when participants rated situations immediately
after they experienced them. This pattern suggests that, for older adults, the
salience of thoughts and feelings increases over time. In addition, in their actual
recall of the situations, older adults reported more thoughts and feelings and
evaluative statements than did younger adults, whereas younger adults reported
more colors, objects, actions, spatial references, and nonvisual sensory references
than did older adults.
Days later, Hashtroudi et al. (1990) asked the older and younger adults to
make source attributions for each situation, indicating whether they participated
in it or just imagined participating in it. Older adults seemed to emphasize their
thoughts and feelings when making these source judgments, as their condence
in their recollections correlated more highly with their ratings of associated
thoughts and feelings for each event than those of younger adults (Johnson &
Multhaup, 1992).
Externally generated emotional information also seems to gain advantage in
memory as one ages. One study suggesting that emotional information gains
salience examined memory for a narrative that contained equivalent numbers
of neutral (e.g., Celia sipped some more sherry) and emotional (e.g., and
looked very searchingly at Mrs. Oliver) phrases (Carstensen & Turk-Charles,
1994). Four age groups from 20 to 83 years old were tested. Each successive age
group recalled a greater portion of emotional information, so that a signicant
linear trend across the life span emerged. Older adults also recalled a larger proportion of emotional information than younger adults in a study that examined
memory for advertisements (Fung & Carstensen, in press). Participants saw a
series of advertisements (such as a picture of a camera identied with a brand
name and accompanied by a slogan). Some participants saw a particular advertisement picture and brand name with a slogan that appealed to emotional goals
(e.g., Capture those special moments), whereas others saw it with a slogan that
suggested expanding horizons or achieving success in the future (e.g., Capture
the unexplored world). Younger adults remembered these dierent types of slogans equally well, but older adults remembered the emotional slogans better than
the expanding horizons slogans.
One of the challenges of doing research on emotion and memory is counterbalancing emotional and neutral material. Dierences between memory for the
two types of stimuli can sometimes result from something other than the dierence in emotionality. The age dierences in Fung and Carstensens (in press)
study, however, occurred not only for the slogans but also for the brand names
even though the brand names themselves were not emotional, and each brand
name appeared equally often with each type of slogan.
In addition, another recent study suggests that older adults remember the
same information better if it is presented in an emotionally evaluative frame

282

(Rahhal, May, & Hasher, 2002). Older and younger participants listened to a tape
of a male and a female speaker reading trivia. They were told that all of the statements read by the female speaker were true and all of the statements read by the
male speaker were false (or vice versa). After a short delay, they completed a
source memory test in which they were given a list of the statements that had
been read and were asked either who said the statements (a perceptual source
task) or whether they were true or false (a conceptual, emotionally valenced
source task). As many studies have shown, older adults were less able than
younger adults to indicate who said each statement. Nevertheless, they were just
as accurate as the younger adults at indicating whether the statements were true
or false, even though this information had been conveyed by the gender of the
speaker. This pattern was replicated when the conceptual, emotionally valenced
task changed to one of evaluating the character of a person (good or evil) (Rahhal
et al., 2002) or the safety of an item (May, Rahhal, Leighton, & Berry, 2002).
Thus, for example, participants would be told that everyone the female speaker
described was good and everyone the male speaker described was evil. If they
heard the female speaker describe someone named Sally and were later asked if
Sally was good or evil, they were as good as younger adults at making this judgmentbut if asked whether Sally was described by the male or female, they were
less accurate than younger adults.
Each of the studies reviewed in this section suggests that older adults remember emotional aspects of events better than neutral events and thus conrm a
prediction generated from both the research on emotional goals and on brain
regions associated with emotional processes. None of these studies distinguished
positive from negative information, however, so the predictions regarding emotionally gratifying information were not addressed. I address studies that examine age dierences in memory by valence in a subsequent section.

A Dierent Story: Alzheimers Disease

and Emotional Memory

Unlike normal aging, Alzheimers disease has a dramatic impact on the

amygdala. Compared with age-matched controls, Alzheimers patients show a
signicant reduction in volume of their amygdalae, with estimates ranging from
20% to 35% (Callen, Black, Gao, Caldwell, & Szalai, 2001; Cuenod et al., 1993;
Herzog & Kemper, 1980; Smith, Malcein, et al., 1999) to as much as 55% (Scott,
DeKosky, & Sche, 1991). In addition, Alzheimers disease reduces amygdalar
volume more dramatically than it reduces brain volume overall (Scott et al.,
1991) or other individual structures (Cuenod et al., 1993; Kemper, 1994).
Not surprisingly, this degeneration of the amygdala has an impact on emotional memory. A study examining memory for the Kobe earthquake (which
measured 7.2 on the Richter scale and caused more than 6,000 deaths) found
that among Japanese Alzheimers patients, the accuracy of autobiographical

283

memory for the earthquake was positively correlated with the volume of their
amygdala as measured by an MRI (Mori et al., 1999). Furthermore, compared
with other older adults, patients with Alzheimers disease do not show the enhancement in memory for emotional rather than neutral pictures (AbrisquetaGomez, Bueno, Oliveira, & Bertolucci, 2002; Hamann, Monarch, & Goldstein,
2000; Kensinger, Brierley, Medford, Growdon, & Corkin, 2002) or the enhancement in memory for emotional words (Kensinger et al., 2002). Alzheimers patients are also impaired when learning to associate a neutral stimuli with an
aversive outcome (Hamann, Monarch, & Goldstein, 2002).1
Thus, older adults with Alzheimers disease show a very dierent memory
pattern than that seen with normal aging. Whereas emotion enhances memory
as much as or more for older adults than it does for younger adults, this emotional enhancement decreases or disappears altogether for Alzheimers patients.

Impact of an Aective Focus on

Nonaective Information

Many researchers have noted that older adults memories are more inuenced
by the gist or schematic information about an event than are younger adults
memories. Older adults are more likely to falsely recognize new words that are
semantically associated with studied words (Balota et al., 1999; Isingrini,
Fontaine, Taconnat, & Duportal, 1995; Kensinger & Schacter, 1999; Norman
& Schacter, 1997; Rankin & Kausler, 1979; Smith, 1975; Tun, Wingeld, Rosen,
& Blanchard, 1998), to falsely recognize new pictures that are categorically related to studied pictures (Koutstaal, Schacter, & Brenner, 2001; Koutstaal,
Schacter, Galluccio, & Stofer, 1999), to falsely recognize schema-consistent objects as having been in a scene (Hess & Slaughter, 1990), and to incorrectly attribute schematically related statements to a speaker associated with that schema
(Mather et al., 1999).
Perhaps older adults greater schema reliance occurs because they engage
more often in emotion-focused memory processes. For two reasons, we may assume that being more emotion focused will increase reliance on schematic
knowledge. First, an emotional self-focus seems to shift peoples focus away from
the event itself, instead emphasizing connections between the event and oneself.
That is, asking people to rate how they feel when hearing statements enhances
memory for the semantic content but impairs memory for associated contextual
details, such as who made the statement (Johnson, Nolde, & De Leonardis, 1996;
Mather et al., 1999; Suengas & Johnson, 1988). Lacking memory of specic perceptual and contextual details associated with the event in memory, people may
rely more on their schematic knowledge about the event to help reconstruct it
later. In addition, the thoughts and inferences generated while thinking about
ones reactions to an event may draw on schematic knowledge about the event
and may then later become confused with the event itself. In summary, an emo-

284

tional self-focus may increase how schematically an event is remembered, as a

result of how attention is allocated and how related information becomes associated with the event.
One study that supports the hypothesis that an emotional focus can lead related thoughts and inferences to be confused in memory with the event itself
asked older and younger adults to participate in a play and then review it
(Hashtroudi, Johnson, Vnek, & Ferguson, 1994). In the rst part of the experiment, two participants were prompted to repeat or think about lines of a play.
The experimenter read each line of the play to the pair of participants and then
indicated which one should say or think the line. There were three review conditions. In the factual focus condition, participants were asked to talk and then
think about what had been said by both participants in each scene. In the aective focus condition, participants were asked to talk and then think about how
they felt during each scene. Finally, in the control condition, participants were
asked to talk and then think about anything regarding the play. After rehearsing the scenes, half of the participants in each review condition were given a
source-monitoring test, in which individual statements from the play were presented, as well as semantically related distractors. These participants were asked
to indicate whether the statement had been said or thought and which person
had said or thought itor whether the statement had not been in the play. Examining participants source-monitoring accuracy revealed a signicant interaction between age and type of focus. Whereas the type of focus did not aect
younger adults source-monitoring accuracy, older adults were signicantly less
accurate in the aective and control conditions than in the factual focus condition. The other half of the participants recalled the play. Both older and younger
participants in the aective focus condition remembered less of the play. In addition, participants in the aective focus condition recalled more inferences about
and elaborations on the play. In fact, 51% of what the older adults recalled and
38% of what the younger adults recalled in the aective focus condition included
elaborative information (compared to less than 30% in both other conditions).
Thus, an emotional focus can lead ones inferences and related thoughts about
an event to become confused with the event itself when one is remembering.
This combination of greater recall of inferences and reduced recall of perceptual information after focusing on the emotional aspects of events may lead to more
schema-reliant memories after focusing on ones feelings. A study examining
memory for choices is consistent with this possibility (Mather & Johnson, 2000).
In this study, older and younger adults made a series of two-option choices (e.g., a
choice between two job candidates). Each option had positive and negative features (e.g., Seemed quite motivated, Has little professional experience). After
making the choices, one group of participants reviewed how they felt about each
choice, another group reviewed the details of the choice options, and a third did
an unrelated task. After a delay, participants were given memory tests for each
decision that included the features from each option intermixed with new features.

285

They were asked whether each feature had been associated with the rst option,
the second option, or neither option. One piece of schematic knowledge about the
decisions that might have inuenced memory is the belief that chosen options were
better than rejected options. If this belief were to aect memory, it should do so by
increasing the number of choice-supportive memory attributions. Positive features
should more likely be attributed to the chosen than to the rejected options, and
negative features should show the opposite pattern. In averages across conditions,
older adults were more choice-supportive than younger adults. The focusing condition did not aect how choice-supportive older adults were. Younger adults,
however, were signicantly more choice-supportive in the emotional-review condition than in the other two conditions and were as choice-supportive as the older
adults only in the emotional-review condition.
However, increased inuence of general knowledge or beliefs on memory after
engaging in aective focus may occur only for aectively signicant situations
such as choices. Remembering that the option you chose was the better option
should make you feel less regret and more pleasure. Reviewing a choice may
therefore activate emotional goals that would not be activated in other contexts.
To investigate whether an aective review would increase the inuence of schematic knowledge on memory, even when the schema should not lead to more
positive emotion, Mather and Johnson (2003) conducted a similar study with
dierent materials. They used a story that included several unstated inferences
as the to-be-remembered material rather than choice options. Participants read
the story and then were asked either to review how they felt about it, to review
its details, or to do an unrelated task. Both older and younger adults in the
aective-review condition were very likely to falsely recognize the unstated inferences from the story. The factual review, however, helped reduce younger
adults schema reliance much more than it reduced older adults schema reliance. Thus, across two studies, older adults were more schema-reliant than
younger adults, except when participants were told to focus on their feelings. This
pattern suggests that emotional processing increases schema reliance.
Interestingly, the amygdala may help mediate this eect, as patients with
amygdala damage have poorer memory for gist and superior memory for detail
(Adolphs, Denburg, & Tranel, 2001). In the Adolphs et al. study, participants
were shown a series of photographs accompanied by brief descriptions. Normal
controls and patients with unilateral amygdala damage all showed better
memory for general information associated with aversive photographs than for
their visual details. In contrast, a patient with bilateral amygdala damage had
better memory for the detail than for the general information. Along these same
lines, studies with undergraduates have found that emotional stimuli enhance
memory for central information but have a detrimental eect on peripheral details (Burke, Heuer, & Reisberg, 1992; Christianson, 1992; Reisberg & Heuer,
1992). As outlined earlier, the amygdala shows relatively little decline with age;

286

thus, the enhancement it provides for memory for general information about
emotional scenes is probably maintained.

Flashbulb Memories and Aging

People often remember the circumstances of learning about shocking events with
great vividness and detail. Brown and Kulik (1977) called these ashbulb memories because they often seem as clear and vivid as if a ashbulb camera went o
and recorded every detail of the news. To help explain their ndings of extremely
vivid memories for surprising and emotionally arousing events, Brown and Kulik
suggested that on recognizing that an event is unexpected and important, the limbic system issues a Now print! order for memory (Livingston, 1967).
Subsequent research has conrmed that, although ashbulb memories are
far from exact snapshots of the surprising moment, one remembers the circumstances of learning about shocking events better than those of ordinary events
(e.g., Christianson, 1989). The strength of emotional reactions helps predict how
accurate later recall will be (Conway et al., 1994; Pillemer, 1984), especially over
a longer delay (Schmolck, Bualo, & Squire, 2000).
Most of what we know about aging and memory predicts that older adults
will be less likely to have accurate and vividly remembered ashbulb memories.
According to Brown and Kuliks (1977) original formulation, a ashbulb
memory includes information for contextual information associated with
when the person rst learned about the event, such as the source of the information, the place, the other people present at the time, and what else was happening at the time. Older adults have particular diculty remembering this type
of contextual and source information (e.g., Hashtroudi, Johnson, & Chrosniak,
1989; McIntyre & Craik, 1987; Spencer & Raz, 1995), suggesting that they will
be less likely to have ashbulb memories.
Nevertheless, evidence about age dierences in the vividness and accuracy
of ashbulb memories is mixed. Wright, Gaskell, and OMuircheartaigh (1998)
conducted surveys that included several thousand British participants representative of the larger population. Participants were asked about their memories for Margaret Thatchers resignation as prime minister of Great Britain and
for their memories of a semi-nal England Football Association Cup match in
which 96 people were crushed to death in full view of a national television
audience. To the surprise of the researchers, 19 months after Thatchers resignation and 36 months after the football tragedy, self-ratings for memory clarity for these events were greater the older the respondent was, until after age
75, when they returned to the levels in the younger adult cohort. In contrast,
12 years after President Kennedys assassination, Yarmey and Bull (1978)
found less detailed accounts of the circumstances of learning of the assassination in people older than 66.

287

Neither the Wright et al. (1998) nor the Yarmey and Bull (1978) studies included objective measures of accuracy. Studies of younger adults ashbulb
memories show that objective measures are essential because many highly condently recalled ashbulb memories include quite inaccurate vivid details (e.g.,
Neisser & Harsch, 1992; Schmolck et al., 2000). In particular, among younger
adults, ashbulb memories are quite susceptible to distortion in which memory
for the original circumstances becomes confused with other related memories.
For example, when recalling how they learned about the space shuttle Challenger
disaster 3 years after it occurred, many undergraduates remembered learning
about it from a TV broadcast, even though they actually had learned about it
from a friend or some other source (Neisser & Harsch, 1992). Older adults are
more likely than younger adults to exhibit memory distortion as a result of
misattributing information they learned from one source to another source (e.g.,
Henkel et al., 1998; Mather & Johnson, 2000; Mather et al., 1999; Mitchell,
Johnson, & Mather, 2003). So older adults ashbulb memories likely will be
more susceptible to memory distortion caused by source monitoring errors. Thus,
although the participants in the Wright et al. study remembered the shocking
events more vividly the older they were, they also probably remembered them
less accurately.
Two additional studies examining age dierences in ashbulb memories
measured recall within a few weeks of the event and again some months later
(Cohen, Conway, & Maylor, 1994; Davidson & Glisky, 2002). Though participants rst memory reports were probably not exact representations of what
happened, these studies could at least assess the consistency of the ashbulb
memories over timea reasonable proxy for measuring objective accuracy.
These studies provide mixed evidence for the hypothesis that older adults have
less accurate ashbulb memories. In the rst study, groups of younger and older
adults recounted how they heard the news of Thatchers resignation (Cohen
et al., 1994). They were tested within 2 weeks of the resignation and again
11 months later. Older adults responses at the two test points were signicantly
less likely to be consistent than those of younger adults.
In the second study examining age dierences in ashbulb memory consistency, groups of younger and older adults recounted how they learned about
the deaths of Princess Diana and Mother Theresa within 3 weeks of each incident and again about 6 months later (Davidson & Glisky, 2002). In this study,
there were no signicant dierences between the consistency of younger and
older adults memories. In addition, among the older adults, there were no
signicant correlations between the consistency of their memories for the
deaths and scores on a battery of tests associated with medial temporal lobe
function or a battery of tests associated with frontal lobe function. In contrast,
memory consistency for a control event (the most interesting event in their own
lives that occurred on the same weekend when Princess Diana died) was correlated with performance on these batteries (although only marginally with

288

the frontal battery). A number of studies examining older adults source

memory performance have found signicant positive correlations between
source attribution accuracy and performance on tests of frontal lobe function
(Craik et al., 1990; Glisky et al., 1995; Mather et al., 1999) and sometimes
also between source accuracy and performance on tests of medial temporal lobe
function (Henkel et al., 1998; Mather et al., 1999). The fact that memories
were as accurate for the surprising public events among older adults with low
frontal function as among those with high frontal function suggests that other
brain regions (such as the amygdala) may be involved in forming and consolidating memories of surprising events. Processes mediated by these regions may
compensate for decline in frontal regions.
In summary, the research on older adults and ashbulb memories is somewhat surprising because older adults do not always show less accurate memories than younger adults, as one might expect from their memory performance
in similar contexts. The emotional aspect of shocking events may be a critical
component of this greater-than-expected accuracy. More research is needed,
however, in which older and younger adults memory for shocking events is
compared with memory for interesting but nonemotional events. The emotional
compensation hypothesis outlined earlier in the chapter predicts that age dierences in forgetting and distortion should be greater for the control events than
for the shocking events. Also of interest is whether the valence of the event interacts with age dierences at all. Most ashbulb memory studies have looked
at memories for negative events, but memories for positive surprising events can
also have ashbulb-like qualities (Scott & Ponsoda, 1996). As outlined in the
following section, valence seems to have a dierent impact on younger and older
adults memories.

Evidence for Increasingly Gratifying

Memories With Age

For younger adults, negative information dominates positive information in

many contexts (for reviews, see Baumeister, Bratslavsky, Fickenauer, & Vohs,
2001; Rozin & Royzman, 2001). Negative information is processed more thoroughly, negative impressions are more dicult to disconrm, and negative emotions have more impact. Nevertheless, the evidence is mixed for younger adults
memory, with some studies nding a bias for negative and some for positive information. There is growing evidence, however, that with age comes a shift toward favoring positive rather than negative information in memory. As I show
in this section, evidence for this shift has been seen both in longitudinal studies
investigating personal autobiographical memories and in lab studies investigating memory for emotional pictures or verbal stimuli.
Among younger adults, memories for highly negative events seem to be accessible for a longer time than memories of highly positive events (Berntsen,

289

2001, 2002). A survey of over a thousand Danish citizens, however, revealed

that this pattern in which negative memories last longer than positive memories reverses itself in the late 30s (Berntsen & Rubin, 2002). By age 60, there is
dramatically reduced recall of remote negative events compared to recall of remote positive events. When asked to describe memory for an extremely happy
event, older adults tended to pick events from their 20s, whereas when asked to
describe a memory for an extremely sad or traumatic event, they picked more
recent events. This pattern of recall for sad and traumatic events is contrary to
ndings indicating that, in general, people remember information encoded during adolescence and early adulthood better than information encountered earlier or later in life (Rubin, Rahhal, & Poon, 1998). Berntsen and Rubin (2002)
conclude that the likely explanations of the reduced recall of remote sad and traumatic events are reduced rehearsal of negative events or repression.
The decreasing accessibility of negative long-term personal memories across
the life span may help explain an intriguing nding from a longitudinal study
(Field, 1981). Sixty adults were interviewed four times over a 44-year period and
were asked during each interview to rate how happy their childhood had been.
There was a highly signicant trend over time, as participants rated their childhood as increasingly happier the older they became.
A study of political supporters of Ross Perot indicates that memory for the
intensity of past sadness decreases with age (Levine & Bluck, 1997). After Perot
abruptly withdrew from the presidential race in July 1992, some of his supporters were asked to rate their initial emotional reactions. After the elections that
November, they were asked to recall their initial emotional reactions. There were
no age dierences in the initial reports of emotion intensity. Nevertheless, the
follow-up questionnaire revealed that, among those participants who still wished
that Perot had been elected, older adults remembered experiencing less intense
sadness than younger adults did. The older the participants, the more they underestimated how sad they had been when Perot withdrew from the race. Interestingly, there were no age dierences for supporters who no longer wished that
Perot had been electedthese supporters all tended to underestimate how sad
they had been.
A study in which older and younger adults were asked about recent memories also found that younger adults remember negative emotions more intensely
than older adults. Older and younger adults carried electronic pagers for a week
and lled out a questionnaire about their current emotions each time they were
paged at random intervals throughout each day (Carstensen et al., 2000). At
the end of the week, they were asked to recall the percentage of time they had
experienced positive and negative emotions and their concomitant intensities
(Kennedy, 2002). There were no age dierences in the accuracy of estimates for
the percentage of time emotions had been experienced, but younger adults were
signicantly more likely than older adults to overestimate how intense their
negative aect had been.

290

Decreases for negative information are also seen in memory for emotional
pictures (Charles, Mather, & Carstensen, 2003). Younger (1829 years), middleaged (4153 years), and older (6580 years) adults watched a slide show of negative, positive, and neutral pictures displayed in a random order. Half the pictures
were of people (e.g., a couple mourning at a tombstone, a happy family, scuba
divers) and half did not contain people (e.g., a slice of pizza with a cockroach on
it, two bunnies, a chair). The participants included equal numbers of men and
women, European Americans and African Americans, and white- and bluecollar workers within each condition. After seeing the pictures and completing
a 15-minute ller task, participants were asked to recall as many of the pictures
as they could. Then, they completed a recognition memory test in which they
saw a series of pictures and indicated which ones they had seen before. The proportion of total recall that consisted of negative pictures decreased across the life
span, whereas the proportion that consisted of positive pictures increased. Recognition accuracy for negative pictures similarly decreased. These age dierences
were consistent across the dierent genders, ethnicities, and socioeconomic statuses represented among the participants.2
Another study focusing on memory for emotional pictures examined whether
there are age dierences in brain activity at the time of encoding the pictures
(Mather et al., 2001). Younger and older adults watched positive, negative, and
neutral pictures while in a functional magnetic resonance imaging scanner. For
each picture, they rated how emotionally arousing it was for them. After their scan,
they were given a recognition memory test for the pictures in which they were
asked whether they had seen each test picture while they were in the scanner. They
were then asked to indicate whether they remembered it vividly or just knew it
had been shown in the scanner (R/K; e.g., Gardiner & Java, 1993). Older adults
rated the negative pictures as less arousing and rated the positive pictures as more
arousing than the younger adults did. As in the study by Charles et al. (2003), there
was a signicant age by valence interaction in recognition memory performance,
indicating that older adults showed a disproportionate disadvantage in recognizing the negative pictures. Furthermore, when older adults did actually recognize
a negative picture from the scanner session, they were less likely to say they vividly remembered it than they were for the positive pictures they recognized.
As I said earlier, the amygdala appears to be relatively unscathed by aging.
Mather et al. (2001) found that for both older and younger adults, the amygdala
was more active while emotional pictures were displayed than while neutral
pictures were displayed. There was an interaction, however. Older adults had
greater amygdala activity while seeing positive pictures, whereas the younger
adults had equivalent activation for the two types of emotional pictures. In addition, older adults showed more prefrontal activity while viewing positive pictures. These dierences in brain activity at the time of encoding suggest that older
adults bias to remember less negative information is at least partially a result of
processes operating at the time of encoding.

291

Another study suggests that biased attention at the time of encoding contributes to the increasingly emotionally gratifying memories seen with age (Mather
& Carstensen, in press). On each trial, older and younger adults were shown an
emotional and a neutral face side by side for 1 second. Then the faces disappeared,
and a dot probe appeared behind one of the faces. Participants had to indicate
the side of the screen the dot was on. In two experiments, older adults responded
faster if the dot was behind a neutral face than if it was behind a negative face.
They also showed a trend to respond faster if the dot was behind a positive face
than if it was behind a neutral face. Younger adults did not show any attentional
biases. In a later forced-choice recognition memory test, older adults were more
accurate at identifying which positive faces they had seen before than which
negative faces they had seen before, whereas this was not the case for younger
adults.
Increasingly emotionally gratifying memory with age was also found in two
studies focusing on memory for choices (Mather & Carstensen, 2003; Mather &
Johnson, 2000). These studies suggest that processes operating at the time of
retention and retrieval also contribute to older adults emotionally gratifying
memory biases. In the rst study, as described earlier in this chapter, older adults
were more choice-supportive than younger adults when attributing features to
options from past choices (Mather & Johnson, 2000). They were more likely to
attribute positive features to the options they had chosen than to the other options and were more likely to attribute negative features to the rejected options.
The age dierence in choice-supportiveness obtained even when older and
younger adults were equated for overall recognition and source attribution accuracy by decreasing the delay between encoding and test for the older adults.
Older adults choice-supportive biases obtained not only for features actually
associated with the options but also for new features misattributed to one of the
two options. Older adults choice-supportive biases for new features indicate that
at least some of the processes supporting emotionally gratifying memories operate at the time of retrieval.
The importance of retrieval processes is further highlighted by the second study
of age dierences in memory for decisions (Mather & Carstensen, 2003). Half the
participants were asked to recall the features from their choices before completing
a recognition and source attribution test for the features, whereas the other half
took the recognition test rst. Overall, there were age by valence interactions for
both recall and recognition. Younger adults remembered negative features from
the choice options much better than positive features. Older adults showed a less
extreme advantage for the negative features than the positive features. Particularly interesting, however, was the dierential impact of taking a recall test on later
recognition of positive and negative features for older and younger adults. Younger
adults were less likely to recognize positive features if they received the recall test
rst, whereas the test order did not have much of an impact on older adults recognition of positive features. Taking a recall test rst had the opposite eect for

292

recognition accuracy of negative features. Younger adults accuracy for negative

features was not aected by having taken the test, whereas older adults were less
likely to correctly recognize negative features if they had done the recall test rst.
This pattern suggests that for younger adults, attempting to recall information that
includes both positive and negative aspects may inhibit later memory for the positive aspects, whereas for older adults, recall may inhibit later memory for the negative aspects. Repeated retrieval may make older adults memories more positive
and younger adults memories more negative.
In addition, half of the participants in each condition completed the memory
tests after 10 minutes, whereas the other half completed the tests after 2 days.
Comparisons across these two delay groups revealed that older adults recognition accuracy for negative features declined more than did that of younger adults,
whereas their decline in accuracy for positive features was not much greater than
that of younger adults. Thus, older adults tended to forget the negative features
more quickly than the positive features. Thus, factors operating over the retention interval also play a role in making older adults memories more gratifying.
Apparently, the salience of emotional goals also plays an important role in
making memories more gratifying. Mather and Johnson (2000) found that
younger adults were more choice-supportive in their memories of their recent
decisions if they had been asked to review their feelings than if they had been
asked to review the details or to do an unrelated task. Only the emotion-focused
younger adults were as choice-supportive as the older adults were across all of
the conditions, suggesting that older adults were more likely to focus on emotion even when not explicitly told to think about their feelings.
Kennedy, Mather, and Carstensen (in press) also found that making emotional
goals salient increases gratication. In their study, 300 nuns between the ages
of 47 and 102 completed a questionnaire in which they were asked to remember their responses to a questionnaire they had completed 14 years earlier about
their health practices and medical history. Participants in the emotion-focus
condition completed a retrospective questionnaire in which they were asked to
rate their current feelings after each subset of questions. At the same points in
the questionnaire, participants in the accuracy-focus condition were reminded
to answer the questions as accurately as they could and asked to complete ratings regarding their subjective sense of remembering. There was also a control
condition questionnaire without any reminders. For age comparisons, the control group was divided into middle-aged (4765) and older (79102) groups.
Older adults showed a positive bias when remembering their past health practices and illnesses, whereas middle-aged adults showed a negative bias. Both
older and younger participants in the emotion-focused condition had a positive
retrospective memory bias, whereas both older and younger participants in the
accuracy-focused condition had a negative bias. Thus, older adults and emotionfocused participants showed the same pattern in their retrospective biases. These
ndings remained signicant after controlling for current mood and current

293

ratings of physical and mental health and emotional experience. Furthermore,

making emotional goals salient not only led to a positive memory bias but also
led emotion-focused participants to be in a better mood than accuracyfocused
participants after completing the questionnaire.
Further evidence that older adults positive memory biases contribute to their
enhanced well-being comes from a study comparing emotion while one is reminiscing with others about past events to emotion while one is engaged in other
social activities (Pasupathi & Carstensen, in press). In contexts in which participants were engaged in mutual reminiscing, older adults experienced signicantly
more positive emotion than younger adults. In contrast, in social contexts that
did not involve reminiscing, age was not signicantly correlated with positive
emotion. A second study found that age was associated with improved emotional
experience only during reminiscing about positive experiences. These ndings
suggest that, for older adults, remembering positive events can be an eective
way to regulate emotion.

Summary: Aging, Emotion,

and Memory

In contrast with functioning in the physical and cognitive domains, emotional

functioning does not deteriorate with age. Instead, if anything, it actually grows
more eective across the life span. Changing time perspective (Carstensen et al.,
1999), greater experience and maturity (Labouvie-Vief, 1998), and relatively
well-maintained brain regions associated with emotional processing may all
contribute to this intact functioning and help older adults remember emotional
information better than neutral information. Similarly, the pattern of fewer
neurological changes in brain regions associated with emotion than in other
regions suggests that emotional processes may be available to compensate for
decline in other processes that contribute to memory formation and retention.
This emotional compensation hypothesis is supported by studies in which
memory for emotional information is maintained better in old age than memory
for neutral information. For example, older adults remember their own internal
emotional reactions more vividly than younger adults do (Hashtroudi et al., 1990),
and a larger proportion of the externally perceived information they remember
consists of emotional information (Carstensen & Turk-Charles, 1994; Charles
et al., 2003; Fung & Carstensen, in press). Furthermore, at the time of retrieval,
older adults can remember the source of information as well as younger adults if
the source is identied through its emotional rather than nonemotional aspects
(May et al., 2002; Rahhal et al., 2002). In addition, older adults errors and biases resemble those of younger adults under emotional conditions (Kennedy et al.,
2002; Mather & Johnson, 2000, 2003), suggesting that, in general, older adults
are more likely to engage in an emotional focus when encoding or retrieving events.

294

Studies that focus on memory for positive and negative information separately, however, indicate only partial support for the emotional compensation
hypothesis, as aging does not lead to a relative enhancement for all types of
emotional information. Instead, the enhancement in memory for emotional
material among older adults seems to be driven by enhanced memory for positive information. In fact, memory for negative information declines with age
(e.g., Charles et al., 2003; Mather et al., 2001).
This bias to remember positive relatively better than negative events with age
is consistent with the hypothesis that older adults should remember information consistent with their emotional goals. Both in lab studies with controlled
stimuli and in eld studies examining memories for real-life events, older adults
memories are more emotionally gratifying than younger adults memories. In
comparison with memories of younger adults, a larger proportion of what older
adults remember is positive and a smaller proportion is negative (Charles et al.,
2003; Mather et al., 2001; Mather & Carstensen, 2003). Older adults memories for specic events or time periods are also more likely than younger adults
memories to be distorted in an emotionally gratifying direction (Field, 1981;
Levine & Bluck, 1997; Mather & Carstensen, 2003; Mather & Johnson, 2000).
For older adults, negative memories fade faster than positive memories, whereas
this is not the case for younger adults (Berntsen & Rubin, in press; Mather &
Carstensen, 2003).
These gratifying memories seem to be a critical component of older adults
successful emotion regulation processes. For example, older adults experience
more positive emotion than younger adults while reminiscing with other people,
but there is no signicant age dierence in positive emotion in other social contexts (Pasupathi & Carstensen, in press). Thus, memory and emotional functioning seem to interact in mutually benecial ways. Emotional processes can help
older adults remember information they might otherwise have forgottenand,
in turn, emotionally gratifying biases in memory enhance emotional well-being.

Notes
Thanks to Laura Carstensen, Faye Crosby, Paula Hertel, Marcia Johnson, Noah
Mercer, Karen Mitchell, Jennifer Taylor, and Avril Thorne for their comments on
previous versions of this chapter.
1. Contrasting with these ndings, one study found that Alzheimers patients
showed as much of a memory benet for emotional stories as did normal older controls (Kazui et al., 2000). However, the control participants were at ceiling in their
memory accuracy for the emotional aspects of the story, and the patients had relatively mild dementia (their average Mini-Mental State Examination score was 22.6).
2. However, a study testing older and younger adults using a similar set of emotional pictures did not nd an age by valence interaction (Kensinger et al., 2002).
It is not clear why the results of these studies dier.

295

References
Abrisqueta-Gomez, J., Bueno, O. F. A., Oliveira, M. G. M., & Bertolucci, P. H. F.
(2002). Recognition memory for emotional pictures in Alzheimers patients.
Acta Neurologica Scandinavica, 105, 5154.
Adolphs, R., Cahill, L., Schul, R., & Babinsky, R. (1997). Impaired declarative
memory for emotional material following bilateral amygdala damage in humans. Learning and Memory, 4, 291300.
Adolphs, R., Denburg, N. L., & Tranel, D. (2001). The amygdalas role in long-term
declarative memory for gist and detail. Behavioral Neuroscience, 115, 983992.
Agren, M. (1998). Life at 85 and 92: A qualitative longitudinal study of how the
oldest old experience and adjust to the increasing uncertainty of existence. International Journal of Aging and Human Development, 47, 105117.
Baddeley, A. (1998). Human memory: Theory and practice. Rev. ed. Boston: Allyn and
Bacon.
Balota, D. A., Cortese, M. J., Duchek, J. M., Adams, D., Roediger, H. L., McDermott,
K. B., et al. (1999). Veridical and false memories in healthy older adults and in
dementia of the Alzheimers type. Cognitive Neuropsychology, 16, 361384.
Baumeister, R. F., Bratslavsky, E., Fickenauer, C., & Vohs, K. D. (2001). Bad is stronger than good. Review of General Psychology, 5, 323370.
Bechara, A., Damasio, A. R., Damasio, H., & Anderson, S. W. (1994). Insensitivity
to future consequences following damage to human prefrontal cortex. Cognition, 50, 715.
Bechara, A., Damasio, H., & Damasio, A. R. (2000). Emotion, decision making and
the orbitofrontal cortex. Cerebral Cortex, 10, 295307.
Bechara, A., Damasio, H., Tranel, D., & Anderson, S. W. (1998). Dissociation of
working memory from decision making within the human prefrontal cortex.
Journal of Neuroscience, 18, 428437.
Bechara, A., Tranel, D., Damasio, H., & Adolphs, R. (1995). Double dissociation of
conditioning and declarative knowledge relative to the amygdala and hippocampus in humans. Science, 269, 11151118.
Berntsen, D. (2001). Involuntary memories of emotional events: Do memories of
traumas and extremely happy events dier? Applied Cognitive Psychology, 15,
135158.
Berntsen, D. (2002). Tunnel memories for autobiographical events: Central details
are remembered more frequently from shocking than from happy experiences.
Memory & Cognition, 30, 10101020.
Berntsen, D., & Rubin, D. C. (2002). Emotionally charged autobiographical memories across the lifespan: The recall of happy, sad, traumatic, and involuntary
memories. Psychology and Aging, 17, 636652.
Bigler, E. D., Anderson, C. V., & Blatter, D. D. (2002). Temporal lobe morphology in
normal aging and traumatic brain injury. American Journal of Neuroradiology,
23, 255266.
Blanchard-Fields, F. (1998). The role of emotion in social cognition across the adult
life span. In K. W. Schaie & M. P. Lawton (Eds.), Annual review of gerontology
and geriatrics: Focus on emotion and adult development (Vol. 17, pp. 238265).
New York: Springer.

296

Blanchard-Fields, F., Camp, C., & Casper Jahnke, H. (1995). Age dierences in problem-solving style: The role of emotional salience. Psychology and Aging, 10,
173180.
Blanchard-Fields, F., & Camp, C. J. (1990). Aect, individual dierences, and real
world problem solving across the adult life span. In T. Hess (Ed.), Aging and
cognition: Knowledge organization and utilization (pp. 461497). Amsterdam:
North-Holland.
Blaney, P. H. (1986). Aect and memory: A review. Psychological Bulletin, 99, 229
246.
Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129148.
Bradley, B. P., Mogg, K., & Williams, R. (1995). Implicit and explicit memory for
emotion-congruent information in clinical depression and anxiety. Behaviour
Research & Therapy, 33, 755770.
Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 7399.
Burke, A., Heuer, F., & Reisberg, D. (1992). Remembering emotional events.
Memory & Cognition, 20, 277290.
Cabeza, R. (2002). Hemispheric asymmetry reduction in older adults: The HAROLD
model. Psychology and Aging, 17, 85100.
Cacioppo, J. T., Berntsen, G. B., Klein, D. J., & Poehlmann, K. M. (1998). Psychophysiology of emotion across the life span. In K. W. Schaie & M. P. Lawton (Eds.),
Annual review of gerontology and geriatrics: Focus on emotion and adult development (Vol. 17, pp. 2774). New York: Springer.
Cahill, L., Babinsky, R., Markowitsch, H. J., & McGaugh, J. L. (1995). The amygdala
and emotional memory. Nature, 377, 295296.
Cahill, L., Haier, R. J., Fallon, J., Alkire, M. T., Tang, C., Keator, D., et al. (1996).
Amygdala activity at encoding correlated with long-term, free recall of emotional information. Proceedings of the National Academy of Sciences, USA, 93,
80168021.
Callen, D. J. A., Black, S. E., Gao, F., Caldwell, C. B., & Szalai, J. P. (2001). Beyond
the hippocampus: MRI volumetry conrms widespread limbic atrophy in AD.
Neurology, 57, 16691674.
Canli, T., Zhao, Z., Brewer, J. B., Gabrieli, J. D. E., & Cahill, L. (2000). Event-related
activation in the human amygdala associates with later memory for individual
emotional response. Journal of Neuroscience, 20, RC99.
Canli, T., Zhao, Z., Desmond, J. E., Glover, G. H., & Gabrieli, J. D. E. (1999). fMRI
identies a network of structures correlated with retention of positive and negative emotional memory. Psychobiology, 27, 441452.
Carstensen, L. L. (1992). Social and emotional patterns in adulthood: Support for
socioemotional selectivity theory. Psychology and Aging, 7, 331338.
Carstensen, L. L. (1995). Evidence for a life-span theory of socioemotional selectivity. Current Directions in Psychological Science, 4, 151156.
Carstensen, L. L., & Charles, S. T. (1998). Emotion in the second half of life. Current
Directions in Psychological Science, 7, 144149.
Carstensen, L. L., & Fredrickson, B. L. (1998). Inuence of HIV status and age on
cognitive representations of others. Health Psychology, 17, 494503.
Carstensen, L. L., Gross, J. J., & Fung, H. H. (1998). The social context of emotional

297

experience. In K. W. Schaie & M. P. Lawton (Eds.), Annual review of gerontology

and geriatrics: Focus on emotion and adult development (Vol. 17, pp. 325352).
New York: Springer.
Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking time seriously:
A theory of socioemotional selectivity. American Psychologist, 54, 165181.
Carstensen, L. L., Pasupathi, M., Mayr, U., & Nesselroade, J. R. (2000). Emotional
experience in everyday life across the adult life span. Journal of Personality and
Social Psychology, 79, 644655.
Carstensen, L. L., & Turk-Charles, S. (1994). The salience of emotion across the adult
life course. Psychology and Aging, 9, 259264.
Charles, S. T., Mather, M., & Carstensen, L. L. (2003). Aging and emotional memory:
The forgettable nature of negative images for older adults. Journal of Experimental Psychology: General, 132, 310324.
Charles, S. T., Reynolds, C. A., & Gatz, M. (2001). Age-related dierences and change
in positive and negative aect over 23 years. Journal of Personality and Social
Psychology, 80, 136151.
Cherry, B. J., Hellige, J. B., & McDowd, J. M. (1995). Age dierences and similarities
in patterns of cerebral hemispheric asymmetry. Psychology and Aging, 10, 191
203.
Christianson, S.-. (1989). Flashbulb memories: Special, but not so special. Memory
& Cognition, 17, 435443.
Christianson, S.-. (1992). Emotional stress and eyewitness memory: A critical review. Psychological Bulletin, 112, 284309.
Coey, C. E., Wilkenson, W. E., Parashos, I. A., Soady, S. A., Sullivan, R. J., Patterson,
L. J., et al. (1992). Quantitative cerebral anatomy of the aging human brain:
A cross-sectional study using magnetic resonance imaging. Neurology, 42,
527536.
Cohen, G., Conway, M. A., & Maylor, E. A. (1994). Flashbulb memories in older
adults. Psychology and Aging, 9, 454463.
Conway, M. A., Anderson, S. J., Larsen, S. F., Donnelly, C. M., McDaniel, M. A.,
McClelland, A. G., et al. (1994). The formation of ashbulb memories. Memory
& Cognition, 22, 326343.
Cowell, P. E., Turetsky, B. I., Gur, R. C., Grossman, R. I., Shtasel, D. L., & Gur, R. E.
(1994). Sex dierences in aging of the human frontal and temporal lobes. Journal of Neuroscience, 14, 47484755.
Craik, F. I. M., Morris, L. W., Morris, R. G., & Loewen, E. R. (1990). Relations between source amnesia and frontal lobe functioning in older adults. Psychology
and Aging, 5, 148151.
Cuenod, C. A., Denys, A., Michot, J. L., Jehenson, P., Forette, F., Kaplan, D., et al.
(1993). Amygdala atrophy in Alzheimers diseaseAn in-vivo magnetic-resonance-imaging study. Archives of Neurology, 50, 941945.
Damasio, H., Grabowski, T., Frank, R., Galaburda, A. M., & Damasio, A. R. (1994).
The return of Phineas Gage: Clues about the brain from the skull of a famous
patient. Science, 264, 11021105.
Davidson, P. S. R., & Glisky, E. L. (2002). Is ashbulb memory a special instance of
source memory? Evidence from older adults. Memory, 10, 99111.

298

Davidson, R. J., & Irwin, W. (1999). The functional neuroanatomy of emotion and
aective style. Trends in Cognitive Sciences, 3, 1121.
Davidson, R. J., Putnam, K. M., & Larson, C. L. (2000). Dysfunction in the neural
circuitry of emotion regulationA possible prelude to violence. Science, 289,
591594.
DeCarli, C., Murphy, D. G., Gillette, J. A., Haxby, J. V., Teichberg, D., Schapiro,
M. B., et al. (1994). Lack of age-related dierences in temporal lobe volume of
very healthy adults. American Journal of Neuroradiology, 15, 689696.
Diehl, M., Coyle, N., & Labouvie-Vief, G. (1996). Age and sex dierences in strategies
of coping and defense across the life span. Psychology and Aging, 11, 127139.
Diener, E., & Suh, M. E. (1998). Subjective well-being and age: An international
analysis. In B. Tucker & J. Libby (Eds.), Annual review of gerontology and geriatrics (Vol. 17, pp. 304324). New York: Springer.
Dimitrov, M., Phipps, M., Zahn, T. P., & Grafman, J. (1999). A thoroughly modern
Gage. Neurocase, 5, 345354.
Dimond, S. J., Farrington, L., & Johnson, P. (1976). Diering emotional response
from right and left hemispheres. Nature, 261, 690692.
Easterbrook, J. A. (1959). The eect of emotion on cue utilization and the organization of behavior. Psychological Review, 66, 183201.
Ellis, R. J., & Oscar-Berman, M. (1989). Alcoholism, aging, and functional cerebral
asymmetries. Psychological Bulletin, 106, 128147.
Evert, D. L., & Oscar-Berman, M. (2001). Selective attentional processing and the right
hemisphere: Eects of aging and alcoholism. Neuropsychology, 15, 452461.
Fabiani, M., & Friedman, D. (1997). Dissociations between memory for temporal
order and recognition memory in aging. Neuropsychologia, 35, 129141.
Field, D. (1981). Retrospective reports by healthy intelligent elderly people of personal events of their adult lives. International Journal of Behavioral Development,
4, 7797.
Field, D. (1997). Looking back, what period of your life brought you the most satisfaction? International Journal of Aging & Human Development, 45, 169194.
Folkman, S., Lazarus, R. S., Pimley, S., & Novacek, J. (1987). Age dierences in stress
and coping processes. Psychology and Aging, 2, 171184.
Fredrickson, B. L., & Carstensen, L. L. (1990). Choosing social partners: How old
age and anticipated endings make people more selective. Psychology and Aging,
5, 335347.
Fung, H. H., & Carstensen, L. L. (in press). Sending memorable messages to the old:
Age dierences in preferences and memory for emotionally meaningful advertisements. Journal of Personality and Social Psychology.
Fung, H. H., Carstensen, L. L., & Lutz, A. M. (1999). Inuence of time on social preferences: Implications for life-span development. Psychology and Aging, 14, 595
604.
Fung, H. H., Lai, P., & Ng, R. (2001). Age dierences in social preferences among
Taiwanese and mainland Chinese: The role of perceived time. Psychology and
Aging, 16, 351356.
Gardiner, J. M., & Java, R. I. (1993). Recognition memory and awareness: An experiential approach. European Journal of Cognitive Psychology, 5, 337346.

299

George, M. S., Ketter, T. A., Parekh, P. I., Horwitz, B., Herscovitch, P., & Post, R. M.
(1995). Brain activity during transient sadness and happiness in healthy
women. American Journal of Psychiatry, 152, 341351.
Gilboa, E., Roberts, J. E., & Gotlib, I. H. (1997). The eects of induced and naturally
occurring dysphoric mood on biases in self-evaluation and memory. Cognition
& Emotion, 11, 6582.
Glisky, E. L., Polster, M. R., & Routhieaux, B. C. (1995). Double dissociation between
item and source memory. Neuropsychology, 9, 229235.
Goldstein, G., & Shelly, C. (1981). Does the right hemisphere age more rapidly than
the left? Journal of Clinical Neuropsychology, 3, 6578.
Gomez-Isla, T., Price, J. L., McKeel Jr., D. W., Morris, J. C., Growdon, J. H., & Hyman,
B. T. (1996). Profound loss of layer II entorhinal cortex neurons occurs in very
mild Alzheimers disease. Journal of Neuroscience, 16, 44914500.
Good, C. D., Johnsrude, I. S., Ashburner, J., Henson, R. N. A., Friston, K. J., &
Frackowiak, R. S. J. (2001). A voxel-based morphometric study of aging in 465
normal adult human brains. NeuroImage, 14, 2136.
Grafman, J., Vance, S. C., Weingartner, H., Salazar, A. M., & Amin, D. (1986). The
eects of lateralized frontal lesions on mood regulation. Brain, 109, 11271148.
Gross, J. J. (2001). Emotion regulation in adulthood: Timing is everything. Current
Directions in Psychological Science, 10, 214219.
Gross, J. J. (2002). Emotion regulation: Aective, cognitive, and social consequences.
Psychophysiology, 39, 281291.
Gross, J. J., Carstensen, L. L., Pasupathi, M., Tsai, J., Skorpen, C. G., & Hsu, A. Y. C.
(1997). Emotion and aging: Experience, expression, and control. Psychology
and Aging, 12, 590599.
Hamann, S. B. (2001). Cognitive and neural mechanisms of emotional memory.
Trends in Cognitive Sciences, 5, 394400.
Hamann, S. B., Cahill, L., McGaugh, J. L., & Squire, L. R. (1997). Intact enhancement of declarative memory for emotional material in amnesia. Learning &
Memory, 4, 301309.
Hamann, S. B., Ely, T. D., Grafton, S. T., & Kilts, C. D. (1999). Amygdala activity
related to enhanced memory for pleasant and aversive stimuli. Nature Neuroscience, 2, 289293.
Hamann, S. B., Monarch, E. S., & Goldstein, F. C. (2000). Memory enhancement
for emotional stimuli is impaired in early Alzheimers disease. Neuropsychology,
14, 8292.
Hamann, S. B., Monarch, E. S., & Goldstein, F. C. (2002). Impaired fear conditioning in Alzheimers disease. Neuropsychologia, 40, 11871195.
Hashtroudi, S., Johnson, M. K., & Chrosniak, L. D. (1989). Aging and source monitoring. Psychology and Aging, 4, 106112.
Hashtroudi, S., Johnson, M. K., & Chrosniak, L. D. (1990). Aging and qualitative
characteristics of memories for perceived and imagined complex events. Psychology and Aging, 5, 119126.
Hashtroudi, S., Johnson, M. K., Vnek, N., & Ferguson, S. A. (1994). Aging and the
eects of aective and factual focus on source monitoring and recall. Psychology and Aging, 9, 160170.

300

Henkel, L. A., Johnson, M. K., & De Leonardis, D. M. (1998). Aging and source monitoring: Cognitive processes and neuropsychological correlates. Journal of Experimental Psychology: General, 127, 118.
Herzog, A. G., & Kemper, T. L. (1980). Amygdaloid changes in aging and dementia. Archives of Neurology, 37, 625629.
Hess, T. M., & Slaughter, S. J. (1990). Schematic knowledge inuences on memory
for scene information in young and older adults. Developmental Psychology, 26,
855865.
Hutner, N., & Oscar-Berman, M. (1996). Visual laterality patterns for the perception of emotional words in alcoholic and aging individuals. Journal of Studies
on Alcohol, 57, 144154.
Isingrini, M., Fontaine, R., Taconnat, L., & Duportal, A. (1995). Aging and encoding in memory: False alarms and decision criteria in a word-pair recognition
task. International Journal of Aging and Human Development, 41, 7988.
Jack, C. R. Jr, Petersen, R. C., OBrien, P. C., & Tangalos, E. G. (1992). MR-based
hippocampal volumetry in the diagnosis of Alzheimers disease. Neurology, 42,
183188.
Jack, C. R. Jr, Petersen, R. C., Xu, Y. C., Waring, S. C., OBrien, P. C., Tangalos,
E. G., et al. (1997). Medial temporal atrophy on MRI in normal aging and very
mild Alzheimers disease. Neurology, 49, 786794.
Johnson, H. M., & Magaro, P. A. (1987). Eects of mood and severity on memory
processes in depression and mania. Psychological Bulletin, 101, 2840.
Johnson, M. K., & Multhaup, K. S. (1992). Emotion and MEM. In S.-. Christianson
(Ed.), The handbook of emotion and memory: Current research and theory (pp. 33
66). Hillsdale, NJ: Erlbaum.
Johnson, M. K., Nolde, S. F., & De Leonardis, D. M. (1996). Emotional focus and
source monitoring. Journal of Memory and Language, 35, 135156.
Johnson, M. K., & Raye, C. L. (2000). Cognitive and brain mechanisms of false memories and beliefs. In D. L. Schacter & E. Scarry (Eds.), Memory, brain, and belief
(pp. 2586). Cambridge, MA: Harvard University Press.
Johnson, R. C., Cole, R. E., Bowers, J. K., Foiles, S. V., Nikaido, A. M., Patrick, J. W.,
et al. (1979). Hemispheric eciency in middle and later adulthood. Cortex, 15,
109119.
Jones, C. J., & Meredith, W. (2000). Developmental paths of psychological health
from early adolescence to later adulthood. Psychology and Aging, 15, 351
360.
Kazui, H., Mori, E., Hashimoto, M., Hirono, N., Imamura, T., Tanimukai, S., et al.
(2000). Impact of emotion on memory: Controlled study of the inuence of
emotionally charged material on declarative memory in Alzheimers disease.
British Journal of Psychiatry, 177, 343347.
Kemper, T. L. (1994). Neuroanatomical and neuropathological changes during
aging and dementia. In M. L. Albert & E. J. E. Knoepfel (Eds.), Clinical neurology
of aging (2nd ed., pp. 367). New York: Oxford University Press.
Kennedy, Q. (2002). Age dierences in the reconstruction of long-term autobiographical memory: Testing a motivational hypothesis. Unpublished dissertation, Stanford
University, Stanford, CA.

301

Kennedy, Q., Mather, M., & Carstensen, L. L. (in press). The role of motivation in
the age-related positive bias in autobiographical memory. Psychological Science.
Kensinger, E. A., Brierley, B., Medford, N., Growdon, J. H., & Corkin, S. (2002). Eects
of normal aging and Alzheimers disease on emotional memory. Emotion, 2,
118134.
Kensinger, E. A., & Schacter, D. L. (1999). When true memories suppress false
memories: Eects of aging. Cognitive Neuropsychology, 16, 399415.
Kimbrell, T. A., George, M. S., Parekh, P. I., Ketter, T. A., Podell, D. M., Danielson,
A. L., et al. (1999). Regional brain activity during transient self-induced anxiety and anger in healthy adults. Biological Psychiatry, 15, 454465.
Kolb, B., & Whishaw, I. Q. (1990). Fundamentals of human neuropsychology. New
York: W. H. Freeman.
Koutstaal, W., Schacter, D. L., & Brenner, C. (2001). Dual task demands and gistbased false recognition of pictures in younger and older adults. Journal of
Memory and Language, 44, 399426.
Koutstaal, W., Schacter, D. L., Galluccio, L., & Stofer, K. A. (1999). Reducing gistbased false recognition in older adults: Encoding and retrieval manipulations.
Psychology and Aging, 14, 220237.
Labouvie-Vief, G. (1998). Cognitive-emotional integration in adulthood. In K. W.
Schaie & M. P. Lawton (Eds.), Annual review of gerontology and geriatrics: Focus
on emotion and adult development (Vol. 17, pp. 206237). New York: Springer.
Labouvie-Vief, G., DeVoe, M., & Bulka, D. (1989). Speaking about feelings: Conceptions of emotion across the life span. Psychology and Aging, 4, 425437.
Lane, R. D., Reiman, E. M., Ahern, G. L., Schwartz, G. E., & Davidson, R. J. (1997).
Neuroanatomical correlates of happiness, sadness, and disgust. American Journal of Psychiatry, 154, 926933.
Lane, R. D., Reiman, E. M., Bradley, M. M., Lang, P. J., Ahern, G. L., Davidson, R. J.,
et al. (1997). Neuroanatomical correlates of pleasant and unpleasant emotion.
Neuropsychologia, 35, 14371444.
Lawton, M. P., Kleban, M. H., & Dean, J. (1993). Aect and age: Cross-sectional comparisons of structure and prevalence. Psychology and Aging, 8, 165175.
Lawton, M. P., Kleban, M. H., Rajagopal, D., & Dean, J. (1992). Dimensions of aective experience in three age groups. Psychology and Aging, 7, 171184.
Levenson, R. W., Carstensen, L. L., & Gottman, J. M. (1994). The inuence of age
and gender on aect, physiology, and their interrelations: A study of long-term
marriages. Journal of Personality and Social Psychology, 67, 5668.
Levenson, R. W., Friesen, W. V., Ekman, P., & Carstensen, L. L. (1991). Emotion,
physiology, and expression in old age. Psychology and Aging, 6, 2835.
Levine, L. J., & Bluck, S. (1997). Experienced and remembered emotional intensity
in older adults. Psychology and Aging, 12, 514523.
Light, L. L. (1996). Memory and aging. In E. L. Bjork & R. A. Bjork (Eds.), Memory
(pp. 444490). San Diego: Academic Press.
Light, L. L. (2000). Memory changes in adulthood. In S. H. Qualls & N. Abeles (Eds.),
Psychology and the aging revolution: How we adapt to longer life (pp. 7397).
Washington, DC: American Psychological Association.
Lim, K. O., Zipursky, R. B., Murphy, G. M. Jr, & Pfeerbaum, A. (1990). In vivo

302

quantication of the limbic system using MRI: Eects of normal aging. Psychiatry Research, 35, 1526.
Livingston, R. B. (1967). Reinforcement. In G. C. Quarton, T. Melnechuck, & F. O.
Schmitt (Eds.), The neurosciences: A study program (pp. 568576). New York:
Rockefeller University Press.
Masterman, D. L., & Cummings, J. L. (1997). Frontal-subcortical circuits: The anatomic basis of executive, social and motivated behaviors. Journal of Psychopharmacology, 11, 107114.
Mather, M., Canli, T., English, T., Golarai, G., Fleischman, D., Gabrieli, J. D. E., et al.
(2001, November). An fMRI study of age-related changes in encoding and remembering emotional material. Paper presented at the annual conference of the Society for Neuroscience, San Diego, CA.
Mather, M., & Carstensen, L. L. (2003). Implications of the positivity effect for older
peoples memories about health care decisions. Manuscript submitted for
publication.
Mather, M., & Carstensen, L. L. (in press). Aging and attentional biases for emotional
faces. Psychological Science.
Mather, M., & Johnson, M. K. (2000). Choice-supportive source monitoring: Do our
decisions seem better to us as we age? Psychology and Aging, 15, 596606.
Mather, M., & Johnson, M. K. (2003). Aective review and schema reliance in older
and younger adults. American Journal of Psychology, 116, 169189.
Mather, M., Johnson, M. K., & De Leonardis, D. M. (1999). Stereotype reliance in
source monitoring: Age dierences and neuropsychological test correlates.
Cognitive Neuropsychology, 16, 437458.
May, C. P., Rahhal, T., Leighton, E., & Berry, E. (2002, April). Aging, source memory,
and emotional cues. Paper presented at the Cognitive Aging Conference,
Atlanta, GA.
McConatha, J. T., & Huba, H. M. (1999). Primary, secondary, and emotional control across adulthood. Current Psychology, 18, 164170.
McDaniel, M. A., Glisky, E. L., Rubin, S. R., Guynn, M. J., & Routhieaux, B. C. (1999).
Prospective memory: A neuropsychological study. Neuropsychology, 13, 103
110.
McGaugh, J. L. (2000). MemoryA century of consolidation. Science, 287, 248251.
McGaugh, J. L., Ferry, B., & Vazdarjanova, A. (2000). Amygdala: Role in modulation of memory storage. In J. P. Aggleton (Ed.), The amygdala (pp. 391423).
New York: Oxford University Press.
McIntyre, J. S., & Craik, F. I. M. (1987). Age dierences in memory for item and
source information. Canadian Journal of Psychology, 41, 175192.
Mineka, S., & Nugent, K. (1995). Mood-congruent memory biases in anxiety and
depression. In D. L. Schacter (Ed.), Memory distortions: How minds, brains, and
societies reconstruct the past (pp. 173193). Cambridge, MA: Harvard University Press.
Mitchell, K. J., Johnson, M. K., & Mather, M. (2003). Source monitoring and suggestibility to misinformation: Adult age-related dierences. Applied Cognitive
Psychology, 17, 107119.
Mittenberg, W., Seidenberg, M., Oleary, D. S., & Digiulio, D. V. (1989). Changes in

303

cerebral functioning associated with normal aging. Journal of Clinical and Experimental Neuropsychology, 11, 918932.
Moreno, C. R., Borod, J. C., Welkowitz, J., & Alpert, M. (1990). Lateralization for the
expression and perception of facial emotion as a function of age. Neuropsychologia, 28, 199209.
Mori, E., Ikeda, M., Hirono, N., Kitagaki, H., Imamura, T., & Shimomura, T. (1999).
Amygdalar volume and emotional memory in Alzheimers disease. American
Journal of Psychiatry, 156, 216222.
Moscovitch, M., & Olds, J. (1982). Asymmetries in spontaneous facial expressions
and their possible relation to hemispheric specialization. Neuropsychologia, 20,
7182.
Moscovitch, M., & Winocur, G. (1995). Frontal lobes, memory, and aging. Annals
of the New York Academy of Sciences, 769, 119150.
Mroczek, D. K. (2001). Age and emotion in adulthood. Current Directions in Psychological Science, 10, 8790.
Mu, Q., Xie, J., Wen, Z., Weng, Y., & Shuyun, Z. (1999). A quantitative MR study of
the hippocampal formation, the amygdala, and the temporal horn of the lateral ventricle in healthy subjects 40 to 90 years of age. American Journal of
Neuroradiology, 20, 207211.
Nebes, R. D. (1990). Hemispheric specialization in the aged brain. In C. B. Trevarthen
(Ed.), Brain circuits and functions of the mind: Essays in honor of Roger W. Sperry
(pp. 364370). New York: Cambridge University Press.
Nebes, R. D., Madden, D. J., & Berg, W. D. (1983). The eect of age on hemispheric
asymmetry in visual and auditory identication. Experimental Aging Research,
9, 8791.
Neisser, U., & Harsch, N. (1992). Phantom ashbulbs: False recollections of hearing the news about Challenger. In E. Winograd & U. Neisser (Eds.), Aect and
accuracy in recall: Studies of ashbulb memories (pp. 931). Cambridge: Cambridge University Press.
Norman, K. A., & Schacter, D. L. (1997). False recognition in younger and older
adults: Exploring the characteristics of illusory memories. Memory & Cognition,
25, 838848.
Nyberg, L., Cabeza, R., & Tulving, E. (1996). PET studies of encoding and retrieval:
The HERA model. Psychonomic Bulletin & Review, 3, 135148.
Obler, L. K., Woodward, S., & Albert, M. L. (1984). Changes in cerebral lateralization in aging. Neuropsychologia, 22, 235240.
Ohnishi, T., Matsuda, H., Tabira, T., Asada, T., & Uno, M. (2001). Changes in brain
morphology in Alzheimer disease and normal aging: Is Alzheimer disease an
exaggerated aging process? American Journal of Neuroradiology, 22, 16801685.
Oxman, T. E., Barrett, J. E., Sengupta, A., & Williams, J. W. (2000). The relationship of aging and dysthymia in primary care. American Journal of Geriatric Psychiatry, 8, 318326.
Pasupathi, M., & Carstensen, L. L. (in press). Age and emotional experience during
mutual reminiscing. Psychology and Aging.
Phillips, L. H., & Della Sala, S. (1998). Aging, intelligence, and anatomical segregation in the frontal lobes. Learning and Individual Dierences, 10, 217243.

304

Pillemer, D. B. (1984). Flashbulb memories of the assassination attempt on President Reagan. Cognition, 16, 6380.
Prull, M. W., Gabrieli, J. D. E., & Bunge, S. A. (2000). Age-related changes in memory:
A cognitive neuroscience perspective. In F. I. M. Craik & T. A. Salthouse (Eds.),
The handbook of aging and cognition (2nd ed., pp. 91153). Mahwah, NJ: Erlbaum.
Rahhal, T., May, C. P., & Hasher, L. (2002). Truth and character: Sources that older
adults can remember. Psychological Science, 13, 101105.
Raine, A., Buschbaum, M., & LaCasse, L. (1997). Brain abnormalities in murderers
indicated by positron emission tomography. Biological Psychiatry, 42, 495
508.
Rankin, J. L., & Kausler, D. H. (1979). Adult age dierences in false recognitions.
Journal of Gerontology, 34, 5865.
Raz, N. (2000). Aging of the brain and its impact on cognitive performance: Integration of structural and functional ndings. In F. I. M. Craik & T. A. Salthouse (Eds.),
Handbook of aging and cognition (2nd ed., pp. 190). Mahwah, NJ: Erlbaum.
Raz, N., Gunning, F. M., Head, D., Dupuis, J. H., McQuain, J., Briggs, S. D., et al.
(1997). Selective aging of the human cerebral cortex observed in vivo: Dierential vulnerability of the prefrontal gray matter. Cerebral Cortex, 7, 268282.
Regier, D. A., Boyd, J. H., Burke, J. D., Rae, D. S., Myers, J. K., Kramer, M., et al.
(1988). One-month prevalence of mental disorders in the United States: Based
on 5 epidemiologic catchment-area sites. Archives of General Psychiatry, 45,
977986.
Reisberg, D., & Heuer, F. (1992). Remembering the details of emotional events. In
E. Winograd & U. Neisser (Eds.), Aect and accuracy in recall: Studies of ashbulb memories (pp. 163190). Cambridge, MA: Harvard University Press.
Reisberg, D., Heuer, F., McLean, J., & Oshaughnessy, M. (1988). The quantity, not
the quality, of aect predicts memory vividness. Bulletin of the Psychonomic
Society, 26, 100103.
Reminger, S. L., Kaszniak, A. W., & Dalby, P. R. (2000). Age-invariance in the asymmetry of stimulus-evoked emotional facial muscle activity. Aging, Neuropsychology, & Cognition, 7, 156168.
Rolls, E. T. (2000). The orbitofrontal cortex and reward. Cerebral Cortex, 10, 284
294.
Rozin, P., & Royzman, E. B. (2001). Negativity bias, negativity dominance, and contagion. Personality and Social Psychology Review, 5, 296320.
Rubin, D. C., & Kozin, M. (1984). Vivid memories. Cognition, 16, 115.
Rubin, D. C., Rahhal, T., & Poon, L. W. (1998). Things learned in early adulthood
are remembered best. Memory & Cognition, 26, 319.
Salat, D. H., Kaye, J. A., & Janowsky, J. S. (2001). Selective preservation and degeneration within the prefrontal cortex in aging and Alzheimers Disease. Archives
of Neurology, 58, 14031408.
Schacter, D. L., Kazniak, A. W., Kihlstrom, J. F., & Valdiserri, M. (1991). The relation between source memory and aging. Psychology and Aging, 6, 559568.
Schaie, K. W., & Schaie, J. P. (1977). Clinical assessment and aging. In J. E. Birren
& K. W. Schaie (Eds.), Handbook of the psychology of aging (pp. 692723). New
York: Van Nostrand Reinhold.

305

Schmolck, H., Bualo, E. A., & Squire, L. R. (2000). Memory distortions develop over
time: Recollections of the O. J. Simpson trial verdict after 15 and 32 months.
Psychological Science, 11, 3945.
Schwartz, G. E., Davidson, R. J., & Maer, F. (1975). Right hemisphere lateralization
for emotion in the human brain. Science, 190, 186288.
Scott, D., & Ponsoda, V. (1996). The role of positive and negative aect in ashbulb
memory. Psychological Reports, 79, 467473.
Scott, S. A., DeKosky, S. T., & Sche, S. W. (1991). Volumetric atrophy of the
amygdala in Alzheimers disease: Quantitative serial reconstruction. Neurology, 41, 351356.
Scoville, W. B., & Milner, B. (1957). Loss of recent memory after bilateral hippocampal lesions. Journal of Neurology, Neurosurgery and Psychiatry, 20, 1121.
Shimamura, A. P. (1995). Memory and frontal lobe function. In M. S. Gazzaniga
(Ed.), The cognitive neurosciences (pp. 803813). Cambridge, MA: MIT Press.
Smith, A. D. (1975). Partial learning and recognition memory in the aged. International Journal of Aging and Human Development, 6, 359365.
Smith, C. D., Malcein, M., Meurer, K., Schmitt, F. A., Markesbery, W. R., & Pettigrew,
L. C. (1999). MRI temporal lobe volume measures and neuropsychologic function in Alzheimers disease. Journal of Neuroimaging, 9, 29.
Smith, J., Fleeson, W., Geiselmann, B., Settersten, R. A., & Kunzmann, U. (1999).
Sources of well-being in very old age. In P. B. Baltes & K. U. Mayer (Eds.), The
Berlin Aging Study: Aging from 70 to 100 (pp. 450471). New York: Cambridge
University Press.
Spencer, W. D., & Raz, N. (1995). Dierential eects of aging on memory for content and context: A meta-analysis. Psychology and Aging, 10, 527539.
Suengas, A. G., & Johnson, M. K. (1988). Qualitative eects of rehearsal on memories for perceived and imagined complex events. Journal of Experimental Psychology: General, 117, 377389.
Sullivan, E. V., Marsh, L., Mathalon, D. H., Lim, K. O., & Pfeerbaum, A. (1995).
Age-related decline in MRI volumes of temporal lobe gray matter but not hippocampus. Neurobiology of Aging, 16, 591606.
Tisserand, D. J., Visser, P. J., van Boxtel, M. P. J., & Jolles, J. (2000). The relation between global and limbic brain volumes on MRI and cognitive performance in
healthy individuals across the age range. Neurobiology of Aging, 21, 569576.
Tsai, J. L., Levenson, R. W., & Carstensen, L. L. (2000). Autonomic, subjective, and
expressive responses to emotional lms in older and younger Chinese Americans and European Americans. Psychology and Aging, 15, 684693.
Tun, P. A., Wingeld, A., Rosen, M. J., & Blanchard, L. (1998). Response latencies
for false memories: Gist-based processes in normal aging. Psychology and Aging,
13, 230241.
Watkins, P. C., Mathews, A., Williamson, D. A., & Fuller, R. D. (1992). Mood-congruent memory in depression: Emotional priming or elaboration? Journal of
Abnormal Psychology, 101, 581586.
Weissman, M. M., Leaf, P. J., Bruce, M. L., & Florio, L. (1988). The epidemiology of
dysthymia in ve communities: Rates, risks, co-morbidity, and treatment.
American Journal of Psychiatry, 145, 815819.

306

West, M. J., Coleman, P. D., Flood, D. G., & Troncoso, J. C. (1994). Dierences in the
pattern of hippocampal neuronal loss in normal ageing and Alzheimers disease. Lancet, 17, 769772.
West, R. L. (1996). An application of prefrontal cortex function theory to cognitive
aging. Psychological Bulletin, 120, 272292.
Wright, D. B., Gaskell, G. D., & OMuircheartaigh, C. A. (1998). Flashbulb memory
assumptions: Using national surveys to explore cognitive phenomena. British
Journal of Psychology, 89, 103121.
Yarmey, A. D., & Bull, M. P. (1978). Where were you when President Kennedy was
assassinated? Bulletin of the Psychonomic Society, 11, 133135.
Young, M. P., Scannell, J. W., Burns, G. A., & Blakemore, C. (1994). Analysis of
connectivity: Neural systems in the cerebral cortex. Reviews in the Neurosciences,
5, 227250.
Zacks, R. T., Hasher, L., & Li, K. Z. H. (2000). Human memory. In F. I. M. Craik &
T. A. Salthouse (Eds.), The handbook of aging and cognition (pp. 293357). Mahwah,
NJ: Erlbaum.

307

10

. , ,
. , .

ate one afternoon in Abilene, Texas, Margarita Chavez was piling

her three children, ages 1 month, 2 years, and 6 years, into the car
after a shopping trip. As Margarita went to return her shopping cart, an unfamiliar woman rushed to the car, removed the infant, and put her into her own
car. When Margarita turned around, just a few seconds later, the woman had
started to drive away. The frantic mother tried to stop the car and was dragged
nearly 40 feet through the parking lot, sustaining minor injuries. Meanwhile, a
bystander, 13-year-old Robert Gann, also tried to stop the car by slamming his
st into the windshield. Despite their eorts, the woman drove away, abducting
the infant (Baby snatched, 2002). The witnesses to this crime, Margarita, the
two children, and Robert then faced the problem of explaining to the police
what had just happened and describing the kidnapper and her car. Could
they provide an accurate description of this highly emotional experience? How
would their emotional state at the time of the event inuence their subsequent
memory? How might the childrens eyewitness memory dier from that of
Margarita, an adult? Questions such as these are central to the forensic evaluation of eyewitness accounts. The answers to these questions, moreover, often
have critical implications for all involved, including victims, defendants,
and legal professionals. Fortunately, psychological science can provide highly
pertinent insight into these issues.
In this chapter, we address these questions and others in light of current
theory and research on emotion and memory. We focus primarily on memory
for negative emotional events because of the centrality of negative emotions to
the forensic context. We begin by reviewing research on the inuence of emotion on eyewitness memory and the implications of this research for the legal
system. Then, we discuss research on memory for highly emotional, or traumatic, events, including studies of repressed and recovered memories. Because
the experience of trauma is often associated with clinical symptomatology, such

308

308

as posttraumatic stress disorder and dissociation, we also consider the inuence

of psychopathology on eyewitness memory capabilities. Finally, we address developmental considerations in the study of emotion and eyewitness memory. We
conclude with a discussion of potential avenues for future research.

Emotion and Eyewitness Memory

Individuals who become involved in the legal system are often asked to recall
highly emotional and negative information. Witnesses to a murder or victims
of an assault, for instance, may be asked to recount the details of that crime to
police ocers, attorneys, or jurors. Moreover, both victims and witnesses may
be under considerable emotional distress when encoding these details. How is
negative emotional information remembered, and what is the inuence of an
individuals emotional state (i.e., distress) on his or her memory? In this section,
we consider the inuence of negative emotion on eyewitness memory. Topics
include the accuracy of emotional memories, memory for central versus peripheral details of an emotional event, the durability of emotional memories over long
delays, and the potentially biasing eect of emotion on memory.

Accuracy of Negative

Emotional Memories

Does negative emotion enhance or impair memory? Findings across many studies, using a variety of methodologies, suggest that the answer to this question
partially depends on how emotion is assessed (e.g., emotional stimuli vs. emotional distress of participants, self-report vs. physiological measures) and the kind
of information being recalled (e.g., central vs. peripheral details of an event).
In one line of research, participants are presented with a series of objectively
rated stimuli (e.g., words, pictures, stories) ranging in emotionality, and their
memory for these stimuli is assessed. Researchers can then determine how emotional valence (i.e., positive vs. negative vs. neutral) inuences memory. The
majority of research on memory for emotional stimuli suggests that adults remember negatively valenced stimuli better than emotionally neutral stimuli and
sometimes better than positively valenced stimuli (e.g., Anooshian & Hertel,
1994; Bradley, Greenwald, Petry, & Lang, 1992; Canli, Desmond, Zhao, &
Gabrieli, 2002; Doerksen & Shimamura, 2001; Heuer & Reisberg, 1990; LaBar
& Phelps, 1998; Nagae & Moscovitch, 2002; Ochsner, 2000). Thus, individuals who witness (or experience) a negative event, such as an assault, may be particularly accurate when recalling that event.
In one study, Ochsner (2000) examined recognition memory for pictures of
negative, positive, and neutral emotional valence. Results across three experiments consistently indicated that negative pictures were remembered more accurately than either neutral or positive pictures. Moreover, correct recognition

309

of negative pictures was more likely to be associated with a conscious recollective

experience (i.e., a remember judgment; Tulving, 1985) than were neutral or
positive pictures. These ndings suggest that negative emotion may enhance
memory by increasing the distinctiveness of emotional stimuli and/or the vividness with which they are recollected.
In a second line of research, participants memory for an emotional event
is assessed in relation to their emotional state (i.e., distress) during that event.
In these studies, distress is measured in dierent ways, including self-report
(e.g., Christianson & Hubinette, 1993), observer report (e.g., Alexander, Goodman, Schaaf, et al., 2002), and physiological indices (e.g., Kirschbaum, Wolf,
May, Wippich, & Hellhammer, 1996). Because victims and witnesses to a crime
likely experience distress during the event, it is important to understand
how distress inuences eyewitness memory. For instance, if distress enhances
(or impairs) memory for an event, the emotional state of a victim or witness
is one factor that could be used to evaluate the credibility of his or her report.
Perhaps because of diering methodologies, research ndings are sometimes
inconsistent regarding the relation between experienced (or observed) distress and
memory. A number of studies suggest that memory is enhanced both by selfreported distress (Yuille & Cutshall, 1986) and distress rated (or inferred) by
observers (e.g., Alexander, Goodman, Schaaf, et al., 2002; Goodman, Hirschman,
Hepps, & Rudy, 1991; Quas et al., 2001; Yuille, Davies, Gibling, Marxsen, & Porter, 1994), thus complementing ndings from studies of memory for emotional
stimuli. Other studies, however, suggest that both self-reported (e.g., Christianson
& Hubinette, 1993) and observer-reported (e.g., Quas et al., 1999) distress may
be unrelated to memory. Observer-rated distress has also been negatively related
to memory (Vandermaas, Hess, & Baker-Ward, 1993).
Studies using physiological measures of stress have yielded similarly mixed
ndings: Kirschbaum et al. (1996) administered cortisol, which causes physiological changes indicative of stress, to half of a sample of 40 male adults. Participants then studied a list of (not necessarily emotional) nouns. Compared to
those given a placebo, individuals given cortisol showed decrements on a subsequent memory test. In other studies, however, adrenal stress hormones have
facilitated emotional memory (e.g., Cahill, Prins, Weber, & McGaugh, 1994; van
Stegeren, Everaerd, Cahill, McGaugh, & Gooren, 1998). Finally, additional studies suggest that, consistent with the Yerkes-Dodson law (Yerkes & Dodson, 1908),
only moderate levels of stress hormones may enhance memory, whereas very high
or low levels may impair memory (e.g., chapters 2 and 3, this volume; Lupien &
McEwen, 1997; Quirarte, Roozendaal, & McGaugh, 1997).
Thus, although most studies suggest that memory is enhanced for aectively
negative compared to neutral (and sometimes positive) stimuli, research is less
consistent regarding the inuence of an individuals emotional state on his or
her memory for a stressful event. In fact, in a recent survey of experts in psychology and law, only 50% indicated that they would testify in court about the eects

310

of stress on memory, yet over 75% would testify about other topics, such as the
eects of line-up instructions and the inuence of postevent information (Kassin,
Tubb, Hosch, & Memon, 2001). Thus, when evaluating eyewitness reports of
stressful events, legal factnders may not have a consensus of research to inform
their judgments. Given the forensic importance of the inuence of emotion on
memory, further research is crucial to reconcile inconsistencies in ndings and
methodologies from previous studies.

Memory for Central Versus

Peripheral Details

In an attempt to account for previous inconsistent ndings in the emotion and

memory literature, Christianson (1992) proposed that emotion results in a narrowing of attention (Easterbrook, 1959), which enhances memory for central
details of an event while impairing memory for peripheral details. A number of
studies support this hypothesis (e.g., Berntsen, 2002; Bornstein, Liebel, &
Scarberry, 1998; Burke, Heuer, & Reisberg, 1992; Christianson & Loftus, 1991;
Safer, Christianson, Autry, & Osterlund, 1998; Wessel & Merckelbach, 1994,
1997, 1998; but see Wessel, van der Kooy, & Merckelbach, 2000), including
those of real-life crime victims (e.g., Christianson & Hubinette, 1993), suggesting that the inuence of stress on memory may depend in part on the type of information being recalled. Christianson and Hubinette, for instance, interviewed
witnesses to a bank robbery and compared their accounts with information obtained from police reports. Some individuals had been bystanders during the
robbery, whereas the robbers had threatened others (i.e., victims) with guns.
Although accuracy was relatively high for both groups, the victims, who presumably experienced greater stress during the robbery, recounted the central
details of the event more accurately than the bystanders. Participants were also
asked to report their level of stress experienced during the crime; however, their
ratings were unrelated to memory for the event, perhaps because both bystanders and victims reported relatively high levels of stress.
The trade-o between central and peripheral details may also explain the
phenomenon of weapon focus, in which victims of a crime involving a weapon
tend to focus on that weapon at the expense of other details of the crime, such as
the perpetrators face (e.g., Loftus, Loftus, & Messo, 1987; Pickel, 1999; Shaw &
Skolnick, 1999; Stanny & Johnson, 2000). As a result, a victims ability to correctly identify a briey glimpsed stranger in a line-up may be compromised when
a weapon is involved (see Steblay, 1992, for a meta-analytic review), at least in
simulated crime situations. However, the extrapolation of weapon focus to reallife crime situations may be more complicated.
A meta-analysis of studies on weapon focus suggests that, although the eect
is relatively small, it is more pronounced when stress is high and in contexts that
more closely approximate real-life situations (Steblay, 1992). Although this

311

nding suggests that weapon focus may be more prevalent among actual crime
victims than is indicated by laboratory studies, more recent studies of real-life
crimes suggest an opposite pattern of results, in which victims of crimes involving
a weapon recall more information about the perpetrator than do victims of weaponless crimes (Cooper, Kennedy, Herv, & Yuille, 2002; Tollestrup, Turtle, & Yuille,
1994). Cooper et al. (2002), for instance, examined memories of sexual assault
among female prostitutes; those who were assaulted with a weapon provided more
detail about the crime than those assaulted without a weapon. Although these
ndings seemingly contradict the weapon-focus hypothesis, because the accuracy
of the information provided by the real-life victims could not be veried, the possibility remains that the presence of a weapon increases only the amount (as opposed
to the accuracy) of the information provided. Thus, future research should examine the applicability of weapon-focus research to real-life crime situations, particularly given the consensus in the scientic community that a weapon impairs a
victims ability to identify the perpetrator (Kassin et al., 2001): Kassin et al. reported
that 87% of experts agreed that weapon focus was a reliable nding, and 77% indicated that they would testify to that eect in court.
It is also important to note that, according to the weapon-focus hypothesis,
insofar as victims focus on the weapon at the expense of the perpetrators face,
their memory for that weapon should be enhanced (e.g., Kramer, Buckhout, &
Eugenio, 1990). That is, compromised memory for the perpetrator should coincide with improved memory for the weapon. Most studies of weapon focus, however, have not examined memory for the weapon (but see, e.g., Shaw & Skolnick,
1999), making it somewhat dicult to interpret previous ndings of impaired
memory for the perpetrator. Although individuals purportedly orient toward a
weapon, stress induced by the weapon may compromise memory in general, for
both the perpetrator and the weapon.
Nevertheless, the idea that emotion enhances memory for central details at
the expense of memory for peripheral details has received considerable empirical support. Further research needs to be conducted, however, on real-life crime
victims to determine whether these ndings extend beyond the laboratory. Clarifying the mechanism underlying weapon focus is important not only to address
theoretical issues surrounding distress and memory but also to evaluate eyewitness accounts of actual crimes involving weapons.

Long-Term Durability

of Emotional Memories

In many instances, victims or witnesses of a crime may not be asked to recount

their experience for several days, weeks, or even years after the crime. The durability of emotional memories over time is thus an important concern in forensic settings. Several studies of memory for emotional experiences, after delays
ranging from a few months to more than 10 years, suggest that such memories

312

can be highly accurate (e.g., Burgwyn-Bailes, Baker-Ward, Gordon, & Ornstein,

2001; Christianson & Hubinette, 1993; Pillemer, 1984; Quas et al., 1999, 2001;
Shrimpton, Oates, & Hayes, 1998; Yuille et al., 1994). Moreover, distress experienced during the event may continue to facilitate memory even after long delays. One study examined young adults memories of abuse-related legal
involvement after a delay of approximately 13 years (Quas et al., 2001). As children, the young adults had been observed while waiting to testify at the courthouse, and their visible level of distress had been rated. In addition, details about
each childs legal experience (e.g., whether he or she actually testied, who else
was in the courtroom) had been documented and thus could be compared to his
or her current report. The accuracy of the young adults reports of their experience in court, including details about testifying (for those who took the stand),
was positively related to their level of distress rated at the time of legal involvement. That is, individuals who were more distressed during the initial experience provided more accurate details about it after a delay, suggesting that
emotion may facilitate memory retention over time.
In another study, police trainees participated in a simulated crime event that
was either stressful or nonstressful (Yuille et al., 1994). When their memory for
the event was tested after a delay of either 1 or 12 weeks, trainees in the stressful
condition provided more accurate information than those in the nonstressful condition. Moreover, memory for the stressful event was more resistant to decay over
time. These ndings reinforce the notion that the enhancing eect of emotion on
memory increases over time. Neurological research suggests that this eect may
be due to physiological processes involving the amygdala that support the consolidation of emotional memories (chapter 2, this volume; Cahill & McGaugh, 1998;
Hamann, 2001). Another possibility is that individuals who are stressed during an
event are more likely to discuss their experience with others, a process that often
facilitates memory (e.g., Fivush, Haden, & Reese, 1996; Quas et al., 1999).
Thus, emotion appears to enhance memory retention, as suggested by evidence that individuals who are more stressed during an event are less likely to
forget that event over time. Not only can emotional events be remembered after
a relatively long delay but they may be remembered more accurately than less
emotional events. However, there is less consensus about the long-term durability of memory for events that are highly emotional, or traumatic.

Emotional Memory Biases

Although many studies indicate that emotion enhances memory, others suggest
the opposite, and this discrepancy may reect dierences across studies in the types
of information for which memory is assessed (i.e., central vs. peripheral details;
Christianson, 1992). Discrepancies may also result from dierences in the type of
eyewitness response considered (i.e., correct vs. incorrect). Some studies have revealed that emotional stimuli may be associated with increased rates of incorrect

313

responses, or false alarms, as well as increased rates of correct responses, when

recognition memory is assessed (e.g., Windmann & Krger, 1998; Windmann &
Kutas, 2001). Thus, in response to negative emotional information, individuals
may relax the criterion they use to decide whether information is old or new.
Recent event-related brain potential (ERP) data suggest that this phenomenon,
the emotion-induced recognition bias (Windmann & Kutas, 2001), occurs automatically, rather than through a strategic, deliberate mnemonic process. Nevertheless,
it may be possible to reduce this bias through practice or training.
Limited evidence suggests that the experience of stress prior to encoding may
also increase false-alarm rates for neutral information (Payne, Nadel, Allen,
Thomas, & Jacobs, 2002). This idea may be particularly relevant to real-life crime
situations, in which a victim or witness experiences stress during an event and
then may be asked not only about details regarding the crime itself but also about
details following the crime. Payne et al. induced stress by instructing participants
to give an extemporaneous speech, about which they believed they would later
be evaluated. After the stress induction, participants studied several lists of words.
When their memory for those words was then tested, participants who had experienced the stress induction were more likely than those who had not to falsely
recognize words semantically related to those they had studied. Correct recognition, however, was unaected by stress. (False recognition of unrelated words
occurred infrequently and was unrelated to stress condition.) According to the
authors, increased rates of false alarms may result from the impact of stress on
brain regions involved in long-term memory, such as the hippocampus. As we
discuss later, some evidence suggests that chronic exposure to stress, as evidenced
by PTSD symptomatology, may be similarly associated with increased false
alarms. [See chapter 3Eds.]
Heightened false alarm rates may also result from the inuence of emotion
on source-monitoring processes (Johnson, Nolde, & Leonardis, 1996). According to Johnson et al., focusing on ones own emotions while encoding an event
(which may occur during real-life stressful situations; Gross, 2002) can impair
an individuals ability to discriminate the source of a particular memory. Sourcemonitoring diculties, in turn, are associated with more false alarms (Johnson,
Hashtroudi, & Lindsay, 1993).
A possible emotion-induced recognition bias may have important implications for the forensic context. For instance, in forensic interviews, individuals
may be asked yes/no questions (e.g., Did he hit you?) about a stressful experience. If negative experiences bias respondents to respond armatively to such
questions, open-ended questions (e.g., What did he do?), which are less likely
to result in false alarms, may be more useful. For similar reasons, an emotioninduced bias may inuence the outcome of a line-up if, for example, victims are
biased to falsely recognize potential perpetrators. This possibility highlights the
importance of constructing scientically sound line-ups that limit false identications (see Wells et al., 1998). In addition, the potentially stress-inducing

314

nature of forensic situations themselves may be relevant. That is, eyewitnesses

may be stressed during a line-up or while testifying not only because of their past
experiences but also because of the stress associated with recounting those experiences to others.

Summary and Conclusions

In summary, emotion inuences eyewitness memory in a variety of ways.
Whereas research on memory for emotional stimuli generally suggests that
memory for central information is enhanced by emotion, particularly negative
emotion, research on emotional stress and memory has yielded mixed ndings.
Other research, however, suggests that these inconsistent ndings may be due,
in part, to dierences across studies in the type of information recalled (i.e., central vs. peripheral details) or to the type of eyewitness responses considered (i.e.,
correct or false alarms). In addition, the few long-term studies of distress and
memory suggest that emotion may enhance memory consolidation over time
and thus preserve long-term memory. Many of these ndings have been used to
inform forensic practice. To improve the forensic applicability of this research,
however, future research should (1) clarify inconsistencies in the stress and
memory literature, including the applicability of weapon-focus research to reallife crime situations; (2) further investigate the long-term durability of emotional
memories; and (3) examine the prevalence of emotion-induced recognition
biases outside the laboratory.

Memory for Traumatic Events

The majority of research we have discussed concerns memory for events or
stimuli that range in emotionality. However, whether these ndings apply to
memory for events that are highly emotional, or traumatic, is currently a topic
of considerable debate (e.g., American Psychological Association [APA] Working Group, 1998; Williams & Banyard, 1999). Some researchers have hypothesized that memory for trauma is governed by special mechanisms (e.g.,
dissociation, repression; Freyd, 1996; Terr, 1994) that are not involved in the
processing of nontraumatic information and that may cause traumatic memories to be inaccessible for extended periods. In the following sections, we consider
the argument that special mechanisms underlie memory for traumatic events,
as well as research on repressed and recovered memories.

Trauma and Memory: A Case

for Special Mechanisms?

At the upper end of the stress continuum are highly emotional, often traumatic,
events (Metcalfe & Jacobs, 2000). A traumatic experience is threatening to the

315

life or well-being of an individual (Brewin, Dalgleish, & Joseph, 1996), engenders feelings of helplessness or fear (Foa, Zinbarg, & Rothbaum, 1992), and results in altered levels of stress hormones, such as cortisol, epinephrine, and
norepinephrine (American Psychiatric Association, 1994; Cooper, 1986). Examples of traumatic events include (but are not limited to) physical or sexual
assault, automobile accidents, witnessing a violent crime, and military combat.
Although some would predict enhanced memory for at least the central details
of even very traumatic events, due to their highly stressful nature (e.g.,
Christianson, 1992; Goodman et al., 1991; Metcalfe & Jacobs, 2000; Shobe &
Kilhstrom, 1997), others have argued that special mechanisms, such as repression, may be invoked when processing traumatic experiences (e.g., S. Freud,
1915; Freyd, 1996; Terr, 1994). Unfortunately, because it is dicult to study
memory for traumatic events in the laboratory, relatively little empirical evidence
exists to evaluate either of these claims; however, a few studies have assessed
participants memories for real-life traumatic events, such as child sexual abuse
(CSA; e.g., Goodman et al., 2003; Williams, 1994). [Also see chapter 4Eds.]
According to the trauma-specic mechanisms view, events that are particularly traumatic (e.g., severe assault, abuse perpetrated by a close family member) are less likely to be recalled due to psychological defense mechanisms that
inhibit conscious awareness of the experience (e.g., A. Freud, 1936; Freyd,
1996). Some support for this prediction comes from a study of womens disclosure of CSA (Williams, 1994). Williams interviewed 129 women who had been
seen at a hospital in the early 1970s because of alleged sexual abuse. When
questioned approximately 17 years later about a variety of personal topics, including CSA, 38% of the women failed to disclose the documented abuse. Individuals who did not disclose were more closely related to the perpetrator of the
abuse than were individuals who disclosed. Although nondisclosure could reect
psychological processes other than repression, such as reluctance to discuss sensitive topics, these ndings are consistent with the idea that the more severe the
abuse, the less likely it is to be remembered. Furthermore, based on the rate of
nondisclosure in her sample, Williams concluded that forgetting CSA is relatively
common (see also Widom & Morris, 1997).
In a more recent study, however, Goodman et al. (2003) examined CSA disclosure in a sample of men and women with documented sexual abuse histories,
after a delay of approximately 13 years. Only 19% of this sample failed to disclose the documented abuse, a rate considerably lower than that Williams
(1994) obtained. Moreover, Goodman et al. found that abuse severity was positively related to disclosure, indicating that more severe abuse was less likely to
be forgotten. The positive relation between abuse severity and disclosure is consistent with the idea that salient events are more likely to be remembered (Howe,
1997, 2000), despite their traumatic nature. Also consistent with the similar
mechanisms idea is the nding of age-related dierences in CSA disclosure
(Goodman et al., 2003; Williams, 1994); individuals who were younger at the

316

time of the abuse, particularly those who experienced the abuse before the oset
of infantile amnesia (i.e., 35 years of age), were less likely to disclose.
Yet, additional predictors of disclosure may be unique to traumatic events.
For instance, higher scores on the Dissociative Experiences Scale (DES; Bernstein
& Putnam, 1986) were associated with a decreased likelihood of disclosure
(Goodman et al., 2003), which may reect memory impairment resulting from
coping mechanisms associated with trauma.
Despite evidence that many adults can recall traumatic childhood experiences, the accuracy of those recollections has rarely been investigated. In one
exception, Alexander, Goodman, Quas, et al. (2002) compared participants
current reports of the abuse with information obtained when the abuse was initially reported. Results indicated that abuse severity was positively related to the
accuracy of participants memories for details of the abuse. That is, individuals
who experienced more severe abuse made fewer commission errors in an interview about their documented CSA experience approximately 12 to 21 years after
the abuse. In addition, participants who nominated CSA as their most traumatic
lifetime event had better memory for the abuse, independent of abuse severity.
These ndings, like those regarding CSA disclosure, are consistent with the idea
that similar memory mechanisms (e.g., event salience) may be operating for both
traumatic and nontraumtic events (Shobe & Kihlstrom, 1997). However,
Alexander, Goodman, Quas, et al. also identied potentially trauma-specic predictors of CSA memory, related to the severity of posttraumatic stress disorder
symptomatology.
Thus, although many adults can recall traumatic events they experienced in
childhood, some individuals do appear to lack conscious recollection of such experiences, and some potential trauma-specic predictors of memory have been
identied (e.g., dissociation; Goodman et al., 2003). The possibility therefore remains that repression or other psychological defenses could be operating, at least
in some cases. In the following section, we review research on repressed memories, including research on memories that have been repressed and later recovered.

Repressed and Recovered Memories

That some individuals fail to disclose documented CSA or other traumatic experiences (e.g., Goodman et al., in press; Williams, 1994) suggests that the memories for these experiences could have been repressed. This possibility is further
bolstered by ndings from retrospective studies, in which sizeable percentages
of individuals reporting histories of CSA also report having had complete to partial amnesia for their abusive experience for some time (e.g., Briere & Conte,
1993; Epstein & Bottoms, 1998; Elliott & Briere, 1995). Briere and Conte, for
instance, found that nearly 60% of their clinical sample described time periods
when they had no memory for self-reported CSA experiences. Unfortunately, in
these studies, no independent corroboration of the crime (or lack of memory for

317

it) exists, so the data must be interpreted cautiously (Schooler & Eich, 2000).
Moreover, ndings from prospective studies (Edelstein, Ghetti, Redlich, Cordn,
& Raskauskas, 2003; Williams, 1995) suggest that forgetting rates may be considerably lower than those reported in retrospective accounts.
Retrospective studies are also limited by their reliance on participants assessments of prior memory states, which may be fallible (Henry, Mott, Caspi, Langley, & Silva, 1994; Hyman & Loftus, 1998; Read, 1999; Schooler, Ambadar, &
Bendiksen, 1997). To illustrate this point, Schooler et al. interviewed four women
who reported having forgotten a sexual abuse experience. Friends and family
members of these women were then questioned to obtain independent corroboration of the abuse and subsequent forgetting. In at least two of the cases, these
corroborative interviews suggested that the women had not entirely forgotten
the abuse; instead, they appeared to have forgotten having previously remembered it.
Moreover, somewhat ironically, estimates of amnesia may increase with active eorts at reconstruction (Read & Lindsay, 2000). Read and Lindsay asked
adult participants to characterize their memories of childhood events, such as
going to summer camp and high school graduation. They were asked whether
there was ever a time when they had less memory for the event than they had
currently. One group of participants was asked to spend as much time as possible during the next month thinking about an assigned event and attempting
to recall as many details of that event as possible (reminiscence group). The second group received the same instructions but was also asked to speak to family
members about the event, to try to nd photographs of it, and to return to the
site of the event if possible (enhanced group). Results indicated that, when
memory characterizations were assessed after a 1month delay, the enhanced
group was more likely to report prior periods of poor memory for the assigned
event (70%) than the reminiscence group (35%). These ndings suggest that
attempts to reconstruct a memory, which may occur during the process of
therapy (Lindsay & Read, 1995), may alter participants perceptions of past
memory states.
In fact, some researchers have argued that false recovered memories of CSA
can be evoked in therapy (Bottoms, Shaver, & Goodman, 1996; Poole, Lindsay,
Memon, & Bull, 1995), particularly when suggestive therapeutic techniques are
used, such as the examination of the clients self-reported dreams (Mazzoni &
Loftus, 1998; Mazzoni, Lombardo, Malvagia, & Loftus, 1999) or hypnosis (Loftus
& Ketcham, 1991; Woodall, 1999). Because corroborating memories recovered
during therapy is dicult, however, the veracity of such reports remains a highly
controversial topic (e.g., Alpert, Brown, & Courtois, 1998; Ornstein, Ceci, &
Loftus, 1998).
Recognizing the division in the scientic community over repressed and recovered memories of childhood abuse (APA Working Group, 1998), many state
courts have refused to admit recovered memories as evidence (e.g., Engstrom v.

318

Engstrom, 1997; Kohout v. Charter Peachford Hospital, 1998). In other states, however, individuals reporting recovered memories have been allowed to testify
about their experiences (Brown, Schein, & Whiteld, 1999; Woodall, 1999).
In fact, some states have extended the statute of limitations for cases involving
repressed memories, such that the time period for ling charges starts from the
time the memory is recovered, rather than from the time the abuse allegedly occurred (Bowman & Mertz, 1996; Brown, Schein, & Hammond, 1998; Gothard
& Ivker, 2000).

Summary and Conclusions

In summary, although further research is clearly needed, extant ndings indicate that many adults can remember traumatic childhood events. Moreover,
recall of these traumatic events and the accuracy of the reports are often predicted by factors such as age and the salience of the event that also predict
memory for nontraumatic events. Additional predictors have been identied,
however, such as dissociation (Goodman et al., 2003), which may be specic to
memory for trauma. In addition, because some individuals do not appear to remember their traumatic experiences, and because others report prior periods of
no memory, we cannot rule out the possibility that these memories were repressed. This possibility, and the phenomena of repressed and recovered memories, warrants further research. As the intense debate over repressed and
recovered memories continues, a challenge for future research will be to clarify
the conditions under which traumatic memories may be lost and later recovered
and to develop criteria to evaluate the validity of such memories.

Symptoms of Psychopathology:

Implications for Eyewitness Memory

As we discussed earlier, individuals who become involved in the legal system

often have experienced or witnessed a distressing or even traumatic event. Because a number of studies suggest associations between traumatic experiences
and a variety of emotional problems, such as PTSD (e.g., Bremner et al., 1992;
Feeny, Zoellner, & Foa, 2000) and high levels of dissociation (e.g., Chu, 1998;
Putnam & Trickett, 1997), we must consider the inuence of psychopathology
on eyewitness memory. That is, how might clinical symptomatology aect an
individuals ability to recount his or her past emotional experiences or to process new emotional information? We focus our attention on PTSD and dissociation because of their strong links with both trauma and memory; however, other
emotional problems likely inuence eyewitness memory capabilities, including
depression (chapter 6) and panic disorder (Coles & Heimberg, 2002; chapter 5,
this volume).

319

PTSD and Eyewitness Memory

PTSD is characterized by a reaction of intense fear, helplessness, or horror to a
traumatic event (e.g., rape, assault), accompanied by persistent intrusive recall
of the experience, avoidance of reminders of the event, and heightened anxiety
(American Psychiatric Association, 1994). Individuals with PTSD often relive
the traumatic event through intrusive thoughts, ashbacks, and nightmares, so
that, according to network models of memory (Bower, 1992), trauma-related
information is easily activated in memory. Increased activation may, in turn,
enhance memory for trauma-related information. Reliving a traumatic event via
intrusive thoughts and ashbacks may also enhance memory by facilitating
rehearsal of the event. Thus, PTSD may be associated with memory biases for
trauma-related information (e.g., Coles & Heimberg, 2002).
In line with this hypothesis, PTSD has been associated with heightened activation of trauma-related information in an emotional Stroop paradigm (see
Williams, Mathews, & MacLeod, 1996, for a review). In this paradigm, participants are shown a word (e.g., assault) printed in color, and their task is to say
aloud the color of the word. Longer latencies to respond presumably reect increased activation of or attention to the printed word, which interferes with color
naming. Findings indicate that individuals with PTSD, compared to those without PTSD (who are often matched on trauma experience, age, or education level,
or more than one of these), tend to take longer to name the color of the word
when emotional or trauma-related words are presented (Bryant & Harvey, 1995;
Cassiday, McNally, & Zeitlin, 1992; McNally, Amir, & Lipke, 1996; McNally,
Clancy, Schacter, & Pitman, 2000; McNally, Kaspi, Riemann, & Zeitlin, 1990;
McNally, Metzger, Lasko, Clancy, & Pitman, 1998; Moradi, Taghavi, NeshatDoost, Yule, & Dalgleish, 1999). Thus, individuals with PTSD may have diculty
inhibiting the activation of emotional or trauma-related words, which increases
their response time on the emotional Stroop task. This heightened activation
should, in turn, enhance memory for trauma-related information.
To examine directly the hypothesis that PTSD symptomatology enhances
memory for trauma-related information, Vrana, Roodman, and Beckham (1995)
compared the performance of Vietnam veterans with and without PTSD on a
memory task involving emotional words (some of which were related to combat) and neutral words. Individuals with PTSD remembered more emotional
words than those without PTSD; however, there was no dierence in memory
for the neutral words. These ndings suggest that individuals with PTSD have a
memory bias in favor of emotional words, at least when trauma-related words
are included.
Furthermore, PTSD has been associated with enhanced memory for real-life
traumatic events. In a study of long-term memory for documented CSA experiences, Alexander, Goodman, Quas, et al. (2002) found that individuals who
nominated CSA as their most traumatic experience remembered the documented

320

case relatively accurately. For individuals who nominated an event other than
CSA (e.g., death of a loved one) as their most traumatic experience, memory
accuracy was positively related to the number of PTSD criteria met. One interpretation of these ndings is that, for individuals who selected CSA as their most
traumatic event, the experience was particularly salient and emotional, so that
longterm memory accuracy was preserved. This may not have been the case
for victims who reported a non-CSA event as their most traumatic experience
and met few PTSD criteria; for these individuals, the CSA experience may not
have been particularly salient or emotional, and, as a result, their memory was
less accurate. Individuals who nominated a non-CSA event and met more PTSD
criteria, however, evinced memory as accurate as those nominating a CSA event.
For these individuals, greater PTSD symptomatology may be associated with
hypervigilance to trauma cues, which would facilitate CSA memory. In addition,
although memory for traumatic events other than CSA was not investigated in
this study, individuals nominating other events could show enhanced memory
for that particular event.
Despite the potentially memory-enhancing eects of PTSD, additional research suggests that PTSD may also be associated with inconsistency of memory
reports for traumatic experiences (Southwick, Morgan, Nicolaou, & Charney,
1997). Southwick et al. examined memories for combat experience among Operation Desert Storm veterans and found that, compared to veterans with fewer
symptoms, those with more PTSD symptoms had less consistent memory reports
for their combat experience over a 2-year period. Because inconsistent reports
often are deemed less credible (Berman & Cutler, 1996; Fisher & Cutler, 1992),
individuals with PTSD may be at a disadvantage when recalling their traumatic
experiences in a forensic setting, even if their accounts are accurate. However,
Southwick et al. did not examine memory accuracy per se.
In contrast to research on memory for traumatic or emotional information,
research on memory for nonemotional information suggests that PTSD may be
associated with a variety of mnemonic decits. For instance, PTSD symptoms,
particularly the chronic intrusion of traumatic memories, may interfere with
working memory functioning (Galletly, Clark, McFarlane, & Weber, 2001; Uddo,
Vasterling, Brailey, & Sutker, 1993), which is critical for making inferences, reasoning, and thinking abstractly (Baddeley, 1986; Miyake & Shah, 1999). Galletly
et al. suggest that chronic processing of trauma-related memories may deplete
cognitive resources necessary for other functions, including working memory.
Alternatively, the chronic intrusion of unwanted memories may reect a failure to inhibit irrelevant information (McNally, 1998), which may, in turn, impair working memory.
Long-term memory decits have also been demonstrated among individuals
with PTSD (Jenkins, Langlais, Delis, & Cohen, 1998; Moradi, Taghavi, NeshatDoost, Yule, & Dalgleish, 2000). Jenkins et al. found that rape victims with PTSD
performed more poorly on a delayed free-recall task compared to rape victims,

321

matched on age and education level, without PTSD. These memory decits were
reduced, however, when participants were provided with retrieval cues, suggesting that PTSD symptomatology may impede the retrieval of nontraumatic memories by making them less accessible. Individuals with the disorder may thus have
greater diculty than other individuals when asked to recount their experiences,
particularly when few retrieval cues are provided. However, these ndings have
not been replicated using trauma-related information, which is arguably more
directly relevant to forensic settings.
Recent ndings suggest that these memory diculties may be the result of
physiological processes associated with trauma. Some evidence indicates that
the hippocampus, a brain region centrally involved in memory consolidation
(Brewer, Zhao, Desmond, Glover, & Gabrieli, 1998; Otten & Rugg, 2002; Wagner
et al., 1998; chapter 3, this volume), is reduced in volume among adults with
PTSD, including combat veterans (Bremner et al., 1995; Gurvits et al., 1996;
Villarreal et al., 2002) and abuse victims (Bremner et al., 1997; but see Bonne
et al., 2001; De Bellis, Hall, Boring, Frustaci, & Moritz, 2001; see Hull, 2002,
for a review of studies linking PTSD and hippocampal volume). Moreover, animal studies indicate that hippocampal atrophy may result from the chronic stress
(Magarinos, McEwen, Flugge, & Fuchs, 1996; McEwen & Magarinos, 1997; Uno,
Tarara, Else, Suleman, & Sapolsky, 1989), that many trauma victims likely experience, and that may at least partially explain the link between trauma and
hippocampal volume.
As a whole, these ndings suggest that the experience of trauma can lead to
physiological processes (e.g., stress) that may aect neurological functioning
(Magarinos et al., 1996; McEwen & Magarinos, 1997; Uno et al., 1989). Neurological changes, in turn, may impair a traumatized individuals ability to form
new memories (Bremner, 2001). Although further research is necessary to test
this model, particularly given the lack of prospective data linking trauma and
hippocampal damage in humans (but see Gilbertson et al., 2002), it may ultimately provide a comprehensive explanation for the memory alterations observed among traumatized individuals. Now, however, it is unclear whether this
model can adequately explain the ndings that PTSD may enhance memory for
trauma-related stimuli and experiences.
Thus, the inuence of PTSD on memory appears to depend on the nature of the
to-be-remembered information. Preliminary evidence suggests that individuals
with PTSD may be more attentive to trauma-related (or emotional) information
and may remember that information better than individuals without PTSD. Research on memory for nonemotional information, however, indicates that PTSD
is associated with decits in working memory, as well as long-term memory. Given
the contrast between these sets of ndings, future research should investigate conditions in which PTSD may compromise memory as well as those in which PTSD
may enhance memory. Further research is also warranted to examine the role of
the hippocampus in PTSD-related mnemonic functioning. In the following section,

322

we turn to dissociation (believed by some to be linked to PTSD), which has also

been linked with trauma, mnemonic functioning, and hippocampal volume.

Dissociation and Eyewitness Memory

Dissociation is dened by DSM-IV as a disruption in the usually integrated functions of consciousness, memory, identity, or perception of the environment. The
disturbance may be sudden or gradual, transient, or chronic(American Psychiatric Association, 1994, p. 477). Dissociation has been associated with the experience of childhood trauma, such as CSA (e.g., Chu, 1998; Putnam & Trickett,
1997) and with memory diculties (see Eisen & Lynn, 2001, for a review). Although some kinds of dissociative experiences are relatively common, even in
nonclinical populations, more severe and frequent experiences may be considered pathological. In the majority of studies discussed here, dissociation is measured on a continuum, reecting individual dierences in dissociative tendencies,
as opposed to dissociation as a clinical disorder per se. Recent research suggests
that individuals with dissociative tendencies may have diculty maintaining
sustained attention to nonemotional stimuli (Freyd, Martorello, Alvarado,
Hayes, & Christman, 1998) and may also show decits in long-term memory
(DePrince & Freyd, 1999; Goodman et al., 2003), both of which have implications for the veracity of eyewitness reports. Sustained attention, for instance,
plays an important role in both the encoding and retrieval of information (e.g.,
Craik, Govoni, Naveh-Benjamin, & Anderson, 1999).
Regarding long-term memory, DePrince and Freyd (1999) found that DES
scores were negatively related to the number of trauma-related words (e.g., incest, victim) remembered in a free-recall task. Memory for neutral words, however, was unrelated to DES scores. These ndings suggest that dissociation may
limit the processing of certain types of material (e.g., that which is emotional or
threatening) and that dissociative individuals may not necessarily be impaired
when processing aectively neutral information. Yet, as we will describe in later
sections, dissociation has also been linked with increased false-memory rates in
tasks using neutral stimuli.
In addition, Goodman et al. (2003) found that DES scores predicted failure to
disclose CSA after a delay of approximately 13 years. In a sample of young adults
with documented sexual abuse histories, individuals scoring high on the DES
(measured concurrently with disclosure) were less likely to disclose the abuse
during an interview about childhood trauma. Although there are a number of
potential explanations for the link between dissociation and disclosure, including an unwillingness to talk about the abuse, one possibility is that highly dissociative individuals had poorer memory for the experience (or poorer memory
in general). Consistent with this idea is evidence that, among participants who
did disclose, those with higher dissociation scores were more likely to report
having forgotten the abusive experience for some time (Edelstein et al., 2003).

323

Recent ndings indicate that, like PTSD, dissociation may be associated with
a reduction in hippocampal volume among CSA victims (Stein, Koverola, Hanna,
Torchia, & McClarty, 1997). Although further research is warranted, these ndings, together with those related to PTSD, suggest that trauma may inuence
memory in a variety of ways. First, the stress inherent during the experience of
a traumatic event can aect later memory for it. Second, that stress, particularly
if chronic, can inuence physiological processes related to the hippocampus,
which may then aect memory for both the traumatic event and subsequent
events. Third, trauma may result in the development of clinical disorders such
as PTSD and dissociation, which may in turn aect memory directly, or indirectly
through the inuence of the hippocampus (or other psychological means) on
mnemonic functioning. Finally, some combination of these processes may increase an individuals susceptibility to false-memory formation.

Dissociative Tendencies, PTSD,

and False-Memory Formation

Both PTSD and dissociation have also been associated with increased falsememory rates. Concerns about false memories are particularly relevant to forensic settings, because false memories for traumatic experiences could potentially
result in mistaken accusations, and even imprisonment of innocent individuals.
Yet misclassifying a true memory as false could hinder the prosecution of oenders and further traumatize victims. Thus, if we can understand factors that may
lead to false-memory formation, we may be able to eliminate them.
To investigate false memories in the laboratory, many researchers have employed the DRM paradigm, an experimental procedure commonly used to
study false recall and recognition. In this paradigm, introduced by Deese (1959)
and modied by Roediger and McDermott (1995) and Read (1996), participants
are presented with a series of semantically associated words (e.g., bed, pillow,
dream) organized into lists, the majority of which are aectively neutral. At test,
large numbers of participants falsely recall and recognize words that were not
studied (e.g., sleep) but are semantically related to the studied words (e.g.,
Roediger & McDermott, 1995; Roediger, McDermott, & Robinson, 1998).
Several studies using the DRM paradigm suggest that trauma victims, particularly those with associated psychopathology, are more prone to false memories than are nontraumatized individuals (Bremner, Shobe, & Kihlstrom, 2000;
Clancy, Schacter, McNally, & Pitman, 2000; Zoellner, Foa, Brigidi, & Przeworski,
2000). Bremner et al., for instance, compared women with self-reported histories of CSA, both with and without a diagnosis of PTSD, to a control group of
nonabused men and women without PTSD (matched on age and education).
Women with sexual abuse histories and a diagnosis of PTSD were more likely to
falsely recognize nonpresented words than were individuals in the other two

324

groups. In fact, women with PTSD were more likely than other participants to
falsely recall and recognize unstudied words regardless of their semantic relation to the studied items, a result that perhaps reects an overly liberal criterion
for distinguishing between studied and unstudied items.
Findings from DRM studies also suggest that high scorers on the DES are more
prone to false-memory formation (Clancy et al., 2000; Winograd, Peluso, &
Glover, 1998). Although the DRM paradigm has been criticized for lacking
external validity (e.g., Freyd & Gleaves, 1996), the ndings just discussed are bolstered by results from other studies, using other false-memory paradigms, that
indicate a positive relation between DES scores and the tendency to assent to false
autobiographical events (Hyman & Billings, 1998; Qin, 1999). Hyman and Billings presented college students with information about their childhood, obtained
from the students parents, and asked participants to recall as many details about
the events as possible. One of the events described, however, had not actually
occurred. Results indicated that high scorers on the DES were more likely than
low scorers to assent to the false autobiographical event. According to the
authors, DES scores may reect willingness to engage in constructive memory
processes. That is, individuals scoring high on this measure may have made more
eorts to remember details about events from the past, some of which did not
occur. Attempts at reconstruction, in turn, may have increased the likelihood
of assenting to a false event.
Alternatively, false memories may result from a lax criterion used to decide
whether items are old or new (Hekkanen & McEvoy, 2002; Roediger & McDermott,
1999). That is, as individuals become more willing to characterize items as old,
they are more likely to falsely endorse unstudied items as studied. Interestingly,
Hekkanen and McEvoy found that DES scores were positively related to the criterion
used by participants to discriminate between old and new items: Highly dissociative individuals, more so than those with lower dissociation scores, tended to adopt
a more lax criterion, resulting in a greater number of items characterized as old.
Thus, dissociation may predict false memory rates because of the inuence of dissociative tendencies on decision-making criteria. Of course, a lax criterion also may
be responsible for responses to the DES and false-memory rates. Nevertheless, it is
interesting to reconsider the emotion-induced recognition bias (Windmann &
Kutas, 2001) discussed earlier, in which emotional stimuli are associated with
higher false-alarm rates (i.e., a more liberal criterion). Perhaps the heightened rates
of false memories among dissociative individuals (or those with PTSD) is the result of a similar aective process, in which even neutral stimuli are perceived in
emotional terms. Windmann and Krger (1998) and Windmann, Sakhavat, and
Kutas (2002) have proposed that this kind of process may underlie the memory
performance of individuals with panic disorder.
Another potential explanation for the link between psychopathology and
false-memory formation comes from research suggesting that trauma history

325

is associated with overgeneral memories (McNally, Lasko, Macklin, & Pitman, 1995; McNally, Litz, Prassas, Shin, & Weathers, 1994). McNally et al.
have found that traumatized individuals with PTSD have more diculty than
other individuals recalling specic personal memories (e.g., the rst time I
went to Hawaii) in response to cue words (e.g., happy, angry). Instead, participants with PTSD are more likely to recall general memories (e.g., the yearly
vacations I took), perhaps reecting a reliance on categorical or gist memory
traces. Insofar as PTSD is related to a tendency to rely on gist memories, and
reliance on gist is associated with false-memory formation (Brainerd & Reyna,
1998; Brainerd, Reyna, & Poole, 2000), individuals with PTSD may be more
susceptible to false memories.
Thus, both PTSD and dissociative tendencies have been associated with increased false-memory rates, possibly as a result of the criterion used to discriminate between old and new information or because of a tendency to recall
overgeneral memories. Insofar as these ndings extend to real-life situations,
individuals with PTSD and dissociative tendencies may at times have diculty
discriminating between details of events that did and did not occur, so their eyewitness memory capabilities could be impaired. This possibility highlights the
importance of forensic interview techniques that limit false-memory formation,
particularly when individuals with PTSD or dissociation are interviewed. Perhaps false-alarm rates could be reduced by teaching individuals to shift their
response criterion. However, few studies have examined the extent to which
PTSD and dissociation are associated with false memories for traumatic or even
emotional information; thus, further research is necessary to examine the applicability of extant ndings to the forensic domain.

Summary and Conclusions

In summary, recent evidence links PTSD with enhanced memory for trauma-related information, as well as increased activation of such information, which may
serve a memory-enhancing function. With respect to memory for nontraumarelated information, a growing area of research indicates that both PTSD and dissociative tendencies are associated with decits in working memory, long-term
memory, and susceptibility to false memories. Together, these ndings suggest that
the relation between psychopathology and eyewitness memory may be complex
and that, in some cases, considering the nature of the to-be-remembered information may be important. In a forensic setting, an individuals legal involvement may
have been precipitated by a traumatic event, which may, in turn, result in
psychopathology.
Despite recent advances, however, many questions remain about the association between psychopathology and memory. First, only a handful of studies have
examined the inuence of PTSD or dissociation on memory for emotional material,

326

including memory for the precipitating traumatic event, and even fewer of these
have examined false-memory formation. Because memory for a particular traumatic experience may be central to a legal prosecution (or defense), future research
should investigate the inuence of PTSD on memory for trauma. Second, research
linking PTSD and dissociation with false-memory formation has often employed
the DRM paradigm, which may have limited external validity. Further research is
necessary to examine the applicability of DRM ndings to real-life situations. Third,
studies of CSA-related memory functioning have predominantly included female
participants. Although this practice may be warranted, given that women are
reportedly sexually abused more often than are men (Finkelhor, 1984), it is unclear how current ndings may extend to sexually abused men. If men and women
process emotional information dierently (e.g., Canli et al., 2002), considering
gender dierences in this area, as well as in the context of emotion and memory
research more generally, may be important.

Developmental Considerations
A potential inuence on eyewitness memory capabilities is the age of the individual whose memory is being examined. Because childrens developing cognitive and emotional abilities may aect the way they process and remember
emotional events, we focus now on childrens eyewitness memory and the developmental implications of stressful and traumatic life experiences on memory.
Extant research suggests that, like adults, children can remember stressful
personal experiences quite accurately (e.g., Alexander, Goodman, Schaaf, et al.,
2002; Peterson & Bell, 1996), even over relatively long delays (e.g., Peterson &
Whalen, 2001); however, developmental factors may play a role in the veracity
of their reports. The quality of childrens memories for negative emotional experiences may dier from that of adults because of developmental dierences in
biological processes (e.g., brain development; Bremner & Vermetten, 2001),
social-emotional abilities (e.g., understanding and interpretation of emotions and
emotional events; Stein & Liwag, 1997; Campos, Mumme, Kermoian, & Campos, 1994), and cognitive skills (e.g., speed of processing: Kail, 2000; knowledge
base: Bjorklund, 1987; strategy use: Miller & Seier, 1994; executive function:
Diamond, Prevor, Callender, & Druin, 1997). These important developments
contribute to dierences in how children process and recall emotional information, how they cope with and report particularly traumatic experiences, and how
clinical disorders may inuence their memory accuracy. In the following sections, we discuss childrens memory for emotional and traumatic events and the
inuence of clinical disorders on childrens eyewitness memory. We focus on
similarities and dierences between the memory performance of children and
that of adults, as well on potential reasons for observed dierences.

327

Emotion and the Accuracy

of Childrens Eyewitness Memory

Few published studies have examined childrens memory for emotional versus nonemotional information. Existing research indicates that, like adults,
children are more likely to remember emotional than nonemotional information (Davidson & Jergovic, 1996; Davidson, Luo, & Burden, 2001; but see
Moradi et al., 2000). For instance, in one investigation of childrens ability to
recall stories ranging in emotionality (Davidson et al., 2001), 6- to 11-yearold children heard four stories, two highly emotional and two less so. Each story
consisted of three emotional incidents (i.e., happiness, sadness, anger); the
stories diered in the intensity with which the emotion was evoked. The results of two experiments showed that, across age groups, children recalled more
highly emotional behaviors than less emotional behaviors from the stories, suggesting that emotional content enhanced memory. No signicant dierences
were found between recall for positive and negative information. These ndings
are consistent with those obtained with adults (e.g., Heuer & Reisberg, 1990;
Ochsner, 2000), in which emotional stories and other stimuli are associated
with better memory than are nonemotional stimuli. [For related arguments,
see chapters 1and 8Eds.]
Although Davidson et al. (2001) found that emotional content enhanced
memory for both older and younger children, age dierences were nevertheless
evident. For the shorter stories, older children remembered more low-emotion
information than did younger children, yet there were fewer developmental
dierences between older and younger childrens memory for the highly emotional information. For the longer stories, older children remembered more
emotional and nonemotional information than younger children. One explanation for these ndings is that both older and younger children attend preferentially to the highly emotional information, resulting in enhanced recall. For the
younger children, because of their more limited cognitive abilities, greater attention to emotional information may hamper recall of the less emotional information. In contrast, because of their greater processing/memory capacities, older
children may have less diculty processing and remembering both emotional
and nonemotional information. These dierences may have been magnied for
the longer stories.
The degree of stress experienced by children during an emotional event has
also been associated with their later memory for that event (e.g., Alexander,
Goodman, Schaaf, et al., 2002; Bahrick, Parker, Fivush, & Levitt, 1998;
Goodman et al., 1991; Vandermaas et al., 1993); however, ndings have been
somewhat inconsistent regarding the nature of this relation. Vandermaas
et al., for instance, examined childrens memory for a dental procedure and
found that high levels of anxiety, as rated by the researchers, impaired the
memory performance of 7- to 8-year-olds, but not 4- to 5-year-olds. Childrens

328

self-reports and parental reports of anxiety, however, were unrelated to childrens memory. In contrast, several studies of childrens memory for inoculations indicate that observer-rated stress is associated with better memory (e.g.,
Alexander, Goodman, Schaaf, et al., 2002; Goodman et al., 1991; Shrimpton
et al., 1998; also chapter 8 here).
As with studies of memory for emotional stimuli, however, age dierences
characterize childrens abilities to recount stressful experiences (e.g., Bahrick
et al., 1998; Goodman et al., 1991). Age-related improvements in both the
amount of information recalled and the accuracy of that information have been
found in studies of childrens memory for inoculations (e.g., Alexander,
Goodman, Schaaf, et al., 2002; Goodman et al., 1991), natural disasters (Bahrick
et al., 1998), and sniper attacks (Pynoos & Nader, 1989). Moreover, children
who are older at the time of the event remember central details more accurately
than do younger children (Goodman et al., 1991).
In summary, although some studies suggest that childrens memory is enhanced by negative emotion, others suggest that negative emotion impairs (or
is unrelated to) childrens memory. This inconsistency, mirrored in the adult literature, may be due to the many dierent ways that emotion is assessed and the
variety of events for which memory is tested. Thus, further research is clearly
needed before a denitive answer can be reached about the inuence of emotion
on childrens memory. Further research is also needed to address the similarities and dierences in childrens and adults emotional memories. For instance,
do children exhibit emotional memory biases (e.g., Windmann & Kutas, 2001)
similar to those of adults? Investigation of developmental dierences in the tendency to falsely recognize emotional stimuli may provide insight into the mechanisms underlying this bias.
In addition, it is unclear whether the gender dierences found in adults emotional memory (e.g., Canli et al., 2002) apply also to children. In general, gender dierences in emotional memory have not been demonstrated with children;
however, there are several reasons to believe that they may exist. For example,
mothers talk more about emotions with female than with male children (e.g.,
Fivush, 1998; Kuebli, Butler, & Fivush, 1995), and parent-child conversation
has been found to facilitate childrens memory (chapter 8 here). Also, female children elaborate more when discussing emotional experiences than do boys (e.g.,
Fivush, Brotman, Buckner, & Goodman, 2000), which may facilitate memory
retention over time.

Childrens Memory

for Traumatic Events

As discussed earlier, research suggests that most adults can remember traumatic
childhood experiences, such as CSA, even after long delays (Goodman et al.,

329

2003; Williams, 1994). In this section, we will consider childrens memory for
traumatic experiences. Childrens memory may be particularly vulnerable to
distortion or loss because they do not have the cognitive capacity to cope with
such highly stressful experiences. Thus, when children experience traumas, they
may be more likely than adults to put up defenses such as dissociation or repression, which could result in memory distortions (Armsworth & Holaday, 1993).
Alternatively, although memories for traumatic events may dier signicantly
from memories for nontraumatic events, many of the same memory principles
may apply for both children and adults (Pezdek & Taylor, 2002). That is, like
adults, children may remember traumatic experiences quite accurately, and their
memory may be facilitated by factors such as the salience of the event and the
age at which it was experienced. In the following sections, we discuss two lines
of research that have examined childrens memories for documented traumatic
experiences. The rst concerns childrens memory for highly stressful medical
procedures, and the second concerns childrens memory for abuse.
Childrens Memory for Medical Procedures Several studies have examined
childrens memory for a highly stressful and invasive medical procedure, voiding cystourethrogram uoroscopy (VCUG; Brown et al., 1999; Goodman, Quas,
Batterman-Faunce, Riddlesberger, & Kuhn, 1994, 1997; Merritt, Ornstein, &
Spicker, 1994; Ornstein, Baker-Ward, Myers, Principe, & Gordon, 1995; Quas
et al., 1999). Results from these studies indicate that children as young as 3 years
old can remember these experiences relatively well, even after a delay of several
years. As with studies of less stressful events (e.g., laboratory play sessions: Price
& Goodman, 1990; classroom demonstrations: Powell & Thompson, 1996),
however, childrens age at the time of the event is an important predictor of their
later memory (e.g., Brown et al., 1999; Quas et al., 1999). Quas et al., for instance, found that most children who experienced the VCUG procedure at age 3
or older evinced at least some memory for it, even after a delay of 3 years. In
contrast, most children who experienced the procedure before age 3 were unable to demonstrate a clear memory for the event.
Studies of childrens memories for stressful injuries and resulting emergency
room procedures have also yielded age dierences (e.g., Burgwyn-Bailes et al.,
2001; Peterson, 1999; Peterson & Bell, 1996; Peterson & Rideout, 1998; Peterson
& Whalen, 2001; Terr, 1988). Like adults, however, children were more accurate
about central than peripheral details of the injury experience (Peterson & Bell,
1996; Peterson & Whalen, 2001), and, consistent with adult research indicating that stress enhances memory for central details, children who were highly
stressed during the injury had better memory for central details of the hospital
event (Peterson & Whalen, 2001).
Moreover, childrens memories for central details of highly stressful events
are relatively well retained over time (e.g., Burgwyn-Bailes et al., 2001;
Peterson, 1999; Peterson & Whalen, 2001). Burgwyn-Bailes et al., for in-

330

stance, examined 3- to 7-year-old childrens memory for emergency room procedures shortly following their injury and again after a 1-year delay. Childrens
recall after 1 year was nearly as accurate as it had been right after the procedure: 72% of specied features of the procedure were remembered accurately
after 1 year, compared to 78% shortly after the procedure (but see Goodman
et al., 1991). Even after a delay of approximately 5 years, Peterson and Whalen
found that childrens recall of the central components of their injury experience was highly accurate. Children who were as young as 2 years old at the
time of the injury fairly accurately remembered the event, although they were
still less accurate than the older children. Memory for the hospital visit, however, was less accurate after the delay.
Childrens Memory for Abuse Although few studies of childrens memory for
documented abuse exist, extant research suggests that children can accurately
(though not necessarily completely) remember their experiences (e.g., Bidrose
& Goodman, 2000; Orbach & Lamb, 1999; Terr, 1988). Bidrose and Goodman,
for instance, examined the memory of four girls (ages 8 to 15) who had been
involved in sexually abusive acts with several men. The acts were documented
by one of the perpetrators, who photographed or audiotaped (or both) many of
the abusive incidents. The girls testimony, obtained from police interviews and
court transcripts, could thus be compared to the photographs and audiotapes to
determine the accuracy of the victims accounts. Overall, about 80% of the girls
allegations could be veried, so their memory for the abuse was highly accurate,
especially as some acts may not have been documented. Omissions of information (i.e., failing to disclose a particular documented act) were more likely than
were allegations of nonveriable acts.
In another case study, Orbach and Lamb (1999) examined the accuracy of a
13-year-old victims report of a sexual abuse incident. The victims account was
obtained from an investigative interview and was compared to an audiotaped
record of the abuse incident and to the victims sisters account of what happened
immediately prior to and after the abuse. Over 50% of the details the victim reported were corroborated by the audiotaped evidence, and additional details were
conrmed by the perpetrators or sisters reports. Thus, although this study included only one child, the ndings support those of Bidrose and Goodman (2000)
in suggesting that childrens eyewitness memories for abuse experiences can be
quite accurate (see also Jones & Krugman, 1986). In addition, consistent with
previous research on memory for emotional experiences, most of the corroborated details were determined to be central (as opposed to peripheral) to the event
(i.e., plot-related details that specied the outline of the sexual event).
In summary, although only limited research exists, ndings consistently suggest that, like adults, children can remember highly distressing, and perhaps even
traumatic, events. In addition, in some studies, children have remembered central details of traumatic experiences more accurately than peripheral details, a

331

nding paralleled in the adult literature. However, research suggests that, as with
less traumatic events, age dierences are evident in the accuracy of childrens
recall, with older children demonstrating better memory. As we describe in the
next section, trauma may also have a more indirect inuence on childrens
memory, through clinical disorders such as PTSD and dissociation.

Clinical Disorders and Childrens

Eyewitness Memory

Despite some childrens resilience in the face of trauma, other children may develop clinical symptomatology, such as PTSD and dissociation, as a result of their
early experiences (e.g., Putnam, 1996, 1997). These emotional problems, in
turn, may have implications for childrens memory accuracy (Eisen & Goodman,
1998; Eisen & Lynn, 2001). For instance, like adults, children with PTSD may
have enhanced memory for trauma-related information but impaired memory
for nontrauma-related information, whereas dissociative children may show impaired memory for both types of information. Or the relation between psychopathology and memory may change with age.
To examine the inuence of emotional content and PTSD on childrens memory,
Moradi and colleagues (2000) compared memory for words that were negative,
positive, or neutral in children with and without PTSD. The clinical group was
diagnosed with nonabuse-related PTSD and was matched to a control group on
age, verbal intelligence, and reading ability. Participants were presented with a
series of words on a computer screen and were then administered a recall test, followed by a recognition test. Findings showed that participants with PTSD recalled
fewer words than controls; however, this main eect was qualied by an interaction between group and word type, such that control participants recalled more
neutral than positive and negative words. In contrast, PTSD participants recalled
more neutral and negative compared to positive words. Thus, compared to the
control participants, individuals with PTSD showed better memory for the negative words than the neutral words. Measures of recall bias (i.e., preferential memory
for negative vs. neutral or positive words) were unrelated to participants age. In
addition, although some of the negative words were trauma-related, no reliable
dierences in memory were found between trauma-related and nontrauma-related
words for either group. Regarding recognition, signal detection analyses revealed
that participants recognized neutral words more accurately than positive words
and positive words more accurately than negative words, but there were no group
dierences. Thus, as in studies of adults memory for nonemotional information,
when group dierences were evident, participants with PTSD remembered fewer
words than controls. Also consistent with the adult literature, PTSD participants
evinced a negative emotion bias in recall.
In an investigation of general cognitive functioning, Beers and DeBellis
(2002) compared the performance of children with maltreatment-related PTSD

332

with a matched group of non-PTSD, nonmaltreatment controls. Results

showed that children with PTSD performed more poorly on tasks measuring
attention and executive function. A similar study compared the cognitive performance of adolescents with and without PTSD (Moradi et al., 1999). Results
indicated that adolescents with PTSD evinced decits in overall memory performance, including immediate and delayed story recall. These ndings, consistent with those obtained in research with adults, suggest that PTSD may
impair general memory functioning.
Few studies of childrens dissociation and memory exist. In one study (Eisen,
Goodman, Qin, & Davis, 1998; Eisen, Qin, Goodman, & Davis, 2002), children
were questioned about a medical exam they had experienced as a result of suspected child physical or sexual abuse. Childrens memories and suggestibility for
the exam were investigated in relation to their level of dissociation, as assessed
by the Childrens Perceptual Alteration Scale (CPAS; Evers-Szostak & Sanders,
1992), the Adolescent-Dissociative Experiences Scale (A-DES; Armstrong,
Putnam, Carlson, Libero, & Smith, 1997), or the Child Dissociative Checklist
(CDC; Putnam, Helmers, & Trickett, 1993), or more than one of these. In contrast to research with adults, this study found no signicant relation between
dissociation and childrens memories or suggestibility for the exam (but see Eisen,
Qin, Goodman, & Davis, 1999).
Thus, further research is clearly needed to examine the inuence of PTSD
and dissociation on childrens memories for emotional and nonemotional information. Extant studies indicate that, similar to ndings with adults, PTSD
is associated with impairments in general memory functioning. Findings
are less clear, however, regarding the relation between PTSD and childrens
memories for emotional information. Similarly, whether dissociation in children is associated with memory for emotional or nonemotional information
remains unknown. Finally, because some studies linking PTSD and dissociation with memory, in either child or adult populations, have not fully considered the role of other clinical disorders (e.g., depression, general anxiety),
separating the inuence of one form of psychopathology from that of another
is dicult.

Summary and Conclusions

In summary, despite their less than fully developed emotional and cognitive
abilities, children can provide accurate reports about highly stressful and even
traumatic experiences. In addition, their accuracy, like that of adults, is aected
by factors such as behavioral stress and emotional content of the event. Developmental change in the relation between emotion and memory has rarely been
investigated, however. Further research in this area can focus on developmental dierences in the eects of emotion, stress, and concomitant clinical disorders on information processing and, ultimately, memory reports.

333

Conclusions and Implications for

Future Research

We began this chapter by describing victims/witnesses of a crimethe abduction of an infantand asking about their ability to recount the details of that
crime in a forensic setting. Fortunately, in this example, the witnesses were able
to provide enough information to local police to apprehend the kidnapper, and
the infant was returned quickly and safely to her mother. Thus, as in many
crimes, the testimony of eyewitnesses was critical to police eorts. And, although
the kidnapper in this case confessed to the abduction, such is not always the case.
Eyewitness accounts are often extremely important for factnders determinations in a variety of forensic settings, including criminal trials. Given the inherent emotional nature of many crimes, we need to understand how emotion
inuences eyewitness accounts. As a whole, the research throughout this chapter clearly indicates that emotion can powerfully inuence eyewitness memory.
Findings from studies of emotion and memory thus have the potential to inform
the forensic community, including police ocers, attorneys, judges, and jurors.
Based on our review of the literature, we recommend several avenues for forensically relevant research: First, examining the extent to which the emotioninduced recognition bias found for negative stimuli extends to real-life negative
emotional experiences is important. Are victims or witnesses of crimes, for instance, more prone to false recognition about a previously experienced stressful
or traumatic event? Second, because recovered memories have been allowed as
evidence in some states, researchers should evaluate the extent to which traumatic experiences may be forgotten and the conditions under which they may
be later recalled. Third, further research should examine the inuence of trauma
on memory for emotional information, including links between PTSD, dissociation, and true versus false (emotional) memories. In particular, whether PTSD
enhances memory for trauma-related information should be determined. Finally,
longitudinal research on emotion and memory is needed. Many of these issues
have not been adequately examined from a developmental perspective, including the ways in which the relation between emotion and memory may change
with development. Expanding the scope of current emotion and memory research in these directions, and continuing to accurately and cautiously disseminate relevant ndings to the legal community, will be important challenges for
current and future researchers.

Note
This material is based upon work supported by the National Science Foundation
under Grant No. 9602125. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily
reflect the views of the National Science Foundation.

334

References
Alexander, K. W., Goodman, G. S., Quas, J. A., Ghetti, S., Edelstein, R. S., Redlich,
A. D., et al., (2002, September). Long-term memory for child sexual abuse and legal
experiences. Paper presented at the European Conference on Psychology and
Law, Leuven, Belgium.
Alexander, K. W., Goodman, G. S., Schaaf, J. M., Edelstein, R. S., Quas, J. A., &
Shaver, P. R. (2002). Childrens memory and suggestibility: Relations with
attachment and cognitive inhibition. Journal of Experimental Child Psychology,
83, 262290.
Alpert, J. L., Brown, L. S., & Courtois, C. A. (1998). Comment on Ornstein, Ceci, and
Loftus (1998): Adult recollections of childhood abuse. Psychology, Public Policy,
and Law, 4, 10521067.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders. 4th ed. Washington, DC: American Psychiatric Association.
Anooshian, L. J., & Hertel, P. T. (1994). Emotionality in free recall: Language
specicity in bilingual memory. Cognition & Emotion, 8, 503514.
APA Working Group. (1998). Final conclusions of the American Psychological
Association Working Group on investigation of memories of childhood abuse.
Psychology, Public Policy, and Law, 4, 933940.
Armstrong, J., Putnam, F., Carlson, E., Libero, D., & Smith, S. (1997). Development
and validation of a measure of adolescent dissociation: The Adolescent Dissociative Experiences Scale. Journal of Nervous and Mental Disease, 185, 17
Armsworth, M. W., & Holaday, M. (1993). The eects of psychological trauma on
children and adolescents. Journal of Counseling and Development, 72, 4956.
Baby girl snatched at Texas Wal-Mart. (2002, August 15). Cable News Network.
Retrieved August 15, 2002, fromhttp://www.cnn.com/2002/US/08/13/
texas.infant.abduction.
Baddeley, A. D. (1986). Working memory. Oxford: Oxford University Press.
Bahrick, L. E., Parker, J. F., Fivush, R., & Levitt, M. (1998). The eects of stress on
young childrens memory for a natural disaster. Journal of Experimental Psychology: Applied, 4, 308331.
Beers, S. R., & DeBellis, M. D. (2002). Neuropsychological function in children with
maltreatment-related posttraumatic stress disorder. American Journal of Psychiatry, 159, 483486.
Berman, G. L., & Cutler, B. L. (1996). Eects of inconsistencies in eyewitness testimony
on mock-juror decision making. Journal of Applied Psychology, 81, 170177.
Bernstein, E. M., & Putnam, F. W. (1986). Development, reliability, and validity of
a dissociation scale. Journal of Nervous and Mental Disease, 174, 727735.
Berntsen, D. (2002). Tunnel memories for autobiographical events: Central details
are remembered more frequently from shocking than happy experiences.
Memory & Cognition, 16, 10101020.
Bidrose, S., & Goodman, G. S. (2000). Testimony and evidence: A scientic case study
of memory for child sexual abuse. Applied Cognitive Psychology, 14, 197213.
Bjorklund, D. F. (1987). How age changes in knowledge base contribute to the development of childrens memory: An interpretive review. Developmental Review,
7, 93130.

335

Bonne, O., Brandes, D., Gilboa, A., Gomori, J. M., Shenton, M. E., Pitman, R. K.,
et al. (2001). Longitudinal MRI study of hippocampal volume in trauma survivors with PTSD. American Journal of Psychiatry, 158, 12481251.
Bornstein, B. H., Liebel, L. M., & Scarberry, N. C. (1998). Repeated testing in eyewitness memory: A means to improve recall of a negative emotional event.
Applied Cognitive Psychology, 12, 119131.
Bottoms, B. L., Shaver, P. R., & Goodman, G. S. (1996). An analysis of ritualistic and
religion-related child abuse allegations. Law and Human Behavior, 20, 134.
Bower, G. H. (1992). How might emotions aect learning? In S.-. Christianson
(Ed.), Handbook of emotion and memory: Research and theory (pp. 331). Hillsdale,
NJ: Erlbaum.
Bowman, C. G., & Mertz, E. (1996). A dangerous direction: Legal intervention in
sexual abuse survivor therapy. Harvard Law Review, 190, 549639.
Bradley, M. M., Greenwald, M. K., Petry, M. C., & Lang, P. J. (1992). Remembering
pictures: Pleasure and arousal in memory. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 18, 379390.
Brainerd, C. J., & Reyna, V. F. (1998). Fuzzy-trace theory and childrens false memories. Journal of Experimental Child Psychology, 71, 81129.
Brainerd, C. J., Reyna, V. F., & Poole, D. A. (2000). Fuzzy-trace theory and false
memory: Memory theory in the courtroom. In D. F. Bjorklund (Ed.). Falsememory creation in children and adults: Theory, research, and implications (pp. 93
127). Mahwah, NJ: Erlbaum.
Bremner, J. D. (2001). Hypotheses and controversies related to eects of stress on
the hippocampus: An argument for stress-induced damage to the hippocampus in patients with posttraumatic stress disorder. Hippocampus, 11, 7581.
Bremner, J. D., Randall, P., Scott, T. M., Bronen, R. A., Seibyl, J. P., Southwick, S. M.,
et al. (1995). MRI-based measurement of hippocampal volume in combat-related
posttraumatic stress disorder. American Journal of Psychiatry, 152, 973981.
Bremner, J. D., Randall, P., Vermetten, E., Staib, L., Bronen, R. A., Mazure, C., et al.
(1997). Magnetic resonance imaging-based measurement of hippocampal volume in posttraumatic stress disorder related to childhood physical and sexual
abuse: A preliminary study. Biological Psychiatry, 41, 2332.
Bremner, J. D., Shobe, K. K., & Kihlstrom, J. F. (2000). False memories in women
with self-reported childhood sexual abuse: An empirical study. Psychological
Science, 11, 333337.
Bremner, J. D., Southwick, S., Brett, E., Fontana, A., Rosenheck, R., & Charney, D.
(1992). Dissociation and posttraumatic stress disorder in Vietnam combat veterans. American Journal of Psychiatry, 149, 328332.
Bremner, J. D., & Vermetten, E. (2001). Stress and development: Behavioral and
biological consequences. Development and Psychopathology, 13, 473489.
Brewer, J. B., Zhao, Z., Desmond, J. E., Glover, G. H., & Gabrieli, J. D. E. (1998). Making .memories: Brain activity that predicts how well visual experience will be
remembered. Science, 281, 11851187.
Brewin, C. R., Dalgleish, T., & Joseph, S. (1996). A dual representation theory of
posttraumatic stress disorder. Psychological Review, 103, 670686.

336

Briere, J., & Conte, J. R. (1993). Self-reported amnesia for abuse in adults molested
as children. Journal of Traumatic Stress, 6, 2131.
Brown, D. A., Salmon, K., Pipe, M.-E., Rutter, M., Craw, S., & Taylor, B. (1999).
Childrens recall of medical experiences: The impact of stress. Child Abuse and
Neglect, 23, 209216.
Brown, D., Schein, A. W., & Hammond, D. C. (1998). Memory, trauma treatment,
and the law. New York: W. W. Norton.
Brown, D., Schein, A. W., & Whiteld, C. L. (1999). Recovered memories: The current weight of the evidence in science and in the courts. Journal of Psychiatry
and Law, 27, 5156.
Bryant, R. A., & Harvey, A. G. (1995). Processing threatening information in posttraumatic stress disorder. Journal of Abnormal Psychology, 104, 537541.
Burgwyn-Bailes, E., Baker-Ward, L., Gordon, B. N., & Ornstein, P. A. (2001).
Childrens memory for emergency medical treatment after one year: The impact of individual dierence variables on recall and suggestibility. Applied Cognitive Psychology, 15, S25S48.
Burke, A., Heuer, F., & Reisberg, D. (1992). Remembering emotional events.
Memory & Cognition, 20, 277290.
Cahill, L, & McGaugh, J. L. (1998). Mechanisms of emotional arousal and lasting
declarative memory. Trends in Neuroscience, 21, 294299.
Cahill, L., Prins, B., Weber, M., & McGaugh, J. L. (1994). -Adrenergic activation
and memory for emotional events. Nature, 371, 702704.
Campos, J. J., Mumme, D. L., Kermoian, R., & Campos, R. G. (1994). A functionalist
perspective on the nature of emotion. Monographs of the Society for Research in
Child Development, 59, 284303.
Canli, T., Desmond, J. E., Zhao, Z., & Gabrieli, J. D. E. (2002). Sex dierences in the
neural basis of emotional memories. Proceedings of the National Academy of Sciences, 99, 1078910794.
Cassiday, K. L., McNally, R. J., & Zeitlin, S. B. (1992). Cognitive processing of trauma
cues in rape victims with post-traumatic stress disorder. Cognitive Therapy and
Research, 16, 283295.
Christianson, S.-. (1992). Emotional stress and eyewitness memory: A critical review. Psychological Bulletin, 112, 284309.
Christianson, S.-., & Hubinette, B. (1993). Hands up: A study of witnesses emotional reactions and memories associated with bank robberies. Applied Cognitive Psychology, 7, 365379.
Christianson, S.-., & Loftus, E. (1991). Remembering emotional events: The fate
of detailed information. Cognition & Emotion, 5, 81108.
Chu, J. A. (1998). Dissociative symptomatology in adult patients with histories of
childhood physical and sexual abuse. In J. D. Bremner & C. R. Marmar (Eds.),
Trauma, memory, and dissociation (pp. 179203). Washington, DC: American
Psychiatric Press.
Clancy, S. A., Schacter, D. L., McNally, R. J., & Pitman, R. K. (2000). False recognition in women reporting recovered memories of sexual abuse. Psychological
Science, 11, 2631.

337

Coles, M. E., & Heimberg, R. G. (2002). Memory biases in the anxiety disorders:
Current status. Clinical Psychology Review, 22, 587627.
Cooper, A. M. (1986). Toward a limited denition of psychic trauma. In A. Rothstein
(Ed.), The reconstruction of trauma: Its signicance in clinical work (pp. 4156).
Madison, CT: International Universities Press.
Cooper, B. S., Kennedy, M. A., Herv, H. F., & Yuille, J. C. (2002). Weapon focus in
sexual assault memories of prostitutes. International Journal of Law and Psychiatry, 25, 181191.
Craik, F. I. M., Govoni, R., Naveh-Benjamin, M., & Anderson, N. D., (1999). The
eects of divided attention on encoding and retrieval processes in human
memory. Journal of Experimental Psychology: General, 12, 159180.
Davidson, D., & Jergovic, D. (1996). Childrens memory for typical and atypical actions
in script-based stories. Journal of Experimental Child Psychology, 55, 104126.
Davidson, D., Luo, Z., & Burden, M. J. (2001). Childrens recall of emotional
behaviours, emotional labels, and nonemotional behaviours: Does emotion
enhance memory? Cognition & Emotion, 15, 126.
De Bellis, M. D., Hall, J., Boring, A. M., Frustaci, K., & Moritz, G. (2001). A pilot longitudinal study of hippocampal volumes in pediatric maltreatment-related
posttraumatic stress disorder. Biological Psychiatry 50, 305309.
Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in
immediate recall. Journal of Experimental Psychology, 58, 1722.
DePrince, A. P., & Freyd, J. J. (1999). Dissociative tendencies, attention, and
memory. Psychological Science, 10, 449452.
Doerksen, S., & Shimamura, A. P. (2001). Source memory enhancement for emotional words. Emotion, 1, 511.
Diamond, A., Prevor, M. B., Callender, G., & Druin, D. P. (1997). Prefrontal cortex
cognitive decits in children treated early and continuously for PKU. Monographs of the Society for Research in Child Development, 62, 1205.
Easterbrook, J. A. (1959). The eect of emotion on cue utilization and the organization of behavior. Psychological Review, 66, 183201.
Edelstein, R. S., Ghetti, S., Redlich, A. D., Cordn, I. M., & Raskauskas, J. L. (2003,
April). Do child sexual abuse victims forget their experiences? Poster presented at the
Biennial Society for Research in Child Development Conference, Tampa, FL.
Eisen, M. L., & Goodman, G. S. (1998). Trauma, memory, and suggestibility in children. Development and Psychopathology, 10, 717738.
Eisen, M. L., Goodman, G. S., Qin, J., & Davis, S. (1998). Memory and suggestibility
in maltreated children: New research relevant to evaluating allegations of
abuse. In S. Lynn & K. McConkey (Eds.), Truth in memory (pp. 163189). New
York: Guilford Press.
Eisen, M. L., & Lynn, S. J. (2001). Dissociation, memory, and suggestibility in adults
and children. Applied Cognitive Psychology, 15, S49S73.
Eisen, M., Qin, J. J., Goodman, G. S., & Davis, S. (1999). The impact of dissociation,
trauma, and stress arousal on the memory and suggestibility of physically and sexually abused, and neglected children. Washington, DC: National Center on Child
Abuse and Neglect.
Eisen, M. L., Qin, J., Goodman, G. S., & Davis, S. (2002). Memory and suggestibility

338

in maltreated children: Age, stress arousal, dissociation, and psychopathology.

Journal of Experimental Child Psychology, 83, 167212.
Elliott, D. M., & Briere, J. (1995). Posttraumatic stress associated with delayed recall of sexual abuse: A general population study. Journal of Traumatic Stress, 8,
629647.
Engstrom v. Engstrom, Second Appellate District, California, Case No. VC016157,
March, 1997.
Epstein, M. A., & Bottoms, B. L. (1998). Memories of childhood sexual abuse: A survey of young adults. Child Abuse and Neglect, 22, 12171238.
Evers-Szostak, M., & Sanders, S. (1992). The Childrens Perceptual Alteration Scale
(CPAS): A measure of childrens dissociation. Dissociation, 1, 9197.
Feeny, N. C., Zoellner, L. A., & Foa, E. B. (2000). Anger, dissociation, and posttraumatic stress disorder among female assault victims. Journal of Traumatic Stress,
13, 89100.
Finkelhor, D. (1984). Child sexual abuse: New theory and research. New York: Free
Press.
Fisher, R. P., & Cutler, B. L. (1992, September). The relation between consistency and
accuracy of witness testimony. Paper presented at the Third European Conference on Law and Psychology, Oxford, England.
Fivush, R. (1998). Childrens recollections of traumatic and nontraumatic events.
Development and Psychopathology, 10, 699716.
Fivush, R., Brotman, M., Buckner, J. P., & Goodman, S. H. (2000). Gender dierences in parent-child emotion narratives. Sex Roles, 42, 233253.
Fivush, R., Haden, C. A. & Reese, E. (1996). Remembering, recounting and reminiscing: The development of autobiographical memory in social context. In
D. Rubin (Ed.), Remembering our past: An overview of autobiographical memory
(pp. 377397). New York: Cambridge University Press.
Foa, E. B., Zinbarg, R., & Rothbaum, B. O. (1992). Uncontrollability and
unpredictability in post-traumatic stress disorder: An animal model. Psychological Bulletin, 112, 218238.
Freud, A. (1936). The ego and the mechanisms of defense (C. Baines, Trans.). New York:
International Universities Press.
Freud, S. (1915/1957). Repression. In J. Strachey (Ed.), The standard edition of the
complete psychological works of Sigmund Freud. Vol. 14. London: Hogarth.
Freyd, J. J. (1996). Betrayal trauma: The logic of forgetting childhood abuse. Cambridge,
MA: Harvard University Press.
Freyd, J. J., & Gleaves, D. H. (1996). Remembering words not presented in lists: Implications for the recovered/false memory controversy? Journal of Experimental Psychology: Learning, Memory, and Cognition, 22, 811813.
Freyd, J. J., Martorello, S. R., Alvarado, J. S., Hayes, A. E., & Christman, J. C. (1998).
Cognitive environments and dissociative tendencies: Performance on the Standard Stroop Task for high versus low dissociators. Applied Cognitive Psychology,
12, S91S103.
Galletly, C., Clark, C. R., McFarlane, A. C., & Weber, D. L. (2001). Working memory
in posttraumatic stress disorder: An event-related potential study. Journal of
Traumatic Stress, 14, 295309.

339

Gilbertson, M. W., Shenton, M. E., Ciszewski, A., Kasai, K., Lasko, N. B., Orr, S. P.,
et al. (2002). Smaller hippocampal volume predicts pathologic vulnerability
to psychological trauma. Nature Neuroscience, 5, 12421247.
Goodman, G. S., Ghetti, S., Quas, J. A., Edelstein, R. S., Alexander, K. W., Redlich,
A. D., et al. (2003). A prospective study of memory for child sexual abuse: New
ndings relevant to the repressed/lost memory controversy. Psychological Science, 14, 113118.
Goodman, G. S., Hirschman, J. E., Hepps, D., & Rudy, L. (1991). Childrens memory
for stressful events. Merrill-Palmer Quarterly, 37, 109157.
Goodman, G. S., Quas, J. A., Batterman-Faunce, J. M., Riddlesberger, M. M., & Kuhn,
J. (1994). Predictors of accurate and inaccurate memories of traumatic events
experienced in childhood. Consciousness and Cognition, 3, 269294.
Goodman, G. S., Quas, J. A., Batterman-Faunce, J. M., Riddlesberger, M. M., & Kuhn,
J. (1997). Childrens reactions to and memory for a stressful event: Inuences
of age, anatomical dolls, knowledge, and parental attachment. Applied Developmental Science, 1, 5475.
Gothard, S., & Ivker, N. A. C. (2000). The evolving law of alleged delayed memories
of childhood sexual abuse. Child Maltreatment, 5, 176189.
Gross, J. J. (2002). Emotion regulation: Aective, cognitive, and social consequences. Psychophysiology, 39, 281291.
Gurvits, T. V., Shenton, M. E., Hokama, H., Ohta, H., Lasko, N. B., Gilbertson,
M. W., et al. (1996). Magnetic resonance imaging study of hippocampal volume in chronic, combat-related posttraumatic stress disorder. Biological Psychiatry, 40, 10911099.
Hamann, S. (2001). Cognitive and neural mechanisms of emotional memory. Trends
in Cognitive Science, 5, 394400.
Hekkanen, S. T., & McEvoy, C. (2002). False memories and source-monitoring problems: Criterion dierences. Applied Cognitive Psychology, 16, 7385.
Henry, B., Mott, T. E., Caspi, A., Langley, J., & Silva, P. A. (1994). On the remembrance of things past: A longitudinal evaluation of the retrospective method.
Psychological Assessment, 6, 92101.
Heuer, F., & Reisberg, D. (1990). Vivid memories of emotional events: The accuracy of remembered minutiae. Memory & Cognition, 18, 496506.
Howe, M. L. (1997). Childrens memory for traumatic experiences. Learning and Individual Dierences, 9, 153174.
Howe, M. L. (2000). The fate of early memories: Developmental science and the retention of childhood experiences. Washington, DC: American Psychological Association.
Hull, A. M. (2002). Neuroimaging ndings in post-traumatic stress. British Journal
of Psychiatry, 181, 102110.
Hyman, I. E., & Billings, F. J. (1998). Individual dierences and the creation of false
childhood memories. Memory, 6, 120.
Hyman, I. E., & Loftus, E. F. (1998). Errors in autobiographical memory. Clinical
Psychology Review, 18, 933947.
Jenkins, M. A., Langlais, P. J., Delis, D., & Cohen, R. (1998). Learning and memory

340

in rape victims with posttraumatic stress disorder. American Journal of Psychiatry, 155, 278279.
Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitoring. Psychological Bulletin, 114, 328.
Johnson, M. K., Nolde, S. F., & De Leonardis, D. M. (1996). Emotional focus and
source-monitoring. Journal of Memory and Language, 35, 135156.
Jones, D., & Krugman, R. (1986). Can a 3-year-old bear witness to her own assault?
Child Abuse and Neglect, 10, 253258.
Kail, R. (2000). Speed of information processing: Developmental change and links
to intelligence. Journal of School Psychology, 38, 5161.
Kassin, S. M., Tubb, V. A., Hosch, H. M., & Memon, A. (2001). On the general acceptance of eyewitness testimony research. American Psychologist, 56, 405416.
Kirschbaum, C., Wolf, O. T., May, M., Wippich, W., & Hellhammer, D. H. (1996).
Stress- and treatment-induced elevations of cortisol levels associated with impaired declarative memory in healthy adults. Life Sciences, 58, 14751483.
Kohout v. Charter Peachford Hospital, Georgia Supreme Court, Case No. S98C1773,
September 1998.
Kramer, T. H., Buckhout, R., & Eugenio, P. (1990). Weapon focus, arousal, and eyewitness memory: Attention must be paid. Law and Human Behavior, 14, 167184.
Kuebli, J., Butler, S., & Fivush, R. (1995). Mother-child talk about past emotions:
Relations of maternal language and child gender over time. Cognition & Emotion, 9, 265283.
LaBar, K. S., & Phelps, E. A. (1998). Arousal-mediated memory consolidation: Role
of the medial temporal lobe in humans. Psychological Science, 9, 490493.
Lindsay, D. S., & Read, J. D. (1995). Memory work and recovered memories of
childhood sexual abuse: Scientic evidence and public, professional, and personal issues. Psychology, Public Policy, and the Law, 1, 846908.
Loftus, E. F., & Ketcham, K. (1991). Witness for the defense: The accused, the eyewitness and the expert who puts the memory on trial. New York: St. Martins Press.
Loftus, E. F., Loftus, G. R., & Messo, J. (1987). Some facts about weapon focus. Law
and Human Behavior, 11, 5562.
Lupien, S. J., & McEwen, B. S. (1997). The acute eects of corticosteroids on cognition: Integration of animal and human model studies. Brain Research Reviews,
24, 127.
Magarinos, A. M., McEwen, B. S., Flugge, G., & Fuchs, E. (1996). Chronic psychosocial stress causes apical dendritic atrophy of hippocampal CA3 pyramidal
neurons in subordinate tree shrews. Journal of Neuroscience, 16, 35343540.
Mazzoni, G. A., & Loftus, E. F. (1998). Dream interpretation can change beliefs about
the past. Psychotherapy: Theory, Research, Practice, Training, 35, 177187.
Mazzoni, G. A., Lombardo, P., Malvagia, S., & Loftus, E. F. (1999). Dream interpretation and false beliefs. Professional Psychology: Research and Practice, 30, 4550.
McEwen, B. S., & Magarinos, A. M. (1997). Stress eects on morphology and function
of the hippocampus. Annals of the New York Academy of Sciences, 821, 271284.
McNally, R. J. (1998). Experimental approaches to cognitive abnormality in posttraumatic stress disorder. Clinical Psychology Review, 18, 971982.

341

McNally, R. J., Amir, N., & Lipke, H. J. (1996). Subliminal processing of threat cues
in posttraumatic stress disorder? Journal of Anxiety Disorders, 10, 115128.
McNally, R. J., Clancy, S. A., Schacter, D. L., & Pittman, R. K. (2000). Cognitive
processing of trauma cues in adults reporting repressed, recovered, or continuous memories of childhood sexual abuse. Journal of Abnormal Psychology,
109, 355359.
McNally, R. J., Kaspi, S. P., Riemann, B. C., & Zeitlin, S. B. (1990). Selective processing of threat cues in posttraumatic stress disorder. Journal of Abnormal Psychology, 99, 398402.
McNally, R. J., Lasko, N. B., Macklin, M. L., & Pitman, R. K. (1995). Autobiographical memory disturbance in combat-related posttraumatic stress disorder. Behavior Research and Therapy, 33, 619630.
McNally, R. J., Litz, B. T., Prassas, A., Shin, L. M., & Weathers, F. W. (1994). Emotional priming of autobiographical memory in post-traumatic stress disorder.
Cognition & Emotion, 8, 351367.
McNally, R. J., Metzger, L. J., Lasko, N. B., Clancy, S. A., & Pitman, R. K. (1998).
Directed forgetting of trauma cues in adult survivors of childhood sexual abuse
with and without posttraumatic stress disorder. Journal of Abnormal Psychology, 107, 596601.
Merritt, K. A., Ornstein, P. A., & Spicker, B. (1994). Childrens memory for a salient
medical procedure: Implications for childrens testimony. Pediatrics, 94, 1723.
Metcalfe, J., & Jacobs, W. J. (2000). Hot emotions in human recollection: Toward a
model of traumatic memory. In E. Tulving (Ed.), Memory, consciousness, and the
brain: The Tallinn Conference (pp. 228242). Philadelphia: Taylor and Francis.
Miller, P. H., & Seier, W. L. (1994). Strategy utilization deciencies in children:
When, where, and why. In H. W. Reese (Ed.), Advances in child development and
behavior (Vol. 25, 107156), San Diego: Academic Press.
Miyake, A., & Shah, P. (1999). Models of working memory: Mechanisms of active maintenance and executive control. Cambridge: University Press.
Moradi, A. R., Taghavi, R., Neshat-Doost, H. T., Yule, W., & Dalgleish, R. (1999).
Performance of children and adolescents with PTSD on the Stroop colour-naming task. Psychological Medicine, 29, 415419.
Moradi, A. R., Taghavi, R., Neshat-Doost, H. T., Yule, W., & Dalgleish, T. (2000).
Memory bias for emotional information in children and adolescents with posttraumatic stress disorder: A preliminary study. Journal of Anxiety Disorders, 14,
521534.
Nagae, S., & Moscovitch, M. (2002). Cerebral hemispheric dierences in memory
of emotional and nonemotional words in normal individuals. Neuropsychologia,
40, 16011607.
Ochsner, K. N. (2000). Are aective events richly recollected or simply familiar? The
experience and process of recognizing feelings past. Journal of Experimental
Psychology: General, 129, 242261.
Orbach, Y., & Lamb, M. E. (1999). Assessing the accuracy of a childs account of
sexual abuse: A case study. Child Abuse and Neglect, 23, 9198.
Ornstein, P. A., Baker-Ward, L., Myers, J., Principe, G. F., & Gordon, B. N. (1995).
Young childrens long-term retention of medical experiences: Implications for

342

testimony. In F. E. Weinert & W. Schneider (Eds.), Memory performance and competencies: Issues in growth and development (pp. 349371), Hillsdale, NJ: Erlbaum.
Ornstein, P. A., Ceci, S. J., & Loftus, E. F. (1998). Adult recollections of childhood
abuse: Cognitive and developmental perspectives. Psychology, Public Policy, and
Law, 4, 10251051.
Otten, L. J., & Rugg, M. D. (2002). The birth of a memory. Trends in Neurosciences,
25, 279281.
Payne, J. D., Nadel, L., Allen, J. J. B., Thomas, K. G. G., & Jacobs, W. J. (2002). The
eects of experimentally induced stress on false recognition. Memory, 10, 16.
Peterson, C. (1999). Childrens memory for medical emergencies: Two years later.
Developmental Psychology, 35, 14931506.
Peterson, C., & Bell, M. (1996). Childrens memory for traumatic injury. Child Development, 67, 30453070.
Peterson, C., & Rideout, R. (1998). Memory for medical emergencies experienced
by 1 and 2year-olds. Developmental Psychology, 34, 10591072.
Peterson, C., & Whalen, N. (2001). Five years later: Childrens memory for medical
emergencies. Applied Cognitive Psychology, 15, 1724.
Pezdek, K., & Taylor, J. (2002). Memory for traumatic events in children and adults.
In M. L. Eisen, J. A. Quas, & G. S. Goodman (Eds.), Memory and suggestibility in
the forensic interview (pp. 165183). Mahwah, NJ: Erlbaum.
Pickel, K. L. (1999). The inuence of context on the weapon focus eect. Law and
Human Behavior, 23, 299311.
Pillemer, D. B. (1984). Flashbulb memories of the assassination attempt on President Reagan. Cognition, 16, 6380.
Poole, D. A., Lindsay, D. S., Memon, A., & Bull, R. (1995). Psychotherapy and the
recovery of memories of childhood sexual abuse: U.S. and British practitioners
beliefs, practices, and experiences. Journal of Consulting and Clinical Psychology,
63, 426437.
Powell, M. B., & Thomson, D. M. (1996). Childrens memory of an occurrence of a
repeated event: Eects of age, repetition, and retention interval across three
question types. Child Development, 67, 19882004.
Price, D. W., & Goodman, G. S. (1990). Visiting the wizard: Childrens memory for
a recurring event. Child Development, 61, 664680.
Putnam, F. W. (1996). Posttraumatic stress disorder in children and adolescents.
In L. Dickstein, M. Riba, & J. Oldham (Eds.), Review of psychiatry (Vol. 15,
pp. 447467). Washington, DC: American Psychiatric Press.
Putnam, F. W. (1997). Dissociation in childhood adolescents: A developmental perspective. New York: Guilford Press.
Putnam, F. W., Helmers, K., & Trickett, P. K. (1993). Development, reliability, and
validity of a child dissociation scale. Child Abuse and Neglect, 17, 731741.
Putnam, F. W., & Trickett, P. K. (1997). The psychobiological eects of sexual
abuse: A longitudinal study. Annals of the New York Academy of Sciences, 821,
150159.
Pynoos, R. S., & Nader, K. (1989). Childrens memory and proximity to violence.
Journal of the American Academy of Child and Adolescent Psychiatry, 28, 236
241.

343

Qin, J. (1999). Adults memories of childhood: True versus false reports. Unpublished
doctoral dissertation, University of California, Davis.
Quas, J. A., Goodman, G. S., Bidrose, S., Pipe, M.-E., Craw, S., & Ablin, D. (1999).
Emotion and memory: Childrens long-term remembering, forgetting, and suggestibility. Journal of Experimental Child Psychology, 72, 235270.
Quas, J. A., Goodman, G. S., Redlich, A. D., Ghetti, S., Cordn, I. M., & Jones,
D. P. H. (2001, April). Long-term memory for testifying in criminal court as children. Paper presented at the Biennial Meeting Society for Research in Child Development, Minneapolis, MN.
Quirarte, G. L., Roozendaal, B., & McGaugh, J. L. (1997). Glucocorticoid enhancement of memory storage involves noradrenergic activation in the basolateral
amygdala. Proceedings of the National Academy of Sciences of the United States of
America, 94, 1404814053.
Read, J. D. (1996). From passing thoughts to false memories in 2 minutes: Confusing real and illusory events. Psychonomic Bulletin & Review, 3, 105111.
Read, J. D. (1999). The recovered/false memory debate: Three steps forward, two
steps back? Expert Evidence, 7, 124.
Read, J. D., & Lindsay, D. S. (2000). Amnesia for summer camps and high school
graduation: Memory work increases reports of prior periods of remembering
less. Journal of Traumatic Stress, 13, 129147.
Roediger, H. L. III, & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 803814.
Roediger, H. L. III, & McDermott, K. B. (1999). False alarms and false memories.
Psychological Review, 106, 406410.
Roediger, H. L. III, McDermott, K. B., & Robinson, K. J. (1998). The role of associative processes in creating false memories. In M. A. Conway, S. E. Gathercole, &
C. Cornoldi (Eds.), Theories of memory (Vol. 2, pp. 187245). Sussex, UK: Psychology Press.
Safer, M. A., Christianson, S., Autry, M. W., & Osterlund, K. (1998). Tunnel memory
for traumatic events. Applied Cognitive Psychology, 12, 99117.
Schooler, J. W., Ambadar, Z., & Bendiksen, M. (1997). A cognitive corroborative
case study approach for investigating discovered memories of sexual abuse. In
J. D. Read & D. S. Lindsay (Eds.), Recollections of trauma: Scientic evidence and
clinical practice (pp. 379387). New York: Plenum Press.
Schooler, J. W., & Eich, E. (2000). Memory for emotional events. In E. Tulving &
F. I. M. Craik (Eds.), The Oxford handbook of memory (pp. 379392). New York:
Oxford University Press.
Shaw, J. I., & Skolnick, P. (1999). Weapon focus and gender dierences in eyewitness accuracy: Arousal versus salience. Journal of Applied Social Psychology, 29,
23282341.
Shobe, K. K., & Kihlstrom, J. F. (1997). Is traumatic memory special? Current Directions in Psychological Science, 6, 7074.
Shrimpton, S., Oates, K., & Hayes, S. (1998). Childrens memory of events: Eects
of stress, age, time delay, and location of interview. Applied Cognitive Psychology, 12, 133143.

344

Southwick, S. M., Morgan, C. A., Nicolaou, A. L., & Charney, D. S. (1997). Consistency of memory for combat-related traumatic events in veterans of Operation
Desert Storm. American Journal of Psychiatry, 154, 173177.
Stanny, C. J., & Johnson, T. C. (2000). Eects of stress induced by a simulated shooting on recall by police and citizen witnesses. American Journal of Psychology,
113, 359386.
Steblay, N. M. (1992). A meta-analytic review of the weapon focus eect. Law and
Human Behavior, 16, 413424.
Stein, M. B., Koverola, C., Hanna, C., Torchia, M. G., & McClarty, B. (1997). Hippocampal volume in women victimized by childhood sexual abuse. Psychological
Medicine, 27, 951959.
Stein, N. L., & Liwag, M. D. (1997). Childrens understanding, evaluation, and
memory for emotional events. In P. W. van den Broek & P. J. Bauer (Eds.), Developmental spans in event comprehension and representation: Bridging ctional and
actual events (pp. 199235), Hillsdale, NJ: Erlbaum.
Terr, L. C. (1988). What happens to early memories of trauma? A study of twenty
children under age ve at the time of the documented traumatic events. American Journal of Psychiatry, 148, 1020.
Terr, L. C. (1994). Unchained memories. New York: Basic Books.
Tollestrup, P. A., Turtle, J. W., & Yuille, J. C. (1994). Actual victims and witnesses
to robbery and fraud: An archival analysis. In D. F. Ross, J. D. Read, & M. P.
Toglia (Eds.), Adult eyewitness testimony: Current trends and developments. New
York: Press Syndicate of the University of Cambridge.
Tulving, E. (1985). Memory and consciousness. Canadian Psychologist, 26, 112.
Uddo, M., Vasterling, J. J., Brailey, K., & Sutker, P. B. (1993). Memory and attention in combat-related post-traumatic stress disorder. Journal of Psychopathology and Behavioral Assessment, 15, 4352.
Uno, H., Tarara, R., Else, J. G., Suleman, M. A., & Sapolsky, R. M. (1989). Hippocampal damage associated with prolonged and fatal stress in primates. Journal
of Neuroscience, 9, 17051711.
van Stegeren, A. H., Everaerd, W., Cahill, L., McGaugh, J. L., & Gooren, L. J. (1998).
Memory for emotional events: Dierential eects of centrally versus peripherally acting beta-blocking agents. Psychopharmacology, 138, 305310.
Vandermaas, M. O., Hess, T. M., & Baker-Ward, L. (1993). Does anxiety aect
childrens reports of memory for a stressful event? Applied Cognitive Psychology,
7, 109127.
Villarreal, G., Hamilton, D. A., Petropoulos, H., Driscoll, I., Rowland, L. M., Griego,
J. A., et al. (2002). Reduced hippocampal volume and total white matter volume in posttraumatic stress disorder. Biological Psychiatry, 52, 119125.
Vrana, S. R., Roodman, A., & Beckham, J. C. (1995). Selective processing of traumarelevant words in posttraumatic stress disorder. Journal of Anxiety Disorders, 9,
515530.
Wagner, A. D., Schacter, D. L., Rotte, M., Koutstaal, W., Maril, A., Dale, A. M., et al.
(1998). Building memories: Remembering and forgetting of verbal experiences
as predicted by brain activity. Science, 281, 11881191.
Wells, G. L., Small, M., Penrod, S., Malpass, R. S., Fulero, S. M., & Brimacombe,

345

C. A. E. (1998). Eyewitness identication procedures: Recommendations for

lineups and photospreads. Law and Human Behavior, 22, 603648.
Wessel, I., & Merckelbach, H. (1994). Characteristics of traumatic memories in
normal subjects. Behavioral and Cognitive Psychotherapy, 22, 315324.
Wessel, I., & Merckelbach, H. (1997). The impact of anxiety on memory for details
in spider phobics. Applied Cognitive Psychology, 11, 223232.
Wessel, I., & Merckelbach, H. (1998). Memory for threat-relevant and threatirrelevant cues in spider phobics. Cognition & Emotion, 12, 93104.
Wessel, I., van der Kooy, P., & Merckelbach, H. (2000). Dierential recall of central
and peripheral details of emotional slides is not a stable phenomenon. Memory,
8, 95109.
Widom, C. S., & Morris, S. (1997). Accuracy of adult recollections of childhood victimization, Part 2: Childhood sexual abuse. Psychological Assessment, 9, 3446.
Williams, L. M. (1994). Recall of childhood trauma: A prospective study of womens
memories of child sexual abuse. Journal of Consulting and Clinical Psychology,
62, 11671176.
Williams, L. M. (1995). Recovered memories of abuse in women with documented
sexual victimization histories. Journal of Traumatic Stress, 8, 649673.
Williams, L. M., & Banyard, V. L. (1999). Trauma and memory. Thousand Oaks, CA:
Sage.
Williams, M., Mathews, A., & MacLeod, C. (1996). The emotional Stroop task and
psychopathology. Psychological Bulletin, 120, 324.
Windmann, S., & Krger, T. (1998). Subconscious detection of threat as reected
by an enhanced response bias. Consciousness and Cognition, 7, 603633.
Windmann, S., & Kutas, M. (2001). Electrophysiological correlates of emotioninduced recognition bias. Journal of Cognitive Neuroscience, 13, 577592.
Windmann, S. Sakhavat, Z., & Kutas, M. (2002). Electrophysical evidence reveals
aective evaluation decits early in stimulus processing in patients with panic
disorder. Journal of Abnormal Psychology, 111, 357369.
Winograd, E., Peluso, J. P., & Glover, T. A. (1998). Individual dierences in susceptibility to memory illusions. Applied Cognitive Psychology, 12, S5S27.
Woodall, J. (1999). The nature of memory: Controversies about retrieved memories and the law of evidence. Journal of Psychiatry and Law, 26, 151218.
Yerkes, R., & Dodson, J. (1908). The relation of strength of stimulus to rapidity of
habit-formation. Journal of Comparative Neurology of Psychology, 18, 459482.
Yuille, J. C., & Cutshall, J. L. (1986). A case study of eyewitness memory of a crime.
Journal of Applied Psychology, 71, 291301.
Yuille, J. C., Davies, G., Gibling, F., Marxsen, D., & Porter, S. (1994). Eyewitness
memory of police trainees for realistic role plays. Journal of Applied Psychology,
79, 931936.
Zoellner, L. A., Foa, E. B., Brigidi, B. D., & Przeworski, A. (2000). Are trauma victims more susceptible to false memories? Journal of Abnormal Psychology, 109,
517524.

346

11

A Qualitative Study of Oral Testimony
.

oused in archives throughout the world are videotaped oral testimonies documenting the memories of individuals who have lived
through prolonged trauma. The testimonies in these archives consist of unconstrained recall of personal experiences in the context of larger historical events.
The most extensive of these collections have been provided by survivors of the
Holocaust, though sizeable collections in this country are from veterans of World
War II and the Vietnam War.1 The testimonies in these archives support research
in oral history, but they also constitute a valuable resource for cognitive researchers studying memory for traumatic events.
My research took place at the Fortuno Video Archive for Holocaust testimonies at Yale University, a collection of more than 4,000 testimonies from survivors of the Holocaust. Each testimony typically consists of a single Holocaust
survivor spontaneously recalling events in the presence of two nonintrusive
interviewers. Though there are no time limits, the length of most testimonies is
between 40 minutes and 3 hours, with an average length of 2 hours. Survivors
speak from memory, and, in many cases, their testimony is their rst extended
recall for more than 40 years. In the testimonies I studied, survivors came from
12 dierent European countries, varying in economic background, level of education, religiousness, language, and culture. They ranged in age from infancy
to 43 when World War II started and from 40 to 95 when giving testimony.
I focus in this chapter on emotional memory in survivors of prolonged trauma.
To describe and explain the characteristics of emotional memory, I called on my
analysis of more than 200 hours of oral testimony given by Holocaust survivors.
My general strategy was to study and categorize the phenomenal experience of
the survivors: their memories of emotional events, their emotional experiences,
and their observations about memory and emotion.
347

A Summary of the Methods

My methodology is rooted in the inductive methods of ethology and population
biology (Hubbell, 1999; Linton, 1986). More recently, I have drawn on techniques elaborated over the past decade for the qualitative study of complex psychological phenomena in natural settings (Tolman & Brydon-Miller, 2001).
In analyzing the testimonies, I used a repeated study procedure, viewing each
testimony in its entirety twice. During the Wrst viewing, I attended to the general content and organization of the memories and to the observations of the
survivor. I then re-viewed the testimony, annotating passages and describing the
remembered episodes. I selectively viewed the testimony up to 10 more times,
focusing on the annotated passages, transcribing the words, and detailing the
contextual and extra-linguistic features.
After viewing the videotaped testimonies, I categorized the content and structure of the transcribed testimony. I parsed the transcribed portions into passages
that were comparable to paragraphs in a written text and labeled each passage
with the categorical terms that characterized the passage. I then tabulated each
category. As the analysis proceeded, the number of categories grew and I subdivided the categories into more precisely dened and more detailed units. Through
this inductive analysis, the particulars from the individual testimonies coalesced
into general patterns of traumatic memory and the relationships between traumatic memory and strong emotion (Kraft, 2002).

Concepts of Accuracy
My approach to studying traumatic memory assumes that the witnesses who
provided the oral testimony tried to communicate their memories honestly. In
fact, the data collection procedures at the Yale Archive are especially conducive
to unconstrained and detailed testimony. This chapter does not address issues
of accuracy at length because my focus is on the phenomenal experience of traumatic memory and not on the correspondence between historical events and
memory for those events. Nonetheless, oral testimony can contribute to theoretical and practical discussions of accuracy and can support assessments of accuracy in long-term traumatic memory. Thus, it is important to introduce and
delineate the dierent methods that can be employed with oral testimony to assess the accuracy of memory.
Over the past 30 years, myriad facts about the Holocaust have been extensively described and meticulously documented. By comparing the information
in the testimonies to these documented facts, one can assess the historical truth
of information in oral testimony (Spence, 1982). As an outgrowth of this assessment, memory researchers could also categorize the types of information reliably remembered in testimony. In such an assessment, one must separate
documented instances in which the initial encoding of events was awed.2

348

In a seminal study, Wagenaar and Groeneweg (1990) conducted an assessment of historical accuracy in testimony provided by survivors from Camp Erika
in the Netherlands. Some of these survivors were interviewed twice, once in the
mid-1940s and again in the mid-1980s; other survivors were interviewed just
once in the mid-1980s, primarily in the context of trials of camp guards.
Wagenaar and Groeneweg compared the recall of inmates interviewed once with
known facts, and they compared the two sets of testimony provided by the survivors interviewed twice. They found a high level of remembering across the
40-year interval between events and interviews. The rst sentence of the Results
section makes this nding clear: The most striking aspect of the testimonies is
that witnesses agreed about the basic facts, which is demonstrated by a comparison of 55 longer interviews (p. 80). After noting the remarkable degree of remembering, the authors then identify and tabulate specic kinds of errors
witnesses made (p. 80).
In fact, sometimes oral testimony can identify existing errors in historical
documents. Homan and Homan (1994) described how oral testimony corrected an inaccurately captioned photograph on display at Yad Vashem, the
ocial Holocaust museum in Israel. The photograph was labeled Gardelegen,
a concentration camp in which 150 inmates were killed. The oral testimony of
the second author proved that the photograph was not of Gardelegen. Rather, it
depicted a barn near Hannover into which Nazis herded prisoners, then burned
them to death.
A variation of the method used in Wagenaar and Groeneweg (1990) is to
compare selected testimonies to other, corroborating testimonies. In my study
of child survivors, I assessed the accuracy of Holocaust testimony by comparing
the memories of the child survivors to those of older witnesses, in the few cases
that permitted it. As with other memory-based methods of verifying accuracy,
this method assumes the accuracy of the corroborating (or refuting) memories
(Kraft, 1996, 2002).
Categorizing the metamemory statements of witnesses constitutes another
assessment method. While providing testimony, witnesses make direct statements about what they can and cannot remember about particular events. They
also use probabilistic terms and qualifying statements, which indicate uncertainty (Kraft, 2000, 2002). These metamemory statements and the probabilistic terms can then be used to identify the types of information that are clear and
accessible to the witnesses and the types that are unclear and inaccessible. In
fact, Winograd (1994) has concluded that rather than debating the general question of accuracy, memory researchers should focus on identifying the conditions
conducive to accuracy and those likely to lead to distortion.
A third assessment of accuracy involves examining the consistency and inconsistency of information in the testimonies. Though this approach does not assess
accuracy per se, consistency in the testimonies is a suggestive indicator of accuracy. When Winograd (1992) considered the relationship between accuracy and

349

consistency in ashbulb memories, he concluded, If people are inconsistent in

reporting the same memory on two occasions, it is dicult to argue for accuracy
(p. 1). This method arises from the obverse of Winograds conclusionadmittedly a more tentative one: if people consistently report the same memory on two
occasions, it is dicult to argue for inaccuracy. In my study of oral testimony, I
assessed consistency by comparing the information in dierent testimonies provided by the same witnesswith the dierent testimonies separated by 8 years
in most cases. In a variation of this approach, I also examined consistencies and
inconsistencies within each testimony (Kraft, 2002).
One diculty with naturalistic studies of accuracy is the problem of deciding
a priori what someone else will or should remember. In the laboratory, of course,
such decisions can be made on the basis of the specic materials to be remembered. With traumatic events in the world, social or forensic considerations may
determine what information is helpful to remember, but failure to recall such
information only indicates the incompleteness of memory, not inaccuracy. In
general, what people do remember from traumatic events, they remember vividly, consistently, and permanently.
For an introduction to experimentation and theory on accuracy and emotional
memory, refer to Aect and Accuracy in Recall (Winograd & Neisser, 1992), Memory
Distortions and Their Prevention (Intons-Peterson & Best, 1998), and, more recently,
a comprehensive investigation of research on eyewitness testimony conducted by
Wells et al. (2000), and the 2000 edition of The Annual Review of Psychology, which
contains a concise history of research approaches in experimental psychology that
pertain to accuracy and distortion in memory, including a section on autobiographical memory (Koriat, Goldsmith, & Pansky, 2000).

Excerpts From the Testimonies

My plan in this chapter is to describe, exemplify, and explain the characteristics of
emotion and memory that emerged from qualitative analysis of more than 130
dierent testimonies of Holocaust survivors. To provide a sense of the inductive
process in this study, I begin by presenting a set of excerpts from these testimonies.
For some reason . . . I could not get any tears. . . . Why werent we screaming maniacs at that point, those of us who saw it daily? (Testimony of Irene
W., 1982)
There is one thing I have to say: that throughout my experience then, I
dont remember feeling fear. . . . What I remember feeling is numbness. (Testimony of Meir V., 1992)
I didnt feel anything. I didnt even feel the elation that I thought I was
going to feel. . . . It really didnt make any dierence. So we are liberated, so
what? (Testimony of Daniel F., 1980)

350

You was indierent to everything. You was like a vegetable. . . . I didnt

care. . . . I didnt care if they would kill me in a minute. (Testimony of Zoltan
G., 1979)
Only very highly charged, painful memories are with me. Specic memories. (Testimony of Clemens L., 1990)
To this day, I cannot really hear German. I have not changed about that
at all. When I hear the language . . . it does something to my insides. They
turn over. (Testimony of Eva B., 1979)
I dont want to think about it because I cant bear to think about it. (Testimony of Zezette L., 1980)
I am going over it if I am speaking about it or not. There is not really one
day of my life or one night that I dont think about it. (Testimony of Nadia R.,
1995)
I think the problem is . . . Im afraid if I open it up, Im going to have nightmares that I had for years and years, and I will not allow this, . . . Im afraid it
might destroy me. (Testimony of Martin S., 1988)
It is dicult to talk about feelings. . . . Now I feel more horrible than I felt
at the time. (Testimony of Alex H., 1983)
The tortures days and nights. Its something that we have. Its in our mind.
You cant forget. . . . I dont want to live with that pain. But its there. Its
there. It forms its own entity. And it surfaces whenever it wants to. I go on a
train and I will cry. I will read something and Ill be right back there where I
came from. And I cant erase it. Im not asking for it. It comes by itself. (Jacob
K. in the Testimony of Bessie and Jacob K., 1983)
Flashes come through about . . . my experiences. I still have dreams. (Testimony of Abraham P., 1984)
The screams when they took away the babies, I can never forget this. This
is with me constantly. Walking and taking away babies from a mothers
hands. (Testimony of Nina S., 1996)
Its still fty years since it was the end of the war, and many things, because of such time elapsed, are somewhere in the back of my mind. Now, I
could never forget the things that happened there, but they are somewhere
in the back. And suddenly everything comes to the front. (Testimony of Henry
S., 1996)
How can you tell a child? How can I tell my child? My young son, to tell
him that another human being burned his grandparents. How can you tell
such horror? (Testimony of Violet S., 1991)
What we learned was rage. . . . I dont know wherehow all that takes
shape. What surprises me is that were able to contain it. (Testimony of Robert K., 1984)

As I studied the relationships between traumatic memory and strong emotion, broad categories emerged that became the organizing topics of this chapter: (1) general characteristics of traumatic memory that inform the study of
emotion and memory, (2) recall of emotions that occurred at the time of the

351

events, (3) the re-experiencing of emotion in the present, (4) emotions that endure over the course of peoples lives, and (5) the palliative eects of distraction.

Characteristics of

Traumatic Memory

Analysis of oral testimony reveals two general characteristics of traumatic

memory pertinent to the study of emotion and memory: episodic structure and
duality of representation. With oral testimony, the primary unit of remembering is the episode: a discrete event with a start and a nish, a small story. The
episode is usually a narrative, although it can be impressionistic, and the episode may be placed within a precisely described spatial layout. The basic elements
of the episode are vivid perceptual images, deeply felt emotions, and physiological
experiences (Kraft, 1998a). These basic elements are then used to construct the
episodic testimony. Martin S. precisely summarizes the structure of his thinking: I just know episodes. Specics. That comes back to mind repeatedly (Testimony of Martin S., 1988).
Typically, survivors recall individual episodes one at a time, proceeding from
one episode to the next. During testimony, these encapsulated episodes are organized in sequences, like beads on a string, with dierent sequences of episodes
connected to one another by thin strands that are either chronological and geographic or thematic or rhetorical (Kraft, 2002).
The elements of episodic memory and the episodes themselves constitute two
separate levels of memorial representation. At one level, memory represents the
original phenomenal experience in the form of images, emotions, and bodily sensations. These original memories are then integrated into a narrative, episodic
construction of events, creating memorys second level. In their seminal article,
Brown and Kulik (1977) posited a ashbulb memory for highly consequential events that is xed for a very long time, and conceivably permanently, varying in complexity with consequentiality but, once created, always there, and in
need of no further strengthening. . . . The memory is not narrative and not even
in verbal form, but represented in other, perhaps imaginal, ways (p. 85). The
term ashbulb memory quickly migrated into the lexicon of cognitive psychology, and the underlying concept became the subject of considerable debate and
research, especially about accuracy, durability, and status as a distinctive
memory system. Arising from this debate a decade later was the elaboration of a
fundamentally dierent, nonnarrative, and imaginal level of memory.
Dierent pairs of terms introduced in the literature on memory reect the
two distinct levels of memory suggested by Brown and Kulik (1977). Flashbulb
memory corresponds to deep memory, to be distinguished from common
memory (Langer, 1991, pp. 56). It is situationally accessible memory as
distinguished from verbally accessible memory (Brewin, Dalgleish, & Joseph,

352

1996, p. 676). Similarly, Nadel (1994) proposed two separate memory systems
on the basis of task-driven, computational dierences and rates of learning: (1)
a system with rapid acquisition of information about specic episodes that preserves spatial information as a primary way of distinguishing one episode from
another and (2) a slower system that learns incrementally and generalizes
across dierent instances. According to Nadel, the episodic system represents
specic episodes sparsely, so that they do not interfere with one another, and
provides a rapidly forming template (p. 57) to serve as a basis for gradually
changing the representations in the longer-lasting general system.
To formalize the distinction in the two types of memory, Reyna and her colleagues introduced fuzzy-trace theory, which posits verbatim memory and gist
memory (Reyna, 1998; Reyna & Titcomb, 1997). Verbatim memory is the original representation of the events; gist memory structures and summarizes verbatim memory and is in linguistic form. Distortions, errors, and intrusions do
occur, but they occur within gist memory, not within verbatim memory. Verbatim memory endures, resistant to distortion and assimilation (Reyna &
Titcomb, 1997).
I refer to the two levels of memory as core memory and narrative memory
(Kraft, 2002). Core memory is the representation of the original phenomenal
experience in the form of perceptual, emotional, and physiological experience:
visual images, sounds, smells, tastes, emotions, and bodily sensations. I chose
core because the word specically refers to the most basic and activating
elements in a structure, whereas verbatim applies specically to language
and does not describe the powerful images, emotions, and physiological experiences represented in this basic level of memory. Core memory is both explicit
and implicit, consisting of representations of imagery and emotion that can be
consciouslythough not deliberatelyexperienced during extended recall, as
well as implicit representations of emotional, physiological, and bodily responses that can be triggered by external cues.3
Narrative memory is constructed from the images in core memory, shaped
in accordance with narrative conventions and conveyed primarily in language.
The word narrative characterizes how people structure episodes in personal
memory for the purpose of thinking about oneself in silent remembering and
communicating with others to tell the events of the past (Sarbin, 1986). The
following excerpt illustrates how narrative memory draws on imagery in core
memory and how testimony can be analyzed to learn about the underlying
sources of information. In testimony, Renee G. describes an incident that occurred while she was a teenager surviving as a slave laborer in a small ghetto:
At one point, I bent down because I found a picture of my family, so I must
have stopped to look at it and all of a sudden I heard a big whip over my legs
and I just remember the sting of it. And I just froze for a minute and got up
and stood straight. . . . I was probably by that time, completely bereft of emo-

353

tion . . . sort of half-dead, not here and not there. . . . And I just stared at the
German into the face. I remember big, red face. I would probably recognize him
today. (Testimony of Renee G., 1980)

The excerpt suggests contributions from core memory as well as narrative

constructions. Based on images provided by core memory, Renee G. clearly remembers the sting of the whip and the German guards big, red face. The
statements that she must have stopped to look at the photograph and that she
was probably bereft of emotion indicate that she is constructing a reason for
stopping her work and for freezing after being whipped. The reason for the uncharacteristic pause in her work appears to be constructed from general knowledge of her actions at the time and not based in core memory. Renees assessment
that she was bereft of emotion also appears to be based on a schematized representation of her general emotional state while subsisting in the ghetto (Testimony of Renee G., 1980).
Those who give oral testimony about the Holocaust usually call on narrative
memory to tell of the horrible events, and the testimony is structured and coherent. Sometimes, though, when describing a scene, the survivor may be drawn
into core memory, losing contact with narrative memory. While providing testimony, these survivors appear immersed in visualizing the events, where the
outside world dissolves and the survivors re-experience the events of the past.
When a person is fully visualizing, or back there, as survivors often say, the
images are conveyed in raw form, unstructured, and testimony can seem incoherent to the listener. In terms of subjective experience, back there refers to a level
of consciousness similar to hypnagogic sleep, with the images of memory possessing the erce vividness of dreams. As the observer of a movie experiences
being in the scene while simultaneously maintaining a disconnected awareness
of the surrounding theater, the survivor experiences a loss of self within the
immediate environment and an immersion in the past, while only vaguely aware
of the present.
After an hour of testimony, Jolly Z. says, Im sure we are guided by inner
needs to see or to deny things around us. Um, but, um. [She shakes her head and
looks down for ve seconds.] I lost my thought. Im back there again. [She gestures backwards with her nger.] . . . Im just back there. Jolly is asked what
she sees when she is back there, and she replies: Mud. And just grey, and mud.
[She shakes her head.] Mmm, bodies. [She sighs.] Oh, I know what I was trying
to say. During this testimony, Jolly loses the present, momentarily slipping into
the horrors of the remembered past.
Jolly describes the experience of being back there: I feel myself to be there. I
see the mud around me. I smell it. Smell is very important. I smell it. I see the
bodies, dead and alive. Im there. I see all the details. Im there. Im very visual.
Im there. I see the sun or the rain. I feel the wet clothes. Im there. Jolly explains that she is seventeen or eighteen years old when she submerges in this

354

memory, a dierent self: Im not here. . . . I dont even know about myself now.
Im there. . . . Somebody else talks out of me. . . . You see its not me. Its that
person who experienced it who is talking about those experiences. And maybe
thats what I am referring to. Because I am there. It is that part of me. Not now
(Testimony of Jolly Z., 1988).
Hanna F. accesses core memory while describing an episode at Auschwitz.
When inmates died during the night, they were taken out of the barracks and
dumped onto a pile of bodies. One night, Hanna had to go to the latrine, which
meant walking past this pile of corpses. She says, And the rats were standing
and eating the peoples faces. Eating. They were having a. [She stops and does
not nish her sentence. After 13 seconds of silence, she abruptly begins again.]
Anyway, I had to do my job. I was just looking [at] whats happening to a
human being (Testimony of Hanna F., 1987). During the silence in her testimony, Hanna appears to be immersed in the images of core memory, reliving
the events of the past. Other survivors experience similar immersion into the
altered reality of atrocity, and they refer to this experiential change in their
testimony. Myra L. interrupts her own testimony to comment: Im talking a little
incoherent because these things, these things. Im coming back to the transports
from the ghetto (Testimony of Myra L., 1984).
In comparing laboratory research on trauma with naturalistic studies of
trauma outside the laboratory, one fact is clear: simulations of traumatic events
fortunately lack many of the essential characteristics of trauma: the suddenness,
the overwhelming emotionincluding shock, helplessness, and terror, the violation of ones body and psyche, physical pain, confusion, betrayal, and humiliation, a ood of information that has not been perceived before and that is not
attended to eectively, an inability of existing schemas to predict outcomes, a
ight response that cannot be enacted, a hyperaroused state of mind, and a complex of neurophysiological responses. Terr (1994) wrote succinctly, You cant
replicate trauma in an experimental lab. You cant simulate murders without
terrorizing your research subjects (p. 52).
In terms of subjective experience, the initial exposure to trauma leaves people
bereft of understanding, unable to apply prior learning, and unable to comprehend. Contact between ongoing events and existing knowledge breaks apart,
resulting in unguided perception of the events. Those who are victimized see an
unbelievable reality of horror in eeting moments of painful clarity, ultimately
creating memories that are sensory and relivable, in the form of smells, sounds,
scenes, emotions, and physiological sensations (Kraft, 1998b).
One emerging consensus from the vast clinical literature on PTSD is that severe trauma leads to memories qualitatively dierent from memories for events
that do not threaten ones health and well-being. Memories from severe trauma
are distinctively fragmented, intrusive, and durable and can be accompanied by
recurrent nightmares and ashbacks that involve vivid sensory experiences and
high physiological arousal. Moreover, some aspects of traumatic memories ap-

355

pear to remain distinct from other memories and to be constant over time (Brewin
et al., 1996; Kraft, 2002; Terr, 1994; van der Kolk, 1996). In terms of Nadels
(1994) two-system framework, the fundamental dierence in episodic memory
between the formation of traumatic memories and the formation of nontraumatic
memories is that traumatic events are not represented sparsely. They are learned
quickly and represented in persistent detail, creating a disruption in the process
of assimilating specic memories into more general representations.

Recall of Emotional Experiences

For Holocaust survivors, the most frequently recalled emotional experiences are
fear and numbness. The survivors remember experiencing extreme emotion
or the absence of emotion. Both meanings of the word petrify accurately convey the victims remembered emotional state during severe trauma: to paralyze
with terror and to convert organic matter into a stony replica.

Fear
Survivors clearly recall their fear within specic episodes. Blanche H. remembers being sent into hiding after her older sister was taken away by the Nazis.
Accompanied by an older woman, Blanche took a train to her uncles house, and
she and the woman were the only two civilian passengers on the train. The rest
were uniformed German soldiers. Blanche says, That was about a three-hour
train ride, and I am thinking . . . I was going to die. I shook like a leaf and the
mere fact that I was shaking I felt would give myself away to them (Testimony
of Blanche H., 1992). Similarly, Esther W. describes standing appell at Skarzysko,
lining up and being counted for hours every morning, while German guards
berated them: The fear. The fear. You could have died from the fear alone (Testimony of Esther W., 1994).
Performing slave labor as a young boy, Martin S. proved his worth by operating four machines by himself, tooling bullets. He excelled at his labor, but he
felt guilty because others were doing what they could to sabotage the German
war eort. So Martin began making defective bullets, cutting them so they would
jam in the ries. One day, after making boxes and boxes of defective bullets,
Martin noticed the camp inspector approaching, so he quickly adjusted the
machines, turning out several dozen good bullets. By chance, when the inspector examined Martins work, he chose one of the properly engineered bullets.
Martin describes how he felt: Oh, the fear was unbelievable. I thought I had had
it. He continues: We all went through many terrifying moments (Testimony
of Martin S., 1986).
Sabina S. tells of her childhood hiding from the Nazis. And we heard above
us the Germans yelling, searching for us and looking for us. . . . You couldnt

356

breathe. You couldnt talk. You were always afraid they were going to discover
you. So there was always that fear that theyre there to kill you (Testimony
of Sabina S., 1985). Kluger (2001) provides a metaphoric description of her
emotions while waiting in a special barracks at Auschwitz, either to be gassed
or sent to a labor camp: Fear infected me like a poisonous illness. My mind
was like a theater in which a re alarm had gone o and panic has broken out
(p. 111).

Numbness
In recalling her arrival at Auschwitz, Jolly Z. provides a glimpse of the unthinking and unfeeling response when people confront unprecedented horrors: When
these things actually happen, you dont think, you just act. Your feelings, the
reactions, come later. When the selection took place, you probably just think,
which way did he say to go, so Ill go right or left. There are no emotions. There
is no observing reaction. Theres just action. The rest comes later (Testimony
Jolly Z., 1988).
In the testimonies I studied, more than 75% of the survivors who describe their
emotional state during the horrors say that they were numb (Kraft, 1998b).
They use a variety of phrases to elaborate this state of numbness: in a trance,
like a piece of wood, frosted over, frozen, like a stone, hibernating, like
a vegetable, like robots, in a catatonic state.
Eva L. describes her initial memory of leaving the freight car at Auschwitz and
being immediately separated from her mother:
We couldnt believe that everything was happening so fast. And then, we just
didnt feel anything, didnt feel anything. Was just like in a trance. I just didnt
feel anything. They pushed me into a bath, they stripped me, they shaved my
hair. They tookstripped me of my clothes. They just gave me a little rag, just
a little blouse that covered my front. The back was ripped. The rest was naked.
And I was just pushed, like in a trance, I didnt care (Testimony of Eva L., 1982).

The numbness is so alien and so pervasive that some survivors say they were
given drugs. Nina S. describes the aftermath of her arrival at Auschwitz, after
losing her mother during the initial selection. She says, They gave us tranquilizers because the next day nobody was crying. She explains: Suddenly we became all calm at the same time. They gave us some soup. And everybody said
there was something in it because we became like dierent people. The next day
nobody missed their parents. Nobody was crying (Testimony of Nina S., 1996).
Chana F. describes the quick steps to chronic numbness:
If you ask people about feelings, I ask myself in the beginning the same thing.
The rst ordinance was tragic: we cried, we pulled our hair, whats going to
be. The second was bad. The third you take already, and then it just comes to

357

you. . . . We were numb. We were already numb. We were numb the ve

years. There were no feelings (Testimony of Chana F. and Henry F., 1980).

The persistent numbness caused by constant horror showed itself clearly at

liberation. After struggling to live through the camps, some survivors expected
to feel elation when liberated, but that expectation often could not override their
adaptive numbness. In fact, liberation only highlighted the survivors blunted
emotional state. Emina N. describes her feeling of being drugged and her subsequent response to liberation:
It was a day before the liberation. . . . I was walking through the dead [bodies]. I wanted to go to the toilet. I have to walk through all the piles of dead
people and I dont believe that I had any feelings. It didnt bother me at all. I
didnt. Either way, like it would be some dirt or something, I walked away.
I think it was either something wrong with me, with us, with everybody, or
maybe they gave us something, like medication, or something that you were
just numb. And, in fact, when I was liberated, I didnt even want to go out.
(Emina N. in the Testimony of Emina N. and Miriam W., 1979).

Celia K. summarizes her emotional experience just after liberation: There

werent any emotions at all. You couldnt hate. You couldnt feel (Testimony
of Celia K., 1980). Leon W. agrees: I dont think I was happy. . . . I dont think
we had any feelings for dancing or joy or smiling (Testimony of Leon W., 1979).

Associated Emotions
Although fear and numbness are often the only emotional states that survivors
talk about, they were not the only emotions to be experienced. People may not
remember other emotions during horric events because shock may have prevented them from being aware of these emotions or because the strong memory
of fear and numbness obscures memory for other emotions. In particular, adaptive numbness can remain long after the horrors end, so the experience of other
strong emotions may arise only after many years. Clemens L. describes an image from early childhood, just before he went into hiding. While he was playing
with other children in a courtyard near his apartment, the owner of the apartment building came over to him and demanded that he pull down his pants.
Clemens complied, and the owner saw he was circumcised. After Clemens told
his mother about the incident, the two of them quickly escaped. Clemens goes
on to describe his emotions then and now: It seems like I dont remember the
feelings at the time. I think when I speak about those images, I have. Im speaking from the point of view of a stunned person. I have feelings now of course. . . .
I do have very strong feelings about it. And I think I guess I have more feelings
about it than I thought I did (Testimony of Clemens L., 1990).
Izzard (1991) describes patterns of emotion, combinations of fundamental
emotions that occur together with the same motivational impact. Although fear

358

and numbness predominate in the memories of Holocaust survivors, sometimes

complex emotions are remembered, even within the connes of atrocity. After
being transported by truck to a large building, Esther W. and others were forced
to undress and were told they were going to the showers. They had all heard rumors about gas chambers, so there was the fear that they were about to die. Esther
says, Once we went in, naked, the things were put aside. Everything was taken
away from us, and the doors were locked. And there were little windows and the
Germans stood, looked in the windows. And we were waiting for the gas to come.
And all of a sudden, hot water came. Hot water came, so, of course, I wouldnt say
joy, but at least you werent gassed (Testimony of Esther W., 1994).
After adapting to the routine of horror, some survivors remember the reemergence of motivating emotion growing out of their numbness and helplessness. Eva L. worked to maintain cooperation within her group of ve prisoners
at Auschwitz, using hatred and creative thoughts of revenge to sustain her:
We started building hatred. Hatred, and all kinds of ideas what we going to
do to the Germans when we live through this. And that hatred made us survive and made us live. The thoughts that we had . . . were sickening. . . . We
said we going to cut them into pieces. We going to put salt on them. We going
to tie them to two horses and let the two horses run. The most horrible, horrible things that could come to mind. We fed ourselves with that hatred, and
that made us go. (Testimony of Eva L., 1982)

Separation of Emotional Memory

and Event Memory

Analysis of oral testimony indicates that emotional experience is represented separately from other event information, such as location, time, people, and sequences
of action. With child survivors in particular, memory for strong emotion and
memory of specic events can be unconnected. They can recall emotional events
without remembering the emotional experience, and they can recall emotional
experience without recalling the events themselves (Kraft, 1996).
Susan S. talks about separating from her mother as a child and going into
hiding: To be very honest I must have repressed a lot of the feelings. I really dont
remember missing my mother. Susan knows that she must have felt strong
emotions, and she clearly remembers not knowing where her mother was, but
she does not recall the emotions she felt (Testimony of Susan S., 1991).
Conversely, Clemens L. says, My emotions are so powerful. Im surprised Im
feeling them as much as I am here. So thats there. But the placement of it, the
specic memory of it, thats dicult (Testimony of Clemens L., 1990). Similarly,
Hilda S. primarily recalls emotion but is unsure about the associated events. From
her early childhood, Hilda describes her memory of Kristallnacht (the night of the
broken glass), which began November 9, 1938, a night during which synagogues
and Jewish businesses throughout Germany were damaged and burned: There

359

were res all over the city and obviously we all knew about that and there was talk
in the family about these events. Whether I saw it or whether I heard it, I really
cant remember. But I do clearly recall being very frightenedbeing terried is
more the word for itby what was happening (Testimony of Hilda S., 1993).
Amelia B. describes what she remembers of the 3 days when she was in a cattle
car being transported to Auschwitz: The funny part is I dont recollect the train
ride so clearly. I mean I remember going into the train and coming out from the
train. But for the three days really forgot what really happened. But I remember
seeing my father. He was so scared, he was so. And somehow we were sitting
around and it was like utter despair (Testimony of Amelia B., 1994). For Amelia,
the emotions inside the cattle car are accessible, though the events that gave rise
to these emotions are not.

Repression
Survivors can remember emotions without specic event information and specic event information without the associated emotions. It is also the case that
both kinds of information can be inaccessible. As Bessie K. was transported from
the ghetto in Kovno to a concentration camp in Estonia, her baby was taken from
her. She details the shock and dissociation:
As I look back, I think that for a while I was in a daze. I didnt know what was
happening, actually. I saw they [were] taking away the men separate, the
children separate, and the women separate. So, I had the baby, and I took the
coats what I had with the bundles, and I wrapped around the baby and I put
it on my left side because I saw the Germans were saying left and right. And
I went through with the baby. But the baby was short of breath, started to
choke, to choke, it started to cry. So the German called me back. He says What
do you have there? in German. Now, I didnt know what to do because everything was so fast. Everything happened so suddenly. I wasnt prepared for it.
To look back, the experience, was I think I was numb. Or something happened
to me, I dont know. But I wasnt there even. And he stretched out his arms I
should hand him over the bundle. And I hand him over the bundle. And this
was the last time I had the bundle with me. (Bessie K. in the Testimony of
Bessie K. and Jacob K., 1983)

As a result of this dissociative horror, Bessie K. repressed the memory of losing

her child. She was taken from the camp in Estonia, detained in a eld for 2 weeks
with other inmates, then placed on a ship to Stutthof in Danzig. In Stutthof, Bessie
happened to see the doctor who operated on her in the ghetto following the birth
of her baby. She describes their meeting:
And when she saw me there, she was so happy to see me. Right away she said,
What happenedwheres the baby? What happened to the baby? And right

360

there I said, What baby? [Bessie pauses.] I said to the doctor, What baby?
I didnt have a baby. I dont know of any baby. Thats what it did to me. [She
points to her head with her index nger.] (Bessie K. in the Testimony of Bessie
K. and Jacob K. 1983).

Repressing a memory is not simply forgetting, nor is it the failure to recall a

specic horrible and consequential incident in the context of numerous similar
incidents. Repression involves memory for a complex event that is inaccessible
but potentially retrievable. Repression can then be dened as the inability to recall a distinctive event represented in memory that is generally agreed to be consequential even after repeated prompts to remember. Moreover, the distinctive experience
of dissociation appears to be a necessary factor for the occurrence of repression
(van der Kolk, 1996). Using this denition, I have identied four documented
cases of repression, the most dramatic being that of Bessie K. and her baby. It
should be noted, however, that the study of oral testimony is not well suited for
uncovering repressed memories. Simply put, if survivors are repressing important events, by denition, they are unable to describe these events in testimony,
and the repression remains concealed (W. Banks, personal communication,
November 12, 1994). Identication of repression occurs when the repressed
event is remembered before testimony is given or when historical documents or
other survivors independently identify the event in question. For more information on methods for studying repressed memories, refer to the debated methodology of Williams (Williams, 1994a, 1994b; Loftus, Garry, & Feldman, 1994);
the research of Briere and Conte (1993), Herman and Schatzow (1987), Loftus
(1993), Terr (1994, 1996), and van der Kolk and Fisler (1995); and the reviews
and comments of van der Kolk (1996), Burgess, Hartman, and Baker (1995),
Loftus (1994), and Olio (1994).

Re-experiencing Emotion
Multiple Systems
Based on the representation of emotion in memory, Robinson (1995) outlined
two broad explanations for recalling and re-experiencing emotions. The rst
proposes that information about emotion is a part of the overall cognitive representation of an event in personal memory. The emotional information must
then be decoded and described for the emotion to be re-experienced by the
rememberer. In fact, this approach is similar to that of William James (1890/
1950), who conjectured that emotions cannot be directly revived from
memory, but must be recreated. Reliving of emotion during a particular event
follows retrieval of memory for that event and subsequent thoughts about the
memory.
Robinson (1995) discusses a contrasting explanation that begins with the idea

361

that emotional experience involves complex states consisting of physiological

arousal, perceptual representations, knowledge, and evaluation, with dierent
representational systems for each type of information. Within this multiple systems framework, re-experienced emotions can be the direct result of accessing
an implicit emotional representation, separate from constructive processes. The
multiple systems framework posits that emotional experiences can be remembered
in four ways: (1) remembering how one felt but not re-experiencing the feeling,
(2) cognitively generated aect, (3) spontaneous aect, and (4) implicit memory.
Within this framework, people re-experience emotions for three reasons: (1) the
interpretation of a specic memory in such a way that emotions are cognitively
generated, (2) the direct accessing of core memory representations of past emotions, and (3) the triggering of implicit memories.

Re-experiencing Emotion

During Recall

During testimony, emotional experience that follows from a description of ones

thought processes suggests cognitive generation based on the interpretation of a
specic memory. Emotional experience that is sudden and surprising suggests the
direct accessing of a core representation within a specic memory. Even when people
do not initially recall strong emotions, the process of extended, detailed recall can
expose core memories and lead to the experiencing of the represented emotions.
Many times the recalling of a specic event produces sudden emotionusually grief or anger. The display of emotion can seem abrupt and unexpected until
one realizes that the emotions are released with the disclosure of a specic episode in memory. Judith G., a young survivor of the Vilna ghetto and Riga, explains: The scars are there. You dont talk about it as much. You dont discuss
it. But when you think about it, there is a great emotional charge that carries
forward (Testimony of Judith G., 1992).
After 9 minutes of at, dispassionate description, Sally H. tells of a specic
incident in the ghetto when a rabbi is humiliated and tortured. It was horrible. . . . They were pulling his beard, and blood was coming out. And no one
would help him. She then cries, abruptly and unexpectedly (Testimony of Sally
H., 1979). After 40 minutes of testimony, Amelia B. cries for the rst time as she
talks about a friend who was walking behind her at Auschwitz: I had a little
friend. Shes here. And she walked behind me. [Amelia starts crying suddenly,
then composes herself.] And she said, Well never come out from here. She
didnt. [Amelia cries again.] (Testimony of Amelia B., 1994).
When survivors like Sally H. and Amelia B. cry while giving testimony, they
often apologize, indicating that they cannot control the pain of the memory and
that the specic episodic memories are the source of the emotion.4 In these examples and others with sudden grief, the emotional experience appears to arise
directly from emotional representations in core memory.

362

Leon H. gave testimony in 1985, 40 years after the end of World War II. After
testifying for more than 2 hours, describing one atrocity after another, Leon
concludes: Angry all the time. Face looks angry. I dont know how to straighten
my face out. He tells of his mother who was partially paralyzed from a stroke,
though she still worked in the Lodz ghetto, weaving rags. One day, two SS men
broke into their apartment and demanded that she tear up the oorboards so they
could search for valuables. She was physically unable to do so quickly enough,
so the SS men brutally kicked her, and she died soon after. Kicked her so hard,
she hardly could breathe, Leon says. That was my anger! His face shakes, and
he cries. Then he stops (Testimony of Leon H., 1985). Leons sudden display of
emotion appears to be the result of accessing a core emotional representation in
memory, formed at the time he witnessed the beating of his mother.
Other displays of anger during testimony appear to be cognitively generated.
Clemens L. describes hiding in a convent near the end of the war:
I remember my mother coming to visit me on Sundays. I remember one time
she came to visit me, she brought me an egg. [He pauses, becoming emotional.]
I seem like Im more sad now, thinking about it, than I was then. But to bring a
goddamnto be thankful for a fucking egg is really preposterous to me. It
really pisses me o, to think about it that way. But, I must have been eternally
thankful to get this one fucking egg (Testimony of Clemens L., 1990).

The abrupt coarseness of the language seems out of character for this witness,
and the anger appears to erupt without warning. But based on what Clemens
says, his anger arises only after he evaluates the deprived conditions of his own
childhood in terms of present schemas of decency and responsibility in the care
of children.
When relating a specic incident during which she experienced humiliation,
Esther W. cries suddenly, her emotion emerging in concert with the remembered
humiliation. While in the slave labor camp of Skarzysko, Esther performed 12-hour
shifts with two other women, working a machine that engineered bullets. She
recalls the time that she and the other two women were accused of sabotage:
One day the three of us were called, taken away from the machine, brought
into a shed. . . . We had to bend down [she gestures] and were beaten [one]
by one. And two ocers were standing and waiting and looking while we were
beaten. And when it came to me to be beaten, unfortunately I made in my
pants, or whatever. And they stopped. From fear, I urinated. Im not ashamed
to say it but. [She cries.] . . . Its a funny thing. Im not crying because I was
beaten, but the humiliation (Testimony of Esther W., 1994).

Esther W. shows emotionally generated emotion, crying in response to the remembered humiliation at the hands of her tormentors.
In contrast, some survivors recall their experiences dispassionately, describing gruesome events and cruelties that they personally witnessed, without ex-

363

pressing emotion while recalling. Arnold C. atly describes the aftermath of allied bombing at Zeldenlager: In the morning, there were arms and legs all over
the place, on the wires, on the barbed wire, got caught. I must admit that it was
the rst and only place where I saw cannibalism. I saw two people take a piece
of meat from a body and try to make a re and cook it. The German ocer who
walked by, who saw it, shot them immediately. [He then gives an arming nod
to the camera, as if to say, I witnessed this and I can talk about it.] (Testimony of
Arnold C., 1983). In telling the events, Arnold C. calls on narrative memory, not
core memory. He does not access core emotional representations, nor does he
show cognitively generated disgust or anger. His motivation is to tell the events
clearly and directly.
Repeated Testimony Fourteen survivors gave testimony at the Fortuno Video
Archive more than once, and analysis of their later testimony reveals the emotional
aftermath of extended recall itself. All the survivors say that giving testimony is
profoundly distressing and that they are surprised at the strength of their distress.
Daniel F. evaluates his rst testimony: I didnt realize that its going to just take
me to the depths of depression for months. I didnt realize it (Testimony of Daniel
F., 1988). Eva B. agrees: When I did the rst tape here, I didnt expect to get as
upset as I did get . . . I cant imagine what the reaction would be (Testimony of
Eva B., 1988). Hanna F. says that she needed to be tranquilized afterward because
of the powerful emotion released during her testimony (Testimony of Hanna F.,
1988). In fact, little is remembered from the testimony itself except the emotional
pain. In her second testimony, Dori K. succinctly summarizes this pattern of recall. She says, I remember very well how I felt. And I remember the interview.
But I dont remember, really, what I said (Testimony of Dori K., 1988).
What accounts for this emotional response to extended recall of trauma? The
survivors themselves theorize that the emotional pain is caused by the length of
the testimony, the concentration on a long sequence of events. Daniel F. explains:
Its the segment. It was a lengthy, in-depth kind of thing, where I felt that I
was back there. And I was depressed for quite a while afterwards. . . . Your
mind starts to focus on a particular segment of time, and youre trying to pick
out all the events in that segment of time. And then after the interview nishes, you continue trying to nd pieces of memory of those events (Testimony
of Daniel F., 1988).

For Daniel F., recalling the details of an event accessed core emotional memories while also encouraging rumination about the memories afterward.
Alan Z. said he does not cry when he talks to individuals and to larger groups
about the loss of his family during the Holocaust: Only when I go back that far
is there a lot of detail. You see, when you go to speak somewhere to a school or
to the synagogue, I dont go into these details where it makes me emotional (Testimony of Alan Z., 1984).

364

Though a subset of traumatic memories remains painfully active, other core

memories lie dormant, below the surface of consciousness. Linton (1986) drew
a similar distinction with normal memories between what pours out of memory
and what can be elicited (p. 65). Accessible with eort, these dormant traumatic
memories can be the most threatening to the survivor, and awareness that such
memories exist leads to a reluctance to give testimony. Once retrieved into consciousness, these memories engender emotional pain. Detailed recall can mean
experiencing the pain of core memory in accessible memories and opening associated memories that are otherwise not immediately accessible. As survivors
engage in extended recall, they relate episodes recalled only because they are
directly connected to immediately accessible episodes.
Sally H. says, Took me quite a while to come down from that interview. . . .
You dig up memories you want to bury (Testimony of Sally H., 1989). Eva B. says,
It was much more alive in me and much more burdensome to me than I had realized. I really thought I had it good and buried and it wasnt going to ever come
up again or bother me again (Testimony of Eva B., 1988). Dori K. says, It was
the rst time I had spoken at such length. . . . It was a very painful experience
(Testimony of Dori K., 1988). Lengthy testimony can revive dormant memories
of horror, which can remain active long after the testimony ends and create new
emotional disturbance. Martin S. worries that his dreaded nightmares will return
because of memories reactivated during testimony (Testimony of Martin S., 1988).
Some survivors avoid the pain of memory by omitting details, talking about
their lives in cryptic labels. Ruth A. rst encountered this avoidance of detail
during initial contact with survivors after the war:
The only question was, Where were you during the war? I was in a concentration camp. Thats it. I was in the partisans. Thats it. I was hiding
in thesome place. Thats it. Nobody spoke any details. It seems that the
people wanted to block it out from their mind. They did not want to talk about
it, just vaguely question: What is your name? Where did you live before the
war? Where were you during the war? And that was the whole conversation. (Testimony of Ruth A., 1994)

Renee G. says that her rst testimony was the rst time her memories came
out full force, explaining, It was almost like a release of emotions, with tension and fear and crying. By saying almost, Renee carefully qualies the process of release; there was tension and fear and crying, but not catharsis
(Testimony of Renee G., 1988). Eva B. agrees: I dont know that I ever will really
get it out of my system, but at least, to live with it more peacefully (Testimony
of Eva B., 1988). Traumatic memory seems to be a self-generating source of
emotional pain. Specic memories release emotional disturbance in the form of
depression and nightmares, especially when an extended sequence of episodes
has been recalled, but the power of emotional memory is not diminished through
the release of emotions during testimony.

365

Clinical psychologists who work with survivors of trauma promote the therapeutic value of talking about the trauma, allowing the victimized to examine
their emotions in detail and to regain control of their own individual narratives
(Harber & Pennebaker, 1992; Jano-Bulman, 1985). For Holocaust survivors,
however, the therapeutic value of speaking their memories is not a given. Survivors themselves disagree: some avoid opportunities to talk and others carefully
welcome the opportunity to communicate the horrors to receptive listeners.
Although obvious almost to the point of tautology, if imposed silence is exacerbating emotional disturbance, then talking can be therapeutic. In such cases,
conditions should enable these survivors to communicate their memories, without social discouragements, with listeners recognizing and helping to manage
the inevitable pain of remembrance. Depending on the desired audience, survivors could actively seek out environments that are receptive to remembering:
family meetings among survivors and their children, mediated by family therapists; ocial gatherings of Holocaust survivors; instructional programs for teaching young people about the Holocaust; archives for collecting oral testimony. To
force the reticent to talk is not helpful, but encouraging them to enter settings
that are conducive to talking can be therapeutic.

Re-experiencing Emotion Due

to External Cues

Emotions can be experienced suddenly during the course of daily life. Particular
events and situations in the world today can trigger retrieval of past emotional
experience, producing strong emotional responses. Many survivors give examples
of fears connected to specic Holocaust memories, many arising from what was
lost in the traumatic pastthe most pervasive fear concerning the safety of their
children. In her testimony, Nina S. evinces this fear while revealing its source in
memory: I sometimes look out there and I say, God All Mighty, it might happen
again. Whats going to happen to the babies? I used to say when my children were
born. I used to pray nothing should happen until they are six years old because in
my mind, up to six years old, they going to take them, they going to kill them.The
source of Ninas fear resides in her specic memories of the Lodz Ghetto, which she
explains: They tore the children, the babies away from the mothers. And there
were screams. Could you imagine? Babies. Up to six years old they tore them away
from mothers hands. Mothers were crazy. But they had no choice. . . . They were
going from one house to the next. And we sat there and we listened to it. (Testimony of Nina S., 1996). For Nina, fears about the well-being of her childrenand
later her grandchildrengrow from these memories. Many years later, when Nina
became a mother, these memories evoked insuppressible fear.
Oral testimony reveals specic fears that are nearly universal among Holocaust
survivors. Jolly Z. describes one such fear: When I came over to this country, even
on the railroads, even a conductor, I suspected. Because he had a uniform (Testi-

366

mony of Jolly Z., 1988). Martin S. concurs, providing more detail: A sharp, black,
dark-blue, black type uniform still makes me freeze no matter where I see it. He
adds, Uniforms, to this day, especially when theyre very sharply dressed, is something that changes my attitude immediately (Testimony of Martin S., 1988).
Survivors fear that food will no longer be available, they fear dogs, they fear being
trapped, and they fear abduction, with each fear linked to specic memories.
News of current catastrophes can revive emotional memories of the traumatic
past. Leon H. summarizes: If I look at you it reminds me of something. If I look
at world, reminds me of something (Testimony of Leon H., 1985). When those
in the United States and Europe belatedly learned of the genocide in Cambodia,
survivors reacted. Edith P. says, When I learned about Cambodia, I went into a
depression. It pains me terribly that the world has not learned (Testimony of
Edith P., 1980). When President Reagan visited Bitburg, Jolly Z. says she was
physically sick for the rst time since the war. She concludes that the emotion
in Holocaust memory does not pale, especially when it is triggered by something (Testimony of Jolly Z., 1983).
Reports in newspapers and on television summon past atrocities. Leo G. says,
You do open up a paper and you do listen to the radio and you watch it and it
hits you from every side, and everything is constantly back. Your mind walks
back and names and places and faces, and then you translate all this to your
family (Testimony of Leo G., 1988). Sabina G. says, Any little news, political
news, can disturb me. And can bring thoughts: Whats going to be? How its
going to be? Will we, will we have it again? (Testimony of Sabina G., 1984).
Prolonged trauma splits the self-concept, creating the experience of two separate selves, each supported by memories that remain irreconcilable. When events
in the world connect these two sets of memoriesHolocaust and post-Holocaust
this connection is profoundly threatening to the survivors current self-concept
and can cause a strong emotional response. Current events perceived as similar
to the events of the Holocaust can create deep despairevents such as the cruelty and devastation in Cambodia, Bosnia, and Rwanda. Seeing people in uniform or seeing large dogs can generate fear. When events in the present lives of
the survivors make contact with the traumatic memories of the Holocaust, emotional responses can result, in the form of fear and depression.
Sometimes, events in the world tap into the well of implicit core memories,
suddenly and without warning. When this happens, Holocaust memories can
ood into consciousness, unwanted and unrestrained, with responses that are
unavoidable and physical. Sally H. tells about a panic attack she suered while
visiting the Holocaust museum in Detroit. When she saw the ame in the museum, she became disoriented and frightened, and then she panicked and ran
(Testimony of Sally H., 1989).
Celia K. is clear about the extent and power of her memories. I keep on just,
not rehashing it. Im not looking for it. Its just automatically keeps on coming
to me. And somehow everything is vivid, very much alive, very much. Celia

367

relates a recent incident. A bonre had been built to burn garbage in her back
yard. Unfortunately, the re vividly retrieved the memory of her cousins who
were burned alive in a small synagogue. I became so hysterical. Was screaming nonstop. I could not stop. . . . I was screaming until my jaw came out. I could
not put my jaw back in. And I could not control myself. Celia then had to be
restrained (Testimony of Celia K., 1987).
Celina R. talks of overwhelming emotion the one time she visited Auschwitz:
Then we went. In Auschwitz there is one place that they collect hair and
glasses, eyeglasses [She gestures], and I saw that mountain of glasses. My little
brother wore glasses . . . and when I saw those glasses I was hysterical. I was so
hysterical. I just couldnt stop crying. All those things come back (Testimony
of Celina R., 1995).
Many survivors also speak of automatic physical responses of disgust when
hearing the language of their tormentors. Ernest R. says, When I do travel in
Europe and I hear Germans speaking, it does bother my ear. . . . Even until this
date, I cannot forget it . . . I will never forget it (Testimony of Ernest R., 1987).
Celia K. says, Today, after so many years, if I hear Ukrainian spoken or
Lithuanian, I cringe. I get very sick. I cant be in the same room with them . . .
everything opens up all over again (Testimony of Celia K., 1980). These responses are not the result of thoughtful resentment, and they are not the result
of attitudes learned over time. They arise from connections in core memory between the survivors specic memories of trauma and the language of their tormentors. Kluger (2001) observed that the strongest bond between an individual
and a place is the language of that place (p. 205).
Events in the world that trigger emotional responses can also be temporal.
Isabella L. describes a recurring change in mood every May, the time of year she
was deported from the ghetto in Kisvarda, Hungary, to Auschwitz:
For decades, the end of May, which is coming upon us soon, was an unbearable time for me. I never knew why suddenly I got terribly depressed and I
couldnt cope with the days in May. And I never quite remembered why. And
then it would dawn on me. . . . Then in June, I would have a dierent outlook on life. I would be more hopeful. But it would hit me like a clock. It reappeared every May (Testimony of Isabella L., 1989).

For Isabella, the time of year retrieves a representation of the original events,
which in turn connects to core memory, leading to a release of emotion and the
resulting eects on her mood. For many years, the mood change came without
her conscious awareness of its source.
Hilda S. says, I lived through some very dicult periods. For instance, when
my oldest daughter was eight, the age I was when I left home, I had a real tough
time. [She cries for the rst time, then stops.] I was always very. I guess, I guess,
I was always very obsessed with what my mother went through. Sending two
children away (Testimony of Hilda S., 1993).

368

According to Holocaust survivors, strong emotional and physical responses

can occur during daily life without warning and without reason. These emotional and bodily responses appear to arise from implicit representations in core
memory; to the survivors, the responses seem automatic and immutable. However, after repeatedly experiencing these unbidden responses, the survivors are
able to attribute their responses to specic memories of the traumatic events. They
do not use the language of cognitive psychology, but their meaning is clear.
Beatrice S. describes witnessing the absolute horror of family members being
brutally murdered. She hid with her father in a small cellar, watching as a Gestapo ocer shot her mother and crushed her younger brother to death beneath
his boots. As she witnessed the murder, Beatrices father clamped his hand over
her mouth to prevent her from screaming. In her testimony, Beatrice describes
the continuing physiological eects of this horror: To this day, I have trouble
breathing, and I think its because what happened (Testimony of Beatrice S.,
1982). After surviving the war, Nadia R. discovered that she experienced powerful waves of panic in the company of other Jewish people. She says, For the
rst time . . . after the concentration camp, I was surrounded with Jews. It
brought me such an anxiety attack . . . I started to hyperventilate (Testimony
of Nadia R., 1995). The implicit, somatic representations of suocation for
Beatrice S. and fear for Nadia R. are repeatedly activated and become evident
to the survivors, who then can attribute these responses to their memories of
specic events.
When a current event (spatial or temporal) causes a survivor to access implicit representations in core memory, the represented emotional and bodily responses can be triggered and experienced. The activating event acts as a tuning
fork causing an implicit memory representation to resonate, generating strong
responses. Survivors are initially surprised by the triggered responses, but after
becoming aware of the connection to particular worldly conditions, they construct intuitive theories to account for their overwhelming emotional and bodily
responses. These theories provide an explanation for the emotional ooding, but
they do not make contact with the implicit memories and do not alleviate the
problem of uncontrollable responses. Awareness of this connection does not
prevent the overwhelming emotions any more than awareness of the physics of
pressure waves and resonant frequencies can prevent unwanted resonant vibration. About the persistence of such responses, Martin S. concludes, Things like
that, I just cannot seem to shakeand probably never will (Testimony of
Martin S., 1986).

Enduring Emotion
Some emotions endure for the remainder of the survivors lives. Though not
constant, these emotions are repeated and persistent over many years. When

369

discussing long-term emotional eects, survivors frequently refer to their sadness and occasionally their anger, but the enduring emotion that they analyze
most elaborately and question most intensely is guilt.
In the literature on survivors of widespread catastrophes, there is considerable discussion and disagreement about guilt (e.g., Danieli, 1988; Hartman,
1994, Hass, 1995; Levi, 1986/1989). From the oral testimony in this study, it
is clear that Holocaust survivors experience guilt for extended time, often beginning shortly after liberation and continuing for many years. Many talk about
current feelings of guilt. As described by the survivors, guilt assumes several
forms, with multiple causes and dierent developmental paths.
Many survivors evince a questioning, retrospective why me guilt, most often
deemed survivor guilt. Menachem S. was 1 of only 2 children out of possibly 4,000
who survived his concentration camp; all the others were killed. As a result,
Menachem experiences a questioning guilt, which he expresses in its most concise form: You ask yourself the questions: Why? Why did I survive? (Testimony
of Menachem S., 1979). Similarly, Nina S. asks, Why are we the ones to live? And
so many wonderful people did not survive? (Testimony of Nina S., 1996).
Michael G. lost his family during the Holocaust, and memories of the events
that led to his familys murder continue to interrogate him: To live with it and
to justify your living. I feel, at many times, I feel guilty. I live with the guilt.
And its not because Ive done anything wrong. Its just a guilt there: why. Ive
lived with that guilt for a long time. I ask myself many times when I do certain
things and I say, well, I could have helped . . . if. If whatever (Testimony of
Michael G., 1992). Survivors describe this kind of guilt as arising from attempts
to make sense of events that make no senseevents that violate moral principles as well as the laws of probability. Their very existence is based on surviving in an environment without morality and for no reason other than luck.
Danieli (1988) suggested that this type of survivor guilt is, in part, an unconscious attempt to prevent oneself from remembering the absolute helplessness
experienced during the Holocaust. According to Danieli, guilt presupposes
choice and the ability to decide ones fate. The unyielding pain of guilt is thus
more acceptable to the survivors belief system than the experience of intolerable passivity and helplessness, and guilt serves as a defense against the
memory of a nightmare world devoid of choice.
In some cases, guilt assumes a form that many observers can imagine: an intense regret for failing to prevent the deaths of others. After the brother of Abe
L. escaped the ghetto, Abe rescued his brothers young daughters but was unable to save his brothers two sons, ages 6 and 7. Abe confesses: These two boys,
I cant get over it, until this moment. Day and night, I cant go over it. Not being
melodramatic, these two boys are killing me (Testimony of Abe L., 1990). Abe
persists in thinking about what he could have done dierently to save his
brothers sons.

370

Survivors sometimes try to diminish the impact of guilt with conceptual

knowledge. Some survivors have learned to resolve the conict of living while
so many others died by emphasizing the concept of destiny. Samuel B. explains:
This stupid feeling of guilt. . . . How is this: if I am alive, that somebody died
because of me? Because I was supposed to be alive? All these things were very,
very heavy somewhere sitting in me. And it took me years and years to understand, to liberate myself somehow from all this. It was a very crippling feeling. Trying to understand that nothing of all that was my own personal
choice. It was all imposed on me by some games of destiny. (Testimony of
Samuel B., 1995)

After years of inner conict, many survivors ultimately incorporate the knowledge that responsibility lies with those who imposed the desperate conditions and
not with their own seemingly cowardly or selsh responses. As with Samuel B.,
Clara L. shed her conict and replaced it with a strong belief in destiny. She says,
I was one of those who was destined to survive. And I feel no guilt feelings. I did
whatever I could for my parents. I did whatever I could for my sister. And whoever was around me. More I could not do. So I dont feel guilty about it to remain alive (Testimony of Clara L., 1993).
Eva L. understands the impossible conditions that killed others in her family,
and she assimilates the concept of destiny, yet feelings of guilt persist. In the Lodz
ghetto, Eva lay in a coma for 5 weeks with typhus before nally reviving. To aid
her recovery, Evas father sold his food ration for medicine, and she credits him
for saving her life. Soon after, her father died of starvation. She says, He swell
up, and he died. And I could not help him. I couldnt do anything for him. After
describing her fathers death, Eva reveals a resolution of her underlying feelings
of helplessness: It was meant for me to live. Because under such circumstances,
nobody can survive. But Eva also describes the persistent, lingering eects of
her familys death: Sometimes I feel like guilt. Why am I the one, that I am here?
. . . And the rest of my family is not? She laments, I can never push it away. I
can never chase it away (Testimony of Eva L., 1982).
Horowitz and Reidbord (1992) provided one explanation for the etiology of
feeling simultaneously guilty and not guilty. When a person is victimized by
overwhelming trauma, the trauma destabilizes his or her normal experience of
a single self-concept. To accommodate the traumatic events, the normally coherent sense of self splits into multiple self-concepts, and these multiple selves
then permit people to process the traumatic events along several parallel
streams (p. 353). In fact, many survivors speak of split selves, simultaneously
experiencing the reality of the traumatic past as well as the normal present, as
reected in their emotional experience. The result many years later of these parallel streams of encoding is the experience of simultaneous conicting emotions:
guilty and not guilty, happy and not happy.

371

For survivors who were eyewitnesses to the murder of their own families,
strong feelings of guilt arose at the time of the murder and are represented with
the events. As I described earlier, Beatrice S. witnessed such an atrocity while
hiding from the Gestapo with her family:
It was very cold and dark and my brother started crying and my mother said,
Because of him well all get killed. So she went out with him and she covered us. She covered the top up in the cellar, so they wouldnt see theres an
opening. And she went to our next door neighbor, who was Christian
Polish, and the Gestapo shot her. And my brother was wounded. And I saw
through the little window there, and my father put his hand on my mouth
[she gestures], I shouldnt scream. . . . And he was wounded, and the Gestapo
came with his boots and just stepped on him.

Beatrices memory of this atrocity splits her self, causing her to feel simultaneously guilty and not guilty for the rest of her life. As a result, Beatrice says,
Its like there are two of me. Theres one that wants to have a zest for life, wants
to live and enjoy. And theres another part of me that I have guilt feelings (Testimony of Beatrice S., 1982).
Kochevit P. relates a similar atrocity for which she claims partial responsibility, consequently suering overwhelming guilt. As a young child, Kochevit
was taken in by a neighbor who saw the SS soldiers coming with dogs looking
for Jews:
And I saw the family, my mother, my brother, my grandmother, and my
aunt, and they shoot them down, on the place. . . . I saw this. . . . I was in
the cellar and I saw through a small window. [She gestures, ducking her
head down and peering out.] And my brother was begging, Please. And
he tried to run away. He said, Please. Dont shoot me. And he put his [She
gestures with her right hand in front of her head, then stops talking.] I cant
speak about this.

Kochevit then struggles to nish her account:

Was a very long. It took me a very long time to believe what happened. I
was a little girl. I said that they shot them, but they will never come back?
And I saw this with my eyes. [She looks down and pauses.] I suered a lot after
this. I was very sorry that I run away, as a child. Because I was blaming myself that everybody is dead. Only I am alive. I was very much disturbed. (Testimony of Kochevit P., 1979)

For those who were eyewitnesses to their own familys murder, the concept of
destiny fails because of the persistent memory of useless proximity. These survivors describe experiencing guilt at the time of the events, guilt that is represented
in the memory for these events.

372

Other guilt feelings seem cognitively generated, arising from interpretations

of past behavior. Abraham P. vividly describes his arrival at Auschwitz. After
getting o the train transport, Abraham told his younger brother to go with his
mother, and they were both sent to the gas chamber. I feel like I killed him,
Abraham says. I cant get it out of my head. It hurts me. It bothers me and I
dont know what to do. I feel that I am responsible for that. He then describes
his long-term thought processes: In one day . . . there were so many things have
happened to us, we really couldnt sort them out. And Im still trying to sort out
that day (Testimony of Abraham P., 1984).
Some survivors feel guilt for what they consider their own unfairness to other
victims. Edith P. confesses intense guilt about a single insulting comment she
made to an inmate at Auschwitz, an older woman who shoved Edith as they
entered their barracks. After years of suering from extreme deprivation, brutal
beatings, and daily humiliations, Edith refers to this brief insult as one of my
most painful episodes from Auschwitz. An oense that in normal times might
demand an apology intensies into an enduring insult, with Holocaust survivors interpreting their thoughtlessness or selshness as contributing to the misery of their fellow victims, however excusable and transitory. The interpretation
of the insult alters their self-concept by placing these survivors in the role of victimizer, along with its associated characteristics. To this day, Ediths rudeness
haunts her, and every year at Yom Kippur, Edith asks forgiveness from the
woman (Testimony of Edith P, 1987).

Enduring Emotion as the Result

of Cognitive Discontinuity

In part, the persistence of emotion stems from discrepancies between fundamental expectations and the actual events, the eort to account for these discrepancies, and the inability to resolve them in the case of extended atrocity.
According to the schema-based framework of Harber and Pennebaker (1992),
emotions arise when there is a conict between our expectations and our direct encounters with events in the world. Emotions then direct attention to the
unsuspecting schemas so they can be restructured to accommodate the discordant events. When expectations are realigned to account for the violating
events, the emotions dissipate. When expectations and the remembered events
are not aligned, emotions remain. With survivors of extended atrocity, the gulf
between schema-based expectations and remembered experience is so wide
that the discrepancy remains unresolved, and the emotions persist.
Violet S. tries to take on the untenable perspective of the architects of genocide, but she is unable to, and the irreconcilable dierence between her own view
and that of the murderers generates great anger: Children. Children. Imagine
children to be alive? I mean this is. To this day I mean, it wasnt just a mess of,
bunch of, hoodlums that run them. These were the top intelligence, scientically,

373

who planned to burn our children alive. I never, never forgive them. Never. I tried
(Testimony of Violet S., 1991). Hilda S. presents the same pattern. Her fundamental beliefs about the treatment of human beings cannot be restructured to
account for the unprecedented cruelties she experienced, and her emotion persists. Hilda says, I have always obviously been very bitter about what happened
to the Jewish people. I carry a lot of hatred. And I think Im always trying to
understand how could it have happened (Testimony of Hilda S., 1993).
After the war, after survivors re-entered stable social settings, their emotional
state conicted with the emotional state of others around them. This interpersonal disparity paralleled and reinforced the cognitive disparity within the individual survivor. For example, after Julien E. immigrated to the United States and
began attending high school, his own emotional state conicted with that of his
peers. Julien says he was unable to take part in the seemingly frivolous experiences of his fellow classmates because of his sense of continuing mourning,
which, he says, made it impossible to partake fully in the lighter side of life enjoyed by peers in school and elsewhere (Testimony of Julien E., 1995).

Distracting Emotional Memory

Instead of asking why survivors of atrocity experience strong emotion, perhaps
we should ask how they manage not to. Certainly, survivors can avoid talking
about their traumatic memories, but to avoid thinking about these memories is
another matter. What survivors discuss most frequently is the importance of
meaningful distraction in their daily lives. Such distraction eectively decreases
the pain of emotional memory in four ways: by focusing ones thinking on activities in the present, by diverting attention away from potential triggers of traumatic memories, by decreasing the likelihood of cognitively generated emotion,
and by allowing recent memories to overlay older, traumatic memories, helping to redene the remembered self in normal life.
Raising children, building careers, establishing friendships, and carrying out
civic responsibilities direct the attention of Holocaust survivors away from the
traumatic past and toward the challenges of normal life. Alan Z. says that he
suered great fears and nightmares after liberation, but these fears and nightmares went away with the distraction of work and family. He says, After I came
to the United States, and I started to work, everything disappeared. I mean, my
life changed drastically, when I was in the United States. My work, I was involved
in my business. And raising my family. And it just moved away from me (Testimony of Alan Z., 1984). Many survivors make similar observations about the
relationship between the emotional pain of traumatic memory and the distractions of normal life.
Intuitively, survivors distinguish between the cognitive and the emotional.
Rena C. says, Intellectually I manage very well. I went to college and so on.

374

Work-wise Im okay. But emotionally, I am not okay (Testimony of Rena C.,

1992). Karl S. says, I believe I am a successful professional in my eld, as a technician. But as a person, as a Jew, I feel Im sitting on a volcano (Testimony of
Karl S., 1980). Survivors do not use the same terminology as memory researchers, but they clearly distinguish between cognitive strategies for managing emotional pain and the experience of emotion.
Some survivors describe overriding their emotional distress with resolute cognitive strategies. Near the end of his testimony, Harry F. says, I had to make a
decision. Is this going to bother me for the rest of my life, or what? And I decided
that I couldnt let it bother me for the rest of my life because that would be an admission that the other side had won the war. He says, I stopped suering from
the consequences one week after it was over. Its something you never ever forget. Its impossible. But we have to educate the next generations about mans inhumanity to man and then go on from there (Testimony of Harry F., 1995).
While living in Czechoslovakia after the war, Nadia R. actively suppressed her
memories and hid her Jewish identity, which she interpreted as a sign of wellbeing. It was totally buried in me, she says, and I felt that is a good thing. That,
I thought, was my proof how well I came across the whole trauma. During the
23 years that Nadia lived in Czechoslovakia after the war, her closest friends were
unaware of her background or her religion (Testimony of Nadia R., 1995).

Emotional Memory Getting Worse

The power of distraction becomes most evident when it diminishes. Later in life,
when the healthy distractions of work and raising children go away, many survivors experience this loss of distraction by experiencing their memories as growing stronger and more distressing. Near the end of his testimony, Abe L. says:
But the worst of all, certain people, for some reason, some reason, certain gures, certain people is in you, in your brain. Certain people, they stay with
you and they cant get away, they cant, they just cant get away. Anyone, if
he thinks, he sees the hole in his heart, isis not getting smaller, is getting
bigger. And the Holocaust itself, I thought when years go by, passes by, these
will go away. We will have forgotten. No way. Its getting closer. The Holocaust is getting nearer, and not farer [sic]. Its getting nearer. (Testimony of
Abe L., 1990).

Isabella L. says, Its much harder now. I feel my head is lled with garbage. All
these images and sounds. And my nostrils are lled with the stench of burning
esh (Testimony of Isabella L., 1989). Eva B. agrees: Its going to be there until
the day I die. She says, Its more painful; its more alive rather than less so
(Testimony of Eva B., 1988).
Leon W. explains why the torment of memory is worse: I think about it more
now, maybe because I get older . . . I think a lot about it. Too much I think. Alex

375

H. says that for many years after the war, he was so involved in the ght for a
new existence that he did not think about the past. Beginning with no family,
no schooling, and the wrong language, Alex says the daily ght to establish himself used all his energy. In fact, he suppressed his time in the concentration camp
until 3 years before he came in to give testimony. He describes the result of having accomplished his goals: My past is starting to haunt me . . . and I feel so
depressed, very often. That I actually feel that I, very often feel that I lived long
enough. [He speaks softly and slowly, looking down] (Testimony of Alex H.,
1985).
In the laboratory, Reisberg and Heuer (1992) showed the importance of retention interval in assessing emotional memory, with strong emotion slowing
the process of forgetting with memory tests delayed between 1 and 2 weeks.
Outside the laboratory in naturalistic studies of emotional memory, retention
intervals can be considerably longer. Using the birth of a younger sibling,
Sheingold and Tenney (1982) examined retention intervals from 1 month up
to 16 years, with the results showing only a slight decrease in memory over this
interval. Over longer periods of time, some emotional memories appear virtually permanent, showing no decline. In my study of oral testimony, recall intervals ranged from 35 to 55 years, with witnesses reporting no loss of memory for
traumatic events (Kraft, 2002). Over these longer intervals, changes in the
thoughts about ones emotional memories can be experienced as changes in the
memories themselves. After looking back over the terrain of normal life and recent world events and realizing the implications of the past atrocities, ones interpretation of memory changes. The images remain, but the memories seem more
horrifying.
Eva L. ruminates about why she survived, with her focus on memory. First,
she talks of her memories of the Holocaust: The older I get, the more memories
are haunting me. I keep asking myself, why, what is it? Is it fate that I am here,
of the whole family, of the whole wonderful culture, only I am here from a whole
family? . . . Its getting harder, its getting worse with age. I sit and I work, and I
think about things that I went through and I wonder. Later Eva says, I think
when I was younger, it was easier because I was occupied with the children. I
was busy with them. And I try, I try to push away those thoughts, but now they
come more often, and it . . . haunts me (Testimony of Eva L., 1982).
Eva L. assesses her life raising children after the war, and she cannot reconcile the two sets of memories: I myself cannot believe that I was there and I was
able to live through a horror like this and yet come here to this country and have
a family and live a normal life, more or less, a normal life, and bring up my children a normal way. And I wonder and I lay awake. And when I lay awake, I relive
lot of things (Testimony of Eva L., 1982).
A contributing factor to the increasing pain of Holocaust memory is the loss
of normal, pre-Holocaust memories. When Eva tries to recapture her preHolocaust memories of times when she was happy, she fails. She says, Nights

376

that I am up and I cannot sleep, I try to see my parents. Its kind of like, so many
years, the pictures fading away and I wont remember them. Comes a holiday,
I try to copy, and remember how it was at home. I can never copy (Testimony
of Eva L., 1982). A fading happier childhood, a decrease in worldly distraction,
and the laser clarity of the remembered horrors combine to worsen the torment
of traumatic memory. Eva works at remembering the faces of her parents and
the rituals of Jewish holidays, but these general personal memories formed long
ago during normal times are not strong enough to endure. Memory for neutral
events strengthens when rehearsed through thoughts or actions, yet these events
were simply not repeated enough to endure. Prevented from recreating past rituals as they were in the world, Eva endeavors to reconstruct them in her thoughts,
but the faded images of these comforting rituals lack the singular emotional and
perceptual vividness of traumatic memories. She has no diculty remembering
the atrocities that followed, and she can no longer distract herself from these
memories.

Rebound Eects
In their summary of research on thought suppression, Wegner and Schneider
(1989) documented the fact that distraction from thinking about unwanted
thoughts can produce rebound eects. After participants actively suppressed
thinking about a particular image (e.g., a white bear) by distracting themselves
with thoughts of a variety of other topics, they later experienced more thoughts
about the suppressed image than a control group who did not suppress. Wenzla
and Wegner (2000) summarized three processes that may be pertinent to suppression of thoughts about real-world traumatic events and subsequent increases
in the frequency, intensity, and duration of these thoughts. First, images used
to distract can become associated with the unwanted thoughts of trauma and
can later serve as reminders of these thoughts. For Holocaust survivors, many
years of actively suppressing images of the ghettos and the camps by distracting
themselves with thoughts of other events may return the images with more frequency and more force many years later. Moreover, when attention is focused
on ones children, there may be associated thoughts of the children who were
lost, either ones own children or siblings, nieces, nephews, and cousins. Raising children simultaneously serves as an eective distractor and a reminder of
the children who never had the chance to live.
Wenzla and Wegner (2000) also describe ironic process theory, which involves two strategies: (1) actively seeking out other thoughts in an attempt to
suppress unwanted thoughts and (2) a monitoring process that remains vigilant
for failures in the suppression of unwanted memories and thoughts. Although
this monitoring process serves a useful function, notifying the individual when
to expend more cognitive eort on suppression, when active suppression diminishes or goes away entirely, the monitoring process continues, highlighting the

377

unwanted memories and thoughts. Wenzla and Wegner propose a metacognitive

process that depends on the underlying belief that suppression should be successful. When it is not, people become even more focused on suppressing and even more
critical of themselves when suppression fails. This view is supported by testimony
from survivors who say that they felt successful when hiding their past, that
hiding the past from oneself and others was a sign of having endured the trauma.
Associated with the belief that suppression should be successful is the corollary that
a reversal in the eectiveness of thought suppression is a sign of declining emotional
health and personal failure.
Holocaust survivors repeatedly demonstrate the pattern of distracting or
suppressing traumatic memories for many years and then experiencing a
strong return or reformation of painful emotion. Fifty years after the end of the
war, Aladar M. talks of his nightmares: Even more lately than in the beginning . . . because in the beginning, it was more like I have to work, I have to
make a living, I have to do things. I did not have the time to think about it.
There was no time (Testimony of Aladar M., 1995). After 23 years, Nadia R.
actively reversed her initial decision to suppress memories of the Holocaust and
chose to accept her past. With that acceptance, she says her former self-concept
returned, a self-concept dened immediately after liberation from the concentration camp of Terezin and laden with fears and anxiety (Testimony of
Nadia R., 1995).
Sybilla F. decided to be positive and not dwell on her suering during the
war. She says that this attitude did not come quickly but that she was able to
use strategies to overcome her emotions. After the war, after Sybilla came out of
hiding, her remaining family did not talk about her brothers death or the deaths
of her mothers family. She says, To break down and then to feel terrible would
not have been the way forward. A few years before her testimony in 1991, however, Sybilla went through serious depression for a year and a half, which she
describes: Everything went black. [She gestures, with her hands spread wide.]
And everything negative. And it just happened. . . . It was just something that
happened (Testimony of Sybilla F., 1991).

Conclusions
The most pervasive nding in the study of Holocaust testimony is the extraordinary persistence of emotional memory. More than 50 years after the events,
emotional memory remains vivid and powerful, with specic memories causing
survivors to cry suddenly, to break down uncontrollably, to become enraged.
Recent memory does not weaken or conceal older memories of atrocity, and time
does not diminish their potency. Recalling past traumatic experiences does not
reduce emotional pain, and there is no cathartic release.

378

During testimony, survivors appear to experience emotion through cognitive

generation or through direct accessing of core emotional representations. Cognitive generation is suggested if emotions are displayed after interpretation of the
memories on the part of the survivortypically a comparison of the remembered
events with events in normal life or an assessment of the choices the perpetrators made. Core representations are implicated if emotion occurs suddenly in the
testimony and is followed by an apology or by the survivors observations that
the emotion was unexpected. When the poet Theodore Roethke wrote, We think
by feeling. What is there to know? he embraced both the concept of core emotional representations and their primacy. In daily life, emotions can be reexperienced suddenly and without warning if events in the world activate
implicit core memories of trauma, releasing emotional and bodily responses.
One puzzling pattern in the testimony concerns numbness. So predominant
in the remembered emotions of survivors, it is largely absent during recall and
diuse in the daily lives of the survivors. Numbness is remembered as an allencompassing feeling inhabiting ones entire self. Survivors sometimes say that
the source of the numbness was externalin some cases incorrectly attributing it to an ingested substancebut they remember the numbness as an experience involving the entire self. For such numbness to be re-experienced fully,
similar external conditions need to be reinstated, conditions that are impossible
during testimony and unlikely in the present lives of the survivors.
Complicating the re-experience of numbness is the duality of the survivors
self-concepts in their lives today. Consider a few phrases that survivors use to
describe their experience of self: a double existence, another world, a schizophrenic division, two worlds, two dierent planets, double lives. Many
years after the events, survivors have the phenomenal experience of two separate selves, with residual numbness in the past traumatized self. The child survivor, Renee H., explicitly addresses this split in her self before and after the war:
What is left is . . . two separate units in ones experience. And so there is the
me that is the wartime and prewartime me and me that is the postwartime.
Its like having an era before and after. And thatwhile they are all connected
in myselfthey are not reconcilable. And it took me a long, long time to realize that not only are they not reconcilable, I dont want them to be reconcilable. That I wanted them to be separate. (Testimony of Renee H., 1979)

The diuse experience of numbness in daily life shows itself in the duality of self
after prolonged trauma, in the self-concept constructed during the period of extended trauma. When survivors speak of numbness later in life, they refer to the
self-concepts they constructed during the Holocaust. The experience of Alina Z.
is representative: I was like stone. I couldnt believe. And something stay in me.
Im very happy, but I cant show a lot. I cant cry with tragedy. Now, she says
her children cannot tell when she is happy: I said Im happy. Im very happy,

379

but I just cant show the feeling. Cant show. When Im happy, they think Im
not happy (Testimony of Alina Z., 1993).
When survivors speak of emotional masks, they are referring to hiding their
Holocaust selves. Recalling her time in Auschwitz, Violet S. says, Many times,
people that dont know me, they look . . . they have no idea. After describing her
escape to Prague during a death march from Gross-Rosen to Bergen-Belsen,
Beatrice S. admits in her testimony that her children do not know the real me,
conceding that she is playing the part that is expected of her: You put on a smile
and you go. You just become an actress (Testimony of Beatrice S., 1982).
For all survivors, grief persists throughout their lives. For many, guilt and
anger remain. For some survivors, nightmares continue even to this day. One
strategy that survivors use to reduce this emotional pain is distraction. But the
palliative eects of distraction are temporary, and the emotional pain of deeply
traumatic memory remains, often growing stronger as distractions diminish.
According to these survivors, an eective long-term strategy for accommodating the emotional pain of traumatic memory is to nd meaning through communication of their memories.
The child survivor and psychiatrist Robert K. presents a symposium for high
school students every year. In his testimony he asserts, Nothing that I have ever
heard from any psychiatrist or psychoanalyst in the world . . . matches one
twenty-minute presentation by one of my survivor panelists in terms of their
therapeutic well-being. He points out that these survivors have guarded their
memories for decades, hiding the memories within themselves. He then describes
the survivors therapeutic revelation when speaking in public:
They discover that there are ve hundred students there who listen to them
and you should see the listening that goes on when they speak. And they see
perhaps, perhaps theyre doing something with their experience that some
of those ve hundred will be touched by them, to perhaps make even this
much [He holds his thumb and forenger, one inch apart.] of a shift in their
perception of life, their philosophy. That gives the survivor great hope. And
Ive seen [them] go through it and get depressed and break down and be in
tears and stay depressed for two weeks. And tell me that it was the best day of
their lives. (Testimony of Robert K., 1984).

Fivush, Haden, and Reese (1995) documented the importance of children

talking about their experiences to develop their skills in recounting memory
and in learning the value of reminiscing. They proposed that language and
memory can become inextricably intertwined in the development of autobiographical memory (p. 355). In chapter 8 in this volume, Fivush and Sales
document complex linguistic interactions between parents and children that
are specically directed at the childrens memories for personal events. In stark
contrast, those who are repeatedly traumatized in childhood learn the value
of not talking about the past. They learn not to reminisce. According to the

380

testimony of child survivors of the Holocaust, much remembering is unspoken,

often insistent and unbidden. Very often, memories come to mind without
being narrated or expressed. Later in life, these survivors discover that it can
be therapeutic to unlearn the lesson of silent remembering and to nd ways to
communicate their memories.
Recalling traumatic events in public exposes painful emotion but does not
diminish it. Documenting the traumatic events does not ease the pain of memory,
nor does it provide meaning to the memories of suering. Change occurs when
survivors reinterpret the function of memory. No longer meant to be hidden, traumatic memories are meant to be communicated: to educate others and to commemorate the lives of those who were lost. Translating memories of atrocity into
understandable narratives gives meaning to the act of recallingan act that formerly provided only torment. Narrating the traumatic events of the past provides
meaning to the telling of these events, allowing the survivors to coexist with the
emotional pain of their memories.

Notes
I thank Daniel Reisberg, whose thoughtful suggestions and skillful editing strengthened and claried this chapter. I gratefully acknowledge permission from the Yale
University Library to reproduce oral testimonies from the Fortuno Video Archive
for Holocaust Testimonies. Permission must be obtained before reproducing any of
the testimony quoted in this chapter. Please contact Manuscripts and Archives, Yale
University Library, and cite the specic testimony to be reproduced.
1. A list of archives, primarily in the United States, can be obtained from the Oral
History Association at the following ground address: James Sleight, Associate Director, H-Net, Humanities and Social Sciences OnLine, Michigan State University,
East Lansing, Michigan 48824-1120; or at the following Web page: www2.hnet.msu.edu/~oralhist/projects.html.
2. If the original interpretation of events is faulty in some way, memory can
accurately represent this faulty interpretation. For example, Jolly Z. insists incorrectly, They put a chemical in our food so that we did not menstruate later. There
was a chemical in our food and we did not have menstrual periods (Testimony of
Jolly Z. and Rosalie W., 1979). Adele W. draws the same conclusion: And you know
how we suered until we got back our periods. We were. They gave us some certain pills to lose our periods. . . . I was very, very skinny, but I was like a balloon.
Blown up. Because we didnt have our periods (Testimony of Adele W., 1982).
These witnesses remember accurately that their periods went away, but they misinterpret the reason and accurately remember this misinterpretation.
3. The concepts of ashbulb memory (Brown and Kulik, 1977), deep memory
(Langer, 1991), situationally accessible memory (Brewin et al., 1996), verbatim
memory (Reyna, 1998), core memory (Kraft, 2002), and emotional memory (chapter
3 here) all converge on the idea of two distinct representational systems in personal event
memory. Although the dierent systemic distinctions do not t together cleanly,
the gist is that two levels of memory can be distinguished on the basis of phenomenal

381

experience, clinical analysis, laboratory studies, and neurophysiological processes.

A case can be made for further subdividing personal event memory, but in the context of Ockhams razor, the results of my analysis of oral testimony encourage an integration of these distinctions to support a dual representation theory.
4. The following are characteristic examples of survivors crying during testimony
and then expressing surprise at their emotion. Renee G. describes how her younger
brother desperately pleaded for a hiding place to save his life. She then surprises herself when she begins to cry: I didnt think I would cry today. Sorry (Testimony of
Renee G., 1988). Josephine B. describes how her father escaped to England during
the war and how the family was reunited afterward. When she talks of the reunion,
she unexpectedly cries: And then he was there. And he had a suitcase with him. I
always. I mean I love that in my memory. Big suitcase. And there was a pair of shoes
he had for each of us. Josephine then begins to cry and says to herself, Stupid, for
breaking down (Testimony of Josephine B., 1992). Zezette L. discusses how she never
talks with her brother about the Holocaust: He knows I went back to Auschwitz. He
cannot talk. He knows that I now talk, and forgive me. [Zezette begins to cry.] I dont
I dont usually become.Thank God, learned not to cry this way. But, I guess Im talking a little more than usual (Testimony of Zezette L., 1980).

References
Briere, J., & Conte, J. (1993). Self-reported amnesia for abuse in adults molested as
children. Journal of Traumatic Stress, 6, 2131.
Brewin, C., Dalgleish, T., & Joseph, S. (1996). A dual representation theory of posttraumatic stress. Psychological Review, 103, 670686.
Brown, R., & Kulik, J. (1997). Flashbulb memories. Cognition, 5, 7399.
Burgess, A. W., Hartman, C. R., & Baker, T. (1995). Memory presentations of childhood sexual abuse. Journal of Psychosocial Nursing, 33, 916.
Danieli, Y. (1988). Treating survivors and children of survivors of the Nazi Holocaust. In F. M. Ochberg (Ed.), Post-traumatic therapy and victims of violence
(pp. 278294). New York: Brunner/Mazel.
Fivush, R., Haden, C., & Reese, E. (1995). Remembering, recounting, and reminiscing: The development of autobiographical memory in social context. In D. C.
Rubin (Ed.), Remembering our past: Studies in autobiographical memory (pp. 341
359). New York: Cambridge University Press.
Harber, K. D. & Pennebaker, J. W. (1992). Overcoming traumatic memories In
S.-. Christianson (Ed.), The handbook of emotion and memory: Research and theory
(pp. 359387). Hillsdale, NJ: Erlbaum.
Hartman, G. H. (Ed.). (1994). Holocaust remembrance: The shapes of memory. Cambridge, MA: Blackwell.
Hass, A. (1995). The aftermath: Living with the Holocaust. New York: Cambridge University Press.
Herman, J. L. (1993). Sequelae of prolonged and repeated trauma: Evidence for a
complex posttraumatic syndrome (DESNOS). In J. Davidson & E. Foa (Eds.),
Posttraumatic stress disorder: DSM-IV and beyond (pp. 213228). Washington,
DC: American Psychiatric Press.

382

Herman, J. H., & Schatzow, E. (1987). Recovery and verication of memories of

childhood sexual trauma. Psychoanalytic Psychology, 4, 114.
Homan, A., & Homan, H. (1994). Reliability and validity in oral history, In
J. Jerey & G. Edwall (Eds.), Memory and History (pp. 107130). Lantham, MD:
University Press of America.
Horowitz, M. J., & Reidbord, S. P. (1992). Memory, emotion, and response to trauma.
In S.-. Christianson (Ed.), The handbook of emotion and memory: Research and
theory (pp. 343358). Hillsdale, NJ: Erlbaum.
Hubbell, S. (1999). Waiting for Aphrodite. New York: Houghton Miin.
Intons-Peterson, M. J., & Best, D. L. (Eds.). (1998). Memory distortions and their prevention. New York: Wiley.
Izzard, C. E. (1991). The psychology of emotions. New York: Plenum.
James, W. (1890/1950). The principles of psychology. Vol. 2. New York: Dover.
Jano-Bulman, R. (1985). The aftermath of victimization: Rebuilding shattered
assumptions. In C. Figley (Ed.), Trauma and its wake (pp. 1535). New York:
Brunner/Mazel.
Kluger, R. (2001). Still alive: A Holocaust girlhood remembered. New York: The Feminist Press at the City University of New York.
Koriat, A., Goldsmith, M., & Pansky, A. (2000). Toward a psychology of memory
accuracy. In S. T. Fiske, D. L. Schacter, & C. Zahn-Waxler (Eds.), The Annual
Review of Psychology, 51, (pp. 481537), Palo Alto, CA: Annual Reviews.
Kotre, J. (1995). White gloves. New York: Free Press.
Kraft, R. N. (1996, November). Memory in young children who survived the Holocaust.
Paper presented at the 37th annual meeting of the Psychonomic Society,
Chicago.
Kraft, R. N. (1998a). Holocaust memory: Persistent recall of traumatic events. In
R. R. Homan, M. F. Sherrick, & J. S. Warm (Eds.), Viewing psychology as a whole
(pp. 375 398). Washington, DC: American Psychological Association.
Kraft, R. N. (1998b, November). Levels of awareness during traumatic events. Paper
presented at the 39th Annual Meeting of the Psychonomic Society, Dallas.
Kraft, R. N. (2000, November). Revealing mistakes in personal memory. Paper presented at the 41st Annual Meeting of the Psychonomic Society, New
Orleans.
Kraft, R. N. (2002). Memory perceived: Recalling the Holocaust. Westport, CT: Praeger.
Langer, L. L. (1991). Holocaust testimonies: The ruins of memory. New Haven, CT: Yale
University Press.
Levi, P. (1989). The drowned and the saved (R. Rosenthal, Trans.). New York: Vintage. (Original work published 1986).
Linton, M. (1986). Ways of searching and the contents of memory. In David C. Rubin
(Ed.), Autobiographical memory (pp. 5067). New York: Cambridge University
Press.
Loftus, E. F. (1993). The reality of repressed memories, American Psychologist, 48,
518537.
Loftus, E. F. (1994). The repressed memory controversy. American Psychologist, 49,
443445.
Loftus, E. F., Garry, M., & Feldman, J. (1994). Forgetting sexual trauma: What does

383

it mean when 38% forget? Journal of Consulting and Clinical Psychology, 62,
11771181.
Nadel, L. (1994). Multiple memory systems: What and why, an update. In
D. Schacter & E. Tulving (Eds.), Memory systems (pp. 3963). Cambridge, MA:
MIT Press.
Olio, K. A. (1994). Truth in memory. American Psychologist, 49, 442443.
Reisberg, D., & Heuer, F. (1992). Remembering the details of emotional events. In
E. Winograd & U. Neisser (Eds.), Aect and accuracy in recall: Studies of ashbulb memories (pp. 162189). New York: Cambridge University Press.
Reyna, V. F. (1998). Fuzzy-trace theory and false memory. In M. J. Intons-Peterson
& D. L. Best (Eds.), Memory distortions and their prevention (pp. 1527). New
York: Wiley.
Reyna, V. F., & Titcomb, A. L. (1997). Constraints on the suggestability of eyewitness testimony: A fuzzy-trace theory analysis. In D. G. Payne & F. G. Conrad
(Eds.), Intersections in basic and applied memory research (pp. 157174).
Mahwah, NJ: Erlbaum.
Robinson, J. A. (1995). Perspective, meaning, and remembering. In D. C. Rubin
(Ed.), Remembering our past: Studies in autobiographical memory (pp. 199217).
New York: Cambridge University Press.
Sarbin, T. R. (1986). The narrative as a root metaphor for psychology. In T. R. Sarbin
(Ed.), Narrative psychology: The storied nature of human conduct (pp. 321), New
York: Praeger.
Sheingold, K., & Tenney, Y. T. (1982). Memory for a salient childhood event. In
U. Neisser (Ed.), Memory observed: Remembering in natural contexts (pp. 201
212). San Francisco: Freeman.
Spence, D. P. (1982). Narrative truth and historical truth: Meaning and interpretation
in psychoanalysis. New York: Wiley.
Terr, L. (1994). Unchained memories: True stories of traumatic memories, lost and found.
New York: Basic Books.
Terr, L. (1996). True memories of childhood trauma: Flaws, absences, and returns.
In K. Pezdek & W. P. Banks (Eds.) The recovered memory/false memory debate
(pp. 6980). San Diego: Academic Press.
Testimony of Abe L. (1990, May). Videotape recording number T-1394, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Abraham P. (1984, August). Videotape recording number T-731,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Adele W. (1982, October). Videotape recording number T-213, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Aladar M. (1995, May). Videotape recording number T-2854,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Alan Z. (1984, June). Videotape recording number T-284, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Alex H. (1983, July). Videotape recording number T-210, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.

384

Testimony of Alina Z. (1993, April). Videotape recording number T-2045, Fortuno

Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Amelia B. (1994, December). Videotape recording number T-2740,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Arnold C. (1983, March). Videotape recording number T-363, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Beatrice S. (1982). Videotape recording number T-72, Fortuno Video
Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Bessie K. & Jacob K. (1983, May). Videotape recording number T-206,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Blanche H. (1992, April). Videotape recording number T-2068, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Celia K. (1980, February). Videotape recording number T-36,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Celia K. (1987, October). Videotape recording number T-970,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Celina R. (1995, September). Videotape recording number T-3131,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Chana F. and Henry F. (1980, March). Videotape recording number
T-81, Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Clara L. (1993, May). Videotape recording number T-1850, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Clemens L. (1990, March). Videotape recording number T-1315,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Daniel F. (1980, February). Videotape recording number T-153, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Daniel F. (1988, April). Videotape recording number T-978, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Dori K. (1988, January). Videotape recording number T-969, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Edith P. (1980, February). Videotape recording number T-107,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Edith P. (1987, December). Videotape recording number T-974,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Emina N. and Miriam W. (1979). Videotape recording number T-144,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.

385

Testimony of Ernest R. (1987, May). Videotape recording number T-897, Fortuno

Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Esther W. (1994, August). Videotape recording number T-2609,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Eva B. (1979, May). Videotape recording number T-1, Fortuno Video
Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Eva B. (1988, April). Videotape recording number T-1101, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Eva L. (1982, May). Videotape recording number T-71, Fortuno Video
Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Hanna F. (1980, February). Videotape recording number T-18,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Hanna F. (1987, October). Videotape recording number T-971,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Harry F. (1995, March). Videotape recording number T-2761,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Henry S. (1996, July) Videotape recording number T-3380, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Hilda S. (1993, June). Videotape recording number T-2148, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Irene W. (1982, November). Videotape recording number T-65,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Isabella L. (1989, May). Videotape recording number T-1270,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Jolly Z. (1983, January). Videotape recording number T-220, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Jolly Z. (1988, January). Videotape recording number T-971, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Jolly Z. & Rosalie W. (1979, August). Videotape recording number
T-34, Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Josephine B. (1992, May). Videotape recording number T-869, Fortuno
Video Archive for Holocaust Testimonies,.New Haven, CT: Yale University.
Testimony of Judith G. (1992, September). Videotape recording number T-1879,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Julien E. (1995, January). Videotape recording number T-2742, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Karl S. (1980, April). Videotape recording number T-173, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.

386

Testimony of Kochevit P. (1979). Videotape recording number T-85, Fortuno Video

Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Leo G. (1980, March). Videotape recording number T-158, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Leo G. (1988, March). Videotape recording number T-977, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Leon H. (1985, November). Videotape recording number T-628, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Leon W. (1979, May). Videotape recording number T-2, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Martin S. (1986, January). Videotape recording number T-641, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Martin S. (1988, May). Videotape recording number T-1091, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Mary E. (1984, March). Videotape recording number T-260, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Meir V. (1992, December). Videotape recording number T-1892,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Menachem S. (1979, July). Videotape recording number T-152, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Michael G. (1992, August). Videotape recording number T-1880,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Myra L. (1984). Videotape recording number T-299, Fortuno Video
Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Nadia R. (1995, September). Videotape recording number T-3132,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Nina S. (1996, March). Videotape recording number T-3260, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Rena C. (1992, March). Videotape recording number T-2293,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Renee G. (1980, March). Videotape recording number T-5, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Renee G. (1988, March). Videotape recording number T-976, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Renee H. (1979, August). Videotape recording number T-50, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Robert K. (1984, May). Videotape recording number T-318, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Ruth A. (1994, October). Videotape recording number T-2678, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Sabina G. (1984, June). Videotape recording number T-287, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.

387

Testimony of Sabina S. (1985, November). Videotape recording number T-623,

Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Sally H. (1979, August). Videotape recording number T-3, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Sally H. (1989, March). Videotape recording number T-1154, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Samuel B. (1995, October). Videotape recording number T-618,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Susan S. (1991, November). Videotape recording number T-876,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Sybilla F. (1991, May). Videotape recording number T-1855, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Violet S. (1991, September). Videotape recording number T-1650,
Fortuno Video Archive for Holocaust Testimonies. New Haven, CT: Yale
University.
Testimony of Zezette L. (1980, July). Videotape recording number T-100, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Testimony of Zoltan G. (1979, August). Videotape recording number T-35, Fortuno
Video Archive for Holocaust Testimonies. New Haven, CT: Yale University.
Tolman, D. L., & Brydon-Miller, M. (Eds.). (2001). From subjects to subjectivities. New
York: New York University Press.
van der Kolk, B. A. (1994). The body keeps the score: Memory and the evolving psychobiology of posttraumatic stress. Harvard Review of Psychiatry, 1, 253265.
van der Kolk, B. A. (1996). Trauma and memory. In B. A. van der Kolk, A. C.
McFarlane, & L. Weisaeth (Eds.), Traumatic stress: The eects of overwhelming experience on mind, body, and society (pp. 279302). New York: Guilford.
van der Kolk, B. A., & Fisler, R. (1995). Dissociation and the fragmentary nature of
traumatic memories: Overview and exploratory study. Journal of Traumatic
Stress, 8, 505525.
Wagenaar, W., & Groeneweg, J. (1990). The memory of concentration camp survivors. Applied Cognitive Psychology, 4, 7787.
Wegner, D. M., & Schneider, D. J. (1989). Mental control: The war of the ghosts in
the machine. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought (pp. 287
305). New York: Guilford.
Wells, G. L., Malpass, R. S., Lindsay, R. C. L., Fisher, R. P., Turtle, J. W., & Fulero,
S. M. (2000). From the lab to the police station: A successful application of eyewitness research. American Psychologist, 55, 581598.
Wenzla, R. M. & Wegner, D. M. (2000). Thought suppression. Annual Review of
Psychology, 51, 5991.
Williams, L. M. (1994a). Recall of childhood trauma: A prospective study of womens
memories of child sexual abuse. Journal of Consulting and Clinical Psychology,
62, 11671176.
Williams, L. M. (1994b). What does it mean to forget child abuse? A reply to Loftus,

388

Garry, and Feldman (1994). Journal of Consulting and Clinical Psychology, 62,
11821186.
Winograd, E. (1992). Introduction. In E. Winograd & U. Neisser (Eds.) Aect and
accuracy in recall: Studies of ashbulb memories (pp. 15). New York: Cambridge
University Press.
Winograd, E. (1994). The authenticity and utility of memories. In U. Neisser &
R. Fivush (Eds.), The remembering self: Construction and accuracy in the selfnarrative (pp. 243251). New York: Cambridge University Press.
Winograd, E., & Neisser, U. (Eds.). (1992). Aect and accuracy in recall: Studies of
ashbulb memories. New York: Cambridge University Press.

389

This page intentionally left blank

abuse victims, physical

memory retrieval in, 133, 265,
333
See also child sexual abuse (CSA)
accessible memory
in episodic memory, 352353, 362,
381n.3
in Holocaust survivors, 360, 362,
365
accuracy-focused strategies, for aging
memory, 293294
adaptive numbness, in Holocaust
survivors, 358359, 379380
adrenaline. See epinephrine
adrenergic activity, in memory, 19,
104105, 108
adrenocorticotrophic hormone
(ACTH), release with stress, 80,
82
adult-guided reminiscing. See parentchild reminiscing
affect
emotional valence of, 4, 2124; in
elderly, 274275, 284287; visual
vs. verbal, 32, 279280, 286
with PTSD: as extreme, 144, 325;
isolation of, 95, 9799, 101,
111; negative vs. positive, 137
138
in re-experiencing emotion, 362,
382n.4
with schizophrenia, 221224

aging, emotional memory with, 272

295
affective focus and nonaffective
information, 284287
Alzheimers disease, 5960, 6364,
283284, 295n.1
deficits of, 59, 63, 82, 91, 272
emotional compensation hypothesis
of, 273274, 281282, 289, 294
295
flashbulb memories and, 287289
goal-directed emotional memory
hypothesis of, 272, 274275, 281,
286, 293294
as increasingly gratifying, 265n.2,
272273, 275, 289295
neurodegeneration in, 5960, 63
65
neutral vs. emotional information,
260264, 273, 275276, 281284
processing changes with, 272281;
arousal decreases, 275276;
neural mechanism maintenance,
276281; regulation effectiveness
increases, 273275; well-being
improvement, 272273
of shocking events, 287289
summary of predictions for, 281,
294
valence factor, 274275, 279, 281
287, 295n.2
allostasis, stress and, 7879

Alzheimers disease (AD)

amygdala role in, 5960
emotional memory and, 283284,
295n.1
hippocampus role in, 5960
neurodegenerative characteristics
of, 6364
ambiguity, role in emotional memory,

187

amino acids, excitatory, mediation of

stress, 45, 8991

amnesia
with amygdala damage, 4849
of child sexual abuse, 317319
forgetting as illusion of, 130136,
149

misremembering of, 2728

nonhuman primate model of, 4445
psychogenic, 134135
stress-induced (see forgetting)

amygdala, role in emotional memory, 12

autobiographical, 57, 5960
in elderly, 276277, 283, 286287
emotion influence, 105, 147
gender differences in, 6061
for gist vs. detail, 5256
in humans, 4761, 66
as integrative, 6568
isolation of affect and, 98100
lesion studies of, 4652, 66
neurodegenerative characteristics

of, 6465
neuroimaging of, 48, 5658, 61, 63
in nonhuman animals, 4447
in schizophrenia, 224
stress impact on, 96, 98100
amygdala lesions, 4652
disorders associated with, 4647
memory damage from, 66; bilateral,
4849; unilateral, 4952
anger

in Holocaust survivors, 363, 380

influence on memory, 2324, 30;
impact of accurate remembering,
3132
parent-child reminiscing about,
246249
animals
acute and chronic memory stressors
in, 8891
emotional memory in, 4447

anime, for stimuli emotionality study,

2122
antecedent-focused strategies, for

emotional control, 274

anxiety and anxiety disorders, 155180
defense mechanisms for, 9798, 102
encoding with, 156158, 161163,
167, 172173, 176177, 179180
implicit memory in, 171175,
180n.1

influence on memory: as bias, 93,

155, 157158, 161162, 176179;
for emotional information, 158
175; encoding tasks and, 179
180; intensity variable, 2526;
theoretical implications of, 175
180

memory deficits in, 155158

recall and: of autobiographical

memories, 158160; of
experimentally presented stimuli,

160167
recognition memory in, 167171
selective memory effects in, 155180
trait: high level, 158161, 167, 171
172, 175; low level, 171172; vs.
state, 156
trauma-induced, 97, 102103, 111
arousal, role in memory, 13, 36
in children, 254255, 260262
cognitive research on, 106107,
109, 113n.10
context of, 1820, 24, 4344
in elderly, 275276
in emotional continuum with

nonlinear effects, 43, 45, 54, 56,

106, 109
emotional valence vs., 2223, 54,
56, 146, 156
heightened with PTSD, 97, 99100,
146, 320321
as increasing function, 2426, 35
interruption of, 1213, 16, 20
intrinsic vs. extraneous, 1920, 24,
4344
as narrowing attention, 7, 1718, 35
physiological, 1820, 275
selective in anxiety disorders, 156
157, 161, 167
visual vs. thematic, 89, 1418, 35
36

artifact, in memory narrowing, 1013

atrocity(ies). See trauma and traumatic
memory
attention
affective focus of, by elderly, 284
287
avoidant, by abuse victims, 139141
control of: habit as detriment to,
200205, 210n.4n.5; impaired in
nonhabitual tasks, 195200,
210n.3; opposing habit through,
205209
narrowing of, 7, 1418, 35
selective, with anxiety disorders,
156157, 161, 167, 179180
stimulus salience and, 24
with traumatic events, 320;
sustained vs. dissociative, 323324
See also memory narrowing
attention magnets, 15, 17, 21
autobiographical memory, 243
amygdala and, 57, 5960
with anxiety disorders, 158160
depressed states and, 187, 189, 193
195; in Holocaust survivors, 364,
367, 380
detailedness and coherency of, 34,
30, 9497, 262, 264
dysregulation factors of, 59, 6364,
91, 112n.1
emotional development and, 244
245
parent-child reminiscing and, 243
244
PTSD impact on, 94, 97, 130, 136
137
in schizophrenia, 235
as unique process, 42, 6768, 95
autonomic nervous system, PTSD
impact on, 9799
aversive experiences. See negative
experiences
avoidance
in childrens traumatic memory,
255256, 260
in sexual abuse encoding, 139141
awareness
of previous trauma, in forgetting vs.
remembering, 132, 134, 148149
in schizophrenia (see conscious
awareness)

subjective: in normal controls, 230

232; prefrontal cortex role in,
101102; in schizophrenia, 219
221, 233
background details, in memory
categorization, 810
basolateral amygdala (BLA), emotional
memory role, 105
behavioral guilt, in Holocaust
survivors, 373
benzodiazepines, influence on
memory, 90
beta-blockers, impact on emotional
memory, 1213, 16, 20, 46, 105
106
bias. See memory bias
biopsychology, of trauma, 76113
definitions for, 7780
effects on brain, 8084
effects on forgetting: hippocampus
and, 94100; prefrontal cortex
and, 100111
effects on memory, 12, 2021, 45,
8494
isolation of affect with, 95, 98101,
111
overview of, 21, 76, 111112
taxonomy for, 7677
body memories, 97
body state changes, in schizophrenia,
221, 224
boundary extension, 10
brain
aging effects on, 280; amygdala,
276277, 283, 286287;
hemispheric asymmetry of,
279280; prefrontal cortex,
277279
effects of stress on, 8084;
forgetting as, 94111; memory
manifestations of, 8494
See also specific anatomy
brain imaging studies. See specific
imaging technique
brain lesion(s), impact on memory,
4652, 66
brain state changes
with aging, hemispheric asymmetry
of 279280, 221, 224225
in schizophrenia, 221, 224225

catecholamines, memory modulation

and, 6667, 97
categorization schemes
for emotional memory research, 8
10, 13, 108, 326
for relevant traumatic memory, 27
28
causal relations, of emotion and
memory vividness, 2930
in children, 263264
central details, of events (centrality)
aging and, 272, 281, 286
cognitive research on, 106111
in eyewitness memory, 311312; of
children, 330332
role in memory, 710, 16
valence factor of, 2122, 2829
child sexual abuse (CSA), memory
retrieval of
accuracy of, 331332
after repression, 317319
avoidant encoding impact on, 139
141
cognitive studies of, 136139, 143
dissociation impact on, 323325,
327
mechanisms of, 9193, 110, 130
132, 135, 316
reminiscing factor, 265, 318, 321
severity-disclosure correlation in,
316317
childrens memories, 242266
autobiographical, development of,
242244
cognitive functioning and, 327328,
330, 333
coping factors: developmental
aspects of, 244245, 328333;
gender socialization as, 245250,
265, 329; narrative skills as, 264
265; stressful memory and, 258,
262264
emotional development and, 244
245; in eyewitness memory, 327
333
of emotional events: co-construction
of, 242243, 248, 263264;
parent-child reminiscing about,
245250
of medical procedures, 252253,
260, 264265, 328, 330331

of negative vs. positive experiences,

242243, 246, 260264, 328
329; clinical disorders and, 330,
332333
research strategies for, 265266
of stressful and traumatic events,
250262; accuracy of, 241, 251
252, 330332; arousal role, 254
255, 260262; avoidance in,
255256, 260; cognitive
functioning and, 253, 258260,
264265; emotional processing
in, 245, 247250, 258260, 262
264; Hurricane Andrew, 250,
254260; interpersonal, 265;
interviewing strategies, 251, 263
264; parent-child reminiscing
about, 242243, 262264;
valence and, 242243, 260262
choice-supportive attributes, of
memory in elderly, 286, 292293
closed-ended questions, for
interviewing children, 251
co-construction, of emotional memory
in children, 242243, 248, 263
264
cognition and cognitive functioning
with anxiety disorders, 155158
attention control with: habit as
detriment to, 200205, 210n.4
n.5; impaired in nonhabitual
tasks, 195200, 210n.3; opposing
habit through, 204209, 366
in children, 327328, 330, 333
in elderly, 279280, 286287
emotional memory and, 155; in oral
testimonies, 347, 361, 381n.1; in
re-experiencing emotion, 361
363, 382n.4; research on, 106
111, 113n.10n.11
habitual: as control factor, 200
209, 210n.4n.5; in depressed
states, 45, 186188; in Holocaust
survivors, 364, 366, 369
in Holocaust survivors:
discontinuity of, 373374;
distraction strategies for, 374
378, 380381; habitual vs.
controlled, 364, 366, 369
intrusive, color-naming paradigm
for, 141142, 165

in schizophrenia, 217, 221, 223,

235n.1

traumatic memory and, 93, 9799,

362; in children, 253, 258260;
control factor, 79, 84; research
on, 136139, 143
coherence. See memory coherence
color-naming paradigm, for intrusive

cognition, 141142, 165

commission errors, in traumatic

memory, 86, 100, 110, 317

computerized tomography (CT), of

amygdala, 48
concentration ability, with depression,
196197
concentration camps. See Holocaust
survivors
conscious awareness, in schizophrenia

emotional influence on, 218, 221,

230232; normal controls of,
230232
recognition memory with

subjective, 219221, 232233

research implications, 233235
conscious recollection, 230232, 234
in Holocaust survivors, 354355,
365, 368
consistent memory, in Holocaust

survivors, 349350, 355356

context

of arousal stimulus, 1820, 24;

affective processing by elderly,

284287; neural mechanism links

to, 4344, 67
in episodic memory, 77; stress
impact on, 9497, 99, 103, 110
111, 112n.5
control and controllability

of cognitive attention: habit as

detriment to, 200205, 210n.4

n.5; impaired in nonhabitual
tasks, 195200, 210n.3; opposing
habit through, 204209, 366
of stress, 79, 84, 101
coping mechanism(s)

in children: developmental aspects

of, 244245, 328333; gender

socialization impact on, 245250,
265, 329; narrative skills impact
on, 264265; stressful memory
and, 258, 262264

for threats: avoidant encoding as,

139141; selective attention as,

156158, 161, 167
core memory, 109
in episodic memory, 352353, 381n.3
in Holocaust survivors, 353354; reexperiencing emotions and, 362
365, 382n.4
corticosterone, memory modulation

and, 4546, 85
in animals, 8889
corticotropin-releasing factor (CRF),

release with stress, 80, 82

cortisol level

in emotional memory, 105, 109

stress impact on, 82, 8587, 112n.3;
memory coherence and, 9697,
101, 147
cortisone

anxiety disorders with, 97

impact on memory retrieval, 8688
crime, influence on memory

in children, 251, 261

legal system and, 132133, 309,
312315, 318319, 324, 334
in perpetrators, 135
prefrontal cortex role, 278
See also eyewitness memory

crying, by Holocaust survivors

re-experiencing during recall, 362

365, 382n.4
re-experiencing with external cues,

368, 379
recall of, 357358
culture. See social-cultural theory
Cushings syndrome, stress hormones
and, 82, 84, 101, 112n.3
cytokines, proinflammatory, influence

on memory, 90
date rape, 1516
decision making, emotional

prefrontal cortex role in, 62

with PTSD dissociative tendencies,

325

declarative memory
as emotional memory, 4344
neural processing mechanisms for,
4243, 45; as integrated, 6568
deep memory, in episodic memory,
352, 381n.3

Deese-Roediger-McDermott (DRM)
paradigm, for stress effect on false
memory, 109110, 143144, 327
defense mechanisms, for memory with
anxiety, 9798, 102
delusions, role of emotional processing
in. See schizophrenia
dementia. See Alzheimers disease (AD)
dendrites, stress-induced atrophy of,
81, 90, 99
prevention of, 89, 91
dentate gyrus, effects of stress on, 81
82
depression and depressed mood, 186
210
in elderly, 272273
habits of memory with, 186, 188
195; in autobiographical
experience, 193195, 364, 367,
380; congruency of, 191193
habits of thought with, 186188
influence on memory, 29, 93, 101,
186
memory as object in, 189, 195;
appropriate use of, 197200,
209n.2; effortful construction of,
196197; for mood-congruency,
191193
mood-congruent recall with, 189
193, 209n.1
mood-congruent suppression with,
201203, 208, 210n.5
negative thinking associated with,
186188, 192
with PTSD, 137138
self-control with: as impaired in

nonhabitual tasks, 195205,

210n.3n.5; opposing habit
through, 205209
detailedness, of memory. See memory
detail
details, of events
background, 810
central (see central details)
peripheral (see peripheral details)
dexamethasone, impact on memory
retrieval, 8586, 110
directed-forgetting paradigm
for anxiety disorders, 164
for depressed states, 202203,
210n.5

for trauma-related material, 139

141
disasters. See natural disasters;
shocking events
disengagement, impaired, in depressed
states, 201205, 210n.4n.5
dissociation
of memory tasks in schizophrenia,
218219, 221
with traumatic memory, 28, 129
130, 132133, 135; avoidant
encoding with, 139141; avoidant
recall for, 255; in children, 330,
332333; cognitive studies of,
136139, 143; false-memory
formation and, 318319, 324
326; semantic confusion about,
148149; in witnesses and
victims, 317, 319, 323324, 330
distinctiveness, role in memory
construction, 11
distraction
for emotional memory in Holocaust
survivors, 374378; as getting
worse, 375377; importance of
meaningful, 374375, 380381;
rebound effects of, 377378
for habitual thoughts, 188, 196
197, 208
distress. See stress
doctor/mechanic stimuli, memory
construction and, 6, 12, 19, 105
domain-specific material, memory
system for, 43, 52
dopamine
influence on memory, 8990
in schizophrenic state, 224225
dual representation theory, of
episodic memory, 352353,
356, 381n.3
durability, of emotional memory. See
long-term memory
dysphoria, 186
habits of memory with, 186, 188
195, 209n.1
habits of thought with, 187188
self-control with: as impaired in
nonhabitual tasks, 195205,
210n.3n.5; opposing habit
through, 205209
dysthymia, 273

Easterbrook hypothesis, of arousal and

memory, 7, 14, 1618, 35
aging and, 275276
elaboration hypothesis
in parent-child reminiscing:
autobiographical memory and,
243244, 246248, 262; impact
on coping, 258, 262264, 318
of threats and anxiety disorders,

174, 176178
See also memory enhancement
elderlys memories. See aging
emotion and emotionality
in eyewitness memory, 308315, 334
in Holocaust survivors, 351352,
354359, 381; with liberation,
357; re-living of, 352, 355, 361
369, 382n.4
impact on memory: biological, 12,
2021; in children, 327329;
detailedness and coherency in,
2830, 9497, 112n.5, 264; in
elderly, 281284; intensely
negative variable, 2628;
intensity factor, 4, 2426, 3233;
as positive promotion, 78, 10, 33,
35; prefrontal cortex role in, 100
102; reaction assessment in, 30
33; in schizophrenia, 222223,
225; socialization factor, 244
245; source factor, 11, 1315, 282,
314; type factor, 1820, 97, 111,
112n.6; as undermining, 10, 35;
valence factor, 2024, 28, 54,
146, 156, 203; visual vs. thematic,
89, 1418, 3536
emotion-criticizing style, 248, 250
emotion-explaining style, 248
emotion-focused strategies, of elderly
impact on nonaffective information,
284287
as increasingly gratifying, 265n.2,
272273, 275, 289295
for interpersonal problem-solving,
274275
emotion-induced recognition bias,
314315, 325
emotional blunting
in schizophrenia, 222223; word
free-recall investigation of, 227
230

stress-induced, 97100, 102; in

Holocaust survivors, 357358,
379380
emotional compensation hypothesis,
for aging, 273274, 281282,
289, 294295
emotional control. See self-regulation
emotional development, in children,
244245
eyewitness memory and, 327333
emotional disturbances, in
schizophrenia, 221225
of affect, 221224
of body state, 221, 224
of brain state, 221, 224225
cognitive evaluation of, 217, 221,
223, 235n.1
research implications, 225
emotional goals, with aging, 272,
274275, 281, 286, 293
294
emotional masks, of Holocaust
survivors, 380
emotional memory
accuracy of, 413, 3536;
detailedness and coherency vs.,
2830, 9497, 112n.5, 262
anxiety disorders and, 155175
in children, 242266
in depression, 186210
durability of (see long-term memory)
in elderly, 272295
eyewitness, 308334
in Holocaust survivors, 347382
literature review of, 103111
neuroanatomy of, 12, 4268
in nonhuman animals, 4447
research perspectives for (see
research)

in schizophrenia, 217236
traumatic stress impact on, 98
100, 112n.6113n.8 (see also
stress; trauma and traumatic
memory)
emotional numbing/numbness
in Holocaust survivors, 357358,
379; as adaptive, 358359, 379
380
with PTSD, 97100
emotional pain, in Holocaust
survivors, 365366, 381

emotional processing
by children, 258260; of negative
experiences, 245, 247250, 258
260
by elderly, 272281; arousal
decrease, 275276; neural
mechanisms for, 276281;
regulation effectiveness increase,
273275; well-being
improvement, 272273
in schizophrenia, 217, 221, 223,
235n.1

emotional reaction(s)

as extreme, with traumatic memory

retrieval, 144, 146148

inhibition of: frontal cortex role in,

100102; during stress, 79, 84,

101

remembering: accurate assessment

with, 3133; influence on
memory, 3031
emotional scenes, as threatening, 156
157

emotional self-concept

in children, development of, 245,

248, 250
in elderly, 273275
in Holocaust survivors, 367, 371,
373, 378380
emotional stimuli/events, memory for,

336
with amygdala damage, 4853
arousal continuum with nonlinear

effects, 43, 45, 54, 56, 106, 109

in children, 327333
cognitive research on, 106111,
113n.10

construction of, 6, 810

declarative, 4344
in elderly, 273, 275276, 281283
intrinsic vs. extraneous, 1820, 24,
4344
neural processing in: in humans,

4763; as integrative, 6568, 76

77; in nonhuman animals, 4447
parent-child reminiscing about,

245250, 318
prefrontal cortex role, 6263
PTSD and dissociation impact on, 327
Sternberg task for evaluation of,

217, 223, 235n.1

emotional valence
in children, 260264, 281283
in elderly, 274275, 279, 281287,
295n.2

in eyewitness memory, 309311,

321

impact on memory, 2024, 2829,

54, 146, 156, 203
in schizophrenia, 222223, 225
230, 236n.2
in survivors (see Holocaust

survivors)

emotional well-being
in elderly, 272273, 294295
in Holocaust survivors, 355356
emotionality effect
in memory, 4, 2426, 3233
in schizophrenia, 217, 227230;
conscious awareness and, 232
233; research implications, 233
235

encoding, of memory

with aging, 291292, 294

in amnesia confirmation, 131132,
135136, 149
amygdala role in, 66, 147
arousal impact on, 106107
avoidant style for coping, 139141
as flawed in Holocaust survivors,

348, 353, 381n.2

initial, 34, 56, 60
integrative modulation of, 6568,
7677
prefrontal cortex role in, 63, 100
in schizophrenia, 219, 226
as selective in anxiety disorders,

156158, 161163, 167, 172173,

176177, 179180
stress impact on, 88, 9497, 99,
109, 139, 149, 323
epilepsy, intractable, temporal

lobectomy for, 48
epinephrine, memory modulation and,

19, 4546, 66
episode(s), basic elements of

remembering, 352
episodic memory, 77
cognitive research on, 106111
disruption of: with PTSD, 9899,
102103, 111, 113n.8; in
schizophrenia, 218220

hippocampal function and, 80

retrieval process of, 95; in Holocaust
survivors, 134, 146, 356361,
365366
stress impact on, 84, 111; acute, 85
88; chronic, 9194, 99, 112n.4
in traumatic memory, 352, 362
event material(s), in emotional
memory

ambiguity of, 187

comparisons across studies, 89
role of nature of, 4, 67
trauma-related, 139141
exercise, influence on memory, 1920
explicit memory, 77
dissociation of in schizophrenia,
218219
in habits of thought, 189, 195;
appropriate use of, 197200,
209n.2; effortful construction of,
196197
hippocampal function and, 80
stress impact on, 147148, 171, 175
expression, of emotions, parent-child
reminiscing impact on, 245250
external arousal/cues
in episodic memory, 353; of Holocaust
survivors, 361, 366369
in memory construction, 1820, 24
rumination and, 187, 364, 376
extraneous emotion, in memory
construction, 1820, 24, 4344
eyewitness memory, 308334
in children, 251252, 261262;
accuracy of, 328329; age as
factor, 327328; clinical disorders
and, 332333; of medical
procedures, 252253, 260, 264
265, 328, 330331; of sexual
abuse, 9193, 110, 130132, 135,
265, 316319, 331332; for
traumatic events, 330332
of crime, 7, 107, 135, 251, 261, 278
developmental considerations of,
327333
emotion and, 308315, 334;
accuracy of negative events, 308
311, 328329; biases in, 313315,
325, 329; central vs. peripheral
details in, 311312; long-term
durability of, 312313

as false, with dissociation, 318319,

324326
legal system impact on, 132133,
309, 312315, 319, 324, 334
psychopathology implications for,
319327; in children, 330, 332
333; dissociative tendencies and,
318319, 323326, 330, 332;
PTSD and, 319324, 332
research strategies for, 308309,
315, 319, 327, 333
survivor guilt with, 372
for traumatic events, 315319; in
children, 330332; repression and
recovery of, 317319; special
mechanisms in, 316317
fading
aging and, 289291, 295
of mishaps and triumphs, 3132,
3435
false alarms, memory biases in, 314
315, 325
false memory
of emotional vs. neutral events, 33
34, 109111
in schizophrenia, 220221
in space alien abductees, 144146
of traumatic events, 142143;
cognitive studies on, 136139,
143; with dissociative tendencies,
318319, 324326
fear
in Holocaust survivors, 356367;
associated emotions of, 358360;
re-experiencing with external
cues, 366369
parent-child reminiscing about,
246249
trauma-induced, 102103
See also phobia(s)

fear conditioning, classical, 4445

feelings, toward emotional event
accuracy in remembering, 45, 10,
20, 3031, 103
influence on memory, 3132
flashbacks, trauma-induced, 4, 94,
102103
flashbulb memories
accuracy of, 3435, 42, 289
aging and, 287289

flashbulb memories (continued)

consequentiality of, 35
in Holocaust survivors, 352, 381n.3
role of emotion in, 56, 32, 111
specialized production process for,

34, 42
forebrain

role in emotional memory, 63, 65, 67

see also prefrontal cortex (PFC)

forensic context, of emotional memory

emotion-induced recognition bias

in, 314315, 325

eyewitness, 251, 256, 258, 308309,
311, 321
false-memory in, 315, 324, 326
forgetting

difficulty of: with depressed states,

186, 202203; for trauma-related

material, 2627, 94100, 139
140, 164, 210n.5
in schizophrenia, 227, 233
stress-induced, 100111; cognitive
approach to, 106111; emotion
and memory literature on, 103
111; frontal cortex models of,
100103; learning/
neuroscientific approach to, 104
106

of traumatic events, 129149;

amnesia vs., 130136, 149, 318;
cognitive research on, 136139,
143; complex mechanisms of, 26
27, 94100, 164, 210n.5;
dispelling semantic confusion,

148149; distortion with, 141

143; evidence adduced for, 130
136; flashbacks and, 4, 94,
102103; of material related to,
139141; recovery mechanisms
of, 2728; repression vs., 361;
sexual abuse and, 317319; in
space alien abductees, 144146;
types of memory in, 94, 96, 111,
112n.5113n.7
fragmentation, of memory

emotion influence on, 103104

stress-induced, 9495, 102103, 111
functional relevance. See plot-relevant
items
funny memories. See positive
experiences

fuzzy trace theory, 109

of episodic memory, 353
GABA, mediation of stress on memory,
45, 8990
gender differences, in memory

amygdala activity and, 6061

in children, 244250, 265, 329
in elderly, 283
generalized anxiety disorder (GAD)

selective memory in, 158159, 161

163, 168, 172173, 176
trauma-induced, 97, 111
gist information

amygdala and memory of, 5256

cognitive research on, 107111,
113n.10n.11, 143
emotional valence of, 2, 108, 143,
146, 156; in elderly, 284287
in episodic memory, 353, 381n.3
gender differences and, 61
in memory categorization, 810, 13,
108, 326
glucocorticoid-administration studies

in animals, 8889, 91
in humans, 8587; limitations of,
8788
glucocorticoid antagonists, impact on

memory, 85
glucocorticoids

hippocampal function and:

disruption of, 8081, 9596;

evidence for volume reduction,

8283; potentiation of, 82, 90

as indicator of stress, 7980
memory modulation and: in

animals, 8891; emotion

influence on, 108109; in
humans, 6667, 8586, 99
prefrontal cortex function and, 101
glucose

influence on memory: amygdala

metabolism, 56; serum level, 19,

25, 46
influence on mental health, 278
glutamate, mediation of stress on

memory, 45, 8991

goal-directed emotional memory

hypothesis, for aging, 272, 274

275, 281, 286, 293294
granule cells, effects of stress on, 81

gratification, emotional, as increasing

with aging, 265n.2, 272273,
275, 289295
grief
accurate recall of, 31
in Holocaust survivors (see crying)
guided imagery protocol
memory distortion with, 142143
for memory distraction, 377
guilt, in Holocaust survivors
behavioral, 373
survivor, 369373
habits of memory
in depressed states, 188195, 364
in Holocaust survivors, 364, 376;
distraction strategies for, 366,
374378, 380381
self-control with: as detriment to,
200205, 210n.4n.5; as strategy
to oppose, 205209, 366
habits of thought
cognitive control for, 205209; in
Holocaust survivors, 366, 374
378; as impaired in nonhabitual
tasks, 195205, 209n.2210n.3
as depression feature, 187188
influence on memory, 45, 186
hallucinations
neurobiology of, 102103
in space alien abductees, 145
happiness
in Holocaust survivors, 350, 358,
376377
parent-child reminiscing about, 246
recall of: with amygdala damage,
5053; vividness vs. accuracy of,
45, 10, 20, 103
See also positive experiences
hatred, in Holocaust survivors, 359,
374
hemispheric asymmetry, of emotion,
with aging, 279280
herpes simplex encephalitis, 47
hippocampus
effects of stress on, 8082; evidence
for atrophy, 8284, 112n.2;
fragmentation and, 9597;
glucocorticoid-administration
studies of, 8589, 91; hormonal
mediation of, 8081, 8891

role in emotional memory: in

elderly, 276277; in humans, 61
62, 147; as integrative, 6568, 77;
isolation of affect and, 98100;
neurodegenerative characteristics
of, 6465; in nonhuman animals,
44; stress impact on, 98100
Holocaust survivors, emotional
memory in, 347382
destiny and, 371
distraction for, 374378; as getting
worse, 375377; importance of
meaningful, 374375, 380381;
rebound effects of, 377378
as enduring: guilt, 369373; as
result of cognitive discontinuity,
373374
impact of sharing vs. not sharing,
366, 380381
physiological arousal with, 355,
361362, 368369, 375, 382n.4
qualitative study of, 347381;
accuracy concepts, 348350,
381n.2; conclusions from, 378
381; of distraction, 374378; of
endurance, 369374, 378;
methodology, 347350, 381n.1; of
re-experiencing, 352, 355, 361
369, 382n.4; of recall, 356361;
testimony excerpts, 350352;
trauma, 352356, 381n.3
re-experiencing emotion through,
361369; due to external cues,
366369; examples of, 352, 355;
multiple systems for, 361362;
during recall, 362366, 382n.4;
in repeated testimony, 364366
recall of, 134, 146, 356361, 381;
associated emotions, 358359;
events separate from emotions,
359360; fear, 356357, 360,
366369; forgetting vs., 360361;
numbness with, 357359, 379
380; re-experiencing emotion
during, 362366, 382n.4
repression of, 360361
self-concept with, 367, 371, 373,
378380
homeostasis, stress and, 78
hope, emotional memory for
generation of, 36

hopelessness, 29
hormonal modulation, of memory
system
emotions and, 104106
in humans, 6667, 8081
in nonhuman animals, 4546, 8891
horrific events. See trauma and
traumatic memory
humor

accurate recall of, 31

influence on memory, 21
Hurricane Andrew, childrens recall
of, 250, 254260
hydrocortisone, impact on memory
retrieval, 8688
hyperarousal, in Holocaust survivors,
with re-living emotional memory,
355, 361362, 368369, 382n.4
hypermnesia, with trauma, 102
hypervigilance, with PTSD, 97, 99
100, 321
hypothalamic-pituitary-adrenal (HPA)

axis

cognitive research on, 109, 113n.10

impact of aging on, 91
stressor activation of, 7980, 89
imagery protocol

guided: memory distortion with, 142

143; for memory distraction, 377
script-driven, for space alien

abductees, 145146
imagination inflation paradigm, for

memory distortion, 142143, 352

imaging studies, of brain. See specific

imaging technique

implicit memory, 77
anxiety disorders and, 171175,
180n.1

dissociation of in schizophrenia,

218219
stress impact on, 147148, 362
incomplete memory, in Holocaust

survivors, 349
inferences

anxiety disorders and ability to

draw, 156, 321

schema-based: for memory recall in

elderly, 284287; for memory

reconstruction, 34, 108, 110111,
113n.11, 194

inhibition

of emotional response, frontal cortex

role in, 100102

impaired, in depressed states, 200
205, 210n.4n.5
of traumatic memory recovery, 148
149

intensity factor, of emotional memory,

4, 2426, 3233
intentionality

of forgetting, 139140, 164, 202

203, 210n.5
in habits of thought, 187, 189, 195,
197

interleukin-I (IL-I), influence on

memory, 90
interpersonal problems, emotionfocused strategies for solving,
274275
interpersonal violence, childrens

recall of, 130, 265

interpretative memory, in Holocaust

survivors, 376, 379, 381

interviewing strategies

for childrens memories, 251, 263

264

for forensics, 314, 326

intrinsic emotion, in memory

construction, 1820, 4344

intrusion errors, in traumatic memory,

30, 34, 110, 166, 321, 353

intrusive recollection, of traumatic

material, 140142, 166, 321,

353

inverted-U function, of memory and

stress. See Yerkes-Dodson law

ironic-process theory, 208

for memory distraction, 377378
Jewish persecution. See Holocaust
survivors
laboratory studies

of emotional memory, 67;

induction methods, 1518
of traumatic memory, 136139, 355
language, as childrens trauma recall

factor, 253, 258260, 264

impact on coping, 264265
lateralization, of emotional cognition,

with aging, 279280

learning theory
of habits of thought, 187, 189, 195,
197, 208
of stress-induced forgetting, 104
106
legal system, influence on traumatic
memory construction, 132133,
309, 312315, 318319, 324, 334
lipoid proteinosis, 47
long-term memory, emotional
dissociation impact on, 323327
eyewitness, 312313
in Holocaust survivors, 352, 355
356, 378
mechanisms of, 5960, 6365, 82,
90, 157
with PTSD, 322323
long-term potentiation (LTP), as
hippocampal response, 82, 90
lost memories, 2628
magnetic resonance imaging (MRI)
of amygdala, 48, 5658, 61, 277, 284
of hippocampal atrophy, 8283,
112n.2, 277
of prefrontal cortex, 63, 100
for research on aging, 284, 291
malingering, 135
material, for memories. See event
material(s)
media, influence of, 5
medial temporal lobe memory system
in humans, 4243, 4761; cool vs.
hot, 98100
integrative mechanisms of, 6568, 77
neurodegenerative characteristics
of, 6465
in nonhuman animals, 4447
medical procedures, childrens
memory for, 328, 330331
recall of, 252253, 260, 264265
memory
age-related decline in, 59, 63, 82, 91
effects of stress on (see stress; trauma
and traumatic memory)
as object, 189, 195; appropriate use
of, 197200, 209n.2; effortful
construction of, 196197; in
mood-congruency, 191193; in
mood-suppression, 201202, 208,
210n.5

physiological models of, 1820, 45,

82, 107, 275
taxonomy for nature of, 7677
memory accuracy
in children: of child sexual abuse,
331332; of stressful events, 241,
251252, 330332
of emotional events, 413, 21, 35
36; detailedness and coherency
vs., 2830, 9497, 112n.5, 262; in
feelings toward, 45, 10, 20, 30
31, 103
in eyewitness memory, 308311,
328329
in flashbulb memories, 3435, 42,
289, 352
of happiness vs. sadness, 45, 10,
20, 103
in Holocaust survivors, 348350,
381n.2
of personal vs. public events, 45,
13, 31, 33
strategies for, in aging memory,
293294
in traumatic memory, 45, 3435,
42, 289
vividness vs., with negative
experiences, 5, 308311
memory assessment
role in emotions, 3033
white-noise method of, 173174
memory bias
anxiety-related, 155, 157158;
encoding tasks contribution to,
179180; with experimentally
presented stimuli, 160167;
limited evidence for, 176178;
with panic, 178179
decision making and, 62, 325
326
in depressed states, 186195, 201
in elderly, 275, 289294
in eyewitness memory, 313315,
325, 329
research coding schemes, 15
memory coherence, 2830
autobiographical, 34, 30, 9497,
262, 264
stress impact on, 9497, 101,
112n.5, 112n.7, 147, 262,
264

memory-consolidation, postevent

glucocorticoid impact on, 88, 105

neural mechanisms for, 44, 49, 67
68

physiological arousal impact on, 20,

45, 107
memory construction
emotional context of, 45
role in perpetuating schizophrenia,
217218, 221, 230233
role of scrutiny in, 1112
memory deficits

in anxiety disorders, 155158

in depressed states, 196, 198,
209n.2; self-control for
improving, 205209
in elderly, 59, 63, 82, 91, 272
in schizophrenia, 218221;
subjective states of, 219221,
233; task dissociation, 218219,
221

with stress: acute, 8588; in

animals, 4647, 8891; chronic,
9194, 99, 112n.4; correlation to
severity, 317, 322; frontal cortex
role in, 100103; in humans, 66
67, 112n.4n.5, 113n.9n.11;
mnemonic, 321323; paradoxical
with PTSD, 9294, 9899, 110
111, 129136; as selective vs.
general, 9394
with trauma, 12, 2021, 45;
biopsychology of, 8494, 102
103, 112n.1n.3, 112n.6n.8
See also forgetting
memory detail

amygdala role in, 52, 5456

background, 810
coherence with, 2830, 262, 264;
autobiographical, 34, 30, 94
97, 262, 264; of emotional
events, 2830, 9497, 112n.5,
262

emotion influence on, 2830, 109

110

gender differences in, 61, 244250,

265, 329
stress impact on, 7, 9497, 112n.5
memory distortion, with trauma, 142
143, 352353

memory enhancement

in anxiety disorders, 156157

human neural mechanisms for, 43,
6162; hormonal, 8081; as
integrative, 6568, 106, 109; with
PTSD, 99100
with PTSD, 320323
See also elaboration hypothesis

memory errors
in episodic memory, 352353
of traumatic events: in children,
251252; emotion influence on,
106107, 109, 111; initial
encoding of, 348349, 353,
381n.2; intrusion, 30, 34, 36, 110,
166, 321, 353; omission vs.
commission, 86, 100, 110, 317;
reconstruction, 5, 2728, 3536,
95

memory narrowing

arousal role in, 7, 1518, 35

as artifactual, 1013
emotion source impact on, 11, 1315
emotion valence impact on, 2122,
203

studies of, 710

memory organization, temporal

framework for, 810, 96

See also spatio-temporal memory

memory reconstruction

schema-based inferences for, 34,

108, 110111, 113n.11; with
aging, 284287; with depression,
194

of traumatic events, errors in, 5, 27

28, 3536, 95
memory recovery. See recall and

recollection

memory rehearsal
with emotional events, 107108
of flashbulb events, 3435
promotion of retention, 45, 376
377

with traumatic event, 320, 376

memory suppression

as emotional control, 274

for Holocaust survivors, 374378; as
getting worse, 375377;
importance of meaningful, 374
375; rebound effects of, 377378

mental disorders. See psychopathology;

specific disorder
metamemory, in Holocaust survivors,
349
mineralization, in amygdala lesions,
4748
mineralocorticoids, in memory
modulation, 101
misleading questions, for interviewing
children, 251
mood-congruency, with depression
indirect tests of, 191193, 198,
209n.1
memory suppression and, 201203,
208, 210n.5
recall and, 186, 189193; induction
of, 190, 209n.1
mood disorders
habits of memory with, 186, 188
195
habits of thought with, 187188
influence on memory, 93, 156, 186,
273
narrative memory, 109, 146
in episodic memory of Holocaust,
352354, 364, 366, 380381
narrative skill, in children
impact on coping, 264265
as trauma recall factor, 253, 258
260
narrative smoothing, 109111
narrowing of attention, 7, 1418, 35
See also memory narrowing
natural disasters
recall by children, 250, 254260,
265; parent-child reminiscing
impact on, 262264
recall by elderly, 283284
negative experiences
amnesia for, 2628
biological effects of, 21
cognitive research on, 106111,
136139, 143
emotional resolution of, by children,
245, 247250, 258260
emotional valence of, 2224, 29, 54,
56, 203; in children, 242243,
246, 260264, 328329, 332
333; conscious awareness and,

230232; in elderly, 273275,

289291, 295, 295n.1; in
schizophrenia, 222223, 225,
227230, 233
eyewitness memory accuracy of,
308311
fading memory of, 3132; in elderly,
289291, 295
habitual memory of, 188195
parental-child reminiscing about,
242243, 246250
recall of, 103104, 106; with
amygdala damage, 5053; with
anxiety disorders, 156167, 170
171; by children, 250262;
vividness vs. accuracy of, 45, 10,
20, 308311
See also stress; trauma and

traumatic memory

negative-priming paradigm, of
inhibition in depressed states,
201202, 210n.4
negative thinking, associated with
depression, 186188, 192
neural modulation, of memory system
cool vs. hot, 98100
in elderly, 276281
emotions and, 104106, 109,
113n.10, 147
in humans, 43, 6162, 100
integrative mechanisms of, 6568,
7677
in nonhuman animals, 4547
neuroanatomy, of emotional memory,
12, 4268
cool vs. hot systems in, 98100
fragmentation role, 9597, 113n.7
n.8
gender differences in, 6061
human amygdala in, 4761
human hippocampus in, 6162
impact of aging on, 276280
integrative systems, 6568, 7677
literature review of, 103111
neurodegeneration and, 6365
in nonhuman animals, 4447
prefrontal cortex in, 44, 6163
processing systems, 4244
stress impact on, 21, 4547, 7879,
8285, 112n.3, 316, 323324

neurobiology

of isolation of affect, 95, 98101, 111

of trauma-induced flashbacks, 102
103

neurodegeneration
amygdala and, 5960
emotional memory and, 6365
hippocampus and, 5960
neurogenesis, stress-induced, 81
neuroimaging

in emotional memory research: of

amygdala, 48, 5658, 61; of

prefrontal cortex, 63
See also specific imaging technique

neurons, effects of stress on, 8182

neuropsychological evaluation

of amygdala: for gist vs. detail

memory, 52, 5456; with lesion

damage, 4853, 66;
neuroimaging correlation to, 56
58, 61, 63
for emotional memory, 104106
for PTSD, 9294
neuroscientific research, on emotion
and memory, 104106
neutral stimuli/events
emotional valence of, 2123, 30,
107108, 203; in children, 242
243, 246, 260264, 328329,
332333; conscious awareness
and, 230233; in elderly, 260
264, 273, 275276, 281284; in
Holocaust survivors, 376377
eyewitness memory accuracy of,
308311
false memories of, 3334, 109
memory construction of, 6, 89, 11
recall with amygdala damage, 48
53, 59
NMDA receptor blockade, for

prevention of dendrite atrophy,

91

nonaffective information, affective

processing by elderly, 284287
nondeclarative memory, neural
processing mechanism for, 4243
in animals, 4445
nonemotional events. See neutral
stimuli/events
nonhabitual tasks, impaired cognitive
control in, 195200, 210n.3

noradrenergic neurotransmitters,

memory modulation and, 4546,

67, 89
norepinephrine, influence on memory,

19, 89, 96, 101

emotion and, 105
numbness. See emotional blunting;

emotional numbing/numbness

obsessive compulsive disorder (OCD),

selective memory in, 158, 164

165, 169, 173174, 178
omission errors, in traumatic memory,

86, 100, 110, 317

open-ended questions

for forensics, 314

for interviewing children, 251, 263
oral testimony(ies)
of Holocaust survivors, qualitative
study of, 347381, 381n.2
382n.4
traumatic memory characteristics
of, 352356, 381n.3
of war survivors, 347, 381n.1
orbitofrontal cortex

emotional memory and, 6263, 65

impact of aging on, 278280
original phenomenal experience, in
episodic memory, 352353,
381n.3
overgeneral memories, with PTSD,

326

panic disorder

memory bias in, 158, 167, 178179

retrieval of threatening information

in, 165166, 170, 174

parent-child reminiscing
about emotional events, 245250
about stressful events, 242243,
262264, 318
autobiographical memory and, 243
244

style vs. content in, 244, 247, 262

264

parents

as abusers, 130, 265

as co-constructors of childrens

memory, 242243, 248, 263264

Pavlovian conditioning, classical, 44
45

perceptual encoding hypothesis, of

threats and anxiety disorders,
176178
perceptual priming, for memory in
schizophrenia, 218220, 223
peripheral details, of events
cognitive research on, 106111, 286
in eyewitness memory, 311312; of
children, 330332
role in memory, 710, 16
valence factor of, 2122, 28
perpetrators, of crime
memory retrieval in, 135
parents as, 130, 265
personal events, memory accuracy of,
4
personality profile, in traumatic
memory retrieval, 137139
phenytoin (Dilantin), for prevention of
dendrite atrophy, 91
phobia(s)
social, selective memory in, 158
159, 163165, 168169, 173
spider, memory research on, 1011
physical abuse victims, memory
retrieval in, 133, 265, 333
physiological arousal
as decreasing with aging, 275276
emotion intensity variable, 2526
intensely negative variable, 2628
in memory construction, 1820, 45,
82, 107, 275
with re-experiencing emotion in
Holocaust survivors, 355, 361
362, 368369, 375, 382n.4
with stress, 79, 98100, 102
pleasant memories. See positive
experiences
plot-relevant items
emotional valence of, 21, 23
in memory categorization, 89, 107
Pollyanna tendency, in schizophrenia,
217, 226227, 232, 234, 236n.2
positive experiences
accurate remembering of feelings
about, 45, 10, 20, 31, 103
biological effects of, 21
emotional valence of, 2124, 29, 54,
56, 203; in children, 242243,
246, 260264, 328329, 332
333; conscious awareness and,

230232; in elderly, 273275,

289295; in schizophrenia, 217,
222233, 236n.2
eyewitness memory accuracy of,
308311
fading memory of, 3132
parent-child reminiscing about,

246

positron emission tomography (PET)

of aging brain, 276277
research applications of, 5657, 100,
136
post-traumatic stress disorder (PTSD)
affect with: as extreme, 144, 146
148; isolation of, 9799, 101, 111;
negative vs. positive, 137138
cognitive studies of, 136139
diagnostic criteria for, 9192, 135,
320
eyewitness memory and, 319323;
with dissociation, 318319, 324
326
false memory in, 136139, 142146,
318319, 324326
heightened arousal with, 97100,
146, 320321
hippocampal atrophy with, 8384,
112n.2
material-related to: forgetting, 139
141; intrusion of, 140142, 166,
321, 353
memory deficits with, 9294, 110
111, 129136; in children, 332
333; correlation to severity, 317,
322; dissociative tendencies
impact on, 318319, 323327;
mnemonic, 321323; as
paradoxical, 9899
memory enhancement with, 320
323
prevalence rates of, 92
selective memory in, 158160, 164
165, 169, 173174
stress hormone levels with, 84,
112n.3
potentiation, stress-induced forms of,
82, 90
predictions, of future, emotional
memory for, 36
prednisone, impact on memory
retrieval, 8586, 110

prefrontal cortex (PFC)

role in emotional memory: in
elderly, 277279; in humans, 61
63, 100; as integrative, 6568;
isolation of affect and, 98101; in
nonhuman animals, 44; stress
impact on, 100111
stress-induced forgetting in, 100
111; cognitive approach to, 106
111; emotion and memory
literature on, 103111; learning/
neuroscientific approach to,
104106; theoretical models of,
100102; trauma-related, 102
103
primed burst potentiation (PBP), by
hippocampus, 82
problem solving
by elderly, 274275, 279280
emotional memory for generation of,
30, 36
psychiatric questionnaires, in
traumatic memory retrieval, 137
138
psychopathology, impact on emotional
memory
in children, 330, 332333
as dysfunctional dysregulation, 91,
112n.1
See also specific disorder
public events, memory accuracy of, 4
5, 13, 31, 33
reappraisal, as emotional control,
274275
recall and recollection
with anxiety disorders, as selective,
157158, 167; of autobiographical
memories, 158160; of
experimentally presented stimuli,
160167
conscious, 230232
of emotional memory: accuracy of,
413, 21, 3536; autobiographical,
57, 5960; as biased by elderly,
279, 289294; with bilateral
amygdala damage, 4849, 66; by
children, 250262; cognitive
studies on, 136139, 143; failure
with intensely negative, 2628;
with frontal cortex damage, 100

102; hippocampal atrophy and,

8384, 112n.2; impairment with
stress, 8394, 112n.2, 112n.4
113n.11; intrusive, 140142;
semantic confusion about, 148
149; sharing vs. not sharing
impact on, 102, 113n.8, 366, 380
381; with unilateral amygdala
damage, 4952; vividness vs., 45,
10, 20, 29, 103; working backward
from, 3334
of emotional reactions, 3033
of explicit vs. implicit memory, 77,
147148, 175
habitual, of negative experiences,
189195
in Holocaust survivors: associated
emotions, 358359; of emotions,
134, 146, 356361, 381; events
separate from emotions, 359
360; of fear, 356357, 360, 366
369; forgetting vs., 360361;
numbness with, 357359, 379
380; re-experiencing emotion
during, 362366, 382n.4
mood-congruent, 186, 189193,
209n.1
of nonaffective information, by
elderly, 284287
in schizophrenia: emotion and, 226
230; memory task dissociation
and, 218219, 221; subjective
states of consciousness and, 219
221, 233
of traumatic memory, 27, 94100,
110111, 129139, 250; after
repression of, 2728, 132, 317
319; in witnesses, 315319
recall error. See memory errors
recognition memory
anxiety disorders and, 167171
in children, 251252
in depressed states, 197200,
210n.3
in schizophrenia: conscious
awareness and, 219221;
emotional words influence on,
232233
reconstruction errors, in traumatic
memory, 5, 2728, 3536, 95
red-out, 135

re-experiencing, of emotional reality

in Holocaust survivors, 352, 355,
361369, 382n.4
sharing vs. not sharing impact on,
102, 113n.8, 366, 380381
rehearsal. See memory rehearsal
relational memories
as elderly focus, 274275
neural processing mechanism for, 43
parent-child reminiscing and, 245
246, 248250, 263
Remember/Know (R/K) procedure, for
memory impairment
in normal controls, 230232
in schizophrenia, 219220, 232234
reminiscing
of child sexual abuse, 265, 318,
320321
by elderly, 294295
by Holocaust survivors, 364, 376;
distraction strategies for, 364,
376, 380381
vs. talking about present emotions,
3, 245
See also parent-child reminiscing
repression, of traumatic memory
in children, 317, 330, 332333
cognitive studies on, 136139, 143
forgetting vs., 361
in Holocaust survivors, 360361
mechanisms of, 43, 98, 129130,
133, 135
recall after, 2728, 132, 317319
semantic confusion about, 148
149
research, on emotional memory
categorization schemes for, 810,
13, 15, 2728, 108, 236
children and, 265266
cognitive approach to, 106111,
113n.10n.11, 286; on forgetting
trauma, 136139, 143
eyewitness, 308309, 315, 317319,
327, 333334
further needs for, 13, 24, 26, 33,
111, 289
generalizability of, 1315, 86
in Holocaust survivors, 360361
learning approach to, 104106
neuroimaging in, 48, 5658, 61, 63,
100, 136, 284, 291

neuroscientific approach to, 104

106
oral testimonies used in, 347,
381n.1
perspectives for, 3, 67, 13, 1518
schizophrenia and, 225, 233236
response-focused strategies, for
emotional control, 274
retention
factors influencing, 45
in Holocaust survivors, 376377
retention interval, in assessing
emotional memory, 376
retrieval, of memory. See recall and
recollection
retrieval cues
for memory: with anxiety disorders,
159167, 178; in children, 251
254; semantic confusion with
traumatic, 148149, 159
in mood-congruent memory, 186
187, 191193, 202
right-hemisphere hypothesis, of
problem solving by elderly, 279
280
rumination. See habits of thought
sadness
parent-child reminiscing about,
246250
recall of, 2324, 290; vividness vs.
accuracy of, 45, 10, 20
in survivors (see crying)

See also negative experiences

satisfaction, emotional, in elderly,

273
as increasingly gratifying, 265n.2,
272273, 275, 289295
schematization, for memory
reconstruction, 34, 108, 110111,
113n.11
with depression, 194
in elderly, 284287
in Holocaust survivors, 373374
schizophrenia, 217236
conscious awareness in, 218, 233
235; emotional influence on, 221,
230232; normal controls of
emotional influence, 230232;
recognition memory and, 219
221, 232233

schizophrenia (continued)
emotional disturbances in, 221225;
of affect, 221224; of body state,
221, 224; of brain state, 221, 224
225; cognitive evaluation of, 217,
221, 223, 235n.1; research
implications, 225
emotional memory interaction with,
225233; conclusions about, 233
235; conscious awareness
influence on, 221, 230233; as
influential, 217218; literature
review of, 225227, 236n.2; recall
performance in, 227230
memory impairment in, 218221;
subjective states of, 219221, 233;
task dissociation, 218219, 221
scripts
for memory reconstruction, 110
111, 113n.11
by space alien abductees, 145146
selective memory
in anxiety disorders, 155180
in schizophrenia, 218219
self-concept. See emotional selfconcept
self-control. See self-regulation
self-defining, in emotional selfconcept, 245, 248
self-focus, emotional, in elderly, 284
287
self-in-relation, in emotional selfconcept, 245, 248
self-referential tasks
of habitual memory, 189190, 194
195, 203
in selective memory with anxiety,
172177
self-regulation, of emotions
by children, 245, 248, 250
by elderly, 273275
impaired: with habits of thinking, 200
205, 210n.4n.5; in nonhabitual
tasks, 195200, 210n.3
opposing habit through, 205209
semantic memory
in anxiety disorders, 161, 166, 174,
177
in elderly, 284285
with PTSD dissociative tendencies,
324326

recall implications of, 77, 193, 220,

231
serotonin, influence on memory, 89
sexual abuse victims
children as (see child sexual abuse
(CSA))
memory in, 1516
sharing, of traumatic memories,
impact on recollection, 102,
113n.8, 366, 380381
shocking events, flashbulb memories
of

accuracy of, 5, 3435, 42, 289

aging and, 287289
role of emotion in, 56, 32, 111
short-term memory, impairment of,
hippocampal atrophy with, 83
84, 112n.2
signal detection analysis, in selective
memory, 168169
simulation. See laboratory studies
situationally accessible memory, in
episodic memory, 352353,
381n.3
social-cultural theory
of childrens memories, 243244;
emotional development, 243244,
327, 329; parent-child
reminiscing context, 245250
Holocaust applications of, 373374,
376377
social phobias, selective memory in,
158159, 163165, 168169, 173
socialization, in emotional
development
of children, 244245, 327, 329
of elderly, 274275, 294
gender impact on, 245250, 265,
329
socioemotional selectivity theory,
274275
source-monitoring, in emotional
memory, 11, 1315
with aging, 282283, 285, 287289
eyewitness, 314315
space alien abductees, emotional
memory in, 144146
spatio-temporal memory
with aging, 279280
in episodic memory, 352353, 369
framework for organization, 810, 96

glucocorticoid impact on, 85, 88, 96

hippocampal function in, 80, 9596
spider phobics, memory research on,
1011
startle reflex, with PTSD, 97, 99100
Sternberg task, for cognitive
evaluation, of emotional events,
217, 223, 235n.1
stimuli
doctor/mechanic, memory
construction and, 6, 12, 19, 105
emotional (see emotional stimuli/
events)
emotional valence of (see negative
experiences; positive experiences)
experimentally presented, selective
recall with anxiety, 160167
nonemotional (see neutral stimuli/
events)
stimuli emotionality study, animation
for, 2122
stress
biopsychology of: effects on brain,
8084; effects on forgetting in
hippocampus, 94100; effects on
forgetting in prefrontal cortex,
101102; effects on memory, 84
94
chronic, 314; comorbidities impact,
29, 9394, 101; episodic memory
and, 9194, 99, 112n.4; intensity
variable, 2526
coping mechanisms for: avoidance
as, 139141, 156158, 161, 167;
in children, 244250, 258, 262
265, 328333
definition of, 7780, 250251
effects on memory, 8494; acute,
8588, 110, 314; in animals, 46
47, 8891; in children, 250264,
328332; chronic, 2526, 29,
9194, 101, 314; data review,
8485, 112n.3; emotion
influence on, 103111, 113n.10;
eyewitness, 309313; in humans,
6667, 321322; long-term
durability, 312313, 322323;
as reversible, 88, 91, 112n.4;
schizophrenia and, 222223;
traumatic, 12, 2021, 45, 84
94

physiological arousal with, 79, 98

100, 102
See also trauma and traumatic
memory
stress hormones
as indicator of stress, 7879, 82
memory modulation and, 21, 4547,
85; emotion influence on, 104
106, 110, 147; in witnesses and
victims, 316
with PTSD, 84, 112n.3
stress receptors, in frontal cortex, 101
stressor, definition of, 7880, 112n.1
Stroop color-naming paradigm
anxiety disorders and, 165, 168
for intrusive cognition, 141142
PTSD memory bias in, 320321
subjectivity, of emotionality
in normal controls, 230232
prefrontal cortex role in, 101102
research implications, 234235
in schizophrenia, 219221, 233
235
suicide, 16, 29
survivor guilt, 369373
survivors. See Holocaust survivors;
victims
target event
emotions differential favor of, 24, 28
peak vs. end snapshots of, 33
temporal lobectomy
for intractable epilepsy, 48
memory impact of, 4951, 54
temporal memory
in Holocaust survivors, 368369
See also spatio-temporal memory
thematic information
cognitive research on, 106111,
113n.10n.11
emotional valence of, 2223
in memory construction, 89, 14
18, 3536
thought suppression. See habits of
thought; memory suppression
threat(s)
anxiety disorders and: information
retrieval with, 160168, 171172,
174175; memory elaboration
with, 176178; panic as bias for,
178179

threat(s) (continued)
coping mechanisms for: avoidant
encoding as, 139141; in
children, 245250, 258, 262265;
selective attention as, 156158,
161, 167
memory bias with, 178179, 313
315, 325
transfer-appropriateness, in habitual
memory, 196197, 199, 207
trauma and traumatic memory
biopsychology of, 76113; effects on
brain, 8084; effects on
forgetting, 112n.5113n.8;
hippocampus and, 94100;
prefrontal cortex and, 100111;
effects on memory, 8494,
112n.5; emotional memory in,
12, 2021, 45; hippocampal
atrophy studies, 8384, 112n.2;
overview of, 21, 76, 111112,
112n.1; taxonomy for, 7677
construction of: avoidant encoding,
139141; emotion influence, 106
111, 112n.6113n.8; legal system
and, 132133, 309, 312315,
318319, 324, 334; sharing and,
102, 113n.8, 366, 380381
definitions of, 7880, 112n.1
dissociation with, 28, 129130,
132133, 135; avoidance and,
139141, 255; cognitive studies of,
136139, 143; false-memory
formation and, 136139, 142
146, 318319, 324326; semantic
confusion about, 148149; in
witnesses and victims, 317, 319,
323324
eyewitness, 315319; amnesia
research on, 317319; special
mechanisms of, 316317
flashbacks of, 4, 94, 102103
forgetting, 129149; amnesia vs.,
130136, 149; cognitive research
on, 136139; common
assumptions about, 129130;
complex mechanisms of, 2627,
94100; dispelling semantic
confusion, 148149; evidence
adduced for, 130136; later
recovery of, 2728, 132, 317319;

of material related to, 139141; in

space alien abductees, 144146;
types of memory in, 94, 96, 111,
112n.5113n.7
in Holocaust survivors:
characteristics of, 352356,
381n.3; qualitative study methods
for, 347352, 381n.1n.2
intrusion of, 140142, 166, 321,
353
memory distortion with, 142143
neurobiological account of, 97100
prolonged (see Holocaust survivors)
recall and recovery of: after
repression, 2728, 132, 317319;
in children, 250262, 330332;
cognitive studies on, 136139,
143; disruptive reactions and,
9899, 102103, 111, 113n.8;
in elderly, 287289; errors in, 5,
30, 33, 3536, 166, 251; filling
in with, 110111; as healing,
130; with repetitive events, 130;
as special, 146148; vividness vs.
accuracy of, 45, 10, 20
recategorization of relevant, 2728
repression of, 43, 98, 129130, 133,
135139; recovery mechanisms,
28, 132, 317319
survivor guilt in, 369373
See also post-traumatic stress
disorder (PTSD)
tunnel memories, 22
uncinate fasciculus fiber bundle, in
autobiographical memory
retrieval, 60
undoing, as anxiety defense, 9798,
102
unfamiliar/unusual stimulus, memory
construction of, 11, 15
unpleasant memories. See negative
experiences
Urbach-Weithe disease, 12, 4748
vagus nerve, body-state information
transmission through, 46, 67
valence, of emotion. See emotional
valence
ventral striatum, role in emotional
memory, 65, 67

verbatim memory
in episodic memory, 352353, 381n.3
stress impact on, 85, 8788
victims, emotional memory in

amnesia research on, 317319

children as, 130, 251252, 261,
265 (see also child sexual abuse
[CSA])

distress impact on, 309311, 313

legal system impact on, 132133,
309, 312315, 318319, 324, 334
of persecution (see Holocaust

survivors)

of sexual abuse, 9193, 110, 130

132, 135141, 265
space alien abductees as, 144146
special mechanisms of, 316317
of trauma (see trauma and traumatic

memory)
weapon impact on, 311312
violence, childrens recall of, 251252,
261262
interpersonal, 265 (see also child
sexual abuse [CSA])
visual information
basic-level, in memory
categorization, 89
processing of with aging, 279280,
286
visually induced emotion
in memory construction, 89, 14
18, 3536
valence factor of, 2223, 28

vividness
emotion influence on, 2930; in
children, 263264
memory accuracy vs., 45, 10, 20;
with negative experiences, 5,
308311
war veterans, emotional memory in,

9193, 95, 146

cognitive studies on, 137, 142,
320321
oral testimonies of, 347, 381n.1
recall of, 134135, 159, 174
weapon focus effect

hypodermic needle example, 17

18

on memory construction:

eyewitness, 311312; negative,

7, 10, 14, 107; positive,
22

white-noise method, of implicit

memory assessment, 173174

witnesses, emotional memory of. See

eyewitness memory

working memory, 100

PTSD impact on, 320322
Yerkes-Dodson law, of arousal and
memory, 2426, 35
challenges to, 2526, 254
in children, 254255
stress effect as, 4647, 146147,
253