You are on page 1of 1073

HANDBOOK OF

CHILD PSYCHOLOGY
SIXTH EDITION

Volume Two: Cognition, Perception, and Language

Volume Editors

DEANNA KUHN and ROBERT SIEGLER

Editors-in-Chief

WILLIAM DAMON and RICHARD M. LERNER

John Wiley & Sons, Inc.


HANDBOOK OF
CHILD PSYCHOLOGY
HANDBOOK OF
CHILD PSYCHOLOGY
SIXTH EDITION

Volume Two: Cognition, Perception, and Language

Volume Editors

DEANNA KUHN and ROBERT SIEGLER

Editors-in-Chief

WILLIAM DAMON and RICHARD M. LERNER

John Wiley & Sons, Inc.


Copyright 2006 by John Wiley & Sons, Inc. All rights reserved.

Published by John Wiley & Sons, Inc., Hoboken, New Jersey.


Published simultaneously in Canada.

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, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States
Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy
fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 646-8600, or on the web
at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons,
Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008.

Limit of Liability/ Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no
representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any
implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives
or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a
professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages,
including but not limited to special, incidental, consequential, or other damages.

This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold with the
understanding that the publisher is not engaged in rendering professional services. If legal, accounting, medical, psychological or any other
expert assistance is required, the services of a competent professional person should be sought.

Designations used by companies to distinguish their products are often claimed as trademarks. In all instances where John Wiley & Sons,
Inc. is aware of a claim, the product names appear in initial capital or all capital letters. Readers, however, should contact the appropriate
companies for more complete information regarding trademarks and registration.

For general information on our other products and services please contact our Customer Care Department within the United States at (800)
762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic
books. For more information about Wiley products, visit our web site at www.wiley.com.

Library of Congress Cataloging-in-Publication Data:

Handbook of child psychology / editors-in-chief, William Damon & Richard M. Lerner.


6th ed.
p. cm.
Includes bibliographical references and indexes.
Contents: v. 1. Theoretical models of human development / volume editor,
Richard M. Lerner v. 2. Cognition, perception, and language / volume editors,
Deanna Kuhn, Robert Siegler v. 3. Social, emotional, and personality development /
volume editor, Nancy Eisenberg v 4. Child psychology in practice / volume editors, K.
Ann Renninger, Irving E. Sigel.
ISBN 0-471-27287-6 (set : cloth)
ISBN 0-471-27288-4 (v. 1 : cloth) ISBN 0-471-27289-2 (v. 2 : cloth)
ISBN 0-471-27290-6 (v. 3 : cloth) ISBN 0-471-27291-4 (v. 4 : cloth)
1. Child psychology. I. Damon, William, 1944 II. Lerner, Richard M.
BF721.H242 2006
155.4dc22
2005043951

Printed in the United States of America.


10 9 8 7 6 5 4 3 2 1
In memory of Paul Mussen, whose generosity of spirit
touched our lives and helped build a field.
Contributors

Karen E. Adolph Michael Cole


Department of Psychology Department of Communication
New York University University of California, San Diego
New York, New York La Jolla, California

Glenda Andrews Michelle de Haan


Department of Psychology Cognitive Neuroscience Unit
Griffith University Institute of Child Health
Queensland, Australia London, England

Martha E. Arterberry Sam Franklin


Department of Psychology Department of Human Development
Gettysburg College Columbia University Teachers College
Gettysburg, Pennsylvania New York, New York

Patricia J. Bauer
Howard Gardner
Department of Psychological and Brain Sciences
Graduate School of Education
Duke University
Harvard University
Durham, North Carolina
Cambridge, Massachusetts

Sarah E. Berger
Department of Psychology David C. Geary
City University of New York Department of Psychological Sciences
College of Staten Island University of Missouri
Staten Island, New York Columbia, Missouri

Cara H. Cashon Susan A. Gelman


Department of Psychological and Brain Sciences Department of Psychology
University of Louisville University of Michigan
Louisville, Kentucky Ann Arbor, Michigan

Leslie B. Cohen Susan Goldin-Meadow


Department of Psychology Department of Psychology
University of Texas University of Chicago
Austin, Texas Chicago, Illinois

vii
viii Contributors

Graeme S. Halford Seana Moran


School of Psychology Graduate School of Education
University of Queensland Harvard University
Brisbane, Australia Cambridge, Massachusetts

Paul L. Harris Yuko Munakata


Graduate School of Education Department of Psychology
Harvard University University of Colorado
Cambridge, Massachusetts Boulder, Colorado

Katherine Hilden Charles A. Nelson III


Department of Counseling, Richard David Scott Chair in Pediatrics
Educational Psychology, Boston Childrens Hospital
and Special Education Developmental Medicine Center Laboratory
Michigan State University of Cognitive Neuroscience
East Lansing, Michigan Harvard Medical School
Boston, Massachusetts
Janellen Huttenlocher
Center for Early Childhood Research Nora S. Newcombe
University of Chicago Department of Psychology
Chicago, Illinois Temple University
Philadelphia, Pennsylvania
Charles W. Kalish
Department of Educational Psychology Michael Pressley
University of Michigan College of Education
Ann Arbor, Michigan Michigan State University
East Lansing, Michigan

Frank Keil
Jenny R. Saffran
Department of Psychology
Department of Psychology
Yale University
University of Wisconsin
New Haven, Connecticut
Madison, Wisconsin

Philip J. Kellman Robert S. Siegler


Department of Psychology Department of Psychology
University of California Carnegie-Mellon University
Los Angeles, California Pittsburgh, Pennsylvania

Deanna Kuhn Kathleen M. Thomas


Department of Human Development Institute of Child Development
Columbia University Teachers College University of Minnesota
New York, New York Minneapolis, Minnesota

Jeffrey L. Lidz Michael Tomasello


University of Maryland Max Planck Institute for Evolutionary Anthropology
College Park, Maryland Leipzig, Germany
Contributors ix

Sandra R. Waxman Lynne A. Werner


Department of Psychology Department of Speech and Hearing Sciences
Northwestern University University of Washington
Evanston, Illinois Seattle, Washington

Janet F. Werker Ellen Winner


Department of Psychology Department of Psychology
University of British Colombia Boston College
Vancouver, British Colombia, Canada Chestnut Hill, Massachusetts
Preface to Handbook of Child Psychology,
Sixth Edition
WILLIAM DAMON

Scholarly handbooks play several key roles in their dis- not only has endured over time but has evolved into a
ciplines. First and foremost, they reflect recent changes thriving tradition across a number of related academic
in the field as well as classic works that have survived disciplines.
those changes. In this sense, all handbooks present their All through its history, the Handbook has drawn on,
editors and authors best judgments about what is most and played a formative role in, the worldwide study of
important to know in the field at the time of publication. human development. What does the Handbooks history
But many handbooks also influence the fields that they tell us about where we, as developmentalists, have been,
report on. Scholarsespecially younger oneslook to what we have learned, and where we are going? What
them for sources of information and inspiration to guide does it tell us about what has changed and what has re-
their own work. While taking stock of the shape of its mained the same in the questions that we ask, in the
field, a handbook also shapes the stock of ideas that will methods that we use, and in the theoretical ideas that we
define the fields future. It serves both as an indicator draw on in our quest to understand human development?
and as a generator, a pool of received knowledge and a By asking these questions, we follow the spirit of the sci-
pool for spawning new insight. ence itself, for developmental questions may be asked
about any endeavor, including the enterprise of studying
human development. To best understand what this field
THE HANDBOOKS LIVING TRADITION has to tell us about human development, we must ask how
the field itself has developed. In a field that examines
Within the field of human development, the Handbook of continuities and changes, we must ask, for the field itself,
Child Psychology has served these key roles to a degree what are the continuities and what are the changes?
that has been exceptional even among the impressive The history of the Handbook is by no means the whole
panoply of the worlds many distinguished scholarly story of why the field is where it is today, but it is a fun-
handbooks. The Handbook of Child Psychology has had a damental part of the story. It has defined the choices
widely heralded tradition as a beacon, organizer, and en- that have determined the fields direction and has influ-
cyclopedia of developmental study for almost 75 years enced the making of those choices. In this regard, the
a period that covers the vast majority of scientific work Handbooks history reveals much about the judgments
in this field. and other human factors that shape a science.
It is impossible to imagine what the field would look
like if it had not occurred to Carl Murchison in 1931 to
assemble an eclectic assortment of contributions into THE CAST OF CHARACTERS
the first Handbook of Child Psychology. Whether or not
Murchison realized this potential (an interesting specu- Carl Murchison was a scholar/impresario who edited
lation in itself, given his visionary and ambitious na- The Psychological Register; founded and edited key psy-
ture), he gave birth to a seminal publishing project that chological journals; wrote books on social psychology,

xi
xii Preface to Handbook of Child Psychology, Sixth Edition

politics, and the criminal mind; and compiled an assort- which still is fresh today, Bhler described intricate
ment of handbooks, psychology texts, autobiographies of play and communication patterns among toddlers, pat-
renowned psychologists, and even a book on psychic be- terns that developmental psychology would not redis-
liefs (Sir Arthur Conan Doyle and Harry Houdini were cover until the late 1970s. Bhler also anticipated the
among the contributors). Murchisons initial Handbook critiques of Piaget that would appear during the socio-
of Child Psychology was published by a small university linguistics heyday of the 1970s:
press (Clark University) in 1931, when the field itself
Piaget, in his studies on childrens talk and reasoning, em-
was still in its infancy. Murchison wrote:
phasizes that their talk is much more egocentric than so-
cial . . . that children from 3 to 7 years accompany all their
Experimental psychology has had a much older scientific
manipulations with talk which actually is not so much in-
and academic status [than child psychology], but at the
tercourse as monologue . . . [but] the special relationship
present time it is probable that much less money is being
of the child to each of the different members of the house-
spent for pure research in the field of experimental psy-
hold is distinctly reflected in the respective conversations.
chology than is being spent in the field of child psychol-
(Buhler, 1931, p. 138)
ogy. In spite of this obvious fact, many experimental
psychologists continue to look upon the field of child psy-
chology as a proper field of research for women and for Other Europeans included Anna Freud, who wrote on
men whose experimental masculinity is not of the maxi- The Psychoanalysis of the Child, and Kurt Lewin,
mum. This attitude of patronage is based almost entirely who wrote on Environmental Forces in Child Behavior
upon a blissful ignorance of what is going on in the and Development.
tremendously virile field of child behavior. (Murchison, The Americans whom Murchison chose were equally
1931, p. ix) notable. Arnold Gesell wrote a nativistic account of his
twin studies, an enterprise that remains familiar to us
Murchisons masculine allusion, of course, is from an- today, and Stanfords Louis Terman wrote a comprehen-
other era; it could furnish some good material for a social sive account of everything known about the gifted
history of gender stereotyping. That aside, Murchison child. Harold Jones described the developmental ef-
was prescient in the task that he undertook and the way fects of birth order, Mary Cover Jones wrote about chil-
that he went about it. At the time Murchison wrote the drens emotions, Florence Goodenough wrote about
preface to his Handbook, developmental psychology was childrens drawings, and Dorothea McCarthy wrote
known only in Europe and in a few forward-looking about language development. Vernon Joness chapter on
American labs and universities. Nevertheless, Murchison childrens morals focused on the growth of character,
predicted the fields impending ascent: The time is not a notion that was to become lost to the field during the
far distant, if it is not already here, when nearly all com- cognitive-developmental revolution, but that reemerged
petent psychologists will recognize that one-half of the in the 1990s as the primary concern in the study of
whole field of psychology is involved in the problem of moral development.
how the infant becomes an adult psychologically Murchisons vision of child psychology included an
(Murchison, 1931, p. x). examination of cultural differences as well. His Hand-
For his original 1931 Handbook, Murchison looked to book presented to the scholarly world a young anthropol-
Europe and to a handful of American centers (or field ogist named Margaret Mead, just back from her tours of
stations) for child research (Iowa, Minnesota, the Uni- Samoa and New Guinea. In this early essay, Mead wrote
versity of California at Berkeley, Columbia, Stanford, that her motivation in traveling to the South Seas was to
Yale, Clark). Murchisons Europeans included a young discredit the views that Piaget, Levy-Bruhl, and other
genetic epistemologist named Jean Piaget, who, in an nascent structuralists had put forth concerning ani-
essay on Childrens Philosophies, quoted extensively mism in young childrens thinking. (Interestingly,
from interviews with 60 Genevan children between the about a third of Piagets chapter in the same volume was
ages of 4 and 12 years. Piagets chapter would provide dedicated to showing how Genevan children took years
American readers with an introduction to his seminal to outgrow animism.) Mead reported some data that she
research program on childrens conceptions of the called amazing: In not one of the 32,000 drawings
world. Another European, Charlotte Bhler, wrote a ( by young primitive children) was there a single case
chapter on childrens social behavior. In this chapter, of personalization of animals, material phenomena, or
Preface to Handbook of Child Psychology, Sixth Edition xiii

inanimate objects (Mead, 1931, p. 400). Mead parlayed in the Handbook tradition, made an appearance as au-
these data into a tough-minded critique of Western psy- thor of a major chapter ( by far the longest in the book)
chologys ethnocentrism, making the point that animism on prenatal and perinatal growth. Three other physio-
and other beliefs are more likely to be culturally in- logically oriented chapters were added as well: one on
duced than intrinsic to early cognitive development. neonatal motor behavior, one on visual-manual func-
This is hardly an unfamiliar theme in contemporary psy- tions during the first 2 years of life, and one on physio-
chology. Mead also offered a research guide for develop- logical appetites such as hunger, rest, and sex.
mental fieldworkers in strange cultures, complete with Combined with the Goodenough and Gesell shifts in
methodological and practical advice, such as the follow- focus, these additions gave the 1933 Handbook more of a
ing: Translate questions into native linguistic categories; biological thrust, in keeping with Murchisons long-
dont do controlled experiments; dont do studies that standing desire to display the hard science backbone of
require knowing ages of subjects, which are usually un- the emerging field.
knowable; and live next door to the children whom you Leonard Carmichael was president of Tufts Univer-
are studying. sity when he organized Wileys first edition of the
Despite the imposing roster of authors that Murchison Handbook. The switch from a university press to the
assembled for the 1931 Handbook of Child Psychology, long-established commercial firm of John Wiley &
his achievement did not satisfy him for long. Barely 2 Sons was commensurate with Carmichaels well-
years later, Murchison put out a second edition, of which known ambition; indeed, Carmichaels effort was to
he wrote: Within a period of slightly more than 2 years, become influential beyond anything that Murchison
this first revision bears scarcely any resemblance to the might have anticipated. The book (one volume at that
original Handbook of Child Psychology. This is due time) was called the Manual of Child Psychology, in
chiefly to the great expansion in the field during the past keeping with Carmichaels intention of producing an
3 years and partly to the improved insight of the editor advanced scientific manual to bridge the gap between
(Murchison, 1933, p. vii). The tradition that Murchison the excellent and varied elementary textbooks in this
had brought to life was already evolving. field and the scientific periodical literature
Murchison saw fit to provide the following warning in (Carmichael, 1946, p. viii).
his second edition: There has been no attempt to sim- The publication date was 1946, and Carmichael com-
plify, condense, or to appeal to the immature mind. This plained that this book has been a difficult and expensive
volume is prepared specifically for the scholar, and its one to produce, especially under wartime conditions
form is for his maximum convenience (Murchison, (Carmichael, 1946, p. viii). Nevertheless, the project was
1933, p. vii). It is likely that sales of Murchisons first worth the effort. The Manual quickly became the bible of
volume did not approach textbook levels; perhaps he re- graduate training and scholarly work in the field, avail-
ceived negative comments regarding its accessibility. able virtually everywhere that human development was
Murchison exaggerated when he wrote that his sec- studied. Eight years later, now head of the Smithsonian
ond edition bore little resemblance to the first. Almost Institution, Carmichael wrote, in the preface to the 1954
half of the chapters were virtually the same, with minor second edition, The favorable reception that the first
additions and updating. (For the record, though, despite edition received not only in America but all over the
Murchisons continued use of masculine phraseology, world is indicative of the growing importance of the
10 of the 24 authors in the second edition were women.) study of the phenomena of the growth and development of
Some of the authors whose original chapters were the child (Carmichael, 1954, p. vii).
dropped were asked to write about new topics. So, for Carmichaels second edition had a long life: Not until
example, Goodenough wrote about mental testing rather 1970 did Wiley bring out a third edition. Carmichael was
than about childrens drawings, and Gesell wrote a gen- retired by then, but he still had a keen interest in the
eral statement of his maturational theory that went well book. At his insistence, his own name became part of the
beyond the twin studies. title of the third edition; it was called, improbably,
But Murchison also made some abrupt changes. He Carmichaels Manual of Child Psychology, even though it
dropped Anna Freud entirely, auguring the marginaliza- had a new editor and an entirely different cast of authors
tion of psychoanalysis within academic psychology. and advisors. Paul Mussen took over as the editor, and
Leonard Carmichael, who was later to play a pivotal role once again the project flourished. Now a two-volume set,
xiv Preface to Handbook of Child Psychology, Sixth Edition

the third edition swept across the social sciences, gener- plained.) In any case, Carmichael acknowledged the
ating widespread interest in developmental psychology roots of his Manuals, if not always their original editor.
and its related disciplines. Rarely had a scholarly com- His choice to start with those roots is a revealing part
pendium become both so dominant in its own field and so of the Handbooks history, and it established a strong
familiar in related disciplines. The set became an essen- intellectual legacy for our present-day descendants of
tial source for graduate students and advanced scholars the early pioneers who wrote for the Murchison and
alike. Publishers referred to Carmichaels Manual as the Carmichael editions.
standard against which other scientific handbooks were Although Leonard Carmichael took the 1946 Manual
compared. in much the same direction established by Murchison
The fourth edition, published in 1983, was now re- back in 1931 and 1933, he did bring it several steps fur-
designated by John Wiley & Sons to become once again ther in that direction, added a few twists of his own, and
the Handbook of Child Psychology. By then, Carmichael dropped a couple of Murchisons bolder selections.
had passed away. The set of books, now expanded to four Carmichael first appropriated five Murchison chapters
volumes, became widely referred to in the field as the on biological or experimental topics, such as physiologi-
Mussen handbook. cal growth, scientific methods, and mental testing. He
added three new biologically oriented chapters on ani-
mal infancy, physical growth, and motor and behavioral
WHAT CARMICHAEL CHOSE FOR THE maturation (a tour de force by Myrtal McGraw that in-
NOW EMERGENT FIELD stantly made Gesells chapter in the same volume obso-
lete). Then he commissioned Wayne Dennis to write an
Leonard Carmichael, who became Wileys editor for adolescence chapter that focused exclusively on physio-
the project in its now commercially funded and ex- logical changes associated with puberty.
panded versions (the 1946 and 1954 Manuals), made On the subject of social and cultural influences in de-
the following comments about where he looked for his velopment, Carmichael retained five of the Murchison
all-important choices of content: chapters: two chapters on environmental forces on the
child by Kurt Lewin and by Harold Jones, Dorothea Mc-
Both as editor of the Manual and as the author of a spe- Carthys chapter on childrens language, Vernon Joness
cial chapter, the writer is indebted . . . [for] extensive
chapter on childrens morality (now entitled Character
excerpts and the use of other materials previously pub-
DevelopmentAn Objective Approach), and Margaret
lished in the Handbook of Child Psychology, Revised Edi-
tion. (1946, p. viii)
Meads chapter on primitive children (now enhanced
by several spectacular photos of mothers and children
Both the Handbook of Child Psychology and the Handbook
from exotic cultures around the world). Carmichael also
of Child Psychology, Revised Edition, were edited by Dr.
stayed with three other Murchison topics (emotional de-
Carl Murchison. I wish to express here my profound appre-
ciation for the pioneer work done by Dr. Murchison in pro-
velopment, gifted children, and sex differences), but he
ducing these handbooks and other advanced books in selected new authors to cover them. But Carmichael
psychology. The Manual owes much in spirit and content dropped Piaget and Bhler.
to the foresight and editorial skill of Dr. Murchison. Carmichaels 1954 revision, his second and final edi-
(1954, p. viii) tion, was very close in structure and content to the 1946
Manual. Carmichael again retained the heart of Murchi-
The first quote comes from Carmichaels preface to sons original vision, many of Murchisons original
the 1946 edition, the second from his preface to the authors and chapter topics, and some of the same mate-
1954 edition. We shall never know why Carmichael rial that dated all the way back to the 1931 Handbook.
waited until the 1954 edition to add the personal tribute Not surprisingly, the chapters that were closest to
to Carl Murchison. Perhaps a careless typist dropped Carmichaels own interests got the most significant up-
the laudatory passage from a handwritten version of the dating. Carmichael leaned toward the biological and
1946 preface and its omission escaped Carmichaels physiological whenever possible. He clearly favored ex-
notice. Or perhaps 8 years of further adult development perimental treatments of psychological processes. Yet he
increased Carmichaels generosity of spirit. (It also still kept the social, cultural, and psychological analyses
may be possible that Murchison or his family com- by Lewin, Mead, McCarthy, Terman, Harold Jones, and
Preface to Handbook of Child Psychology, Sixth Edition xv

Vernon Jones, and he even went so far as to add one new on Carmichaels last Manual, Paul Mussen wrote, The
chapter on social development by Harold and Gladys 1954 edition of this Manual had only one theoretical
Anderson and one new chapter on emotional develop- chapter, and that was concerned with Lewinian theory
ment by Arthur Jersild. which, so far as we can see, has not had a significant
The Murchison and Carmichael volumes make for lasting impact on developmental psychology (Mussen,
fascinating reading, even today. The perennial themes of 1970, p. x). The intervening years had seen a turning
the field were there from the start: the nature-nurture away from the norm of psychological research once
debate; the generalizations of universalists opposed by fondly referred to as dust-bowl empiricism.
the particularizations of contextualists; the alternating The Mussen 1970 editionor Carmichaels Manual,
emphases on continuities and discontinuities during on- as it was still calledhad a new look and an almost
togenesis; and the standard categories of maturation, entirely new set of contents. The two-volume edition
learning, locomotor activity, perception, cognition, lan- carried only one chapter from the earlier books,
guage, emotion, conduct, morality, and cultureall Carmichaels updated version of his own long chapter
separated for the sake of analysis, yet, as authors on the Onset and Early Development of Behavior,
throughout each of the volumes acknowledged, all some- which had made its appearance under a different title in
how inextricably joined in the dynamic mix of human Murchisons 1933 edition. Otherwise, as Mussen wrote
development. in his preface, It should be clear from the outset . . .
These things have not changed. Yet, much in the early that the present volumes are not, in any sense, a revision
editions is now irrevocably dated. Long lists of chil- of the earlier editions; this is a completely new Manual
drens dietary preferences, sleeping patterns, elimina- (Mussen, 1970, p. x).
tion habits, toys, and somatic types look quaint and And it was. In comparison to Carmichaels last edi-
pointless through todays lenses. The chapters on chil- tion 16 years earlier, the scope, variety, and theoretical
drens thought and language were written prior to the depth of the Mussen volumes were astonishing. The
great contemporary breakthroughs in neurology and field had blossomed, and the new Manual showcased
brain / behavior research, and they show it. The chapters many of the new bouquets that were being produced.
on social and emotional development were ignorant of The biological perspective was still strong, grounded by
the processes of social influence and self-regulation that chapters on physical growth ( by J. M. Tanner) and phys-
soon would be revealed through attribution research and iological development ( by Dorothy Eichorn) and by
other studies in social psychology. Terms such as cogni- Carmichaels revised chapter (now made more elegant
tive neuroscience, neuronal networks, behavior genetics, by some excerpts from Greek philosophy and modern
social cognition, dynamic systems, and positive youth de- poetry). But two other cousins of biology also were rep-
velopment were of course unknown. Even Meads rendi- resented, in an ethological chapter by Eckhard Hess and
tion of the primitive child stands as a weak straw in a behavior genetics chapter by Gerald McClearn. These
comparison to the wealth of cross-cultural knowledge chapters were to define the major directions of biologi-
available in todays cultural psychology. cal research in the field for at least the next 3 decades.
Most telling, the assortments of odd facts and norma- As for theory, Mussens Handbook was thoroughly
tive trends were tied together by very little theory permeated with it. Much of the theorizing was organ-
throughout the Carmichael chapters. It was as if, in the ized around the approaches that, in 1970, were known
exhilaration of discovery at the frontiers of a new field, as the three grand systems: (1) Piagets cognitive-
all the facts looked interesting in and of themselves. developmentalism, (2) psychoanalysis, and (3) learning
That, of course, is what makes so much of the material theory. Piaget was given the most extensive treatment.
seem odd and arbitrary. It is hard to know what to make He reappeared in the Manual, this time authoring a
of the lists of facts, where to place them, which ones comprehensive (and, some say, definitive) statement of
were worth keeping track of and which ones are expend- his entire theory, which now bore little resemblance to
able. Not surprisingly, the bulk of the data presented in his 1931/1933 sortings of childrens intriguing verbal
the Carmichael manuals seems not only outdated by expressions. In addition, chapters by John Flavell, by
todays standards but, worse, irrelevant. David Berlyne, by Martin Hoffman, and by William
By 1970, the importance of theory for understanding Kessen, Marshall Haith, and Philip Salapatek all gave
human development had become apparent. Looking back major treatments to one or another aspect of Piagets
xvi Preface to Handbook of Child Psychology, Sixth Edition

body of work. Other approaches were represented as quite on center stage. Research topics now ranged from
well. Herbert and Ann Pick explicated Gibsonian the- childrens play to brain lateralization, from childrens
ory in a chapter on sensation and perception, Jonas family life to the influences of school, day care, and dis-
Langer wrote a chapter on Werners organismic theory, advantageous risk factors. There also was coverage of
David McNeill wrote a Chomskian account of language the burgeoning attempts to use developmental theory as
development, and Robert LeVine wrote an early version a basis for clinical and educational interventions. The
of what was soon to become culture theory. interventions usually were described at the end of chap-
With its increased emphasis on theory, the 1970 Man- ters that had discussed the research relevant to the
ual explored in depth a matter that had been all but ne- particular intervention efforts, rather than in whole
glected in the books previous versions: the mechanisms chapters dedicated specifically to issues of practice.
of change that could account for, to use Murchisons old This brings us to the efforts under the present edito-
phrase, the problem of how the infant becomes an adult rial team: the Handbooks fifth and sixth editions ( but
psychologically. In the process, old questions such as really the seventh and eighth editions, if the germinal
the relative importance of nature versus nurture were re- two pre-Wiley Murchison editions are counted). I must
visited, but with far more sophisticated conceptual and leave it to future commentators to provide a critical sum-
methodological tools. mation of what we have done. The volume editors have
Beyond theory building, the 1970 Manual addressed an offered introductory and/or concluding renditions of
array of new topics and featured new contributors: peer their own volumes. I will add to their efforts here only
interaction (Willard Hartup), attachment (Eleanor Mac- by stating the overall intent of our design and by com-
coby and John Masters), aggression (Seymour Feshback), menting on some directions that our field has taken in
individual differences (Jerome Kagan and Nathan Kogan), the years from 1931 to 2006.
and creativity (Michael Wallach). All of these areas of in- We approached our editions with the same purpose
terest are still very much with us in the new millennium. that Murchison, Carmichael, and Mussen before us had
If the 1970 Manual reflected a blossoming of the shared: to provide, as Mussen wrote, a comprehen-
fields plantings, the 1983 Handbook reflected a field sive and accurate picture of the current state of knowl-
whose ground cover had spread beyond any boundaries edgethe major systematic thinking and researchin
that could have been previously anticipated. New the most important research areas of the psychology of
growth had sprouted in literally dozens of separate lo- human development (Mussen, 1983, p. vii). We as-
cations. A French garden, with its overarching designs sumed that the Handbook should be aimed specifically
and tidy compartments, had turned into an English gar- for the scholar, as Murchison declared, and that it
den, a bit unruly but glorious in its profusion. Mussens should have the character of an advanced text, as
two-volume Carmichaels Manual had now become the Carmichael defined it. We expected, though, that our
four-volume Mussen Handbook, with a page-count in- audiences may be more interdisciplinary than the read-
crease that came close to tripling the 1970 edition. erships of previous editions, given the greater tendency
The grand old theories were breaking down. Piaget of todays scholars to cross back and forth among fields
was still represented by his 1970 piece, but his influence such as psychology, cognitive science, neurobiology,
was on the wane throughout the other chapters. Learning history, linguistics, sociology, anthropology, educa-
theory and psychoanalysis were scarcely mentioned. Yet tion, and psychiatry. We also believed that research-
the early theorizing had left its mark, in vestiges that oriented practitioners should be included under the
were apparent in new approaches, and in the evident con- rubric of the scholars for whom this Handbook was
ceptual sophistication with which authors treated their intended. To that end, for the first time in 1998 and
material. No return to dust bowl empiricism could be again in the present edition, we devoted an entire vol-
found anywhere in the set. Instead, a variety of classical ume to child psychology in practice.
and innovative ideas were coexisting: Ethology, neurobi- Beyond these very general intentions, we have let
ology, information processing, attribution theory, cul- chapters in the Handbooks fifth and sixth editions take
tural approaches, communications theory, behavioral their own shape. We solicited the chapters from authors
genetics, sensory-perception models, psycholinguistics, who were widely acknowledged to be among the leading
sociolinguistics, discontinuous stage theories, and con- experts in their areas of the field, although we know
tinuous memory theories all took their places, with none that, given an entirely open-ended selection process and
Preface to Handbook of Child Psychology, Sixth Edition xvii

no limits of budget, we would have invited a large num- has a way of moving in alternating cycles (or spirals, for
ber of other leading researchers whom we did not have those who wish to capture the progressive nature of sci-
the spaceand thus the privilegeto include. With entific development). In our time, developmental study
very few exceptions, every author whom we invited has cycled away from classic topics such as motivation
agreed to accept the challenge. Our only real, and great, and learningnot in the sense that they were entirely
sadness was to hear of the passing of several authors forgotten, or that good work ceased to be done in such
from the 1998 edition prior to our assembly of the pres- areas, but in the sense that they no longer were the most
ent edition. Where possible, we arranged to have their prominent subjects of theoretical reflection and debate.
collaborators revise and update their chapters. Some of the relative neglect was intentional, as scholars
Our directive to authors was simple: Convey your got caught up in controversies about whether psycholog-
area of the field as you see it. From then on, the authors ical motivation was a real phenomenon worthy of
took center stagewith, of course, much constructive study or whether learning could or should be distin-
feedback from reviewers and volume editors. No one guished from development in the first place. All this has
tried to impose a perspective, a preferred method of in- changed. As the contents of our current edition attest,
quiry, or domain boundaries on any of the chapters. The developmental science always returns, sooner or later, to
authors expressed their views on what researchers in concepts that are necessary for explaining the heart of
their areas attempt to accomplish, why they do so, how its concerns, progressive change in individuals and so-
they go about it, what intellectual sources they draw on, cial groups over time, and concepts such as learning and
what progress they have made, and what conclusions motivation are indispensable for this task. Among the
they have reached. exciting features of this Handbook edition are the ad-
The result, in my opinion, is still more glorious pro- vances it presents in theoretical and empirical work on
fusion of the English garden genre, but perhaps con- these classic concepts.
tained a bit by some broad patterns that have emerged The other concept that has met some resistance in
over the past decade. Powerful theoretical models and recent years is the notion of development itself. For
approachesnot quite unified theories, such as the some social critics, the idea of progress, implicit in the
three grand systemshave begun once again to organize notion of development, has seemed out of step with
much of the fields research and practice. There is great principles such as equality and cultural diversity. Some
variety in these models and approaches, and each is genuine benefits have accrued from that critique; for
drawing together significant clusters of work. Some example, the field has worked to better appreciate di-
have been only recently formulated, and some are com- verse developmental pathways. But, like many critique
binations or modifications of classic theories that still positions, it led to excesses. For some, it became ques-
have staying power. tionable to explore issues that lie at the heart of human
Among the formidable models and approaches that development. Growth, advancement, positive change,
the reader will find in this Handbook are the dynamic achievement, and standards for improved performance
system theories, the life span and life course ap- and conduct, all were questioned as legitimate subjects
proaches, cognitive science and neuronal models, the of investigation.
behavior genetics approach, person-context interaction Just as in the cases of learning and motivation, no
theories, action theories, cultural psychology, and a doubt it was inevitable that the fields center of gravity
wide assortment of neo-Piagetian and neo-Vygotskian sooner or later would return to broad concerns of devel-
models. Although some of these models and approaches opment. The story of growth from infancy to adulthood is
have been in the making for some time, they have now a developmental story of multifaceted learning, acquisi-
come into their own. Researchers are drawing on them tions of skills and knowledge, waxing powers of attention
directly, taking their implied assumptions and hypothe- and memory, growing neuronal and other biological ca-
ses seriously, using them with specificity and control, pacities, formations and transformations of character
and exploiting their implications for practice. and personality, increases and reorganizations in the un-
Another pattern that emerges is a rediscovery and derstanding of self and others, advances in emotional and
exploration of core processes in human development behavioral regulation, progress in communicating and
that had been underexamined by the generation of re- collaborating with others, and a host of other achieve-
searchers just prior to the present one. Scientific interest ments documented in this edition. Parents, teachers, and
xviii Preface to Handbook of Child Psychology, Sixth Edition

other adults in all parts of the world recognize and value and current set of scientific theories and findings
such developmental achievements in children, although available in the field today.
they do not always know how to understand them, let
alone how to foster them. February 2006
The sorts of scientific findings that the Handbooks Palo Alto, California
authors explicate in their chapters are needed to pro-
vide such understanding. The importance of sound sci- REFERENCES
entific understanding has become especially clear in
recent years, when news media broadcast story after Bhler, C. (1931). The social participation of infants and toddlers. In
story based on simplistic and biased popular specula- C. Murchison (Ed.), A handbook of child psychology. Worcester,
MA: Clark University Press.
tions about the causes of human development. The
Carmichael, L. (Ed.). (1946). Manual of child psychology. New York:
careful and responsible discourse found in these chap- Wiley.
ters contrasts sharply with the typical news story about Carmichael, L. (Ed.). (1954). Manual of child psychology (2nd ed.).
the role of parents, genes, or schools in childrens New York: Wiley.
growth and behavior. There is not much contest as to Mead, M. (1931). The primitive child. In C. Murchison (Ed.), A
handbook of child psychology. Worcester, MA: Clark University
which source the public looks to for its information and Press.
stimulation. But the good news is that scientific truth Murchison, C. (Ed.). (1931). A handbook of child psychology. Worces-
usually works its way into the public mind over the long ter, MA: Clark University Press.
run. The way this works would make a good subject for Murchison, C. (Ed.). (1933). A handbook of child psychology (2nd
ed.). Worcester, MA: Clark University Press.
developmental study some day, especially if such a
Mussen, P. (Ed.). (1970). Carmichaels manual of child psychology.
study could find a way to speed up the process. In the New York: Wiley.
meantime, readers of this edition of the Handbook of Mussen, P. (Ed.). (1983). Handbook of child psychology. New York:
Child Psychology will find the most solid, insightful Wiley.
Acknowledgments

A work as significant as the Handbook of Child Psychol- colleagues at Publications Development Company for
ogy is always produced by the contributions of numerous undertaking the enormous task of copy editing and pro-
people, individuals whose names do not necessarily ap- ducing the thousands of pages of the Sixth Edition.
pear on the covers or spines of the volumes. Most impor- Their professionalism and commitment to excellence
tant, we are grateful to the more than 150 colleagues were invaluable resources and provided a foundation
whose scholarship gave life to the Sixth Edition. Their upon which the editors work was able to move forward
enormous knowledge, expertise, and hard work make productively.
this edition of the Handbook the most important refer- Child development typically happens in families. So
ence work in developmental science. too, the work of editors on the Handbook moved along
In addition to the authors of the chapters of the four productively because of the support and forbearance of
volumes of this edition, we were fortunate to have been spouses, partners, and children. We thank all of our
able to work with two incredibly skilled and dedicated loved ones for being there for us throughout the several
editors within the Institute for Applied Research in years on which we have worked on the Sixth Edition.
Youth Development at Tufts University, Jennifer Davi- Numerous colleagues critiqued the chapters in man-
son and Katherine Connery. Their can-do spirit uscript form and provided valuable insights and sug-
and their impressive ability to attend to every detail gestions that enhanced the quality of the final
of every volume were invaluable resources enabling products. We thank all of these scholars for their enor-
this project to be completed in a timely and high mous contributions.
quality manner. William Damon and Richard M. Lerner thank the
It may be obvious, but we want to stress also that John Templeton Foundation for its support of their re-
without the talent, commitment to quality, and profes- spective scholarly endeavors. In addition, Richard M.
sionalism of our editors at John Wiley & Sons, this edi- Lerner thanks the National 4-H Council for its support
tion of the Handbook would not be a reality and would of his work. Nancy Eisenberg thanks the National Insti-
not be the cutting-edge work we believe it to be. The tute of Mental Health, the Fetzer Institute, and The In-
breadth of the contributions of the Wiley staff to the stitute for Research on Unlimited LoveAltruism,
Handbook is truly enormous. Although we thank all Compassion, Service ( located at the School of Medi-
these colleagues for their wonderful contributions, we cine, Case Western Reserve University) for their sup-
wish to make special note of four people in particular: port. K. Ann Renninger and Irving E. Sigel thank
Patricia Rossi, Senior Editor, Psychology, Linda Wit- Vanessa Ann Gorman for her editorial support for Vol-
zling, Senior Production Editor, Isabel Pratt, Associate ume 4. Support from the Swarthmore College Provosts
Editor, and Peggy Alexander, Vice President and Pub- Office to K. Ann Renninger for editorial assistance on
lisher. Their creativity, professionalism, sense of bal- this project is also gratefully acknowledged.
ance and perspective, and unflagging commitment to the Finally, in an earlier form, with Barbara Rogoff s en-
tradition of quality of the Handbook were vital ingredi- couragement, sections of the preface were published in
ents for any success we may have with this edition. Human Development (April 1997). We thank Barbara
We are also deeply grateful to Pam Blackmon and her for her editorial help in arranging this publication.

xix
Preface to Volume Two
Cognition, Perception, and Language
DEANNA KUHN and ROBERT S. SIEGLER

When we were approached about beginning work on the publication of the 5th and 6th editions of the Handbook
sixth edition of the Handbook of Child Psychology, the is only half as long as the interval between the 4th and
two of us had the same startled reaction: What? It 5th editions, there certainly seemed to be as much new
seemed not that long ago that we had regained our stride material to be examined in this edition as there had in
after the rush of attention devoted to completing the the previous one, again perhaps a reflection of the rap-
fifth edition, and we couldnt believe it was time to start idly growing field but also a reflection of the new per-
again. Now, another 3 years has passed, and once again spectives that are introduced. As volume editors we had
we are in the final phases of completing the manuscript a critical decision to make at the outset of planning the
for the volume, this time the sixth edition; at this point, sixth editionwhether to ask authors of chapters in the
we have greater appreciation for the timeliness and fifth edition to update their chapters or whether to com-
value of a new edition that will appear 8 years after the mission new chapters. We decided that we would com-
previous one. mission all new chapters because there were so many
The field is growing at a prodigious rate. The number important topics that had not been covered in the earlier
of specialized journals has increased as the field itself edition and so many leading authorities with unique and
not only grows but becomes more segmented and spe- important perspectives who had not previously con-
cialized. New journals target a much narrower range of tributed a chapter to this Handbook. Soliciting entirely
authors and readers than do the traditional journals new chapters also guaranteed freshly construed integra-
Child Development and Developmental Psychology. Few tions from an early twenty-first-century vantage point
researchers claim expertise and a command of the rele- for all chapters.
vant literature in more than a very few subfields. The Cognition, Language, and Perception volume of
In this climate, high-quality Handbook chapters, writ- the sixth edition of the Handbook of Child Psychology
ten by leaders in their specialties, serve a very important contains 22 chapters, compared to 19 in the fifth edi-
function. They offer an efficient and effective means to tion. In a number of cases, chapters in the sixth edition
escape the tunnel vision that has become an academic oc- represent new perspectives on literatures that were also
cupational hazard. The danger is that of becoming in- included in the fifth editionfor example the chapters
creasingly focused on ones own narrow specialty until it by Nelson, Thomas, and de Haan on neural bases of cog-
absorbs all ones awareness. Access to broad, integrative nition, Kellman and Arterberry on infant visual percep-
Handbook chapters reduces the likelihood of the field be- tion, Adolph and Berger on motor development, Waxman
coming a collection of diligent scholars who know more and Lidz on early word learning, Halford and Andrews
and more about less and less, their own work ultimately on problem solving and reasoning, and Cole on cognition
diminished by lack of nourishment from analogous in cultural context. Comparison of the corresponding
themes, insights, and questions that emerge in other areas. chapters from the 1998 and current editions makes clear
The present volume offers such a set of broad, inte- the value of offering readers a different perspective on
grative contributions. Although the interval between the same research area. Comparison of the two editions

xxi
xxii Preface to Volume Two

also makes clear that the chapters in the fifth edition are 4. An emphasis on individual dif ferences. How do chil-
by no means obsolete; they continue to offer unique in- dren become so different from each other?
sights and integrative visions. 5. Predictions and recommendations for the future with
Soliciting new chapters also allowed us to include new respect to your topic.
topics that had not been the focus of chapters in the corre-
sponding volume of the previous Handbook but that repre-
sent areas of growing importance in the field. These All of our authors were conscientious and good hu-
include Goldin-Meadows chapter on noverbal communi- mored about meeting deadlines and responding to sug-
cation, Gearys chapter on mathematical development, gestions for modifications, with the result that we had a
Winners chapter on development in the arts, Newcombe final set of manuscripts in hand while all were still fresh
and Huttenlochers chapter on spatial cognition, and and timely. This favorable set of circumstances gave us
Sieglers chapter on microgenetic studies of learning. the opportunity to react to them as a whole. In this re-
The rapid growth of these areas, and the many intriguing spect we would like to share several reactions. The major
theoretical debates within them, attests to the vigor of the one connects to a remark that appeared in our earlier
field of child development. commentary in connection with the fifth edition, and
Other differences between the current and the previ- that was the detection of a growing dissatisfaction with
ous volume reflect a variety of other currents within the task-bound paradigms on the part of the chapters au-
field and within the Handbook. We chose to omit two top- thorsin other words, lines of work whose explanatory
ics that were the subject of chapters in the previous edi- power derives from a single experimental task. Such par-
tion: representation and individual differences. Both adigms leave unanswered the question of whether their
topics have become so fundamental that they warranted explanatory power extends at all beyond that particular
treatment within the context of almost all of the individ- research context.
ual chapters, rather than being limited to a single chapter. In the present volume, we see more pronounced and
We divided another topicmemoryinto two chapters multiple kinds of evidence of advances regarding task-
because a clear division has arisen between event mem- bound paradigms. One form of evidence is the fact that
ory, which is dominant in infancy and early childhood, the authors of many of the chapters bring multiple levels
and intentional, strategic memory, which becomes of crit- of explanation to bear on their topic and their more spe-
ical interest in later childhood. Finally, chapters by two cific claimsneurological, biographical, anthropologi-
other authorsone on cognitive science and cognitive cal, evolutionary, species comparative, and historical
development by Frank Keil and one on exceptional cogni- comparative, as well as behaviors amenable to more tra-
tive development by Seana Moran and Howard Gardner ditional examination in a laboratory. None of the authors
are new to the volume, although Keil and Gardner each looks to a single narrow form of evidence as exhaustive
contributed chapters to other volumes of the fifth edition input for addressing the questions at hand. Tomasello,
of the Handbook. for example, emphasizes the impossibility of studying
We asked authors in the sixth edition to emphasize language development apart from the context of develop-
five themes. These themes help to unify the diverse top- ing cognitive and social skills, and Cohen and Cashon
ics covered in the volume. They are: argue that infant cognition research must incorporate
neuroscience. And certainly, the combination of experi-
mental and naturalistic methodologies is no longer un-
1. The history of the area. How has the study of your usual. At the level of the entire volume, the chapters
topic evolved over the years? added to the sixth edition are signs of a broadened per-
2. The contemporary status of the topic. Here you should spective: Both nonverbal and aesthetic dimensions of
feel free to express your own appropriately supported cognitive development are included as offering valuable
opinions, while acknowledging that others exist. insight into what develops.
3. An emphasis on mechanisms of change. The time has A second form of evidence is the number of authors
passed when it is sufficient to describe development, who take an explicitly historical view. They decline to
and we must now seek to understand how this devel- view their topic only in terms of the research paradigms
opment occurs. most popular at the moment and instead ask how these
Preface to Volume Two xxiii

paradigms connect to ones in use at earlier times. Harris, combe and Huttenlocher remind us in their chapter of
for example, asks how current theory-of-mind research the value of this extended ontogenetic perspectivea
(and the particular false-belief task from which it arose) value that is often overlooked in the current focus on
connects to research of several decades ago on role-taking searching for cognitive origins of competencies in the
and perspective-taking. Cohen and Cashon suggest that earliest years of life. Ultimately, we can only fully un-
recent infant cognition research continues the study of ob- derstand a competency to the extent that we understand
ject permanence and causality associated with Piagets where it is headed, that is, understand its mature
early work. Pressley and Hilden probe the connections be- state. As Saffran, Werker, and Werner make clear, this
tween the current research they examine and early mem- often entails research investigating the nature of adult
ory strategy research, while Kuhn and Franklin, Siegler, competencies.
and Geary examine relations between contemporary and Fourth, and finally, we see in these chapters an
earlier classical learning research, and Kuhn and Franklin awareness that new, enriched perspectives, and new
do the same with respect to adolescent cognition. Other forms of evidence, brought to bear on a topic have the
authors, for example Tomasello, Newcombe and Hutten- potential to provide new insights into old questions,
locher, and Munakata, make explicit note of the limited questions that may have been regarded as at a standstill.
explanatory power of constrained task formats. Keil makes this point explicitly in his chapter in arguing
A third and related form of evidence is the number for an expanded interdisciplinary approach. But this
of authors who take an extended view ontogenetically, same awareness is reflected in other chapters as well.
as well as historically. In other words, even if their own Like Tomasello, who draws our attention to individual
area of investigation is largely limited to the early years differences, Newcombe and Huttenlocher note that indi-
of life, they ask what the later implications are or what vidual differences, long regarded by many simply as nui-
further developmental course the investigated compe- sance variance to be avoided, to the contrary stand to
tencies may follow. Kuhn and Franklins chapter explic- provide insight into the fundamental nature of the com-
itly takes the perspective of this later development, as mon competencies and developmental pathways that are
does Moran and Gardners, but the authors of a number of interest.
of other chapters reflect their awareness that develop- We are privileged to have worked with this excep-
ment does not stop, or cease to be interesting at age 3, or tional set of authors and thank them for their dedication
5, or 10. Authors who have characteristically focused to their task and their contribution to the final product.
their attention on early development, such as Gelman We have said enough that we should at this point con-
and Kalish, Harris, and Bauer, now ask what might be clude our own remarks and let the authors speak for
expected in the years following early childhood. New- themselves.
Contents

Section One: Foundations


1 NEURAL BASES OF COGNITIVE DEVELOPMENT 3
Charles A. Nelson III, Kathleen M. Thomas, and Michelle de Haan

2 THE INFANTS AUDITORY WORLD: HEARING, SPEECH, AND THE


BEGINNINGS OF LANGUAGE 58
Jenny R. Saffran, Janet F. Werker, and Lynne A. Werner

3 INFANT VISUAL PERCEPTION 109


Philip J. Kellman and Martha E. Arterberry

4 MOTOR DEVELOPMENT 161


Karen E. Adolph and Sarah E. Berger

5 INFANT COGNITION 214


Leslie B. Cohen and Cara H. Cashon

Section Two: Cognition and Communication


6 ACQUIRING LINGUISTIC CONSTRUCTIONS 255
Michael Tomasello

7 EARLY WORD LEARNING 299


Sandra R. Waxman and Jeffrey L. Lidz

8 NONVERBAL COMMUNICATION: THE HANDS ROLE IN TALKING


AND THINKING 336
Susan Goldin-Meadow

xxv
xxvi Contents

Section Three: Cognitive Processes


9 EVENT MEMORY 373
Patricia J. Bauer

10 INFORMATION PROCESSING APPROACHES TO DEVELOPMENT 426


Yuko Munakata

11 MICROGENETIC ANALYSES OF LEARNING 464


Robert S. Siegler

12 COGNITIVE STRATEGIES 511


Michael Pressley and Katherine Hilden

13 REASONING AND PROBLEM SOLVING 557


Graeme S. Halford and Glenda Andrews

14 COGNITIVE SCIENCE AND COGNITIVE DEVELOPMENT 609


Frank Keil

15 CULTURE AND COGNITIVE DEVELOPMENT IN PHYLOGENETIC,


HISTORICAL, AND ONTOGENETIC PERSPECTIVE 636
Michael Cole

Section Four: Conceptual Understanding


and Achievements
16 CONCEPTUAL DEVELOPMENT 687
Susan A. Gelman and Charles W. Kalish

17 DEVELOPMENT OF SPATIAL COGNITION 734


Nora S. Newcombe and Janellen Huttenlocher

18 DEVELOPMENT OF MATHEMATICAL UNDERSTANDING 777


David C. Geary

19 SOCIAL COGNITION 811


Paul L. Harris
Contents xxvii

20 DEVELOPMENT IN THE ARTS: DRAWING AND MUSIC 859


Ellen Winner

21 EXTRAORDINARY ACHIEVEMENTS: A DEVELOPMENTAL AND


SYSTEMS ANALYSIS 905
Seana Moran and Howard Gardner

Section Five: The Perspective beyond Childhood


22 THE SECOND DECADE: WHAT DEVELOPS (AND HOW) 953
Deanna Kuhn and Sam Franklin

Author Index 995

Subject Index 1029


SEC T ION ONE

Foundations
CHAPTER 1

Neural Bases of Cognitive Development


CHARLES A. NELSON III, KATHLEEN M. THOMAS, and MICHELLE DE HAAN

WHY DEVELOPMENTAL PSYCHOLOGISTS SHOULD Visuospatial Working Memory 33


BE INTERESTED IN NEUROSCIENCE 4 Visuospatial Recognition and Recall Memory 34
BRAIN DEVELOPMENT AND NEURAL Spatial Navigation 35
PLASTICITY: A PRCIS TO OBJECT RECOGNITION 35
BRAIN DEVELOPMENT 5 Face/Object Recognition 35
Brain Development 5 Occipito-Temporal Cortex 35
Summary 12 Amygdala 38
NEURAL PLASTICITY 13 Role of Experience 38
Developmental Plasticity 14 Is There a Visuospatial Module? 39
Adult Plasticity 16 EXECUTIVE FUNCTIONS 39
What Are the Effects of Enriched Environments on Domains of Executive Function 40
Brain Development and Function? 17 Working Memory Revisited 40
What Is the Difference between Development Inhibitory Control 41
and Plasticity? 17
Attentional Control 41
THE NEURAL BASES OF
NONEXECUTIVE ASPECTS OF ATTENTION 43
COGNITIVE DEVELOPMENT 19
Alerting, Vigilance, or Arousal 43
Memory 19
Orienting 43
Disorders of Memory 26
Nondeclarative Memory 27 THE FUTURE OF DEVELOPMENTAL
COGNITIVE NEUROSCIENCE 44
SPATIAL COGNITION 31
REFERENCES 45
Mental Rotation 31
Spatial Pattern Processing 32

The goal of this chapter is to review what is known about discussion of how experience influences the develop-
the neural bases of cognitive development. We begin by ingand when appropriate, developedbrain. Within
discussing why developmental psychologists might be this discussion on experience-dependent changes in
interested in the neural bases of behavior (with particu- brain development, we briefly touch on two issues we
lar reference to cognitive development). Having estab- consider to be essential for all developmental psycholo-
lished the value of viewing child development through gists: whether the mechanisms that underlie develop-
the lens of the developmental neurosciences, we provide mental plasticity differ from those that underlie adult
an overview of brain development. This is followed by a plasticity, and more fundamentally, what distinguishes
plasticity from development.
The writing of this chapter was made possible, in part, by
With this basic neuroscience background behind us,
grants from the NIH to the first author (NS34458, NS329976),
we turn our attention to specific content areas, limiting
and from the John D. and Catherine T. MacArthur Foundation
(through their funding of a research network on Early Experi-
ourselves to domains in which there is a corpus of knowl-
ence and Brain Development) and to the second author from edge about the neural underpinnings of cognitive devel-
the NIMH (MH02024). The first author wishes to thank Lisa opment. We discuss learning and memory, face/object
Benz for assistance in literature reviews, Eric Knudsen for recognition, attention /executive functions, and spatial
sharing his insights into neural plasticity, and members of the cognition, including illustrative examples from both typi-
Laboratory of Developmental Cognitive Neuroscience. cal and atypical development. We conclude the chapter

3
4 Neural Bases of Cognitive Development

with a discussion of the future of developmental cogni- Our major argument in this regard is that our under-
tive neuroscience. standing of cognitive development will improve as the
mechanisms that underlie development are elucidated.
WHY DEVELOPMENTAL This, in turn, should permit us to move beyond the de-
PSYCHOLOGISTS SHOULD BE scriptive, black box level to the level at which the actual
INTERESTED IN NEUROSCIENCE cellular, physiologic, and eventually, genetic machinery
will be understoodthe mechanisms that underlie de-
Prior to the ascendancy of Piagetian theory, the field velopment. An example follows.
of cognitive development, such as it was, was domi- A number of distinguished cognitive developmental-
nated by behaviorism (for discussion, see Goldman- ists and cognitive theorists have proposed or at least
Rakic, 1987; Nelson & Bloom, 1997). As students of implied that elements of number concept (Wynn, 1992;
the history of psychology are well aware, behaviorism Wynn, Bloom, & Chiang, 2002), object permanence
eschewed the nonobservable; therefore, the study of the (Baillargeon, 1987; Baillargeon, Spelke, & Wasserman,
neural bases of behavior was not pursued for the simple 1985; Spelke, 2000), and perhaps face recognition
reason that neural processes could not be observed. (Farah, Rabinowitz, Quinn, & Liu, 2000) reflect what
Through the 1950s and 1960s, Piagetian theory gradu- we refer to as experience-independent functions; that is,
ally came to replace behaviorism as the dominant the- they reflect inborn traits (presumably coded in the
ory of cognitive development. Despite being a biologist genome) that do not require experience to emerge. We
by training, Piaget and, subsequently, his followers see several problems with this perspective. First, these
primarily concerned themselves with developing a arguments seem biologically implausible because they
richly detailed cognitive architecture of the mind represent sophisticated cognitive abilities; if they were
albeit a brainless mind. We do not mean this in a pejo- coded in the genome, they would surely be polygenic
rative sense, but rather, to imply that the zeitgeist of traits and would not reflect the action of a single gene.
the time was to develop elegant models of cognitive Given that we now know the human genome consists of
structures, with little regard for (a) whether such approximately 30,000 to 40,000 genes, it seems highly
structures were biologically plausible or ( b) the neuro- unlikely that we could spare the genes to separately code
biological underpinnings of such structures. (At that for number concept, object permanence, or face recog-
time, there was no way to observe the living childs nition. After all, those 40,000 genes must be involved in
brain directly.) Throughout the late 1970s and into myriad other events (e.g., the general operation of the
the last decade of the twentieth century, neo- and non- body as a whole) far more important than subserving
Piagetian approaches came into favor. Curiously, a these aspects of cognitive development.
prominent theme of a number of investigators writing A second concern about this nativist perspective is
during this time was that of nativism; we say curiously that it is not developmental. To say something is innate
because inherent in nativism is the notion of biological essentially closes the door to any discussion of mecha-
determinism, yet those touting a nativist perspec- nism. More problematic is that genes do not cause be-
tive rarely if ever grounded their models and data in haviors; rather, genes express proteins that in turn work
biological reality. It was not until the mid-1990s that their magic through the brain. It seems unlikely that
neurobiology began to be inserted into a discussion of behaviors that are not absolutely essential to survival
cognitive development, as reflected in Mark Johnsons (of the species, not the individual) have been directly
eloquent contribution to this Handbooks fifth edition coded in the genome, given the limited number of genes
(Johnson, 1998). This perspective has become more that are known to exist. Far more likely is that these
commonplace, although the field of developmental cog- behaviors are subserved by discrete or distributed neu-
nitive neuroscience is still in its infancy. (For recent ral circuits in the brain. And, these circuits, in turn,
overviews of this field generally, see de Haan & John- most likely vary in the extent to which they depend on
son, 2003a; Nelson & Luciana, 2001.) Moreover, we experience or activity for their subsequent elaboration.
have observed that it is still not clear to many develop- We make three general points: (1) The value added
mental psychologists why they should be interested of thinking about behavior in the context of neurobiol-
in the brain. This is the topic to which we next direct ogy is that doing so provides biological plausibility to
our attention. our models of behavior (to be discussed further later).
Brain Development and Neural Plasticity: A Prcis to Brain Development 5

(2) Viewing behavioral development through the lens of both molecular and electrical signals that arise sponta-
neuroscience may shed light on the mechanism(s) under- neously in growing neural networks. By spontaneously,
lying behavior and behavioral development, thereby we mean signals that are inherent in the circuitry and
moving us beyond the descriptive level to the process are entirely independent of any outside influence.
level. (3) When we insert the molecular biology of brain These molecular and electrical signals establish neural
development into the equation, a more holistic view of pathways and patterns of connections that are remark-
the child becomes possiblegenes, brain, and behavior. ably precise and that make it possible for animals to
This, in turn, permits us to move beyond simplistic no- carry out discrete behaviors beginning immediately
tions of gene-environment interactions to talk about the after birth. They also underlie instinctive behaviors
influence of specific experiences on specific neural cir- that may appear much later in life, often associated
cuits, which in turn influence the expression of particu- with emotional responses, foraging, sex, and social in-
lar genes, which influence how the brain functions and teractions. Beyond the scope of this chapter, but cer-
how the child behaves. tainly worth considering, is which human behaviors
fall into this category of instinctive. Our bias is that
these are most likely going to be behaviors that have
BRAIN DEVELOPMENT AND enormous implications for survival or reproductive
NEURAL PLASTICITY: A PRCIS fitness, such as the ability to experience fear, to avoid
TO BRAIN DEVELOPMENT predators or to experience pleasure and conversely,
the reduction of displeasure, to become attached to a
Before discussing the details of neural development, it is caregiver.
important to understand that brain development, at To return to our discussion of nativism, there is no
the species level, has been shaped over many thousands question that genetics has an enormous influence over
of generations by selective pressures that drive evolu- who we are. To a large extent, human characteristics re-
tion. According to Knudsen (2003b), this portion of bio- flect evolutionary learning, which is exhibited in pat-
logical inheritance is responsible for nearly all the terns of neural connections and interactions that have
genetic influences that shape the development and func- been shaped adaptively by evolution over thousands of
tion of the nervous system, the majority of which have generations. In addition to adaptive capacities, however,
proven to be adaptive for the success of any given genetics (through mutations) can lead to deficits in
species. These influences determine both the properties brain function, such as impairments of sensation, cogni-
of individual neurons and the patterns of neural connec- tion, emotion, and movement. We provide examples of
tions. As a result, these selective pressures delimit both in subsequent sections of this chapter.
an individuals cognitive, emotional, sensory, and motor Genes specify the properties of neurons and neural
capabilities. connections to different degrees in different pathways
There is, however, a small portion of biological and at different computational levels. On one hand, the
inheritance that is unique to the individual, resulting extent of genetic determination reflects the degree to
from the novel combination of genes that the child re- which the information processed at a particular connec-
ceives from the parents. Because there is no history to tion is predictable from one generation to the next. On
this gene pattern, any new phenotype that is produced the other hand, because many aspects of an individuals
has never been subjected to the forces of natural world are not predictable, the brains circuitry must rely
selection and is unlikely to confer any selective advan- on experience to customize connections to serve the
tage for that individual. However, this small portion of needs of the individual. Experience shapes these neural
biological inheritance is particularly important for connections and interactions but always within the con-
driving evolutionary change, as novel combinations of straints imposed by genetics.
genes or mutations that do confer a selective advantage
will increase in the gene pool, while those that result in
Brain Development
maladaptive phenotypes will die out (Knudsen, 2003a,
2003b). The construction and development of the human brain oc-
The brain develops according to a complex array curs over a very protracted period beginning shortly after
of genetically programmed influences. These include conception and, depending on how we view the end of
6 Neural Bases of Cognitive Development

Figure 1.1 Overview of human brain development, beginning the 15th prenatal week and continuing to term and then adult.

development, continuing through at least the end of ado- tion until approximately Day 26 (Sidman & Rakic,
lescence (for overviews, see Figure 1.1 and Table 1.1). 1982). The rostral portion of the tube eventually forms
Shortly after conception, embryonic tissue forms the brain, and the caudal portion develops into the
from the two-celled zygote (specifically, from the blas- spinal cord.
tocyst, the ball of cells created through multiplying Cells trapped inside the tube typically go on to
cells). The outer layer of the embryo gives rise to, comprise the central nervous system (CNS); however,
among other things, the central ( brain and spinal cord) there is a cluster of cells trapped between the outside
and peripheral nervous systems. It is this outermost of the neural tube and the dorsal portion of the ecto-
layer we will be concerned with in this chapter. dermal wall that is referred to as the neural crest.
Neural crest cells typically develop into the autonomic
Neural Induction and Neurulation nervous system (ANS).
Typical Development A fair amount is now known about the genes that reg-
The process of transforming the undifferentiated tissue ulate many aspects of brain development, including neu-
lining the dorsal side of the ectoderm into nervous sys- rulation. Much of this knowledge is based on studies of
tem tissue is referred to as neural induction. In contrast, invertebrates and vertebrates, in which alterations in
the dual processes called primary and secondary neuru- morphogenesis are observed after genes are selectively
lation refer to the further differentiation of this neural deleted ( knock-out ) or in a more recently developed
tissue into, respectively, the brain and the spinal cord method, added ( knock-in). Although at first glance,
(for recent review of neural induction and neurulation, one might be suspicious about the generalizability of
see Lumsden & Kintner, 2003). such work to the human, reassurance can be found in the
The thin layer of undifferentiated tissue that lines observation that humans share more than 61% of their
the ectoderm is gradually transformed into an increas- genes with fruit flies and 81% with mice. Of course, not
ingly thick layer of tissue that will become the neural everything we know is based on animal models: Increas-
plate. A class of chemical agents referred to as trans- ingly our knowledge of the molecular biology of brain
forming growth factors is responsible for the subsequent development is based on careful genetic analysis of nerv-
transformation of this undifferentiated tissue into nerv- ous system tissue that has failed to develop correctly.
ous system tissue (Murloz-Sanjuan & Brivanfou, 2002). The patterning of the neuroaxis (i.e., head to tail) is
What one observes morphologically is the shift from for the most part completed by about the 5th prenatal
neural plate to neural tube. Specifically, the neural week. Based on mouse studies, many of the transcrip-
plate buckles, forming a crease down its longitudinal tion factors1 responsible for this process are well known.
axis. The tissue then folds inward, the edges rise up, As reviewed by Levitt (2003), some of the genes involved
and a tube is formed. This process begins on approxi-
mately Day 22 of gestation (Keith, 1948), fusing first at 1
Transcription factors refer to the regulation of the proteins
the midsection and progressing outward in either direc- involved in the transcription of other genes.
Brain Development and Neural Plasticity: A Prcis to Brain Development 7

TABLE 1.1 Neurodevelopmental Timeline from Conception through Adolescence


Developmental Event Timeline Overview of Developmental Event

Neurulation 1824 prenatal days Cells differentiate into one of three layers: endoderm, mesoderm, and ectoderm,
which then form the various organs in the body.
The neural tube (from which the central nervous system is derived) develops from
the ectoderm cells; the neural crest (from which the autonomic nervous system is
derived) lies between the ectodermal wall and the neural tube.
Neuronal migration 624 prenatal weeks Neurons migrate at the ventricular zone along radial glial cells to the cerebral cortex.
The neurons migrate in an inside-out manner, with later generations of cells
migrating through previously developed cells.
The cortex develops into six layers.
Synaptogenesis 3rd trimester Neurons migrate into the cortical plate and extend apical and basilar dendrites.
adolescence Chemical signals guide the developing dendrites toward their final location, where
synapses are formed with projections from subcortical structures.
These connections are strengthened through neuronal activity, and connections with
very little activity are pruned.
Postnatal neurogenesis Birthadulthood The development of new cells in several brain regions, including:
Dentate gyrus of the hippocampus
Olfactory bulb
Possibly cingulate gyrus; regions of parietal cortex
Myelination 3rd trimestermiddle age Neurons are enclosed in a myelin sheath, resulting in an increased speed of action
potentials.
Gyrification 3rd trimesteradulthood The smooth tissue of the brain folds to form gyri and sulci.
Structural development Birthlate adulthood The prefrontal cortex is the last structure to undergo gyrification during uterine life.
of the prefrontal cortex The synaptic density reaches its peak at 12 months; however, myelination of this
structure continues into adulthood.
Neurochemical Uterine lifeadolescence All major neurotransmitter systems undergo initial development during uterine life
development of the and are present at birth.
prefrontal cortex Although it is not well studied in humans, it is thought that most neurotransmitter
systems do not reach full maturity until adulthood.

in dorsal patterning include members of the emx, Pax, to as spina bifida) may be due to mutations in the Pax1
and ihx families of genes, whereas nkx and dlx gene gene (Hof et al., 1996). Importantly, neural tube defects
families may play a role in ventral patterning. generally have been associated with a deficiency of folic
acid, and supplementing the diet of pregnant women
Atypical Development with this nutrient appears to have reduced the incidence
Unfortunately, errors do occur in neural induction and of these disorders in the general population.
neurulation. Neural tube defects are disorders of primary
Proliferation
neurulation and include craniorachischisis totalis (in
which there is a total failure of neural induction), anen- Once the neural tube has closed, cell division leads to a
cephaly (where the anterior portion of the neural tube massive proliferation of new neurons (neurogenesis),
fails to close completely), holoprosencephaly (where generally beginning in the 5th prenatal week, and peak-
there is one undifferentiated forebrain, rather than a ing between the 3rd and 4th prenatal months (Volpe,
forebrain that has two halves), and most commonly, 2000; for review, see Bronner-Fraser & Hatten, 2003).
myelomeningocele (where the posterior portion of the The term massive barely captures this process. It has
neural tube fails to close normally). Holoprosencephay been estimated, for example, that at its peak, several
appears to be due to mutations in the sonic hedgehog hundred thousand new nerve cells are generated each
transcription factor (ZIC2 genes; e.g., S. A. Brown minute (M. Brown, Keynes, & Lumsden, 2001). Prolif-
et al., 1998), whereas myelomeningocele (also referred eration begins in the innermost portion of the neural
8 Neural Bases of Cognitive Development

tube, referred to as the ventricular zone (Chenn & Mc- cephalon (which will give rise to the medulla). The mes-
Connell, 1995), a region that is derived from the encephalon gives rise to the midbrain.
subependeymal location that lines the neural tube. In a Finally, our knowledge of the molecular biology of
process called interkinetic nuclear migration, new neu- cell proliferation is gradually advancing. The Foxg1
ral cells travel back and forth between the inner and gene has been implicated in this process (Hanashima,
outer portions of the ventricular zone. The new cell Li, Shen, Lai, & Fishell, 2004), but undoubtedly many
first travels toward the outer portion of the ventricular other genes are involved as well.
zonethe so-called S phase of mitosiswhere DNA is
synthesized, creating a duplicate copy of the cell. Once Disorders of Neurogenesis
the S phase has been completed, the cell migrates There are a numner of examples of errors of cell prolif-
downward toward the innermost portion of the ventric- eration, although most fall into two categories: mi-
ular zone where it divides into two cells (for a generally croencephaly (small brain) and macroencephaly ( large
accessible description of these phases, see Takahashi, brain; see Volpe, 2000). Microencephaly is generally
Nowakowski, & Caviness, 2001). Each of these new thought to occur during the asymmetric period of cell
cells then begins the process again. As cells divide, a division, generally the 6th through 18th weeks of gesta-
new zone is created, the marginal zone, which contains tion. The cause of the microencephalies can be either
processes (axons and dendrites) from the cells of the genetic or environmental; examples of the latter include
ventricular zone. During the second phase of prolifera- rubella, irradiation, maternal alcoholism, excessive vi-
tion, neurons actually begin to form. However, for each tamin A, and human immunodeficiency virus (e.g., Ko-
dividing cell, only one daughter cell will continue to di- zlowski et al., 1997; Warkany, Lemire, & Cohen, 1981).
vide; the nondividing cell goes on to migrate to its final Macroencephaly is generally thought to be due to an un-
destination (Rakic, 1988). derlying genetic disorder; for example, if genes that reg-
Before turning to disorders of proliferation, three ulate normal cellular proliferation do not turn off, this
points should be noted. First, with the exception of cells failure could lead to an overproduction of new cells. As-
that comprise the olfactory bulb, the dentate region of suming survival, both micro- and macrocencephaly
the hippocampus, and possibly regions of the neocortex, generally lead to varying degrees of mental and physical
virtually every one of the estimated 100 billion neurons retardation /impairment.
we possess (Naegele & Lombroso, 2001) are of prenatal
origin (see section on postnatal neurogenesis); glia fol- Postnatal Neurogenesis
low this same general pattern, although the development With the exception of cells in the olfactory bulb, it was
of glia (with the exception of radial glial cells; see sec- assumed until recently that the nervous system at birth
tion on migration that follows) lags somewhat behind had virtually all the neurons it would ever have, and that
neuronal development. What needs to be underscored no new neurons were added. Like many aspects of neu-
about this observation is its importance in the context of roscience, we have now had to revise this view, due in
plasticity: Unlike all other cells, the brain generally does part to the advent of new methodsspecifically, the use
not make new neurons after birth, which means that the of new staining methods for looking at DNA turnover
brain does not repair itself in response to injury or dis- (e.g., 5-Bromo-2-deoxyuridine [BrdU]). Based on work
ease by making new neurons. with both human (e.g., Gage, 2000) and nonhuman
Second, as cells continue to proliferate, the general primates (e.g., Bernier, Bdard, Vinet, Lvesque, & Par-
shape of the neural tube undergoes a dramatic transfor- ent, 2002; Gould, Beylin, Tanapat, Reeves, & Shors,
mationspecifically, three distinct vesicles are formed: 1999; Kornack & Rakic, 1999) and rodents (Gould et al.,
the proencephalon (forebrain), mesencephalon (mid- 1999), there is now widespread agreement that in certain
brain), and rhombencephalon ( hindbrain). Further prolif- regions of the brain, new cells are added for many years
eration leads to the proencephalon splitting into the after birth. Where this agreement breaks down is in de-
telencephalon, which will give rise to the cerebral hemi- termining precisely which regions experience this post-
spheres, and the diencephalon, which gives rise to the natal birth of new neurons. There is little controversy
thalamus and hypothalamus. The rhombencephalon will that the dentate gyrus of the hippocampus experiences
in turn give rise to the metencephalon (from which this process; controversy exists, however, concerning re-
the pons and cerebellum are derived) and the myelen- gions of the neocortex, such as the cingulate gyrus and
Brain Development and Neural Plasticity: A Prcis to Brain Development 9

segments of the parietal cortex (for review and discus- 2002b). (For an enlightened discussion of this topic,
sion, see Gould & Gross, 2002). There is also a recent the reader is encouraged to see a special issue of the
report of postnatal neurogenesis in the amygdala, Journal of Neuroscience, 2002, vol. 22, issue 3.) Resolu-
piriform cortex, and inferior temporal cortex in nonhu- tion of this matter will prove vitally important to our
man primates (Bernier et al., 2002), although to date understanding of plasticity and to the development of
these findings have not been replicated (for a highly interventions and therapeutics; regarding the latter, for
readable review of postnatal neurogenesis generally, see example, there is strong evidence that exercise (in ro-
Barinaga, 2003). dents) leads to an up-regulation of neurogenesis in the
Particularly relevant to this chapter is the observa- dentate gyrus of the hippocampus (for a general discus-
tion that the addition of such cells can be influenced by sion of the possible link between postnatal neurogene-
experience (e.g., Gould et al., 1999; Mirescu, Peters, & sis and therapeutics, see Lie, Song, Colamarino, Ming,
Gould, 2004). For example, the number of cells pro- & Gage, 2004).
duced in the rodent dentate gyrus is increased when rats
Cell Migration
are placed in contexts in which demands are placed on
learning and memory. Similarly, the presence of the The cortex proper (arguably the seat of cognition) is
hormone prolactin in pregnant rats appears to increase formed by a process whereby newly formed cells mi-
the number of new cells in the forebrain subventricular grate out beyond their birthplace to ultimately give rise
zone. Because this zone gives rise to olfactory neurons, to a six-layered cortex. As discussed by M. Brown et al.
it has been assumed that this adaptive response may (2001), the ventricular zone (the epithelium that lines
facilitate olfactory recognition of offspring (see Shingo the lateral ventricular cavities) gives rise to cells that
et al., 2003). In contrast, stress in adulthood (e.g., the undergo cell division, with the postmitotic cell migrat-
presence of novel odors such as the smell of a fox) ing through the intermediate zone to its final point of
appears to down-regulate neurogenesis in the rat dentate destination. The cells born earliest take up residence in
gyrus. Moreover, when rodents are housed together, the preplate (the first layer of cortical neurons), which
dominance hierarchies develop, and the dominant ani- subsequently divides into the subplate and the marginal
mals produce more new neurons than the nondominant zone, both of which are derived from the cortical plate.
animals (Gould, 2003). Interestingly, if these same ani- The postmitotic cells move in an inside-out (ventricu-
mals are then housed in complex environments (see sub- lar-to-pial) direction, such that the earliest migrating
sequent section on plasticity), there is an up-regulation cells occupy the deepest layer of the cortex (and play an
of neurogenesis in the dentate, although the dominant important role in the establishment of cortical connec-
animals continue to fare better than the nondominant tions), with subsequent migrations passing through the
animals. Finally, inflammation (which commonly oc- previously formed layer(s) (note that this rule applies
curs following brain injury) can lead to a down-regula- only to the cortex; the dentate gyrus and the cerebellum
tion of dentate hippocampal cells (Ekdahl, Claasen, are formed in an outside-in pattern). At approximately
Bonde, Kokaia, & Lindvall, 2003). Collectively, experi- 20 weeks gestation, the cortical plate consists of three
ence appears to have a profound influence on postnatal layers, and by the 7th prenatal month the final contin-
neurogenesis. gent of six layers can be seen (Marin-Padilla, 1978).
Postnatally derived cells may differ in several re- There are two types of migratory patternsradial
spects from prenatally derived cells. For example, the and nonradial (generally tangential). Radial migration
former differentiate and appear to be normal in all re- generally refers to the propagation of cells from the ven-
spects, although these cells may have a relatively short tricular zone outward, or from the deepest to most su-
half-life (Gould, Vail, Wagers, & Gross, 2001). Fur- perficial layers of the cortex. Approximately 70% to
thermore, it is not yet clear how such cells might work 80% of migrating neurons use this radial pathway. In
their way into existing synaptic circuits (although contrast, cells adopting a tangential migratory pattern
Song, Stevens, & Gage, 2002, have demonstrated that (generally interneurons for the cortex and nuclei of the
postnatally derived cells both express neurotransmitter brain stem) move along a tangential (across) path.
and form synaptic connections, which is a requirement Pyramidal neurons, the major projection neurons in
for cell functioning). Finally, this revised view of neu- the brain, along with oligodendrocytes and astrocytes,
rogenesis is not without its critics (e.g., Rakic, 2002a, enlist radial glial cells to migrate through the layers of
10 Neural Bases of Cognitive Development

cortex (Kriegstein & Gtz, 2003), whereas cortical The Development of Axons
interneurons (for local connections) migrate via tangen-
Growth cones, small structures that sit on top of axons,
tial migration within a cortical layer (Nadarajah &
appear to play a key role both in developing the axon and
Parnavelas, 2002). Radial migration is particularly note-
in guiding it to its target (e.g., Raper & Tessier-Lavigne,
worthy for several reasons. First, there are different
1999; for review, see Raper & Tessier-Lavigne, 2003).
types of radial migration. Locomotion is characterized
Using cues from the extracellular matrix surrounding the
by migration along a radial glial fiber. In somal translo-
neuron (Jessell, 1988), and possibly local gene expression
cation the cell body (soma) of a cell advances toward
(Condron, 2002), the growth cone directs the axon to-
the pial surface by way of a leading process. Finally,
ward some targets and away from others.
cells that move from the intermediate zone (IZ) to the
As Raper and Tessier-Lavigne (1999) discuss, lamel-
subventricular zone (SVZ) appear to migrate using mul-
lipodia and filopodia play primary roles in axon guid-
tipolar migration (Tabata & Nakajima, 2003). A number
ance. Lamellipodia are thin fan-shaped structures,
of genes are involved in the regulation of migratory
whereas filopodia are long, thin spikes that radiate for-
movement (see Hatten, 2002, and Ridley et al., 2003 for
ward. In both cases, these structures provide the axon
recent reviews).
with the ability to move through parenchymal ( brain)
tissue micrometer by micrometer until the axon is within
Disorders of Cell Migration synapse range of a neighboring neuron. They do so by
There are many examples of disorders of cell migration sampling the local environment for molecular cues (e.g.,
(for review, see Naegele & Lombroso, 2001; Tanaka & aminin, tenascin, collagen; see Bixby & Harris, 1991; de
Gleeson, 2000; Volpe, 2000). Subcortical band hetero- Castro, 2003; Erskine et al., 2000; Hynes & Lander,
topia (also known as X-linked lissencephaly), in which 1992; Schachner, Taylor, Bartsch, & Pesheva, 1994, for
neural tissue is misplaced (e.g., cell bodies where only discussion).
axons should exist) has been associated with mental Over and above these molecular cues, there are also
retardation and epilepsy, and with the DCX gene. molecules that sit on the surface of established axons and
Schizencephaly is a case in which there is a cleft in the act as guides (Tessier-Lavigne & Goodman, 1996). An
frontal cortex, whereas in holoprosencephaly the telen- example is cell adhesion molecules (CAMs; Rutishauser,
cephalon fails to cleave, resulting in one undifferenti- 1993; Takeichi, 1995). Whether the molecules reside in
ated hemisphere. Perhaps the best known disorder of the extracelluar matrix or on the axon proper, axons are
migration is agenesis of the corpus callosum, in which guided toward (attractant cues) or away from (repellant
the corpus callosum (the major bundle of fibers that cues) neighboring neurons (Tessier-Lavigne & Good-
connects the two hemispheres) is partially or entirely man, 1996).
absent. As a class, disorders of cell migration generally
result in varying degrees of behavioral or psychological The Development of Dendrites
disturbance; a case in point may be the psychiatric dis-
Recent work has indicated that the gene calcium-regu-
order schizophrenia, which has been hypothesized to
lated transcriptional activator (CREST) plays an impor-
result from a migrational disorder (see Elvevg &
tant role in the development of dendrites (Aizawa et al.,
Weinberger, 2001).
2004). The earliest dendrites make their appearance as
thick processes with few spines (small protuberances) ex-
The Growth and Development of Processes
tending from the cell body. As dendrites mature, spine
Axons and Dendrites
number and density increase, thereby increasing the
Once a neuron has completed its migratory journey, it chance of making contact with a neighboring axon. Not
generally proceeds along one of two roads: The cell can surprisingly, dendrites grow and develop in conjunction
differentiate and develop axons and dendrites (the topic with axons. Dendritic sprouting begins to occur at ap-
of this section), or it can be retracted through the nor- proximately 15 weeks, about the same time that axons
mative process of apoptosis (programmed cell death), a reach the cortical plate. Between the 25th and 27th weeks
phenomenon that is widespread (e.g., 40% to 60% of all of gestation, dendritic spines begin appearing on both py-
neurons may die; for review, see Oppenheim & John- ramidal and nonpyramidal neurons. This sprouting con-
son, 2003). tinues to expand through the 24th postnatal month in
Brain Development and Neural Plasticity: A Prcis to Brain Development 11

some cortical regions (Mrzljak, Uylings, Kostovic, & enced a typical, full-term birth. This suggests a highly
VanEden, 1990). Additionally, there appears to be an regularized process with little influence by experience.
overproduction of both axons and dendrites during devel- As we demonstrate, however, the same cannot be said
opment with the final number achieved through competi- for synaptic pruning and the cultivating of synaptic
tive elimination. circuits, both of which are strongly influenced by the
environment.
Disorders of Axon and Dendrite Development
There are a number of environmental agents that lead to Synaptic Pruning. The process of retracting
disorders associated with errors in axonal and dendritic synapses until some final (and presumably optimal)
development, including oxygen deprivation, toxins, mal- number has been reached is dependent in part on the
nutrition, or genetic anomalies (Webb, Monk, & Nelson, communication among neurons. Pruning appears to fol-
2001). In addition, genetic disorders such as Angelman low the Hebbian principle of use/disuse: Thus, more
syndrome, fragile X syndrome, autism, and Duchenne active synapses tend to be strengthened and less active
muscular dystrophy show possible errors in dendritic synapses tend to be weakened or even eliminated
development (Volpe, 1995). (Chechik, Meilijson, & Ruppin, 1999). Neurons orga-
nize and support synaptic contact through neurotrans-
Synaptogenesis mitter receptors ( both excitatory and inhibitory) on the
Background. Synapses generally refer to the point presynaptic cell (the cell attempting to make contact)
of contact between two neurons. Depending on the re- and through neurotrophins expressed by the postsynap-
ceiving neuron, the resulting action can be excitatory tic cell (the cell on which contact is made). Synapses are
(promoting an action potential) or inhibitory (reducing modulated and stabilized by the distribution of excita-
the likelihood of an action potential). tory and inhibitory inputs (Kostovic, 1990). The adjust-
ments that are made in the pruning of synapses can
Development. The first synapses are generally ob- either be quantitative (reducing the overall number of
served by about the 23rd week of gestation (Molliver, synapses) or qualitative (refining connections such that
Kostovic, & Van der Loos, 1973), although the peak of incorrect or abnormal connections are eliminated; for
production does not occur until sometime in the 1st review, see Wong & Lichtman, 2003).
year of life (for review, see Webb et al., 2001). It is now As has been thoroughly reported in both the lay and
well known that there is a massive overproduction of scientific press, the pruning of synapses appears to vary
synapses distributed across broad regions of the brain, by area. As seen in Figure 1.2, synapse numbers in the
followed by a gradual reduction in synapses; it has been human occipital cortex peak between 4 and 8 months of
estimated that 40% more synapses are produced than age and are reduced to adult numbers by 4 to 6 years of
exist in the final (adult) complement of synapses (see age. In contrast, synapses in the middle frontal gyrus of
Levitt, 2003). The peak of the overproduction varies by the human prefrontal cortex reach their peak closer to
brain area. For example, in the visual cortex, a synaptic
peak is reached between roughly the 4th and 8th postna-
tal months (Huttenlocher & de Courten, 1987), whereas 100
90
in the middle frontal gyrus (in the prefrontal cortex) the
80
peak synaptic density is not obtained until after 15 post- 70
Percent Max

natal months (Huttenlocher & Dabholkar, 1997). 60


There is evidence that the overproduction of synapses 50
is largely under genetic control, although little is known 40
30
about the genes that regulate synaptogenesis. For exam-
20
ple, Bourgeois and colleagues (Bourgeois, Reboff, & 10
Rakic, 1989) have reported that being born prematurely 0
or even removing the eyes of monkeys prior to birth has 0 2 4 6 8 10 1 2 4 6 8 10 20 30 40 50 60 70 80
Months Years
little effect on the overproduction of synapses in the
monkey visual cortex. Thus, in both cases the absolute Figure 1.2 Synapse overproduction and elimination in the
number of synapses is the same as if the monkey experi- occipital and frontal cortices.
12 Neural Bases of Cognitive Development

one to 112 years of age, but are not reduced to adult num- such work, it was revealed that myelination begins pre-
bers until mid to late adolescence. Unfortunately, these natally with the peripheral nervous system, motor roots,
data are based on relatively few brains (thus leaving sensory roots, somesthetic cortex, and the primary vi-
open the question of the range of individual differences) sual and auditory cortices (in this chronological order).
and relatively old methods (i.e., density of synapses per During the first postnatal year, regions of the brain stem
unit area, which increases the risk that nonsynaptic and myelinate, as does the cerebellum and splenium of cor-
even nonneuronal elements may be counted, such as glial pus callosum; by 1 year, myelination of all regions of the
cells). We should expect improved figures in the years to corpus callosum is underway.
come with advances in new methods, a point that applies Although staining for myelin is undoubtedly the most
to much of the literature reviewed thus far. sensitive metric for examining the course of myelina-
tion, an obvious disadvantage to this procedure is that it
Synaptic Plasticity. The initial development of can only be done on a relatively small number of post-
synapses is generally quite tenuous. Thus, it is only after mortem brains; in addition, as is the case with human
a synapse has been activated repeatedly that it becomes synaptogenesis, it is also of concern how representative
stabilized; frequently, synapses that do not receive con- these brains are of the general population. Fortunately,
firmation are eliminated or reabsorbed (Changeux & advances in magnetic resonance imaging (MRI ) have
Danchin, 1976). Moreover, synapse stabilization also now made it possible to acquire detailed information
depends on chemical communication, such as the local about myelination in living children; importantly, sev-
release of neurotrophins by postsynaptic cells (Huang & eral longitudinal studies have examined the course of
Reichardt, 2001) or through the activation of receptors myelination from early childhood through early adult-
such as glutamate that in turn mediate postsynaptic acti- hood (Giedd, Snell, et al., 1996; Giedd, Vatuzis, et al.,
vation of cortical cells. Finally, even in the mature 1996; Jernigan, Trauner, Hesselink, & Tallal, 1991; Paus
brain, synaptic plasticity is driven by synaptic activity. et al., 1999; Sowell et al., 1999; Sowell, Thompson,
For example, reducing activity in a given neuron before Holmes, Jernigan, & Toga, 2000). The results of this
a synapse has been formed leads to a reduction in func- work paint the following picture: The pre- through
tional inputs to that neuron (from neighboring neurons); postadolescent period witnesses an increase in gray mat-
in contrast, reducing activity to a given neuron after a ter volume, followed by a decrease, whereas white mat-
synapse has been established leads to an increase in ter shows first a decrease and then an increase.2 During
synaptic input (Burrone, OByrne, & Murthy, 2002). this same age period (pre-post adolescence), particular
changes of note occur in the dorsal, medial, and lateral
Disorders of Synaptogenesis regions of the frontal lobes, whereas relatively smaller
There are several disorders of synaptogenesis, with frag- changes are observed in the parietal, temporal, and oc-
ile X syndrome being perhaps the best known. In fragile cipital lobes. This suggests, not surprisingly, that the
X, it appears that the disruption in the transcription of most dramatic changes in myelination occur in the
the FMR1 gene may be responsible for the abnormality frontal lobes through the adolescent period (for a gen-
in synapses (perhaps due in part to an exuberance in eral overview, see Durston et al., 2001).
dendritic spines, e.g., Churchill, Beckel-Mitchener,
Weiler, & Greenough, 2002). Summary

Myelination Overall, brain development begins within weeks of con-


ception and continues through the adolescent period.
Myelin is a lipid/protein substance that wraps itself
around an axon as a form of insulation and, as a result, 2
Gray matter is composed primarily of cell bodies and neu-
increases conduction velocity. Oligodendroglia produce
ropil, whereas white matter consists mostly of axons. An in-
myelin in the CNS, whereas schwann cells produce crease in gray matter, therefore, suggests that there is an
myelin in the ANS. Myelination occurs in waves begin- increase in new neurons, whereas a decrease in gray matter
ning prenatally and ending in young adulthood (and in would likely be due to an increase in apoptosis (and corre-
some regions, as late as middle age; see Benes, Turtle, sponding loss of neurons). An increase in white matter refers
Khan, & Farol, 1994). Historically, myelin was exam- to an increase in the number of myelinated axons or simply
ined in postmortem tissue using staining methods. From the amount of myelin present on such axons.
Neural Plasticity 13

This general statement does not do justice to the age- ration of an open field; (2) relative time on the open
specific changes that occur during the first 2 decades of arms of a plus maze; (3) latency to find a hidden plat-
life. Thus, the assembly of the basic architecture of the form in the Morris water maze; and (4) acoustic startle
brain occurs during the first two trimesters of fetal life, prepulse inhibition. Animals that had been cross-
with the last trimester and the first few postnatal years fostered prenatally or postnatally did not exert any ef-
reserved for changes in connectivity and function. The fects on the expected phenotype. However, mice that
most prolonged changes occur in the wiring of the brain had been prenatally and postnatally cross-fostered ex-
(synaptogenesis) and in making the brain work more ef- hibited the same behavioral phenotype as the adopted
ficiently (myelination), both of which show dramatic, strain, despite differing genetically from the adopted
nonlinear changes from the preschool period through the strain. The fact that the mice that had been cross-fos-
end of adolescence. tered prenatally did not show this effect supported the
authors contention that the effects of the combined
cross-fostering must be due to nongenetic factors, and
NEURAL PLASTICITY the powerful role of experience on gene expression.
Moving to the human, Turkheimer, Haley, Waldron,
Having established how the anatomical and physiologi- DOnofrio, and Gottesman (2003) examined IQ in 7-
cal properties of the brain are laid down over the first 2 year-old twins, a substantial number of whom were
decades of life, we may have inadvertently created the drawn from families living at or below the poverty level.
impression that these events simply unfold of their own The authors reported that the heritability of IQ varied
accordthat is, they are largely maturational in nature. nonlinearly as a function of socioeconomic standard
Although that is true to a great extent prenatally, it (SES). Thus, among twins living in impoverished envi-
vastly underrepresents the powerful role of experience ronments, a substantial portion of the variance was ac-
in sculpting the fine architecture of the brain. Before counted for by environmental factors, with relatively
providing specific examples of experience-induced little variance accounted for by genetics; in contrast,
changes in brain development and brain function, we this effect was nearly completely reversed among twins
must first provide some background. living in affluent families.
It is important to understand at the outset that expe- Collectively, these two studies demonstrate the
riences dont just happen to the brain; rather, experi- powerful role of the environment in moderating and me-
ence is the product of an ongoing, reciprocal interaction diating effects of genes on behavior. Animal studies,
between the environment and the brain. Second, experi- adopting both knockout and knock-in molecular tools,
ence has typically been defined by the properties of the coupled with careful rearing studies will undoubtedly
environment in which an individual lives. Here it must shed light on the pattern of gene expression and non-
be stressed that experience is not simply a function of genetic inheritance patterns.
the environment per se, but is the result of a complex, Within this context, any given experience can vary
bidirectional interaction between that environment and enormously under identical environmental conditions,
the developing brain. Third, experience interacts im- depending on the history, maturation, and state of the
portantly with genetics. Two examples suffice, one individuals brain. Even the experience that results from
from the rodent literature, one from the human litera- a simple physical manipulation, for example, may vary
ture. Regarding the former, Francis, Szegda, Campbell, widely depending on the background and state of the in-
Martin, and Insel (2003) cross-fostered two strains of dividual involved.
mice with one another, either prenatally (in which The relative maturity of the brain also has an enor-
hours-old embryos from one strain were implanted in mous impact on experience. Different areas of the cen-
the mothers of the other strain) and postnatally (in tral nervous system mature at different rates. A young
which newborn pups from one strain were placed with child who is exposed to information before his or her
the mothers of the other strain). The offspring of the brain is capable of processing that information will not
two original (noncross-fostered) strains served as con- have the same experience as an older child who has more
trols. All animals were tested at 3 months of age. As ex- advanced capability. A less mature brain is affected
pected, the control animals differed reliably from one largely by more fundamental features of the environ-
another on four dimensions: (1) differences in explo- ment, such as patterned light or the speech train. As the
14 Neural Bases of Cognitive Development

brain matures and changes with experience, more de- species, although we focus mainly on rodents, monkeys,
tailed aspects of the environment influence it. Thus, as and humans. Finally, we limit our discussion to percep-
an individuals brain changes, particularly during the tual and cognitive functions, and do not discuss aspects
early developmental periods, the same physical environ- of plasticity from social and emotional development.
ment can result in very different experiences.
Certain properties of the brain differ dramatically Developmental Plasticity
across individuals and within individuals over time.
Therefore, because experience is defined as the inter- Auditory Function
action of the brain with the environment, a scientific de- Sur and Leamey (2001) published a rather amazing ex-
scription of an experience must include a description of ample of plasticity in the auditory system. Normal reti-
the background, developmental stage, and state of the nal inputs pass through the lateral geniculate nucleus
brain as well as a description of the specific experience (LGN) of the thalamus on their way to the visual cortex,
to which the individual is exposed. By the same token, whereas normal auditory inputs pass through the medial
an analysis of the effect of an experience also must take geniculate nucleus (MGN) on their way to the auditory
into account variability in those same factors. cortex. In this study, retinal projections in the young fer-
The impact of experience on the brain is not constant ret were routed through the MGN. When electrophysio-
throughout life. As the brain passes through its different logical recordings were made from the auditory cortex,
developmental stages, its sensitivity to experience varies normal visual responses were observed; for example,
accordinglyhence the concept of sensitive periods (see these auditory cells showed orientation selectivity to vi-
Knudsen, 2003a). Early experience often exerts a partic- sual stimuli.
ularly strong influence in shaping the functional proper- Recently, Cheng and Merzenich (2003) reported that
ties of the immature brain. Many neural connections pass infant rats reared in an environment of continuous, mod-
through a period during development when the capacity erate-level noise were delayed in the organization of
for experience-driven modification is greater than it is in the auditory cortex; specifically, the auditory receptive
adulthood. Thus, individual capabilities reflect the com- fields differed from what would be expected under
bined influences of both evolutionary learning and per- normal auditory rearing conditions. When these rats
sonal experience. reached young adulthood, the auditory cortex appeared
In the following sections, we will deconstruct plastic- to be very similar to infantsthat is, the auditory recep-
ity into two types: developmental and adult. By this, we tive fields had not yet crystallized into the adult pattern
simply mean plastic processes that operate during the and had retained their juvenile appearance. To test the
time of brain development (presumably the first 2 hypothesis that rearing the rats in continuous noise from
decades of life) versus the postdevelopmental period birth delayed the closing of the critical period for orga-
(when the brain is certainly capable of changing but not nizing the auditory cortex, the authors exposed some of
in the same fashion as the childs brain). We revisit this the same rats to a 7 kHz tone train for 2 weeks. This led
simplistic view at the end of this section by discussing to a reorganization of the auditory cortex, supporting
whether the plastic processes that operate during devel- the hypothesis that degraded acoustic inputs early in life
opment are the same as or different from those that op- delayed the organizational maturation of the auditory
erate during adulthood. cortex. This led the authors to speculate that early expo-
We, and others, have recently provided overviews to sure to abnormal auditory inputs (e.g., loud noise) could
the myriad ways experience influences the developing contribute to some auditory and/or linguistic delays in
and developed brain (Black, Jones, Nelson, & Gree- human children.
nough, 1998; Cline, 2003; de Haan & Johnnson, 2003a; It has been known for some time that human adults
Grossman, Churchill, Bates, Kleim, & Greenough, 2002; have great difficulty in discriminating speech contrasts
Huttenlocher, 2002; Knudsen, 2003b, in press; Nelson, from various nonnative languages. This differs dramati-
1999, 2000a, 2002; Nelson & Bloom, 1997). Our pur- cally from the ceiling-level ability to discriminate
pose is simply to provide a brief overview to this topic, speech contrasts from their own (English) language. Im-
beginning with a discussion of developmental plasticity portantly, between 6 and 12 months of life, the infants
and proceeding to a discussion of adult plasticity. ability to discriminate phonemes from languages to
Within each of these sections, we draw on a range of which they are not exposed declines dramatically (for
Neural Plasticity 15

review, see Saffran, Werker, & Werner, Chapter 2, this been exposed to either sign language early in life
Handbook, this volume; Werker & Vouloumanos, 2001). (among those born congenitally deaf ) or individuals ex-
It appears however, that the door does not shut com- posed to spoken language early in life but who subse-
pletely on retaining the ability to discriminate nonnative quently lost their hearing than among congenitally deaf
contrasts; for example, Kuhl and colleagues (Kuhl, Tsao, individuals not exposed to either signed or spoken lan-
& Liu, 2003) have reported that if before 12 months of guage early in life. Furthermore, Newman, Bavelier,
age infants are given additional experience with speech Corina, Jezzard, and Neville (2002) have reported that
sounds in a nonnative language, this ability is retained. ASL is represented in the same location as spoken lan-
Collectively, these speech data have been interpreted to guage only among individuals who acquired ASL before
suggest that the speech system remains open to experi- puberty. Thus, although there does not appear to be a
ence for a certain period, but if experience in a particu- sensitive period for acquiring ASL in general, there
lar domain (such as hearing speech contrasts in different does appear to be a sensitive period for representing
languages) is not forthcoming, the window begins to ASL in the same regions of the brain as spoken lan-
close within the 1st year of life. guage, and for signing with a high level of proficiency.

Linguistic Function Visual Function


There has long been great controversy over the sensitive The development of stereoscopic depth perception is
period for acquiring a second (or third or fourth) lan- made possible by the development of ocular dominance
guage. Dehaene et al. (1997) have reported that the neu- columns, which represent the connections between each
ral representation of a second language is identical to eye and layer IV of the visual cortex. If the two eyes are
that of a first language in truly bilingual individuals. not aligned properly, or cant move together (vergence
However, if linguistic competence of the second lan- movements), then the ocular dominance columns that
guage is not as strong as the first, then the functional support normal stereoscopic depth perception fail to de-
neuroanatomy ( based on PET data) is different. Since velop normally. If this situation is not corrected by 4 to 5
the bilinguals studied in this work had acquired their years of agewhen the number of synapses begins to
second language early in life, the initial conclusion was reach adult valuesthe child will not develop normal
that the second language needed to be acquired early to stereoscopic vision. Thus, normal visual input to an in-
be represented in the brain in the same location as the tact visual system during a sensitive period is necessary
first language. This conclusion has recently been chal- for the development of binocular vision.
lenged, however. Perani et al. (1998) asked whether it Another example of sensitive periods for visual de-
was the age at which the second language was acquired velopment can be found in an elegant series of studies by
that was the critical variable, or the proficiency with Maurer and colleagues (Maurer, Lewis, Brent, & Levin,
which this language was spoken. In this study, the age at 1999). These authors have reported on a longitudinal
which the second language had been mastered covaried study of children with cataracts. A particularly attrac-
with how well this language was spoken. The authors re- tive feature of this work is that some infants are born
ported that it was the latter dimension that was most with cataracts, and others acquire them a few years
critical. Thus, regardless of when the second language after birth; moreover, some have their cataracts re-
was acquired, speaking this language with equal profi- moved in the first months of life, and others undergo
ciency as the first language led to shared neural repre- this procedure a few years later. The authors have re-
sentation for both languages. Importantly, similar ported that among infants born with cataracts that are
findings have been obtained with congenitally deaf indi- removed and new lenses placed within months of birth,
viduals who are proficient in sign language; thus, the even just a few minutes of visual experience leads to a
areas of the brain involved in sign are identical to those dramatic change in visual acuity. The longer the
of hearing speakers using spoken language (Petitto cataracts are left untreated, reducing the amount of vi-
et al., 2000). sual experience, the less favorable the outcome. Note,
Is there a critical period for acquiring American sign however, that although most visual functions undergo
language (ASL)? Mayberry, Lock, and Kazmi (2002) dramatic improvements following early cataract re-
have reported that proficiency in sign language among moval, some aspects of face processing remain impaired
deaf individuals was far greater among subjects who had for years (Le Grand, Mondloch, Maurer, & Brent,
16 Neural Bases of Cognitive Development

2001). From this observation comes the idea of a differ- have reported (using MEG) that the somatosensory cor-
ential sensitive period: specifically, the sensitive period tex representing the fingers of the left hand (used for
for visual acuity may differ from that of face processing. fine finger board movements) in highly proficient
stringed instrument players is larger than (a) the analo-
Learning and Memory Function gous area in the opposite hemisphere representing the
Hormones have long been associated with mediating right hand (used to bow) and ( b) the analogous area in
cognitive differences between males and females due to nonmusicians. Similarly, Stewart et al. (2003) reported a
their organizing effect on the brain. This association study in which musically naive subjects were scanned
can even be observed prenatally. For example, Shors and using functional magnetic resonance imaging (fMRI )
Miesegaes (2002) capitalized on previous findings in before and after they had been taught to read music and
which exposure to stressful and traumatic events en- play keyboard. When subjects played melodies from mu-
hances new learning in adult male rats, but impairs new sical notation after training, activation was seen in a
learning in adult female rats. In the present study, the cluster of voxels within the right superior parietal cor-
authors performed two experiments. In the first, male tex consistent with the view that music reading involves
rats were castrated at birth, whereas female rats were spatial sensorimotor mapping. Third, Draganski et al.
injected with testosterone; both groups were then tested (2004) have recently reported that individuals who had
on a hippocampal-dependent learning task (trace eye- been given 3 months to learn to juggle showed increased
blink conditioning). The castrated male rats still exhib- neural activation (as revealed by fMRI ) bilaterally in
ited enhanced learning, as did the testosterone-injected the midtemporal area and in the left posterior intrapari-
females (opposite the normal pattern of response, in etal sulcus; importantly, 3 months after this group of
which reduced learning is typically observed for fe- jugglers stopped juggling, there was a decrease in acti-
males). In the second experiment, pregnant female rats vation. There was a dose/response effect between jug-
were injected with a testosterone antagonist, thereby de- gling performance and gray matter changes. Finally,
priving the offspring of testosterone. The male and fe- nonjugglers over the 6-month period showed no changes
male offspring were subsequently tested as adults on the in brain activation.
same learning task following exposure to a stressor
( loud noise bursts). Male and female rats now both ex-
Visual and Auditory Function
hibited reduced learning, pointing to the powerful role
of early experience with testosterone. As was the case for motor functions, just a few examples
will be provided for adult visual plasticity. First, it
has been well established that congenitally blind indi-
Adult Plasticity viduals show activation of the visual cortex when read-
ing Braille or performing other tactile discrimination
For many years, it was commonly believed that once the functions (Lanzenberger et al., 2001; Sadato et al.,
brain was fully developed (thought to be the end of ado- 1998; Uhl, Franzen, Lindinger, Lang, & Deecke, 1991).
lescence), its ability to be molded by experience and to Interestingly, a Braille-reading woman blind since birth
recover from injury was greatly limited. This view has who subsequently suffered a bilateral stroke of the vi-
been turned on its head in recent years; indeed, research sual cortex in adulthood lost the ability to read Braille
on so-called adult plasticity is currently receiving although other somatosensory functions remained intact
tremendous attention in the field of neuroscience. In (Hamilton, Keenan, Catala, & Pascual-Leone, 2000).
this section, we provide an overview to several domains And, Sadato, Okada, Honda, and Yonekura (2002) have
in which the adult brain is now known to be plastic. now reported that among Braille-reading individuals
who lost their sight after age 16, activation of the V1
Motor Learning cortex was absent, whereas such activation was present
Countless examples could be offered about the plasticity among those who lost their sight before age 16. Such
of motor functions, many of which have been reviewed findings are consistent with a sensitive period for visual
elsewhere (e.g., Nelson, 1999, 2000a; Nelson & Bloom, function. Finally, Finney, Fine, and Dobkins (2001) have
1997). Three will suffice. First, Elbert and colleagues reported that deaf individuals show activation of the au-
(Elbert, Pantev, Wienbruch, Rockstroh, & Taub, 1995) ditory cortex to visual stimuli.
Neural Plasticity 17

Collectively, it appears that the somatosensory, vi- Overall, there are now ample data to suggest that
sual, and auditory cortices are capable of reorganizing learning and memory are correlated with changes in the
well into adulthood, although there is some hint that there brain at multiple levels, from changes in pre- and postsy-
may still be a sensitive period for such reorganization. naptic functioning mediated by glutomate receptors to
the molar changes at the level of anatomy. There is no
sense that there is a sensitive period for learning and
Learning and Memory Function
memory to occur (for a tutorial on the development of
No area of plasticity has received more attention in the learning and memory, see Nelson, 1995, 2000b). Indeed,
neuroscience literature than that of learning and mem- there is some sense that activities that engage the learn-
ory. It has been known for over 30 years, for example, ing and memory system may confer some protection on
that rats raised in complex laboratory environments lifelong learning and memory function (see Nelson,
(those containing lots of toys and social contacts) per- 2000b, and later for discussion).
form certain cognitive tasks better than rats reared in
isolation (e.g., Greenough, Madden, & Fleischmann,
1972). Looking closely at the brain reveals that (a) sev- What Are the Effects of Enriched Environments
eral regions of the dorsal neocortex are heavier and on Brain Development and Function?
thicker and have more synapses per neuron; ( b) dendritic
Although a great deal is known about the effects of dep-
spines and branching patterns increase in number and
rivation on brain development and function, very little is
length; and (c) there is increased capillary branching,
known about enrichment. Indeed, in the work on com-
thereby increasing blood and oxygen volume (for exam-
plex environments discussed earlier, such environments,
ples of both original findings and overviews, see Black,
while enriched relative to typical laboratory conditions,
Jones, Nelson, & Greenough, 1998; Greenough & Black,
are still impoverished relative to real-world environ-
1992; Greenough et al., 1972; Greenough, Juraska, &
ments. Over and above the challenge of defining what is
Volkmar, 1979; Kolb & Whishaw, 1998).
meant by enrichment (e.g., what may be considered en-
Moving up to the systems level of the nervous system,
riched in one context may be impoverished in another),
Erickson, Jagadeesh, and Desimone (2000) have re-
there is also the challenge of employing measures that
ported a study in which multicolored complex stimuli
possess the sensitivity to detect such effects. One exam-
(some familiar, some novel) were presented to monkeys.
ple of enrichment effects on human brain function can
Neuronal activity was recorded from the perirhinal cor-
be found in a recent study by Colcombe et al. (2004).
tex, an area known to be strongly involved in episodic
These authors have reported that highly aerobically fit
memory (see later section on memory). After only 1 day
older adults or providing aerobic training to older adults
of experience viewing the stimuli, performance of
who were previously unfit was associated with improved
neighboring neurons became highly correlated, whereas
performance on a test of executive function, and with
viewing novel stimuli revealed little correlated neuronal
increases in task-related activity in the superior and
activity. These findings suggest that visual experience
middle frontal gyrus, the superior parietal lobule, and
leads to functional changes in an area of the brain known
reduced activity in the anterior cingulate cortex, regions
to be involved in memory.
all associated with attentional control. This not only
Perhaps the most dramatic example of experience-in-
speaks to the plasticity of these cortical regions well
duced changes in the neural architecture underlying
into the life span, but also demonstrates that enrichment
memory is the now well-known London cabdriver
confers benefits on brain function over and above typi-
study. Here structural MRI scans were obtained from
cal living conditions (in which many older Americans
London cabdrivers, all highly expert at navigating the
lack cardiovascular fitness).
streets of London. The authors (Maguire et al., 2000)
reported that the posterior hippocampus ( hypothesized
to serve as the storage location for spatial representa-
What Is the Difference between Development
tions) of these drivers was larger than in a comparison
and Plasticity?
group. Not surprisingly, there was also a positive corre-
lation between hippocampal volume and the amount of At a deep neurobiological level, one might argue that
experience the driver had. the molecular processes that underlie plasticity (e.g.,
18 Neural Bases of Cognitive Development

changes in the neurochemical profile of a synapse, gyrus may contain a total of 1.5 million cells, with
anatomical changes such as growth of an axon or 250,000 of these being generated anew each month).
sprouting new dendritic spines) are no different in the Another way to illustrate the difference between de-
developing brain than they are in the mature brain. velopmental versus adult plasticity can be found at a
Once the cellular machinery is in place, it generally more systems or behavioral level. The goal of infancy is
operates the same regardless of the age of its container. to develop neural circuits in the service of some behav-
Similarly, the sprouting of new dendritic spines in re- ior. In the adult, however, these systems are already in
sponse to complex environments may be the same place and must simply be reconfigured for a different,
regardless of the age of the brain involvedand the albeit related, purpose such as acquiring a second or
molecular events that underlie changes in dendritic third language. Thus, second language learning may in
function are also likely the same or vastly similar. fact be fundamentally different from first language
However, there are still a number of fundamental dif- learning because, in the former case, there is already a
ferences in the way plastic processes might work in the scaffold on which to build, whereas in the latter case
developing brain and in the developed brain. there is not. This example builds on our first point, as
First, the local cellular, anatomical, and metabolic second language learning may involve either rewiring or
environment in which plastic processes operate are extending existing neural circuits to a new, albeit re-
vastly different early in the life span versus later in the lated, domain, whereas learning a first language surely
life span. Thus, the newborn brain has countless more involves new wiring.
neurons and synapses than an adult brain, and many of There is a third point that illustrates the possible dif-
these are not yet committed to particular functions. ference between developmental versus adult plasticity:
Thus, an axon that is growing toward its target has a specifically, whether there is a difference between de-
very different terrain to negotiate in the newborn brain velopment versus plasticity. Developmental psycholo-
than in the adult brain. Similarly, modifying already- gists are familiar with the principles that underlie
committed synapses (i.e., rewiring the brain) is quite behavioral change across age. Yet neuroscientists are
different from committing synapses to a particular cir- also aware that the molecular, anatomic, physiologic,
cuit for the first time (i.e., wiring the brain). and neurochemical changes that underlie changes in
An example of this point can be found in a paper behavior generally are likely operating in the back-
by Carleton, Petreanu, Lansford, Alvarez-Buylla, and ground. Thus, if the processes that mediate changes in
Lledo (2003). These authors report that the development behavior in general at the cellular/atomical level are
of the electrophysiological properties of neurons born in the same as those that mediate changes across age, then
adulthood differ from those born pre- or perinatally. For what is the difference between development and plas-
example, spiking activity (the excitability of the cell) ticity? We would contend that these terms are funda-
was delayed in late-born versus early-born cells; that is, mentally the same; the difference, however, is that we
this was not observed until the cell was nearly fully ma- view plasticity as lifelong, whereas we view develop-
ture. The authors argue that this could be due to the ment as something that happens over the first 2 (or so)
need to make sure the late-born new cell does not inter- decades of life. If this is true, it then raises the possi-
rupt the existing circuit until it is ready to be part of that bility that the plastic processes that exist throughout
circuit. Suffice to say, this illustrates a fundamental dif- life may be different from those operating to direct de-
ference between neurogenesis in the developing brain velopment during the first 2 decades of life (and surely
versus the developed brain. there are differences within this developmental plastic
A second example of this point can be found in the process within this timeframe).
work on postnatal neurogenesis. For example, Gould Admittedly, our answer to the question of whether
(2003) has reported that the number of prenatally de- developmental and adult plasticity are different is not
rived cells tends to be fairly stable across development, fully satisfying; we remain somewhat on the fence, al-
with only a slight reduction with aging. In contrast, though leaning toward arguing for a difference. It is our
cells derived in adulthood (see earlier section on post- hope that a more satisfying answer to this question will
natal neurogenesis) tend to be massively produced be found in the coming years as research in neural plas-
but relatively short-lived (e.g., the adult rodent dentate ticity increases. But, it is important to ponder this ques-
The Neural Bases of Cognitive Development 19

tion as developmental theory gradually becomes more pus increases in size from the first to the 2nd year of life
neurobiological in its orientation. as the bird reaches sexual maturity, whereas no such in-
crease is observed among nonmigrants. Thus, the ability
of the hippocampus to adapt to the birds environmental
THE NEURAL BASES OF niche represents an example of the evolutionary signifi-
COGNITIVE DEVELOPMENT cance of memory.
A similar and equally impressive example of mem-
Having laid the foundation for how the human brain de- ory can be found in a report by Tomizawa et al. (2003).
velops, we now turn our attention to the neural bases of These authors note that the hormone oxytocin, which
specific cognitive functions. As is the case with many of has typically been associated with inducing labor and
the topics targeted for review in this chapter, the reader facilitating caregiving behavior, has also been associ-
is encouraged to consult more comprehensive treatises ated with two cognitive functions: social recognition
(e.g., de Haan & Johnson, 2003b; Johnson, 2001; Nelson (Ferguson, Aldag, Insel, & Young, 2001; Ferguson
& Luciana, 2001). Also note that because the focus of et al., 2000) and improved spatial memory (Kinsley
our discussion is on the neural bases of cognitive devel- et al., 1999). Tomizawa et al. (2003) demonstrated that
opment, we restrict our discussion to the literature that in hippocampal slices perfused with oxytocin, long-
directly relates a specific cognitive ability to brain de- term potentiation (LTP)3 was facilitated, and among
velopment; the basic cognitive developmental behavioral hippocampal slices taken from mice that had previ-
literature is thoroughly reviewed in other chapters con- ously given birth (and in which no exogenous oxytocin
tained in this and previous volumes. was administered), CREB phosphorylation (the gene
expression cascade believed to contribute to memory
formation) was increased. Moreover, spatial memory
Memory could be improved in mice that had never given birth
but were treated with oxytocin. In contrast, an oxy-
Why Is Memory Important? tocin antagonist administered to mice who had given
We begin our discussion of brain and cognitive func- birth impaired spatial memory.
tions with a discourse about memory and lead with the A final example of the adaptive significance of mem-
question of why memory is important. ory may be found in the observation that the young of
Memory is a cornerstone ability on which many gen- many species appear to have a proclivity to attend dis-
eral cognitive functions are assembled. Our knowledge proportionately to novelty; indeed, Nelson (2000b) has
of the world is predicated on a store of information that speculated that (a) this proclivity may be obligatory
is acquired through acts of memory. Thus, it would seem and ensures that infants continually add to their knowl-
important to delineate the developmental trajectory of edge base, and ( b) this preference for novelty early in
this ability, particularly since our understanding of the life may set the stage for the lifelong facility for mem-
cognitive and neural processes that underlie intelligence ory formation.
is not well established. In light of the ontogenetic and evolutionary impor-
The abilities to encode and subsequently recall infor- tance of memory, let us turn to what is known about the
mation are likely ones that have been conserved across neural bases of memory. Before addressing this question,
species and have proved reproductively adaptive. Three a brief tutorial on what we mean by memory is in order.
examples suffice. First, in migratory birds, it is impera-
Memory Systems
tive that a bird return to its breeding ground to give
birth to and provide care for the next generation. How Tulving (1972) opened the door to questioning the pre-
does a bird remember the location of the breeding vailing dogma that memory was a unitary trait by pro-
ground after many months? Mettke-Hofmann and Gwin- posing that there are two memory systems, which he
ner (2003) demonstrated that migrant birds can recall termed semantic and episodic. Tulvings argument was
the location of a particular feeding site for 12 months,
whereas nonmigrant birds can do so for only 2 weeks. 3
Based on Hebbian principles, LTP is thought to reflect one
More importantly, among migrant birds, the hippocam- of the molecular mechanisms underlying memory formation.
20 Neural Bases of Cognitive Development

based on behavioral dissociations, although over the Like many dichotomies, this distinction between
next 30 years this was augmented by experimental work memory systems is not wholly satisfying, as there are
with nonhuman animals and neuroimaging data from many gray areas that fall between and within systems.
human adults. However, perhaps the turning point was As Cycowicz (2000) discusses, the distinction between
the data derived from patient H. M. As Eichenbaum and recognition memory (a type of explicit memory) and
colleagues discuss, H. M. (who underwent a bilateral repetition priming (a type of implicit memory) has more
resection of the temporal lobes for relief of intractable to do with the instructions issued to the subject than
epilepsy; see Scoville & Milner, 1957, for original anything inherent in memory itself. Thus, in tests of
report, or Corkin, 2002, for recent report) provided recognition memory, subjects are explicitly asked to
crucially important data to the argument in favor of mul- identify previously experienced stimuli or events (e.g.,
tiple memory systems (Eichenbaum, 2002; Eichenbaum by verbal response, pressing a button), whereas in tests
et al., 1999). For example, although H. M. suffers from a of repetition priming, subjects are typically asked to
severe deficit in encoding new facts, he can learn new perform an incidental task (e.g., press button A if the
motor skills, although he has no conscious recollection stimulus is oriented normally and button B if the stimu-
of having done so. lus is inverted), and priming is inferred by speeded re-
Since Tulvings seminal paper in 1972, countless action times to previously experienced (albeit often
studies of rodents, monkeys, and humans have collec- unconsciously) stimuli. In addition, there can be overlap
tively pointed to a major dissociation of two types of in the neural circuits that underlie each system (e.g., the
memory: Explicit or declarative memory, on one hand, visual cortex is required in both visual recognition
and implicit or nondeclarative memory (also sometimes memory and repetition priming for visual stimuli). Still,
referred to as procedural memory) on the other. The the field of memory has by and large accepted this dis-
standard definition of the former generally refers to tinction, various pitfalls notwithstanding.
memory for facts and events that can be brought to con-
The Development of Memory Systems
scious awareness and that can be expressed explicitly. In
Some Background
contrast, implicit or nondeclarative memory generally
refers to the acquisition of skills or procedures that are Drawing on data from both juvenile and mature nonhu-
expressed nonconsciously through motor activity or man primates, neuropsychological and neuroimaging
changes in processing speed. Not surprisingly, there are data with adult humans, and the limited neuroimaging
different types of memory within each of these systems; literature with developing humans, Nelson (1995)
thus, traditional recognition and recall memory fall proposed that explicit qua explicit memory begins to
under the rubric of explicit memory, whereas both prim- develop some time after the first half year of life, as in-
ing and procedural learning fall under the rubric of im- ferred from tasks such as deferred imitation, cross-
plicit memory. modal recognition memory, delayed nonmatch-to-sample
Additional evidence for the segregation of these two (DNMS), modified oddball designs in which event-
memory systems can be found in the neuroscience liter- related potentials (ERPs) are recorded, and preferential
ature. For example, at least in the adult, the explicit looking and habituation procedures that impose delays
memory system appears to involve select neocortical between familiarization (or habituation) and test. As is
areas (e.g., visual cortex for visual explicit memory), the the case with the adult, this system depends on a distrib-
cortical areas that surround the hippocampus (e.g., en- uted circuit that includes neocortical areas (such as the
torhinal cortex, parahippocampal gyrus), and the hip- inferior temporal cortical area TE), the tissue surround-
pocampus proper. In contrast, nondeclarative memory ing the hippocampus (particularly the entorhinal cortex),
involves circuits specific to the particular type of mem- and the hippocampus proper. However, Nelson also pro-
ory (although the medial temporal lobe tends not to be posed that the development of explicit memory is pre-
involved). The acquisition of habits and skills appears to ceded by an earlier form of memory referred to as
depend on the neostriatal system, whereas sensory-to- preexplicit memory. What most distinguishes preexplicit
motor adaptations and the adjustment of reflexes appear from explicit memory is that the former (a) appears at or
to depend disproportionately on the cerebellum (for re- shortly after birth, ( b) is most evident in simple novelty
cent reviews, see Eichenbaum, 2003). preferences (often reflected in the visual paired compar-
The Neural Bases of Cognitive Development 21

TABLE 1.2 Neural Bases of Memory Development


General System Subsystems Tasks Neural Systems Related to Tasks
Implicit memory Procedural learning Serial reaction time task Striatum, supplementary motor
(nondeclarative memory) association, motor cortex, frontal cortex
Implicit memory Procedural learning Visual expectation paradigm Frontal cortex, motor areas
(nondeclarative memory)
Implicit memory Conditioning Conditioning Cerebellum, basal ganglia
(nondeclarative memory)
Implicit memory Perceptual representation Perceptual priming paradigms Modality dependent: parietal cortex,
(nondeclarative memory) system occipital cortex, inferior temporal cortex,
auditory cortex
Explicit memory system Preexplicit memory Novelty detection in habituation Hippocampus, possibly entorhinal cortex
and paired comparison tasks
Explicit memory system Semantic (generic Semantic retrieval, word priming, Left prefrontal cortex, anterior cingulate
knowledge) and associative priming
Explicit memory system Semantic (generic Semantic retrieval, word priming, Hippocampal cortex
knowledge) and associative priming
Explicit memory system Episodic (autobiographical) Episodic encoding Left prefrontal cortex, left orbitoprefrontal
cortex
Explicit memory system Episodic (autobiographical) Recall and recognition Right prefrontal, anterior cingulate,
parahippocampal gyrus, entorhinal cortex

ison procedure), and (c) is largely dependent on the hip- baums model (2003) is nicely representative of the field,
pocampus proper. and of our own thinking about development.4 According
These proposals (subsequently updated and elabo- to Eichenbaum, the parahippocampal region (composed
rated by Nelson & Webb, 2003) were built less on direct of the entorhinal, perirhinal, and parahippocampal cor-
visualization of brain-behavior relations than on the in- tices) is disproportionately involved in mediating the rep-
tegration of data from many sources. This renders the resentation of isolated items, and is capable of holding
model a useful heuristic, albeit one that would benefit these items in memory for short periods (e.g., several
from more data and less speculation. The challenge is minutes). In contrast, the hippocampus proper appears to
that relatively little is known about the development of play a more prominent role in mediating the relational
the circuitry purported to be involved in different mem- and representational properties of the stimuli: For exam-
ory systems and different types of memory; similarly, ple, during the time the parahippocampal region is hold-
there are relatively few investigators using neuroimag- ing individual items in memory, the hippocampus would
ing tools of any sort to examine brain-memory relations. be comparing these items to others already in memory;
Nevertheless, advances over the past decade in the test- similarly, the hippocampus will be involved in extracting
ing of developing nonhuman and human primates have similar information across multiple exemplars or across
provided much needed additional information. multiple contexts. Data from studies of human adults and
Table 1.2 summarizes our current thinking about the
4
neural bases of memory development. We focus first on Extending adult models of memory function to infants or
the development of explicit memory, and then turn our children is problematic for many reasons, not the least of
attention to implicit memory. which is that we are uncertain about whether the structure-
function relations that hold in the adult will hold in the devel-
Current FindingsExplicit Memory in the oping child. Nevertheless, it is more useful to begin with the
adult configuration and then modify and expand accordingly
Mature Brain
than to start from the ground up. As attractive as the latter is,
Although there is no consensus in the adult literature on there are currently too few studies on brain-memory relations
the neural circuitry involved in explicit memory, Eichen- to permit adequate model building.
22 Neural Bases of Cognitive Development

developing and adult monkeys with select hippocampal pocampal damage shows impairments on the VPC task
lesions indicate that the hippocampus also mediates nov- but relatively intact recognition memory (Pascalis,
elty preferences (e.g., Manns, Stark, & Squire, 2000; Hunkin, Holdstock, Isaac, & Mayes, 2004). Together,
McKee & Squire, 1999). Finally, the neocortex is likely these results suggest that recognition memory per se
the storage site for long-term information, and in the case may be mediated by extra-hippocampal tissue (a point
of the frontal cortex, the site that facilitates storage to which we will return), whereas novelty preferences
through mnemonic strategies (e.g., chunking, and so on; are likely mediated by the hippocampus proper. The
for review of the neuroimaging literature on declarative monkey data are only partially consistent with this
memory, see Yancey & Phelps, 2001). view, as they suggest that the hippocampus and sur-
rounding cortex all play a role in novelty preferences,
Current FindingsExplicit Memory
although differentially so, at least in the adult.
in the Developing Brain Further evidence for the role of the hippocampus in
As is the case with studies of the mature organism, it is encoding the relations among stimuli (versus encoding
critical to consider the task being used to evaluate mem- individual stimuli, which may be the domain of the
ory in deriving an understanding of what structure or parahippocampal region) can be found in a recent paper
circuit is involved. As discussed by Nelson (1995), Nel- by A. J. Robinson and Pascalis (2004). These authors
son and Webb (2003), and most recently, Hayne (2004), tested groups of 6-, 12-, 18-, and 24-month-old infants
the very same task, used in different ways, could impose using the VPC. Rather than evaluate memory for indi-
different task demands on the subject and recruit differ- vidually represented stimuli, the authors required in-
ent underlying circuitry (see discussion of the DNMS fants to encode the properties of stimuli in context.
task later in this chapter). During familiarization, stimuli were presented against
There is now very good evidence from the monkey one background, and during testing the same stimulus,
literature that lesions of the hippocampus lead to dis- paired with a novel stimulus, was presented against the
ruptions in visual recognition memory, at least under same or a different background. The authors report that
certain circumstances, and at least as inferred from although all age groups showed strong novelty prefer-
novelty preferences. Pascalis and Bachevalier (1999) ences when the background was the same, only the 18-
have reported that in monkeys, neonatal lesions of the and 24-month-olds showed novelty preferences when the
hippocampus ( but not amygdala; see Alvarado, Wright, backgrounds were changed. This would suggest that this
& Bachevalier, 2002) impair visual recognition mem- particular function of the hippocampusstudying the
ory as tested in the Visual Paired Comparison (VPC) relations among stimuli or encoding stimuli in context
procedure, at least when the delay between familiar- is somewhat slower to develop than recognizing stimuli
ization and test is more than 10 seconds. Nemanic, Al- in which the context is the same.5 Support for this claim
varado, and Bachevalier (2004) have reported similar can be found in the developmental neuroanatomy litera-
effects in adult monkeys. Specifically, lesions of the ture. Specifically, although the hippocampus proper, the
hippocampus, perirhinal cortex, and the parahip- entorhinal cortex, and the connections between them are
pocampal cortex all impair performance on the VPC, known to mature early in life (e.g., Serres, 2001), it is
although differentially so. Thus, lesions of the perirhi- also known that the development of dentate gyrus of the
nal cortex, parahippocampal cortex, and hippocampus hippocampus matures more slowly (see Serres, 2001), as
proper lead to impairments when the delay between fa- does the perirhinal cortex (see Alvarado & Bachevalier,
miliarization and test exceeds 20 seconds, 30 seconds, 2000). Thus, if Pascalis and colleagues are correct that
and 60 seconds, respectively. These findings are con- their context-dependent task is dependent on the hip-
sistent with those from the human adult, where, for ex- pocampus, then in theory the delayed maturation ob-
ample, individuals with known damage to the served on this task reflects the delayed maturation of
hippocampus also show deficits in novelty preferences
under short delay conditions (McKee & Squire, 1993). 5
In theory, an alternative interpretation of these data is that
Importantly, in the human adult work, the same indi- the backgrounds were more salient to older versus younger in-
viduals who hours later fail to show novelty prefer- fants, and thus, the novelty preferences have more to do with
ences do show intact recognition memory (Manns stimulus salience generally than with relational memory qua
et al., 2000). Moreover, patient Y.R. with selective hip- memory.
The Neural Bases of Cognitive Development 23

some specific region of the hippocampus, such as the the infancy literature in which recognition memory is
dentate gyrus. typically inferred from novelty preferences?
This last observation underscores an important point, First, it may be unwise to assume that all tasks that
which is that although explicit memory emerges some- tap novelty preferences by default place the same de-
time between 6 and 12 months of life, it is far from fully mands on memory. For example, our view is that in tasks
developed by this age. The fact that even very young in- that require the subject to generalize discrimination
fants (i.e., a few months of age) do quite well on stan- across multiple exemplars of the same category of stim-
dard preferential looking paradigms suggests that there uli (e.g., to distinguish male faces from female faces),
is either enough hippocampal function to subserve task novelty preferences may depend on the ability to exam-
performance (as would be inferred from the monkey ine the relations among stimuli, and thus, depend dis-
work performed by Bachevalier and colleagues) or that proportionately on the hippocampus. In contrast, if the
perhaps the parahippocampal region (about which virtu- task simply requires one to discriminate two individual
ally nothing is known developmentally) is mediating exemplars (e.g., one female face from another), then
task performance. Thus, consistent with the neuro- perhaps the parahippocampal region is involved. Sec-
science literature, the full, adultlike expression of ex- ond, novelty preferences and recognition memory may
plicit memory awaits the subsequent development of not represent the same or different processes as much as
subregions of the hippocampus along with the connec- related ones. Thus, in the VPC procedure, perceptual
tions to and from associated neocortical areas. support is provided by presenting the familiar and novel
stimuli simultaneously. In so doing, recognition may be
Novelty Preferences
facilitated at very short delays, perhaps due to some
Because of the prominent role novelty preferences have iconic store rather than the need to compare the novel
played in evaluating memory in infants and young chil- stimulus to one stored in memory. This may occur in
dren, it is worth discussing at some length the putative short-term memory and be supported by the parahip-
neural bases of such preferences. As stated, we have pocampal region.
concluded from monkey data that the hippocampus plays Whereas it is easy to dissociate novelty preferences
a prominent role in novelty preferences, primarily based from recognition memory in children or adults (in whom
on the perturbations observed in such preferences when instructions can be given), the same is not true for in-
the hippocampus is lesioned. However, we need not re- fants, particularly when behavioral measures are em-
strict ourselves to data from monkeys. Neuroimaging ployed. However, using ERPs, Nelson and Collins (1991,
studies with human adults have also reported that the 1992) appeared to dissociate these processes to some de-
hippocampus is involved in novelty preferences (see gree. Four- through 8-month-old infants were initially
Dolan & Fletcher, 1997; Strange, Fletcher, Hennson, familiarized to two stimuli; during the test trials, one of
Friston, & Dolan, 1999; Tulving, Markowitsch, Craik, these stimuli was presented on a random 60% of the tri-
Habib, & Houle, 1996). In contrast, Zola et al. (2000) als (frequent-familiar), the other on a random 20% of the
reported that hippocampal lesions in mature monkeys trials (infrequent-familiar), and on each of the remain-
did not affect novelty preferences at 1-second delays, ing 20% of the trials, a different novel stimulus (infre-
and thus, that impairment (i.e., decline in performance) quent-novel) was presented. In theory, if recognition
at subsequent delays was due to problems in memory, not memory is independent of how often a stimulus is seen
novelty detection. Similarly, Manns et al. (2000) have (i.e., its probability of occurrence), then all infants
shown that among intact human adults, novelty prefer- should treat the two familiar stimuli as equivalent, re-
ences and recognition memory are both intact shortly gardless of how often they are presented during the test
after familiarization, although with increasing delay trials. The authors reported that it was not until 8 months
novelty preferences disappear, and recognition memory that infants responded equivalently to the two classes of
remains intact. This work, coupled with that from Zola familiar events and differently to the novel events. In
et al. (2000), argues for a dissociation between novelty contrast, at 6 months infants responded differently to the
preferences and recognition memory and raises two two classes of familiar events and differently yet again to
questions. First, are novelty preferences truly mediated the novel events. These findings were interpreted to re-
by the hippocampus or perhaps by extra-hippocampal flect an improvement in memory from 6 to 8 months,
tissue, and second, what does this dissociation say about specifically, the ability to ignore how often a stimulus is
24 Neural Bases of Cognitive Development

seen (inherent in novelty responses) from whether the visual recognition memory as inferred from the DNMS;
stimulus is familiar (inherent in recognition memory). importantly, data from the Nemanic et al. (2004) study
At this point in time, it is difficult to say with cer- suggests that lesions of the hippocampus and parahip-
tainty whether, in human children, it is the hippocampus pocampal regions have little effect on DNMS perfor-
or parahippocampal region that subserves novelty pref- mance. Of note is the observation that these data
erences and whether novelty preferences reflect a sub- contradict those reported by other groups (e.g., Zola
routine involved in recognition memory or reflect a et al., 2000), where hippocampal lesions in adult mon-
proxy for recognition memory per se. These are ques- keys do lead to impairments on DNMS performance.
tions that await further study. Our group has reported hippocampal activation in
human adults tested with the DNMS task (Monk et al.,
Delayed Non-Match-to-Sample 2002). What is to account for these differences? First, it
What about other tasks that have been thought to reflect could be that there is a fundamental difference in struc-
explicit memory, such as the DNMS task? Here a subject ture function relations in young versus mature subjects
is presented with a sample stimulus, and following some (see footnote 4); thus, the same function could be sub-
delay (during which the stimulus cannot be observed), served by a different structure in the developing versus
the sample and a novel stimulus are presented side by the developed individual. Second, it could mean that the
side. In the case of monkeys, the animal is rewarded for demands of the DNMS task differ (or are interpreted
retrieving the novel stimulus; in the case of humans, differently) depending on developmental age (e.g., the
some investigators implement a similar reward system reward value of reinforcing stimuli could differ; how the
and essentially adopt the animal-testing model (e.g., see child/juvenile monkey interprets the task demands
Overman, Bachevalier, Sewell, & Drew, 1993), whereas could be different from how the adult /mature monkey
others have modified the task such that no reward is ad- interprets the task demands). Third, earlier studies of
ministered and looking at the novel stimulus rather than hippocampal lesions in adult monkeys could have in-
reaching for it serves as the dependent measure (see A. cluded lesions of the surrounding cortex, thus making it
Diamond, 1995).6 In the classic DNMS task, it is gener- difficult to distinguish between impairments due to the
ally reported that monkeys do not perform at adult levels hippocampus or perirhinal cortex.7 Finally, in the Monk
until they approach 1 year of age, and performance et al. (2002) neuroimaging work, it was difficult to dis-
among children does not begin to resemble adults until tinguish the hippocampus from adjacent cortex, and
they are approximately 4 years of age (assuming the therefore, it is possible that the surrounding cortex was
standard 14 ratio of monkey to human years, these data most involved in performance on the DNMS task.
are remarkably consistent). Interestingly, Pascalis and Elicited Imitation
Bachevalier (1999) have reported that neonatal lesions
of the hippocampus do not impair performance on the Deferred or elicited imitation is a task that has now re-
DNMS task, suggesting that the DNMS likely depends ceived considerable attention by memory researchers. In
on extra-hippocampal structures and does not depend on the case of humans, a sequence of events is modeled,
just novelty preferences (i.e., unlike the VPC task, the and following some delay, and without benefit of prac-
DNMS task requires the subject to coordinate action tice, the subject is then presented with the props used in
schemes with what is represented in memory, and as 7
Until approximately the late 1990s, most surgical approaches
well, inhibit a response to the familiar stimulus). Sup-
to lesioning the hippocampus involved going through the sur-
port for this can be found in studies reported by
rounding cortex, thereby damaging this region (e.g., entorhi-
Mlkov, Bachevalier, Mishkin, and Saunders (2001) nal cortex) and fibers of passage. As a result, it was often
and Nemanic et al. (2004), in which lesions of the difficult to ascertain whether it was lesions of the hippocam-
perirhinal cortex in adult monkeys led to impairments in pus proper that led to perturbations in performance, lesions of
the surrounding cortex, or the severing of connections be-
6
It must be acknowledged that by modifying the task in this tween the hippocampus and surrounding cortex. Recent
way, the task requirements change dramatically, and thus, it methodological advances now employ the administration of
is difficult to argue that the two tasks are the same. This neurotoxins directly to the targeted structure, destroying cell
makes it difficult to compare across species. bodies but leaving fibers of passage intact.
The Neural Bases of Cognitive Development 25

the original sequence. Correct scores are given for the Historically, it has long been known that lesions of the
number of steps in the sequence that are correctly re- frontal cortex perturb the ability to reproduce a se-
called. Used extensively by Bauer and colleagues (see quence of events in the correct order (for review, see
Bauer, Chapter 9, this Handbook, this volume), Hayne Lepage & Richer, 1996). In addition to recalling the se-
and colleagues (e.g., Collie & Hayne, 1999; Hayne, quence in which objects are presented, the participants
Boniface, & Barr, 2000), and Meltzoff and colleagues must also encode the physical actions performed on the
(e.g., Meltzoff, 1985, 1988, 1995), among others, the objects. Here again, the frontal cortex may be involved;
prevailing evidence is that there is dramatic improve- Nishitani and Hari (2000) found that observing and imi-
ment in this ability starting at 6 months and continuing tating manual actions activated the left inferior pre-
through at least 24 months. In this case, improvement is frontal cortex as well as the premotor cortex and the
defined as the ability to reproduce longer sequences of occipital cortex. The supplementary motor area has also
objects over increasing delays. Indeed, as infants tra- been implicated in motor sequencing and may encode
verse their first birthday, under certain circumstances the numerical order of components (Clower & Alexan-
memory for temporal order is retained over months. der, 1998). Collectively, although the hippocampus may
Of particular relevance to this chapter is a trio of underlie elements of the elicited imitation task (perhaps
related observations. First, McDonough, Mandler, the recognition or recall memory of specific compo-
McKee, and Squire (1995) have reported that human nents, such as recognizing a familiar object or recalling
adults with bilateral damage to the medial temporal lobe a familiar sequence; see Carver et al., 2000), other com-
(including the hippocampus) perform poorly on this ponents of the task may be under control of the pre-
task. Similarly, individuals who sustained bilateral hip- frontal cortex (and presumably the basal ganglia for the
pocampal damage during childhood show impairments coordination of the reaching and memory components).
on the same task, although interestingly, the impair- As with much of the developmental literature on mem-
ments are not so severe as those observed in adults by ory, further insight into the neural correlates of elicited
McDonough et al. (1995; Adlam, Vargha-Khadem, imitation will have to await our ability to image the
Mishkin, & de Haan, 2004). Second, Eichenbaum and brain in action.
colleagues (e.g., Agster, Fortin, & Eichenbaum, 2002;
What of the Role of Frontal Cortex in
Fortin, Agster, & Eichenbaum, 2002) have reported that
Explicit Memory?
rodents trained to remember a sequence of odors show
impairments in distinguishing one sequence from an- In the context of development, virtually nothing is
other when the hippocampus proper is lesioned. Finally, known about the role of the prefrontal cortex in explicit
our group (e.g., DeBoer, Georgieff, & Nelson, in press; memory. It is well known in the adult, however, that
de Haan, Bauer, Georgieff, & Nelson, 2000) has demon- damage to frontal cortex impairs performance on tasks
strated impairments in elicited imitation in populations that require strategies or organizational manipulation
of infants presumed to have experienced pre- or perina- (for discussion, see Yancey & Phelps, 2001). The ability
tal damage to the hippocampus (e.g., those born very to transform, manipulate, or evaluate memories appears
prematurely or experiencing hypoxia-ischemia prena- to be a distinctly frontal lobe function (see Milner,
tally). Consistent with conclusions drawn by Bauer and 1995). In this context, it should not be surprising to
colleagues (e.g., Bauer, Wenner, Dropik, & Wewerka, learn that frontal lobe damage has little if any effect on
2000; Bauer, Wiebe, Carver, Waters, & Nelson, 2003; recognition memory but can have a dramatic impact on
Carver, Bauer, & Nelson, 2000), Nelson (1995) and recall memory, presumably because recall involves ma-
Hayne (2004), these data would appear to support the nipulating and evaluating information. Memory tasks
view that performance on elicited or deferred imitation particularly affected by damage to the frontal lobes in-
tasks depend on the integrity of the hippocampus. How- clude source memory (e.g., knowing who presented the
ever, as recently discussed by Nelson and Webb (2003), information or where the information was presented;
like the DNMS task, the elicited imitation task is com- see Janowsky, Shimamura, & Squire, 1989), frequency
plex and comprises a collection of subtasks. Infants memory (e.g., which item is presented most or least
must encode not just the properties of the objects but as often; see M. L. Smith & Milner, 1988), and consistent
well, the order in which these objects are presented. with our discussion of elicited imitation, memory for
26 Neural Bases of Cognitive Development

temporal order (e.g., which item was presented most re- neocortical areas that are thought to store such memo-
cently; see Butters, Kaszniak, Glisky, Eslinger, & ries, and the improved communication between the neo-
Schacter, 1994). What are the implications of these ob- cortex and the medial temporal lobe (MTL). It is most
servations for memory development? In brief, the abil- likely these changes that usher in the end of infantile
ity to use strategies to encode and retrieve information amnesia (for elaboration, see Nelson, 1998; Nelson &
and the ability to perform mental operations on the con- Carver, 1998).
tents of ones memory are skills that develop beginning
in the preschool period and continue through the imme-
Disorders of Memory
diate preadolescent period (for review of the develop-
ment of prefrontal cortical functions, see Luciana, Recent evidence suggests that bilateral hippocampal
2003). Not surprisingly, this is a period of rapid devel- damage sustained during childhood can lead to specific
opment of the prefrontal cortex, including synapse elim- impairments in memory, a condition labeled developmen-
ination and myelination. Thus, the changes that have tal amnesia (Gadian et al., 2000; Temple & Richardson,
been observed in memory across the preschool and ele- 2004; Vargha-Khadem et al., 1997).8 Interestingly, the
mentary school years (for review, see Flavell, Miller, & neuropsychological profile of these cases appears to dif-
Miller, 1993; Siegler, 1991) are likely not due to further fer somewhat from the amnesia observed in adults follow-
maturation of medial temporal lobe structures, but ing medial temporal lobe damage. One difference is that,
rather to maturation of frontal lobe structures and im- unlike adult amnesiacs who show impairments in both
portantly, to increased connectivity between the medial recognition and recall, in at least one intensively studied
temporal lobe and the prefrontal cortex. With the ability case of developmental amnesia, delayed recognition
to employ fMRI across this age range, it would be our memory was relatively preserved, whereas delayed recall
hope that this tool will one day be applied to test this was severely impaired (Baddeley, Vargha-Khadem,
connectivity hypothesis. Mishkin, 2001). A second difference is that, unlike adult
amnesiacs who tend to show impairments in both episodic
Summary
and semantic memory, patients with developmental am-
Piecing together heterogeneous sources of information, nesia tend to show severe impairments in episodic mem-
it appears that an early form of explicit memory emerges ory with semantic memory remaining fairly intact
shortly after birth (assuming a full-term delivery). This (Vargha-Khadem et al., 1997).
preexplicit memory is dependent predominantly on the The reason for this difference in profile between de-
hippocampus. As infants enter the second half of their velopmental and adult-onset cases remains a matter of
1st year of life, hippocampal maturation, coupled with debate. One possibility is that the patients with develop-
development of the surrounding cortex, makes possible mental amnesia generally show a milder memory impair-
the emergence of explicit memory. A variety of tasks ment than adult amnesiacs because their damage is more
have been used to evaluate explicit memory, some selective. Many of the adult cases reported in the litera-
unique to the human infant, others adopted from the ture have widespread damage including additional areas
monkey. Based on such tasks, one observes a gradual within the temporal lobe and also extra-temporal brain
improvement in memory across the first few years of regions. By contrast, at least within the medial temporal
life, most likely due to changes in the hippocampus lobe, the perirhinal, entorhinal, and parahippocampal
proper (e.g., dentate gyrus), to the surrounding cortex cortices of patients with developmental amnesia appear
(e.g., parahippocampal cortex), and to increased con- intact as measured by volumetric analysis (Schoppik
nectivity between these areas. The changes one observes et al., 2001). It is possible that these regions normally
in memory from the preschool through elementary subserve recognition memory and semantic recall, and
school years are likely due to changes in prefrontal cor- their preservation in developmental amnesia allows for a
tex, and connections between the prefrontal cortex and relative preservation of these skills (Vargha-Khadem,
the medial temporal lobe. Such changes make possible
the ability to perform mental operations on the contents 8
Here it must be stressed that in such cases the brain damage
of memory, such as the ability to use strategies to en- is not generally confined to the hippocampus, but rather, is
code and retrieve information. Finally, changes in long- somewhat more diffuse, involving a number of medial tempo-
term memory are likely due to the development of the ral lobe structures (such as the rhinal cortex).
The Neural Bases of Cognitive Development 27

Gadian, & Mishkin, 2001). Consistent with this view are Mishkin, 2001). However, while intact individuals
reports that adult patients with more selective hip- showed primarily left hippocampal activation, the pa-
pocampal lesions also show some preserved semantic tient showed bilateral activation, as well as a different
learning (e.g., Verfaellie, Koseff, & Alexander, 2000). pattern of connectivity of the hippocampus with other
However, others have reported that adult patients with brain areas compared to controls (Maguire et al., 2001).
lesions reportedly restricted to the hippocampus show a It is thus possible that the remaining areas of function-
more general memory deficit (e.g., Manns & Squire, ing hippocampus organize with other brain regions to
1999) and that a patient with more widespread damage mediate the preserved memory abilities. If this is true,
sustained at 6 years of age also showed relatively pre- then one might expect that age of injury would affect the
served semantic learning (Brizzolara, Casalini, Monta- ultimate outcome, with early injury leading to better
naro, & Posteraro, 2003). These findings would suggest outcome than later injury due to greater plasticity. A re-
that age of injury, rather than extent, may be critical. cent report does not support this prediction, showing
Understanding the separate and combined contribution that delayed memory does not differ for children who
of the selectivity of damage within the MTL, the degree sustained their injury before 1 year of age and those
of extra-hippocampal damage (e.g., abnormalities in the who sustained injury after 6 years of age (Vargha-Kha-
putamen, posterior thalamus, and right retrosplenial cor- dem et al., 2003). However, those with earlier lesions
tex are reported in developmental amnesia; Vargha- performed better than those with later lesions on some
Khadem et al., 2003) and age of injury will be important tests of immediate memory, suggesting that some plas-
questions for future studies. ticity may be operating.
Interestingly, for early onset injuries the degree of
hippocampal damage appears to affect the pattern of
Nondeclarative Memory
memory impairment (Isaacs et al., 2003). Compared with
controls, patients with developmental amnesia (average Nondeclarative or implicit forms of learning and mem-
bilateral hippocampal volume reduction of 40%, ranging ory functions represent an essential aspect of human
from 27% to 56%) were impaired on tests of delayed re- cognition by which information and skills can be
call, whereas preterm children (average reduction of 8% learned through mere exposure or practice, without re-
to 9%, ranging to 23%) showed no deficit in delayed re- quiring conscious intention or attention and eventually
call ( but did show a deficit in prospective memory and becoming automatic. Although controversy exists
route following). These differences cannot be accounted regarding the definition of nondeclarative memory
for by general differences in ability level as the groups or learning, performance on most nondeclarative
were matched on IQ. However, the possibility of differ- tasks does not appear to depend on medial temporal
ences in extra-hippocampal damage, which may vary as a lobe structures. Patients like H. M., mentioned earlier,
function of degree of hippocampal damage, is not known. demonstrate severe deficits in explicit memory consis-
Another possible explanation for the differing tent with known insults to or disruption of medial tem-
profiles between child- and adult-onset cases is that poral lobe memory systems (see Figure 1.3). However,
early-onset cases show milder impairments because the these patients are not impaired on classic tests of
plasticity of the developing brain allows for compensa- implicit memory and learning, such as perceptual
tion of function. In children with neonatal damage, priming or serial reaction time (SRT) learning (Milner,
memory impairments are not typically noted until chil- Corkin, & Teuber, 1968; Shimamura, 1986; Squire,
dren are of school age (Gadian et al., 2000). This would 1986; Squire & Frambach, 1990; Squire, Knowlton, &
appear surprising given that the hippocampus is be- Musen, 1993; Squire & McKee, 1993).
lieved to play a critical role in memory from birth, as A multitude of tasks have emerged in the cognitive
discussed earlier. A possible explanation is that there is literature to assess nondeclarative cognitive functions
a degree of compensation that allows relatively good (see Seger, 1994; Reber, 1993). Reber (1993) proposes
initial memory skills but that ultimately leads to a mild making a distinction between two broad categories of
form of memory impairment. A neuroimaging case nondeclarative function: implicit memory (the end-state
study has shown that the remaining hippocampal tissue representation of knowledge available to an individual,
in developmental amnesia appears to be functional of which he or she is unaware) and implicit learning
during memory tasks (Maguire, Vargha-Khadem, & (the unintentional and unconscious acquisition of
28 Neural Bases of Cognitive Development

Similarly, priming in the context of an object classifica-


tion task is associated with decreased activity in infe-
rior temporal cortex and ventral prefrontal cortex
(Buckner et al., 1998), regions previously linked to ex-
plicit object identification.
Although a number of studies have examined visual
priming performance in infancy and early childhood
(Drummey & Newcombe, 1995; Hayes & Hennessy,
1996; Parkin & Streete, 1988; Russo, Nichelli, Giber-
toni, & Cornia, 1995; Webb & Nelson, 2001), few have
applied brain-imaging methods to examine the brain
bases of this function. In general, behavioral evidence
overwhelmingly suggests that young children and even
infants show priming effects that are very similar to
those of adults, supporting the idea that implicit learn-
Figure 1.3 Diagram of the medial temporal lobe memory ing and memory functions may show very little develop-
system. mental change while explicit learning and memory
demonstrate protracted development. Webb and Nelson
knowledge). Implicit learning involves learning of un- (2001) used ERP measures to assess visual perceptual
derlying rules and structure in the absence of any con- priming in 6-month-old infants compared with adults.
scious awareness of those regularities. Such learning is Although developmental differences were observed in
slow and requires repeated exposure to the information ERP components between adults and infants, these dif-
to be retained. In contrast, implicit memory may occur ferences were similar to differences observed during
with a single exposure to a stimulus and results in an in- explicit memory function. However, ERP evidence of
creased processing efficiency for subsequent presenta- memory (an activity difference between new and re-
tions of that stimulus or closely related stimuli. Not peated stimuli) was observed only at an early ERP com-
only do these categories differ in their basic cognitive ponent typically associated with attention, and not at
nature (representation versus acquisition of knowledge), the late component typically associated with explicit
but they seem to differ in their underlying neural sub- memory (Nelson, 1994; Nelson & Monk, 2001; see De-
strates as well. In a now well-known classification of Boer, Scott, & Nelson, 2004 for review of infant ERP
memory systems, Squire (1994) identified three pri- components).
mary forms of implicit learning and memory: priming,
procedural learning (skills and habits), and classical Implicit Sequence Learning
conditioning (associative learning), each relying on sep- In contrast to priming, implicit learningalso termed
arable neural systems. habit learning, skill learning, or procedural learning
involves the slow acquisition of a knowledge base or be-
Visual Priming havioral skill set over time. In everyday life, learning to
Evidence for this taxonomy comes from animal and ride a bicycle involves the gradual acquisition of a skill
human lesion studies as well as neuroimaging methods. that is very difficult or impossible to describe verbally.
For example, priming, or improvements in detecting or Although intentionally trying to learn the skill, the
processing a stimulus based on a recent exposure to that learner is typically unaware of what exactly is being
stimulus, appears to rely on neocortical brain regions learned. Implicit learning has most frequently been
relevant to the stimulus of interest. That is, when tested using sequence learning (e.g., Nissen & Bullemer,
healthy adults are presented with a visual stimulus re- 1987) or artificial grammar learning paradigms (Reber,
peatedly, areas of extrastriate visual cortex show de- 1993). In the SRT task, individuals are asked to map a
creased activity, presumably reflecting diminished set of spatial or object stimuli onto an equal number of
processing requirements when reactivating the sensory response buttons. Reaction times to match the stimulus
circuit (Schacter & Buckner, 1998; Squire et al., 1992). and the button are recorded. Unknown to the partici-
The Neural Bases of Cognitive Development 29

pant, whereas stimuli often appear in random order, at Malone (1995) report age-related improvements in co-
other times, the order of stimulus presentation follows variation learning, with older children showing larger
a predictable and repeating sequence. Implicit learning learning effects than younger children. However,
is assumed when participants show reaction time im- Lewicki (1986) found no evidence of age-related learn-
provements during the sequential trials compared with ing effects on the original version of the same task. Sim-
random trials, despite no conscious awareness of the un- ilarly, Thomas and Nelson (2001) found that, although
derlying regularity. the size of the implicit learning effect was similar be-
Patients with temporal lobe amnesia perform nor- tween 4-, 7-, and 10-year-olds who showed evidence of
mally on sequence learning tasks. However, patients sequence learning on an SRT task, older children were
with basal ganglia damage, such as those with Parkin- more likely to show learning. In fact, the probability of
sons or Huntingtons diseases, have been shown to be significant learning was inversely related to age, with
impaired on the serial reaction time task (Ferraro, fully one-third of the youngest age group showing no evi-
Balota, & Connor, 1993; Heindel, Salmon, Shults, dence of learning (whereas all 10-year-olds showed a
Walicke, & Butters, 1989; Knopman & Nissen, 1991; learning effect). Evidence for age-related improvements
Pascual-Leone et al., 1993). Importantly, these patients in implicit sequence learning come from an infant ana-
perform normally on measures of explicit memory as logue of the SRT task: visual expectancy formation. In
well as on measures of perceptual and conceptual prim- this task, infants are shown a repeating pattern of visual
ing (Schwartz & Hastroudi, 1991), providing support for stimuli, and eye movements are recorded to determine
the separability of implicit learning and implicit mem- whether the infant learns to anticipate the location of the
ory at the neural systems level. Neuroimaging data upcoming stimuli over time. Although we cannot rule out
provide supporting evidence for the role of subcortical the possibility that this behavior is explicitly learned, the
structures in serial reaction time learning. Common task has many similarities to the adult SRT paradigm. In-
findings across a number of laboratories (Bischoff- terestingly, although infants as young as 2 and 3 months
Grethe, Martin, Mao, & Berns, 2001; Grafton, Hazel- of age can show reliable visual expectancy formation
tine, & Ivry, 1995; Hazeltine, Grafton, & Ivry, 1997; (Canfield, Smith, Brezsnyak, & Snow, 1997; Haith,
Schendan, Searl, Melrose, & Ce, 2003) demonstrate dif- Hazan, & Goodman, 1988), older infants are able to
ferential activity in frontalbasal gangliathalamic learn more complicated sequential relationships than
circuits for sequential trials compared with random younger children (Clohessy, Posner, & Rothbart, 2002; P.
trials. Further evidence suggests that connections Smith, Loboschefski, Davidson, & Dixon, 1997).
among these fronto-striatal loops and fronto-cerebellar In a recent pediatric imaging study of the SRT task,
loops may be an important aspect of implicit learning. Thomas et al. (in press) compared the neural systems
Pascual-Leone et al. (1993) observed that, although subserving implicit sequence learning in 7- to 11-year-
adults with basal ganglia insults demonstrated signifi- old children and adults. Overall, results from adults
cant reductions in implicit sequence learning, adults were consistent with previous neuroimaging studies im-
with cerebellar degeneration showed no evidence of plicating fronto-striatal circuitry in visuomotor se-
learning on the SRT task. quence learning. In particular, activity in the basal
Significant controversy exists regarding the develop- ganglia was positively correlated with the size of the im-
mental trajectory of implicit learning. In a study of SRT plicit learning effect (greater learning was associated
learning in 6- to 10-year-old children and adults, Meule- with increased activity in the caudate nucleus). Al-
mans, van der Linden, and Perruchet (1998) observed though children and adults showed many of the same re-
equivalent learning of a ten-step spatial sequence across gions of activity, relative group differences were
age groups, despite overall reaction time differences observed overall, with children showing greater subcor-
with age. These data support the notion that implicit cog- tical activity and adults showing greater cortical activ-
nition may mature very early in infancy and show little ity. Consistent with recent findings from adult SRT
variation or improvement with age (Reber, 1992). How- studies (Schendan et al., 2003), both adults and children
ever, other measures of implicit pattern learning or con- showed activity in the hippocampus despite no explicit
tingency learning, as well as SRT data from an alternate awareness of the sequence. This activity is unlikely to
group, are less clear. Maybery, Taylor, and OBrien- be either necessary or sufficient for implicit sequence
30 Neural Bases of Cognitive Development

learning given the adult literature indicating spared per- implicit learning resulting from early insults to basal
formance following lesions to the hippocampus. Instead, ganglia circuitry.
this activity may reflect a sensitivity to stimulus nov-
Conditioning or Associative Learning
elty, a function of the hippocampus discussed earlier in
this chapter. Children showed an inverse pattern of hip- In contrast to priming and implicit sequence learning,
pocampal activity as compared to adults (random trials the existing knowledge regarding the brain bases of con-
elicited greater hippocampal activity than sequence tri- ditioning derives largely from animal studies. In condi-
als for child participants). Interestingly, the SRT task tioning paradigms, the subject learns the contingency
used here produced significant developmental differ- between two previously unrelated stimuli through either
ences in the magnitude of the learning effect, with chil- association with a cue or signal (classical conditioning)
dren demonstrating significantly less learning than or an outcome or reward (instrumental conditioning).
adults. Unlike prior behavioral studies, this effect was Studies of classical eyeblink conditioning in rabbits
not driven by a difference in the percentage of nonlearn- have implicated the cerebellum and its connections to
ers in the two age groups. Rather, despite significant in- brain stem nuclei (particularly the interpositus nucleus)
dividual learning in both groups, adults learned to a as key regions necessary for normal conditioning to
greater extent with the same degree of exposure. occur (Woodruff-Pak, Logan, & Thompson, 1990).
Finally, some evidence exists to address the effects of Lesions of either of these structures can completely pre-
basal ganglia insults early in development. Although early vent acquisition of the conditioned response, while
insults may lead to lasting impairments in functions sub- lesions of the cerebrum have no effect on either acquisi-
served by the affected systems, the plastic nature of the tion of the response or its maintenance (Mauk &
developing brain may also allow for redistribution of Thompson, 1987). The hippocampus also is not required
function to other regions which are unaffected. Structural for classical delay eyeblink conditioning. Patients with
neuroimaging studies have identified childhood attention- temporal lobe amnesia show normal performance on
deficit/hyperactivity disorder (ADHD), as well as perina- delay conditioning paradigms (Woodruff-Pak, 1993)
tal complications such as intraventricular hemorrhage as whereas those with lesions of the lateral cerebellum are
risk factors for disrupted basal ganglia circuitry. Castel- severely impaired in acquiring the conditioned eye-blink
lanos et al. (2001, 2002) have reported decreases in cau- response (Woodruff-Pak, Papka, & Ivry, 1996). In con-
date volume in children with ADHD compared with trast, emotional fear conditioning (pairing a neutral
nonaffected controls. Similarly, functional imaging stud- stimulus with a shock) appears to rely heavily on the
ies of attentional control (see Executive Functions, later amygdala (LaBar, Gatenby, Gore, LeDoux, & Phelps,
in this chapter) suggest a lack of typical basal ganglia ac- 1998; Pine et al., 2001).
tivity in children with ADHD (Vaidya et al., 1999). A re- Developmentally, newborn infants (10 days old) are
cent paper addressing reading disabilities suggests a able to acquire a conditioned eyeblink response. There-
possible link between reduced motor sequence learning fore, like implicit sequence learning, the basic mecha-
and symptoms of ADHD (Waber et al., 2003). Thomas nism supporting conditioned responding must be present
and colleagues reported evidence of significant decre- and functional at birth, although this in no way implies
ments in sequence learning for 6- to 9-year-old children that its function is fully mature. Developmental im-
diagnosed with ADHD (Thomas, Vizueta, Teylan, Ec- provements are observed in conditioning paradigms. For
card, & Casey, 2003). These authors also examined im- example, with age, children can handle longer and longer
plicit sequence learning in children with perinatal delays between the neutral signal and the unconditioned
histories of intraventricular hemorrhage (IVH), or bleed- stimulus (Orlich & Ross, 1968). Such developmental ef-
ing into the lateral ventricles at birth. Children whose fects might be expected given the very protracted matu-
IVH was moderate ( bilateral grade II or more severe) ev- rational trajectory of the cerebellum (Keller et al.,
idenced significant decrements in the magnitude of the 2003). Much less is known regarding the development of
implicit learning effect. In contrast, children whose peri- the amygdala response in fear conditioning. Similarly,
natal IVH had been relatively mild (unilateral grade II or although significant behavioral research has been con-
less severe) showed no difference in learning from a full- ducted addressing instrumental conditioning in infancy
term, age- and gender-matched control group. Together, (Rovee-Collier, 1997a), relatively little is known regard-
these studies suggest a potential long-term deficit in ing the brain systems underlying this form of learning.
Spatial Cognition 31

Rovee-Collier (1997b) and colleagues have observed in- Carlson, & Mann, 1982). This is consistent with evi-
creases in the retention duration for conditioning with dence that the parietal lobe begins to obtain adultlike
age. That is, older children remember the contingencies status at about 10 to 12 years of age (Giedd et al., 1999).
for longer periods than younger children do (Hartshorn Neuroimaging studies comparing activation in children
et al., 1998). Few, if any, studies have examined the and adults confirm that parietal areas are activated in 8-
brain bases of this form of learning in childhood. to 12-year-old children as in adults (Booth et al., 1999,
2000; Chang, Adleman, Dienes, Menon, & Reiss, 2002;
Shelton, Christoff, Burrows, Pelisari, & Gabrieli, 2001).
SPATIAL COGNITION
Parietal areas are activated during mental rotation tasks
regardless of individual differences in ability level
Spatial cognition refers to a range of abilities involved
(Shelton et al., 2001). However, differences between
in perceiving, remembering, and mentally manipulating
children and adults in the overall pattern of activation
spatial relations and orientation at several scales includ-
have been noted. For example, in one study, adults and
ing features of a single object, objects and their context,
9- to 12-year-olds were asked to mentally rotate a letter
and oneself in space.
or number from one of four different rotations to its
upright position and then to determine whether it was
Mental Rotation
forward or backward (Booth et al., 2000). Both age
Mental rotation is a form of visual imagery involving groups showed similar overall levels of performance and
. . . the imagined circular movement of a given object showed the typical pattern of increasing reaction
about an imagined pole in either 2 or 3 dimensional time and errors with increasing degree of rotation. De-
space (Ark, 2002, p. 1). In adults, the superior parietal spite the similarities in behavioral performance, adults
lobe appears to be involved in the actual act of imaginary showed more activation in the superior parietal and
rotation, as it is not only active during mental rotation middle frontal areas, whereas children showed more ac-
tasks but its extent of activation is related to reaction tivation in the supramarginal gyrus. The authors suggest
time (Richter et al., 2000; Richter, Ugurbil, Georgopou- that this may reflect strategy differences, with adults
los, & Kim, 1997). Activations in visual area MT (M. S. engaging in more mental rotation and children engaging
Cohen et al., 1996) and motor and premotor areas in the in more processing of the noncanonical orientations of
frontal cortex (Johnston, Leek, Atherton, Thacker, & the letters and numbers. The idea that the neural
Jackson, 2004; Richter et al., 2000; Windischberger, processes underlying mental rotation may differ even
Lamm, Bauer, & Moser, 2003) have also been observed when performance is equated is echoed in results of
in some, but not all, studies. These activations suggest electroencephalogram (EEG) studies, in which greater
that some of the processes involved in mental rotation parietal activation was reported in 8-year-old boys than
overlap with processes involved in actual physical rota- girls during 2-D mental rotation despite similar levels of
tion of a stimulus: MT activation occurs when subjects behavioral performance (Roberts & Bell, 2002).
observe objects moving, and premotor and motor areas It is possible that some of the differences observed
are activated when subjects physically move objects. In between children and adults are related not to develop-
support of this idea, one study demonstrated bilateral mental change in mental rotation abilities per se, but to
premotor activation during mental rotation of pairs of failure to match on gender. In adults, differences in
hands but only left premotor activation during mental ro- brain areas activated during mental rotation have been
tation of tools (Vingerhoets, de Lange, Vandemaele, Del- noted between men and women (Roberts & Bell, 2002).
baere, & Achten, 2002). This pattern suggests that Even when performance is equated, women tend to show
participants were using motor imagery and imagined more bilateral activation and men more right-lateralized
moving both hands in the hand condition and imagined activation in parietal regions (Jordan, Wustenberg,
using and/or moving the tools with their preferred (right) Heinze, Peters, Jncke, 2003); women also tend to show
hand in the tool condition (Vingerhoets et al., 2002). greater activation in frontal regions (Thomsen et al.,
Between 8 and 12 years of age, childrens mental ro- 2000; Weiss et al., 2003).
tation abilities become adultlike in several respects Consistent with neuroimaging results in children,
(e.g., nature of errors, relation of reaction time to degree which have reported bilateral parietal activation, both
of rotation; see Dean, Duhe, & Green, 1983; Waber, right and left hemisphere lesions in children appear to
32 Neural Bases of Cognitive Development

impair performance on mental rotation tasks (Booth typically developing 12- to 14-year-olds performing spa-
et al., 2000). Adults tested in those neuroimaging stud- tial processing tasks (Moses et al., 2002). Specifically,
ies also showed bilateral activation, although other stud- children, like adults, showed faster reaction time to
ies have reported evidence for either left or right global than to local stimuli in the left visual field/right
lateralization (Harris & Miniussi, 2003; Roberts & Bell, hemisphere but the reverse pattern of reaction time to
2002, 2003; Zacks, Gilliam, & Ojemann, 2003). Chil- stimuli in the right visual field/ left hemisphere; they
dren with unilateral lesions showed activations in simi- also showed the adultlike pattern of lateralized occipito-
lar regions as healthy children and adults, but only in the temporal activation for the two tasks. However, not all
intact hemisphere (Booth et al., 2000). children displayed this mature pattern. A subgroup of
children failed to show the expected lateral differences
in the split-visual field task and also showed no task-
Spatial Pattern Processing
related lateral differences during the fMRI study, only a
Spatial analysis involves the ability both to segment a generally greater right-hemisphere activation during
pattern into a set of constituent parts and to integrate both types of task (Moses et al., 2002). The authors sug-
these parts into a coherent whole (Stiles, Moses, Pas- gest that this pattern indicates that a greater degree of
sarotti, Dick, & Buxton, 2003). In adults, the left infe- brain lateralization is linked to a greater degree of skill
rior temporal gyrus and parts of the left fusiform gyrus or more mature abilities.
are more active than the homologous right hemisphere With respect to brain lesions, several studies of pedi-
during pattern segmentation ( local processing), while atric patients with right or left hemisphere damage con-
the homologous areas in the right are more active than firm the importance of these areas for global and local
those in the left during pattern integration (global pro- processing respectively. When children are shown hier-
cessing; Martinez et al., 1997). In addition, occipito- archical forms and asked to reproduce them, children
parietal regions in the area of the intraparietal sulcus, with left hemisphere injury have difficulty reproducing
particularly on the right, are active during both types of local elements, whereas children with right hemisphere
task, with larger areas activated during pattern integra- lesions have difficulty reproducing the global structure
tion than segmentation (Martinez et al., 1997; Sasaki (adults: Delis et al., 1986; children: Stiles, Bates, Thal,
et al., 2001). This latter finding is consistent with the Trauner, & Reilly, 1998); these impairments persist to
zoom lens analogy of selective attention: Attention to school age. By contrast, children as young as 4 can re-
local features activates cortical foveal representations, produce both levels accurately (Dukette & Stiles, 1996).
whereas attention to global aspects activates peripheral On other tasks such as block construction and copying
areas (Sasaki et al., 2001). Studies of adult patients with and memory for complex figures, children with either
brain damage have revealed parallel results, with pat- type of lesion show delays. And, even when children
tern integration being linked to right hemisphere dam- eventually produce relatively accurate constructions,
age and pattern segmentation to left hemisphere damage the strategies they use to produce them are often simpli-
(Delis, Roberston, & Efron, 1986). fied and more representative of those seen in typically
Behavioral studies indicate that processing of global developing children at a younger age.
information develops more quickly than processing of While damage to the developing brain can produce a
local information, with both types reaching adult levels pattern of deficit similar to that seen in adults, these ef-
by 14 years of age. Roe, Moses, and Stiles (1999) tested fects may be less pronounced and show more improve-
children aged 7 to 14 years using hierarchical stimuli ment with time than in adults. While performance on
and found that, while the left visual field advantage for block construction and complex pattern memory are
global levels appeared sooner than the right visual field clearly delayed, children eventually can produce accu-
advantage for local levels, both are present in adultlike rate responses. To investigate the brain bases of this
form by age 14 ( but see Mondloch, Geldart, Maurer, & plasticity, Stiles, Moses, Roe, et al. (2003) investigated
de Schonen, 2003). There appears to be a gradual emer- two children with frank parietal lesions and white mat-
gence of the left hemisphere bias for local levels of ter loss more posteriorly, one with right hemisphere in-
forms between 7 and 14 years of age. volvement and one with left hemisphere involvement.
In children, neuroimaging studies have shown that During tasks requiring pattern segmentation and inte-
areas similar to those activated in adults are activated in gration, patients showed lateralization of activity to the
Spatial Cognition 33

contralesional hemisphere, suggesting that the intact & Logie, 1999, p. 28). It allows humans . . . to maintain
hemisphere can take on some abilities normally sub- a limited amount of information in an active state for a
served by the damaged one. brief period of time and to manipulate that information
Children initially seem to show bilateral activation (E. E. Smith & Jonides, 1998, p. 12061). For adults, the
during both global and local processing, and pediatric number of items is estimated at 1 to10 over a time span
patients with unilateral lesions appear to be able to use a of 0 to 60 seconds (E. E. Smith & Jonides, 1998).
single hemisphere to process both types of pattern; Working memory is often divided into three compo-
these facts seem at odds with other findings suggesting nents: the central executive that controls and regulates
that hemispheric biases in spatial processing are already the working memory system, and two domain-specific
apparent very early in life. For example, 8-month-olds slave systems responsible for processing information in
tend to process configural aspects more than featural the phonological (phonological loop) or visuospatial (vi-
aspects of faces (e.g., Schwarzer & Zauner, 2003), and suospatial sketchpad) forms (Baddeley, 1986; Baddeley
by a similar age they show a right-hemisphere bias in the & Hitch, 1974).
processing of faces (de Haan & Nelson, 1997, 1999; de Visuospatial working memory improves with age
Schonen & Mathivet, 1990) but not objects (de Haan & into adolescence (e.g., Luciana & Nelson, 1998). What
Nelson, 1999). More direct evidence comes from studies drives these changes? One factor that is known to be re-
in which the sensitivity of the right versus the left hemi- lated to improvements in visuospatial working memory
sphere to featural / local changes compared with config- is phonological encoding. From about 8 years of age,
ural changes has been assessed by split visual field children increasingly use verbal strategies during visual
presentation with infants. These studies also suggest memory tasks (e.g., object memory); improvements in
that the left visual field/right hemisphere is more sensi- phonological encoding facilitate use of this strategy
tive to global changes while the right visual field/ left (Pickering, 2001a). However, this cannot completely ac-
hemisphere is more sensitive to local changes (Deruelle count for developmental changes in visuospatial work-
& de Schonen, 1991, 1995). ing memory as there are also developmental changes on
The bilateral activation seen in some young children tasks that are more purely visuospatial (e.g., Isaacs &
during both local and global processing may reflect less Vargha-Khadem, 1989). These changes may be due to
a lack of hemisphere bias and more an immature pattern (a) changes in knowledge base, ( b) changes in strategy,
in which task difficulty demands recruiting all available (c) changes in processing speed, and (d) changes in at-
resources for the task at hand. Similarly, when patients tentional focus (reviewed in Pickering, 2001a).
with unilateral lesions are presented with these tasks, With respect to neural substrates, studies with adults
they devote all their limited resources to complete them. show that a frontoparietal network (especially on the
With development, in this view there is increasing spe- right) plays an important role in visuospatial working
cialization such that improved performance is related to memory. There may also be a dorsal-ventral distinction
recruitment of a smaller set of brain areas that are able in frontal areas for visual versus spatial processing
to perform a particular task most efficiently (Johnson, (Ruchkin, Johnson, Grafman, Canoune, & Ritter, 1997;
2001; Stiles, Moses, Roe, et al., 2003). In other words, Sala, Rama, & Courtney, 2003). This has been studied
there is a selective recruitment of resources rather than a mainly for visuospatial working memory, but there is
recruitment of all available resources. Experience with some evidence for activation of similar pathways during
processing the different types of stimuli is thought in audiospatial working memory tasks (Martinkauppi,
part to drive this process. In support of this view, neu- Rama, Aronen, Korvenoja, & Carlson, 2000; though sin-
roimaging studies with adults have shown that percep- gle-cell studies suggest different but parallel popula-
tual learning is associated with an increased efficiency tions of neurons are involved, Kikuchi-Yorioka &
in the brain areas activated as expertise is acquired Sawaguchi, 2000). The visual superior posterior parietal
(Gauthier, Tarr, Anderson, Skudlarski, & Gore, 1999). and premotor areas may be involved in rehearsal (that
can reactivate rapidly decaying contents of the storage
component), while inferior posterior parietal and ante-
Visuospatial Working Memory
rior occipital regions may mediate storage (whose con-
Working memory is the moment to moment monitoring, tents decay rapidly). A frontoparietal network also
processing and maintenance of information (Baddeley appears to be involved in children, and in this gross
34 Neural Bases of Cognitive Development

anatomical sense, there are similarities between chil- increase transmission locally within parietal cortex and
dren and adults (Casey et al., 1995; Kwon, Reiss, & also its communication with frontal cortex. This could
Menon, 2002; Nelson, Lin, et al., 2000; Thomas et al., result in more stable frontoparietal activity during
1999; Zago & Tzourio-Mazoyer, 2002). However, there working memory delays and cue periods, and thus less
are also changes that occur with development. Older resistance to interference (Klinberg et al., 2002). Synap-
children show higher activation in the superior frontal tic pruning and axonal pruning could also result in less
and intraparietal cortex than younger children. Working competition with input from other areas, leaving the
memory capacity is significantly correlated with brain frontoparietal network more stable.
activity in these same regions (Klingberg, Forssberg, & Several developmental disorders also implicate a
Westerberg, 2002; Kwon et al., 2002), suggesting that frontoparietal network in visuospatial working memory.
increases in working memory with age are related to in- With respect to lesions, right frontal cortex damage at 7
creases in activation in these regions. Regions thought years has been related to impairments in visuospatial
to underlie the phonological loop are activated during working memory (Eslinger & Biddle, 2000). Turners
spatial tasks (Kwon et al., 2002). syndrome (a sex chromosome disorder affecting girls, in
These studies collectively suggest that increases in which one of the two X chromosomes is missing) is asso-
working memory are associated with increases in acti- ciated with deficits in visuospatial working memory
vation in the frontoparietal network. One question that (Cornoldi, Marconi, & Vecchi, 2001; Haberecht et al.,
arises is whether this represents mainly the maturation 2001). In Turners syndrome, imaging studies show de-
or coming online of the frontoparietal network. In sup- creased activation in the supramarginal gyrus, dorsolat-
port of this view, diffusion tensor imaging studies in eral prefrontal cortex, and the caudate nucleus, with a
children suggest that developmental increases in gray suggestion that these abnormalities in frontoparietal and
matter activation in frontal and parietal cortices could frontostriatal circuits are related to poor working mem-
be due to maturation of white matter connections (Ole- ory performance (Haberecht et al., 2001). Individuals
sen, Nagy, Westerberg, & Klingberg, 2003). However, with fragile X syndrome are also known to have impair-
studies with adults also show increased activity in these ments in visuospatial working memory and appear to be
areas with improvements in working memory following unable to modulate activation in the prefrontal and pari-
training (Olesen, Westerberg, & Klingberg, 2004). etal cortex in response to an increasing working memory
These results show that increased activity in these areas load. These deficits may be related to a lower level of
can also occur as a result of experience and practice. FMRP expression in individuals with fragile X syndrome
The extent to which the increases in activation in the compared with typical subjects (Kwon et al., 2001).
working memory network seen in adults following train-
ing and during normal development reflect similar or
Visuospatial Recognition and Recall Memory
different mechanisms requires further investigation. In
children, diffusion tensor imaging studies suggest that In addition to the frontoparietal network implicated in
developmental increases in gray matter activation in spatial processing and working memory, medial temporal
frontal and parietal cortices could be due to maturation lobe structures, particularly in the right hemisphere, have
of white matter connections (Olesen et al., 2003). Thus, been implicated in visuospatial memory in adults.
there is evidence of increased activation of the working The question of whether a similar hemispheric bias exists
memory network with increases in age and working in children has been investigated in neuropsychological
memory abilities. studies of children with unilateral temporal lobe epilepsy.
The developmental increases in working memory per- In one study, children who had either early (0 to 5 years)
formance coincide with the timing of several neurode- or later (5 to 10 years) onset of seizures were tested
velopmental processes, including a decrease in synaptic with visual and verbal tasks; children with early-onset
density (Bourgeois & Rakic, 1993), axonal elimination epilepsy performed generally poorly, whereas children
(LaMantia & Rakic, 1990), changes in global cerebral with later onset showed a material-specific pattern
metabolism (Chugani & Phelps, 1986), myelination of deficit (Lespinet, Bresson, NKaoua, Rougier, &
(Paus et al., 1999), and changes in catecholamine recep- Claverie, 2002). Two other studies investigating children
tor structure and density. The inferior parietal cortex is following temporal lobectomy noted that visuospatial
one of the last areas to myelinate, a change that would deficits were both pronounced and persistent (Hepworth
Object Recognition 35

& Smith, 2002; Mabbott & Smith, 2003). These results cessing faces. Neuropsychological studies provided the
suggest that hemispheric biases ( left verbal, right visu- first evidence to support this view, with reports of dou-
ospatial) develop at some point after 5 years of age. ble dissociation of face and object processing. That is,
there are patients who show impaired face processing
Spatial Navigation but relatively intact general vision and object processing
(with the occasional exception of color vision; reviewed
The ability to remember locations and routes in the vi- in Barton, 2003), and other patients who show the oppo-
sual environment relies on both body-centered (e.g., site pattern of deficit (e.g., Moscovitch, Winocur, &
ocentric) and environment-centered (allocentric) spa- Behrmann, 1997). These studies also hinted that dam-
tial information. In adults, the hippocampus appears age to the right hemisphere might be necessary for the
to be involved in spatial memory (Astur, Taylor, Mame- face-processing impairments to be observed. More re-
lak, Philpott, & Sutherland, 2002; Nunn, Polkey, & cently, ERP, MEG, and fMRI methods have been used
Morris, 1998), particularly with respect to allocentric to identify the pathways involved in face processing in
spatial memory (Abrahams, Pickering, Polkey, & Mor- the intact brain. These studies have confirmed and
ris, 1997; Holdstock et al., 2000; Incisa della Rocchetta extended findings from brain-injured patients, indicat-
et al., 2004). It has been suggested that the hippocam- ing that a distributed network of regions in the brain me-
pus is involved in the consolidation of allocentric infor- diate face processing: occipito-temporal regions are
mation into long-term memory, rather than the initial important for the early perceptual stages of face pro-
encoding of allocentric spatial information (Holdstock cessing, with more anterior regions, including areas of
et al., 2000). the temporal and frontal cortices and the amygdala, in-
A case study of a patient with perinatal selective hip- volved in processing aspects such as identity and emo-
pocampal damage also shows deficits in allocentric spa- tional expression (Adolphs, 2002; Haxby, Hoffman, &
tial processing. This patient was tested in a virtual Gobbini, 2002). In this section, we focus mainly on the
reality town and showed a massive additional impair- involvement of occipito-temporal cortex and the amyg-
ment when tested from a shifted viewpoint compared dala, as these are the areas for which the most develop-
with a mild, list length-dependent impairment when mental data are available.
tested from the original viewpoint (King, Burgess, Hart-
ley, Vargha-Khadem, & OKeefe, 2002). The same pa-
tient was impaired on all topographical tasks on a Occipito-Temporal Cortex
second virtual reality study and on his recall of the con- In adults, a network including the inferior occipital
text-dependent questions. In contrast, he showed normal gyrus, the fusiform gyrus, and the superior temporal sul-
recognition of objects from the virtual town, and of top- cus is important for the early stages of face processing
ographical scenes (Spiers, Burgess, Hartley, Vargha- (Haxby et al., 2002). In this view, the inferior occipital
Khadem, & OKeefe, 2001). These results suggest a gyrus is primarily responsible for early perception of fa-
relative lack of developmental plasticity following hip- cial features, while the fusiform gyrus and the superior
pocampal lesions, at least with respect to allocentric temporal sulcus are involved in more specialized pro-
spatial encoding. cessing (Haxby et al., 2002). In particular, the fusiform
gyrus is thought to be involved in the processing of in-
OBJECT RECOGNITION variant aspects of faces (such as the perception of
unique identity), whereas the superior temporal sulcus is
Face/Object Recognition involved in the processing of changeable aspects (such as
perception of eye gaze, expression, and lip movement;
Among the numerous visual inputs that we receive each Haxby et al., 2002; Hoffman & Haxby, 2000).
moment, the human face is perhaps one of the most Perhaps the most intensively studied of these regions
salient. The importance of the many signals it conveys is an area of fusiform gyrus labeled the fusiform face
(e.g., emotion, identity, direction of eye gaze), together area (Kanwisher, McDermott, & Chun, 1997; Puce, Al-
with the speed and ease with which adults typically pro- lison, Asgari, Gore, & McCarthy, 1996). This region is
cess this information, are compelling reasons to suppose more activated to faces compared with other objects or
that there may exist brain circuits specialized for pro- body parts (Kanwisher et al., 1997; Puce et al., 1996).
36 Neural Bases of Cognitive Development

Although some investigators have argued against the faces is likely mediated by subcortical mechanisms
view of face-specific patches of cortex and have instead (e.g., superior colliculus; for a review of the evidence,
emphasized the distributed nature of the representation see Johnson & Morton, 1991), and that cortical mecha-
of object feature information over the ventral posterior nisms do not begin to emerge until 2 to 3 months of age.
cortex (Haxby et al., 2001), even these authors acknowl- At this early age, cortical areas are thought to be rela-
edge that the response to faces appears unique in certain tively unspecialized (Johnson & Morton, 1991; Nelson,
ways (e.g., extent of activation, modulation by attention; 2001). One possible role of the earlier-developing sub-
Ishai, Ungerleider, Martin, & Haxby, 2000). cortical system is to provide a face-biased input to the
Although these studies appear to suggest that particu- slower-developing occipito-temporal cortical system,
lar regions of cortex are devoted specifically to face pro- and to provide one mechanism whereby an initially more
cessing, this interpretation has been questioned. In broadly tuned processing system becomes increasingly
particular, it has been argued that the supposed face- specialized to respond to faces during development
specific cortical areas are not specific to faces per se, but (Johnson & Morton, 1991; Nelson, 2001).
instead are recruited for expert-level discrimination of The only functional neuroimaging study to investi-
complex visual patterns, whether faces or other classes of gate face processing in human infants confirms that oc-
objects (R. Diamond & Carey, 1986; Gauthier et al., cipito-temporal cortical pathways are involved by 2 to 3
1999). In this view, the mechanisms active during devel- months of age. In this study, 2-month-olds positron
opment of face processing could be the same as those ob- emission tomography (PET) activation in the inferior
served in adults learning an equally challenging visual occipital gyrus and the fusiform gyrus, but not the supe-
perceptual task. In support of this view, studies have rior temporal sulcus, was greater in response to a human
shown that the fusiform face area is also activated by face than to a set of three light diodes (Tzourio-
nonface objects (e.g., cars) if the subjects are experts Mazoyer et al., 2002). These results demonstrate that
with that category (Gauthier, Skudlarski, Gore, & An- areas involved in face processing in adults can also be
derson, 2000), and activation in the fusiform face areas activated in infants by 2 months of age, although they do
increases following training of expertise with a category not address the question of whether these areas are
of visual forms (Gauthier et al., 1999). specifically activated by faces rather than by other vi-
Developmental studies can provide important infor- sual stimuli. It is interesting that the superior temporal
mation to constrain the claims of the different sides of sulcus, suggested to be involved in the processing of in-
this debate. For example, by studying when and how formation relevant to social communication, was not ac-
face-specific brain responses emerge, developmental tivated in this study. One possible explanation is that the
studies can provide some hints as to whether and how stimuli (static and neutral) were not optimal for activat-
much experience might be needed for these responses to ing processing in the superior temporal sulcus. However,
emerge. Behavioral studies provide a suggestion that the observation that activation in the superior temporal
face-processing pathways may be functional from very sulcus has been found in adults even in response to
early in life: newborn babies move their eyes, and some- static, neutral faces (e.g., Kesler-West et al., 2001) ar-
times their heads, longer to keep a moving facelike pat- gues against this interpretation. It is possible that the su-
tern in view than several other comparison patterns perior temporal sulcus plays a different role in the
(Johnson, Dzuirawiec, Ellis, & Morton, 1991). While face-processing network in infants than in adults, since
there is a debate as to whether this reflects a specific re- in primates its connectivity with other visual areas is
sponse to facelike configurations or a lower-level visual known to differ in infants compared with adult monkeys
preference (e.g., for patterns with higher density of ele- (Kennedy, Bullier, & Dehay, 1989).
ments in the upper visual field, see Turati, Simion, Mi- Event-related potential studies support the idea that
lani, & Umilta, 2002; see also Banks & Ginsburg, 1985; cortical mechanisms are involved in face processing
Banks & Salapatek, 1981), there is some agreement from at least 3 months of age. However, these studies
among the divergent views that the ultimate result is that also suggest that when cortical mechanisms become in-
facelike patterns are preferred to other arrangements volved in infants processing of faces, they are less
from the first hours to days of life. tuned in to faces than is the mature system. Two stud-
While this might seem to support the notion that ies have shown that face-responsive ERP components
face-specific cortical areas are active from birth, the are more specific to human faces in adults than in in-
prevailing view is that this early preferential orienting to fants (de Haan, Pascalis, & Johnson, 2002; Halit, de
Object Recognition 37

Haan & Johnson, 2003). In adults, the N170, a negative procedure discriminate only between individual human
deflection over occipito-temporal electrodes that peaks faces (Pascalis, de Haan, & Nelson, 2002). The results
approximately 170 ms after stimulus onset, is thought to also suggest that the structural encoding of faces may be
reflect the initial stage of the structural encoding of the dispersed over a longer time in infants than in adults. It
face. Although the location in the brain of the genera- is possible that, as the processes become more auto-
tor(s) of the N170 remains a matter of debate, it is gen- mated, they are carried out more quickly and/or in par-
erally believed that regions of the fusiform gyrus allel rather than in serial fashion.
(Shibata et al., 2002), the posterior inferior temporal Interestingly, the spatial distribution of the N170 and
gyrus (Bentin et al., 1996; Shibata et al., 2002), lateral P400 both change from 3 to 12 months, with maxima
occipito-temporal cortex (Schweinberger, Pickering, shifting laterally for both components (de Haan et al.,
Jentzsch, Burton, & Kaufman, 2002), and the superior 2002; Halit et al., 2003). In addition, the maxima of
temporal sulcus (Henson et al., 2003) are involved. The these components appear more superior than in adults, a
N170 is typically of larger amplitude and/or has a longer result consistent with studies in children finding a shift
latency for inverted than upright faces (Bentin et al., from superior to inferior maximum of the N170 with age
1996; de Haan et al., 2002; Eimer, 2000; Itier & Taylor, (Taylor, Edmonds, McCarthy, & Allison, 2001). This
2002; Rossion et al., 2000), a pattern that parallels be- might reflect a change in the configuration of generators
havioral studies showing that adults are slower at recog- underlying these components with age.
nizing inverted than upright faces (Carey & Diamond, Investigations with older children also support the
1994). In adults, the effect of inversion on the N170 is view of a gradual specialization of face processing.
specific for human faces and does not occur for inverted ERP studies suggest that there are gradual, quantitative
compared with upright exemplars of nonface object cat- improvements in face processing from 4 to 15 years of
egories (Bentin et al., 1996; Rebai, Poiroux, Bernard, & age rather than stagelike shifts (Taylor, McCarthy, Sal-
Lalonde, 2001; Rossion et al., 2000), even animal (mon- iba, & Degiovanni, 1999). The ERP studies also provide
key) faces (de Haan et al., 2002). evidence that there is a slower maturation of configural
Developmental studies have identified two compo- than featural processing: Responses to eyes presented
nents, the N290 and the P400, believed to be precursors alone matured more slowly than responses to eyes pre-
of the N170. Both components are maximal over poste- sented in the configuration of the face (Taylor et al.,
rior cortex, with the N290 peaking at about 290 ms after 2001). There have not been many functional imaging
stimulus onset and the P400 about 100 ms later. The studies of face processing during childhood, but one
N290 shows an adultlike modulation of amplitude by study examining facial identity processing suggests that
stimulus inversion by 12 months of age: inversion in- changes in the network are activated in 10- to 12-year-
creases the amplitude of the N290 for human but not olds compared with adults (Passarotti et al., 2003).
monkey faces (Halit et al., 2003). The P400 has a Children showed a more distributed pattern of activa-
quicker latency for faces compared with objects by 6 tion compared with adults. Within the fusiform gyrus,
months of age (de Haan & Nelson, 1999) and shows an children tended to show more activation in lateral areas
adultlike effect of stimulus inversion on peak latency by of the right hemisphere than did adults: In the left hemi-
12 months of age: It is of longer latency for inverted than sphere, children showed greater lateral than medial ac-
upright human faces but does not differ for inverted tivation, whereas adults showed no difference for the
compared with upright monkey faces (Halit et al., two areas. In addition, children showed twice as much
2003). At 3 and 6 months, the N290 is unaffected by in- activation of the middle temporal gyrus as adults. The
version; and the P400, while modulated by inversion, authors interpret these results as suggesting that in-
does not show effects specific to human faces (de Haan creased skill is associated with a more focal pattern of
et al., 2002; Halit et al., 2003). Overall, these findings activation.
suggest that there is a gradual emergence of face- Studies of children with autistic spectrum disorders
selective response over the first year of life (and support the view that atypical activation of occipito-
beyond). This finding is consistent with results of a be- temporal cortex is related to impairments in face pro-
havioral study showing a decrease in discrimination cessing. Autistic spectrum disorders are characterized
abilities for nonhuman faces with age; 6-month-olds can by impairments in processing of social information, in-
discriminate between individual humans and monkeys, cluding faces. Functional imaging studies indicate that
whereas 9-month-olds and adults tested with the same when individuals with autism or Asperger syndrome
38 Neural Bases of Cognitive Development

view faces, they show a diminished response in the 1989; Holstege, van Ham, & Tan, 1986), these results
fusiform gyrus compared with controls (Hubl et al., suggest that by 5 months of age, portions of the amyg-
2003; Pierce, Muller, Ambrose, Allen, & Courchesne, dala circuitry underlying the response to facial expres-
2001; Schultz et al., 2000) and show an increased acti- sions may be functional.
vation of object-processing areas in the inferior tempo- Interestingly, there is evidence that early damage to
ral cortex (Hubl et al., 2003; Schultz et al., 2000). It is the amygdala may have a more pronounced effect on
possible that this reflects a different processing strat- recognition of facial expression than damage sustained
egy in which individuals with autism focus more on later in life. For example, in one study of emotion recog-
featural rather than configural information in the face. nition in patients who had undergone temporal lobec-
In other words, individuals with autism may rely more tomy as treatment for intractable epilepsy, emotion
on general purpose object-processing pathways rather recognition in patients with early, right mesial temporal
than specialized face-processing pathways when view- sclerosis, but not those with left-sided damage or ex-
ing faces. tratemporal damage, showed impairments on tests of
recognition of facial expressions of emotion but not on
Amygdala comparison tasks of face processing (Meletti et al.,
2003). This deficit was most pronounced for fearful ex-
The amygdala is a heterogeneous collection of nuclei lo- pressions, and the degree of deficit was related to the
cated in the anterior temporal lobe. Several studies in age of first seizure and epilepsy onset.
adults have shown that lesions to the amygdala impair
emotion recognition, even when they leave other aspects
Role of Experience
of face processing intact (e.g., identity recognition;
Adolphs, Tranel, Damasio, & Damasio, 1994). Lesion The preceding studies suggest that the cortical system
studies also indicate that recognition of fearful expres- involved in face processing becomes increasingly spe-
sions is particularly vulnerable to such damage (Adolphs cialized for faces throughout the course of development.
et al., 1994, 1999; Broks et al., 1998; Calder et al., Several developmental theories propose that experience
1996). Functional imaging studies in healthy adults and is necessary for this process of specialization to occur
school-age children complement these findings, with (e.g., Nelson, 2001). Only a few studies have directly ex-
some studies showing that the amygdala responds to a amined the role of experience in development of face
variety of positive, negative, or neutral expressions processing. In one series of studies, the face-processing
(Thomas et al., 2001; Yang et al., 2002), and other stud- abilities of patients with congenital cataracts who were
ies suggesting that the amygdala is particularly respon- deprived of patterned visual input for the first months of
sive to fearful expressions (Morris et al., 1996; Whalen life were tested years after this period of deprivation.
et al., 2001). These patients show normal processing of featural infor-
There is indirect evidence that the amygdala plays a mation (e.g., subtle differences in the shape of the eyes
role in processing facial expressions in infants. Balaban and mouth), but show impairments in processing config-
(1995) used the eye-blink startle response (a reflex ural information (i.e., the spacing of features within the
blink initiated involuntarily by sudden bursts of loud face; Le Grand et al., 2001; Geldart, Mondloch, Maurer,
noise) to examine the psychophysiology of infants re- de Schonen, & Brent, 2002). This pattern was specific
sponses to facial expressions. In adults, these reflex to faces in that both featural and configural aspects of
blinks are augmented by viewing slides of unpleasant geometric patterns were processed normally (Le Grand
pictures and scenes, and they are inhibited by viewing et al., 2001). Moreover, when patients were examined
slides of pleasant or arousing pictures and scenes (Lang, whose visual input had been restricted mainly to one
Bradley, & Cuthbert, 1990, 1992). Consistent with the hemisphere during infancy, it was found that visual
adult findings, Balaban found that 5-month-old infants input to the right hemisphere, but not the left hemi-
blinks were augmented when they viewed angry expres- sphere, was critical for an expert level of face processing
sions and were reduced when they viewed happy expres- to develop (Le Grand, Mondloch, Maurer, & Brent,
sions, relative to when they viewed neutral expressions. 2003). These studies suggest that visual input during
As animal studies indicate that the fear potentiation of early infancy is necessary for the normal development of
the startle response is mediated by the amygdala (Davis, at least some aspects of face processing.
Executive Functions 39

Another way in which the role of experience has been with the innate modularity view that certain modules
investigated is by studying children who experience are impaired or intact from the start, since it shows that
atypical early emotional environments. For example, early abilities do not necessarily predict later ones.
Pollak and colleagues have found that perception of the Advocates of the innate modularity view might argue
facial expression of anger, but not other expressions, is that there could still be innate modules, because face
altered in children who are abused by their parents. processing has been noted as a strength even from early
Specifically, they report that, compared with nonabused in life in Williams syndrome. Moreover, adults with
children, abused children show a response bias for anger Williams syndrome can achieve near normal perfor-
(Pollak, Cicchetti, Hornung, & Reed, 2000), identify mance on standardized tests of face recognition. Could
anger based on less perceptual input (Pollak & Sinha, this be an intact innate module? Proponents of the
2002), and show altered category boundaries for anger interactive specialization view suggest that, if so, face
(Pollak & Kistler, 2002). These results suggest that atyp- recognition should be accomplished by the same cogni-
ical frequency and content of their emotional interac- tive mechanisms as in typically developing individuals.
tions with their caregivers result in a change in the basic However, several studies show this is not the case. Indi-
perception of emotional expressions in abused children. viduals with Williams syndrome appear to rely less than
typically developing children on cues from the configu-
Is There a Visuospatial Module? ration of features in the face to recognize identity
(Deruelle, Mancini, Livet, Casse-Perrot, & de Schonen,
There are different perspectives regarding whether vi- 1999). In addition, their ERPs show atypical patterns
suospatial processing is mediated by a specific module within the first several hundred milliseconds of pro-
in the brain. According to the innate modularity view, cessing (Grice et al., 2001.). Together, these results
the cognitive modules observed in adults are genetically argue against the view that the cognitive strengths in
specified and are already present from the beginnings Williams syndrome reflect the operation of intact, spe-
of development. This view predicts that, even early in cific modules.
development, one or more modules can be impaired
without affecting other modules. Williams syndrome,
together with another developmental disorder called
Specific Language Impairment (SLI ), provides support EXECUTIVE FUNCTIONS
for this view. In Williams syndrome, language is a
strength despite lowered intellect; in SLI, language is Most high-level cognitive functions involve executive
impaired despite normal intellect. The double dissocia- processes, or cognitive control functions, such as atten-
tion of language and general intellect across the two de- tion, planning, problem solving, and decision making.
velopmental disorders is taken by some to support the These processes are largely voluntary (as opposed to au-
view that there is a language module that can be selec- tomatic) and are highly effortful. Such functions, in-
tively spared or impaired. Also consistent with this view cluding selective and executive attention, inhibition, and
is the finding that LIMK1 hemizygosity in Williams working memory, are hypothesized to improve with age
syndrome is related to impaired visuospatial construc- and practice, and to vary with individual differences
tive cognition (Frangiskakis et al., 1996). in motivation or intelligence. These cognitive control
An opposing view is that the brain is not initially or- processes have been described as providing a supervi-
ganized into distinct modules and only becomes so sory attention system (Shallice, 1988)a system for
during development as an emergent property of both bio- inhibiting or overriding routine or reflexive behaviors in
logical maturation and experience in the environment. favor of more controlled or situationally appropriate or
One piece of evidence in support of this view is that the adaptive behaviors. Desimone and Duncan (1995) de-
cognitive profile of Williams syndrome is not constant scribe this system as an attentional bias that provides a
over development. Whereas adults and older children mechanism for attending to relevant information by si-
with Williams syndrome show a strength in verbal skills multaneously inhibiting irrelevant information (Casey,
and weakness in number skills, infants show the oppo- Durston, & Fossella, 2001). The ability to override a
site pattern (Paterson, Brown, Gsodl, Johnson, & dominant response or ignore irrelevant information is
Karmiloff-Smith, 1999). This finding is incompatible critical in everyday life, as evidenced by the functional
40 Neural Bases of Cognitive Development

impairments associated with chronic inattention, behav- tional control as well as response inhibition. Whatever
ioral impulsivity, or poor planning and decision making. scheme is used, it is apparent that the classic executive
Classic lesion cases, such as the famous case of function tasks involve more than one of the preceding
Phineas Gage, indicate that injury to the prefrontal cor- aspects of voluntary control or regulation. In the follow-
tex can result in difficulties in behavioral regulation, ing sections, we provide examples of behavioral tasks
such as impulsivity and socially inappropriate behavior thought to tap various aspects of executive function
(Fuster, 1997), as well as disruptions in planning, work- across development, as well as provide select examples
ing memory, and focused attention. Cognitive develop- that evidence the role of specific regions of prefrontal
mentalists will recognize these same functions as cortex in supporting cognitive control. Of course, pre-
showing relatively protracted behavioral maturation, frontal cortex does not act in isolation. Other brain
often not reaching adult levels of performance until late regions are assumed to be integral to the executive func-
adolescence (Anderson, Anderson, Northam, Jacobs, & tion system, providing input and feedback, as well as re-
Catroppa, 2001). It is therefore not surprising that pre- ceiving inputs from prefrontal cortex. Developmental
frontal cortex shows one of the longest periods of devel- improvements in executive function may arise as much
opment of any brain region (A. Diamond, 2002; Luciana, from the development of such functional integration as
2003, for reviews). In fact, the relations between pre- from the architectural and physiological development of
frontal cortex development and the development of exec- prefrontal cortex (Anderson et al., 2001; Anderson,
utive functions is probably one of the clearest relations Levin, & Jacobs, 2002).
in the developmental cognitive neuroscience literature.
However, this does not imply that we fully understand Working Memory Revisited
the instantiation of attention, working memory, or inhi-
bition in the brain. Instead, we have significant evidence Perhaps the task most clearly associated with both child
from lesion and neuroimaging methods to relate subre- development and prefrontal cortex is the classic A-not-B
gions of prefrontal cortex to specific aspects of cogni- task. In this paradigm (or its close cousin, the delayed
tive control in adulthood, as well as a growing body of response task), an infant or animal watches an object
literature addressing the normative and atypical func- being hidden at one of two identical locations, and after
tion of these regions, and their connected networks of some delay is rewarded for retrieving the object. This
structures, in the development of cognitive control. For task requires both holding information in mind across a
reasons of space, only illustrative examples from norma- delay and, on subsequent trials when the hiding location
tive behavioral development, animal models, adult and changes, inhibiting a prepotent tendency to return to
pediatric neuroimaging studies, and atypical develop- a previously correct response location (A. Diamond,
mental populations are provided here (Casey, Durston, 1985). Animal lesion studies support the importance of
et al., 2001; A. Diamond, 2002; Luciana, 2003). DLPFC in successful performance on the A-not-B and
delayed response tasks (A. Diamond & Goldman-Rakic,
Domains of Executive Function 1989; Fuster & Alexander, 1970; Goldman & Rosvold,
1970). In addition, electrophysiological studies indicate
Many researchers have identified working memory and that cells in this region actively respond during the
behavioral inhibition as the primary functions of pre- delay interval, suggesting that DLPFC is involved in
frontal cortex, and by extension, the basic components the maintenance of information in working memory
of executive function (e.g., A. Diamond, 2001). Working (Funahashi, Bruce, & Goldman-Rakic, 1989; Fuster &
memory has typically been associated with dorsolateral Alexander, 1971). Further investigation demonstrates
prefrontal cortex (DLPFC; J. Cohen et al., 1994; Fuster, that lesions to this region impair performance only
1997; Levy & Goldman-Rakic, 2000), while more ven- under delay conditions and not during immediate object
tral regions have been implicated in inhibition of a pre- retrieval (A. Diamond & Goldman-Rakic, 1989). Func-
potent behavioral response (Casey, Trainor, Orendi, tional imaging studies suggest that developmental dif-
et al., 1997; Kawashima et al., 1996; Konishi et al., ferences in working memory function, at least in middle
1999). Other investigators have parsed their definition childhood, may be reflected in less efficient or less
of executive functions somewhat differently in an effort focal activation of DLPFC. That is, pediatric fMRI
to include the voluntary and effortful aspects of atten- studies have demonstrated that children activate similar
Executive Functions 41

regions of DLPFC compared with adults during both Importantly, these same studies highlight the impor-
verbal and spatial working memory tasks, but also may tance of regions beyond the prefrontal cortex. In partic-
activate additional areas of prefrontal cortex, including ular, basal ganglia structures have also been shown to be
ventral lateral regions (VLPFC; Casey et al., 1995; Nel- involved in response inhibition (e.g., Luna et al., 2001),
son, Lin, et al., 2000; Thomas et al., 1999). perhaps particularly so for children (Bunge et al., 2002;
Casey, Trainor, Orendi, et al., 1997; Durston et al.,
2002). Children with ADHD show significantly lower
Inhibitory Control activity in basal ganglia regions during performance of
a go/no-go task than typically developing children
Although working memory has been associated with (Durston et al., 2003; Vaidya et al., 1998) and show high
dorsolateral regions of the prefrontal cortex, the ability rates of false alarms on the task. In a recent study, chil-
to inhibit inappropriate behaviors has typically been as- dren with ADHD showed additional recruitment of dor-
sociated with ventral medial or orbital frontal cortex solateral PFC not observed for the control group who
(Casey, Trainor, Orendi, et al., 1997; Konishi et al., performed at a high rate of accuracy (Durston et al.,
1999). In adults, lesions to ventral prefrontal cortex lead 2003). Interestingly, when taking medication to treat
to impulsive and socially inappropriate behavior (Bar- their inattention and impulsivity, children with ADHD
rash, Tranel, & Anderson, 1994; Damasio, Grabowski, show basal ganglia activity equivalent to the typically
Frank, Galaburda, & Damasio, 1994). One common de- developing group along with parallel improvements in
velopmental measure of response inhibition is the go/no- behavioral performance (Vaidya et al., 1998). Other
go paradigm. In this task, children are asked to respond developmental disorders, such as Tourette syndrome,
to every stimulus except one (e.g., all letters except X). obsessive-compulsive disorder, and childhood-onset
The task is designed such that the majority of trials are schizophrenia also have been associated with disruption
go trials, building up a compelling behavioral re- of frontal-striatal circuitry (connections between basal
sponse tendency. The childs ability to refrain from ganglia and frontal cortex) and impaired performance
making the response at the occurrence of the no-go on tasks involving attentional control.
stimulus is used as an index of inhibitory control. Per-
formance on such tasks has been shown to improve
across the preschool and school-age years (Casey, Attentional Control
Durston, et al., 2001; Ridderinkhof, van der Molen,
Band, & Bashore, 1997). Neuroimaging studies using Beyond the general processes of working memory or in-
the go/no-go paradigm have demonstrated signal in- hibition, many real-world and experimental tasks re-
creases in ventral PFC during periods high in inhibitory quire selectively focusing attention on relevant task
demand (Casey, Forman, et al., 2001; Casey, Trainor, information while simultaneously suppressing interfer-
Orendi, et al., 1997), with correspondingly lower levels ence from salient but irrelevant or misleading informa-
of activity during periods of low inhibitory demand. tion (Casey, Durston, et al., 2001). Perhaps the most
Konishi et al. (1999) observed increased ventral PFC ac- studied adult task of this type is the color-word Stroop
tivation during no-go trials in an event-related fMRI paradigm, in which participants are asked to identity the
paradigm. Recent pediatric neuroimaging studies have color of ink in which a word is written but, in the case of
demonstrated both developmental differences in activa- a color word, have to inhibit a natural tendency to read
tion of ventral prefrontal cortex (Bunge, Dudukovic, the word instead (e.g., the word BLUE presented in
Thomason, Vaidya, & Gabrieli, 2002), with children red ink). Neuroimaging data from the Stroop paradigm
showing reduced signal compared with adults, and in- identify medial prefrontal cortex, specifically the ante-
creasing activation of ventral lateral PFC with increas- rior cingulate cortex, as particularly important in de-
ing inhibitory load (Durston, Thomas, & Yang, 2002). tecting (e.g., Botvinick, Braver, Barch, Carter, &
Durston and colleagues (2002) showed that behavioral Cohen, 2001; Bush et al., 1999; Bush, Luu, & Posner,
performance on the go/no-go task was significantly cor- 2000; Duncan & Owen, 2000; Fan, Flombaum, McCan-
related with activity in inferior frontal cortex, as well as dliss, Thomas, & Posner, 2003; MacDonald, Cohen,
other prefrontal regions, including the anterior cingulate Stenger, & Carter, 2000; Posner & Petersen, 1990) and
gyrus (ACC). perhaps even resolving this type of attentional conflict.
42 Neural Bases of Cognitive Development

DLPFC and other regions of prefrontal cortex may pared with 8- to 10-year-olds and adults. Despite a signif-
also be activated during cognitive conflict depending on icant literature demonstrating developmental improve-
demands of the specific task. In the Simon task, a spa- ments in cognitive control across early and middle
tial conflict is created between the location of a stimu- childhood, fewer studies have addressed the brain bases
lus ( left or right side of the screen) and the required of this development. A. Diamond (2001) has shown that
response ( left or right button; Gerardi-Caulton, 2000). children with presumed functional disruptions of pre-
In a study by Fan et al. (2003), this conflict was associ- frontal cortex (children treated for phenylketonuria or
ated with activation of superior frontal gyrus as well as PKU) are impaired on their performance of Stroop-like
anterior cingulate cortex, while conflict in the Stroop and inhibitory control tasks like the go/no-go task, sug-
task was associated with activity in ventral lateral pre- gesting that typical developmental function is relying on
frontal cortex. When the same adult participants this brain region that is continuing to develop across
performed the Eriksen flanker task, which requires fo- childhood. Similar effects on other executive function
cusing attention on a central stimulus and actively ig- tasks have also been observed in this population (Lu-
noring competing flanking stimuli, these authors ciana, Sullivan, & Nelson, 2001). Likewise Casey, Tot-
observed attention-related activity in premotor cortex. tenham, and Fossella (2002) have demonstrated that
Despite these task-related differences in MR signal, fur- children with psychiatric disorders associated with
ther analyses demonstrated overlapping regions of ac- frontal-striatal circuitry show specific impairments on
tivity in the anterior cingulate gyrus and the left tests of cognitive conflict. Children with obsessive-com-
prefrontal cortex across all three tasks, suggesting some pulsive symptomatology showed deficits when required
common activity related to cognitive control or manage- to inhibit a well-learned response set, but no impairment
ment of cognitive conflict (Fan et al., 2003). Other neu- in a Stroop-like task. In contrast, children with diagnoses
roimaging studies of the Eriksen flanker task have of childhood onset schizophrenia were impaired on the
indicated that prefrontal cortex is differentially acti- Stroop-like task but not on the response selection task or
vated based on the degree of cognitive conflict within on a go/no-go inhibition task, suggesting potential differ-
the same task (Casey et al., 2000; Durston et al., 2003). ences within prefrontal cortex.
Recently, Durston and colleagues (2003) found that A related aspect of cognitive control arises when the
parametric manipulations of conflict on the flanker task individual is required not only to ignore irrelevant infor-
were associated with monotonic increases in both mation, but also to shift among multiple rules for re-
DLPFC and ACC. Of course, additional brain regions sponding. The classic adult task of this type is the
are activated beyond the prefrontal cortex, such as the Wisconsin Card Sorting Task (WCST), in which partici-
superior parietal cortex, perhaps related to the spatial pants must discover the sorting rule simply on the basis of
nature of the task. Applying the concept of a spotlight binary feedback received during card-sorting perfor-
of attention (Posner & Raichle, 1997), the flanker task mance. Healthy adults show rapid acquisition of the ini-
requires narrowing the spatial distribution of attentional tial sorting rule and quickly alter their behavior if and
focus to reduce interference or conflict from the irrele- when the rule is changed. Lesions to the left DLPFC but
vant flanking stimuli. Activation of superior parietal not other regions of prefrontal cortex impair performance
cortex may be related to this spatial feature of the task on switching tasks, resulting in perseverative errors
(also see Orienting, later in this chapter). (Keele & Rafal, 2000; Owen et al., 1993; Shallice &
Behaviorally, the cognitive control tasks previously Burgess, 1991). Neuroimaging studies provide convergent
described show significant developmental changes across evidence for the role of DLPFC and basal ganglia circuits
early and middle childhood, and in some cases, even into in task switching and reversal learning (Cools, Clark, &
adolescence. Casey, Durston, and Fossella (2001) provide Robbins, 2004; Cools, Clark, Owen, & Robbins, 2004).
developmental data demonstrating that, for tasks such as Perhaps not surprisingly, typically developing 3-year-
the Eriksen flanker and Stroop, adultlike performance is old children perform very similarly to adults with
not achieved until early adolescence. A. Diamond (2002) frontal lobe lesions. Zelazo et al. (1996) have described
has shown similar trajectories for Stroop-like tasks in- a dramatic developmental shift in set-switching perfor-
cluding the Simon task, with evidence of protracted de- mance in the preschool years using the dimensional card
velopment. Rueda and colleagues (2004) showed evidence sorting task. This task requires the child to first sort
of increased cognitive conflict in 6- and 7-year-olds com- cards by one criterion (e.g., shape), and then to shift and
Nonexecutive Aspects of Attention 43

sort the cards by another criterion (e.g., color). While 3- subserving arousal, one common view of arousal is Pos-
year-olds have no trouble sorting by the first criterion, ners alerting model, identifying regions of right frontal
whether it is shape or color, they frequently fail to shift and parietal cortices and the norepinephrine system as
their sorting behavior when the criterion changes de- critical players in vigilance (Lewin et al., 1996). In this
spite being able to verbalize the rule, reminiscent of model, alertness leads to quieting of other activity in
some patients with prefrontal cortex damage (REF). the brain through the release of norepinephrine (NE)
However, 5-year-olds generally have no difficulty with from the locus coeruleus. The presence of NE is hy-
this task. Kirkham and colleagues have suggested that pothesized to bring on the state of alertness by enhanc-
this developmental shift is predominantly the result of ing the signal-to-noise ratio in regions where it is
improved inhibitory control (A. Diamond, Kirkham, & acting. As alerting-related activity in right frontal cor-
Amso, 2002). tex increases, activity in the anterior cingulate cortex, a
The developmental neuroimaging data in this domain region associated with target detection decreases. Pos-
are still sparse, although a number of groups are work- ner suggests that this decrease in activity is important
ing in this direction. Additional work will be needed to for the brain to reduce potential sources of interference
assess developmental changes in the recruitment and ef- during the waiting period prior to the appearance of a
ficiency of prefrontal cortex in such cognitive conflict target stimulus (Posner & Raichle, 1997). Arousal or
or cognitive control tasks. It remains to be seen whether sustained attention is associated with decreases in heart
the normal developmental pattern of functional brain ac- rate, both in adults and in infants, reflecting activity of
tivity engages the same circuits known to be disrupted in cardioinhibitory centers in the brain, particularly in or-
adult lesion populations with executive dysfunctions. bitofrontal cortex (Richards, 2001). Using heart rate
measures as an index of the brains arousal system,
Richards and colleagues have demonstrated develop-
NONEXECUTIVE ASPECTS OF ATTENTION mental improvements in sustained attention over the
first 6 months of life (Casey & Richards, 1988;
The previously discussed concepts of attentional control Richards, 1994). Likewise, certain components of the
and cognitive conflict, while well-studied in adult popu- ERP in infancy are thought to reflect automatic alerting
lations, are more difficult to apply to the earliest periods responses, particularly in response to novel stimuli
of development. Although tasks such as the A-not-B or (Courchesne, 1978). ERP studies of the response to fa-
delayed response fall under the broad rubric of executive miliar and novel events in 6-month-olds show that the
functions, the majority of research on infants attention early negative component (Nc) typically elicited by
involves more basic processes, such as maintaining an stimulus novelty is enhanced during periods of sus-
alert state, and orienting to stimuli in the environment. tained attention (as indexed by heart rate deceleration),
While these functions can involve voluntary or con- compared with periods of attention termination
trolled processes (Posner & Raichle, 1997), often they (Richards, 1998). However, Richards found that, during
reflect automatic or obligatory responses. Not unlike the sustained attention, the Nc component did not differen-
memory literature discussed previously, the various tiate among familiar and novel stimuli. Instead, a later
forms of attention have been associated with somewhat component, the positive slow wave (PSW), was en-
separable brain systems (Posner & Raichle, 1997). hanced to familiar but infrequently occurring stimuli
(Richards, 1998), suggesting that this basic alerting
function may modulate the brain response to stimulus
Alerting, Vigilance, or Arousal
novelty, as might be expected behaviorally.
One major function of the attentional system is to main-
tain alertness or vigilance (also called arousal). Vigi- Orienting
lance processes allow the brain to prepare for upcoming
stimuli that may require decision making and/or a be- The third arm of the Posner model of attentional net-
havioral response. Alertness or arousal is associated works is the orienting network. Posner describes both
with improved behavioral performance in a nonspecific overt and covert forms of orienting, or attentional shift-
or unfocused way (Posner & Raichle, 1997). Although ing in adults (Posner & Raichle, 1997; Posner, Walker,
researchers have proposed various neural mechanisms Friedrich, & Rafal, 1984). When a cue stimulus flashes
44 Neural Bases of Cognitive Development

in the periphery, adults tend to automatically shift eye work of brain regions. The developmental trajectory of
gaze to the periphery (overt attention shift). In contrast, orienting behavior has been examined most often using
a centrally presented stimulus can cue the individual to the inhibition of return paradigm. Inhibition of return
either overtly orient ( by shifting eye gaze to fixate) or refers to the tendency to avoid returning attention to a
covertly orient ( by shifting attention without shifting location just previously attended. Adult studies by
eye gaze). Therefore, in adults, centrally presented cues Rafal, Calabresi, Brennan, and Scioloto (1989) indicate
allow the individual to choose to orient in a controlled that inhibition of return occurs only in situations in
manner, whereas peripheral cues tend to elicit automatic which an eye movement is prepared, even if the actual
orienting responses. An extensive neuropsychological movement is never made. Studies of infants 3, 4, 6, 12,
literature implicates parietal cortex in the ability to ori- and 18 months of age showed that, although 6- to 18-
ent attention (e.g., Posner et al., 1984). Adults with pari- month-olds showed inhibition of return equivalent to
etal lesions show a deficit in the ability to disengage adults, 3- and 4-month-olds showed no evidence of inhi-
attention from one stimulus in order to shift attention bition of return (Clohessy, Posner, Rothbart, & Vecera,
elsewhere. Single cell recordings in animals suggest that 1991; Johnson, Posner, & Rothbart, 1994). This age co-
cells in the thalamus and parietal cortex show increased incides with behavioral evidence of the development of
activity during covert shifts of attention (D. Robinson, saccadic eye movements, as well as PET evidence of the
Bowman, & Kertzman, 1995). Lesions in other brain re- age at which parietal metabolism reaches adult levels
gions can mimic the unilateral behavioral neglect ob- (Chugani, Phelps, & Mazziotta, 1987). Hood et al.
served with parietal lesions, but do not appear to disrupt (1998) have demonstrated that 4-month-olds can show
covert orienting (Posner et al., 1984). Functional imag- evidence of covert orienting, however. When a cue is too
ing studies of covert orienting in adults show activity in short in duration to elicit an eye movement, 4-month-
superior parietal cortices, particularly contralateral to olds show improved performance for targets presented
the direction of attention (Corbetta, Kincade, Ollinger, at the cued location (Hood, Atkinson, & Braddick,
McAvoy, & Shulman, 2000). However, in contrast to the 1998). The importance of the development of the abili-
inability to disengage observed with parietal lesions, ties to disengage, shift, and reengage attention is proba-
damage to the midbrain visual system (superior collicu- bly obvious to developmental psychologists. These
lus) involved in saccadic eye movements leads to diffi- abilities are critical for learning new information from
culty in moving between stimuli, a separate step in our environmentsperceptual, cognitive, social, and
visual orienting (Rafal, Posner, Friedman, Inhoff, & emotionaland impairments in this system result in
Bernstein, 1988). Finally, damage to the pulvinar (part significant problems in everyday life.
of the thalamus) yields difficulties in reengaging atten-
tion or amplifying the target location once a movement
has been made (Danziger, Ward, Owen, & Rafal, 2001). THE FUTURE OF DEVELOPMENTAL
Several major developments occur in visual orienting COGNITIVE NEUROSCIENCE
behavior that are assumed to reflect development of the
brain systems supporting orienting in adults. Around 1 Based on our extensive review of the literature, the field
month of age, once infants are able to voluntarily fixate of developmental cognitive neuroscience has clearly ad-
on a stimulus, infants experience a period of obligatory vanced since the fifth edition of the Handbook was pub-
looking, during which they stare fixedly at objects and lished. We know a considerable amount about the neural
may have difficulty disengaging their fixation. This bases of a variety of cognitive abilities, although our
obligatory looking can be distressing when a baby be- knowledge base is uneven both between and within de-
comes fixated on strong visual stimuli, such as high- velopmental periods. We know more about the neural
contrast checkerboard patterns, but is very adaptive for bases of memory in infancy than we do in childhood,
forming social bonds when the baby is fixated on the and we know more about some executive functions than
caregivers face. Obligatory looking disappears around others (e.g., working memory versus planning).
4 months of age when infants develop more voluntary What does the future hold for those interested in this
control of their orienting (Posner & Raichle, 1997), pre- area? For starters, as our knowledge of brain develop-
sumably reflecting development of the orienting net- ment improves, our ability to ground behavior in the
References 45

brain should similarly improve. This, in turn, should Agster, K. L., Fortin, N. J., & Eichenbaum, H. (2002). The hippocam-
permit establishing more biologically plausible models pus and disambiguation of overlapping sequences. Journal of Neu-
roscience, 22, 57605768.
of behavioral development (connectionist models, in
Aizawa, H., Hu, S.-C., Bobb, K., Balakrishnan, K., Ince, G., Gurevich,
particular, should benefit from this advance). Second, I., et al. (2004). Dendrite development regulated by CREST, a cal-
as our knowledge of neural plasticity and molecular bi- cium-regulated transcriptional activator. Science, 303, 197202.
ology increases, we will do a better job of designing Alvarado, M. C., & Bachevalier, J. (2000). Revisiting the maturation
of medial temporal lobe memory functions in primates. Learning
studies that shed light on which behaviors are derived
and Memory, 7(5), 244256.
from experience-dependent versus experience-indepen- Alvarado, M. C., Wright, A. A., & Bachevalier, J. (2002). Object and
dent processes. This, in turn, should lead us away from spatial relational memory in adult rhesus monkeys is impaired by
strongly held nativist beliefs, at least in the context of neonatal lesions of the hippocampal formation but not the amyg-
daloid complex. Hippocampus, 12, 421433.
higher-level cognitive functions. Third, we anticipate a
Anderson, V., Anderson, P., Northam, E., Jacobs, R., & Catroppa, C.
judicious increase in the study of clinical populations, (2001). Development of executive functions through late child-
as such study has the potential to provide converging in- hood and adolescence in an Australian sample. Developmental
formation on typical development. Fourth, given in- Neuropsychology, 20(1), 385406.
creased interest in using neuroimaging tools to study Anderson, V., Levin, H., & Jacobs, R. (2002). Executive functions
after frontal lobe injury: A developmental perspective. In D.
affective development, we anticipate increased interest Stuss & R. Knight (Eds.), Principles of frontal lobe function
in linking cognitive and emotional development in the (pp. 504527). New York: Oxford University Press.
context of brain development. Fifth, increased attention Ark, W. S. (2002, January). Neuroimaging studies give new insight to
mental rotation. Proceedings of the 35th Hawaii International
will be paid to the coregistration of imaging modalities,
Conference on System Sciences, 5, 136, Hawaii.
particularly ERPs with fMRI and optical imaging. Astur, R. S., Taylor, L. B., Mamelak, A. N., Philpott, L., & Suther-
Sixth, as investigators gain more experience in conduct- land, R. J. (2002). Humans with hippocampus damage display se-
ing fMRI studies with children, and as physicists and vere spatial memory impairments in a virtual Morris water task.
Behavioural Brain Research, 132, 7784.
engineers improve MR scanning parameters, the field
Baddeley, A. (1986). Working memory. Oxford, England: Clarendon Press.
should experience a downward shift in the age at which
Baddeley, A., & Hitch, G. J. (1974). Working memory. In G. A.
such studies can be performed. Scanning infants will al- Bower (Ed.), Recent advances in learning and motivation (Vol. 7,
ways prove difficult, but scanning preschool-age chil- pp. 4790). New York: Academic Press.
dren may prove less challenging. Baddeley, A., & Logie, R. H. (1999). Working memory: The multiple-
component model. In A. Miyake & P. Shah (Eds.), Models of work-
These are just a few of the areas of growth we antici-
ing memory (pp. 2861). New York: Cambridge University Press.
pate in the coming decadeothers will surface as the Baddeley, A., Vargha-Khadem, F., Mishkin, M. (2001). Preserved
field evolves. But, we remain optimistic that interest in recognition in a case of developmental amnesia: Implications for
linking brain with behavior in the context of development the acquisition of semantic memory? Journal of Cognitive Neuro-
science, 13(3), 357369.
is now firmly entrenched in developmental psychology.
Baillargeon, R. (1987). Object permanence in 31 2 - and 41 2 -month-old
infants. Developmental Psychology, 23, 655664.
REFERENCES Baillargeon, R., Spelke, E., & Wasserman, S. (1985). Object perma-
nence in 5-month-old infants. Cognition, 20, 191208.
Abrahams, S., Pickering, A., Polkey, C. E., & Morris, R. G. (1997). Balaban, M. T. (1995). Affective inf luences on startle in 5-month-old
Spatial memory deficits in patients with unilateral damage to the infants: Reactions to facial expressions of emotion. Child Devel-
right hippocampal formation. Neuropsychologia, 35, 1124. opment, 66, 2836.
Adlam, A., Vargha-Khadem, F., Mishkin, M., & de Haan, M. (2004). Banks, M. S., & Ginsburg, A. P. (1985). Infant visual preferences: A
Deferred imitation in developmental amnesia. Manuscript submit- review and new theoretical treatment. Advances in Child Develop-
ted for publication. ment and Behavior, 19, 207246.
Adolphs, R. (2002). Recognizing emotion from facial expressions: Banks, M. S., & Salapatek, P. (1981). Infant pattern vision: A new
Psychological and neurological mechanisms. Behavioural and approach based on the contrast sensitivity function. Journal of Ex-
Cognitive Neuroscience Reviews, 1, 2161. perimental Psychology, 31, 145.
Adolphs, R., Tranel, D., Damasio, H., & Damasio, A. (1994). Im- Barinaga, M. (2003). Newborn neurons search for meaning. Science,
paired recognition of emotion in facial expressions following bi- 299, 3234.
lateral damage to the human amygdala. Nature, 372, 669672. Barrash, J., Tranel, D., & Anderson, S. (1994). Assessment of dra-
Adolphs, R., Tranel, D., Hamann, S., Young, A. W., Calder, A. J., matic personality changes after ventromedial frontal lesions. Jour-
Phelps, E. A., et al. (1999). Recognition of facial emotion in nine nal of Clinical and Experimental Neuropsychology, 18, 355381.
individuals with bilateral amygdala damage. Neuropsychologia, Barton, J. J. (2003). Disorders of face perception and recognition.
37, 11111117. Neurologic Clinics, 21(2), 521548.
46 Neural Bases of Cognitive Development

Bauer, P. J., Wenner, J. A., Dropik, P. L., & Wewerka, S. S. (2000). Buckner, R., Goodman, J., Burock, M., Rotte, M., Koutstaal, W.,
Parameters of remembering and forgetting in the transition from Schacter, D., et al. (1998). Functional-anatomic correlates of ob-
infancy to early childhood. Monographs of the Society for Re- ject priming in humans revealed by rapid event-related fMRI.
search in Child Development, 65, 1204. Neuron, 20, 285296.
Bauer, P. J., Wiebe, S. A., Carver, L. J., Waters, J. M., & Nelson, C. A. Bunge, S., Dudukovic, N., Thomason, M., Vaidya, C., & Gabrieli, J.
(2003). Electrophysiological indices of long-term recognition pre- (2002). Immature frontal lobe contributions to cognitive control
dict infants long-term recall. Psychological Science, 14, 629635. in children: Evidence from fMRI. Neuron, 33(2), 301311.
Benes, F., Turtle, M., Khan, Y., & Farol, P. (1994). Myelination of a Burrone, J., OByrne, M., & Murthy, V. N. (2002). Multiple forms of
key relay zone in the hippocampal formation occurs in the human synaptic plasticity triggered by selective suppression of activity
brain during childhood, adolescence, and adulthood. Archives of in individual neurons. Nature, 420, 414418.
General Psychiatry, 51, 477484. Bush, G., Frazier, J., Rauch, S., Seidman, L., Whalen, P., Jenike, M.,
Bernier, P. J., Bdard, A., Vinet, J., Lvesque, M., & Parent, A. et al. (1999). Anterior cingulate cortex dysfunction in attention-
(2002). Newly generated neurons in the amygdala and adjoining deficit / hyperactivity disorder revealed by fMRI and the counting
cortex of adult primates. Proceedings of the National Academy of stroop. Biological Psychiatry, 45(12), 15421552.
Sciences, 99, 1146411469. Bush, G., Luu, P., & Posner, M. (2000). Cognitive and emotional in-
Bischoff-Grethe, A., Martin, M., Mao, H., & Berns, G. (2001). The f luences in anterior cingulate cortex. Trends in Cognitive Sci-
context of uncertainty modulates the subcortical response to pre- ences, 4, 215222.
dictability. Journal of Cognitive Neuroscience, 13(7), 986993. Butters, M. A., Kaszniak, A. W., Glisky, E. L., Eslinger, P. J., &
Bixby, J. L., & Harris, W. A. (1991). Molecular mechanisms of axon Schacter, D. L. (1994). Recency discrimination deficits in frontal
growth and guidance. Annual Review of Cell Biology, 7, 117159. lobe patients. Neuropsychologia, 8, 343353.
Black, J. E., Jones, T. A., Nelson, C. A., & Greenough, W. T. (1998). Calder, A. J., Young, A. W., Rowland, D., Perett, D. I., Hodges, J. R.,
Neuronal plasticity and the developing brain. In N. E. Alessi, J. T. & Etcoff, N. L. (1996). Facial emotion recognition after bilateral
Coyle, S. I. Harrison, & S. Eth (Eds.), Handbook of child and ado- amygdala damage: Differentially severe impairment of fear. Cog-
lescent psychiatry: Vol. 6. Basic psychiatric science and treatment nitive Neuropsychology, 13, 699745.
(pp. 3153). New York: Wiley. Canfield, R., Smith, E., Brezsnyak, M., & Snow, K. (1997). Informa-
Booth, J. R., MacWhinney, B., Thulborn, K. R., Sacco, K., Voyvodic, tion processing through the first year of life: A longitudinal study
J., & Feldman, H. M. (1999). Functional organization of activa- using the visual expectancy paradigm. Monographs of the Society
tion patterns in children: Whole brain fMRI imaging during thee for Research in Child Development, 62(2).
different cognitive tasks. Progress in Neuro-Psychopharmacology Carey, S., & Diamond, R. (1994). Are faces perceived as configura-
and Biological Psychiatry, 23, 669682. tions more by adults than children? Visual Cognition, 1, 253274.
Booth, J. R., MacWhinney, B., Thulborn, K. R., Sacco, K., Voyvodic, Carleton, A., Petreanu, L. T., Lansford, R., Alvarez-Buylla, A., &
J., & Feldman, H. M. (2000). Developmental and lesion effects in Lledo, P.-M. (2003). Become a new neuron in the adult olfactory
brain activation during sentence comprehension and mental rota- bulb. Nature Neuroscience, 6, 507518.
tion. Developmental Neuropsychology, 18, 139169. Carver, L. J., Bauer, P. J., & Nelson, C. A. (2000). Associations be-
Botvinick, M., Braver, T., Barch, D., Carter, C., & Cohen, J. (2001). tween infant brain activity and recall memory. Developmental Sci-
Conf lict monitoring and cognitive control. Psychological Review, ence, 3, 234246.
108(3), 624652. Casey, B., Cohen, J., Jezzard, P., Turner, R., Noll, D., Trainor, R.,
Bourgeois, J. P., & Rakic P. (1993). Changes of synaptic density in et al. (1995). Activation of prefrontal cortex in children during a
the primary visual cortex of the macaque monkey from fetal to nonspatial working memory task with functional MRI. Neuroim-
adult stage. Journal of Neuroscience, 13, 28012820. age, 2, 221229.
Bourgeois, J.-P., Reboff, P. J., & Rakic, P. (1989). Synaptogenesis in Casey, B., Durston, S., & Fossella, J. (2001). Evidence for a mecha-
visual cortex of normal and preterm monkeys: Evidence from in- nistic model of cognitive control. Clinical Neuroscience Research,
trinsic regulation of synaptic overproduction. Proceedings of the 1, 267282.
National Academy of Sciences, 86, 42974301. Casey, B., Forman, S., Franzen, P., Berkowitz, A., Braver, T., Nys-
Brizzolara, D., Casalini, C., Montanaro, D., & Posteraro F. (2003). A trom, L., et al. (2001). Sensitivity of prefrontal cortex to changes
case of amnesia at an early age. Cortex, 39(4/5), 605625. in target probability: A functional MRI study. Human Brain Map-
Broks, P., Young, A. W., Maratos, E. J., Coffey, P. J., Calder, A. J., ping, 13, 2633.
Isaac, C. L., et al. (1998). Face processing impairments after en- Casey, B., & Richards, J. (1988). Sustained visual attention in young
cephalitis: Amygdala damage and recognition of fear. Neuropsy- infants measured by with an adapted version of the visual prefer-
chologia, 36(1), 5970. ence paradigm. Child Development, 59, 15151521.
Bronner-Fraser, M., & Hatten, M. B. (2003). Neurogenesis and mi- Casey, B., Thomas, K., Welsh, T., Badgaiyan, R., Eccard, C., Jen-
gration. In L. R. Squire, F. E. Bloom, S. K. McConnell, J. L. nings, J., et al. (2000). Dissociation of response conf lict, atten-
Roberts, N. C. Spitzer, & M. J. Zigmond (Eds.), Fundamental neu- tional control, and expectancy with functional magnetic
roscience (2nd ed., pp. 391416). New York: Academic Press. resonance imaging (fMRI ). Proceedings of the National Academy
Brown, M., Keynes, R., & Lumsden, A. (2001). The developing brain. of Sciences, USA, 97(15), 87288733.
Oxford, England: Oxford University Press. Casey, B. J., Tottenham, N., & Fossella, J. (2002). Clinical, imaging,
Brown, S. A., Warburton, D., Brown, L. Y., Yu, C. Y., Roeder, E. R., lesion, and genetic approaches toward a model of cognitive con-
Shengel-Rutkowski, S., et al. (1998). Hologrosencephaly due to trol. Developmental Psychobiology, 40(3), 237254.
mutations in ZIC2, a homologue of Drosophila odd-paired. Nature Casey, B., Trainor, R., Orendi, J., Schubert, A., Nystrom, L., Cohen,
Genetics, 20, 180193. J., et al. (1997). A pediatric functional MRI study of prefrontal
References 47

activation during performance of a go-no-go task. Journal of Cog- Colcombe, S. J., Kramer, A. F., Erikson, K. I., Scalf, P., McAuley, E.,
nitive Neuroscience, 9, 835847. Cohen, N. J., et al. (2004). Cardiovascular fitness, cortical plas-
Castellanos, F., Giedd, J., Berquin, P., Walter, J., Sharp, W., Tran, T., ticity, and aging. Proceedings of the National Academy of Science,
et al. (2001). Quantitative brain magnetic resonance imaging in 101, 33163321.
girls with attention-deficit / hyperactivity disorder. Archives of Collie, R., Ettayne, H. (1999). Deferred imitation by 6- and 9-month-
General Psychiatry, 58, 289295. old infants: More evidence for declarative memory. Developmen-
Castellanos, F., Lee, P., Sharp, W., Jeffries, N., Greenstein, D., tal Psychobiology, 35, 8390.
Clasen, L., et al. (2002). Developmental trajectories of brain vol- Condron, B. (2002). Gene expression is required for correct axon
ume abnormalities in children and adolescents with attention- guidance. Current Biology, 12(19), 16651669.
deficit / hyperactivity disorder. Journal of the American Medical Cools, R., Clark, L., Owen, A., & Robbins, T. (2002). Defining the
Association, 288, 17401748. neural mechanisms of probabilistic reversal learning using event-
Chang, K. D., Adleman, N., Dienes, K., Menon, V., & Reiss, A. (2002, related functional magnetic resonance imaging. Journal of Neuro-
October). FMRI of visuospatial working memory in boys with bipo- science, 22(11), 45634567.
lar disorder. Poster session presented at the 49th annual meeting
Cools, R., Clark, L., & Robbins, T. (2004). Differential responses in
of the American Academy of Child and Adolescent Psychiatry,
human striatum and prefrontal cortex to changes in object and
San Francisco, CA.
rule relevance. Journal of Neuroscience, 24(5), 11291135.
Changeux, J.-P., & Danchin, A. (1976). Selective stabilization of de-
Corbetta, M., Kincade, J., Ollinger, J., McAvoy, M., & Shulman,
veloping synapses as a mechanism for the specification of neu-
G. (2000). Voluntary orienting is dissociated from target detec-
ronal networks. Nature, 64(5588), 705712.
tion in human posterior parietal cortex. Nature Neuroscience,
Chechik, G., Meilijson, I., & Ruppin, E. (1999). Neuronal regulation: 3(3), 292297.
A mechanism for synaptic pruning during brain maturation.
Corkin, S. (2002). Whats new with amnesic patient H. M.? Nature
Neural Computation, 11(8), 20612080.
Reviews Neuroscience, 3, 153160.
Chenn, A., & McConnell, S. K. (1995). Cleavage orientation and the
Cornoldi, C., Marconi, F., & Vecchi. T. (2001). Visuospatial working
asymmetric inheritance of Notch1 immunoreactivity in mam-
malian neurogenesis. Cell, 82, 631641. memory in Turners syndrome. Brain and Cognition, 46, 9094.

Cheng, E. F., & Merzenich, M. M. (2003). Environmental noise re- Courchesne, E. (1978). Neurophysiological correlates of cognitive de-
tards auditory cortical development. Science, 300, 498502. velopment: Changes in long-latency event-related potentials from
childhood to adulthood. Electroencephalography and Clinical Neu-
Chugani, H. T., & Phelps, M. E. (1986). Maturational changes in
rophysiology, 45, 468482.
cerebral function in infants determined by 18FDG positron emis-
sion tomography. Science, 231, 840843. Cycowicz, Y. M. (2000). Memory development and event-related
brain potentials in children. Biological Psychology, 54, 145174.
Chugani, H. T., Phelps, M., & Mazziotta, J. (1987). Positron emission
tomography study of human brain functional development. Annals Damasio, H., Grabowski, T., Frank, R., Galaburda, A., & Damasio,
of Neurology, 22(4), 487497. A. (1994). The return of Phineas Gage: Clues about the brain from
the skull of a famous patient. Science, 264, 11021105.
Churchill, J. D., Beckel-Mitchener, A., Weiler, I. J., & Greenough,
W. T. (2002). Effects of fragile X syndrome and an FMR1 knock- Danziger, S., Ward, R., Owen, V., & Rafal, R. (2001). The effects
out mouse model on forebrain neuronal cell biology. Microscopy of unilateral pulvinar damage in humans on ref lexive orienting
Research and Techniques, 57, 156158. and filtering of irrelevant information. Behavioral Neurology,
Churchill, J. D., Grossman, A. W., Irwin, S. A., Galvez, R., 13(3/4), 95104.
Klintsova, A. Y., Weiler, I. J., et al. (2002). A converging-methods Davis, M. (1989). The role of the amygdala and its efferent projec-
approach to fragile X syndrome. Developmental Psychobiology, tions in fear and anxiety. In P. Tyrer (Ed.), Psychopharmacology
40, 323328. of anxiety (pp. 5279). Oxford, England: Oxford University Press.
Cline, H. (2003). Sperry and Hebb: Oil and vinegar? Trends in Neuro- Dean, A. L., Duhe, D. A., & Green, D. A. (1983). The development of
science, 26, 655661. childrens mental tracking strategies on a rotation task. Journal of
Clohessy, A., Posner, M., & Rothbart, M. (2002). Development of the Experimental Child Psychology, 36, 226240.
functional visual field. Acta Psychologia, 106(1/2), 5168. DeBoer, T., Georgieff, M. K., & Nelson, C. A. (in press). Elicited im-
Clohessy, A., Posner, M., Rothbart, M., & Vecera, S. (1991). The de- itation: A tool to investigate the impact of abnormal prenatal envi-
velopment of inhibition of return in early infancy. Journal of Cog- ronments on memory development. In P. Bauer (Ed.), Varieties of
nitive Neuroscience, 3(4), 345350. early experience: Inf luences on declarative memory development.
Clower, W. T., & Alexander, G. E. (1998). Movement sequence-re- Hillsdale, NJ: Erlbaum.
lated activity ref lecting numerical order of components in sup- DeBoer, T., Scott, L., & Nelson, C. (2004). Event-related potentials
plementary and presupplementary motor areas. Journal of in developmental populations. In T. Handy (Ed.), Event-related
Neurophysiology, 80, 15621566. potentials: A methods handbook (pp. 263297). Cambridge,
Cohen, J., Forman, S., Braver, T., Casey, B., Servan-Schreiber, D., & MA: MIT Press.
Noll, D. (1994). Activation of prefrontal cortex in a non-spatial de Castro F. (2003). Chemotropic molecules: Guides for axonal
working memory task with functional MRI. Human Brain Map- pathfinding and cell migration during CNS development. News
ping, 1, 293304. Physiological Sciences, 18, 130136.
Cohen, M. S., Kosslyn, S. M., Breiter, H. C., DiGirolamo, G. J., de Haan, M., Bauer, P. J., Georgieff, M. K., & Nelson, C. A. (2000).
Thompson, W. L., Anderson, A. K., et al. (1996). Changes in cor- Explicit memory in low-risk infants aged 19 months born between
tical activity during mental rotation. A mapping study using func- 27 and 42 weeks of gestation. Developmental Medicine and Child
tional MRI. Brain, 119, 89100. Neurology, 42, 304312.
48 Neural Bases of Cognitive Development

de Haan, M., & Johnson, M. H. (2003a). The cognitive neuroscience of dependence on dorsolateral prefrontal cortex. Experimental Brain
development. London: Psychology Press. Research, 74, 2440.
de Haan, M., & Johnson, M. H. (2003b). Mechanisms and theories of Diamond, A., Kirkham, N., & Amso, D. (2002). Conditions under
brain development. In M. de Haan & M. H. Johnson (Eds.), The which young children can hold two rules in mind and inhibit a
cognitive neuroscience of development (pp. 118). London: Psy- prepotent response. Developmental Psychology, 38, 352363.
chology Press. Diamond, R., & Carey, S. (1986). Why faces are and are not special:
de Haan, M., & Nelson, C. A. (1997). Recognition of the mothers An effect of expertise. Journal of Experimental Psychology: Gen-
face by 6-month-old infants: A neurobehavioral study. Child De- eral, 115, 107117.
velopment, 68, 187210. Dolan, R. J., & Fletcher, P. C. (1997). Dissociating prefrontal and
de Haan, M., & Nelson, C. A. (1999). Brain activity differentiates hippocampal function in episodic memory encoding. Nature,
face and object processing in 6-month-old infants. Developmental 388(6642), 582585.
Psychology, 35, 11131121. Draganski, B., Gaser, C., Busch, V., Schuierer, G., Bogdahn, U., &
de Haan, M., Pascalis, O., & Johnson, M. H. (2002). Specialization of May, A. (2004). Changes in grey matter induced by training. Na-
neural mechanisms underlying face recognition in human infants. ture, 427, 311312.
Journal of Cognitive Neuroscience, 14(2), 199209. Drummey, A., & Newcombe, N. (1995). Remembering versus know-
DeHaene, S., Dupoux, E., Mehler, J., Cohen, L., Paulesu, E., Perani, ing the past: Childrens explicit and implicit memories for pic-
D., et al. (1997). Anatomical variability in the cortical represen- tures. Journal of Experimental Child Psychology, 59, 549565.
tation of first and second language. NeuroReport, 8, 38093815. Duncan, J., & Owen, A. (2000). Common regions of the human
Delis, D. C., Roberston, L. C., & Efron, R. (1986). Hemispheric spe- frontal lobe recruited by diverse cognitive demands. Trends in
cialization of memory for visual hierarchical stimuli. Neuropsy- Neurosciences, 23, 475483.
chologia, 24, 205214. Durston, S., Hulshoff Pol, H. E., Casey, B. J., Giedd, J. N., Buitelaar,
DeRegnier, R.-A., Nelson, C. A., Thomas, K., Wewerka, S., & J. K., & van Engeland, H. (2001). Anatomical MRI of the devel-
Georgieff, M. K. (2000). Neurophysiologic evaluation of auditory oping human brain: What have we learned? Journal of the Ameri-
recognition memory in healthy newborn infants and infants of di- can Academy of Child and Adolescent Psychiatry, 40, 10121020.
abetic mothers. Journal of Pediatrics, 137, 777784. Durston, S., Thomas, K., & Yang, Y. (2002). The development of neu-
Deruelle, C., & de Schonen, S. (1991). Hemispheric asymmetries ral systems involved in overriding behavioral responses: An event-
in visual pattern processing in infancy. Brain and Cognition, related fMRI study. Developmental Science, 5, 916.
16, 151179. Durston, S., Tottenham, N., Thomas, K., Davidson, M., Eigsti, I.,
Deruelle, C., & de Schonen, S. (1995). Pattern processing in in- Yang, Y., et al. (2003). Differential patterns of striatal activation
fancy: Hemispheric differences in the processing of shape and in young children with and without ADHD. Biological Psychiatry,
location visual components. Infant Behavior and Development, 53(10), 871878.
18, 123132. Eichenbaum, H. (2002). The cognitive neuroscience of memory. Lon-
Deruelle, C., Mancini, J., Livet, M. O., Casse-Perrot, C., & don: Oxford University Press.
de Schonen, S. (1999). Configural and local processing of faces Eichenbaum, H. (2003). Learning and memory: Brain systems. In
in children with Williams syndrome. Brain and Cognition, 41, L. R. Squire, F. E. Bloom, S. K. McConnell, J. L. Roberts, N. C.
276298. Spitzer, & M. J. Zigmond (Eds.), Fundamental neuroscience (2nd
de Schonen, S., & Mathivet, E. (1990). Hemispheric asymmetry in a ed., pp. 12991327). New York: Academic Press.
face discrimination task in infants. Child Development, 61, Eichenbaum, H. B., Cahill, L. F., Gluck, M. A., Hasselmo, M. E.,
11921205. Keil, F. C., Martin, A. J., et al. (1999). Learning and memory:
Desimone, R., & Duncan, J. (1995). Neural mechanisms of selective Systems analysis. In M. J. Zigmond, F. E. Bloom, S. C. Landis,
visual attention. Annual Review of Neuroscience, 18, 193222. J. L. Roberts, & L. R. Squire (Eds.), Fundamental neuroscience
Diamond, A. (1985). Development of the ability to use recall to guide (pp. 14551486). New York: Academic Press.
action, as indicated by infants performance on A-not-B. Child Eimer, M. (2000). The face-specific N170 component reflects late stages
Development, 56, 868883. in the structural encoding of faces. NeuroReport, 11, 23192324.
Diamond, A. (1995). Evidence of robust recognition memory early in Ekdahl, C. T., Claasen, J.-H., Bonde, S., Kokaia, Z., & Lindvall, O.
life even when assessed by reaching behavior. Journal of Experi- (2003). Inf lamation is detrimental for neurogenesis in adult brain.
mental Child Psychology, 59, 419456. Proceedings of the National Academy of Sciences, 100,
Diamond, A. (2001). A model system for studying the role of 1362213637.
dopamine in the prefrontal cortex during early development in Elbert, T., Pantev, C., Wienbruch, C., Rockstroh, B., & Taub, E.
humans: Early and continuously treated phenylketonuria. In C. (1995). Increased cortical representation of the fingers of the left
Nelson & M. Luciana (Eds.), Handbook of developmental cogni- hand in string players. Science, 270, 305307.
tive neuroscience (pp. 433472). Cambridge, MA: MIT Press. Elvevg, B., & Weinberger, D. R. (2001). The neuropsychology of
Diamond, A. (2002). Normal development of prefrontal cortex from schizophrenia and its relationship to the neurodevelopmental
birth to young adulthood: Cognitive functions, anatomy, and model. In C. A. Nelson & M. Luciana (Eds.), Handbook of de-
biochemistry. In D. Stuss & R. Knight (Eds.), Principles of velopmental cognitive neuroscience (pp. 577594). Cambridge,
frontal lobe function (pp. 466503). New York: Oxford Univer- MA: MIT Press.
sity Press. Erickson, C. A., Jagadeesh, B., & Desimone, R. (2000). Clustering of
Diamond, A., & Goldman-Rakic, P. (1989). Comparison of human in- perirhinal neurons with similar properties following visual expe-
fants and rhesus monkeys on Piagets A-not-B task: Evidence for rience in adult monkeys. Nature Neuroscience, 3(11), 11431148.
References 49

Erskine, L., Williams, S. E., Brose, K., Kidd, T., Rachel, R. A., creases with expertise in recognizing novel objects. Nature Neu-
Goodman, C. S., et al. (2000). Retinal ganglion cell axon guid- roscience, 2, 568573.
ance in the mouse optic chiasm: Expression and function of robos Geldart, S., Mondloch, C. J., Maurer, D., de Schonen, S., & Brent,
and slits. Journal of Neuroscience, 20(13), 49754978. H. P. (2002). The effect of early visual deprivation on the devel-
Eslinger, P. J., & Biddle, K. R. (2000). Adolescent neuropsychological opment of face processing. Developmental Science, 5(4), 490501.
development after early right prefrontal cortex damage. Develop- Geraldi-Caulton, G. (2000). Sensitivity to spatial conf lict and the de-
mental Neuropsychology, 18, 297329. velopment of self-regulation in children 24 to 30 months of age.
Fan, J., Flombaum, J., McCandliss, B., Thomas, K., & Posner, M. Developmental Science, 3, 397404.
(2003). Cognitive and brain consequences of conf lict. NeuroIm- Giedd, J. N., Blumenthal, J., Jeffries, N. O., Castellanos, F. X., Liu,
age, 18(1), 4257. H., Zijdenbos, A., et al. (1999). Brain development during child-
Farah, M. J., Rabinowitz, C., Quinn, G. E., & Liu, G. T. (2000). Early hood and adolescence: A longitudinal MRI study. Nature Neuro-
commitment of neural substrates for face recognition. Cognitive science, 2, 861863.
Neuropsychology, 17, 117124. Giedd, J. N., Snell, J., Lange, N., Rajapakse, J., Casey, B., Kozuch, P.,
Ferguson, J. N., Aldag, J. M., Insel, T. R., & Young, L. J. (2001). Oxy- et al. (1996). Quantitative magnetic resonance imaging of human
tocin in the medial amygdala is essential for social recognition in brain development: Ages 4 to 18. Cerebral Cortex, 6, 551560.
the mouse. Journal of Neuroscience, 21, 82788285. Giedd, J. N., Vatuzis, A. C., Hamburger, S. D., Lange, N., Ra-
Ferguson, J. N., Young, L. J., Hearn, E. F., Matzuk, M. M., Insel, japakse, J. C., Matsen, D., et al. (1996). Quantitative MRI of the
T. R., Ervinslow, J. T. (2000). Social amnesia in mice lacking the temporal lobe, amygdala, and hippocampus in normal human de-
oxytocin gene. Nature Genetics, 25, 284288. velopment: Ages 4 to 18 years. Journal of Comparative Neurol-
ogy, 366, 223230.
Ferraro, F., Balota, D., & Connor, L. (1993). Implicit memory and the
formation of new associations in nondemented Parkinsons dis- Goldman, P., & Rosvold, H. (1970). Localization of function within
ease individuals and individuals with senile dementia of the the dorsolateral prefrontal cortex of the rhesus monkey. Experi-
Alzheimer type: A serial reaction time (SRT) investigation. Brain mental Neurology, 29, 291304.
and Cognition, 21, 163180. Goldman-Rakic, P. S. (1987). Development of cortical circuitry and
Finney, E. M., Fine, I., & Dobkins, K. R. (2001). Visual stimuli activate cognitive function. Child Development, 58(3), 601622.
auditory cortex in the deaf. Nature Neuroscience, 4, 11711173. Gould, E. (2003, July). Neurogenesis in the adult brain. Paper pre-
Flavell, J. H., Miller, P. H., & Miller, S. A. (1993). Cognitive develop- sented at the Merck Summer Institute on Developmental Disabili-
ment (3rd ed.). Englewood Cliffs, NJ: Prentice-Hall. ties, Princeton University, Princeton, NJ.
Fortin, N. J., Agster, K. L., & Eichenbaum, H. B. (2002). Critical role Gould, E., Beylin, A., Tanapat, P., Reeves, A., & Shors, T. J. (1999).
of the hippocampus in memory for sequences of events. Nature Learning enhances adult neurogenesis in the hippocampal forma-
Neuroscience, 5, 458462. tion. Nature Neuroscience, 2, 260265.
Francis, D. D., Szegda, K., Campbell, G., Martin, W. D., & Insel, Gould, E., & Gross, C. G. (2002). Neurogenesis in adult mammals:
T. R. (2003). Epigenetic sources of behavioral differences in Some progress and problems. Journal of Neuroscience, 22, 619623.
mice. Nature Neuroscience, 6, 445448. Gould, E., Vail, N., Wagers, M., & Gross, C. G. (2001). Adult-gener-
Frangiskakis, J. M., Ewart, A. K., Morris, C. A., Mervis, C. B., ated hippocampal and neocortical neurons in macaques have a
Bertrand, J., Robinson, B. F., et al. (1996). LIM-kinase1 hemizy- transient existence. Proceedings of the National Academy of Sci-
gosity implicated in impaired visuospatial constructive cognition. ences, USA, 98, 1091010917.
Cell, 86, 5969. Grafton, S., Hazeltine, E., & Ivry, R. (1995). Functional mapping of
Funahashi, S., Bruce, C., & Goldman-Rakic, P. (1989). Mnemonic sequence learning in normal humans. Journal of Cognitive Neuro-
coding of visual space in the monkeys dorsolateral prefrontal science, 7(4), 497510.
cortex. Journal of Neurophysiology, 61, 119. Greenough, W. T., & Black, J. E. (1992). Induction of brain structure
Fuster, J. (1997). The prefrontal cortex: Anatomy, physiology, and by experience: Substrates for cognitive development. In M. R.
neuropsychology of the frontal lobe (3rd ed.). Philadelphia: Lip- Gunnar & C. A. Nelson (Eds.), Minnesota Symposia on Child Psy-
pencott-Raven Press. chology: Vol. 24. Developmental behavioral neuroscience
(pp. 155200). Hillsdale, NJ: Erlbaum.
Fuster, J., & Alexander, G. (1970). Delayed response deficit by cryo-
Greenough, W. T., Juraska, J. M., & Volkmar, F. R. (1979). Maze
genic depression of frontal cortex. Brain Research, 61, 7991.
training effects on dendritic branching in occipital cortex of adult
Fuster, J., & Alexander, G. (1971). Neuron activity related to short- rats. Behavioral and Neural Biology, 26, 287297.
term memory. Science, 173, 652654.
Greenough, W. T., Madden, T. C., & Fleishchmann, T. B. (1972). Ef-
Gadian, D. G., Aicardi, J., Watkins, K. E., Porter, D. A., Mishkin, M., fects of isolation, handling, and enriched rearing on maze learn-
& Vargha-Khadem, F. (2000). Developmental amnesia associated ing. Psychonomic Science, 27, 279280.
with early hypoxic-ischaemic injury. Brain, 123, 499507.
Grice, S. J., Spratling, M. W., Karmiloff-Smith, A., Halit, H., Csibra,
Gage, F. H. (2000). Mammalian neural stem cells. Science, G., de Haan, M., et al. (2001). Disordered visual processing and
287(5457), 14331438. oscillatory brain activity in autism and Williams syndrome. Neu-
Gauthier, I., Skudlarski, P., Gore, J. C., & Anderson, A. W. (2000). roReport, 12, 26972700.
Expertise for cars and birds recruits brain areas involved in face Grossman, A. W., Churchill, J. D., Bates, K. E., Kleim, J., & Gree-
recognition. Nature Neuroscience, 3, 191197. nough, W. T. (2002). A brain adaptation view of plasticity: Is synap-
Gauthier, I., Tarr, M. J., Anderson, A. W., Skudlarski, P., & Gore, tic plasticity an overly limited concept. In M. A. Hofman, G. J.
J. C. (1999). Activation of the middle fusiform face area in- Boer, A. J. G. D. Holtmaat, E. J. W. Van Someren, J. Verhaagen, &
50 Neural Bases of Cognitive Development

D. F. Swaab (Eds.), Progress in brain research (Vol. 138, Hoffman, E., & Haxby, J. V. (2000). Distinct representations of eye
pp. 91108). New York: Elsevier Science. gaze and identity in the distributed human neural system for face
Haberecht, M. F., Menon, V., Warsofsky, I. S., White, C. D., Dyer- perception. Nature Neuroscience, 3, 8084.
Friedman, J., Glover, G. H., et al. (2001). Functional neu- Holdstock, J. S., Mayes, A. R., Cezayirli, E., Isaac, C. L., Aggleton,
roanatomy of visuo-spatial working memory in Turner syndrome. J. P., & Roberts, N. (2000). A comparison of egocentric and allo-
Human Brain Mapping, 14, 96107. centric spatial memory in a patient with selective hippocampal
Haith, M., Hazan, C., & Goodman, G. (1988). Expectation and antic- damage. Neuropsychologia, 38, 410425.
ipation of dynamic visual events by 31 2 -month-old babies. Child Holstege, G., van Ham, J. J., & Tan, J. (1986). Afferent projections to
Development, 59, 467479. the orbicularis oculi motoneural cell group: An autoradiographi-
Halit, H., de Haan, M., & Johnson, M. H. (2003). Cortical specialisa- cal tracing study in the cat. Brain Research, 374, 306320.
tion for face processing: Face sensitive-event related potential Hood, B., Atkinson, J., & Braddick, O. (1998). Selection-for-action
components in 3- and 12-month-old infants. Neuroimage, 19, and the development of orienting and visual attention. In J.
11801193. Richards (Ed.), Cognitive neuroscience of attention: A develop-
Hamilton, R., Keenan, J. P., Catala, M., & Pascual-Leone, A. (2000). mental perspective (pp. 219250). Mahwah, NJ: Erlbaum.
Alexia for Braille following a bilateral occipital stroke in an early Huang, E. J., & Reichardt, L. F. (2001). Neurotrophins: Roles in neu-
blind woman. NeuroReport, 11, 237240. ronal development and function. Annual Review of Neuroscience,
Hanashima, C., Li, S. C., Shen, L., Lai, E., & Fishell, G. (2004). 24, 677736.
Foxg1 suppresses early cortical cell fate. Science, 303, 5659. Hubl, D., Bolte, S., Feineis-Matthews, S., Lanfermann, H., Feder-
Harris, I. M., & Miniussi, C. (2003). Parietal lobe contribution to spiel, A., Strik, W., et al. (2003). Functional imbalance of visual
mental rotation demonstrated with rTMS. Journal of Cognitive pathways indicates alternative face processing strategies in
Neuroscience, 15, 315323. autism. Neurology, 61, 12321237.
Hartshorn, K., Rovee-Collier, C., Gerhardstein, P., Bhatt, R., Klein, Huttenlocher, P. R. (2002). Neural plasticity: The ef fects of environ-
P., & Aaron, F. (1998). Developmental changes in the specificity ment on the development of the cerebral cortex. Cambridge, MA:
of memory over the first year of life. Developmental Psychobiol- Harvard University Press.
ogy, 33, 6178. Huttenlocher, P. R., & Dabholkar, A. S. (1997). Regional differences
Hatten, M. E. (2002). New directions in neuronal migration. Science, in synaptogenesis in human cerebral cortex. Journal of Compara-
297, 16601663. tive Neurology, 387(2), 167178.
Haxby, J. V., Gobbini, M. I., Furey, M. L., Ishai, A., Schouten, J. L., Huttenlocher, P. R., & de Courten, C. (1987). The development of
& Pietrini, P. (2001). Distributed and overlapping representa- synapses in striat cortex of man. Human Neurobiology, 6, 19.
tions of faces and objects in ventral temporal cortex. Science, Hynes, R. O., & Lander, A. D. (1992). Contact and adhesive speci-
293, 24252430. ficities in the associations, migrations, and targeting of cells and
Haxby, J. V., Hoffman, E. A., & Gobbini, M. I. (2002). Human neural axons. Cell, 68(2), 303322.
systems for face recognition and social communication. Biologi- Incisa della Rocchetta, A., Samson, S., Ehrle, N., Denos, M., Has-
cal Psychiatry, 51, 5967. boun, D., & Baulac M. (2004). Memory for visuospatial location
Hayes, B., & Hennessy, R. (1996). The nature and development of following selective hippocampal sclerosis: The use of different
nonverbal implicit memory. Journal of Experimental Child Psy- coordinate systems. Neuropsychology, 18, 1528.
chology, 63, 2243. Isaacs, E. B., & Vargha-Khadem, F. (1989). Differential course of de-
Hayne, H. (2004). Infant memory development: Implications for velopment of spatial and verbal memory span: A normative study.
childhood amnesia. Developmental Review, 24, 3373. British Journal of Developmental Psychology, 7, 377380.
Hayne, H., Boniface, J., & Barr, R. (2000). The development of de- Isaacs, E. B., Vargha-Khadem, F., Watkins, K. E., Lucas, A.,
clarative memory in human infants: Age-related changes in de- Mishkin, M., & Gadian, D. G. (2003). Developmental amnesia
ferred imitation. Behavioral Neuroscience, 114(1), 7783. and its relationship to degree of hippocampal atrophy. Proceed-
Hazeltine, E., Grafton, S., & Ivry, R. (1997). Attention and stimulus ings of the National Academy of Sciences, USA, 100(22),
characteristics determined the locus of motor sequence encoding: 1306013063.
A PET study. Brain, 120, 123140. Ishai, A., Ungerleider, L. G., Martin, A., & Haxby, J. V. (2000). The
Heindel, W., Salmon, D., Shults, C., Walicke, P., & Butters, N. (1989). representation of objects in the human occipital and temporal cor-
Neuropsychological evidence for multiple implicit memory sys- tex. Journal of Cognitive Neuroscience, 12(Suppl. 2), 3551.
tems: A comparison of Alzheimers, Huntingtons, and Parkin- Itier, R. J., & Taylor, M. J. (2002). Inversion and contrast polarity re-
sons disease patients. Journal of Neuroscience, 9(2), 582587. versal affect both encoding and recognition processes of unfamiliar
Henson, R. N., Goshen-Gottstein, Y., Ganel, T., Otten, L. J., Quayle, faces: A repetition study using ERPs. NeuroImage, 15(2), 353372.
A., & Rugg, M. D. (2003). Electrophysiological and haemody- Janowsky, J. S., Shimamura, A. P., & Squire, L. R. (1989). Source
namic correlates of face perception, recognition and priming. memory impairment in patients with frontal lobe lesions. Neu-
Cerebral Cortex, 13, 795805. ropsychologia, 8, 10431056.
Hepworth, S., & Smith, M. (2002). Learning and recall of story con- Jernigan, T. L., Trauner, D. A., Hesselink, J. R., & Tallal, P. A.
tent and spatial location after unilateral temporal-lobe excision in (1991). Maturation of human cerebrum observed in vivo during
children and adolescents. Neuropsychology, Development, and adolescence. Brain, 114, 20372049.
Cognition. Section C. Child Neuropsychology, 8, 1626. Jessell, T. M. (1988). Adhesion molecules and the hierarchy of neural
Hof, F. A., Geurds, M. P., Chatkupt, S., Shugart, Y. Y., Balling, R., development. Neuron, 1(1), 313.
Schrander-Stumpel, C. T., et al. (1996). PAX genes and human Johnson, M. H. (1998). The neural basis of cognitive development. In
neural tube defects: An amino acide substitution in PAS1 in a pa- W. Damon, D. Kuhn, & R. S. Siegler (Editor-in-Chief ) & D. Kuhn
tient with spina bifida. Journal of Medical Genetics, 33, 655660. & R. S. Siegler (Vol. Eds.), Handbook of child psychology: Vol. 2.
References 51

Cognition, perception and language (5th ed., pp. 149). New York: Knudsen, E. (2003b). MacArthur Foundation research network on.
WileyPress. Early Experience and Brain Development. Annual report.
Johnson, M. H. (2001). Functional brain development in humans. Na- Knudsen, E. (in press). Sensitive periods in the development of the
ture Reviews Neuroscience, 2, 475483. brain and behavior. Journal of Cognitive Neuroscience.
Johnson, M. H., Dziurawiec, S., Ellis, H., & Morton, J. (1991). New- Kolb, B., & Whishaw, I. Q. (1998). Brain plasticity and behavior. An-
borns preferential tracking of face-like stimuli and its subse- nual Review of Psychology, 49, 4364.
quent decline. Cognition, 40, 119. Konishi, S., Nakajima, K., Uchida, I., Kikyo, H., Kameyama, M., &
Johnson, M. H., & Morton, J. (1991). Biology and cognitive develop- Miyashita, Y. (1999). Common inhibitory mechanism in human
ment: The case of face recognition. Oxford, England: Blackwell. inferior prefrontal cortex revealed by event-related functional
Johnson, M. H., Posner, M., & Rothbart, M. (1994). Facilitation of MRI. Brain, 122, 981999.
saccades toward a covertly attended location in early infancy. Kornack, D. R., & Rakic, P. (1999). Continuation of neurogenesis in
Psychological Science, 5(2), 9093. the hippocampus of the adult macaque monkey. Proceedings of the
Johnston, S., Leek, E. C., Atherton, C., Thacker, N., & Jackson, A. National Academy of Sciences, USA, 98, 57685773.
(2004). Functional contribution of medial premotor cortex to Kostovic, I. (1990). Structural and histochemical reorganization of
visuo-spatial transformation in humans. Neuroscience Letters, the human prefrontal cortex during perinatal and postnatal life.
355, 209212. Progress in Brain Research, 85, 223239.
Jordan, K., Wustenberg, T., Heinze, H. J., Peteres, M., & Jncke, L. Kozlowski, P. B., Brudkowska, J., Kraszpulski, M., Sersen, E. A., Wr-
(2003). Women and men exhibit different cortical activation patterns zolek, M. A., Anzil, A. P., et al. (1997). Microencephaly in chil-
during mental rotation tasks. Neuropsychologia, 40, 23972408. dren congenitally infected with human immunodeficiency virus:
Kanwisher, N., McDermott, J., & Chun, M. M. (1997). The fusiform A gross-anatomical morphometric study. Acta Neuropathologica,
face area: A module in human extrastriate cortex specialized for 93(2), 136145.
face perception. Journal of Neuroscience, 17, 43024311. Kriegstein, A. R., & Gtz, M. (2003). Radial glia diversity: A matter
Kawashima, R., Satoh, K., Itoh, H., Ono, S., Furumoto, S., Grotoh, of cell fate. Glia, 43, 3743.
R., et al. (1996). Functional anatomy of go/no-go discrimination Kuhl, P. K., Tsao, F. M., & Liu, H. M. (2003). Foreign-language expe-
and response selection: A PET study in man. Brain Research, rience in infancy: Effects of short-term exposure and social inter-
728, 7989. action on phonetic learning. Proceedings of the National Academy
Keele, S., & Rafal, R. (2000). Deficits of task set in patients with left of Sciences, 100, 90969101.
prefrontal cortex lesions. In S. Monsell & J. Driver (Eds.), Con- Kwon, H., Menon, V., Eliez, S., Warsofsky, I. S., White, C. D., Dyer-
trol of cognitive processes, attention and performance (Vol. 18). Friedman, J., et al. (2001). Functional neuroanatomy of visuospa-
Cambridge, MA: MIT Press. tial working memory in fragile X syndrome: Relation to
Keith, A. (1948). Human embryology and morphology. London: Ed- behavioral and molecular measures. American Journal of Psychia-
ward Arnold & Company. try, 158, 10401051.
Keller, A., Castellanos, F. X., Vaituzis, A. C., Jeffries, N. O., Giedd, Kwon, H., Reiss, A. L., & Menon, V. (2002). Neural basis of pro-
J. N., & Rapoport, J. L. (2003). Progressive loss of cerebellar vol- tracted developmental changes in visuo-spatial working mem-
ume in childhood-onset schizophrenia. American Journal of Psy- ory. Proceedings of the National Academy of Science, USA, 99,
chiatry, 160(1), 128133. 1333613341.
Kennedy, H., Bullier, J., & Dehay, C. (1989). Transient projection LaBar, K., Gatenby, J., Gore, J., LeDoux, J., & Phelps, E. (1998).
from the superior temporal sulcus to area 17 in the newborn Human amygdala activation during conditioned fear acquisition
macaque monkey. Proceedings of the National Academy of Sci- and extinction: A Mixed-Trial fMRI Study. Neuron, 20, 937945.
ences, USA, 86, 80938097. LaMantia, A. S., & Rakic, P. (1990). Axon overproduction and elimi-
Kesler-West, M. L., Andersen, A. H., Smith, C. D., Avison, M. J., nation in the corpus callosum of the developing rhesus monkey.
Davis, C. E., Kryscio, R. J., et al. (2001). Neural substrates of fa- Journal of Neuroscience, 10, 21562175.
cial emotion processing using fMRI. Brain Research: Cognitive Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1990). Emotion, at-
Brain Research, 11, 213226. tention, and the startle ref lex. Psychological Review, 97, 377395.
Kikuchi-Yorioka, Y., & Sawaguchi, T. (2000). Parallel visuospatial Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1992). A motivational
and audiospatial working memory processes in the monkey dorso- analysis of emotion: Ref lex-cortex connections. Psychological
lateral prefrontal cortex. Nature Neuroscience, 3, 10751076. Science, 3, 4449.
Kinsley, C. H., Madonia, L., Gifford, G. W., Tureski, K., Griffin, Lanzenberger, R., Uhl, F., Windischberger, C., Gartus, A., Streibl,
G. R., Lowry, C., et al. (1999). Motherhood improves learning and B., Edward, V., et al. (2001). Cross-modal plasticity in congeni-
memory. Nature, 402, 137138. tally blind subjects. International Society for Magnetic Resonance
Klingberg, T., Forssberg, H., & Westerberg, H. (2002). Increased Medicine, 9.
brain activity in frontal and parietal cortex underlies the develop- Le Grand, R., Mondloch, C. J., Maurer, D., & Brent, H. P. (2001).
ment of visuospatial working memory capacity during childhood. Early visual experience and face processing. Nature, 410, 890.
Journal of Cognitive Neuroscience, 14(1), 110. Le Grand, R., Mondloch, C. J., Maurer, D., & Brent, H. P. (2003). Ex-
Knopman, D., & Nissen, M. (1991). Procedural learning is impaired pert face processing requires input to the right hemisphere during
in Huntingtons disease: Evidence from the serial reaction time infancy. Nature Neuroscience, 6, 11081112.
task. Neuropsychologia, 29(3), 245254. Lepage, M., & Richer, F. (1996). Inter-response interference con-
Knudsen, E. I. (2003a). Early experience and critical periods. In tributes to the sequencing deficit in frontal lobe lesions. Brain,
L. R. Squire, F. E. Bloom, S. K. McConnell, J. L. Roberts, N. C. 119, 12891295.
Spitzer, & M. J. Zigmond (Eds.), Fundamental neuroscience (2nd Lespinet, V., Bresson, C., NKaoua, B., Rougier, A., & Claverie, B.
ed., pp. 555573). New York: Academic Press. (2002). Effect of age of onset of temporal lobe epilepsy on the
52 Neural Bases of Cognitive Development

severity and the nature of preoperative memory deficits. Neu- Marin-Padilla, M. (1978). Dual origin of the mammalian neocortex
ropsychologia, 40, 15911600. and evolution of the cortical plate. Anatomical Embryology,
Levitt, P. (2003). Structural and functional maturation of the devel- 152, 109126.
oping primate brain. Journal of Pediatrics, 143(Suppl. 4), Martinez, A., Moses, P., Frank, L., Buxton, R., Wong, E., & Stiles, J.
S35S45. (1997). Hemispheric asymmetries in global and local processing:
Levy, R., & Goldman-Rakic, P. (2000). Segregation of working mem- Evidence from fMRI. NeuroReport, 8, 16851689.
ory functions within the dorsolateral prefrontal cortex. Experi- Martinkauppi, S., Rama, P., Aronen, H. J., Korvenoja, A., & Carlson,
mental Brain Research, 133(1), 2332. S. (2000). Working memory of auditory localization. Cerebral
Lewicki, P. (1986). Processing information about covariations that Cortex, 10, 889898.
cannot be articulated. Journal of Experimental Psychology: Learn- Mauk, M., & Thompson, R. (1987). Retention of classically condi-
ing, Memory, and Cognition, 12, 135146. tioned eyelid responses following acute decerebration. Brain Re-
Lewin, J., Friedman, L., Wu, D., Miller, D., Thompson, L., Klein, S., search, 403, 8995.
et al. (1996). Cortical localization of human sustained attention: Maurer, D., Lewis, T. L., Brent, H. P., & Levin, A. V. (1999). Rapid
Detection with functional MR using a vigilance paradigm. Jour- improvement in the acuity of infants after visual input. Science,
nal of Computer Assisted Tomography, 20(5), 695701. 286, 108110.
Lie, D. C., Song, H., Colamarino, S. A., Ming, G.-I., & Gage, F. H. Mayberry, R. I., Lock, E., & Kazmi, H. (2002). Linguistic ability and
(2004). Neurogenesis in the adult brain: New strategies for central early language exposure. Nature, 417, 38.
nervous system diseases. Annual Review of Pharmacology and Maybery, M., Taylor, M., & OBrien-Malone, A. (1995). Implicit
Toxicology, 44, 399421. learning: Sensitive to age but not to IQ. Australian Journal of Psy-
Luciana, M. (2003). The neural and functional development of human chology, 47, 817.
prefrontal cortex. In M. de Haan & M. H. Johnson (Eds.), The McDonough, L., Mandler, J. M., McKee, R. D., & Squire, L. R.
cognitive neuroscience of development (pp. 157179). London: (1995). The deferred imitation task as a nonverbal measure of de-
Psychology Press. clarative memory. Proceedings of the National Academy of Sci-
ences, 92, 75807584.
Luciana, M., & Nelson, C. A. (1998). The functional emergence of
prefrontally-guided working memory systems in 4- to 8-year-old McKee, R. D., & Squire, L. R. (1993). On the development of declar-
children. Neuropsychologia, 36, 272293. ative memory. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 19, 397404.
Luciana, M., Sullivan, J., & Nelson, C. A. (2001). Individual differ-
ences in phenylalanine levels moderate performance on tests of Meletti, S., Benuzzi, F., Rubboli, G., Cantalupo, G., Stanzani
executive function in adolescents treated early and continuously Maserati, M., Nichelli, P., et al. (2003). Impaired facial emotion
for PKU. Child Development, 72, 16371652. recognition in early-onset right mesial temporal epilepsy. Neurol-
ogy, 60, 426431.
Lumsden, A., & Kintner, C. (2003). Neural induction and pattern for-
Meltzoff, A. N. (1985). Immediate and deferred imitation in 14- and
mation. In L. R. Squire, F. E. Bloom, S. K. McConnell, J. L.
24-month-old infants. Child Development, 56, 6272.
Roberts, N. C. Spitzer, & M. J. Zigmond (Eds.), Fundamental neu-
roscience (2nd ed., pp. 363390). New York: Academic Press. Meltzoff, A. N. (1988). Infant imitation and memory: Nine-month-
olds in immediate and deferred tests. Child Development, 59,
Luna, B., Thulborn, K., Munoz, D., Merriam, E., Garver, K., Minshew,
217225.
N., et al. (2001). Maturation of widely distributed brain function
subserves cognitive development. NeuroImage, 13, 786793. Meltzoff, A. N. (1995). What infant memory tells us about infantile
amnesia: Long-term recall and deferred imitation. Journal of Ex-
Mabbott, D. J., & Smith, M. L. (2003). Memory in children with perimental Child Psychology, 59, 497515.
temporal or extra-temporal excisions. Neuropsychologia, 41,
Mettke-Hofmann, C., & Gwinner, E. (2003). Long-term memory for a
9951007.
life on the move. Proceedings of the National Academy of Sciences,
MacDonald, A., Cohen, J., Stenger, V., & Carter, C. (2000). Dissoci- USA, 100, 58635866.
ating the role of the dorsolateral prefrontal and anterior cingulate
Meulemans, T., van der Linden, M., & Perruchet, P. (1998). Implicit
cortex in cognitive control. Science, 288(5472), 18351838.
sequence learning in children. Journal of Experimental Child Psy-
Maguire, E. A., Gadian, D. G., Johnsrude, I. S., Good, C. D., Ash- chology, 69, 199221.
burner, J., Frackowiak, R. S. J., et al. (2000). Navigation-related Milner, B. (1995). Aspects of human frontal lobe function. In H. H.
structural change in the hippocampi of taxi drivers. Proceedings Jasper, S. Riggio, & P. S. Goldman-Rakic (Eds.), Epilepsy and the
of the National Academy of Sciences, 97, 43984403. functional anatomy of the frontal lobe. New York: Raven Press.
Maguire, E. A., Vargha-Khadem, F., & Mishkin, M. (2001). The effects Milner, B., Corkin, S., & Teuber, H. (1968). Further analysis of the
of bilateral hippocampal damage on fMRI regional activations and hippocampal amnesic syndrome: 14-year follow-up study of HM.
interactions during memory retrieval. Brain, 124, 11561170. Neuropsychologia, 6, 215234.
Mlkov, L., Bachevalier, J., Mishkin, M., & Saunders, C. (2001). Mirescu, C., Peters, J. D., & Gould, E. (2004). Early life experience
Neurotoxic lesions of perirhinal cortex impair visual recognition alters response of adult Neurogenesis to stress. Nature Neuro-
memory in rhesus monkeys. NeuroReport, 12, 19131917. science, 7, 841846.
Manns, J. R., & Squire, L. R. (1999). Impaired recognition memory Molliver, M., Kostovic, I., & Van der Loos, H. (1973). The develop-
on the Doors and People Test after damage limited to the hip- ment of synapses in the human fetus. Brain Research, 50, 403407.
pocampal region. Hippocampus, 9(5), 495499. Mondloch, C., Geldart, S., Maurer, D., & de Schonen, S. (2003). De-
Manns, J. R., Stark, C. E., & Squire, L. R. (2000). The visual paired- velopmental changes in the processing of hierarchical shapes
comparison task as a measure of declarative memory. Proceedings continue into adolescence. Journal of Experimental Child Psy-
of the National Academy of Sciences, 97(22), 1237512379. chology, 84, 2040.
References 53

Monk, C. S., Zhuang, J., Curtis, W. J., Ofenloch, I. T., Tottenham, N., Nelson, C. A., & Collins, P. F. (1992). Neural and behavioral corre-
Nelson, C. A., et al. (2002). Human hippocampal activation in the lates of recognition memory in 4- and 8-month-old infants. Brain
delayed matching- and nonmatching-to-sample memory tasks: An and Cognition, 19, 105121.
event-related functional MRI approach. Behavioral Neuroscience, Nelson, C. A., Lin, J., Carver, L. J., Monk, C. S., Thomas, K. M., &
116, 716721. Truwit, C. L. (2000). Functional neuroanatomy of spatial working
Morris, J. S., Frith, C. D., Perrett, K. I., Rowland, D., Young, A. W., memory in children. Developmental Psychology, 36(1), 109116.
Calder, A. J., et al. (1996). A differential neural response in the Nelson, C. A., & Luciana, M. (Eds.). (2001). Handbook of develop-
human amygdala to fearful and happy facial expressions. Nature, mental cognitive neuroscience. Cambridge, MA: MIT Press.
383, 812815.
Nelson, C. A., & Monk, C. S. (2001). The use of event-related poten-
Moscovitch, M., Winocur, G., & Behrmann, M. (1997). What is spe- tials in the study of cognitive development. In C. Nelson & M. Lu-
cial about face recognition? Nineteen experiments on a person ciana (Eds.), Handbook of developmental cognitive neuroscience
with visual object agnosia but normal face recognition. Journal of (pp. 125136). Cambridge, MA: MIT Press.
Cognitive Neuroscience, 9, 555604.
Nelson, C. A., & Webb, S. J. (2003). A cognitive neuroscience per-
Moses, P., Roe, K., Buxton, R. B., Wong, E. C., Frank, L. R., & Stiles, spective on early memory development. In M. de Haan & M. H.
J. (2002). Functional MRI of global and local processing in chil- Johnson (Eds.), The cognitive neuroscience of development
dren. Neuroimage, 16, 415424. (pp. 99125). London: Psychology Press.
Mrzljak, L., Uylings, H. B. M., Kostovic, I., & VanEden, C. (1990).
Nemanic, S., Alvarado, M. C., & Bachevalier, J. (2004). The hip-
Prenatal development of neurons in the human prefrontal cortex:
pocampal /parahippocampal regions and recognition memory: In-
Vol. I. A qualitative golgi study. Journal of Comparative Neurol-
sights from visual paired comparison versus object-delayed
ogy, 271, 355386.
nonmatching in monkeys. Journal of Neuroscience, 24, 20132026.
Murloz-Sanjuan, I., & Brivanfou, A. H. (2002). Neural induction:
Newman, A. J., Bavelier, D., Corina, D., Jezzard, P., & Neville, H. J.
The default model and embryonic stem cells. Nature Reviews Neu-
(2002). A critical period for right hemisphere recruitment in Amer-
roscience, 3(4), 271280.
ican sign language processing. Nature Neuroscience, 5, 7680.
Nadarajah, B., & Parvavelas, J. G. (2002). Models of neuronal migra-
Nishitani, N., & Hari, R. (2000). Temporal dynamics of cortical rep-
tion in the developing cerebral cortex. Nature Reviews Neuro-
resentation for action. Proceedings of the National Academy of
science, 3, 423432.
Sciences, USA, 97, 913918.
Naegele, J. R., & Lombroso, P. J. (2001). Genetics of central nervous
system developmental disorders. Child and Adolescent Psychiatric Nissen, M., & Bullemer, P. (1987). Attentional requirements of
Clinics of North America, 10, 225239. learning: Evidence from performance measures. Cognitive Psy-
chology, 19, 132.
Nelson, C. (1994). Neural correlates of recognition memory in the
first postnatal year of life. In G. Dawson & K. Fischer (Eds.), Nunn, J. A., Polkey, C. E., & Morris, R. G. (1998). Selective spatial
Human behavior and the developing brain (pp. 269313). New memory impairment after right unilateral temporal lobectomy.
York: Guilford Press. Neuropsychologia, 36, 837848.
Nelson, C. A. (1995). The ontogeny of human memory: A cognitive Olesen, P. J., Nagy, Z., Westerberg, H., & Klingberg, T. (2003). Com-
neuroscience perspective. Developmental Psychology, 31, bined analysis of DTI and fMRI data reveals a joint maturation of
723738. white and grey matter in a fronto-parietal network. Brain Re-
Nelson, C. A. (1998). The nature of early memory. Preventive Medi- search: Cognitive Brain Research, 18, 4857.
cine, 27, 172179. Olesen, P. J., Westerberg, H., & Klingberg, T. (2004). Increased pre-
Nelson, C. A. (1999). Neural plasticity and human development. Cur- frontal and parietal activity after training of working memory.
rent Directions in Psychological Science, 8, 4245. Nature Neuroscience, 7, 7579.
Nelson, C. A. (2000a). Change and continuity in neurobehavioral de- Oppenheim, R. W., & Johnson, J. E. (2003). Programmed cell death
velopment. Infant Behavior and Development, 22, 415429. and neurotrophic factors. In L. R. Squire, F. E. Bloom, S. K. Mc-
Nelson, C. A. (2000b). Neural plasticity and human development: The Connell, J. L. Roberts, N. C. Spitzer, & M. J. Zigmond (Eds.),
role of early experience in sculpting memory systems. Develop- Fundamental neuroscience (2nd ed., pp. 499532). New York:
mental Science, 3, 115130. Academic Press.
Nelson, C. A. (2001). The development and neural bases of face Orlich, E., & Ross, L. (1968). Acquisition and differential condition-
recognition. Infant and Child Development, 10, 318. ing of the eyelid response in normal and retarded children. Jour-
nal of Experimental Child Psychology, 6, 181193.
Nelson, C. A. (2002). Neural development and life-long plasticity. In
R. M. Lerner, F. Jacobs, & D. Wetlieb (Eds.), Promoting positive Overman, W. H., Bachevalier, J., Sewell, F., & Drew, J. (1993). A
child, adolescent, and family development: Handbook of program comparison of childrens performance on two recognition mem-
and policy interventions (pp. 3160). Thousand Oaks, CA: Sage. ory tasks: Delayed nonmatch-to-sample-versus-visual-paired-
Nelson, C. A., & Bloom, F. E. (1997). Child development and neuro- comparison. Developmental Psychobiology, 26, 345357.
science. Child Development, 68, 970987. Owen, A., Roberts, A., Hodges, J., Summers, B., Polkey, C., & Rob-
Nelson, C. A., & Carver, L. J. (1998). The effects of stress on brain bins, T. (1993). Contrasting mechanisms of impaired attentional
and memory: A view from developmental cognitive neuroscience. set shifting in patients with frontal lobe damage or Parkinsons
Development and Psychopathology, 10, 793809. disease. Brain, 119, 15971615.
Nelson, C. A., & Collins, P. F. (1991). Event-related potential and Parkin, A., & Streete, S. (1988). Implicit and explicit memory in young
looking time analysis of infants responses to familiar and novel children and adults. British Journal of Psychology, 79, 361369.
events: Implications for visual recognition memory. Developmen- Pascalis, O., & Bachevalier, J. (1999). Neonatal aspiration lesions of
tal Psychology, 27, 5058. the hippocampal formation impair visual recognition memory
54 Neural Bases of Cognitive Development

when assessed by paired-comparison task but not by delayed non- strings, and textures: A functional magnetic resonance imaging
matching-to-sample task. Hippocampus, 9, 609616. study. Journal of Neuroscience, 16, 52055215.
Pascalis, O., de Haan, M., & Nelson, C. A. (2002). Is face pro- Rafal, D., Calabresi, P., Brennan, C., & Scioloto, T. (1989). Saccade
cessing species specific during the first year of life? Science, preparation inhibits reorienting to recently attended locations.
296, 13211323. Journal of Experimental Psychology: Human Perception and Per-
Pascalis, O., Hunkin, N. M., Holdstock, J. S., Isaac, C. L., & Mayes, formance, 15, 673685.
A. R. (2004). Visual paired comparison performance is impaired Rafal, R., Posner, M., Friedman, J., Inhoff, A., & Bernstein, E.
in a patient with selective hippocampal lesions and relatively in- (1988). Orienting of visual attention in progressive supranuclear
tact item recognition. Neuropsychologia, 42, 12301293. palsy. Brain, 111(2), 267280.
Pascual-Leone, A., Grafman, J., Clark, K., Stewart, M., Massaquoi, S., Rakic, P. (1988). Specification of cerebral cortical areas. Science,
Lou, J.-S., et al. (1993). Procedural learning in Parkinsons disease 241, 170176.
and cerebellar degeneration. Annals of Neurology, 34, 594602. Rakic, P. (2002a). Adult neurogenesis in mammals: An identity cri-
Passarotti, A. M., Paul, B. M., Bussiere, J. R., Buxton, R. B., Wong, sis. Journal of Neuroscience, 22, 614618.
E. C., & Stiles, J. (2003). The development of face and location Rakic, P. (2002b). Neurogenesis in adult primate neocortex: An eval-
processing: An fMRI study. Developmental Science, 6, 100117. uation of the evidence. Nature Reviews Neuroscience, 3, 6571.
Paterson, S. J., Brown, J. H., Gsodl, M. K., Johnson, M. H., & Raper, J. A., & Tessier-Lavigne, M. (1999). Growth cones and axon
Karmiloff-Smith, A. (1999). Cognitive modularity and genetic pathfinding. In M. J. Zigmond, F. E. Bloom, S. C. Landis, J. L.
disorders. Science, 28, 23552358. Roberts, & L. R. Squire (Eds.), Fundamental neuroscience
Paus, T., Zijdenbos, A., Worsley, K., Collins, D. L., Blumenthal, J., (pp. 579596). New York: Academic Press.
Giedd, J. N., et al. (1999). Structural maturation of neural path- Raper, J., & Tessier-Lavigne, M. (2003). Growth cones and axon
ways in children and adolescents: In vivo study. Science, 283, pathfinding. In L. R. Squire, F. E. Bloom, S. K. McConnell,
19081911. J. L. Roberts, N. C. Spitzer, & M. J. Zigmond (Eds.), Funda-
Perani, D., Paulesu, E., Galles, N. S., Dupoux, E., Dehaene, S., Betti- mental neuroscience (2nd ed., pp. 449467). New York: Aca-
nardi, V., et al. (1998). The bilingual brain: Proficiency and age of demic Press.
acquisition of the second language. Brain, 121, 18411852. Rebai, M., Poiroux, S., Bernard, C., & Lalonde, R. (2001). Event-re-
Petitto, L. A., Zatorre, R. J., Gauna, K., Nikeiski, E. J., Dostie, D., & lated potentials for category-specific information during passive
Evans, A. C. (2000). Speech-like cerebral activity in profoundly viewing of faces and objects. Internal Journal of Neuroscience,
deaf people processing signed languages: Implications for the 106(3/4), 209226.
neural basis of human language. Proceedings of the National Acad- Reber, A. (1992). The cognitive unconscious: An evolutionary per-
emy of Sciences, 97, 1396113966. spective. Consciousness and Cognition, 1, 93133.
Pickering, S. J. (2001a). Cognitive approaches to the fractionation of Reber, A. (1993). Implicit learning and tacit knowledge: An essay on
visuo-spatial working memory. Cortex, 37, 457473. the cognitive unconscious (Vol. 19). New York: Oxford Univer-
Pickering, S. J. (2001b). The development of visuo-spatial working sity Press.
memory. Memory, 9, 423432. Richards, J. (1994). Baseline respiratory sinus arrhythmia and
Pierce, K., Muller, R. A., Ambrose, J., Allen, G., & Courchesne, E. heart rate responses during sustained visual attention in
(2001). Face processing occurs outside the fusiform face area preterm infants from 3 to 6 months of age. Psychophysiology,
in autism: Evidence from functional MRI. Brain, 124, 20592073. 31, 235243.
Pine, D., Fyer, A., Grun, J., Phelps, E., Szesko, P., Koda, V., et al. Richards, J. (1998). Development of selective attention in young in-
(2001). Methods for developmental studies of fear conditioning fants. Developmental Science, 1, 4551.
circuitry. Biological Psychiatry, 50, 225228. Richards, J. (2001). Attention in young infants: A developmental psy-
Pollak, S. D., Cicchetti, D., Hornung, K., & Reed, A. (2000). Recog- chophysiological perspective. In C. Nelson & M. Luciana (Eds.),
nizing emotion in faces: Developmental effects of child abuse and Handbook of cognitive neuroscience (pp. 321338). Cambridge,
neglect. Developmental Psychology, 36(5), 679688. MA: MIT Press.
Pollak, S. D., & Kistler, D. J. (2002). Early experience is associated Richter, W., Somorjai, R., Summers, R., Jaramasz, M., Menon, R. S.,
with the development of categorical representations for facial ex- Gati, J. S., et al. (2000). Motor area activity during mental rota-
pressions of emotion. Proceedings of the National Academy of Sci- tion studied by time-resolved single-trial fMRI. Journal of Cogni-
ences, USA, 99(13), 90729076. tive Neuroscience, 12, 310320.
Pollak, S. D., & Sinha, P. (2002). Effects of early experience on chil- Richter, W., Ugurbil, K., Georgopoulos, A., & Kim, S.-G. (1997). Time-
drens recognition and facial displays of emotion. Developmental resolved fMRI of mental rotation. NeuroReport, 8, 36973702.
Psychology, 38(5), 784791.
Ridderinkhof, K., van der Molen, M., Band, G., & Bashore, T.
Posner, M., & Petersen, S. (1990). The attention system of the human (1997). Sources of interference from irrelevant information: A
brain. Annual Review of Neuroscience, 13, 2542. developmental study. Journal of Experimental Child Psychology,
Posner, M., & Raichle, M. (1997). Images of mind. New York: Scien- 65, 315341.
tific American Library. Ridley, A. J., Schwartz, M. A., Burridge, K., Firtel, R. A., Ginsberg,
Posner, M., Walker, J., Friedrich, F., & Rafal, R. (1984). Effects of M. H., Borisy, G., et al. (2003). Cell migration: Integrating sig-
parietal injury on covert orienting of attention. Journal of Neuro- nals from front to back. Science, 302, 17041709.
science, 4(7), 18631874. Roberts, J. E., & Bell, M. A. (2002). The effects of age and sex on
Puce, A., Allison, T., Asgari, M., Gore, J. C., & McCarthy, G. (1996). mental rotation performance, verbal performance, and brain elec-
Differential sensitivity of human visual cortex to faces, letter- trical activity. Developmental Psychobiology, 40, 391407.
References 55

Roberts, J. E., & Bell, M. A. (2003). Two- and three-dimensional Schendan, H., Searl, M., Melrose, R., & Ce, S. (2003). An FMRI
mental rotation tasks lead to different parietal laterality for men study of the role of the medial temporal lobe in implicit and ex-
and women. International Journal of Psychophysiology, 50, plicit sequence learning. Neuron, 37(6), 10131025.
235246. Schoppik, D., Gadian, D. G., Connelly, A., Mishkin, M., Vargha-
Robinson, A. J., & Pascalis, O. (2004). Development of f lexible visual Khadem, F., & Saunders, R. C. (2001). Volumetric measurement
recognition memory in human infants. Developmental Science, 7, of the subhippocampal cortices in patients with developmental
527533. amnesia. Society for Neuroscience, 27, 1400.
Robinson, D., Bowman, E., & Kertzman, C. (1995). Covert orienting Schultz, R. T., Gauthier, I., Klin, A., Fulbright, K. A., Anderson,
of attention in macques: Vol. 2. Contributions of parietal cortex. A. W., Volkmar, F., et al. (2000). Abnormal ventral temporal cor-
Journal of Neurophysiology, 74(2), 698712. tical activity during face discrimination among individuals with
autism and Asperger syndrome. Archives of General Psychiatry,
Roe, K., Moses, P., & Stiles, J. (1999). Lateralization of spatial
57, 331340.
processes in school aged children [Abstracts]. Cognitive Neuro-
science Society, 41. Schwartz, B., & Hashtroudi, S. (1991). Priming is independent of
skill learning. Journal of Experimental Psychology: Learning,
Rossion, B., Gauthier, I., Tarr, M. J., Despland, P., Bruyer, R., Memory and Cognition, 17(6), 11771187.
Linotte, S., et al. (2000). The N170 occipito-temporal component
Schwarzer, G., & Zauner, N. (2003). Face processing in 8-month-old
is delayed and enhanced to inverted faces but not inverted objects:
infants: Evidence for configural and analytical processing. Vision
An electrophysiological account of face-specific processes in the
Research, 43, 27832793.
human brain. NeuroReport, 11, 6974.
Schweinberger, S. R., Pickering, E. C., Jentzsch, I., Burton, A. M., &
Rovee-Collier, C. (1997a). Development of memory in infancy. In N. Kaufmann, J. M. (2002). Event-related brain potential evidence
Cowan (Ed.), The development of memory in childhood (pp. 539). for a response of inferior temporal cortex to familiar face repeti-
London: University College London Press. tions. Brain Research: Cognitive Brain Research, 14, 398409.
Rovee-Collier, C. (1997b). Dissociations in infant memory: Rethink- Scoville, W. B., & Milner, B. (1957). Loss of recent memory after bi-
ing the development of implicit and explicit memory. Psychologi- lateral hippocampal lesions. Journal of Neurology, Neurosurgery,
cal Review, 104, 467498. and Psychiatry, 20, 1121.
Ruchkin, D. S., Johnson, R., Jr., Grafman, J., Canoune, H., & Ritter, Seger, C. (1994). Implicit learning. Psychological Bulletin,
W. (1997). Multiple visuospatial working memory buffers: Evi- 115(2), 163196.
dence from spatiotemporal patterns of brain activity. Neuropsy- Serres, L. (2001). Morphological changes of the human hippocampal
chologia, 35(2), 195209. formation from midgestation to early childhood. In C. A. Nelson
Rueda, M. R., Fan, J., McCandliss, B. D., Halparin, J. D., Gruber, & M. Luciana (Eds.), Handbook of developmental cognitive neuro-
science (pp. 4558). Cambridge, MA: MIT Press.
D. B., Lercari, L. P., et al. (2004). Development of attentional net-
works in childhood. Neuropsychologia, 42(8), 10291040. Shallice, T. (1988). From neuropsychology to mental structure. New
York: Cambridge University Press.
Russo, R., Nichelli, P., Gibertoni, M., & Cornia, C. (1995). Develop-
Shallice, T., & Burgess, P. (1991). Higher-order cognitive impair-
mental trends in implicit and explicit memory: A picture comple-
ments and frontal lobe lesions in man. In H. Levin, H. Eisenberg,
tion study. Journal of Experimental Child Psychology, 59, 566578.
& A. Benton (Eds.), Frontal lobe function and dysfunction
Rutishauser, U. (1993). Adhesion molecules of the nervous system. (pp. 125138). Oxford, England: Oxford University Press.
Current Opinion in Neurobiology, 3, 709715. Shelton, A. L., Christoff, K., Burrows, J. J., Pelisari, K. B., & Gabrieli,
Sadato, N., Okada, T., Honda, M., & Yonekura, Y. (2002). Critical J. D. E. (2001). Brain activation during mental rotation: Individual
period for cross-modal plasticity in blind humans: A functional differences. Society for Neuroscience Abstracts, 27, 456.
MRI study. Neuroimage, 16, 389400. Shibata, T., Nishijo, H., Tamura, R., Miyamoto, K., Eifuku, S., Endo,
Sadato, N., Pascual-Leone, A., Grafman, J., Deiber, M. P., Ibanez, V., S., et al. (2002). Generators of visual evoked potentials for faces
& Hallett, M. (1998). Neural networks for Braille reading by the and eyes in the human brain as determined by dipole localization.
blind. Brain, 121, 12131229. Brain Topography, 15, 5163.
Sala, J. B., Rama, P., & Courtney, S. M. (2003). Functional topogra- Shimamura, A. (1986). Priming effects in amnesia: Evidence for a
phy of a distributed neural system for spatial and nonspatial in- dissociable memory function. Quarterly Journal of Experimental
Psychology, 38A, 619644.
formation maintenance in working memory. Neuropsychologia,
41(3), 341356. Shingo, T., Gregg, C., Enwere, E., Fujikawa, H., Hassam, R., Geary,
C., et al. (2003). Pregnancy-stimulated neurogenesis in the adult
Sasaki, Y., Hadijkhani, M., Fischl, B., Liu, A. K., Marrett, S., Dale,
female forebrain mediated by prolactin. Science, 299, 117120.
A. M., et al. (2001). Local and global attention are mapped retino-
Shors, T. J., & Miesegaes, G. (2002). Testosterone in utero and at
topically in human occipital cortex. Proceedings of the National
birth dictates how stressful experience will affect learning in
Academy of Sciences, USA, 98, 20772082.
adulthood. Proceedings of the National Academy of Sciences, 99,
Schachner, M., Taylor, J., Bartsch, U., & Pesheva P. (1994). The per- 1395513960.
plexing multifunctionality of janusin, a tenascin-related molecule. Sidman, R., & Rakic, P. (1982). Development of the human central
Perspectives in Developmental Neurobiology, 2(1), 3341. nervous system. In W. Haymaker & R. D. Adams (Eds.), Histol-
Schacter, D., & Buckner, R. (1998). On the relations among priming, ogy and histopathology of the nervous system. Springfield, IL:
conscious recollection, and intentional retrieval: Evidence from Charles C Thomas.
neuroimaging research. Neurobiology of Learning and Memory, Siegler, R. S. (1991). Childrens Thinking (2nd ed.). Englewood
70, 284303. Cliffs, NJ: Prentice-Hall.
56 Neural Bases of Cognitive Development

Smith, E. E., & Jonides J. (1998). Neuroimaging analyses of human Sur, M., & Leamey, C. A. (2001). Development and plasticity of cor-
working memory. Proceedings of the National Academy of Sci- tical areas and networks. Nature Reviews Neuroscience, 2(4),
ences, USA, 95, 1206112068. 251262.
Smith, M. L., & Milner, B. (1988). Estimation of frequency of occur- Tabata, H., & Nakajima, K. (2003). Multipolar migration: The third
rence of abstract designs after frontal or temporal lobectomy. mode of radial neuronal migration in the developing cerebral cor-
Neuropsychologia, 26, 297306. tex. Journal of Neuroscience, 23(31), 999610001.
Smith, P., Loboschefski, T., Davidson, B., & Dixon, W. (1997). Takahashi, T., Nowakowski, R. S., & Caviness, V. S. (2001). Neocor-
Scripts and checkerboards: The inf luence of ordered visual infor- tical neurogeneisis: Regulation, control points, and a strategy of
mation on remembering locations in infancy. Infant Behavior and structural variation. In C. A. Nelson & M. Luciana (Eds.), Hand-
Development, 13, 129146. book of developmental cognitive neuroscience (pp. 322). Cam-
Song, H., Stevens, C. E., & Gage, F. H. (2002). Neural stem cells bridge, MA: MIT Press.
from adult hippocampus develop essential properties of func- Takeichi, M. (1995). Morphogenetic roles of classic caherins. Current
tional CNS neurons. Nature Neuroscience, 5, 438445. Opinion in Cell Biology, 7, 619627.
Sowell, E. R., Thompson, P. M., Holmes, C. J., Batth, R., Jernigan, Tanaka, T., & Gleeson, J. G. (2000). Genetics of brain development
T. L., & Toga, A. W. (1999). Localizing age-related changes in and malformation syndromes [Related articles]. Current Opinions
brain structure between childhood and adolescence using statisti- in Pediatrics, 12(6), 523528.
cal parametric mapping. Neuroimage, 9, 587597. Taylor, M. J., Edmonds, G. E., McCarthy, G., & Allison, T. (2001).
Sowell, E. R., Thompson, P. M., Holmes, C. J., Jernigan, T. L., & Eyes first! Eye processing develops before face processing in
Toga, A. W. (2000). In vivo evidence for post-adolescent brain children. NeuroReport, 12, 16711676.
maturation in frontal and striatal regions. Nature Neuroscience, Taylor, M. J., McCarthy, G., Saliba, E., & Degiovanni, E. (1999).
2, 859961. ERP evidence of developmental changes in processing of faces.
Spelke, E. S. (2000). Core knowledge. American Psychologist, Clinical Neurophysiololgy, 110, 910915.
55, 12331243. Temple, C. M., & Richardson, P. (2004). Developmental amnesia: A
Spiers, H. J., Burgess, N., Hartley, T., Vargha-Khadem, F., & new pattern of dissociation with intact episodic memory. Neu-
OKeefe, J. (2001). Bilateral hippocampal pathology impairs top- ropsychologia, 42, 764781.
ographical and episodic memory but not visual pattern matching. Tessier-Lavigne, M., & Goodman, C. S. (1996). The molecular biol-
Hippocampus, 11, 715725. ogy of axon guidance. Science, 274, 11231133.
Squire, L. (1986). Mechanisms of memory. Science, 232, 16121619. Thomas, K. M., Drevets, W. C., Whalen, P. J., Eccard, C. H., Dahl, R. E.,
Squire, L. (1994). Declarative and nondeclarative memory: Multiple Ryan, N. D., et al. (2001). Amygdala response to facial expressions
brain systems supporting learning and memory. In D. L. Schacter in children and adults. Biological Psychiatry, 49, 309316.
& E. Tulving (Eds.), Memory systems (pp. 203231). Cambridge, Thomas, K. M., Hunt, R., Vizueta, N., Sommer, T., Durston, S., Yang,
MA: MIT Press. Y., et al. (in press). Evidence of developmental differences in im-
Squire, L., & Frambach, M. (1990). Cognitive skill learning in amne- plicit sequence learning: An fMRI study of children and adults.
sia. Psychobiology, 18, 109117. Journal of Cognitive Neuroscience.
Squire, L., Knowlton, B., & Musen, G. (1993). The structure and or- Thomas, K. M., King, S. W., Franzen, P. L., Welsh, T. F., Berkowitz,
ganization of memory. Annual Review of Psychology, 44, 453495. A. L., Noll, D. C., et al. (1999). A developmental functional MRI
Squire, L., & McKee, R. (1993). Declarative and nondeclarative mem- study of spatial working memory. Neuroimage, 10, 327338.
ory in opposition: When prior events inf luence amnesic patients Thomas, K. M., & Nelson, C. (2001). Serial reaction time learning in
more than normal subjects. Memory and Cognition, 21(4), 424430. preschool- and school-age children. Journal of Experimental Child
Squire, L., Ojemann, J., Miezin, F., Petersen, S., Videen, T., & Psychology, 79, 364387.
Raichle, M. (1992). Activation of the hippocampus in normal hu- Thomas, K. M., Vizueta, N., Teylan, M., Eccard, C., & Casey, B.
mans: A functional anatomical study of memory. Proceedings of (2003, April). Impaired learning in children with presumed basal
the National Academy of Sciences, USA, 89, 18371841. ganglia insults: Evidence from a serial reaction time task. Paper
Stewart, L., Henson, R., Kampe, K., Walsh, V., Turner, R., & Frith, presented at the annual meeting of the Cognitive Neuroscience
U. (2003). Becoming a pianist: An fMRI study of musical liter- Society, New York, NY.
acy acquisition. Annals of the New York Academy of Sciences, Thomsen, T., Hugdahl, K., Ersland, L., Barndon, R., Lundervold, A.,
999, 204208. Smievoll, A. I., et al. (2000). Functional magnetic resonance im-
Stiles, J., Moses, P., Passarotti, A., Dick, F. K., & Buxton, R. B. aging (fMRI ) study of sex differences in a mental rotation task.
(2003). Exploring developmental change in the neural bases of Medical Science Monitor, 6, 11861196.
higher cognitive functions: The promise of magnetic resonance Tomizawa, K., Iga, N., Lu, Y., Moriwaki, A., Matushita, M., Li, S.,
imaging. Developmental Neuropsychology, 24, 641668. et al. (2003). Oxytocin improves long-lasting spatial memory dur-
Stiles, J., Moses, P., Roe, K., Akshoomoff, N. A., Trauner, D. J., Wong, ing motherhood through MAP kinase cascade. Nature neuro-
E. L. R., et al. (2003). Alternative brain organization after prena- science, 6, 384390.
tal cerebral injury: Convergent fMRI and cognitive data. Journal Tulving, E. (1972). Episodic and semantic memory. In E. Tulving &
of the International Neuropsychological Society, 9, 604622. W. Donaldson (Eds.), Organization of memory (pp. 381403).
Strange, B. A., Fletcher, P. C., Hennson, R. N. A., Friston, K. J., & New York: Academic Press.
Dolan, R. J. (1999). Segregating the functions of the human Tulving, E., Markowitsch, H. J., Craik, F. E., Habib, R., & Houle, S.
hippocampus. Proceedings of the National Academy of Sciences, (1996). Novelty and familiarity activations in PET studies of
96, 40344039. memory encoding and retrieval. Cerebral Cortex, 6, 7179.
References 57

Turati, C., Simion, F., Milani, I., & Umilta, C. (2002). Newborns Weiss, E., Siedentopf, C. M., Hofer, A., Deisenhammer, E. A., Hopt-
preferences for faces: What is crucial. Developmental Psychology, man, M. J., Kremser, C., et al. (2003). Sex differences in brain ac-
38, 875882. tivation pattern during a visuospatial cognitive task: A functional
Turkheimer, E., Haley, A., Waldron, M., DOnofrio, B., & Gottes- magnetic resonance imaging study in healthy volunteers. Neuro-
man, I. I. (2003). Socioeconomic status modifies heritability of science Letters, 344, 169172.
IQ in young children. Current Directions in Psychological Science, Werker, J. F., & Vouloumanos, A. (2001). Speech and language pro-
14, 623628. cessing in infancy: A neurocognitive approach. In C. A. Nelson &
Tzourio-Mazoyer, N., de Schonen, S., Crivello, F., Reutter, B., Au- M. Luciana (Eds.), Handbook of developmental cognitive neuro-
jard, Y., & Mazoyer, B. (2002). Neural correlates of woman face science. Cambridge, MA: MIT Press.
processing by 2-month-old infants. NeuroImage, 15, 454461. Whalen, P. J., Shin, L. M., McInerney, S. C., Fisher, H., Wright, C. I.,
Uhl, F., Franzen, P., Lindinger, G., Lang, W., & Deecke, L. (1991). & Rauch, S. L. (2001). A functional MRI study of human amyg-
On the functionality of the visually deprived occipital cortex in dala responses to facial expressions of fear versus anger. Emo-
early blind person. Neuroscience Letters, 124, 256259. tion, 1, 7083.
Vaidya, C., Austin, G., Kirkorian, G., Ridlehuber, H., Desmond, J., Windischberger, C., Lamm, C., Bauer, H., & Moser, E. (2003).
Glover, G., et al. (1998). Selective effects of methylphenidate in Human motor cortex activity during mental rotation. Neuroimage,
attention deficit hyperactivity disorder: A functional magnetic 20, 225232.
resonance study. Proceedings of the National Academy of Sciences, Wong, R. O., & Lichtman, J. W. (2003). Synapse elimination. In L. R.
USA, 95(24), 1449414499. Squire, F. E. Bloom, S. K. McConnell, J. L. Roberts, N. C. Spitzer,
Vargha-Khadem, F., Gadian, D. G., & Mishkin, M. (2001). Dissocia- & M. J. Zigmond (Eds.), Fundamental neuroscience (2nd ed.,
tions in cognitive memory: The syndrome of developmental amne- pp. 533554). New York: Academic Press.
sia. Philosophical Transactions of the Royal Society of London, Woodruff-Pak, D. (1993). Eyeblink classical conditioning in HM:
Biological Sciences, 356(1413), 14351440. Delay and trace paradigms. Behavioral Neuroscience, 107,
Vargha-Khadem, F., Gadian, D. G., Watkins, K. E., Connelly, A., Van 911925.
Paesschen, W., & Mishkin, M. (1997). Differential effects of Woodruff-Pak, D., Logan, C., & Thompson, R. (1990). Neurobiologi-
early hippocampal pathology on episodic and semantic memory. cal substrates of classical conditioning across the lifespan. In A.
Science, 277(5324), 376380. Diamond (Ed.), The development and neural bases of higher cogni-
Vargha-Khadem, F., Salmond, C. H., Watkins, K. E., Friston, K. J., tive functions (Vol. 608, pp. 150173). New York: New York
Gadian, D. G., & Mishkin, M. (2003). Developmental amnesia: Academy of Sciences.
Effect of age at injury. Proceedings of the National Academy of Woodruff-Pak, D., Papka, M., & Ivry, R. (1996). Cerebellar involve-
Sciences, USA, 100(17), 1005510060. ment in eyeblink classical conditioning in humans. Neuropsychol-
Verfaellie, M., Koseff, P., & Alexander, M. P. (2000). Acquisition of ogy, 10, 443458.
novel semantic information in amnesia: Effects of lesion location. Wynn, K. (1992). Addition and subtraction by human infants. Nature,
Neuropsychologia, 38(4), 484492. 358, 749750.
Vingerhoets, G., de Lange, F. P., Vandemaele, P., Deblaere, K., & Wynn, K., Bloom, P., & Chiang, W.-C. (2002). Enumeration of col-
Achten, E. (2002). Motor imagery in mental rotation: An fMRI lective entities by 5-month-old infants. Cognition, 83(3),
study. Neuroimage, 17, 16231633. B55B62.
Volpe, J. J. (1995). Neurology of the newborn (3rd ed.). Philadel- Yancey, S. W., & Phelps, E. A. (2001). Functional neuroimaging and
phia: Saunders. episodic memory: A perspective. Journal of Clinical and Experi-
Volpe, J. J. (2000). Overview: Normal and abnormal human brain de- mental Neuropsychology, 23, 3248.
velopment. Mental Retardation and Developmental Disabilities Re-
Yang, T. T., Menon, V., Eliez, S., Blasey, C., White, C. D., Reid,
search Reviews, 6, 15.
A. J., et al. (2002). Amygdalar activation associated with posi-
Waber, D. P., Carlson, D., & Mann, M. (1982). Developmental and tive and negative facial expressions. NeuroReport, 13(14),
differential aspects of mental rotation in early adolescence. Child 17371741.
Development, 53, 16141621.
Zacks, J. M., Gilliam, F., & Ojemann, J. G. (2003). Selective distur-
Waber, D., Marcus, D., Forbes, P., Bellinger, D., Weiler, M.,
bance of mental rotation by cortical stimulation. Neuropsycholo-
Sorensen, L., et al. (2003). Motor sequence learning and reading
gia, 41, 16591667.
ability: Is poor reading associated with sequencing deficits? Jour-
nal of Experimental Child Psychology, 84(4), 338354. Zago, L., & Tzourio-Mazoyer, N. (2002). Distinguishing visuospatial
working memory and complex mental calculation areas within the
Warkany, J., Lemire, R. J., & Cohen, M. M. (1981). Mental retarda-
parietal lobes. Neuroscience Letters, 331, 4549.
tion and congenital malformations of the central nervous system.
Chicago: Year Book Medical. Zelazo, P. D., Frye, D., & Rapus, T. (1996). An age-related dissocia-
Webb, S. J., Monk, C. S., & Nelson, C. A. (2001). Mechanisms of tion between knowing rules and using them. Cognitive Develop-
postnatal neurobiological development in the prefrontal cortex ment, 11, 3763.
and the hippocampal region: Implications for human develop- Zola, S. M., Squire, L. R., Teng, E., Stefanacci, L., Buffalo, E. A., &
ment. Developmental Neuropsychology, 19, 147171. Clark, R. E. (2000). Impaired recognition memory in monkeys
Webb, S. J., & Nelson, C. A. (2001). Perceptual priming for upright after damage limited to the hippocampal region. Journal of Neuro-
and inverted faces in infants and adults. Journal of Experimental science, 20, 451463.
Child Psychology, 79, 122.
CHAPTER 2

The Infants Auditory World: Hearing, Speech,


and the Beginnings of Language
JENNY R. SAFFRAN, JANET F. WERKER, and LYNNE A. WERNER

INFANT AUDITION 59 Stress and Phonotactic Cues 80


Development of the Auditory Apparatus: Higher-Level Units 81
Setting the Stage 59 LEARNING MECHANISMS 81
Measuring Auditory Development 60 Units for Computations 83
Frequency Coding 61 BUILDING FROM THE INPUT DURING THE
Intensity Coding 63 1ST YEAR 83
Temporal Coding 67 Learning Phonology and Phonotactics 84
Spatial Resolution 69 Word Segmentation 84
Development of Auditory Scene Analysis 70 Beginnings of Word Recognition 88
Implications for the Development of Listening for Meaning 89
Speech Perception 71 Beginnings of Grammar 91
INFANT SPEECH PERCEPTION AND WORD CONCLUSIONS AND FUTURE DIRECTIONS 92
LEARNING: BEGINNINGS OF LANGUAGE 72 Relationship between Auditory Processing and
Emergence of the Field: Phonetic Perception 72 Speech Perception 92
A Preference for Speech 75 Constraints on Learning 93
Perception of the Visible Information in Speech 76 Domain Specificity and Species Specificity 93
Perception of Prosodic Attributes of the The Infants Auditory World 94
Speech Signal 77
REFERENCES 95
Perception of Other Aspects of the Speech Signal 78
IMPLICIT DISCOVERY OF CUES IN THE
INPUT: A DRIVE TO MAKE SENSE OF
THE ENVIRONMENT 80

The auditory world provides a rich source of informa- tion provides a channel for many important sources of
tion to be acquired by the developing infant. We are born inputs, including a variety of critical environmental
with well-developed auditory systems, capable of gath- sounds such as music and spoken language. For these
ering a wealth of knowledge even prior to birth. Audi- reasons alone, it has long been of great interest to char-
acterize the nature of this system as it develops.
We thank Dick Aslin, Suzanne Curtin, LouAnn Gerken, Lin- But studying how infants use their auditory environ-
coln Gray, Jim Morgan, and Erik Thiessen for their helpful ments can tell us more than just how these developmen-
comments on a previous draft. Support for the preparation of
tal processes unfold. In the past decade, studies of
this chapter was provided by grants from NIH (R01 HD37466)
infant audition, speech, and the beginnings of language
and NSF (BCS-9983630) to J. R. S., from the Natural Science
and Engineering Research Council of Canada, Social Sciences
have increasingly begun to bear on central debates in
and Humanities Research Council of Canada, The Canada Re- developmental cognitive science and cognitive neuro-
search Chair Program, and the Human Frontiers Science Pro- science. We are moving beyond such classic questions
gram to J. F. W., and from NIH (R01 DC00396, P30 DC04661) such as whether speech is special (i.e., subserved by a
to L. A. W. dedicated neural system that is not shared by other as-
This chapter is dedicated to the memory of Peter Jusczyk. pects of perception) to begin to study the actual learn-

58
Infant Audition 59

ing mechanisms underlying infants precocious acquisi- auditory processing will result in imprecise representa-
tion of the speech sounds of their native language. tions of speech as well as other sounds, and hence
Similarly, studies of the origins of infants linguistic limit the information that is available to the infant. In
knowledge have moved beyond descriptions of when this section, we consider whether limitations of audi-
infants know about various features of their native tory processing may serve to constrain early speech
language to studies that ask how that learning occurred. perception.
Increasingly, such behavioral studies are paired with The auditory system is designed to locate and iden-
research using psychophysiological methods to study tify sound sources in the environment. Sounds entering
the neural underpinnings of the behavior, experi- the ear are shaped by the structures of the outer ear to
ments using nonhuman animals to probe the species- optimize detection of relevant sounds and to allow de-
specificity of the behavior, and studies using materials termination of a sounds location in space. The sound
drawn from other domains to assess the domain-speci- is then analyzed into frequency bands by the inner ear.
ficity of the behavior. Periodicity, intensity and temporal fluctuations are
In this chapter, we review the state of the art in our represented within each band. This code provides the
field, using the ever-increasing interdisciplinarity of basis of all auditory perception, but the auditory sys-
research on infant audition, speech perception, and tem must calculate some sound characteristics from the
early language acquisition to highlight several themes. basic code. For example, the shape of a sounds spec-
One broad theme is the cause of developmental change. trum is extracted and differences between the ears are
Are changes due to maturation of central and/or pe- calculated in the auditory brainstem. Once all of this
ripheral neural structures? Or are they due to learning coding and calculation, referred to as primary process-
mechanisms, which are continually discovering com- ing, is completed, however, the system must still deter-
plex structure in the environment? A related broad mine which frequency bands emanate from a common
theme concerns the nature of these perceptual and source, on the basis of commonalities in frequency,
learning processes, and the extent to which they are periodicity, intensity, temporal fluctuations, location,
specifically tailored for a single task (e.g., learning and spectral shape. The latter stage of processing is
about speech) as opposed to available more generally known as sound source segregation or auditory scene
for learning across domains. We will also consider con- analysis (Bregman, 1990). Failure to segregate a sound
straints on perception and learningarising from our source from the background makes a listener less
perceptual systems, neural structures, species-specific sensitive to that sound; factors that promote sound
limitations on learning, and domain-specific limita- source segregation tend to make listeners more sensi-
tions on learningthat will help to inform our theories tive to a sound. These processes are likely to undergo
of how these processes are related to other aspects of important developmental change during infancy.
infant development. Finally, we will point to many of Finally, it is important to recognize that attention, mo-
the open questions that continue to drive research in tivation, memory and other cognitive processes influ-
this field, and which we hope to see answered in the ence auditory scene analysis, and in a very real way,
subsequent edition of this Handbook. hearing. These effects are described collectively as
processing efficiency, and they also contribute to au-
ditory development.
INFANT AUDITION
Development of the Auditory Apparatus:
Most infant auditory research has focused on infants Setting the Stage
perception of speech. In subsequent sections of this
chapter, we review many of these studies that demon- In humans, the inner ear begins to function during the
strate that even newborns are capable of making many second trimester of gestation. If humans are like other
phonetic and other speech distinctions. Clearly infants mammals, neural responses to sound are possible as soon
have the auditory capacity to represent some of the as the inner ear begins to transduce sound. The consensus
critical acoustic features of speech. Little is known, is that scalp-recorded auditory evoked potentials and be-
however, about the acoustic information that infants havioral responses can be observed in fetuses and in
use to make these fine-grained distinctions. Immature preterm infants as young as 28 weeks gestational age. The
60 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

possibility of prenatal hearing has important implications range in childhood (Yoshinaga-Itano, Sedey, Coulter, &
for understanding the effects of early experience on neu- Mehl, 1998).
ral development. Fetuses experience with sound is se-
verely limited by the sound transmission properties of Measuring Auditory Development
maternal tissue and amniotic fluid, by the conduction of
sound to the fetal inner ear, and by immaturities Sounds differ in three basic dimensionsfrequency, in-
of the inner ear and auditory nervous system (Smith, tensity, and changes in frequency and intensity over
Gerhardt, Griffiths, & Huang, 2003). Nonetheless, time. The auditory system encodes the frequency and in-
several studies have documented differential fetal re- tensity of sound and extracts information about tempo-
sponsiveness to sounds of different intensities and fre- ral variation. In addition, the auditory system calculates
quencies (e.g., Lecanuet, Gramer-Deferre, & Busnel, additional information about differences between
1988; Shahidullah & Hepper, 1994). Further, several sounds arriving at the two ears, or interaural differ-
studies have demonstrated that prenatal experience with ences. In psychoacoustics, auditory capacities are ap-
sound can influence later auditory responses, at least in proached from two directions. One is to describe the
the immediate postnatal period. This is most dramatically accuracy with which a dimension is coded, or resolution.
shown in the preference for mothers voice shown at birth The other is to describe the function relating the
(DeCasper & Fifer, 1980) and in the preference for a acoustic dimension to its perception. Both approaches
story and/or song heard prenatally (DeCasper & Spence, have been taken in characterizing hearing during in-
1986). At the same time, little is known about the impor- fancy, although the former is more straightforward in a
tance of prenatal experience with sound for auditory or nonverbal subject.
other aspects of development. Except as noted, the studies described in this section
At term, the neonate is believed to have a mature used one of three varieties of discrimination learning
inner ear ( but see Abdala, 2001; for example, Bargones procedures to estimate infants thresholds for detecting
& Burns, 1988; Bredberg, 1968). However, the con- or discriminating between sounds. Each depends upon
duction of sound through the external and middle ear teaching the infant that when a sound occurs or when a
to the inner ear is less efficient in neonates than sound changes in some way, a response will be rein-
in adults (Keefe, Bulen, Arehart, & Burns, 1993; forced by the presentation of an interesting audiovisual
Keefe, Burns, Bulen, & Campbell, 1994; Keefe et al., event. The common reinforcers are mechanical toys and
2000). The transmission of information through the au- video displays. Two of the procedures teach infants to
ditory neural pathway is slow and inefficient (e.g., make a head turn when the appropriate sound occurs. In
Gorga, Kaminski, Beauchaine, Jesteadt, & Neely, one variant, infants learn to turn toward the reinforcer
1989; Gorga, Reiland, Beauchaine, Worthington, & (e.g., Berg & Smith, 1983; Nozza & Wilson, 1984). In
Jesteadt, 1987; Ponton, Moore, & Eggermont, 1996). the other variant, sounds are presented from one of two
The implications of this pattern of immaturities for loudspeakers on a random schedule. The infant learns to
postnatal auditory development are discussed in the turn toward the speaker producing the sound (e.g., Tre-
sections that follow. hub, Schneider, & Edman, 1980). Infants older than
Approximately 2 to 3 in 1,000 infants are born with a about 6 months of age can be successfully tested using
hearing loss, 1 in 1,000 with a severe to profound hear- either of these procedures. However, younger infants do
ing loss. However, 20% to 30% of hearing-impaired chil- not make the crisp directional head turns that older in-
dren develop hearing loss postnatally. These children fants do. To get around this difficulty, observer-based
can be identified, even as neonates, with appropriate conditioning procedures capitalize on whatever re-
hearing screening (Norton et al., 2000), although in the sponse the infant makes to the sound (e.g., Tharpe &
recent past, the average age of identification of hearing- Ashmead, 2001; Werner, 1995). In this method, which
impaired children was 212 years. Disruption of nearly was originally developed to study infants visual acuity
all aspects of development, but particularly of language (Teller, 1979), an observer watching the infant knows
development, is typical in hearing-impaired children. when a sound may be presented but not whether it was in
Recent evidence suggests, however, that early identifica- fact presented. On the basis of the infants response, the
tion of hearing losswith intervention beginning prior observer must judge whether or not the sound was pre-
to 6 months of agefacilitates the development of lan- sented. The infant is reinforced for producing a response
guage skills (signed and/or spoken) within the normal that leads to a correct observer judgment. Infants as
Infant Audition 61

young as 1 month of age have been successfully tested in determining pitch, the perceptual dimension corre-
using this technique. lated with sound frequency. Further, in adults, it is the
Once children are 3 or 4 years old, they can be tested processing of sound in the inner ear that limits the repre-
using a variant of adult psychophysical procedures (e.g., sentation of complex sounds.
Wightman, Allen, Dolan, Kistler, & Jamieson, 1989).
Commonly, three intervals are presented to the child, Frequency Resolution
only one of which randomly contains the signal to be The most common way to assess the resolution of the
processed. The child is asked to choose the interval place code of frequency is to determine the frequencies
containing the different sound. The intervals can be of competing sounds that interfere with a listeners abil-
presented with cartoon indicators, making the whole ity to detect a frequency-specific sound. The phenome-
procedure more like a video game. non of one sounds increasing the difficulty of detecting
another is called masking. The interfering sound is the
Frequency Coding masker; the sound to be detected is called the signal or
probe. When masking occurs, the threshold ratio of
Frequency is coded in the auditory system by two mech-
probe to masker intensity required to detect the probe is
anisms. The basilar membrane in the inner ear vibrates
higher. It has long been known that one sound will only
in response to incoming sound, and because the stiffness
mask another if their frequencies are separated by less
of the membrane varies along its length, each position
than about a third of an octave (there are exceptions to
along the membrane responds maximally to a particular
this rule which are discussed below). It is fairly easy to
frequency. Hair cells are positioned along the length of
understand, then, that masking occurs when the masker
the basilar membrane. Outer hair cells provide mechani-
evokes activity in the same auditory nerve fibers that re-
cal feedback that results in higher amplitude and more
spond to the probe, and that masking provides a method
restricted, or sharper, basilar membrane responses to a
for assessing the quality of the place code for frequency.
given frequency. Each inner hair cell transduces basilar
Both behavioral and electrophysiological measures
membrane motion into a neural response in the auditory
indicate that frequency resolution measured using mask-
nerve fibers that exclusively contact it. Thus activity in
ing is immature at birth, but is mature by about 6 months
a particular auditory nerve fiber indicates the presence
of age. For example, Spetner and Olsho (1990) showed
of frequencies within a band about a third of an octave
that a 4,000 or 8,000 Hz tone was masked by a broader
wide. The frequency content of any sound, then, is repre-
range of frequencies for 3-month-olds than for 6-month-
sented in the pattern of activity across auditory nerve
olds and adults. Three-month-olds frequency resolution
fibers innervating different positions along the basilar
was mature only at 1,000 Hz, and 6-month-olds demon-
membrane. This neural representation of sound is re-
strated mature frequency resolution at all frequencies.
ferred to as the place code. Because the basilar mem-
Studies by Schneider, Morrongiello, and Trehub (1990)
brane vibrates at the frequency of stimulation, the
and Olsho (1985) confirmed that frequency resolution
action potentials in auditory nerve fibers tend to occur
was mature at 6 months of age. Although a few studies
at the frequency of stimulation. Thus, for frequencies
purported to show immature frequency resolution at 4
below 5,000 Hz, the intervals between action potentials
years of age (Allen, Wightman, Kistler, & Dolan, 1989;
provide another code for frequency.1 This phenomenon is
Irwin, Stillman, & Schade, 1986), Hall and Grose
known as phase locking, and it provides the basis for the
(1991) subsequently showed that children of this age had
temporal code for frequency. The bulk of evidence from
mature frequency resolution when thresholds were ap-
adults indicates that both frequency codes are involved
propriately measured. Thus, frequency resolution at low
1
frequencies appears to be mature by birth. At high fre-
Single auditory nerve fibers do not fire on every cycle of a
quencies, frequency resolution becomes adultlike some
continuing periodic sound; at frequencies above about 1,000
time between 3 and 6 months.
Hz, single nerve fibers cannot fire fast enough to provide a
temporal code for frequency. However, each nerve fiber re-
Lack of development of the auditory nervous system
sponds at the same phase of the sound and different nerve appears to be responsible for early immaturity of fre-
fibers randomly respond on different cycles. By combining quency resolution. The consensus is that the inner ear
the responses of many auditory nerve fibers responding to the mechanisms responsible for frequency resolution are
same frequency, a code for frequency can be derived up to mature at birth ( but see Abdala, 2001; for example, Bar-
5,000 Hz. gones & Burns, 1988; Bredberg, 1968). However, the
62 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

mature frequency resolution established in the inner the greatest changes in low-frequency discrimination
ear is not faithfully transmitted through the auditory appear to occur between 4 and 6 years of age.
nervous system. Like the studies based on behavioral The change in high-frequency discrimination be-
measures, several studies based on brainstem evoked- tween 3 and 6 months is consistent with the improve-
potential measures of frequency resolution report matu- ment in frequency resolution observed at this age. The
rity at low frequencies, but not at high frequencies for nature of the prolonged developmental course for low
3-month-old infants (Abdala & Folsom, 1995a, 1995b; frequency discrimination is less obvious. It has been
Folsom & Wynne, 1987). By 6 months, these measures suggested that adults use the temporal code to represent
indicate mature resolution across the frequency range. low frequency tones and the place code to represent high
The parallels between behavioral and neural evoked po- frequency tones (B. C. J. Moore, 1973). One possible ex-
tentials results suggest a neural basis for the immaturi- planation for poor low-frequency discrimination by in-
ties observed early in infancy. fants and children is that they do not use the temporal
code in pure-tone frequency discrimination or that they
Frequency Discrimination
use the temporal code inefficiently. However, Allen,
Frequency resolution is a measure of the precision of the Jones, and Slaney (1998) reported that, compared to
place code for frequency. Discrimination between adults, 4-year-olds detection is more dependent on the
sounds on the basis of frequency, however, is accom- periodicity, or pitchiness of a tone. The temporal code
plished via both the place code and the temporal code is the basis of that sound quality (B. C. J. Moore, 1996).
for frequency. Despite the fact that 6-month-olds have In addition 7-year-olds pure tone frequency discrimina-
mature frequency resolution, frequency discrimination tion is more affected by decreases in tone duration than
remains immature, at least at low frequencies. Olsho is that of adults, which would suggest that children
(1984) first reported that 6-month-olds needed about are more dependent on the temporal code (Thompson
twice the frequency change that adults did to detect a et al., 1999). However, by 7 years of age, children are
change in frequencies below 2,000 Hz. At 1,000 Hz, sev- also good at low-frequency discrimination (Maxon &
eral studies have estimated that 6-month-old infants can Hochberg, 1982). The other possible explanation for
detect a 1.5% to 3% change in frequency, while adults poor low-frequency discrimination is that it takes longer
can detect a change of 1% or less (Aslin, 1989; Olsho, to learn to discriminate between low frequencies than
1984; Olsho, Koch, & Halpin, 1987; Olsho, Schoon, high (Demany, 1985; Olsho, Koch, & Carter, 1988). If
Sakai, Turpin, & Sperduto, 1982; Sinnott & Aslin, infants and young children generally take longer to learn
1985). At higher frequencies, Olsho (1984) reported that a task than adults do, then they might be at a particular
6-month-olds detected frequency changes as well as disadvantage in learning to discriminate between low
adults did. Sinnott and Aslin (1985) and Olsho et al. frequencies. An analysis of infant frequency discrimina-
(1987) reported results generally consistent with that tion by Olsho, Koch, and Carter (1988) suggests that
pattern. Olsho et al. also tested 3-month-old infants and training effects might account for some, but not all, of
found that they performed similarly to 6-month-olds at the difference between infants and adults in low-
low frequencies, but that they had higher frequency dis- frequency discrimination
crimination thresholds at high frequencies.
Several studies of preschool and school-age children Perception of Pitch
show that pure-tone frequency discrimination is not The relative importance of the temporal and place codes
adultlike until 10 years of age (Jensen & Neff, 1993; has been debated extensively in the literature on com-
Maxon & Hochberg, 1982; Thompson, Cranford, & plex pitch perception. A complex tone consists of multi-
Hoyer, 1999). Maxon and Hochberg reported that the ple frequency components, a fundamental frequency and
discrimination of low frequencies was more immature harmonics. The perception of complex pitch is said to be
than that of high frequencies at 4 years, consistent with unitary. That is, although the pitch of a complex tone
the results of the infant studies. However, these investi- generally matches the pitch of its fundamental, the com-
gators did report some improvement in discrimination of plex is perceived as having a single pitch, and the higher
high frequencies between 4 and 12 years. Nonetheless, harmonics contribute to that percept (B. C. J. Moore,
the greatest changes in high-frequency discrimination 1996). Clarkson and her colleagues have carried out an
appear to occur during the first 6 months of life, while impressive series of studies of infants perception of
Infant Audition 63

complex pitch. In many respects, complex pitch percep- level of performance. It is difficult to distinguish inten-
tion in 7- to 8-month-olds appears to be adultlike: In- sity coding effects from processing efficiency effects on
fants are able to categorize complexes on the basis of an absolute measure. Few experiments have been carried
fundamental frequency, even when the fundamental fre- out that allow for this distinction, and a major question
quency component is missing from the complex (Clark- in this area is the relative contributions of auditory
son & Clifton, 1985; Montgomery & Clarkson, 1997). capacities and processing efficiency to age-related
Infants have difficulty categorizing inharmonic com- changes in thresholds.
plexes on the basis of pitch, as do adults (Clarkson & Intensity resolution is typically measured psy-
Clifton, 1995). However, when only high-frequency har- chophysically by finding the smallest intensity change
monics are present, adults are still able to hear the pitch that a listener can detect. When the change detected is
of the missing fundamental, while infants are not (Clark- from no sound to sound, we say we are measuring
son & Rogers, 1995). Because periodicity in the wave- absolute sensitivity. When change is detected in an audi-
form of combined high-frequency harmonics provides ble sound, we say we are measuring intensity discrimi-
the basis of this percept in adults, Clarkson and Rogers nation, increment detection or masking. In the classic
result suggests, again, that infants have difficulty using intensity discrimination paradigm, a listener hears two
the temporal code in pitch perception. or more sounds, and responds to the more intense. In in-
crement detection, the background sound is continuous;
Intensity Coding the listener responds when the intensity of the back-
ground increases. Simultaneous masking is a special
The primary code for intensity in the auditory system is case of intensity discrimination or increment detection
the firing rate of auditory nerve fibers. There are sev- in that the addition of a signal to the masker is detected
eral studies of developing nonhumans that suggest that as an increase in the intensity of the stimulus. All of
immature neurons cannot sustain a response over time these measures largely depend on the same underlying
and that the maximum firing rate achieved by auditory processes, so the expectation is that they will develop
nerve fibers increases with development (Sanes & along a similar course, with one exception: The noise
Walsh, 1998). In humans, evoked potential amplitude in- that limits absolute sensitivity is neural and physiologi-
creases more slowly with increasing intensity in infants cal noise. It is not conducted through the external and
than in adults (Durieux-Smith, Edwards, Picton, & Mc- middle ear. Immaturity of the conductive apparatus will
Murray, 1985; Jiang, Wu, & Zhang, 1990). Further, be- affect the level of a signal played into the ear, but not the
cause the external and middle ear grow during infancy level of the background when absolute threshold is mea-
and childhood, the conduction of sound to the inner sured. In masking, intensity discrimination and incre-
would be expected to improve with age. Thus, there is ment detection, the conductive apparatus affects both
reason to believe that intensity processing would un- the signal and the background, leaving the signal-to-
dergo postnatal developmental change in humans. noise ratio unchanged. Thus, conductive immaturity will
be reflected in absolute thresholds, but not the other
Intensity Resolution
measures of intensity resolution.
Some sensory processes can be measured by comparing The most commonly measured aspect of intensity
thresholds across masking conditions. In the studies of processing is the absolute threshold, the intensity of
infant frequency resolution described earlier (Olsho, sound that is just detectable in a quiet environment.
1985; Schneider et al., 1990; Spetner & Olsho, 1990), 6- Some studies have measured absolute thresholds in in-
month-olds thresholds for detecting a tone were always fants 3 months and younger. Weir (1976, 1979) esti-
higher than those of adults, but their thresholds changed mated the behavioral threshold of neonates, based on
as the frequency or bandwidth of the masker changed their spontaneous responses to tones. The thresholds she
exactly as adults thresholds did. It is the pattern of measured ranged from 68 dB SPL at 250 Hz to 82 dB
change in threshold that indicates resolution. Making it SPL at 2,000 Hz, approximately 30 and 70 dB higher
difficult to judge intensity resolution to be immature is than adult thresholds at these frequencies, respectively.
the fact that measures of intensity resolution tend to be Ruth, Horner, McCoy, and Chandler (1983) and Kaga
absolute measures; they do not depend on a comparison and Tanaka (1980) reported behavioral observation au-
of performance across conditions, but on the absolute diometry thresholds for 1-month-olds that are similar to
64 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

those reported by Weir for neonates. However, thresh- crimination than are adults. Few data are available for
olds measured at 1 to 2 months of age using observer- infants younger than 6 months of age, but there is evi-
based procedures are quite a bit lower, about 40 to 55 dB dence that newborns respond to an intensity change in a
SPL, 35 to 45 dB higher than adults thresholds (Tharpe speech sound as small as 6 dB (Tarquinio, Zelazo, &
& Ashmead, 2001; Trehub, Schneider, Thorpe, & Judge, Weiss, 1990). By 7 to 9 months, Sinnott and Aslin (1985)
1991; Werner & Gillenwater, 1990; Werner & Mancl, found that infants detected intensity differences of 6 dB
1993). The difference between 1-month-olds and between 1,000 Hz tones, while adults could detect dif-
adults threshold is about 10 dB greater at 500 Hz than ferences of about 2 dB. Kopyar (1997) reported that in-
at 4,000 Hz (Werner & Gillenwater, 1990). Whether in- fants of this age detected differences of 9 dB between
fants sensitivity actually improves by 25 dB between tones or between broadband noises. Adults detected dif-
birth and 1 month is not clear, given the differences in ferences of about 4 dB between tones, but about 3 dB be-
the procedures used to assess threshold. tween noises.
By 3 months, thresholds appear to improve by about Intensity discrimination has not been examined in
10 dB at 500 Hz and by nearly 20 dB at 4,000 Hz (Olsho, children between 9 months and 4 years of age. Maxon
Koch, Carter, Halpin, & Spetner, 1988). Compared to and Hochberg (1982) tested intensity discrimination of
adults, 3-month-olds are still about 5 dB less sensitive at tones in children older than 4 years. They found a
the higher frequency. A longitudinal, observer-based steady improvement in the discrimination threshold
study of infants detection thresholds for a broad noise from about 2 dB at age 4 years to about 1 dB at 12
band confirmed an improvement of about 15 dB between years, when the level of tones was near 60 dB above the
1 and 3 months (Tharpe & Ashmead, 2001). Between 3 childs absolute threshold. Thus, by 4 years, intensity
and 6 months, very little improvement is observed in the discrimination appears to be quite good. Only minor
500-Hz threshold, but a further 15 dB improvement is improvements occur thereafter, at least for tones well
seen in the 4,000-Hz threshold (Olsho, Koch, Carter, above absolute threshold.
et al., 1988). Tharpe and Ashmead observed about a 15 Increment detection matures somewhat earlier than
dB improvement in threshold for a noise band between 3 does discrimination between discrete sounds. Several
and 6 months. In the Olsho et al. study, the performance studies have shown that 7- to 9-month-olds can detect 3
difference between 6-month-olds and adults at 4,000 Hz to 5 dB increments in a broad noise band (Berg &
is about 15 dB, while at 500 Hz it is about 20 dB. There Boswell, 1998; Kopyar, 1997; Werner & Boike, 2001),
is general consensus that in the vicinity of 1,000 Hz, 6- under conditions in which adults detect increments of 1
month-olds are about 15 dB less sensitive than adults to 2 dB. Schneider, Bull, and Trehub (1988) reported
(Berg & Smith, 1983; Nozza & Wilson, 1984; Olsho, that 12-month-olds could detect 3 dB increments in a
Koch, Carter, et al., 1988; Ruth et al., 1983; Sinnott, continuous broadband noise, while adults could detect
Pisoni, & Aslin, 1983; Tharpe & Ashmead, 2001; Tre- increments less than 1 dB. Berg and Boswell (2000)
hub et al., 1980). measured increment detection thresholds in 1- and 3-
More extensive work has documented absolute sensi- year-old children, for a 2-octave wide noise band cen-
tivity from 6 months to adulthood. Trehub, Schneider, tered at 4,000 Hz. Their results for 1-year-olds are
Morrengiello, and Thorpe (1988) measured thresholds similar to those reported by Schneider et al.; 3-year-olds
for noise bands centered at different frequencies for lis- appeared to be adultlike in this task.
teners ranging from 6 months of age through the school The only thresholds for detection of an increment in
years to adulthood. They report that threshold improves a tone were reported by Kopyar (1997). Infants do rela-
by about 25 dB at 400 Hz, about 20 dB at 1,000 Hz, but tively worse in detecting tone increments than noise in-
only 10 dB at 10,000 Hz. Further, the higher the fre- crements, requiring increments of 8 dB, compared to 2
quency, the earlier adult levels are achieved: 10 years dB for adults. A number of studies have examined the
or later at 1,000 Hz, but before 5 years of age at 4,000 development of detection of a tone or narrow noise
and 10,000 Hz. band masked by a noise, essentially the detection of an
The ability to detect a change in the intensity of an increment in an ongoing sound. Schneider, Trehub,
audible sound is frequently measured by intensity dis- Morrongiello, and Thorpe (1989) estimated masked
crimination, that is, by asking the listener to respond to thresholds for a 1-octave band noise, centered at fre-
an intensity difference between sounds. Several studies quencies ranging from 400 to 10,000 Hz, masked by a
of infants indicate that they are poorer at intensity dis- broadband noise, in children ranging from 6 months to
Infant Audition 65

10 years of age and adults. At 6 months, the infants de- some means. Children as young as 5 years of age are able
tection threshold was equivalent to a 7 dB increment; to rate the loudness of tones numerically and with line
adults detected a 1 dB increment.2 The age difference length. Moreover, loudness appears to grow with in-
was about the same across the frequency range. Rela- creasing intensity in the same way for children and
tively large improvements in performance were re- adults (Bond & Stevens, 1969; Collins & Gescheider,
ported between 6 and 18 months and between 4 years 1989). The evidence on loudness growth in infants is
and 8 years. There was little difference between 10- sparse. Leibold and Werner (2002) examined the rela-
year-olds and adults. Other studies have examined in- tionship between intensity and reaction time in 7- to 9-
fants and childrens detection of tones in noise (Allen month-olds and adults. Reaction time decreased with
& Wightman, 1994; Bargones, Werner, & Marean, increasing sound intensity in both age groups, but the
1995; Berg & Boswell, 1999; Nozza & Wilson, 1984). rate of decrease was greater for infants than for adults.
They report similar results, although it is clear that This finding suggests that loudness grows more rapidly
there is considerable variability across children in this with increasing intensity in infants, but again, the impli-
task (e.g., Allen & Wightman, 1994). cations of this finding for early audition are not clear.
In summary, absolute threshold, intensity discrimina-
Perception of Timbre tion, detection of tones masked by noise and spectral
Timbre, or sound quality, is determined by the relative shape discrimination all undergo relatively large age-
amplitudes of the components of a complex sound, and related improvements during infancy and the preschool
thus involves the comparison of intensities across fre- years. However, adult levels of performance are not
quency. The physical dimension associated with timbre reached until 8 or 10 years of age. Interestingly, incre-
is referred to as spectral shape. Vowel perception and ment detection in broadband sounds appears to mature
sound localization depend on spectral shape processing. earlier, around 3 years of age. Nozza (1995; Nozza &
A few studies have examined the development of timbre Hensen, 1999) showed that the level at which a noise
perception. Seven-month-olds can discriminate between would just start to mask a tone was 8 dB more intense
sounds of different timbres, complex tones with the for infants than for adults. This clever experiment
same pitch that contain different harmonics (Clarkson, demonstrates that immature thresholds of 8- to 11-
Clifton, & Perris, 1988). Trehub, Endman, and Thorpe month-olds are due largely to changes in sensitivity,
(1990) also showed that infants could categorize tonal rather than performance factors. Several factors are
complexes on the basis of spectral shape. The sharp- known to contribute to age-related improvements in in-
ness of infants representation of spectral shape has not tensity processing. For example, the frequency response
been assessed. Allen and Wightman (1992) used a com- of the infant ear canal changes during infancy. While the
plex sound with a sinusoidal spectral shape to measure adult ear canal conducts sounds best in the range be-
childrens threshold for detecting changes in spectral tween 2,000 and 5,000 Hz, the infant ear canal conducts
shape. They were unable to elicit discrimination be- higher frequency sounds better (Keefe et al., 1994). Fur-
tween such complexes in 4-year-olds. Five and 7-year- ther, the efficiency with which the middle ear conducts
olds performed the task, but only 9-year-olds performed sound into the inner ear has been shown to increase over
as well as adults. These results suggest that spectral a long age period, from birth to perhaps 10 years of age
shape, or timbre, discrimination follows a long develop- (Keefe et al., 1993, 2000; Keefe & Levi, 1996; Okabe,
mental course. It is not clear that performance on this Tanaka, Hamada, Miura, & Funai, 1988). The largest
task generalizes to vowel perception. improvements occur during the 1st year of life, espe-
cially for frequencies over about 1,000 Hz. It has been
Perception of Loudness estimated that the efficiency of the conductive appara-
A final measure of intensity processing is loudness. In tus improves by as much as 20 dB at 3,000 Hz between
adults, loudness is measured by having listeners match birth and adulthood, with about half of that improve-
sounds in loudness or by having them rate loudness by ment occurring during infancy. Age-related improve-
ment in conductive efficiency is smaller at lower
2
If the threshold is expressed as the sound pressure level of frequencies, with about a 5 dB improvement between
the signal added to the masker, 6-month-olds threshold for birth and adulthood (Keefe et al., 1993). Thus, one fac-
detecting a noise band or a tone in noise is 8 to 10 dB higher tor in the development of absolute sensitivity is the de-
than adults. velopment of the conductive apparatus. It is likely that
66 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

much, but not all, of the improvement in high-frequency auditory cortex takes much longer to reach maturity,
absolute thresholds during the first 6 months of life is with anatomical and physiological changes occurring to
due to improvements in conductive efficiency. The adolescence (J. K. Moore, 2002; J. K. Moore & Guan,
conductive improvement is also reflected in thresholds 2001; Ponton et al., 2000). The extent to which these
for evoked neural responses (e.g., Lary, Briassoulis, de changes in the auditory nervous system are reflected in
Vries, Dubowitz, & Dubowitz, 1985; Sininger & Ab- intensity processing beyond infancy is not known.
dala, 1996; Sininger, Abdala, & Cone-Wesson, 1997). Finally, higher-level processes also contribute to age-
Development of the inner ear is probably not a con- related improvements in intensity processing. Such
tributing factor to postnatal auditory development. Most processes fall into the category of efficiency, as opposed
indications are that the inner ear is adultlike in structure to resolution, but nonetheless influence sensitivity. Sev-
by term birth (Bredberg, 1968; Fujimoto, Yamamoto, eral investigators have examined the age differences in
Hayabuchi, & Yoshizuka, 1981; Hoshino, 1990; Igarashi intensity processing, and concluded that only a small
& Ishii, 1979, 1980; Igarashi, Yamazaki, & Mitsui, 1978; portion of the difference in thresholds between infants
Lavigne-Rebillard & Bagger-Sjoback, 1992; Lavigne- and adults can be accounted for by simple lapses of at-
Rebillard & Pujol, 1987, 1988, 1990; Nakai, 1970; Pujol tention (Viemeister & Schlauch, 1992; Werner, 1992;
& Lavigne-Rebillard, 1992). Otoacoustic emissions3 may Wightman & Allen, 1992). Recall, however, that infants
have a higher amplitude in infants and a higher stimulus and children tend to do quite well in detecting incre-
level may be required to elicit emissions, but they are ments in a broadband noise. Werner and Boike (2001)
qualitatively adultlike (Bonfils, Avan, Francois, Trotoux, showed that 7- to 9-month-old infants were more adult-
& Narcy, 1992; Bonfils, Francois, et al., 1992; Brown, like in their detection of an increment in a broadband
Sheppard, & Russell, 1994; Burns, Campbell, & Arehart, noise than they were in detecting a tone masked by a
1994). Abdala and her colleagues (Abdala, 1998, 2001; broadband noise. Werner and Boike argued that because
Abdala & Chatterjee, 2003) have presented data on otoa- infants ultimately achieve the same asymptotic perfor-
coustic emissions from young infants that differ some- mance in detecting tones and noise, the difference can-
what from that seen in adults. However, at this point it is not be accounted for by differences in attentiveness to
not clear that these differences could not be accounted narrow-band and broadband sounds. Bargones and
for by the reduction in input to the cochlea because of Werner (1994) showed that adults tended to listen selec-
middle ear immaturity. tively for a tone at an expected frequency, with the
Werner and her colleagues (Werner, Folsom, & result that they did not hear tones at unexpected frequen-
Mancl, 1993, 1994) have reported that the time that it cies. Infants, in contrast, detected expected and unex-
takes for the neural response to sound to travel through pected frequencies equally well. This suggests that
the brainstem predicts a 3-month-olds absolute thresh- infants listen over a broad band of frequencies, even
old at 4,000 and 8,000 Hz. Thus, another factor in the when they are detecting a narrow-band sound. As a re-
early development of intensity processing is maturation sult, more of the background noise will interfere with
of the primary auditory nervous system. Anatomical and their detection and their thresholds will be higher (Bar-
electrophysiological studies suggest that the auditory gones et al., 1995; Dai, Scharf, & Buus, 1991). Because
brainstem continues to develop throughout infancy infants and adults both listen over a broad frequency
(Gorga et al., 1989; J. K. Moore, Guan, & Shi, 1997; range when detecting broadband sounds, infants will de-
J. K. Moore, Perazzo, & Braun, 1995; J. K. Moore, Pon- tect broadband sounds better, relative to adults. There is
ton, Eggermont, Wu, & Huang, 1996; Ponton, Egger- evidence that 6-year-olds detect expected frequencies
mont, Coupland, & Winkelaar, 1992; Ponton et al., better than unexpected frequencies (Greenberg, Bray, &
1996). Ponton et al. (1996) provide evidence that age-re- Beasley, 1970). The age at which this ability is acquired
lated change in evoked potentials during infancy results has not been established.
largely from increases in synaptic efficiency. Primary To summarize, maturation of the conductive appara-
tus, of primary auditory pathways, and of listening
3
Otoacoustic emissions are sounds that are produced in the strategies are important factors in the development of
ear and transmitted back out into the ear canal. The presence intensity processing. Infants failure to distribute their
of otoacoustic emissions is a good indicator of normal attention to the features of complex sound as adults do
cochlear function. may have additional implications. Given that infants
Infant Audition 67

have relatively little experience with complex sounds 20 ms (Jusczyk, Pisoni, Reed, Fernald, & Myers, 1983;
and dont know which features are most important, their Morrongiello & Trehub, 1987), a change on the order of
broadband approach seems sensible. An interesting 10%. Morrongiello and Trehub also found that 5- to 6-
question is how this broadband listening strategy influ- year-olds could discriminate 15 ms changes, while
ences what infants hear in a natural environment. adults discriminated 10 ms changes in the same condi-
tions. At the same time, two studies of children 4- to 10-
Temporal Coding years-old, using an oddity task (Which of these three
sounds is different?), report that 4-year-olds may need
As Viemeister and Plack (1993) noted the temporal at least a 50% change in the duration of a 300 or 400 ms
pattern of spectral changes is, essentially, the informa- sound to detect the change (Elfenbein, Small, & Davis,
tion substrate of speech and other communication sig- 1993; Jensen & Neff, 1993). Elfenbein et al. found that
nals (p. 116). Temporal resolution is defined as the duration discrimination did not reach adult levels of per-
precision with which a listener can follow rapid changes formance until 10 years of age. There are many differ-
in the intensity or frequency of sound over time. Adult ences among these studies, most notably in the methods
listeners process sounds through a running temporal used to estimate a threshold, making it extremely diffi-
window, which averages the input over about 8 ms cult to decide how good duration discrimination really
(B. C. J. Moore, Glasberg, Plack, & Biswas, 1988). is at a given age.
However, adults also demonstrate temporal integration: The earliest studies of temporal resolution examined
They can integrate or otherwise combine a series of gap detection in infants and in children. In a gap detec-
these 8-ms looks at a sound over 200 to 300 ms while tion task, the duration of an interruption, or gap, in a
maintaining the temporal detail with 8 ms resolution sound is manipulated to find the shortest gap duration
(Viemeister, 1996; Viemeister & Wakefield, 1991). As that can be detected. Werner et al. (1992) reported that
noted, early in development auditory neurons often do the gap detection thresholds in a continuous noise for 3-,
not maintain their reponse to ongoing sound, at least in 6-, and 12-month-old infants were an order of magnitude
nonhuman species (Sanes & Walsh, 1998). Moreover, it worse than those of adults, around 50 compared to 5 ms.
is well established that auditory evoked responses from Trehub, Schneider, and Henderson (1995) reported better
infants and children are more susceptible to adaptation, performance for 6- and 12-month-olds when the gap to
or the effects of prior stimulation, than responses from be detected was between two short tone pips. However,
adults. (Fujikawa & Weber, 1977; Fujita, Hyde, & Al- infants gap detection thresholds were still around 30
berti, 1991; Jiang et al., 1990, 1991; Klein, Alvarez, & ms. Six-month-olds barely achieved 70% correct for gap
Cowburn, 1992; Lasky, 1984, 1991, 1993, 1997; Lasky durations of 28 and 40 ms. Wightman and his colleagues
& Rupert, 1982; Mora, Exposito, Solis, & Barajas, (1989) reported that gap detection thresholds remained
1990; Plessinger & Woods, 1987). Adaptation effects immature at 3.5 years, with the threshold for detecting a
decline with age, but may be seen in the auditory brain- gap in a noise band centered at 2,000 Hz at about 10 ms.
stem response as late as 3 years of age (Jiang et al., Trehub et al. and Wightman et al. both reported that gap
1991). More centrally generated evoked potentials may detection thresholds were mature by 5 years of age. In
exhibit even more dramatic adaptation effects in infants adults, gap detection performance tends to be better for
and children (e.g., Mora et al., 1990). Such effects may high-frequency sounds than for low (Eddins & Green,
be reflected in a reduced ability to follow rapidly chang- 1995). Both infants and children demonstrate adultlike
ing stimuli. How temporal integration might be affected frequency effects on gap detection (Werner et al., 1992;
is unclear. Wightman et al., 1989).
Conclusions about the development of temporal reso- Trehub et al. (1995) suggested that the reason that in-
lution depend strongly on the measure chosen to de- fants had somewhat better thresholds for detecting a gap
scribe it. Further, the interpretation of age effects that between two short sounds than they did for detecting a
have been demonstrated is not always obvious. A good gap in a continuous sound was that the continuous sound
example is the development of duration discrimination. created excess adaptation in the immature auditory sys-
At least two studies indicated that infants as young as 2 tem. If the neural response is low at the onset of the gap,
months of age could discriminate a change in the dura- because of adaptation, it will be more difficult to detect
tion of a repeated 200- or 300-ms-long sound of about the gap. Further, if the immature system takes longer to
68 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

recover from adaptation, the response at the offset of the 5 dB or so higher than those of adults, that still means
gap will be reduced. A psychophysical paradigm for that they are exhibiting 30 dB more masking than
measuring adaptation is forward masking: a relatively adults. Other studies confirm that children at this age
intense sound, the masker, is presented, quickly fol- are more susceptible to backward masking than adults
lowed by a probe sound. If the interval between the ini- (Buss et al., 1999; Rosen, van der Lely, Adlard, & Man-
tial and following sound is less than about 100 ms, the ganari, 2000). Werner (2003) has also reported that 7-
audibility of the probe is reduced relative to the un- and 11-month-olds have higher backward masked
masked condition. The development of forward masked thresholds than adults, although the age difference in
thresholds has been examined in infants and in children. amount of masking is not clear. A recent study, however,
Werner (1999) measured forward masked thresholds for suggests that at least among children, the apparent sus-
a 1-kHz tone in 3- and 6-month-old infants at masker- ceptibility to backward masking may not reflect imma-
probe intervals ranging from 5 to 100 ms. Her results turity of temporal resolution. Hartley and Moore (2002)
showed that the amount of forward masking decreased showed that a listener with normal temporal resolution,
as the interval increased, and in the same way, for each but poor processing efficiency, will be relatively more
age group. The audibility of the probe tone was affected susceptible to backward masking than to forward or si-
more for the 3-month-olds than it was for older listeners. multaneous masking. It should be noted that a similar
However, 6-month-olds were more or less adultlike in conclusion could be drawn about the nature of the per-
the amount of forward masking demonstrated at all in- ceptual deficit associated with language impairment.
tervals. Thus, by this measure of temporal resolution, 3- The gold standard of temporal resolution measures is
month-olds are immature, but 6-month-olds are not. the temporal modulation transfer function (TMTF,
This conclusion argues against the idea that adaptation Viemeister, 1979). Listeners are asked to detect ampli-
effects are responsible for variation in gap detection tude modulation (AM) in a sound. The depth, or amount,
thresholds of 6-month-olds. Buss, Hall, Grose, and Dev of modulation is manipulated to define the threshold for
(1999) also report that the amount of forward masking AM detection. AM detection threshold is estimated over
demonstrated by 5- to 11-year-old children is adultlike. a range of modulation frequencies. For adults, the result
A measure of temporal resolution that has received is a function, the TMTF, with a low-pass characteris-
considerable attention from developmentalists in recent tic: AM detection threshold is fairly constant from
years is backward masking, in which a probe tone is a 4-Hz modulation rate to about 50 Hz. Beyond 50 Hz
masked by a relatively intense masker that follows it by modulation rate, the AM detection threshold grows
a short interval (0 to 50 ms). Tallal and her colleagues poorer at rate of about 3 dB per doubling of modulation
(Tallal, Miller, Jenkins, & Merzenich, 1997; Tallal & frequency. The modulation frequency at which AM de-
Piercy, 1973, 1974) have long argued that a deficit in au- tection begins to deteriorate is taken as the measure of
ditory temporal resolution is the underlying cause of temporal resolution. Hall and Grose (1994) described
specific language impairment. Wright et al. (1997) col- the TMTF of 4- to 10-year-old children. The AM detec-
lected psychophysical data from children with language tion threshold of 4- to 7-year-olds was poorer than that
impairment and typically developing children. They of adults across the range of modulation rates; 9- to 10-
found that children with language impairment did not year-olds were adultlike in this respect. However, the
differ from typically developing children in simultane- shape of the TMTF was the same for all ages; AM detec-
ous masked thresholds. Their thresholds were a little tion began to deteriorate at about 50 to 60 Hz in all age
higher in forward masking, but their thresholds were groups. When the TMTF becomes mature is uncertain.
considerable higher in backward masking. This finding Levi and Werner (1996) reported AM detection thresh-
was taken to support Tallals position and has spurred olds of 3-month-olds, 6-month-olds and adults at two
research in this area. It has also spurred an interest in modulation rates, 4 and 64 Hz. The difference between
the development of backward masking, which had thresholds at the two modulation rates for 3- and 6-
heretofore not been examined. For example, Hartley, month-olds was 3 dB. This difference suggests that in-
Wright, Hogan, and Moore (2000) reported 6-year-olds fants have an adultlike TMTF and mature temporal
backward masked threshold for a 1,000-Hz tone to be 34 resolution.
dB higher than adults. At 10 years, backward masked The development of temporal integration has also
thresholds were nearly 20 dB higher than those of been of recent interest. In adults, increasing the duration
adults. Even if the absolute threshold of 6-year-olds are of a sound by a factor of 10 leads to a little less than 10-
Infant Audition 69

dB decrease in the absolute threshold for that sound. 1981; Morrongiello, 1988; Morrongiello, Fenwick, &
This means that adults are integrating information about Chance, 1990; Morrongiello, Fenwick, Hillier, &
the sound over time nearly perfectly. Sound energy can- Chance, 1994; Morrongiello & Rocca, 1987a, 1990).
not be integrated, however, over intervals longer than The MAA is adultlike, 1 to 2, in 5-year-olds. The
200 to 300 ms. Several studies of infants have reported MAA in elevation decreases from a value greater than
that the maximum interval over which infants integrate 16 at 6 to 8 months to about 4 at 18 months, which is
sound energy in detection is similar to the adult value comparable to the adult MAA in elevation (Mor-
(e.g., Berg & Boswell, 1995; Thorpe & Schneider, rongiello & Rocca, 1987b, 1987c). In adults, the MAA
1987). However, it was also reported that increasing the in azimuth is generally smaller than that in elevation,
duration of sound had a much greater than expected ef- because additional, binaural, cues can be used to local-
fect on infants absolute threshold. For example, Thorpe ize sounds in azimuth. Interestingly, during infancy the
and Schneider found that increasing the duration of a MAA in azimuth is similar to that in elevation (Mor-
noise band by a factor of 6.3 leads to a 20-dB decrease rongiello & Rocca, 1987b, 1987c). That the MAA is
in 6- to 7-month-olds absolute threshold. Berg and similar in the two dimensions suggests that infants may
Boswell argued that infants temporal integration was rely more heavily on spectral shape in sound localiza-
mature, but that infants had difficulty detecting short tion than on binaural differences. Finally, several stud-
duration sounds (see also Bargones et al., 1995). Maxon ies have suggested that infants are sensitive to the
and Hochberg (1982) reported temporal integration data changes in sound intensity that signal a change in sound
for 4- to 10-year-olds. For durations of 50 ms and source distance (Clifton, Perris, & Bullinger, 1991;
longer, thresholds decreased with increasing duration at Morrongiello, Hewitt, & Gotowiec, 1991). The accuracy
an adultlike rate, and thresholds leveled off between 200 with which infants can judge sound source distance has
and 400 ms duration. The only difference between chil- not been examined.
dren and adults was at quite short durations: increasing Humans and other mammals base their judgments of
the duration from 25 to 50 ms leads to a 7-dB decrease a sound sources location on information that first
in threshold at 4 years. By 12 years, the decrease is only reaches their ears; they are able to suppress information
5 dB, but still greater than expected in adults. Thus, as carried in echoes of the original sound. This effect is
children get older they appear to be able to deal with known as the precedence effect. It is known that infants
progressively shorter sounds. The nature of the immatu- also demonstrate this effect (Clifton, Morrongiello, &
rity is not clear; Berg and Boswell suggest that it actu- Dowd, 1984). Interestingly, Litovsky (1997) found that
ally could result from immaturity in the growth of while sound localization is influenced to some degree
neural response with intensity (Fay & Coombs, 1983) or by the presence of echoes in adults, 5-year-olds are
that the immature auditory system is less able to process more affected. This difference suggests that while the
onset responses and hence, transient stimuli. traditional MAA is mature by 5 years, sound localiza-
tion in real environments may continue to be refined be-
Spatial Resolution yond that age.
The mechanisms underlying the development of
Locating sound sources in space involves several sound localization are not completely understood. One
processes including evaluation of spectral shape and in- obvious change that will influence this ability is the
tensity, as well as binaural comparisons. Under normal growth of the head and external ear. As the head grows,
circumstances, spectral shape is the primary cue to po- the size of interaural differences increases (Clifton,
sition in elevation, while binaural time and intensity dif- Gwiazda, Bauer, Clarkson, & Held, 1988), and discrim-
ferences are the primary cues to position in azimuth (the ination of interaural time differences has been shown to
plane that runs through your ears parallel to the ground). improve during infancy (Ashmead, Davis, Whalen, &
Development of the ability to use these cues has been Odom, 1991). Ashmead et al. (1991), however, showed
well studied in infants via measurements of the mini- that immaturity of interaural discrimination does not
mum audible angle (MAA), the threshold for detecting a appear to be great enough to account for early immatu-
change in the position of a sound source. The MAA in rity of sound localization. One possible explanation for
azimuth has been shown to decrease from about 27 at 1 early sound localization immaturity is that infants are
month to less than 5 at 18 months (Ashmead, Clifton, & more dependent on spectral shape than interaural differ-
Perris, 1987; Clifton, Morrongiello, Kulig, & Dowd, ences in determining a sound source location. Infants
70 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

ability to use spectral shape cues will also improve as of the tones in this sequence was reversed, adults had no
the external ear grows. Another explanation is that in- trouble reporting the change. In another sequence, two
fants can process the cues to sound location adequately, pairs of near-frequency tones were repeated. Adults
but that they have not yet developed the ability to trans- heard this sequence as coming from two sources, each of
late a set of acoustic cues into a precise location in space which produced two alternating tones. When the order
(Gray, 1992). In other animals, it has been shown that of this sequence was reversed, adults had difficulty
multimodal experience is required to grow such a map hearing the change. Demany tested 2- to 4-month-olds
of sensory space (e.g., Binns, Withington, & Keating, ability to discriminate an order change in these two
1995; King, Hutchings, Moore, & Blakemore, 1988). sequences, using a habituation /dishabituation task
Further, humans who have only monaural hearing early wherein looking time was the dependent variable. In-
in life may be able to discriminate interaural differences fants appeared to discriminate the order change in the
normally when hearing is restored to the previously de- first sequence, but not in the other sequence, paralleling
prived ear, but still be unable to locate sounds in space adults perception. This result suggests that infants can
(Wilmington, Gray, & Jahrsdorfer, 1994). organize sounds on the basis of frequency.
Besides allowing us to localize sounds with precision, Demanys (1982) study has been criticized on
interaural differences improve our sensitivity to sound. methodological grounds. It is possible to discriminate
In the laboratory this improvement in sensitivity is some of the sequences he used from their reversed ver-
called the masking level difference (MLD): Threshold sion on the basis of the frequency contour, even if the
for a tone presented to both ears is lower if there are in- sequence is not perceived as two parallel streams. Fass-
teraural differences in the masker presented to the two bender (1993) corrected this problem and tested 2- to
ears. It appears that infants derive less benefit from in- 5-month-olds on sequences that adults organized on the
teraural differences than adults do, and that by 5 years basis of frequency, amplitude or timbre. Infants dis-
of age the MLD is adultlike (Hall & Grose, 1990; criminated order changes in the sequences as adults
Nozza, 1987). However, 5-year-olds may still derive less did, supporting the idea that infants group sounds at
benefit from listening with two ears when the listening least qualitatively like adults. In addition, McAdams
situation is complex (Hall & Grose, 1990). and Bertoncini (1997) tested 3- to 4-day-old infants on
sequences that adults segregated on the basis of both
Development of Auditory Scene Analysis location and timbre. Again, infants discriminated order
reversals as adults did, although it is not clear whether
Once the auditory system has analyzed incoming sound the sequences were organized by location, by timbre, or
and extracted information about its spectral shape, tem- by both location and timbre. In this test paradigm,
poral fluctuations and location, it remains to resynthe- note, listeners are never asked to segregate simultane-
size the auditory scene. Information in different ously occurring sounds, as most frequently occurs in
frequency bands must be grouped according to source natural environments. Thus, the conclusions that can be
on the basis of the initial analysis. Moreover, once the drawn about infants sound source segregation are cur-
scene has been reconstructed, the listener may choose to rently limited.
attend to one sound source, while ignoring others. The Only one study has been conducted bearing on the
development of these processes has not been studied ex- issue of sound source segregation in children. Sound in
tensively. A few studies suggest that the process of different frequency bands that fluctuate over time in the
grouping components on the basis of source, called same way tend to be grouped together by adults. In fact,
sound source segregation, is functional in infancy, but it an adult will detect a signal at a lower level if the masker
is not clear how accurately or efficiently it operates. consists of multiple frequency bands with common am-
Demany (1982), for example, used repeating tone se- plitude modulation than when the masker is a single fre-
quences to study source segregation. In one sequence, quency band centered on the signal or if the masker
three of four tones were close in frequency while the noise bands have different amplitude fluctuations. This
fourth was somewhat higher in frequency. Adult listen- effect is known as comodulation masking release
ers perceived this sequence as coming from two sources, (CMR). Grose, Hall, and Gibbs (1993) first showed that
one producing three different low frequencies, the other 4-year-old children derived the same benefit from
producing the single higher frequency tone. If the order adding off-frequency, comodulated frequency bands as
Infant Audition 71

do adults. Hall, Grose, and Dev (1997) subsequently presented. The average threshold of the children who
confirmed this finding in slightly older children. How- could perform the task was much higher than that of
ever, Hall et al. also reported that when the masker band adults. Oh, Wightman, and Lutfi (2001) reported that
centered on the signal frequency and the off-frequency preschool children demonstrated about 50 dB more
comodulated bands were slightly asynchronous, adults masking than adults on average when a varying distant-
CMR was reduced, but childrens CMR was eliminated frequency, two-tone masker was presented with the tone
or became negative. Thus, it would appear that the basic to be detected. Moreover, Wightman, Callahan, Lutfi,
process of grouping frequency bands on the basis of Kistler, and Oh (2003) found that while presenting the
common temporal fluctuations is functional early in varying masker tones to the ear contralateral to the sig-
life, but the process is more easily disrupted in children nal ear eliminated such informational masking in adults,
than it is in adults. These findings may have consider- this manipulation did little to reduce informational
able relevance to childrens listening in modern, com- masking in preschool children. Since acoustic factors
plex sound environments. that increase the listeners ability to perceptually segre-
Finally, to process sound emanating from one among gate the signal and masker typically reduce informa-
several sources, listeners must be able to ignore irrele- tional masking in adults, this finding suggests that
vant sounds. Consider that under normal circumstances, childrens ability to segregate sound sources is not as ro-
the irrelevant sounds in the environment may vary from bust as that of adults. Stellmack, Willihnganz, Wight-
moment to moment in unpredictable ways. One of the man, and Lutfi (1997) quantified the extent to which
most intriguing findings in psychoacoustics is that un- irrelevant information entered into childrens percep-
certainty about the sounds to be ignored makes it tual decisions about intensity, finding that preschool
much more difficult for listeners to detect a known children tended to weight information at different fre-
sound (e.g., Kidd, Mason, & Arbogast, 2002; Neff & quencies equally, even when they were asked to attend
Callaghan, 1988; Neff & Green, 1987; Oh & Lutfi, to a single frequency.
1999), even when the sounds to be ignored are distant in
frequency from the sound to be detected. Reduction in
Implications for the Development of
the audibility of one sound due to the introduction of a
Speech Perception
second sound that does not interfere with the peripheral
processing of the signal sound is called informational The preceding review of infant audition has several im-
masking (Pollack, 1975). plications for their perception of the complex sounds of
In some respects, infants act as if they are uncertain human language. In the first 6 months of postnatal life,
about an irrelevant sound, even when the irrelevant it is likely that the neural representation of sounds is not
sound does not change over time. Werner and Bargones as sharp or detailed as it is in adulthood. This represen-
(1991) showed that 7- to 9-month-olds thresholds for tational limitation may in turn limit infants ability to
detecting a tone increased when a noise band distant in extract information from those sounds. By 6 months of
frequency was presented simultaneously. Adults did not age, infants probably have adultlike representations of
demonstrate masking under the same condition. If com- speech and other complex sounds. However, this is not to
peting distant-frequency tones of varying frequency are say that their perception of complex sounds is adultlike.
presented with the tone to be detected, infants exhibit It is clear that infants do not attend to the information
the same increase in threshold as adults do relative to within complex sounds in the same way that adults do.
the condition in which the competing tones are constant They do not appear to focus on the spectral or temporal
in frequency (Leibold & Werner, 2003). Thus, infants details that are most informative. They identify the spa-
may be equally affected by increased uncertainty, even tial location of a sound source rather grossly, and they
though they are more uncertain than adults under non- may also have difficulty segregating speech from com-
variable conditions. peting sounds. Adult caregivers may compensate for
By contrast, additional uncertainty appears to have these immature processing abilities by exaggerating im-
more dramatic effects on older children than it does on portant details and by speaking to infants in a way that
adults. Allen and Wightman (1995), for example, found makes their speech stand out from background sounds,
that half of 4- to 8-year-olds could not detect a tone at all as indicated at various points in the following sections.
when two competing tones varying in frequency were We return to the possible links between early auditory
72 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

processing and resulting effects on speech and language criminate stimuli from the / ba /-/pa / continuum that
learning at the end of this chapter. constitute between, rather than within, category differ-
ences (according to adult perceptual performance).
Given the difficulty of obtaining labeling data from in-
INFANT SPEECH PERCEPTION AND WORD fants, these findings showing better between than
LEARNING: BEGINNINGS OF LANGUAGE within category discrimination, were taken as evidence
that infants also show categorical perception. A year
later, Morse (1972) extended this work to show that 2-
Emergence of the Field: Phonetic Perception
month-old infants can categorically discriminate / ba /
When the first studies of infant speech perception were versus /ga /, but fail on nonspeech counterparts to these
launched in 1971, a number of studies had been pub- syllabic forms. Similarly, Eimas (1974) showed that in-
lished revealing that adults show categorical perception fants, like adults, only discriminate between stimuli
of speech, but not nonspeech sounds. For example, adults that adults label as instances of different categories,
presented with an equal step-size continuum of stimuli holding acoustical distinctiveness constant. A number
spanning two phonetic categories (e.g., a voicing differ- of additional studies extended this work to other
ence between / b/ and /p/ or a place of articulation dif- consonant types (e.g., Eimas, 1975a; Hillenbrand,
ference between / b/ and /d/) categorically labeled the 1984), to consonants in medial as well as initial posi-
first several steps along the continuum as one phoneme tion (Jusczyk, Copan, & Thompson, 1978), and even to
(e.g., / b/), and the next several steps as the other (e.g., newborn infants (Bertoncini, Bijeljac-Babic, Blum-
/p/), with a very sharp boundary in between. Moreover, stein, & Mehler, 1987). Subsequent studies demon-
their labeling performance predicted discrimination. strated that phonetic categories in young infants show
When presented with pairs of stimuli of equal sized dif- many other properties observed in adults. The bound-
ferences, adults reliably discriminated only those dif- aries between categories are not absolute values, as
ferences to which they were able to assign different would be indicated by auditory models. Instead, they
phonetic category labels. This perceptual skill is very are influenced by other articulatory variables, such as
important to language processing. There are tremendous speaking rate (Miller, 1987). Infants (Eimas & Miller,
variations in the way each individual phoneme is pro- 1992), like adults (Whalen & Liberman, 1987) show a
nounced as a function of the other phonemes around it phenomenon called duplex perception wherein the
(/ b/ is somewhat different in bat than in boot due to exact same stimulus can be simultaneously heard as
the coarticulation from the following vowel), as a func- both speech and nonspeech with categorical perception
tion of speaking rate, and as a function of the voice of the speech percept and continuous (no sharp cate-
quality of the individual speaking. Categorical percep- gory boundaries) perception of the nonspeech percept.
tion allows listeners to treat these differences as equiva- Similar results were found with vowels. Infants, like
lent, and thus to recover the word (and hence the adults, show categorical perception of brief (Swoboda,
meaning) rapidly when listening to others speak. On the Morse, & Leavitt, 1976; Trehub, 1973) but not more ex-
basis of the studies published until 1970, it was believed tended, isolated vowels (Swoboda, Kass, Morris, &
that categorical perception, and perceptual normaliza- Leavitt, 1978), and categorize vowels as equivalent even
tion for speaking rate, vowel context, and so on, was across variations in speaker and gender (Kuhl, 1979). In
unique to humans and unique to speech versus other more recent work, Kuhl and colleagues provided data
types of acoustic signals (see Liberman, Cooper, & suggesting that vowel categories are organized around
Shankweiler, 1967; Repp, 1984). prototypes, with best central instances (see Grieser &
To explore the ontogeny of this capacity, Eimas and Kuhl, 1989). These central instances have been de-
his colleagues (Eimas, Siqueland, Jusczyk, & Vigorito, scribed as magnets which warp the vowel space
1971) published a classic study using the high ampli- (Kuhl, 1991; though see Lotto, Kluender, & Holt, 1998,
tude sucking method, in which infant habituation and for an alternative account). A comprehensive review of
dishabituation are measured via rate of sucking on a this sizeable early work on consonants and vowels can
nonnutritve pacifier (see also Moffit, 1971). Their re- be found in Eimas, Miller, and Jusczyk (1987). Taken
sults demonstrated that 1- and 4-month-old infants, together, these studies led to the claim that speech per-
like English-speaking adults, are better able to dis- ception is special in infants just as it is in adults, and
Infant Speech Perception and Word Learning: Beginnings of Language 73

must therefore reflect the operation of a domain- successfully discriminate the Hindi retroflex-dental dis-
specific ability. tinction and another non-English (Interior Salish, Nth-
A number of studies examining cross-language lakampx glottalized velar versus uvular) distinction, but
speech perception complement the above studies of na- by 10 to 12 months of age English infants were no longer
tive-language speech perception. Languages differ in discriminating non-English distinctions (Werker & Tees,
many properties, including their phoneme inventories. 1984). Confirming that the change was one of mainte-
English, for example, contains a contrast between /r/ nance via language-specific listening exposure, and not
and / l / which is lacking in Japanese, but English lacks simply a general decline at 10 to 12 months for all diffi-
the retroflex / D/ versus dental /d/ distinction that is cult phonetic contrasts, Werker and Tees (1984) showed
used in Hindi and other South Asian languages. A series that infants of the same age (10 to 12 months) raised with
of cross-language speech perception studies in the Hindi or Nthlakampx did successfully discriminate the
1970s revealed that adults have difficulty perceiving contrasts from their native languages.
acoustically similar nonnative contrasts, and are con- In the years since this initial work, there have been a
strained to distinguishing only those differences that number of replications and extensions of this finding.
are used phonemically in the native language (e.g., Several studies have confirmed an effect of listening ex-
Lisker & Abramson, 1971; Strange & Jenkins, 1978) perience on the phonetic differences infants can dis-
whereas young infants discriminate phonetic contrasts criminate by the end of the 1st year of life. A decline in
whether or not they are used in the language they are performance on nonnative contrasts was the common
learning (Aslin, Pisoni, Hennessy, & Percy, 1981; pattern in the first wave of studies (Best, McRoberts,
Lasky, Syrdal-Lasky, & Klein, 1975; Streeter, 1976; Lafleur, & Silver-Isenstadt, 1995; Pegg & Werker,
Trehub, 1976). To capture this pattern of results, Eimas 1997; Tsushima et al., 1994; Werker & Lalonde, 1988;
(1975b) proposed that babies are born with broad- Werker & Tees, 1984). Moreover, the basic pattern of
based, universal sensitivities and that lack of listening findings from these behavioral studies has been repli-
experience leads to loss of unused initial sensitivities. cated by recording event-related potentials (Cheour
Aslin and Pisoni (1980) formalized this view in their et al., 1998; Rivera-Gaxiola, Silva-Pereyra, & Kuhl,
universal theory of speech perception, drawing on 2005). Werker showed in her earliest work that this de-
the notion of maintenance as the perceptual mecha- cline likely involves a reorganization of attention rather
nism accounting for cross-linguistic differences (Got- than a loss of basic discriminatory capacity (Werker &
tlieb, 1976; Tees, 1976). Logan, 1985).
However, these comparisons of infants and adults re- In the past several years, some revealing exceptions
lied on different testing procedures for the two popula- to this pattern of findings have appeared. It is now
tions, and in most cases, tests in different labs on known that vowel perception likely reorganizes at a
different contrasts. Werker, Gilbert, Humphrey, and somewhat younger age than consonant perception (Kuhl,
Tees (1981) addressed this problem by comparing En- Williams, Lacerda, Stevens, & Lindblom, 1992; Polka
glish infants aged 6 to 8 months, English adults, and & Werker, 1994), and that acoustically quite distinct
Hindi adults on their ability to discriminate both an En- contrasts that lie outside the phonological space of the
glish and two (non-English) Hindi consonant contrasts, native language (e.g., click contrasts) may remain dis-
using precisely the same methodologythe conditioned criminable even without listening experience (Best,
head-turn procedurewith the three groups. Their re- McRoberts, & Sithole, 1988). Indeed, there are differ-
sults confirmed the developmental change. All three ences across nonnative contrasts, with some showing the
groups discriminated the English / ba /-/da / contrast, but pattern of decline noted above, and others remaining
only the Hindi adults and the English infants discrimi- discriminable (e.g., Best & McRoberts, 2003; Polka &
nated the two Hindi distinctions. Bohn, 1996), with one influential model suggesting that
Werker and her colleagues subsequently completed a the assimilability to the native language phonology is the
series of studies designed to identify the age at which the best predictor of maintenance versus decline (Best,
change from universal to language-specific phonetic 1994). Moreover, experience not only maintains native
perception might occur (Werker & Tees, 1983), and distinctions, but also seems to improve the sharpness of
found important changes occurring across the 1st year of the native categories (Kuhl, Tsao, Liu, Zhang, & de
life. At 6 to 8 months of age, English-learning infants Boer, 2001; Polka, Colantonio, & Sundara, 2001). This
74 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

observation has led to a reanalysis of the original Eimas (e.g., Novak, Kurtzberg, Kreuzer, & Vaughan, 1989).
(1974) claim of universal phonetic sensitivities from Molfese consistently found the LH advantage for place-
birth, with experience playing primarily a maintenance of-articulation differences reported above, but a
role. Now it appears that although substantial organiza- stronger RH ERP response to a change in voicing
tion may be evident from birth, learning also plays a role (Molfese, Burger-Judish, & Hans, 1991; Molfese &
(Kuhl, 2000; Werker & Curtin, 2005). As we will later Molfese, 1979). The pattern seen with vowels also sug-
address further, an exciting new development is attempt- gests asymmetrical processing, in this case favoring the
ing to ascertain just how powerful learning might be. RH (Cheour-Luhtanen et al., 1995). The findings of
In addition to the empirical evidence suggesting in- early sensitivity to potential phonetic distinctions and
nate biases in phonetic perception, considerable re- the possibility of specialized neural systems subserving
search supported the notion that speech might be this discrimination strengthened the notion that speech
special. Early studies suggested that while speech is is special and computed by dedicated neural systems.
perceived categorically, listeners typically perceive As is the case with any complex research endeavor,
nonspeech analogues in a more continuous fashion (e.g., not all the data fit the pattern so nicely. Shortly after the
Mattingly, Liberman, Syrdal, & Halwes, 1971) and uti- first studies were published revealing categorical-like
lize specialized areas or structures in the left hemi- perception for speech sounds, similar studies were pub-
sphere when engaged in phonetic discrimination tasks lished showing that both adults (Pisoni, 1977) and
(Phillips, Pellathy, & Marantz, 1999; Studdert-Kennedy infants (Jusczyk, Pisoni, Walley, & Murray, 1980) per-
& Shankweiler, 1970). Similar types of findings were ceive some nonspeech sounds categorically. Moreover,
revealed in infant studies (e.g., Eimas et al., 1971). Di- nonhuman animals seem to show similar category
chotic tasks with infants showed a significant right ear boundaries to human infants. Chinchillas (Kuhl &
(LH) advantage in phonetic discrimination tasks by 3 Miller, 1978; Kuhl & Paden, 1983) show categorical
months of age (Glanville, Levenson, & Best, 1977) and perception for both voicing (e.g., /pa /-/ ba /) and place
possibly earlier (Bertoncini et al., 1989). Even in these (e.g., / ba /-/da /) continua, and several other animal
initial reports, however, some results did not support the species can also discriminate between consonants
notion of a right-ear/ left-hemisphere advantage for pho- (Morse & Snowdon, 1975; Waters & Wilson, 1976).
netic discrimination in young infants (see Best, Hoff- Japanese quail show perceptual constancy of consonant
man, & Glanville, 1982; Vargha-Khadem & Corballis, categories across variations in vowels (Kluender, Diehl,
1979, for contradictory studies). & Kileen, 1987), and budgies similarly discriminate
The use of electrophysiology, with the event related consonants (Dooling, Best, & Brown, 1995). Similar
potential (ERP) as the dependent variable, has helped findings have been reported for vowels. Monkeys and
clarify the early neuropsychological work. For example, even cats discriminate /i/ from /u/ (Dewson, 1964), and
Dehaene-Lambertz and Baillet (1998), using ERP, studies with Old World monkeys suggest the same pat-
found brain areas that show activation to a change in tern of vowel perception as seen in humans, with excel-
phonetic category, but not to an equal sized change lent discrimination of distinct vowels and more
within a phonetic category in 3-month-old infants. More confusion of close vowels such as / E/ (as in bet) and /ae/
recently, they have shown the same pattern of findings (as in bat; Sinnott, 1989). It was initially thought that
when multiple voices are used, showing that the infant only humans (adults and infants) show the prototype
brain can extract phonetic categories across variations magnet effect (see Kuhl, 1991), but even rats and birds
in speakers (Dehaene-Lambertz & Pea, 2001). ERP show a warping of their perceptual space to reveal a pro-
studies also consistently reveal asymmetries in phonetic totype organization following brief exposures to vowel
discrimination tasks, but the pattern of asymmetries categories (Kluender, Lotto, Holt, & Bloedel, 1998;
appears to vary with stimulus type and infant age. Some Pons, in press). The animal work raises the strong possi-
studies reveal a left hemisphere (LH) advantage for bility that speech perception is not necessarily a spe-
stop consonant discrimination (Dehaene-Lambertz & cialized human capacity, but perhaps instead reflects
Baillet, 1998; Dehaene-Lambertz & Dehaene, 1994; perceptual biases that are common at least across pri-
Molfese & Molfese, 1979, 1980, 1985) whereas others mates, and perhaps beyond.
indicate bilateral responses at birth, with the emergence To summarize, the first generation of infant speech
of right hemisphere (RH) dominance at 3 months of age perception work took as its starting point the work on
Infant Speech Perception and Word Learning: Beginnings of Language 75

phonetic perception in adults, and was designed to as- comparison with three solo musical instruments playing
sess whether infants showed the same types of responses the same tune.
to phonetic differences as do adults, and if so, whether More recently, a set of studies has examined infants
they used the same underlying neural mechanisms. This preference for acoustic stimuli that have the structural
research led to increasingly sophisticated methods and properties of isolated syllables of human speech in com-
techniques, and to studies with both human and nonhu- parison to carefully matched nonspeech tokens.
man animals. The findings greatly enriched our under- Vouloumanos, Kiehl, Werker, and Liddle (2001) used
standing of the development of speech perception, and complex nonspeech analogues modeled on sine-wave
the explanations offered to explain these findings pro- analogues of speech (Remez, Rubin, Pisoni, & Carrell,
vided rich theoretical fodder for subsequent work. How- 1981). The speech stimuli consisted of the syllable lif
ever, one of the insights guiding research for the past repeated several times in the high pitch, highly modu-
several years has been the realization that there is much lated speech that parents use when speaking to their
more to speech than just phonetic categories, and that infants. The nonspeech counterparts replaced the funda-
infants may be sensitive to many other characteristics of mental frequency and three most intense higher order
the speech around them. frequency components (formants) with a sine wave that
tracked their changes across time. Thus, in contrast to
A Preference for Speech the earlier studies investigating preference for speech
over nonspeech, these stimuli were carefully matched
One of the first questions one might ask is whether in- for duration, timing, fundamental frequency, and area in
fants perceptual systems help them to separate speech the spectrum in which information was presented. How-
from other types of acoustic signals in the environment. ever, human vocal tracts cannot produce the sine wave
An early appearing preference for speech would help in- nonspeech stimuli.
fants orient to just those signals in the environment In the first set of studies, Vouloumanos and Werker
which are essential for language acquisition. Although it (2004) used a sequential preferential looking procedure
is widely believed that infants prefer to listen to speech (e.g., Cooper & Aslin, 1990) to test the listening prefer-
over other sounds from the first moments of life, there ences of 2- to 6-month-old infants. The infants preferred
are actually very few data that specifically address this the speech over the complex nonspeech analogues, lis-
question. Indeed, until very recently the studies upon tening longer on the alternating trials during which
which this widely held belief was founded were not ac- speech versus nonspeech was presented. In the second
tually designed to test infants preference for speech set of studies, newborn infants were tested with these
over nonspeech. For example, one study that is widely same stimuli, with HAS ( high amplitude sucking) as
cited as showing a neonatal preference for speech over the dependent variable. Like their older counterparts,
nonspeech showed that 4- to 5-month-old infants look the newborn infants chose to deliver more HA sucks
longer to a target when it is paired with continuous fe- on the alternating minutes in which speech versus com-
male speech than when it is paired with white noise plex nonspeech was presented (Vouloumanos & Werker,
(Colombo & Bundy, 1981). Today no one would accept 2002). To attempt to rule out a role of experience in
the aversive sound of white noise as an appropriate con- eliciting this preference, Vouloumanos created stimuli
trol for human speech, but in fact Colombo and Bundy that would sound like those available to the fetus by fil-
had not designed the experiment to test for a preference tering them using the filtering characteristics of the
for speech. Rather, they were attempting to develop a uterine wall. Neonates treated the filtered speech and
method for assessing infants responsiveness to different nonspeech as equivalent, even though they discriminate
types of speech sounds. Moreover, what is not noted is the nonfiltered counterparts. This strengthens the possi-
that 2-month-old infants in Columbo and Bundys bility that the preference seen in the newborn is not a di-
(1981) study, in contrast to older infants, did not re- rect result of prenatal listening experience with human
spond differently to speech and white noise. The only speech, and argues instead for an evolutionarily given
other early study directly assessing a listening prefer- perceptual predisposition for sounds that have the struc-
ence for speech was one by Glenn, Cunningham, and tural characteristics of those which could be produced
Joyce (1981) in which 9-month-old infants pulled a lever by a human vocal tract. Moreover, the preference for
more frequently to listen to a female singing a song in communicative signals extends beyond spoken language.
76 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

In a recent study, Krentz and Corrina (2005) have shown ble that combines features of both the heard and seen
that hearing infants show a preference for watching sign stimuli. This effect is robust across many testing condi-
language over carefully matched nonlinguistic gestures. tions and languages (see Green, 1998 for a review) and
Taken together, these studies suggest a broad-based per- has been interpreted as part of our endowment for pho-
ceptual bias for communicative signals. netic perception. Yet, there is also evidence of learning.
These results are corroborated by studies using neu- The McGurk effect is stronger in adults than in children
roimaging techniques. In an event-related fMRI study (Hockley & Polka, 1994; MacDonald & McGurk, 1978;
completed with adults, Vouloumanos et al. (2001) found Massaro, Thompson, Barron, & Laren, 1986), is re-
that the typical speech areas in the left hemisphere of duced further in children who have difficulty articulat-
the temporal cortex are more activated by a change to ing (Desjardins, Rogers, & Werker, 1997), and shows
the speech than a change to the complex nonspeech the same kind of language-specific influences as is seen
stimuli described above. This finding complements in the perception of audible speech with nonnative
many other studies with adults showing greater activa- visemes assimilated to those used in the native lan-
tion of specialized brain areas in the left hemisphere in guage (Massaro, Cohen, & Smeele, 1995; Werker, Frost,
response to speech than to other types of sounds (e.g., & McGurk, 1992).
Benson et al., 2001; Binder et al., 1997; Fiez et al., 1995; Two kinds of studies have explored whether the visi-
Price et al., 1996; Zatorre, Evans, Meyer, & Gjedde, ble information in speech is available prior to learning.
1992; but see also Binder et al., 2000; Zatorre, Meyer, In one, infants are presented with side-by-side displays
Gjedde, & Evans, 1996). of two faces articulating two different syllables. An
To date, only two studies have used imaging tech- acoustic syllable that matches the syllable being articu-
niques to determine if the infant brain responds differ- lated by one of the faces is then presented at mid-line,
ently to speech versus nonspeech. Both of these studies and the amount of time the infant looks at each face is
have contrasted the perception of forward versus recorded. Using this method, Kuhl and Meltzoff (1982)
backwards speech (for related adult studies, see, De- showed that infants of 4.5 months look preferentially at
haene et al., 1997; Wong, Mihamoti, Pisoni, Sehgal, & the face articulating the heard vowel sound (/a / versus
Hutchins, 1999). Using both optical topography (Pea /i/). This finding has since been replicated with other
et al., 2003), and fMRI (Dehaene-Lambertz, Dehaene, vowels (Kuhl & Meltzoff, 1988), with male as well as
& Hertz-Panier, 2002), speech stimuli elicited greater female faces and voices (Patterson & Werker, 1999),
activation in the infants LH than the RH. In the Pea with disyllables (e.g., mama, lulu; MacKain, Stud-
et al. (2003) study neonates were tested, and the in- dert-Kennedy, Speiker, & Stern, 1983), and with a high
creased activation was in the classic language areas over amplitude sucking method (Walton & Bower, 1993).
the temporal lobe. The Dehaene, Dehaene, and Hertz- Moreover, these young infants often display mouth
Pannier (2002) study tested 3-month-olds, eliciting bi- movements themselves that correspond to the concor-
lateral activation to both the forwards and backwards dant bimodal display (Kuhl & Meltzoff, 1988; Patterson
speech over the temporal lobes, with greater LH activa- & Werker, 1999; 2002), suggesting connections includ-
tion more posteriorly. Further experimentation with ing not only the visual and auditory perceptual modali-
nonhuman primates is necessary to determine whether ties, but also articulatory processes. More recently it has
these early perceptual and neural markers of human been shown that the matching effect is equally robust in
speech perception are specific to humans or are instead 2-month-old infants (Patterson & Werker, 2003). The
part of our shared evolutionary history. precocity of this matching ability is particularly strik-
ing when compared to other types of biologically impor-
Perception of the Visible Information in Speech tant information. Infants do not match gender in the face
and voice until 7 to 9 months of age (Walker-Andrews,
Speech perception involves not only the acoustic signal, Bahrick, Raglioni, & Diaz, 1991) even when they are
but also visible articulatory information. The best tested using precisely the same stimuli for which they
known example of this is the McGurk effect (McGurk & show vowel matching (Patterson & Werker, 2002).
MacDonald, 1976). When watching a speaker produce Evidence for the McGurk effect itself is less con-
the syllable /ga / while listening to / ba /, adults typically vincing in the infancy period. Although there are
report perceiving an unambiguous /da / or /tha /, a sylla- reports of infants percepts showing the same kind of
Infant Speech Perception and Word Learning: Beginnings of Language 77

fusion or visual capture as seen by adults when glish and Dutch show greater access to the phoneme
mismatched auditory and visual stimuli are presented (Cutler, Mehler, Norris, & Segui, 1986; Cutler & Norris,
(Burnham & Dodd, 2004; Desjardins & Werker, 2004; 1988; Vroomen, van Zon, & de Gelder, 1996), and Japa-
Rosenblum, Schmuckler, & Johnson, 1997), the effect is nese adults use the mora as the primary unit of segmenta-
not nearly as strong or as consistent as that seen in tion (Otake, Hatano, Cutler, & Mehler, 1993). These
adults. Taken together, these studies suggest that the in- differences not only describe the surface properties of
fant may be endowed from an early age with a percep- languages, but may also provide cues to the underlying
tual system which is sensitive to both heard and seen syntactic structure, that is, the head direction, of the lan-
features of phonetic segments, but that this system is guage (Nespor, Guasti, & Christophe, 1996).
perfected and tuned through experience listening to and Human infants are sensitive to rhythmical differ-
articulating speech. A recent study suggests that we are ences from birth. Since the classic study by Demany,
not the only primate species to use both heard and seen McKenzie, and Vurpilot (1977) showing that 2- to 3-
information in perceiving communicative stimuli month-old infants can discriminate tones based on their
(Ghazanfar & Logothetis, 2003), suggesting that the in- rhythmical sequences, there is growing evidence that
termodality of speech might be deeply ingrained in our young infants can use rhythm to detect the timing char-
evolutionary heritage. acteristics of speech (Fowler, Smith, & Tassinary,
1986). Indeed, there is now considerable evidence that
Perception of Prosodic Attributes of the infants utilize rhythmical characteristics to discrimi-
Speech Signal nate one language from another at birth. In a seminal ex-
periment, Mehler et al. (1988) demonstrated that
One of the fundamental characteristics of human lan- newborns can discriminate French from Russian pro-
guages is their prosodythe musical aspects of speech, duced by a single bilingual speakeras long as one of
including their rhythm and intonation. Languages have these languages is the infants native language. Interest-
classically been categorized according to their predomi- ingly, at least part of this discrimination is based on
nant rhythmic properties into three major types: stress- prosodic characteristics of speech. The same results
timed, syllable-timed, and mora-timed (Abercrombie, were obtained using low-pass filtered speech, in which
1967; Pike, 1945). Stress-timed languages, like English the phonetic characteristics were removed while retain-
and Dutch, tend to alternate between strong and weak ing the prosodic rhythm of the speech samples. Given
syllables, and the strong syllables are roughly equally that these are the same speech characteristics main-
spaced in time, thus the term stress-timed. Languages tained in the uterine environment, an initial interpreta-
like Spanish and Italian, however, use the syllable as the tion of these data was that newborn speech capacities
basic unit of timing; syllables are similarly stressed and are influenced by prenatal maternal exposure. A subse-
roughly equally spaced in time. Finally, languages like quent reanalysis of the initial data, however, revealed
Japanese are timed-based on the mora, a rhythmic unit that discrimination was not limited to the native versus
roughly corresponding (in English) to a consonant fol- an unfamiliar language, but instead was evident as well
lowed by a short vowel ( the contains one mora, while when the French infants listened to English versus Ital-
thee contains two). This nomenclature has been re- ian, two unfamiliar languages (Mehler & Christophe,
fined and quantified by two key properties: percent 1995). This weakened the interpretation that it is prena-
vowel per syllable and the variability in the consonants tal experience that leads to language discrimination.
(Ramus, Nespor, & Mehler, 1999). These rhythmical Indeed, there is now abundant evidence that infants
properties of the language influence adults processing use such rhythmical differences to discriminate among
of speech; speakers of different languages employ differ- a wide variety of languages, suggesting that rhythm may
ent units as their primary unit of segmentation. Speakers be prioritized in infants early speech representations.
of syllable-timed languages (e.g., French, Spanish, Consistent with this hypothesis, infants discrimina-
Catalan, & Portuguese) show a processing advantage for tions are predictable based on languages membership in
the syllable (e.g., Mehler, Dommergues, Frauenfelder, & rhythmic classes. For example, newborns and 2-month-
Segui, 1981; Morais, Content, Cary, Mehler, & Segui, olds can discriminate between two languages from two
1989; Sebastin-Galls, Dupoux, Segui, & Mehler, different classes, but not two languages from the same
1992), speakers of stress-timed languages such as En- class (Mehler et al., 1988; Moon, Cooper, & Fifer, 1993;
78 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

Nazzi, Bertoncini, & Mehler, 1998). This discrimination guages they hear. Such a process would likely facilitate
ability is present for forward speech but not backwards successful bilingual language acquisition by alerting in-
speech, which disrupts rhythmic cues. Nonhuman pri- fants to the fact that their input is drawn not from one
mates show this same pattern of performance, suggest- but two language systems. In the absence of this infor-
ing that processing of the rhythmic characteristics of mation, infants may not have a way to determine that
speech is not specific to humans (Ramus, Hauser, they are hearing multiple languages, leading to potential
Miller, Morris, & Mehler, 2000). Only by 5 months of confusions during the learning process.
age does experience with ones native language allow To begin to address these issues, Bosch and Se-
infants to discriminate it from other languages in the bastin-Galls (2001) assessed the language recognition
same class (e.g., Nazzi, Jusczyk, & Johnson, 2000). abilities of 4-month-olds learning both Spanish and
There is thus an interplay between an inherent bias to at- Catalan. Importantly, these two Romance languages be-
tend to rhythmic distinctions in languages and the learn- long to the same rhythmic category, which should make
ing processes required to distinguish ones own them quite difficult to discriminate. Nevertheless, these
language from others. Even by 5 months of age, infants bilingual-to-be infants were able to discriminate be-
cant distinguish between two unfamiliar languages tween the two languages present in their home environ-
from the same rhythmic class. ments. These results suggest the availability of an early
Languages from different rhythmical classes differ capacity to distinguish languages given simultaneous
not only in their timing characteristics, but also in their bilingual exposure, potentially based on the presence of
intonation contours. And, in some cases, this informa- vowel reduction (since rhythmic cues do not distinguish
tion is also adequate to discriminate two languages that the two languages). More recently, it has been shown
have been low-pass filtered at 400 Hz (which removes that bilingual-exposed infants process languages differ-
phonetic content while preserving intonation cues), in- ently even from birth (see Werker, Weikum, & Yoshida,
cluding English versus Japanese (Ramus & Mehler, in press). When tested on their preference for English
1999). As indicated earlier, young infants are highly over a rhythmically distinct language, Tagalog, English-
sensitive to differences in pitch. Moreover, they are able exposed newborns showed a robust preference for fil-
to use intonation to discriminate vowels (Bull, Eilers, & tered English speech over filtered Tagalog speech.
Oller, 1984; Karzon & Nicholas, 1989), and even to dis- However, newborns exposed to both English and Tagalog
tinguish lists of words differing in pitch contour (Nazzi, prenatally did not show this preference, and indeed,
Floccia, & Bertoncini, 1998). The preference for infant chose to listen equally to filtered English and Tagalog.
directed speech (Cooper & Aslin, 1990; Fernald, 1984; Contrary to hypotheses suggesting that lexical knowl-
Werker & McLeod, 1989) may be explained, at least in edge is needed to engage in language differentiation
part, by this exquisite sensitivity to fundamental fre- (e.g., Genesee, 1989), the basic capacities for language
quency (see Colombo & Horowitz, 1986). differentiation may be in place well prior to the onset of
It is thus of interest to determine whether infants spoken language. However, these results leave open the
ability to discriminate languages is a function of their question of whether infants actually represent the two
perception of rhythm, intonation, or both cues in tan- languages as separate systems, as opposed to discrim-
dem. To do so, Ramus and Mehler (1999) resynthesized inable components of a single system.
natural speech, preserving the rhythm while holding in-
tonation constant. With these stimuli, French newborns
Perception of Other Aspects of the Speech Signal
were still able to discriminate languages from two dif-
ferent rhythmic classes (Ramus, 2002; Ramus et al., In addition to their sensitivity to speech itself, to pho-
2000), although levels of discrimination were somewhat netic segments, to visual speech, and to rhythm and into-
attenuated. These results confirm that while infants nation, even the youngest infants also show impressive
may be using intonation to boost performance, rhythm sensitivities to many other types of information carried
alone is sufficient to distinguish one family of languages by the speech signal. They are sensitive to some kinds
from another. of within phonetic category variation. At 3 to 4 months,
These early language discrimination abilities may be infants show graded perception of VOT (Miller &
particularly useful in bilingual environments. Infants Eimas, 1996), and at 6 months can discriminate within-
exposed to multiple languages may use these rhythmic category differences from along the VOT continuum if
distinctions to segregate the input into the different lan- tested in sufficiently sensitive tasks (McMurray &
Infant Speech Perception and Word Learning: Beginnings of Language 79

Aslin, 2005). Moreover, infants of 6 to 8 months ( but not and are evident in maternal speech across typologically
10 to 12 months) can even treat multiple instances of the distinct languages (Morgan, Shi, & Allopenna, 1996;
voiced, unaspirated [d] versus voiceless, unaspirated [t] Shi, Morgan, & Allopenna, 1998).
(created by removing the s from /sta /) as two separate In a recent series of studies, Shi and colleagues
categories, even though adult English speakers treat showed that infants become increasingly able to use
both of these syllables as equivalently acceptable in- these cues across development. Newborn infants cate-
stances of the phoneme /d/ (Pegg & Werker, 1997). gorically discriminate content from function words,
Sensitivity to within category phonetic variation is even when the words are equated for volume and number
necessary in some language processing tasks (see of syllables (Shi, Werker, & Morgan, 1999). Specific
Werker & Curtin, 2005). One illustration comes from the prenatal listening experience cannot account for this
work on allophone discrimination. Allophones are dif- discrimination capacity; the same pattern of results
ferent phonetic realizations of the same phoneme with emerges when the items are drawn from an unfamiliar
precise phonetic characteristics that vary depending on language. By 6 months of age, infants prefer to listen to
their position in words. By 2 months of age, infants can the content words (Shi & Werker, 2001). This preference
detect the allophonic difference between the unaspi- cannot be accounted for by specific knowledge of highly
rated, unreleased /t / in night rate from the aspirated, familiar items because, again, it is seen even when in-
released, partially retroflex /t / in nitrate (Hohne & fants are tested on words from an unfamiliar language
Jusczyk, 1994), a sensitivity which will ultimately be (Shi & Werker, 2003). Thus the phenomenon must be
useful in word segmentation, as discussed next. based on a developing preference for items with the
Infants are also sensitive to syllable form. Newborn acoustic and phonological patterns seen in content
French infants can count syllables, discriminating words. These findings do not necessarily suggest that in-
lists of bi- versus trisyllabic words even when the words fants are born with knowledge of important grammatical
are modified to have the same overall duration (Bijeljac- categories. They do, however, show that infants percep-
Babic, Bertoncini, & Mehler, 1993). They show better tual biases facilitate dividing words into two fundamen-
discrimination of stimuli that correspond to good syl- tal categories. Subsequently, as infants approach the age
lable formsthose with a vocalic nucleus (/tap/ versus of learning word meanings, they selectively focus on the
/pat /), in comparison to /tsp/ versus /pst / (Bertoncini & louder and generally more salient content word category.
Mehler, 1981). Sensitivity to rhyme (Hayes, Slater, & When they first begin to speak, infants typically omit
Brown, 2000), alliteration (Jusczyk, Goodman, & Bau- function morphemes (e.g., the, -ed, -s), raising the ques-
mann, 1999), and full syllable repetition (Jusczyk, tion of whether infants simply do not perceive them or
Goodman, et al., 1999) have all also been demonstrated whether there are other reasons for the omissions, such
in infants from 7 to 9 months of age. as constraints on speech production (e.g., Gerken &
One such auditory sensitivity may be of use in the ac- McIntosh, 1993). Several lines of evidence suggest that
quisition of grammatically relevant knowledge. Infants infants do in fact perceive these weak items. For exam-
are astonishingly sensitive to the acoustic and phonolog- ple, 11-month-olds, but not 10-month-olds, show a dif-
ical cues that distinguish grammatical classes. Just as ferent pattern of scalp-recorded ERPs to stories that
languages differ in their phoneme inventories and in contained either correct or modified English function
their rhythmical characteristics, they also differ in the morphemes (Shafer, Shucard, Shucard, & Gerken,
number and kinds of grammatical categories words 1998). Similar results using the head-turn preference
might belong to. For example, while English has preposi- procedure with German infants suggest that 7- to 9-
tions, Chinese has postpositions. The languages of the month-olds, but not 6-month-olds, can recognize pre-
world seem to all share a fundamental distinction be- viously familiarized closed class items (Hhle & Weis-
tween open-class ( lexical words, such as nouns, verbs, senborn, 2003). By 11 months of age, infants prefer non-
adjectives, etc.) and closed-class (grammatical words, sense words preceded by a familiar high frequency
such as determiners, prepositions, etc.) categories. function word over that same nonsense word preceded
These classes of words can be distinguished on the basis by a mispronunciation of the function word (e.g., En-
of acoustic and phonological cues such as syllable com- glish infants listen longer to the brink over ke
plexity, syllable number, duration, loudness, and brink ), and do so even for low frequency function
presence of reduced vowels (see Kelly, 1992). These dif- words such as its or her by 13 months of age (Shi,
ferences are magnified in the speech directed to infants, Werker, & Cutler, 2003). Indeed, by 11 months of age it
80 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

appears that familiar high frequency function words Vollmer-Conna, 2003). This fact raises the possibility
such as the facilitate segmentation and learning of that the interplay between paralinguistic and linguistic
new words. Infants show better recognition of nonce factors may be part of, and exclusive to, within-species
word forms if they are first presented in a phrase with a communication.
familiar closed class item such as the (Shi, Werker,
Cutler, & Cruickshank, 2004).
The importance of familiar function words becomes IMPLICIT DISCOVERY OF CUES IN
more pronounced when children are at the peak of learn- THE INPUT: A DRIVE TO MAKE SENSE OF
ing new words, and then shows an apparent decline. In a THE ENVIRONMENT
preferential looking task (side by side pictures) infants
of 18 months were most accurate when the labels for the During the second half of the first year, an explosion oc-
objects were preceded by a correctly pronounced func- curs in infants knowledge of detailed aspects of the
tion word and least accurate when the function word sound structure of their native language(s), as noted in
was mispronounced. Their performance was intermedi- the earlier discussion of age related changes in phonetic
ate for labels with missing function words. Infants of 24 perception. In this section, we document developmental
months showed a similar, but less pronounced pattern, changes in perception of other properties of language.
and by 36 months of age children were able to ignore the The following section addresses the mechanisms that
function word information (Zangl & Fernald, 2003). may be responsible for these learning trajectories.
The role of function words extends beyond segmentation
and identification. By 2 years of age infants are able to
use function words as cues to new versus old informa- Stress and Phonotactic Cues
tion in a sentence context (Shady & Gerken, 1999).
Speech also carries paralinguistic (sometimes called Languages differ greatly in their internal prosodic regu-
indexical) informationcues that convey emotion, larities. By adulthood, speakers use these regularities to
speaker identity, and emphasisto which infants are generate predictions about possible word structures. For
sensitive. They show a preference for infant-directed example, English-speaking adults expect words to be
over adult-directed speech (Cooper & Aslin, 1994; Fer- trochaicto begin with stressed syllablesmirroring
nald, 1984). They discriminate individual voices (De- the distribution of stress in their native language (e.g.,
Casper & Prescott, 1984; Floccia, Nazzi, & Bertoncini, Cutler & Carter, 1987; Cutler & Norris, 1998). This
2000) and show a robust preference for their mothers trochaic bias emerges early in the process of language
voice (DeCasper & Fifer, 1980) from birth, indicating acquisition, well before infants are producing words. For
prenatal learning. Indexical cues may also aid in the example, 9-month-olds prefer to listen to words that ex-
perception of specifically linguistic information. Pho- emplify their native languages stress pattern (Jusczyk,
netic discrimination, for example, is facilitated when Cutler, & Redanz, 1993), and are even sensitive to heavy
the contrasting syllable is produced at the pitch peak in versus light syllables (e.g., syllables with a long vowel
motherese, as shown by Karzon (1985) in infants dis- and/or final consonant versus syllables with only a short
crimination of /marana / versus /malana /. This may be in vowel and no final consonant; Turk, Jusczyk, & Gerken,
part because the distinctiveness of segments may be 1995). These results cannot be explained by recourse to
clarified in motherese, as shown in the acoustic exag- inherent preferences for particular stress patterns, be-
geration of voicing (Ratner & Luberoff, 1984) and of cause 6-month-olds fail to show native-language stress
the vowel space (Kuhl et al., 1997; Ratner, 1984) in in- preferences. This is an example of a potent learning pro-
fant-directed speech. Indeed, research shows that ma- cess; somehow, English-learning infants must have dis-
ternal clarification of the vowel space is correlated with covered a probabilistic prosodic regularity in the input.
superior speech discrimination in 6- to 12-month-old Infants sensitivities are not confined to syllable-
infants (Liu, Kuhl, & Tsao, 2003). Of interest, although level patterns. By 9 months of age, infants have learned a
speech directed to pets has many of the characteristics great deal about the probabilistic phonotactic patterns of
of infant-directed speech, the exaggeration of acoustic their native language: the rates with which certain
cues distinguishing vowels is evident only in speech phoneme sequences occur in particular orders in partic-
directed to human infants (Burnham, Kitamura, & ular positions in syllables and words. For example, the
Learning Mechanisms 81

sequence /ds/ can end, but cannot begin, syllables in En- syllable function words as cues to grammar, have gener-
glish. Phonotactics are not a simple function of pronoun- ally supported the claim that infants are attuned to the
cability; sequences that are legal in some languages are kinds of prosodic variables that are correlated with syn-
illegal in others. Effects of phonotactic structure are tactic structure (for review, see Morgan & Demuth,
observed in studies of adult word recognition (e.g., Vite- 1996). One such prosodic cue is changes in pitch and
vitch & Luce, 1999; Vitevitch, Luce, Charles-Luce, & duration at the ends of clauses in infant-directed speech
Kemmerer, 1997). Similarly, children are affected by (Fisher & Tokura, 1996; Jusczyk et al., 1992). In a clas-
phonotactic probabilities when learning novel object sic study, Hirsh-Pasek et al. (1987) found that 7-month-
names (e.g., Storkel, 2001) and nonword repetition tasks olds listened longer to speech samples with pauses
(Coady & Aslin, in press). By 9 months of age, infants inserted at clause boundaries than sentences with
prefer to listen to phonotactically legal sequences, pauses inserted clause-medially, suggesting that infants
whereas 6-month-olds do not (Friederici & Wessels, detected the disruptions in the latter case. Similar re-
1993; Jusczyk, Friederici, Wessels, Svenkerud, & sults emerge for musical stimuli as well, suggesting that
Jusczyk, 1993), with frequent phonotactic structures detection of prosodic markers serving as unit bound-
preferred over infrequent structures (Jusczyk, Luce, & aries is not limited to language learners (Jusczyk &
Charles-Luce, 1994). Interestingly, infants in bilingual Krumhansl, 1993; Krumhansl & Jusczyk, 1990). More
environments exhibit similar knowledge of phonotactic recent evidence suggests a similar process for phrase
structure (Sebastin-Galls & Bosch, 2002). Differ- units, at least under some circumstances (Soderstrom,
ences were obtained as a function of language domi- Seidl, Kemler Nelson, & Jusczyk, 2003), despite less
nance: infants were most sensitive to phonotactic clear prosodic markers of phrases (Fisher & Tokura,
patterns in their to-be-dominant language, suggesting 1996). Indeed, even newborns have been shown to be
that infants may be limited in the number of phonotactic sensitive to cues correlated with prosodic boundaries
systems they can acquire in parallel. (Christophe, Mehler, & Sebastin-Galls, 2001).
Some types of subsyllabic regularities appear to be Of course, this evidence doesnt demonstrate that
more salient to infants than others. For example, infants know that these prosodic cues point to syntac-
Jusczyk, Goodman, et al. (1999) demonstrated that 9- tic boundaries. And the prosodic phrases to which in-
month-olds were sensitive to sound patterns that re- fants are sensitive correlate only imperfectly with
curred at the beginnings of words, but not the ends of syntactic boundaries (Gerken, Juscyk, & Mandel,
words (for related findings, see Vihman, Nakai, dePao- 1994). Nonetheless, this sensitivity is a prerequisite for
lis, & Hall, 2004). These results suggest that certain the use of prosodic cues to discover syntactic structure.
parts of words may be privileged relative to others in in- There is evidence from adult studies using artificial lan-
fants early speech representations, such that onsets guages that such grouping cues do assist learners in
may contain more detail than codas. Such findings are breaking into syntax (e.g., Morgan, Meier, & Newport,
particularly interesting in light of the conventional wis- 1987; Morgan & Newport, 1981). Other evidence sug-
dom that infants are highly attuned to rhyming, as well gests that prosodic structure helps infants as young as 2
as data suggesting that the ends of words may be privi- months of age to organize and group word sequences in
leged in young childrens lexical representations (Echols memory (e.g., Mandel, Jusczyk, & Kemler Nelson,
& Newport, 1992; Slobin, 1973). 1994). However, the degree to which prosodic structure
facilitates infants discovery of syntactic structure re-
Higher-Level Units mains unknown.

Infants representations of the sound structure of their


language also encompass larger prosodic patterns, span- LEARNING MECHANISMS
ning multiple words. Beginning in the 1980s, re-
searchers have been interested in how such prosodic It is of paramount importance to understand how the
patterns might provide cues to infants to allow them to myriad information in the linguistic environment be-
break into the syntax of their native language. Such comes part of the infants native language knowledge
prosodic bootstrapping accounts, beginning with Gleit- base. Until the advent of new testing techniques, this
man and Wanners (1982) proposal for the use of weak question was only addressable via analyses of production
82 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

from older children, or via logical arguments regarding uvular / k /-/q/ distinction, presumably because ds are
the structure of the problems facing the child and the more frequent in the input than are ks, giving infants
possible solutions that might be part and parcel of the a better opportunity to learn the native language cate-
childs linguistic endowment (e.g., Pinker, 1984, 1989). gory structure. Similarly, Shi et al. (2004) found that in-
Research with computational models also suggested pos- fants recognize and utilize high frequency function
sible ways to structure a learning system that might be words earlier than infrequent ones.
compatible with some of the facts of language acquisi- Infant learning mechanisms are also sensitive to the
tion (Elman, 1990; Rumelhart & McClelland, 1986). statistical distribution of elements in the input. Maye,
In the past decade, researchers have developed exper- Werker, and Gerken (2002) presented 6- and 8-month-
imental methods that help to identify potential learning old infants with stimuli that simulated one of two types
processes. Saffran, Aslin, and Newport (1996) used one of languages. Materials in the unimodal condition col-
such task to determine whether infants could track sta- lapsed a continuum of phonemes into a single category,
tistical properties of speech. Eight-month-old infants such as English does with the two Hindi /d/ sounds, by
listened to a 2 minute continuous sequence of syllables presenting more instances of stimuli from the center of
containing multisyllabic words: for example, golabu- the continuum. Materials in the bimodal condition di-
pabikututibubabupugolabu. . . . They subsequently tested vided the continuum into two categories, such as Hindi
infants ability to discriminate the words of this lan- does with the dental versus retroflex /d/, by including
guage from syllable sequences spanning word bound- more instances of stimuli drawn from closer to the two
aries. Infants success at this task, as evidenced by ends of the continuum. The results suggested that in-
different listening times to the two types of sequences, fants were extremely sensitive to the distributions of the
indicated that they were able to detect and use the statis- exemplars presented during exposure, with different test
tical properties of the speech stream. discrimination exhibited as a function of presentation of
Similar methods can be used to test specific hypothe- the unimodal versus bimodal materials. Distributional
ses about the types of learning mechanisms used by in- statistics can affect category structure, raising the pos-
fants. For example, infants could have succeeded in the sibility that sensitivity to the distributional information
preceding task using two quite different types of statis- in the native language may contribute to the establish-
tics: the probabilities of co-occurrence of syllables (e.g., ment of native language phonetic categories in the 1st
the transitional probability of la given go), or a sim- year of life.
pler computation, frequencies of occurrence (test words Infants also appear to be sensitive to nonstatistical
occurred more often in the input than the other test se- regularities in the input. Marcus, Vijayan, Rao, and
quences). Aslin, Saffran, and Newport (1998) teased Vishton (1999) exposed infants to 3-syllable sentences
apart these two possibilities, demonstrating that infants following a particular pattern (e.g., ga ti ga, li fa li). In-
succeeded at this task even when the test items were fants were then tested on novel sentences that either ex-
matched for frequency of occurrence in the input. A re- emplified or violated the exposure pattern (e.g., wo fe
cent computational analysis of infant directed speech wo versus wo fe fe). Successful discrimination suggested
confirms that probabilities of syllable co-occurrence that the infants acquired abstract information, reflecting
predict word boundaries better than frequencies of syl- knowledge beyond just the specific syllable patterns ob-
lable co-occurrences (Swingley, 2005). served in the input. Marcus et al. (1999) interpreted
We must still track frequencies to discover probabili- their results as evidence for a rule-based learning mech-
ties. And these frequencies are quite salient to infants in anism that detected algebraic rules (operating over vari-
some linguistic domains. For example, infants represent ables). This claim has been controversial, as others have
the frequencies of phonotactic patterns in their native suggested that infants could have performed this task
language (Jusczyk et al., 1994; Mattys, Jusczyk, Luce, without rule-like representations (e.g., Altmann & Di-
& Morgan, 1999). Moreover, infants learn about fre- enes, 1999; Christiansen & Curtin, 1999; Seidenberg &
quent properties of the input before they learn about in- Elman, 1999). Investigators studying adults have simi-
frequent ones. Anderson, Morgan, and White (2003) larly argued that the evidence supports a distinction be-
showed that English infants show a decline at a younger tween rule-based and statistical knowledge (Pea et al.,
age in their perception of the non-English retroflex- 2003). However, it remains difficult to clearly distin-
dental /da /-/ Da / distinction than the non-English velar- guish between the two types of learning systems empir-
Building from the Input during the 1st Year 83

ically (for discussion, see Seidenberg, MacDonald, & development and/or literacy, adults represent both syl-
Saffran, 2003). lables and phonemes (Nygaard & Pisoni, 1995), and
recent adult studies suggest that segmental representa-
tions may serve as the primitives for at least some
Units for Computations
kinds of language learning tasks (Newport & Aslin,
In order to specify the processes that go into the opera- 2004). Of additional interest, there may be differences
tion of any learning mechanism, it is necessary not just across languages in which unit adults use for word seg-
to note the structure of the learning mechanism (e.g., mentation, with French adults showing a pronounced
what computations are performed?) but also to deter- syllable bias and English adults showing sensitivity to
mine the primitives over which those computations are segmental information as well (Cutler, Mehler, Norris,
performed. Consider the simplest possible mechanism, a & Segui, 1983, 1986).
frequency counter that tracks how often some event oc- One interpretation of this body of work is that
curs. Depending on the event in question, the output of the syllable is privileged as a unit of representation
the learning mechanism could be vastly different. For (Bertoncini & Mehler, 1981), and may be used as the
example, if the mechanism is applied to a flock of birds, unit in computations across linguistic input. Another in-
does it compute the total number of birds, or the number terpretation is that both syllables and phonemes are
of birds feet, or the number of swallows versus doves? privileged, but for different types of tasks: the syllable
Each of these primitives, serving as input to the learning is the primary unit for counting, but segmental detail
process, renders a different answer. plays a role in segmentation, at least in stress-timed lan-
The issue of primitives has been prominent in the guages (see Werker & Curtin, 2005). Further research
study of speech representations. Artificial speech will help distinguish between these possibilities.
recognition systems intended to simulate early lan-
guage development have largely focused on the BUILDING FROM THE INPUT DURING
phoneme as the relevant unit for modeling (e.g., Brent & THE 1ST YEAR
Cartwright, 1996; Christiansen, Allen, & Seidenberg,
1998; Jusczyk, 1997). Although some research supports The foregoing review provides some clues regarding the
this unit as important in infant speech perception, other types of linguistic information infants acquire during
work suggests that this idealization may be a mismatch the 1st year, as well as potential learning mechanisms
to infants capabilities. that subserve this acquisition process. We can now turn
Shortly after the field of infant speech perception to the burgeoning literature that puts these two pieces
emerged, researchers began to focus on what the unit of together: studies concerning the acquisition of particu-
representation might be. One long-standing contro- lar linguistic structures. Some of these studies teach in-
versy concerns whether syllables or phonemes (or both) fants new information during a laboratory exposure
are psychologically real to infants. Studies using dis- session. Other studies ask how infants use what theyve
crimination tasks in which either the syllable changes, previously learned about their native language to dis-
or a phonetic feature in the segment changes, provided cover structure in novel input.
convincing evidence that both syllable-level and seg- Many of these studies use artificial nonsense lan-
ment (phoneme) level features are accessed and used guages, a methodology taken from the adult literature, in
by 2- to 4-month-old infants (Eimas & Miller, 1981; which specific cues can be isolated in a way that is im-
Miller & Eimas, 1979). However, studies using similar- possible in natural speech (e.g., Gmez & Gerken, 2000;
ity assessments (Bertoncini, Bijeljac-Babic, Jusczyk, Morgan et al., 1987). Such languages are particularly
Kennedy, & Mehler, 1988; Jusczyk & Derrah, 1987) useful in infant studies because they permit the develop-
yielded a different pattern of results, suggesting that ment of brief exposure materials, fitting the task de-
young infants are sensitive to changes in the number of mands of infants with limited attention spans. On the
syllables in a word, but not the number of phonemes other hand, artificial materials sacrifice ecological va-
(Bijeljac-Babic et al., 1993). Despite the contradictions lidity. An issue currently confronting researchers is the
in the studies with younger infants, older infants in- need to demonstrate that the learning abilities uncov-
clude some subsyllabic structures in their representa- ered using artificial methods are the same as those in-
tions (Jusczyk, Goodman, et al., 1999). Moreover, with fants use when acquiring their native language. While
84 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

ecological validity is always an issue in laboratory learn- to recur cross-linguistically. More generally, studies
ing studies, it is particularly salient given exposure regi- that uncover infants failures may turn out to be as illu-
mens that are so clearly unlike natural language input. minating as those that display infants considerable
strengths at learning, by highlighting constraints on in-
Learning Phonology and Phonotactics fant learning mechanisms.
The acquisition of phonological knowledge is also of
Phonotactic knowledge is a prime candidate for learning great interest, given the rapidity with which infants ac-
studies, because it is so clearly tied to the structure of quire such knowledge in their native language. This
particular languages. Moreover, phonotactics is some- learning process requires infants to integrate different
what different from many of the other features of lan- types of information, an ability that likely emerges be-
guage that infants acquire. Phonotactic patterns are both tween 6 and 9 months of age (e.g., Morgan & Saffran,
general (they apply across the whole language, and are 1995). For example, consider the trochaic biasthe ex-
not specific to known words) and specific (they consist pectation that (at least for English-learning infants)
of segmental patterns, unlike the syllabic and prosodic words begin with stressed syllables. To learn this pat-
patterns that often appear to be the focus of infants at- tern, you must know something about the relationship
tention). Studies of phonotactics in adults suggest a between stressed syllables and their positions within
learning process in which mere exposure to phonotactic words. If you have yet to discover any word boundaries,
regularities influences expectations about possible word it would be impossible to know that stressed syllables
form regularities (Dell, Reed, Adams, & Meyer, 2000; fall in predictable places in words. Thus, to acquire a
Onishi, Chambers, & Fisher, 2002). trochaic bias, infants must learn correspondences be-
Chambers, Onishi, and Fisher (2003) extended their tween stress and word position. This is only possible if
adult studies to include 16.5-month-old infants. The ma- infants first know some trochaic words, which may ex-
terials consisted of nonsense word sequences in which plain the lack of such a bias in 6-month-old infants. In-
consonant positions were restricted; for example, / b/ deed, 612 -month-old infants can be taught a rhythmic
could occur word-initially but not word-finally. Follow- bias by briefly exposing them to word lists exemplifying
ing exposure, infants listened longer to syllables that the bias, and 9-month-olds biases can be similarly al-
were phonotactically legal than those that violated the tered (Thiessen & Saffran, 2004).
exposure patterns. Impressively, infants were also able Other studies investigate how infants acquire more
to learn 2nd order phonotactic regularities, in which the abstract phonological knowledge concerning the stress
presence of one element was conditioned on the pres- assignment patterns of their native language. Gerken
ence of the other (e.g., that / k / begins syllables if and (2004) presented infants with artificial language stimuli
only if the subsequent vowel is /ae/). designed to exhibit particular patterns of metrical
To determine whether certain phonotactic regulari- phonologythe structural principles for stress assign-
ties are harder to learn than others, Saffran and ment in multisyllabic words. Following a brief exposure
Thiessen (2003) exposed infants to two different types to a word list in which certain stress assignments were
of phonotactic patterns. One was consistent with the exemplified, infants were tested to determine whether
types of patterns found cross-linguistically, while the they had inferred stress pattern structures that had not
other was unlike natural language structure. Infants rap- actually occurred in the input. The results suggest that
idly learned regularities of the first type, which in- 9-month-olds generalize to new words using abstract
volved generalizations across sets of acoustic/ linguistic knowledge of possible stress patterns, opening the door
features (such as voicing, the feature that clumps /p/, /t /, to additional studies probing the extent to which infants
and / k / into a separate category from / b/, /d/, and /g/). are able to learn the types of abstract phonological
However, infants failed to learn regularities that disre- structures that typify human languages.
gard such linguistic features (such as the grouping of /p/,
/d/, and / k / versus the grouping of / b/, /t /, and /g/), Word Segmentation
which are unlike natural language patterns. These re-
sults suggest a possible explanation for why languages The problem of how infants discover words in fluent
show the types of patterning that they do. Sound struc- speech, which lacks consistent physical cues to word
tures that are hard for infants to learn may be less likely boundaries (Cole & Jakimik, 1980) has played a promi-
Building from the Input during the 1st Year 85

nent role in studies of early language learning. While in- sounds. To see this statistical structure, consider the fol-
terest in this problem is a relatively recent development lowing example: because the syllable pre precedes a
in the field of language acquisition, there are some no- small set of syllables in English, the probability that pre
table exceptions. Roger Brown began his classic 1973 is followed by ty is quite high. However, because the syl-
volume on language acquisition by describing his own lable ty occurs word-finally, it can be followed by any
problems with word segmentation while taking a Berlitz syllable that can begin an English word. Thus, the prob-
course in Japanese; later he described an early model of ability that ty is followed by ba, as in pretty baby, is ex-
distributional learning in word segmentation by Olivier tremely low. Indeed, infants are sensitive to such
(1968). Gleitman and Wanner (1982) also treated the probabilistic cues, and use them for word segmentation
problem seriously, hypothesizing that stressed syllables (e.g., Aslin, Saffran, & Newport, 1998; Goodsitt, Mor-
may mark words for young learners. These two early gan, & Kuhl, 1993; Saffran et al., 1996).
discussions of the segmentation problem, invoking dis- Several lines of research have converged to suggest
tributional and prosodic cues, were prescient, as these that particular languages contain prosodic cues that facil-
two sources of information are currently at the forefront itate word segmentation. For example, English-learning
of theories regarding infant word segmentation. 712 -month-olds can make use of their knowledge that bi-
In a seminal study, Jusczyk and Aslin (1995) used the syllabic words tend to be trochaic to successfully segment
head-turn preference procedure to determine when in- strong-weak words (those stressed on their first syllable)
fants begin to segment words. They first presented 712 - such as KINGdom from fluent speech, while failing to
month-old infants with a word segmentation problem: segment weak-strong words (those stressed on their sec-
sentences in fluent speech containing a particular target ond syllable) like guiTAR (Jusczyk, Houston, & New-
word (e.g., Mommys cup is on the table. Do you see the some, 1999). In the latter case, infants treat the stressed
cup over there?). Following this familiarization period, syllable TAR as a word. Interestingly, they will com-
infants were tested on the target words (e.g., cup) ver- bine TAR with a subsequent weak syllable if they are
sus novel words (e.g., bike). Each item was played for paired consistently, suggesting the integration of stress-
as long as the infant maintained a head-turn in the direc- based and statistically based strategies. Thus, infants use
tion of a speaker from which the word was played. their expectations about word structure to assist in seg-
Jusczyk and Aslin (1995) found a significant difference mentation (for related results, see also Curtin, Mintz, &
in listening times between the familiar and novel words, Christiansen, 2005; Houston, Jusczyk, Kuijpers, Coolen,
suggesting that the 712 -month-olds discovered the target & Cutler 2000; Houston, Santelmann, & Jusczyk, 2004;
words in fluent speech. Six-month-olds, however, failed Nazzi, Dilley, Jusczyk, Shattuck-Hufnagel, & Jusczyk, in
to show any significant differences between the familiar press).
and novel test items, suggesting either that the ability to Younger infants are unable to take advantage of
segment word from fluent speech develops sometime be- stress-based segmentation cues, demonstrating that this
tween 6 and 712 months of age, or that younger infants knowledge must be learned (e.g., Echols, Crowhurst, &
require additional exposure and/or cues to successfully Childers, 1997; Jusczyk, Houston, et al., 1999; Thiessen
perform the task. Support for the latter view comes from & Saffran, 2003). Further evidence for a learning ac-
a study by Thiessen and Saffran (2003), in which 612 - to count comes from research on languages that incorpo-
7-month-olds successfully performed a word segmenta- rate different stress patterns, such as French (Polka,
tion task in which they received more familiarization Sundara, & Blue, 2002). Artificial language studies also
with the target words, and a study by Bortfeld, Morgan, indicate that stress-based segmentation strategies are
Golinkoff, and Rathbun (2005) demonstrating word seg- learnable (Thiessen & Saffran, 2004). Moreover, infants
mentation by 6-month-olds using additional cues. must learn not to overly focus on stress, which, like all
How do infants solve such a complicated task? A individual cues to word boundaries, is fallible; only by
growing body of evidence suggests that infants are at- 1012 months do infants successfully segment weak-
tuned to a number of cues correlated with word bound- strong words (Jusczyk, Houston, et al., 1999).
aries. One such source of information was initially The fact that infants can use the distribution of stress
suggested in the linguistics literature in the mid-twenti- cues as a cue to word boundaries raises a chicken-and-
eth century (e.g., Harris, 1955), reflecting the observa- egg problem. If stress is a critical cue to word bound-
tion that words consist of predictable sequences of aries, how can infants have discovered the utility of this
86 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

cue prior to knowing words? One must know something criminate the bisyllable /mati/ when the /ma / and the /ti/
about the words of ones native language to discover the are pulled from either a single word or from the final syl-
correlation between stress position and word bound- lable in one word versus the first syllable in the next
aries. One possibility is that infants learn the predomi- (Christophe, Dupoux, Bertoncini, & Mehler, 1994). By 9
nant stress pattern of words in their native language by months of age, infants can detect word boundaries in
hearing words spoken in isolation (e.g., Jusczyk, Hous- contrasts such as nitrates versus night rates, which con-
ton, et al., 1999). This seems intuitively plausible, par- sist of the same sequence of phonemes but different allo-
ticularly given analyses of infant-directed speech phones, suggesting the availability of allophonic cues for
suggesting that a nontrivial proportion of utterances con- segmentation (Jusczyk, Hohne, & Baumann, 1999; Mat-
sist of single words (e.g., Brent & Siskind, 2001). On tys & Jusczyk, 2001). Moreover, at this same age infants
this view, infants might learn words like kitty and phonetic categories reflect sensitivity to position-spe-
mommy by hearing them spoken in isolation, and then cific allophonic variants (Pegg & Werker, 1997). These
use that nascent corpus to discover the stress patterns findings raise the same sort of chicken and egg prob-
characteristic of their native language. However, a recent lem as the stress findingsone must first know some-
analysis suggests that this explanation is unlikely to be thing about words to discover cues correlated with
correct (Swingley, 2005). Only 14% of the bisyllabic ut- internal word structures. It is thus likely not an accident
terances in English spoken to infants are trochaic; most that this ability emerges at roughly the same time for dif-
bisyllables conform to a strong-strong pattern. Thus, in- ferent types of cues. By 9 months of age, infants have
fants must have some other means of discovering the pre- likely segmented enough words using statistical cues and
dominant lexical stress pattern of their native language. other types of information to have developed a suffi-
One possibility is that early access to sequential sta- ciently large corpus to discover these word-internal cues.
tistical segmentation cues provides infants with the be- Phonotactic cues are also correlated with word
ginnings of the corpus they need to subsequently boundaries (e.g., Brent & Cartwright, 1996; Cairns,
discover prosodic regularities. Swingleys (2005) com- Shillcock, Chater, & Levy, 1997; Vitevitch & Luce,
putational analysis suggests that words that are discov- 1998). For example, Mattys et al. (1999) demonstrated
erable via statistical cues render the correct prosodic that infants use the likelihood that particular consonant
template, unlike words heard in isolation. A study by clusters occur within or between words in their native
Thiessen and Saffran (2003) suggests a trajectory of cue language as a segmentation cue. Infants ability to seg-
usage over development consistent with this view. When ment sequences such as nongkuth versus nomkuth was
9-month-olds are confronted with continuous speech in examined. Critically, while the consonant clusters in the
which stress and statistical cues conflict, they follow middle of each sequence are equally likely in English,
the stress cues, as previously demonstrated by Johnson the former is more likely to occur within words (/ngk /)
and Jusczyk (2001). However, 6- to 7-month-olds exhibit while the latter is more likely to span a word boundary
the opposite strategy, relying on statistical cues rather (/mk /). Nine-month-olds used this subtle distinction as a
than stress cues, presumably because they do not yet segmentation cue, inferring word boundaries in the mid-
know their native languages stress pattern. The stress dle of nomkuth but not nongkuth. A segmentation strat-
strategy, then, is presumably bootstrapped from the reg- egy based on phonotactics requires the infant to already
ularities in the initial corpus acquired via sequential sta- know enough words for these regularities to become ap-
tistical cues (Thiessen & Saffran, 2004). parent. Related segmentation cues may require no prior
Other important cues to word boundaries become lexical experience. For example, 12-month-old infants
available to infants beginning around 9 or 10 months of follow the Possible Word Constraint : they generate
age. For example, infants are able to use the distributions segmentations that only create possible words, while
of allophonesthe subtle differences in phonemes that avoiding stranding sequences that are not possible
are a function of the context in which the phoneme oc- words, such as sequences consisting of a single conso-
cursas word boundary cues. Certain sounds only nant (Johnson, Jusczyk, Cutler, & Norris, 2003). This
occur in certain positions in words; the /t / that begins constraint may help infants to segment speech appropri-
English words differs from the /t / that occurs word-me- ately and to avoid errors without requiring a lexicon
dially or word-finally (Church, 1987). Young infants are from which to induce the constraint.
sensitive to allophonic cues which might signal word It should be clear at this point in the discussion that
boundaries, shown, for example, by their ability to dis- no single cue underlies word segmentation. This conclu-
Building from the Input during the 1st Year 87

sion is evident both from the empirical literature ing to the original Jusczyk and Aslin (1995) study, we
demonstrating that infants are sensitive to myriad cues earlier described the results as evidence that infants had
and from the fact that each cue, in isolation, only solves segmented the word cup from the fluent speech. It is
part of the problem for infants. Studies using multiple equally possible, though, that infants test perfor-
cues have largely asked how infants weight conflicting mancediscriminating cup from bikerests on
cues. For example, 6- to 7-month-olds prioritize statis- simply recognizing that the former set of sounds is more
tics over stress, while 9-month-olds prioritize stress familiar than the latter. Doing so would not necessitate
over statistics (Johnson & Jusczyk, 2001; Thiessen & segmentation per se; instead, infants would be respond-
Saffran, 2003). Interestingly, Mattys et al. (1999) found ing based on the familiarity of the sounds, without
that 9-month-olds also prioritize stress over phonotactic having represented cup as a distinct lexical represen-
cues, supporting the hypothesis that, while imperfect, tation. Indeed, one early study suggested that in seg-
stress cues are relatively easy to detect and use mentation tasks, infants pull out metrical feet (a
(Thiessen & Saffran, 2003); Mattys et al. (1999) suggest rhythmical unit) rather than actual words (Myers et al.,
that prosody is an initial cue yielding a coarse first 1996). It is thus of great interest to ask what the output
pass at word boundaries that is subsequently supple- of word segmentation actually is. Saffran (2001) ad-
mented with additional cues such as phonotactic and dressed this issue with respect to the statistical learning
allophonic constraints (p. 482). However, it remains results. When infants respond to golabu during testing,
unknown how such cues are combined in infants emerg- after exposure to golabupabikututipugolabu . . . , are
ing segmentation strategies (see Morgan & Saffran, they treating golabu as a word, or as a familiar sound se-
1995, for an example of a study looking at additive ef- quence? Based on results from a task in which infants
fects of cue combinations). are tested on words like golabu embedded in English
One avenue of research that has effectively explored sentences after exposure, Saffran (2001) suggested that
the use of cue combinations for the discovery of word infants treat these nonsensical patterns as primitive En-
boundaries is the computational literature (for an exten- glish words (i.e., whatever a word is to an 8-month-old,
sive review, see Batchelder, 1997). For example, Chris- in the absence of mapping to meaning).
tiansen et al. (1998), building on the work of Aslin, Recent studies by Curtin et al. (2005) with 7-month-
Woodward, LaMendola, and Bever (1996), examined the old infants further suggest that stress cues are repre-
efficacy of phonotactic cues that predict ends of utter- sented in these newly segmented proto-lexical
ances as a cue to word boundaries in a corpus of child- representations. Curtin et al. used analyses of child-
directed speech. While this cue worked only moderately directed speech to argue that infant learners would be
well in isolation, inclusion of lexical stress cues more successful if they represent stressed and un-
markedly improved the performance of the network. A stressed syllables differently during word segmentation.
different approach to this problem was pursued by In particular, the analyses suggest that incorporation of
Curtin et al. (2005), who found that including stress in- stress into infants representational landscape would re-
formation in a corpus enhanced performance by allow- sult in better distribution-based word segmentation, as
ing the network to represent stressed and unstressed well as an advantage for stress-initial syllable se-
variants of the same syllable as distinct. One of the mes- quences. Results of a behavioral study corroborated
sages provided by the computational literature is that these analyses. In particular, if items in the test phase
more cues are probably better than fewer cues, despite were placed in a sentence context and the target was an
the paradoxical fact that this makes the input more com- exact match (BEdoka) as opposed to a sequence with the
plex. Awaiting future research is the determination of same segments but a different stress pattern (beDOka)
exactly which cues infants attend to, and whether these or another type of nonmatching control sequences, in-
cues are weighted in the manner predicted by the com- fants demonstrated an overwhelming preference for the
putational models. exact match. These results suggest that stress informa-
Throughout this section, we have been discussing tion in the ambient language not only shapes how statis-
word segmentation as though it is clear that infants are tics are calculated over the speech input, but that it is
discovering words in the input, and subsequently repre- also encoded in the representations of parsed speech se-
senting these sound sequences as units, available for quences.
later mapping to meaning. However, it is certainly possi- Once some sequences have been segmented from the
ble that infants are engaged in a simpler process. Return- speech stream to become new lexical entries, can these
88 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

words assist in segmentation of subsequent fluent ity in the input. To explicitly investigate the time course
speech, helping infants to discover other adjacent words of memory for new lexical representations, Jusczyk and
(e.g., Brent & Cartwright, 1996; Dahan & Brent, Hohne (1997) exposed 8-month-olds to stories contain-
1999)? A recent study by Bortfeld et al. (2005) provides ing particular vocabulary items. After 10 days of expo-
evidence that 6-month-old infants can use known words sure to the stories, a 2-week retention interval was
to segment new words from fluent speech. Infants heard introduced, during which infants did not hear the sto-
continuous speech in which the word to be segmented ries. Infants were then tested on their recognition of
appeared adjacent to the infants own name, which in- words from the stories versus similar words that had not
fants recognize early in the 1st year (Mandel, Jusczyk, occurred in the stories. Despite the 2-week retention in-
& Pisoni, 1995). The familiar name served as a strong terval, during which infants heard a vast array of poten-
segmentation cue, providing the first positive evidence tially interfering speech, the infants listened longer to
for word segmentation in infants as young as 6 months. the words from the previously familiarized stories, sug-
By demonstrating that infants prior knowledge alters gesting that auditory representations that were garnered
the manner in which they process new input, the Bort- weeks before were sufficiently robust to support later
feld et al. (2005) results suggest a promising new tact word recognition.
for studies of word segmentation and infant learning How detailed are these early representations? Infants
more generally. appear not to confuse similar sounding words, at least
under certain circumstances. Infants in Jusczyk and
Aslins (1995) experiments did not incorrectly treat
Beginnings of Word Recognition
zeet as familiar after being exposed to feet, sug-
Once infants have segmented words into discrete units, gesting that these early representations are fairly spe-
they are ready to begin recognizing familiar words, cific. Similarly, after repeated exposures to a word and
matching internal representations of words to their in- object, 8-month-old infants show robust evidence of de-
stantiation in subsequent input. This is no simple matter, tecting a change to a new word that differs in only a sin-
because words are not static invariant patterns. The gle phonetic feature (Stager & Werker, 1997).
sounds of any given word are shaped by properties of the Early word representations also appear to include a
speaker (such as speakers voice, sex, speaking rate, and level of acoustic detail that corresponds to the positions
affect) and by the context in which the words are pro- of syllables relative to structural boundaries in sen-
duced (such as coarticulation effects). tences, such as phonological phrases. Acoustic cues cor-
What words might one expect infants to first recog- responding to phonological phrase boundaries are
nize? Mandel et al. (1995) hypothesized that infants detected even by newborn infants (e.g., Christophe
names might be particularly salient. They occur fre- et al., 1994, 2001). By 13 months of age, infants can use
quently, are often presented in isolation, and likely this distinction in word recognition (Christophe, Gout,
carry affective prosody that attracts infants attention. Peperkamp, & Morgan, 2003). For example, infants
Using the preferential listening procedure, 412 -month- trained on the word paper were tested on sentences in
olds heard either their own name or an unfamiliar name. which paper was either a word ( The college with the
Infants preferred to listen to their own names, suggest- biggest paper forms is best ) or in which paper spanned
ing that they matched internal representations of these a phonological phrase boundary ( The butler with the
familiar sounds to the input played during the experi- highest pay performs the best ). Despite the fact that
ment. This does not mean that these infants knew the the syllable sequence was the same in both cases, with
meanings of these sounds. However, by 6 months of age, equivalent statistical and stress cues, the results sug-
infants can recognize highly familiar words based on gested that the infants representations included the sub-
their meanings. When presented with side-by-side video tle acoustic differences between paper and pay per.
displays of their mother and father, infants look longer One source of information that may aid infants in
to the display that matches auditory presentations of speech processing is coarticulation. In order to produce
mommy versus daddy (Tincoff & Jusczyk, 1999). speech as rapidly as we do, whenever we produce a seg-
These results suggest that infants developing lexical ment, syllable, or word, we move the lips, tongue, and
representations are not fleeting, but are built up incre- jaw in a way that maintains the positions required for
mentally and maintained over time, despite the variabil- that segment as well as for both the preceding and fol-
Building from the Input during the 1st Year 89

lowing consonants and vowels. For example, because of nize words. For example, although 14-month-olds can
coarticulation, the phoneme / b/ is different in the word learn to associate two different nonsense words with
beet than in the word boot. Adults are sensitive to two different objects (Schafer & Plunkett, 1998;
this coarticulatory information, but only under some lis- Werker, Cohen, Lloyd, Casasola, & Stager, 1998; Wood-
tening conditions. For example, when words in a string ward, Markman, & Fitzsimmons, 1994), they fail at this
are presegmented by the insertion of pauses, adult lis- same age if the two nonsense words are phonetically
teners show better recognition of those familiar sylla- similar such as bih and dih (Stager & Werker, 1997)
bles that maintain the same coarticulatory information or pin and din (Pater, Stager, & Werker, 2004). Im-
as used during familiarization. However, when the portantly, 14-month-old infants succeed in a virtually
pauses are omitted and adults must rely on only transi- identical task when the word is paired with a visual dis-
tional information, their access to coarticulatory infor- play that is unlikely to evoke labeling (Stager & Werker,
mation is no longer evident (Curtin, Werker, & Ladhar, 1997). Moreover, when an easier variant of the task was
2002). Seven-month-old infants are also sensitive to used in which a single object was paired with a single
coarticulatory information, but under the opposite con- word, the 14-month-olds still failed to notice the change
ditions as adults. When the syllables are presegmented to a phonetically similar word, whereas 8-month-olds
by the insertion of pauses, infants recognition of famil- succeeded in this same task. The failure to learn mini-
iar words is not enhanced by matching coarticulatory mally contrastive words was shown to be short-lived.
cues. However, in tasks that require the infant to segment When tested in exactly the same task, 17- and 20-
syllables from a continuous stream of speech, matching month-old infants succeeded at learning phonetically
coarticulatory information significantly improves per- similar words (Werker, Fennell, Corcoran, & Stager,
formance (Curtin et al., 2002). 2002), as did even infants of 14 months who had partic-
Infants also appear to represent indexical information ularly sizeable vocabularies (Werker et al., 2002; see
that affects word recognition. For example, 712 -month- also Beckman & Edwards, 2000 for a discussion of the
old infants readily recognize words previously heard potential role of vocabulary size). An identical pattern
produced by a speaker of the same sex, but show no evi- of results was obtained using an ERP paradigm in which
dence of word recognition when the target is produced a higher amplitude deflection is seen to known versus
by a speaker of the opposite sex (Houston & Jusczyk, unknown words (Mills et al., 2004).
2000). It seems likely that infants representations Why might 14-month-old infants fail to distinguish
include perceptual features of the speaker, such as com- phonetically similar words in a word-learning task when
ponents of pitch, that make cross-sex matching challeng- they could still discriminate these two words, and when
ing. Similarly, 712 -month-olds represent the affective both younger and slightly older (or even more advanced
state of the speaker, showing word recognition only same-aged infants) succeed? Stager and Werker (1997)
when the affective state of the familiarized words speculated that for the novice word learners, the compu-
matched the targets (Singh, Morgan, & White, 2004). tational demands of linking a word with an object are so
Related arguments are emerging in the field of infant great that attentional resources are not available to uti-
music perception, where researchers are actively inves- lize all the word-level detail that is perceived (see Kah-
tigating the grain at which infants represent musical neman, 1973, for the original postulation of attention as
experiences in memory for subsequent recognition (e.g., a limited resource). However, other interpretations of
Ilari & Polka, 2002; Palmer, Jungers, & Jusczyk, 2001; these findings were that they revealed evidence of a dis-
Saffran, Loman, & Robertson, 2001; Trainor, Wu, & continuity between the representations used in phonetic
Tsang, 2004). versus phonological (or lexical) representations. Indeed,
there is a long-standing tradition in child phonology that
Listening for Meaning posits such a representational discontinuity (see Brown
& Matthews, 1997; Rice & Avery, 1995; Shvachkin,
In contrast to the detailed, multiple levels of informa- 1948). Empirical work by Hall and de Boysson-Bardies
tion available to prelinguistic infants in word recogni- (1994) provided potential support for this discontinuity
tion and segmentation tasks, infants who have begun to hypothesis. They used a word recognition task requiring
assemble a more sizeable lexicon seem to be more selec- infants to listen to lists of highly familiar versus
tive and more limited in which detail they use to recog- unknown words, and found that although infants of 7
90 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

and 11 months both showed a preference when the un- informationphonetic, indexical, coarticulatoryis in-
known words were phonetically dissimilar from the cluded in the lexicon (e.g., Goldinger, 1992), but that not
known words (Hall & de Boysson-Bardies, 1994), if all of this information is used in every task situation
phonetically similar foils were used, only the 7-month- (Werker & Curtin, 2005).
old infants succeeded. Evidence in support of this possibility is provided in
A number of subsequent studies have now discon- recent work with 2.5- and 3-year-olds by Fisher, Church,
firmed the discontinuity hypothesis. When tested in a and Chambers (2004). They demonstrated that children
simpler word recognition procedure wherein infants are represent both abstract and detailed linguistic informa-
shown two pictures and presented with a single word tion pertaining to the specifics of pronunciation of
that either matches one of the objects or is a mispronun- familiar words. For example, their participants repre-
ciation of the same name for that object (e.g., baby sented the distinction between a medial /t / and a more
versus vaby), infants from 20 (Swingley & Aslin, /d/-like flap pronunciation of the same phoneme, despite
2000) down to 14 months (Swingley & Aslin, 2002) can the fact that both pronunciations are legal. Interestingly,
detect the mispronunciation. Sometimes this detection the same pattern of results emerged in a related study
is shown in longer looking to the correct object when the using nonwords, suggesting that even new lexical repre-
word is pronounced correctly (Swingley & Aslin, 2002), sentationsformed after just a few exposures to a
and sometimes it is evident in a shorter latency to look wordare flexible, in that they are both abstract and
away from the mismatch (Swingley & Aslin, 2000), but specific (Fisher, Hunt, Chambers, & Church, 2001).
it is consistently evident. This success is seen for well- These findings suggest that perceptual learning mecha-
known words in the associative task used by Stager and nisms used flexibly throughout life to adapt to new lin-
Werker (1997). If habituated to the word ball paired guistic input may operate from the beginning of the
with the moving object ball, and the word doll word learning process (Fisher et al., 2004).
paired with a visual display of a doll, 14-month-old It may be that in the earliest stages of word learning,
infants detect a switch in the word object pairing in the infants are less able to flexibly select which information
test phase (Fennell & Werker, 2003). Thus, as suggested to attend to. With the attentional resource demands of
by both Stager and Werker (1997) and Swingley and attaching meaning to words, infants at the cusp of word
Aslin (2000), it appears that an attentional resource lim- learning may be captured by that information which is
itation rather than a representational discontinuity ac- most salient. To test this hypothesis, Curtin and Werker
counts for the failure of 14-month-old infants under (cited in Werker & Curtin, 2005) recently tested the
some circumstances. ability of 12-month-old infants to learn words that are
An attentional resource limitation may not fully ex- similar in all respects except stress pattern. They found
plain the preceding findings. It is still necessary to know that these infants, a full 2 months younger than the in-
why it is that phonetic detail is dropped. Is this the only fants who failed to learn phonetically similar words,
detail that infants drop at 14 months, or is other detail could successfully learn to map words such as DObita
also ignored? The word segmentation and recognition versus doBIta (where capitals indicate stress) onto two
studies revealed that 7- to 9-month-old infants utilize different objects.
many different kinds of information in the signal. This With these studies, the links between infant speech
is evident, for example, in their failure to recognize perception, word segmentation, word recognition, and
words if there is a change in speaker gender (Houston & word learning are being much more fully described.
Jusczyk, 2000), or affect (Singh, Bortfeld, & Morgan, Moreover, the infant literature is beginning to interface
2002). However, by 1012 months of age, infants success- much more richly with the large literature on adult lexi-
fully recognize words spoken by opposite-sex speakers cal access. Infancy researchers are no longer restricted
(Houston & Jusczyk, 2000), and are able to ignore to asking questions like what is the unit of representa-
changes in affect and still show evidence of recognizing tion? Instead, the field is now poised to allow the ask-
familiar words (Singh et al., 2002). One increasingly ing of much more nuanced questions such as what
popular account, which is somewhat different from the information is utilized, when, and why?
traditional view in which indexical information is not The advent of new methodologies has allowed re-
part of lexical representations, is that a rich tapestry of searchers to go beyond asking which words infants rec-
Building from the Input during the 1st Year 91

ognize to assess the time course of word recognition. cal response to the change, indicating that at some level
Eye-tracking has become an important tool in assessing in the brain the information is available (Rivera-Gaxiola
adults lexical representations (e.g., Allopenna, Magnu- et al., 2005).
son, & Tanenhaus, 1998; Tanenhaus, Spivey-Knowlton,
Eberhard, & Sedivy, 1995). Adapting these methods to Beginnings of Grammar
study infants, Fernald, Pinto, Swingley, Weinberg, and
McRoberts (1998) assessed infants speed and accuracy Since most infants do not begin combining words gram-
at word recognition over the course of the 2nd year. To matically until the ripe old age of 18 to 24 months or be-
do so, the experimenters measured infants eye move- yond, is there any reason to suspect that the capacity to
ments as they viewed computer-displayed pictures of fa- acquire grammatical structure is present earlier in life?
miliar objects while listening to the names of these Indeed, researchers have demonstrated early evidence
objects. Speed and reliability were correlated with age, for grammatical knowledge of the native language dur-
suggesting that infants lexical representations likely be- ing infancy, as well as precocious abilities to learn new,
come more robust and the cognitive machinery underly- simple, grammatical structures using artificial language
ing word recognition becomes more fluent during the methodologies (see Tomasello, Chapter 6, this Hand-
2nd year. Like adults, 24-month-olds do not need to hear book, this volume, for a review of the literature on sub-
an entire word to recognize it; instead, word recognition sequent aspects of grammar learning). Comprehension
is incremental (Swingley, Pinto, & Fernald, 1999). For studies suggest that infants have a sophisticated grasp of
example, these infants rapidly distinguished doggie from certain syntactic structures by the end of the 2nd year.
tree, correctly fixating on the matching picture, but took For example, Naigles (1990) tested young 2-year-olds in
300 ms longer to distinguish doggie from doll, reflecting a cross-modal matching task that required them to in-
the increased phonetic overlap of the latter pair. Inter- duce the meaning of a new verb. The infants heard either
estingly, infants recognize parts of words just as rapidly transitive structures, such as The duck is kradding the
as whole words, supporting the view that infants, like bunny, or intransitive structures, such as The duck
adults, process words incrementally (Fernald, Swingley, and bunny are kradding. Infants looked longer at a
& Pinto, 2001). This ability appears to be associated video that matched the syntactic structure of the sen-
with infants productive vocabularies, suggesting a link tence they heard. These results suggest that infants can
between lexical growth and the efficiency with which engage in what is known as syntactic bootstrapping:
infants recognize words. These facts about how infants using their prior knowledge of syntactic syntax ( here,
process words are consistent with corpus analyses sug- transitivity) to determine the meaning of kradding.
gesting that the words in infants early vocabularies are Infants morphological knowledge is similarly ad-
sufficiently overlapping in phonological space to neces- vanced. For example, Santelmann and Jusczyk (1998)
sitate detailed lexical representations (e.g., Coady & exposed infants to passages that contained either a
Aslin, 2003). grammatical English dependency between the auxiliary
Experience with particular words appears to enhance verb is and a main verb ending with -ing, or an ungram-
these nascent lexical representations. Church and Fisher matical combination of the modal auxiliary can and a
(1998) observed long-term auditory priming in 2- to 3- main verb ending with -ing. Eighteen-month-olds, but
year-olds very similar to that of adults, showing effects not 15-month-olds, discriminated between the two types
of experience with specific words on subsequent word of passages. These results suggest that by the middle of
identification and repetition. Similar effects emerged in the 2nd year, infants have learned how certain types of
a study with 18-month-olds using a preferential looking discontinuous grammatical dependencies operate in
task, suggesting that just two repetitions of a word as- their native language.
sisted infants in subsequently identifying the target A number of recent studies have employed artificial
word (Fisher et al., 2004). Similarly, neuroimaging tasks grammar methodologies to uncover the learning mecha-
are broadening the range of questions that can be ex- nisms underlying this process. When exposed to word
plored. For example, infants may fail to discriminate a sequences ordered by simple rules (e.g., Marcus et al.,
nonnative contrast given the task demands of a behav- 1999) or finite state grammars (Gmez & Gerken,
ioral task, yet still show evidence of a neurophysiologi- 1999), infants treat test items that violate those patterns
92 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

as novel, even if they are instantiated in new vocabulary. processes unfold, it is evident that they are heavily mul-
For example, Gmez and Gerken (1999) used the head- tidetermined, influenced by factors from the develop-
turn preference procedure to assess whether 12-month- ment of the peripheral auditory system to the nature of
olds could acquire a miniature artificial grammar. our learning mechanisms. We close by considering some
Infants discriminated new grammatical strings from limitations on these processes, which may be very im-
ungrammatical strings after less than 2 minutes of train- portant for future work aimed toward illuminating the
ing, with evidence that they acquired both specific in- nature of infants accomplishments.
formation (e.g., legal beginnings and ends of sentences,
and internal pair-wise combinations) and abstract infor- Relationship between Auditory Processing and
mation (e.g., grammatical structures produced using a Speech Perception
new set of vocabulary). Ongoing research is probing the
circumstances under which infants are more or less That even newborns can discriminate between speech
likely to generalize beyond the input given (e.g., Gmez, sounds and recognize voices has led many to believe that
2002; Gmez & Maye, 2005). hearing does not constrain speech perception or learn-
Saffran and Wilson (2003) extended this line of re- ing during infancy. It is clear, however, that several as-
search to ask how infants might approach learning tasks pects of hearing remain immature early in infancy, and
consisting of multiple levels of information. Twelve- it is likely that these immaturities do constrain speech
month-olds listened to a continuous speech stream in perception to some extent. There are suggestions in the
which the words were ordered via a finite-state gram- literature that 2-month-olds, for example, represent
mar. The infants were thus presented concurrently with speech with less detail than older infants do (Bertoncini
a word segmentation task and a syntax learning task. et al., 1988; Bijeljac-Babic et al., 1993). It is likely that
The results suggest that infants can first segment novel some aspects of speech perception and language learn-
words and then discover syntactic regularities relating ing are delayed until 6 months, when representations of
the new wordsall within the same set of input. Studies the acoustic characteristics of sound are adultlike. In
of this type indicate that artificial learning situations any case, it should be possible to make predictions about
can be scaled up to begin to represent some of the prob- young infants speech discrimination abilities based on
lems confronting learners faced with natural language what is known about their hearing, and to test specifi-
input. For example, Gerken, Wilson, and Lewis (2005) cally to determine whether hearing immaturity has any
performed a hybrid artificial /natural language learning bearing on early speech perception.
study in which infants heard a small subset of Russian A related question is whether infants use the same in-
words marked with correct gender morphology. The re- formation in speech as adults do when they are discrim-
sults demonstrate that certain types of patterns that inating between speech sounds. Because there are
occur in natural language input ( here, redundant cues) multiple cues to phonetic identity, it is possible that in-
play an important role in learning, as indicated by per- fants use cues that they hear better, or that they attend to
formance in this lab-based learning task. more salient cues and ignore others, or that they weight
all cues equally. That infants do not attend to the compo-
nents of a complex sound as adults do in a simple
CONCLUSIONS AND FUTURE DIRECTIONS psychophysical task (Bargones et al., 1995; Bargones &
Werner, 1994; Leibold & Werner, 2003; Werner &
As we hope has been reflected throughout this chapter, Boike, 2001) suggests that their approach to speech may
infants accomplishments in the auditory domain are differ from that of adults. Nittrouers studies of speech
nothing short of extraordinary. In the absence of exter- discrimination in children suggest that preschool chil-
nal guidance or reinforcement, our perceptual systems dren do not, in fact, weight cues to phonetic identity as
hone in on the dimensions of the auditory environment adults do (e.g., Nittrouer, Crowther, & Miller, 1998;
that are most relevant for the development of our com- Nittrouer & Miller, 1997; Nittrouer & Studdert-
municative capacity, and we learn extremely complex Kennedy, 1987). It would be surprising to find, then,
and detailed information about how our auditory envi- that infants weight cues in an adultlike way. There are
ronment is structured, all during our 1st postnatal year. now correlational techniques that can be used to assess
While much remains to be learned about how these the weights that listeners place on various components
Conclusions and Future Directions 93

of a complex sound in making discriminations and these The extent to which similar findings emerge with young
techniques have been successfully applied to young chil- learners is the object of active research
dren (e.g., Stellmack et al., 1997). An interesting prob-
lem in the future will be to apply these techniques to
infants, particularly in the realm of speech perception. Domain Specificity and Species Specificity

Much of the foregoing discussion has focused on learn-


Constraints on Learning ing from the input, and the types of information cap-
tured by infant learning mechanisms. A critical open
Much of the previous discussion has focused on infants question is the degree to which this learning is sub-
remarkable capacity to glean structure from complex served by mechanisms tailored for speech and language.
input. However, it is important to note that demonstra- One possibility is that, perhaps due to the adaptive sig-
tions of powerful learning mechanisms alone do not rep- nificance of human communication systems, we have
resent a satisfying solution to the problems facing young evolved sophisticated learning machinery specifically
language learners. How do learners hone in on the right tailored for language. Alternatively, these early learning
patterns and structures given the massive amount of processes may tap mechanisms that are available for
data in the input? The richness of the stimulus prob- more general tasks.
lem is that there are an infinite number of patterns that A growing body of results suggests that at least one of
an unbiased learner might detect. Clearly, human in- the learning mechanisms we have discussed, sequential
fants are not such learners, and it is incumbent upon re- statistical learning, is quite general. For example, infants
searchers to show not just all the things that infants can can track sequences of musical tones, discovering tone-
learn, but also what infants find more difficult to learn, word boundaries via statistical cues (e.g., Saffran,
to elucidate the limits on learning. It is also possible to 2003a; Saffran & Griepentrog, 2001; Saffran, Johnson,
ask how the structure of the task itself affects the types Aslin, & Newport, 1999), and can learn statistically
of learning that occur, as some types of input may elicit defined visual patterns (e.g., Fiser & Aslin, 2002;
different learning mechanisms than others (e.g., Pea Kirkham, Slemmer, & Johnson, 2002). These findings
et al., 2003; Saffran, Reeck, Niehbur, & Wilson, 2005). and others suggest that at least these basic learning
Thus far, this research strategy has primarily been processes are not tailored solely for language acquisition
carried out with adult learners, with implications to be (e.g., Saffran, 2002, 2003b).
drawn for infant learners. For example, Newport and Another source of evidence bearing on this issue
Aslin (2004) demonstrated that while adults readily comes from studies of nonhuman primates. Hauser and
track the dependencies between adjacent syllables (e.g., his colleagues (Hauser, Newport, & Aslin, 2001;
the probability that pa is followed by bu), they do not do Hauser, Weiss, & Marcus, 2002) have tested cotton-top
so when the relevant dependency skips an intervening tamarins, a new world monkey species, on the linguistic
syllable. Such nonadjacent dependencies are apparently tasks used by Saffran et al. (1996) and Marcus et al.
not automatically tracked by learners. Interestingly, (1999). Intriguingly, the monkeys showed the same pat-
however, adults do detect nonadjacent dependencies tern of performance as human infants, despite their pre-
when the intervening material is different in kind. For sumed lack of evolved abilities to acquire human
example, adults can detect dependencies between two language (Hauser et al., 2001, 2002). Even rats detect
consonants with intervening vowels, or two vowels with some language-relevant patterns (Toro & Trobalan,
intervening consonants (Newport & Aslin, 2004). Be- 2004)! These findings reinforce the view that at least
cause these latter types of structures recur in human some of the learning mechanisms that subserve the be-
languages (in Semitic languages, and in languages like ginnings of language learning are not evolutionary adap-
Turkish that use vowel harmony), while the former do tations specialized for the linguistic domain.
not, Newport and Aslin (2004) suggest that languages Results like these lead immediately to the question of
may be constrained by the limits on human learning. why, if monkeys share our learning machinery, language
That is, only those structures that are learnable by hu- is uniquely human. That is, if monkeys learn like us,
mans persist in our languages. Saffran (2002) makes a shouldnt they be as linguistically sophisticated as we
similar argument based on adult grammatical studies. are? Several avenues of explanation are currently being
94 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

explored. One, of course, is the traditional view that hu- ever, human social interaction was required in the pre-
mans possess innate linguistic knowledge that other sentation of the Mandarin input. When infants received
species lack (e.g., Pinker, 1984). Other investigators are the same input via high-quality DVD recordings, no im-
focusing on the degree to which human learning mecha- pact on their speech perception was observed. These re-
nisms may in fact diverge from those possessed by other sults suggest that, like some species of birds, the
species. For example, tamarins and human adults do not learning system requires a certain type of interactive
show the same pattern of learning of nonadjacencies dis- input to affect perception. If this is the case, then differ-
cussed in the previous section, suggesting that the con- ences in social interaction may also help to explain some
straints on human learning mechanisms may diverge cross-species differences in who learns what.
from those seen in other species (Newport, Hauser,
Spaepen, & Aslin, 2004). Similarly, Hauser, Chomsky,
The Infants Auditory World
and Fitch (2002) have suggested that while humans and
nonhumans may share much of their learning machinery, In this review, we have considered recent developments
humans are differentiated by their ability to perform re- in our understanding of how infants begin to make
cursion operationsthe capacity to generate an infinite sense of their auditory environments. A great deal of
range of expressions from a finite set of elements (see progress has been made in elucidating the basic sensory
also Fitch & Hauser, 2004). On this view, humans and and perceptual mechanisms that provide auditory input
nonhumans should show similar performance when to infant learners, as well as the learning mechanisms
learning about such things as speech contrasts and word that track this input and integrate it with infants exist-
segmentation, and diverge as grammatical complexity ing knowledge.
increases (e.g., Saffran, Hauser, Seibel, Kapf hamer, In future studies, we expect that the relationship be-
Tsao, & Cushman, 2005). While these central questions tween infants auditory abilities and the rest of language
remain to be resolved, their answers are likely to have acquisition (see Tomasello, Chapter 6; Waxman & Lidz,
broad impact on such issues as the modularity of mind Chapter 7, this Handbook, this volume) will become
and the ontogenesis of specific domains of knowledge. clearer. Audition is the gateway to spoken language, and
There is another component to early language learn- infants early accomplishments in acquiring the sound
ing that may be relevant to species differencessocial structure of their native language(s) lay critical ground-
interaction between the speaker and the learner. While work for subsequent learning. Recent studies linking the
such issues as joint attention have played a prominent acquisition of sound structure to later accomplishments
role in the literature on how young children map sound in word learning provide important suggestions about
to meaning (e.g., Harris, Chapter 19; Tomasello, Chap- how infants early abilities are likely to influence later
ter 6, this Handbook, this volume), the role of social in- language learning (e.g., Hollich, Jusczyk, & Luce, 2002;
teraction has not received significant attention in the Saffran & Graf Estes, 2004; Swingley & Aslin, 2002;
literature on how infants acquire sound structure itself. Thiessen, 2004; Werker et al., 2002). For example, early
Certainly, there is ample evidence that caregivers speech perception abilities may predict some aspects of
manipulate the input so it is well tailored to infants per- word learning many months later (Tsao, Liu, & Kuhl,
ceptual predilections (e.g., Kuhl et al., 1997; see Tre- 2004). Similarly, researchers are beginning to investi-
hub, 2003, for related evidence in the domain of music gate the effects of the amelioration of early sensory
perception). The higher pitches and enhanced pitch con- deprivation via cochlear implants on subsequent audi-
tours of infant directed speech are well-established at- tory perception and language learning abilities (Hous-
tention-getters and affect communicators (e.g., Cooper ton, Pisoni, Kirk, Ying, & Miyamoto, in press). Such
& Aslin, 1990; Fernald, 1992). Intriguing new results integrative research enterprises will serve to illuminate
suggest, however, that the role of social interaction ex- the links between the talents of infant listeners in the au-
tends beyond the sound structure of the input. Kuhl, ditory realm and the many linguistic (and nonlinguistic)
Tsao, and Liu (2003) manipulated infants perception of tasks that lie ahead of them. Similarly, much remains to
nonnative speech contrasts such that English-learning be learned about the neural underpinnings of the abili-
infants maintained a Mandarin speech contrast well be- ties described throughout our review, and knowledge
yond the age at which their ability to discriminate the about these neural substrates will help us to better un-
contrast would typically have declined. Critically, how- derstand the behaviors that they subserve. Many fasci-
References 95

nating open questions thus remain, and in the next edi- Aslin, R. N., & Pisoni, D. B. (1980). Some developmental processes in
tion of the Handbook, we hope to read the answersin- speech perception. In G. H. Yeni-Komshian, J. F. Kavanagh, &
C. A. Ferguson (Eds.), Child phonology: Vol. 2. Perception
cluding the answers to the many questions that we do not (pp. 6796). New York: Academic Press.
yet know to ask. Aslin, R. N., Pisoni, D. B., Hennessy, B. L., & Percy, A. J. (1981).
Discrimination of voice onset time by human infants: New find-
ings and implications for the effects of early experience. Child
REFERENCES Development, 52(4), 11351145.
Aslin, R. N., Saffran, J. R., & Newport, E. L. (1998). Computation of
conditional probability statistics by 8-month-old infants. Psycho-
Abdala, C. (1998). A developmental study of distortion product otoa-
logical Science, 9(4), 321324.
coustic emission (2f1-f2) suppression in humans. Hearing Re-
search, 121, 125138. Aslin, R. N., Woodward, J. Z., LaMendola, N. P., & Bever, T. G.
Abdala, C. (2001). Maturation of the human cochlear amplifier: Dis- (1996). Models of word segmentation in f luent maternal speech to
tortion product otoacoustic emission suppression tuning curves infants. In J. L. Morgan & K. Demuth (Eds.), Signal to syntax:
recorded at low and high primary tone levels. Journal of the Bootstrapping from speech to grammar in early acquisition
Acoustical Society of America, 110, 14651476. (pp. 117134). Hillsdale, NJ: Erlbaum.
Abdala, C., & Chatterjee, M. (2003). Maturation of cochlear nonlin- Bargones, J. Y., & Burns, E. M. (1988). Suppression tuning curves for
earity as measured by distortion product otoacoustic emission spontaneous otoacoustic emissions in infants and adults. Journal
suppression growth in humans. Journal of the Acoustical Society of of the Acoustical Society of America, 83, 18091816.
America, 114, 932943. Bargones, J. Y., & Werner, L. A. (1994). Adults listen selectively: In-
Abdala, C., & Folsom, R. C. (1995a). The development of frequency fants do not. Psychological Science, 5, 170174.
resolution in humans as revealed by the auditory brain-stem re- Bargones, J. Y., Werner, L. A., & Marean, G. C. (1995). Infant psy-
sponse recorded with notched-noise masking. Journal of the chometric functions for detection: Mechanisms of immature sen-
Acoustical Society of America, 98, 921930. sitivity. Journal of the Acoustical Society of America, 98, 99111.
Abdala, C., & Folsom, R. C. (1995b). Frequency contribution to the Batchelder, E. O. (1997) Computational evidence for the use of fre-
click-evoked auditory brain stem response in human adults and in- quency information in discovery of the infants first lexicon. Un-
fants. Journal of the Acoustical Society of America, 97, 23942404. published doctoral dissertation, New York, City University.
Abercrombie, D. (1967). Elements of general phonetics. Edinburgh, Beckman, M. E., & Edwards, J. (2000). The ontogeny of phonological
Scotland: Edinburgh University Press. categories and the primacy of lexical learning in linguistic devel-
Allen, P., Jones, R., & Slaney, P. (1998). The role of level, spectral, opment. Child Development, 71(1), 240249.
and temporal cues in childrens detection of masked signals. Jour- Benson, R. R., Whalen, D. H., Richardson, M., Swainson, B., Clark,
nal of the Acoustical Society of America, 104, 29973005. V. P., Lai, S., et al. (2001). Parametrically dissociating speech
Allen, P., & Wightman, F. (1992). Spectral pattern discrimination by and nonspeech perception in the brain using fMRI. Brain and
children. Journal of Speech and Hearing Research, 35, 222233. Language, 78, 364396.
Allen, P., & Wightman, F. (1994). Psychometric functions for chil- Berg, K. M., & Boswell, A. E. (1995). Temporal summation of 500-
drens detection of tones in noise. Journal of Speech and Hearing Hz tones and octave-band noise bursts in infants and adults. Per-
Research, 37, 205215. ception and Psychophysics, 57, 183190.
Allen, P., & Wightman, F. (1995). Effects of signal and masker un- Berg, K. M., & Boswell, A. E. (1998). Infants detection of incre-
certainty on childrens detection. Journal of Speech and Hearing ments in low- and high-frequency noise. Perception and Psy-
Research, 38, 503511. chophysics, 60, 10441051.
Allen, P., Wightman, F., Kistler, D., & Dolan, T. (1989). Frequency Berg, K. M., & Boswell, A. E. (1999). Effect of masker level on
resolution in children. Journal of Speech and Hearing Research, infants detection of tones in noise. Perception and Psy-
32, 317322. chophysics, 61, 8086.
Allopenna, P. D., Magnuson, J. S., & Tanenhaus, M. K. (1998). Berg, K. M., & Boswell, A. E. (2000). Noise increment detection in chil-
Tracking the time course of spoken word recognition: Evidence dren 1 to 3 years of age. Perception and Psychophysics, 62, 868873.
for continuous mapping models. Journal of Memory and Language, Berg, K. M., & Smith, M. C. (1983). Behavioral thresholds for
38, 419439. tones during infancy. Journal of Experimental Child Psychology,
Altmann, G. T. M., & Dienes, Z. (1999). Rule learning by 7-month- 35, 409425.
old infants and neural networks. Science, 284, 875. Bertoncini, J., Bijeljac-Babic, R., Blumstein, S., & Mehler, J. (1987).
Anderson, J. L., Morgan, J. L., & White, K. S. (2003). A statistical Discrimination of very short CV syllables by neonates. Journal of
basis for speech sound discrimination. Language and Speech, the Acoustical Society of America, 82, 3137.
46(2/3), 155182. Bertoncini, J., Bijeljac-Babic, R., Jusczyk, P. W., Kennedy, L. J., &
Ashmead, D. H., Clifton, R. K., & Perris, E. E. (1987). Precision of Mehler, J. (1988). An investigation of young infants perceptual
auditory localization in human infants. Developmental Psychol- representations of speech sounds. Journal of Experimental Psy-
ogy, 23, 641647. chology: General, 117(1), 2133.
Ashmead, D. H., Davis, D., Whalen, T., & Odom, R. (1991). Sound lo- Bertoncini, J., & Mehler, J. (1981). Syllables as units in infant speech
calization and sensitivity to interaural time differences in human perception. Infant Behavior and Development, 4(3), 247260.
infants. Child Development, 62, 12111226. Bertoncini, J., Morais, J., Bijeljac-Babic, R., McAdams, S., Peretz, I.,
Aslin, R. N. (1989). Discrimination of frequency transitions by human & Mehler, J. (1989). Dichotic perception and laterality in
infants. Journal of the Acoustical Society of America, 86, 582590. neonates. Brain and Language, 37(4), 591605.
96 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

Best, C. T. (1994). The emergence of native-language phonological Brown, A. M., Sheppard, S. L., & Russell, P. (1994). Acoustic Dis-
inf luences in infants: A perceptual assimilation model. In J. C. tortion Products (ADP) from the ears of term infants and young
Goodman & H. C. Nusbaum (Eds.), The development of speech adults using low stimulus levels. British Journal of Audiology,
perception: The transition from speech sounds to spoken words 28, 273280.
(pp. 167224). Cambridge, MA: MIT Press. Brown, C., & Matthews, J. (1997). The role of feature geometry in
Best, C. T., Hoffman, H., & Glanville, B. B. (1982). Development of the development of phonemic contrasts. In S. J. Hannahs & M.
infant ear asymmetries for speech and music. Perception and Psy- Young-Scholten (Eds.), Focus on phonological acquisition: Vol.
chophysics, 31(1), 7585. 16. Language acquisition and language disorders (pp. 67112).
Best, C. T., & McRoberts, G. W. (2003). Infant perception of nonna- Amsterdam: Benjamins.
tive contrasts that adults assimilate in different ways. Language Brown, R. (1973). A first language: The early stages. Cambridge,
and Speech, 46(2/3), 183216. MA: Harvard University Press.
Best, C. T., McRoberts, G. W., LaFleur, R., & Silver Isenstadt, J. Bull, D., Eilers, R. J., & Oller, D. K. (1984). Infants discrimination
(1995). Divergent developmental patterns for infants perception of intensity variation in multisyllabic stimuli. Journal of the
of two nonnative consonant contrasts. Infant Behavior and Devel- Acoustic Society of America, 76(1), 1317.
opment, 18(3), 339350. Burnham, D., & Dodd, B. (2004). Audiovisual speech perception by
Best, C. T., McRoberts, G. W., & Sithole, N. M. (1988). Examination prelinguistic infants: Perception of an emergent consonant in the
of perceptual reorganization for nonnative speech contrasts: Zulu McGurk effect. Developmental Psychobiology, 45(4), 202220.
click discrimination by English-speaking adults and infants. Jour- Burnham, D., Kitamura, C., & Vollmer-Conna, U. (2002).
nal of Experimental Psychology: Human Perception and Perfor- Whats new, pussycat? On talking to babies and animals. Sci-
mance, 14(3), 345360. ence, 296(5572), 1435.
Bijeljac-Babic, R., Bertoncini, J., & Mehler, J. (1993). How do 4-day- Burns, E. M., Campbell, S. L., & Arehart, K. H. (1994). Longitudinal
old infants categorize multisyllabic utterances? Developmental measurements of spontaneous otoacoustic emissions in infants.
Psychology, 29(4), 711721. Journal of the Acoustical Society of America, 95, 384394.
Binder, J. R., Frost, J. A., Hammeke, T. A., Bellgowan, P. S. F., Buss, E., Hall, J. W., Grose, J. H., & Dev, M. B. (1999). Development
Springer, J. A., Kaufman, J. N., et al. (2000). Human temporal of adult-like performance in backward, simultaneous, and for-
lobe activation by speech and nonspeech sounds. Cerebral Cortex, ward masking. Journal of Speech Language and Hearing Research,
10(5), 512528. 42, 844849.
Binder, J. R., Frost, J. A., Hammeke, T. A., Cox, R. W., Rao, S. M., & Cairns, P., Shillcock, R., Chater, N., & Levy, J. (1997). Bootstrapping
Prieto, T. (1997). Human brain language areas identified by func- word boundaries: A bottom-up corpus-based approach to speech
tional magnetic resonance imaging. Journal of Neuroscience, segmentation. Cognitive Psychology, 33(2), 111153.
17(1), 353362. Chambers, K. E., Onishi, K. H., & Fisher, C. (2003). Infants learn
Binns, K. E., Withington, D. J., & Keating, M. J. (1995). The devel- phonotactic regularities from brief auditory experiences. Cogni-
opmental emergence of the representation of auditory azimuth in tion, 87(2), B69B77.
the external nucleus of the inferior colliculus of the guinea-pig: Cheour, M., Ceponiene, R., Lehtokoski, A., Luuk, A., Allik, J.,
The effects of visual and auditory deprivation. Developmental Alho, K., et al. (1998). Development of language-specific
Brain Research, 85, 1424. phoneme representations in the infant brain. Nature Neuroscience,
1(5), 351353.
Bond, B., & Stevens, S. S. (1969). Cross-modality matching of
brightness to loudness by 5-year-olds. Perception and Psy- Cheour-Luhtanen, M., Alho, K. l., Kuijala, T., Sainio, K.,
chophysics, 6, 337339. Reinikainen, K., et al. (1995). Mismatch negativity indicates vowel
discrimination in newborns. Hearing Research, 82(1), 5358.
Bonfils, P., Avan, P., Francois, M., Trotoux, J., & Narcy, P. (1992).
Distortion-product otoacoustic emissions in neonates: Normative Christiansen, M. H., Allen, J., & Seidenberg, M. S. (1998). Learning
data. Acta Otolaryngologica, 112, 739744. to segment speech using multiple cues: A connectionist model.
Language and Cognitive Processes, 13, 23.
Bonfils, P., Francois, M., Avan, P., Londero, A., Trotoux, J., & Narcy,
Christiansen, M. H., & Curtin, S. L. (1999). Transfer of learning:
P. (1992). Spontaneous and evoked otoacoustic emissions in
Rule acquisition or statistical learning? Trends in Cognitive Sci-
preterm neonates. Laryngoscope, 102, 182186.
ences, 3, 289290.
Bortfeld, H., Morgan, J., Golinkoff, R., & Rathbun, K. (2005).
Christophe, A., Dupoux, E., Bertoncini, J., & Mehler, J. (1994). Do
Mommy and me: Familiar names help launch babies into speech
infants perceive word boundaries? An empirical study of the
stream segmentation. Manuscript in preparation.
bootstrapping of lexical acquisition. Journal of the Acoustical So-
Bosch, L., & Sebastin-Galls, N. (2001). Early language differentia- ciety of America, 95(3), 15701580.
tion in bilingual infants. In J. Cenoz & F. Genesee (Eds.), Trends
Christophe, A., Gout, A., Peperkamp, S., & Morgan, J. (2003). Dis-
in bilingual acquisition (pp. 7193). Amsterdam: Benjamins.
covering words in the continuous speech stream: The role of
Bredberg, G. (1968). Cellular pattern and nerve supply of the human prosody. Manuscript in preparation.
organ of Corti. Acta Otolaryngologica (Suppl.), 236. Christophe, A., Mehler, J., & Sebastin-Galls, N. (2001). Percep-
Bregman, A. S. (1990). Auditory scene analysis: The perceptual orga- tion of prosodic boundary correlates by newborn infants. Infancy,
nization of sound. Cambridge, MA: MIT Press. 2, 358394.
Brent, M. R., & Cartwright, T. A. (1996). Distributional regularity Church, B. A., & Fisher, C. (1998). Long-term auditory word priming
and phonotactic constraints are useful for segmentation. Cogni- in preschoolers: Implicit memory support for language acquisi-
tion, 61(1/2), 93125. tion. Journal of Memory and Language, 39, 523542.
Brent, M. R., & Siskind, J. M. (2001). The role of exposure to isolated Church, K. W. (1987). Phonological parsing and lexical retrieval.
words in early vocabulary development. Cognition, 81(2), B33B44. Cognition, 25, 5369.
References 97

Clarkson, M. G., & Clifton, R. K. (1985). Infant pitch perception: Evi- Cutler, A., & Norris, D. (1988). The role of strong syllables in seg-
dence for responding to pitch categories and the missing fundamen- mentation for lexical access. Journal of Experimental Psychology:
tal. Journal of the Acoustical Society of America, 77, 15211528. Human Perception and Performance, 14(1), 113121.
Clarkson, M. G., & Clifton, R. K. (1995). Infants pitch perception: Dahan, D., & Brent, M. R. (1999). On the discovery of novel wordlike
Inharmonic tonal complexes. Journal of the Acoustical Society of units from utterances: An artificial-language study with implica-
America, 98(3), 13721379. tions for native-language acquisition. Journal of Experimental
Clarkson, M. G., Clifton, R. K., & Perris, E. E. (1988). Infant timbre Psychology: General, 128(2), 165185.
perception: Discrimination of spectral envelopes. Perception and Dai, H., Scharf, B., & Buus, S. (1991). Effective attenuation of sig-
Psychophysics, 43, 1520. nals in noise under focused attention. Journal of the Acoustical So-
Clarkson, M. G., & Rogers, E. C. (1995). Infants require low-fre- ciety of America, 89, 28372842.
quency energy to hear the pitch of the missing fundamental. Jour- DeCasper, A. J., & Fifer, W. P. (1980). Of human bonding: Newborns
nal of the Acoustical Society of America, 98, 148154. prefer their mothers voices. Science, 208(4448), 11741176.
Clifton, R. K., Gwiazda, J., Bauer, J., Clarkson, M., & Held, R. DeCasper, A. J., & Prescott, P. (1984). Human newborns perception
(1988). Growth in head size during infancy: Implications for of male voices: Preference, discrimination and reinforcing value.
sound localization. Developmental Psychology, 24, 477483. Developmental Psychobiology, 17, 481491.
Clifton, R. K., Morrongiello, B. A., & Dowd, J. M. (1984). A devel- DeCasper, A. J., & Spence, M. J. (1986). Prenatal maternal speech
opmental look at an auditory illusion: The precedence effect. De- inf luences newborns perception of speech sounds. Infant Behav-
velopmental Psychobiology, 17, 519536. ior and Development, 9, 133150.
Clifton, R. K., Morrongiello, B. A., Kulig, J. W., & Dowd, J. M. Dehaene, S., Dupoux, E., Mehler, J., Cohen, L., Paulesu, D., Perani,
(1981). Developmental changes in auditory localization in infancy. D., et al. (1997). Anatomical variability in the cortical represen-
In R. Aslin, J. Alberts, & M. R. Petersen (Eds.), Development of tation of first and second languages. Neuroreport: For Rapid Com-
perception (Vol. 2, pp. 141160). New York: Academic Press. munication of NeuroScience Research, 8, 38093815.
Clifton, R. K., Perris, E. E., & Bullinger, A. (1991). Infants percep-
Dehaene-Lambertz, G., & Baillet, S. (1998). A phonological repre-
tion of auditory space. Developmental Psychology, 27, 187197.
sentation in the infant brain. NeuroReport, 9(8), 18851888.
Coady, J. A., & Aslin, R. N. (2003). Phonlogical neighbourhoods in
Dehaene-Lambertz, G., & Dehaene, S. (1994). Speed and cerebral
the developing lexicon. Journal of Child Language, 30, 441469.
correlates of syllable discrimination in infants. Nature,
Coady, J. A., & Aslin, R. N. (in press). Young childrens sensitivity to 370(6487), 292295.
probabilistic phonotactics in the developing lexicon. Journal of
Dehaene-Lambertz, G., Dehaene, S., & Hertz-Pannier, L. (2002).
Experimental Child Psychology.
Functional neuroimaging of speech perception in infants. Science,
Cole, R., & Jakimik, J. (1980). A model of speech perception. Hills- 298(5600), 20132015.
dale, NJ: Erlbaum.
Dehaene-Lambertz, G., & Pea, M. (2001). Electrophysiological evi-
Collins, A. A., & Gescheider, G. A. (1989). The measurement of
dence for automatic phonetic processing in neonates. NeuroRe-
loudness in individual children and adults by absolute magnitude
port, 12(14), 31553158.
estimation and cross-modality matching. Journal of the Acoustical
Society of America, 85, 20122021. Dell, G. S., Reed, K. D., Adams, D. R., & Meyer, A. S. (2000).
Speech errors, phonotactic constraints, and implicit learning: A
Colombo, J. A., & Bundy, R. S. (1981). A method for the measure-
study of the role of experience in language production. Journal of
ment of infant auditory selectivity. Infant Behavior and Develop-
Experimental Psychology: Learning, Memory, and Cognition,
ment, 4(2), 219223.
26(6), 13551367.
Colombo, J., & Horowitz, F. D. (1986). Infants attentional re-
Demany, L. (1982). Auditory stream segregation in infancy. Infant
sponses to frequency modulated sweeps. Child Development,
Behavior and Development, 5, 261276.
57(2), 287291.
Cooper, R. P., & Aslin, R. N. (1990). Preference for infant-di- Demany, L. (1985). Perceptual learning in frequency discrimination.
rected speech in the first month after birth. Child Development, Journal of the Acoustical Society of America, 78, 11181120.
61(5), 15841595. Demany, L., McKenzie, B., & Vurpillot, E. (1977). Rhythm percep-
Cooper, R. P., & Aslin, R. N. (1994). Developmental differences in tion in early infancy. Science, 266, 718719.
infant attention to the spectral properties of infant-directed Desjardins, R. N., Rogers, J., & Werker, J. F. (1997). An exploration
speech. Child Development, 65(6), 16631677. of why preschoolers perform differently than do adults in audio-
Curtin, S., Mintz, T. H., & Christiansen, M. H. (2005). Stress visual speech perception tasks. Journal of Experimental Child
changes the representational landscape: Evidence from word seg- Psychology, 66(1), 85110.
mentation. Cognition, 96, 233262. Desjardins, R., & Werker, J. F. (2004). Is the integration of heard and
Curtin, S., Werker, J. F., & Ladhar, N. (2002). Accessing coarticu- seen speech mandatory for infants? Developmental Psychobiology,
latory information. Journal of the Acoustical Society of Amer- 45(4), 187203.
ica, 112, 2359. Dewson, J. H. (1964). Speech sound discrimination by cats. Science,
Cutler, A., & Carter, D. M. (1987). The predominance of strong ini- 141, 555556.
tial syllables in the English vocabulary. Computer Speech and Dooling, R. J., Best, C. T., & Brown, S. D. (1995). Discrimination of
Language, 2, 34. synthetic full-formant and sinewave/rala /continua by budgerigars
Cutler, A., Mehler, J., Norris, D., & Segui, J. (1983). A language-spe- (Melopsittacus undulatus) and zebra finches (Taeniopygia guttata).
cific comprehension strategy. Nature, 304(5922), 159160. Journal of the Acoustical Society of America, 97(3), 18391846.
Cutler, A., Mehler, J., Norris, D., & Segui, J. (1986). The syllables Durieux-Smith, A., Edwards, C. G., Picton, T. W., & McMurray, B.
differing role in the segmentation of French and English. Journal (1985). Auditory brainstem responses to clicks in neonates. Jour-
of Memory and Language, 25(4), 385400. nal of Otolaryngology, 14, 1218.
98 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

Echols, C. H., Crowhurst, M. J., & Childers, J. B. (1997). The per- Fisher, C., & Tokura, H. (1996). Acoustic cues to grammatical struc-
ception of rhythmic units in speech by infants and adults. Journal ture in infant-directed speech: Cross-linguistic evidence. Child
of Memory and Language, 36, 202225. Development, 67(6), 31923218.
Echols, C. H., & Newport, E. L. (1992). The role of stress and position Fisher, C. H., Hunt, C. M., Chambers, C. K., & Church, B. (2001). Ab-
in determining first words. Language Acquisition, 2(3), 189220. straction and specificity in preschoolers representations of novel
Eddins, D. A., & Green, D. M. (1995). Temporal integration and tem- spoken words. Journal of Memory and Language, 45(4), 665687.
poral resolution. In B. C. J. Moore (Ed.), Hearing (pp. 207242). Fitch, W. T., & Hauser, M. D. (2004). Computational constraints on
San Diego, CA: Academic Press. syntactic processing in a nonhuman primate. Science, 303, 377380.
Eimas, P. D. (1974). Auditory and linguistic processing of cues for Floccia, C., Nazzi, T., & Bertoncini, J. (2000). Unfamiliar voice dis-
place of articulation by infants. Perceptual Psychophysiology, crimination for short stimuli in newborns. Developmental Science,
16, 513521. 3(3), 333343.
Eimas, P. D. (1975a). Auditory and phonetic coding of the cues for Folsom, R. C., & Wynne, M. K. (1987). Auditory brain stem re-
speech: Discrimination of the {r-l} distinction by young infants. sponses from human adults and infants: Wave 5Tuning curves.
Perception and Psychophysics, 18(5), 341347. Journal of the Acoustical Society of America, 81, 412417.
Eimas, P. D. (1975b). Speech perception in early infancy. In L. B. Fowler, C. A., Smith, M. R., & Tassinary, L. G. (1986). Perception of
Cohen & P. Salapatek (Eds.), Infant perception: From sensation to syllable timing by prebabbling infants. Journal of the Acoustical
cognition (pp. 193231). New York: Academic Press. Society of America, 79, 814825.
Eimas, P. D., & Miller, J. L. (1981). Organization in the perception of Friederici, A. D., & Wessels, J. M. I. (1993). Phonotactic knowledge
segmental and suprasegmental information by infants. Infant Be- of word boundaries and its use in infant speech perception. Per-
havior and Development, 4, 395399. ception and Psychophysics, 54(3), 287295.
Eimas, P. D., & Miller, J. L. (1992). Organization in the perception of Fujikawa, S. M., & Weber, B. A. (1977). Effects of increased stimulus
speech by young infants. Psychological Science, 3, 340345. rate on brainstem electric response (BER) audiometry as a func-
Eimas, P. D., Miller, J. L., & Jusczyk, P. W. (1987). On infant tion of age. Journal of the American Audiology Society, 3, 147150.
speech perception and the acquisition of language. In S. Harnad Fujimoto, S., Yamamoto, K., Hayabuchi, I., & Yoshizuka, M. (1981).
(Ed.), Categorical perception: The groundwork of cognition Scanning and transmission electron microscope studies on the
(pp. 161195). New York: Cambridge University Press. organ of corti and stria vascularis in human fetal cochlear ducts.
Eimas, P. D., Siqueland, E. R., Jusczyk, P., & Vigorito, J. (1971). Archives of Histology Japan, 44, 223235.
Speech perception in infants. Science, 171(968), 303306. Fujita, A., Hyde, M. L., & Alberti, P. W. (1991). ABR latency in in-
Elfenbein, J. L., Small, A. M., & Davis, M. (1993). Developmental fants: Properties and applications of various measures. Acta Oto-
patterns of duration discrimination. Journal of Speech and Hear- laryngologica, 111, 5360.
ing Research, 36, 842849. Genesee, F. (1989). Early bilingual development: One language or
Elman, J. L. (1990). Finding structure in time. Cognitive Science, two? Journal of Child Language, 16, 161179.
14, 179211. Gerken, L. A. (2004). Nine-month-olds extract structural principles
Fassbender, C. (1993). Auditory grouping and segregation processes in required for natural language. Cognition, 93, B89B96.
infancy. Norderstedt, Germany: Kaste Verlag. Gerken, L., Jusczyk, P. W., & Mandel, D. R. (1994). When prosody
Fay, R. R., & Coombs, S. (1983). Neural mechanisms in sound detec- fails to cue syntactic structure: 9-month-olds sensitivity to
tion and temporal summation. Hearing Research, 10, 6992. phonological versus syntactic phrases. Cognition, 51(3), 237265.
Fennell, C. T., & Werker, J. F. (2003). Early word learners ability to Gerken, L. A., & McIntosh, B. J. (1993). The interplay of function
access phonetic detail in well-known words. Language and morphemes and prosody in early language. Developmental Psy-
Speech, 46(2/3), 245264. chology, 29, 448457.
Fernald, A. (Ed.). (1984). The perceptual and af fective salience of Gerken, L. A., Wilson, R., & Lewis, W. (2005). 17-month-olds can
mothers speech to infants. Norwood, NJ: Ablex. use distributional cues to form syntactic categories. Journal of
Fernald, A. (1992). Prosody in speech to children: Prelinguistic and Child Language, 32, 249268.
linguistic functions. Annals of Child Development, 8. Ghazanfar, A. A., & Logothetis, N. K. (2003). Facial expressions
Fernald, A., Pinto, J. P., Swingley, D., Weinberg, A., & McRoberts, linked to monkey calls. Nature, 423, 937938.
G. W. (1998). Rapid gains in speed of verbal processing by infants Glanville, B. B., Levenson, R., & Best, C. T. (1977). A cardiac mea-
in the 2nd year. Psychological Science, 9(3), 228231. sure of cerebral asymmetries in infant auditory perception. Devel-
Fernald, A., Swingley, D., & Pinto, J. P. (2001). When half a word is opmental Psychology, 13(1), 5449.
enough: Infants can recognize spoken words using partial pho- Gleitman, L. R., & Wanner, E. (1982). Language acquisition: The
netic information. Child Development, 72(4), 10031015. state of the state of the art. In E. Wanner & L. R. Wanner (Eds.),
Fiez, J. A., Tallal, P. A., Raichle, M. E., Miezin, F. M., Katz, W., Language acquisition: The state of the art (pp. 348). Cambridge,
Dobmeyer, S., et al. (1995). PET studies of auditory and phono- England: Cambridge University Press.
logical processing: Effects of stimulus type and task condition. Glenn, S. M., Cunningham, C. C., & Joyce, P. F. (1981). A study of
Journal of Cognitive Neuroscience, 7, 357375. auditory preferences on nonhandicapped infants with Downs
Fiser, J., & Aslin, R. N. (2002). Statistical learning of new visual fea- syndrome. Child Development, 52, 13031307.
ture combinations by infants. Proceedings of the National Acad- Goldinger, S. D. (1992). Words and voices: Implicit and explicit
emy of Sciences, 99, 1582215826. memory for spoken words. Dissertation Abstracts International,
Fisher, C., Church, B., & Chambers, K. E. (2004). Learning to iden- 53(6), 3189.
tify spoken words. In D. G. Hall & S. R. Waxman (Eds.), Weaving Gmez, R. L. (2002). Variability and detection of invariant structure.
a lexicon (pp. 340). Cambridge, MA: MIT Press. Psychological Science, 13(5), 431436.
References 99

Gmez, R. L., & Gerken, L. (1999). Artificial grammar learning by ous masking in 6- to 10-year-old children. Journal of Speech Lan-
1-year-olds leads to specific and abstract knowledge. Cognition, guage and Hearing Research, 43, 14021415.
70(2), 109135. Hauser, M. D., Chomsky, N., & Fitch, W. T. (2002). The faculty of
Gmez, R. L., & Gerken, L. A. (2000). Infant artificial language language: What is it, who has it, and how did it evolve? Science,
learning and language acquisition. Trends in Cognitive Sciences, 298(5598), 15691579.
4, 178186. Hauser, M. D., Newport, E. L., & Aslin, R. N. (2001). Segmentation
Gmez, R. L., & Maye, J. (2005). The developmental trajectory of of the speech stream in a nonhuman primate: Statistical learning
non-adjacent dependency learning. Infancy, 7, 183206. in cotton-top tamarins. Cognition, 78, B53B64.
Goodsitt, J. V., Morgan, J. L., & Kuhl, P. K. (1993). Perceptual strate- Hauser, M. D., Weiss, D., & Marcus, G. (2002). Rule learning by cot-
gies in prelingual speech segmentation. Journal of Child Lan- ton-top tamarins. Cognition, 86, B15B22.
guage, 20, 229252. Hayes, R. A., Slater, A., & Brown, E. (2000). Infants ability to cate-
Gorga, M. P., Kaminski, J. R., Beauchaine, K. L., Jesteadt, W., & gorise on the basis of rhyme. Cognitive Development, 15(4), 405419.
Neely, S. T. (1989). Auditory brainstem responses from children Hillenbrand, J. A. (1984). Speech perception by infants: Categoriza-
3 months to 3 years of age: Vol. 2. Normal patterns of response. tion based on nasal consonant place of articulation. Journal of the
Journal of Speech and Hearing Research, 32, 281288. Acoustical Society of America, 75(950), 16131622.
Gorga, M. P., Reiland, J. K., Beauchaine, K. A., Worthington, D. W., Hirsh-Pasek, K., Kemler Nelson, D. G., Jusczyk, P. W., Wright Cas-
& Jesteadt, W. (1987). Auditory brainstem responses from gradu- sidy, K., Druss, B., & Kennedy, L. (1987). Clauses are perceptual
ates of an intensive care nursery: Normal patterns of response. units for young infants. Cognition, 26, 269286.
Journal of Speech and Hearing Research, 30, 311318. Hockley, N. S., & Polka, L. (1994). A developmental study of Audiovi-
Gottlieb, G. (1976). The roles of experience in the development of be- sual Speech Perception Using the McGurk Paradigm. Unpublished
havior and the nervous system. In G. Gottlieb (Ed.), Neural and masters thesis, McGill University, Montreal, Canada.
behavioral specificity (pp. 2553). New York: Academic Press. Hhle B., & Weissenborn J. (2003). German-learning infants ability
Gray, L. (1992). Interactions between sensory and nonsensory factors to detect unstressed closed-class elements in continuous speech.
in the responses of newborn birds to sound. In L. A. Werner & Developmental Science, 6(2), 122127.
E. W. Rubel (Eds.), Developmental psychoacoustics (pp. 89112). Hohne, E. A., & Jusczyk, P. W. (1994). Two-month-old infants sensi-
Washington, DC: American Psychological Association. tivity to allophonic differences. Perception and Psychophysics,
Green, K. P. (1998). The use of auditory and visual information dur- 56, 613623.
ing phonetic processing: Implications for theories of speech per- Hollich, G., Jusczyk, P., & Luce, P. (2002). Lexical neighborhood ef-
ception. In R. Campbell, B. Dodd, & D. Burnham (Eds.), Hearing fects in 17-month-old word learning. Proceedings of the 26th An-
by eye: Vol. 2. Advances in the psychology of speechreading and au- nual Boston University Conference on Language Development
ditory-visual speech (pp. 325). Hove, England: Psychology Press. (pp. 314323). Boston: Cascadilla Press.
Greenberg, G. Z., Bray, N. W., & Beasley, D. S. (1970). Childrens Hoshino, T. (1990). Scanning electron microscopy of nerve fibers in
frequency-selective detection of signals in noise. Perception and human fetal cochlea. Journal of Electron Microscopy Technique,
Psychophysics, 8, 173175. 15, 104114.
Grieser, D., & Kuhl, P. K. (1989). Categorization of speech by in- Houston, D. M., & Jusczyk, P. W. (2000). The role of talker-spe-
fants: Support for speech-sound prototypes. Developmental Psy- cific information in word segmentation by infants. Journal of
chology, 25(4), 577588. Experimental Psychology: Human Perception and Performance,
Grose, J. H., Hall, J. W. I., & Gibbs, C. (1993). Temporal analysis in 26(5), 15701582.
children. Journal of Speech and Hearing Research, 36, 351356. Houston, D. M., Jusczyk, P. W., Kuijpers, C., Coolen, R., & Cutler, A.
Hall, J. W., & Grose, J. H. (1990). The masking level difference in (2000). Cross-language word segmentation by 9-month-olds. Psy-
children. Journal of the American Academy of Audiology, 1, 8188. chonomic Bulletin and Review, 7(3), 504509.
Hall, J. W., & Grose, J. H. (1991). Notched-noise measures of fre- Houston, D. M., Pisoni, D. B., Kirk, K. I., Ying, E. A., & Miyamoto,
quency selectivity in adults and children using fixed-masker-level R. T. (in press). Speech perception skills of deaf infants following
and fixed-signal-level presentation. Journal of Speech and Hear- cochlear implantation: A first report. International Journal of Pe-
ing Research, 34, 651660. diatric Otorhinolaryngology.
Hall, J. W., & Grose, J. H. (1994). Development of temporal resolution Houston, D. M., Santelmann, L., & Jusczyk, P. W. (2004). English-
in children as measured by the temporal modulation transfer func- learning infants segmentation of trisyllabic words from f luent
tion. Journal of the Acoustical Society of America, 96, 150154. speech. Language and Cognitive Processes, 19, 97136.
Hall, J. W., Grose, J. H., & Dev, M. B. (1997). Auditory development Igarashi, Y., & Ishii, T. (1979). Development of the cochlear and the
in complex tasks of comodulation masking release. Journal of blood-vessel-network in the fetus: A transmission electrographic
Speech Language and Hearing Research, 40, 946954. observation. Audiology, Japan, 22, 459460.
Hall, P. A., & de Boysson-Bardies, B. (1994). Emergence of an early Igarashi, Y., & Ishii, T. (1980). Embryonic development of the human
receptive lexicon: Infants recognition of words. Infant Behavior organ of Corti: Electron microscopic study. International Journal
and Development, 17, 119129. of Pediatric Otorhinolaryngology, 2, 5162.
Harris, Z. (1955). From phoneme to morpheme. Language, 31, 190222. Igarashi, Y., Yamazaki, H., & Mitsui, T. (1978). An electronographic
Hartley, D. E. H., & Moore, D. R. (2002). Auditory processing effi- study of inner/outer haircells of human fetuses. Audiology, Japan,
ciency deficits in children with developmental language impair- 21, 375377.
ments. Journal of the Acoustical Society of America, 112, 29622966. Ilari, B., & Polka, L. (2002). Memory for music in infancy: The role of
Hartley, D. E. H., Wright, B. A., Hogan, S. C., & Moore, D. R. (2000). style and complexity. Paper presented at the International Confer-
Age-related improvements in auditory backward and simultane- ence on Infant Studies, Toronto, Ontario, Canada.
100 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

Irwin, R. J., Stillman, J. A., & Schade, A. (1986). The width of the Jusczyk, P. W., Pisoni, D. B., Walley, A., & Murray, J. (1980). Discrim-
auditory filter in children. Journal of Experimental Child Psychol- ination of relative onset time of two-component tones by infants.
ogy, 41, 429442. Journal of the Acoustical Society Of America, 67(1), 262270.
Jensen, J. K., & Neff, D. L. (1993). Development of basic auditory dis- Kaga, K., & Tanaka, Y. (1980). Auditory brainstem response and be-
crimination in preschool children. Psychological Science, 4, 104107. havioral audiometry: Developmental correlates. Archives of Oto-
Jiang, Z. D., Wu, Y. Y., & Zhang, L. (1990). Amplitude change with laryngology, 106, 564566.
click rate in human brainstem auditory-evoked responses. Audiol- Kahneman, D. (1973). Attention and ef fort. Englewood Cliffs, NJ:
ogy, 30, 173182. Prentice-Hall.
Jiang, Z. D., Wu, Y. Y., Zheng, W. S., Sun, D. K., Feng, L. Y., & Liu, Karzon, R. G. (1985). Discrimination of polysyllabic sequences by 1-
X. Y. (1991). The effect of click rate on latency and interpeak in- to 4-month-old infants. Journal of Experimental Child Psychology,
terval of the brain-stem auditory evoked potentials in children 39, 326342.
from birth to 6 years. Electroencephalography and Clinical Neuro- Karzon, R. G., & Nicholas, J. G. (1989). Syllabic pitch perception
physiology, 80, 6064. in 2- to 3-month-old infants. Perception and Psychophysics,
Johnson, E. K., & Jusczyk, P. W. (2001). Word segmentation by 8- 45(1), 1014.
month-olds: When speech cues count more than statistics. Journal Keefe, D. H., Bulen, J. C., Arehart, K. H., & Burns, E. M. (1993). Ear-
of Memory and Language, 44(4), 548567. canal impedance and ref lection coefficient in human infants and
Johnson, E. K., Jusczyk, P. W., Cutler, A., & Norris, D. (2003). Lexi- adults. Journal of the Acoustical Society of America, 94, 26172638.
cal viability constraints on speech segmentation by infants. Cog- Keefe, D. H., Burns, E. M., Bulen, J. C., & Campbell, S. L. (1994).
nitive Psychology, 46(1), 6597. Pressure transfer function from the diffuse field to the human
Jusczyk, P. W. (1997). Finding and remembering words: Some begin- infant ear canal. Journal of the Acoustical Society of America,
nings by English-learning infants. Current Directions in Psycho- 95, 355371.
logical Science, 6(6), 170174. Keefe, D. H., Folsom, R. C., Gorga, M. P., Vohr, B. R., Bulen, J. C., &
Jusczyk, P. W., & Aslin, R. N. (1995). Infants detection of the sound Norton, S. J. (2000). Identification of neonatal hearing impair-
patterns of words in f luent speech. Cognitive Psychology, 29, 123. ment: Ear-canal measurements of acoustic admittance and re-
Jusczyk, P. W., Copan, H., & Thompson, E. (1978). Perception by 2- f lectance in neonates. Ear and Hearing, 21, 443461.
month-old infants of glide contrasts in multisyllabic utterances. Keefe, D. H., & Levi, E. C. (1996). Maturation of the middle and ex-
Perception and Psychophysics, 24(6), 515520. ternal ears: Acoustic power-based responses and ref lectance typ-
Jusczyk, P. W., Cutler, A., & Redanz, N. J. (1993). Infants prefer- manometry. Ear and Hearing, 17, 113.
ence for the predominant stress patterns of English words. Child Kelly, M. H. (1992). Using sound to solve syntactic problems: The
Development, 64, 675687. role of phonology in grammatical category assignments. Psycho-
Jusczyk, P. W., & Derrah, C. (1987). Representation of speech sounds logical Review, 99(2), 349364.
by young infants. Developmental Psychology, 23, 648654. Kidd, G., Jr., Mason, C. R., & Arbogast, T. L. (2002). Similarity,
Jusczyk, P. W., Friederici, A. D., Wessels, J. M., Svenkerud, V. Y., uncertainty, and masking in the identification of nonspeech
& Jusczyk, A. M. (1993). Infants sensitivity to the sound pat- auditory patterns. Journal of the Acoustical Society of America,
terns of native language words. Journal of Memory and Language, 111, 13671376.
32, 402420. King, A. J., Hutchings, M. E., Moore, D. R., & Blakemore, C. (1988).
Jusczyk, P. W., Goodman, M. B., & Baumann, A. (1999). Nine- Developmental plasticity in the visual and auditory representa-
month-olds attention to sound similarities in syllables. Journal of tions in the mammalian superior colliculus. Nature, 332, 7376.
Memory and Language, 40(1), 6282. Kirkham, N. Z., Slemmer, J. A., & Johnson, S. P. (2002). Visual sta-
Jusczyk, P. W., Hirsh-Pasek, K., Nelson, D. G. K., Kennedy, L. J., tistical learning in infancy: Evidence for a domain general learn-
Woodward, A., & Piwoz, J. (1992). Perception of acoustic corre- ing mechanism. Cognition, 83(2), B35B42.
lates of major phrasal units by young infants. Cognitive Psychol- Klein, A. J., Alvarez, E. D., & Cowburn, C. A. (1992). The effects of
ogy, 24, 252293. stimulus rate on detectability of the auditory brain stem response
Jusczyk, P. W., & Hohne, E. A. (1997). Infants memory for spoken in infants. Ear and Hearing, 13, 401405.
words. Science, 277, 19841986. Kluender, K. R., Diehl, R. L., & Killeen, P. R. (1987, September 4). Japa-
Jusczyk, P. W., Hohne, E. A., & Baumann, A. (1999). Infants sensi- nese quail can learn phonetic categories. Science, 237, 11951197.
tivity to allophonic cues to word segmentation. Perception and Kluender, K. R., Lotto, A. J., Holt, L. L., & Bloedel, S. L. (1998).
Psychophysics, 61, 14651476. Role of experience for language-specific functional mappings
Jusczyk, P. W., Houston, D. M., & Newsome, M. (1999). The begin- of vowel sounds. Journal of the Acoustical Society of America,
nings of word segmentation in English-learning infants. Cognitive 104, 35683582.
Psychology, 39(3/4), 159207. Kopyar, B. A. (1997). Intensity discrimination abilities of infants and
Jusczyk, P. W., & Krumhansl, C. L. (1993). Pitch and rhythmic pat- adults: Implications for underlying processes. Unpublished doc-
terns affecting infants sensitivity to musical phrase structure. toral dissertation, University of Washington, Seattle.
Journal of Experimental Psychology: Human Perception and Per- Krentz, U. C., & Corina, D. C. (2005). Preference for language in
formance, 19(3), 627640. early infancy: The human language bias is not speech specific.
Jusczyk, P. W., Luce, P. A., & Charles-Luce, J. (1994). Infants sensi- Manuscript submitted for publication.
tivity to phonotactic patterns in the native language. Journal of Krumhansl, C. L., & Jusczyk, P. W. (1990). Infants perception of
Memory and Language, 33(5), 630645. phrase structure in music. Psychological Science, 1(1), 7073.
Jusczyk, P. W., Pisoni, D. B., Reed, M. A., Fernald, A., & Myers, M. Kuhl, P. K. (1979). Speech perception in early infancy: Perceptual
(1983). Infants discrimination of the duration of a rapid spec- constancy for spectrally dissimilar vowel categories. Journal of
trum change in nonspeech signals. Science, 222, 175177. the Acoustical Society of America, 66(6), 16681679.
References 101

Kuhl, P. K. (1991). Human adults and human infants show a percep- Lavigne-Rebillard, M., & Pujol, R. (1988). Hair cell innervation in
tual magnet effect for the prototypes of speech categories, mon- the fetal human cochlea. Acta Otolaryngologica, Stockholm,
keys do not. Perception and Psychophysics, 50(2), 93107. 105, 398402.
Kuhl, P. K. (2000). Language, mind, and brain: Experience alters per- Lavigne-Rebillard, M., & Pujol, R. (1990). Auditory hair cells in
ception. In M. S. Gazzaniga (Ed.), The new cognitive neuro- human fetuses: Synaptogenesis and ciliogenesis. Journal of Elec-
sciences (2nd ed., pp. 99115). Cambridge, MA: MIT Press. tron Microscopy Technique, 15, 115122.
Kuhl, P. K., Andruski, J. E., Chistovich, I. A., Chistovich, L. A., Lecanuet, J.-P., Granier-Deferre, C., & Busnel, M.-C. (1988). Fetal
Kozhevnikova, E. V., Ryskina, V. L., et al. (1997). Cross-language cardiac and motor responses to octave-band noises as a function
analysis of phonetic units in language addressed to infants, Sci- of central frequency, intensity and heart rate variability. Early
ence, 277, 684686. Human Development, 18, 8193.
Kuhl, P. K., & Meltzoff, A. N. (1982). The bimodal perception of Leibold, L., & Werner, L. A. (2002). Relationship between intensity
speech in infancy. Science, 218(4577), 11381141. and reaction time in normal hearing infants and adults. Ear and
Kuhl, P. K., & Meltzoff, A. N. (1988). Speech as an intermodal object Hearing, 23, 9297.
of perception. In A. Yonas (Ed.), Minnesota Symposia on Child Leibold, L., & Werner, L. A. (2003). Infants detection in the pres-
Psychology: Perceptual development in infancy (pp. 235266). ence of masker uncertainty. Journal of the Acoustical Society of
Hillsdale, NJ: Erlbaum. America, 113, 2208.
Kuhl, P. K., & Miller, J. D. (1978). Speech perception by the chin- Levi, E. C., & Werner, L. A. (1996). Amplitude modulation detection
chilla: Identification function for synthetic VOT stimuli. Journal in infancy: Update on 3-month-olds. Abstracts of the Association
of the Acoustical Society of America, 63, 905917. for Research in Otolaryngology, 19, 142.
Kuhl, P. K., & Padden, D. M. (1983). Enhanced discriminability at Liberman, A. M., Cooper, F. S., Shankweiler, D. P., & Studdert-
the phonetic boundaries for the place feature in macaques. Jour- Kennedy, M. (1967). Perception of the speech code. Psychological
nal of the Acoustical Society of America, 73, 10031010. Review, 74, 431461.
Kuhl, P. K., Tsao, F.-M., & Liu, H.-M. (2003). Foreign-language ex- Lisker, L., & Abramson, A. S. (1971). Distinctive features and laryn-
perience in infancy: Effects of short-term exposure and social in- geal control. Language, 47, 767785.
teraction on phonetic learning. Proceedings of the National Litovsky, R. Y. (1997). Developmental changes in the precedence ef-
Academy of Sciences, USA, 100, 90969101. fect: Estimates of minimum audible angle. Journal of the Acousti-
Kuhl, P. K., Tsao, F.-M., Liu, H.-M., Zhang, Y., & de Boer, B. (2001). cal Society of America, 102(3), 17391745.
Language/Culture/ Mind/ Brain: Progress at the margins between Liu, H.-M., Kuhl, P. K., & Tsao, F.-M. (2003). An association be-
disciplines. In A. Domasio et al. (Eds.), Unity of knowledge: The tween mothers speech clarity and infants speech discrimination
convergence of natural and human science (pp. 136174). New skills. Developmental Science, 6(3), F1F10.
York: New York Academy of Sciences. Lotto, A. J., Kluender, K. R., & Holt, L. L. (1998). Depolarizing the
Kuhl, P. K., Williams, K. A., Lacerda, F., Stevens, K. N., & Lind- perceptual magnet effect. Journal of the Acoustical Society of
blom, B. (1992). Linguistic experience alters phonetic perception America, 103, 36483655.
in infants by 6 months of age. Science, 255, 606608. MacDonald, J., & McGurk, H. (1978). Visual inf luences on speech
Lary, S., Briassoulis, G., de Vries, L., Dubowitz, L. M. S., & Dubowitz, perception process. Perception and Psychophysics, 24, 253257.
V. (1985). Hearing threshold in preterm and term infants by audi- MacKain, K., Studdert-Kennedy, M., Spieker, S., & Stern, D. (1983,
tory brainstem response. Journal of Pediatrics, 107, 593599. March 18). Infant intermodal speech perception is a left-hemi-
Lasky, R. E. (1984). A developmental study on the effect of stimulus sphere function. Science, 219, 13471349.
rate on the auditory evoked brain-stem response. Electroen- Mandel, D. R., Jusczyk, P. W., & Kemler Nelson, D. G. (1994). Does
cephalography and Clinical Neurophysiology, 59, 411419. sentential prosody help infants organize and remember speech in-
Lasky, R. E. (1991). The effects of rate and forward masking on formation? Cognition, 53, 155180.
human adult and newborn auditory evoked response thresholds. Mandel, D. R., Jusczyk, P. W., & Pisoni, D. B. (1995). Infants recog-
Developmental Psychobiology, 24, 2164. nition of the sound patterns of their own names. Psychological
Lasky, R. E. (1993). The effect of forward masker duration, Science, 6(5), 315318.
rise/fall time, and integrated pressure on auditory brain stem Marcus, G. F., Vijayan, S., Rao, S. B., & Vishton, P. M. (1999). Rule
evoked responses in human newborns and adults. Ear and Hear- learning by 7-month-old infants. Science, 283(5398), 7780.
ing, 14, 95103. Massaro, D. W., Cohen, M. M., & Smeele, P. M. T. (1995). Cross-lin-
Lasky, R. E. (1997). Rate and adaptation effects on the auditory guistic comparisons in the integration of visual and auditory
evoked brainstem response in human newborns and adults. Hear- speech. Memory and Cognition, 23(1), 113131.
ing Research, 111, 165176. Massaro, D. W., Thompson, L. A., Barron, B., & Laren, E. (1986). De-
Lasky, R. E., & Rupert, A. (1982). Temporal masking of auditory velopmental changes in visual and auditory contributions to speech
evoked brainstem responses in human newborns and adults. Hear- perception. Journal of Experimental Child Psychology, 41, 93113.
ing Research, 6, 315334. Mattingly, I. G., Liberman, A. M., Syrdal, A. K., & Halwes, T.
Lasky, R. E., Syrdal-Lasky, A., & Klein, R. E. (1975). Vot discrimi- (1971). Discrimination in speech and non-speech modes. Cogni-
nation by 4- to 6-month-old infants from Spanish environments. tive Psychology, 2, 131157.
Journal of Experimental Child Psychology, 20, 215225. Mattys, S. L., & Jusczyk, P. W. (2001). Phonotactic cues for segmen-
Lavigne-Rebillard, M., & Bagger-Sjoback, D. (1992). Development of tation of f luent speech by infants. Cognition, 78(2), 91121.
the human stria vascularis. Hearing Research, 64, 3951. Mattys, S. L., Jusczyk, P. W., Luce, P. A., & Morgan, J. L. (1999).
Lavigne-Rebillard, M., & Pujol, R. (1987). Surface aspects of the de- Phonotactic and prosodic effects on word segmentation in infants.
veloping human organ of corti. Acta Otolaryngologica, 436, 4350. Cognitive Psychology, 38(4), 465494.
102 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

Maxon, A. B., & Hochberg, I. (1982). Development of psychoa- Moore, B. C. J. (1996). Introduction to the psychology of hearing (5th
coustic behavior: Sensitivity and discrimination. Ear and Hear- ed.). New York: Academic Press.
ing, 3, 301308. Moore, B. C. J., Glasberg, B. R., Plack, C. J., & Biswas, A. K. (1988).
Maye, J., Werker, J. F., & Gerken, L. (2002). Infant sensitivity to dis- The shape of the ears temporal window. Journal of the Acoustical
tributional information can affect phonetic discrimination. Cog- Society of America, 83, 11021116.
nition, 82(3), B101B111. Moore, J. K. (2002). Maturation of human auditory cortex: Implica-
McAdams, S., & Bertoncini, J. (1997). Organization and discrimina- tions for speech perception. Annals of Otology Rhinology and
tion of repeating sound sequences by newborn infants. Journal of Laryngology, 111, 710.
the Acoustical Society of America, 102(5, Pt. 1), 29452953. Moore, J. K., & Guan, Y. L. (2001). Cytoarchitectural and axonal
McGurk, H., & MacDonald, J. (1976). Hearing lips and seeing maturation in human auditory cortex. Journal of the Association
voices. Nature, 264(5588), 746748. for Research in Otolaryngology, 2, 297311.
McMurray, B., & Aslin, R. N. (2005). Infants are sensitive to Moore, J. K., Guan, Y. L., & Shi, S. R. (1997). Axogenesis in the
within-category variation in speech perception. Cognition, human fetal auditory system, demonstrated by neurofilament
95(2), B15B26. immunohistochemistry. Anatomy and Embryology, Berlin, 195,
1530.
Mehler, J., & Christophe, A. (1995). Maturation and learning of lan-
guage in the first year of life. In M. S. Gazzaniga (Ed.), The cog- Moore, J. K., Perazzo, L. M., & Braun, A. (1995). Time course of ax-
nitive neurosciences (pp. 943954). Cambridge, MA: MIT Press. onal myelination in human brainstem auditory pathway. Hearing
Research, 87, 2131.
Mehler, J., Dommergues, J., Frauenfelder, U., & Segui, J. (1981).
The syllables role in speech segmentation. Cognitive Psychol- Moore, J. K., Ponton, C. W., Eggermont, J. J., Wu, B. J., & Huang,
ogy, 18, 186. J. Q. (1996). Perinatal maturation of the auditory brain stem re-
sponse: Changes in path length and conduction velocity. Ear and
Mehler, J., Jusczyk, P. W., Lambertz, G., Halsted, N., Bertoncini, J.,
Hearing, 17, 411418.
& Amiel-Tison, C. (1988). A precursor of language acquisition in
young infants. Cognition, 29, 143178. Mora, J. A., Exposito, M., Solis, C., & Barajas, J. J. (1990). Filter ef-
fects and low stimulation rate on the middle-latency response in
Miller, J. L. (1987). Rate-dependent processing in speech perception. newborns. Audiology, 29, 329335.
In A. Ellis (Ed.), Progress in the psychology of language (Vol. 3,
Morais, J., Content, A., Cary, L., Mehler, J., & Segui, J. (1989).
pp. 119157). Hillsdale, NJ: Erlbaum.
Syllabic segmentation and literacy. Language and Cognitive
Miller, J. L., & Eimas, P. D. (1979). Organization in infant speech Processes, 4(1), 57 67.
perception. Canadian Journal of Psychology, 33, 353367.
Morgan, J. L., & Demuth, K. (1996). Signal to syntax: Bootstrapping
Miller, J. L., & Eimas, P. D. (1996). Internal structure of voicing from speech to grammar in early acquisition. Hillsdale, NJ: Erlbaum.
categories in early infancy. Perception and Psychophysics,
Morgan, J. L., Meier, R. P., & Newport, E. L. (1987). Structural pack-
58(8), 11571167.
aging in the input to language learning: Contributions of intona-
Mills, D. L., Prat, C., Zangl, R., Stager, C. L., Neville, H. J., & tional and morphological marking of phrases to the acquisition of
Werker, J. F. (2004). Language experience and the organization language. Cognitive Psychology, 19, 498550.
of brain activity to phonetically similar words: ERP evidence
Morgan, J. L., & Newport, E. L. (1981). The role of constituent struc-
from 14- and 20-month-olds. Journal of Cognitive Neuroscience,
ture in the induction of an artificial language. Journal of Verbal
16(8), 113.
Learning and Verbal Behavior, 20, 6785.
Moffit, A. R. (1971). Consonant cue perception by 20- to 24-week-
Morgan, J. L., & Saffran, J. R. (1995). Emerging integration of se-
old infants. Child Development, 42, 505511.
quential and suprasegmental information in preverbal speech seg-
Molfese, D. L., Burger-Judisch, L. M., & Hans, L. L. (1991). Conso- mentation. Child Development, 66, 911936.
nant discrimination by newborn infants: Electrophysiological dif-
Morgan, J. L., Shi, R., & Allopenna, P. (1996). Perceptual bases of
ferences. Developmental Neuropsychology, 7(2), 177195.
rudimentary grammatical categories: Toward a broader conceptu-
Molfese, D. L., & Molfese, V. (1979). Hemisphere and stimulus dif- alization of bootstrapping. In J. L. M. K. Demuth (Ed.), Signal to
ferences as ref lected in the cortical responses of newborn infants. syntax (pp. 263283). Hillsdale, NJ: Erlbaum.
Developmental Psychology, 15, 505511.
Morrongiello, B. A. (1988). Infants localization of sounds in the hor-
Molfese, D. L., & Molfese, V. (1980). Cortical responses of preterm izontal plane: Estimates of minimum audible angle. Developmen-
infants to phonetic and nonphonetic speech stimuli. Developmen- tal Psychology, 24, 813.
tal Psychology, 16, 574581. Morrongiello, B. A., Fenwick, K., & Chance, G. (1990). Sound local-
Molfese, D. L., & Molfese, V. (1985). Electrophysiological indices of ization acuity in very young infants: An observer-based testing
auditory discrimination in newborn infants: The bases for predict- procedure. Developmental Psychology, 26, 7584.
ing later language development? Infant Behavior and Develop- Morrongiello, B. A., Fenwick, K. D., Hillier, L., & Chance, G.
ment, 8, 197211. (1994). Sound localization in newborn human infants. Develop-
Montgomery, C. R., & Clarkson, M. G. (1997). Infants pitch percep- mental Psychobiology, 27, 519538.
tion: Masking by low- and high-frequency noises. Journal of the Morrongiello, B. A., Hewitt, K. L., & Gotowiec, A. (1991). Infants
Acoustical Society of America, 102, 36653672. discrimination of relative distance in the auditory modality: Ap-
Moon, C., Cooper, R. P., & Fifer, W. P. (1993). Two-day-olds pre- proaching versus receding sound sources. Infant Behavior and De-
fer their native language. Infant Behavior and Development, velopment, 14, 187208.
16(4), 495500. Morrongiello, B. A., & Rocca, P. T. (1987a). Infants localization of
Moore, B. C. J. (1973). Frequency difference limens for short-duration sounds in the horizontal plane: Effects of auditory and visual
tones. Journal of the Acoustical Society of America, 54, 610619. cues. Child Development, 58, 918927.
References 103

Morrongiello, B. A., & Rocca, P. T. (1987b). Infants localization of Nittrouer, S., & Miller, M. E. (1997). Predicting developmental shifts
sounds in the median sagittal plane: Effects of signal frequency. in perceptual weighting schemes. Journal of the Acoustical Society
Journal of the Acoustical Society of America, 82, 900905. of America, 101, 33533366.
Morrongiello, B. A., & Rocca, P. T. (1987c). Infants localization of Nittrouer, S., & Studdert-Kennedy, M. (1987). The role of coarticula-
sounds in the median vertical plane: Estimates of minimal audible tory effects in the perception of fricatives by children and adults.
angle. Journal of Experimental Child Psychology, 43, 181193. Journal of Speech and Hearing Research, 30, 319329.
Morrongiello, B. A., & Rocca, P. T. (1990). Infants localization of Norton, S. J., Gorga, M. P., Widen, J. E., Folsom, R. C., Sininger, Y.,
sounds within hemifields: Estimates of minimum audible angle. Cone-Wesson, B., et al. (2000). Identification of neonatal hearing
Child Development, 61, 12581270. impairment: Summary and recommendations. Ear and Hearing,
Morrongiello, B. A., & Trehub, S. E. (1987). Age-related changes in 21, 529535.
auditory temporal perception. Journal of Experimental Child Psy- Novak, G. P., Kurtzberg, D., Kreuzer, J. A., & Vaughan, H. G. (1989).
chology, 44, 413426. Cortical responses to speech sounds and their formants in normal
Morse, P. A. (1972). The discrimination of speech and non-speech infants: Maturational sequence and spatiotemporal analysis. Elec-
stimuli in early infancy. Journal of Experimental Child Psychol- troencephalography and Clinical Neurophysiology, 73(4), 295305.
ogy, 14(3), 477492. Nozza, R. J. (1987). The binaural masking level difference in infants
Morse, P. A., & Snowdon, C. T. (1975). An investigation of categori- and adults: Developmental change in binaural hearing. Infant Be-
cal speech discrimination by rhesus monkeys. Perception and Psy- havior and Development, 10, 105110.
chophysics, 17(1), 916. Nozza, R. J. (1995). Estimating the contribution of non-sensory fac-
Myers, J., Jusczyk, P. W., Kemler Nelson, D. G., Charles-Luce, J., tors to infant-adult differences in behavioral thresholds. Hearing
Woodward, A. L., & Hirsh-Pasek, K. (1996). Infants sensitivity Research, 91, 7278.
to word boundaries in f luent speech. Journal of Child Language, Nozza, R. J., & Henson, A. M. (1999). Unmasked thresholds and min-
23(1), 130. imum masking in infants and adults: Separating sensory from
Naigles, L. (1990). Children use syntax to learn verb meaning. Jour- nonsensory contributions to infant-adult differences in behavioral
nal of Child Language, 17, 357374. thresholds. Ear and Hearing, 20, 483496.
Nakai, Y. (1970). An electron microscopic study of the human Nozza, R. J., & Wilson, W. R. (1984). Masked and unmasked pure-
fetus cochlea. Practica of Otology, Rhinology and Laryngology, tone thresholds of infants and adults: Development of auditory
32, 257267. frequency selectivity and sensitivity. Journal of Speech and Hear-
Nazzi, T., Bertoncini, J., & Mehler, J. (1998). Language discrimina- ing Research, 27, 613622.
tion by newborns: Toward an understanding of the role of rhythm. Nygaard, L. C., & Pisoni, D. B. (1995). Speech perception: New di-
Journal of Experimental Psychology: Human Perception and Per- rections in research and theory. In J. L. Miller & P. D. Eimas
formance, 24(3), 756766. (Eds.), Handbook of perception and cognition: Vol. 11. Speech,
Nazzi, T., Dilley, L., Jusczyk, A. M., Shattuck-Hufnagel, S., & language, and communication (2nd ed., pp. 6396). San Diego,
Jusczyk, P. (in press). English-learning infants segmentation of CA: Academic Press.
verbs from f luent speech. Language and Speech. Oh, E. L., & Lutfi, R. A. (1999). Informational masking by everyday
Nazzi, T., Floccia, C., & Bertoncini, J. (1998). Discrimination of sounds. Journal of the Acoustical Society of America, 106, 35213528.
pitch contours by neonates. Infant Behavior and Development, Oh, E. L., Wightman, F., & Lutfi, R. A. (2001). Childrens detection
21, 779784. of pure-tone signals with random multitone maskers. Journal of
Nazzi, T., Jusczyk, P. W., & Johnson, E. K. (2000). Language dis- the Acoustical Society of America, 109, 28882895.
crimination by English-learning 5-month-olds: Effects of rhythm Okabe, K. S., Tanaka, S., Hamada, H., Miura, T., & Funai, H. (1988).
and familiarity. Journal of Memory and Language, 43(1), 119. Acoustic impedance measured on normal ears of children. Journal
Neff, D. L., & Callaghan, B. P. (1988). Effective properties of multi- of the Acoustical Society of Japan, 9, 287294.
component simultaneous maskers under conditions of uncertainty. Olivier, D. C. (1968). Stochastic grammars and language acquisition
Journal of the Acoustical Society of America, 83, 18331838. mechanisms. Unpublished doctoral dissertation, Harvard Univer-
Neff, D. L., & Green, D. M. (1987). Masking produced by spectral sity, Cambridge, MA.
uncertainty with multicomponent maskers. Perception and Psy- Olsho, L. W. (1984). Infant frequency discrimination. Infant Behavior
chophysics, 41, 409415. and Development, 7, 2735.
Nespor, M., Guasti, M. T., & Christophe, A. (1996). Selecting Olsho, L. W. (1985). Infant auditory perception: Tonal masking. In-
word order: The rhythmic activation principle. In U. Kleinhenz fant Behavior and Development, 7, 2735.
(Ed.), Interfaces in phonology (pp. 126). Berlin, Germany: Olsho, L. W., Koch, E. G., & Carter, E. A. (1988). Nonsensory fac-
Akademie Verlag. tors in infant frequency discrimination. Infant Behavior and De-
Newport, E. L., & Aslin, R. N. (2004). Learning at a distance: I. Sta- velopment, 11, 205222.
tistical learning of non-adjacent dependencies. Cognitive Psychol- Olsho, L. W., Koch, E. G., Carter, E. A., Halpin, C. F., & Spetner,
ogy, 48(2), 127162. N. B. (1988). Pure-tone sensitivity of human infants. Journal of
Newport, E. L., Hauser, M. D., Spaepen, G., & Aslin, R. N. (2004). the Acoustical Society of America, 84, 13161324.
Learning at a distance: II. Statistical learning of non-adjacent Olsho, L. W., Koch, E. G., & Halpin, C. F. (1987). Level and age ef-
dependencies in a non-human primate. Cognitive Psychology, fects in infant frequency discrimination. Journal of the Acoustical
49, 85117. Society of America, 82, 454464.
Nittrouer, S., Crowther, C. S., & Miller, M. E. (1998). The relative Olsho, L. W., Schoon, C., Sakai, R., Turpin, R., & Sperduto, V.
weighting of acoustic properties in the perception of s +stop clusters (1982). Auditory frequency discrimination in infancy. Develop-
by children and adults. Perception and Psychophysics, 60, 5164. mental Psychology, 18, 721726.
104 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

Onishi, K. H., Chambers, K. E., & Fisher, C. (2002). Learning phono- Meeting of the Acoustical Society of America: Special Session in
tactic constraints from brief auditory experience. Cognition, Memory of Peter Jusczyk, Pittsburgh, PA.
83(1), B13B23. Polka, L., & Werker, J. F. (1994). Developmental changes in percep-
Otake, T., Hatano, G., Cutler, A., & Mehler, J. (1993). Mora or sylla- tion of nonnative vowel contrasts. Journal of Experimental Psy-
ble? Speech segmentation in Japanese. Journal of Memory and chology: Human Perception and Performance, 20(2), 421435.
Language, 32(2), 258278. Pollack, I. (1975). Auditory informational masking. Journal of the
Palmer, C., Jungers, M. K., & Jusczyk, P. W. (2001). Episodic mem- Acoustical Society of America, 57(Suppl. 1), 5.
ory for musical prosody. Journal of Memory and Language, Pons, F. (in press). The effects of distributional learning on rats sen-
45(4), 526545. sitivity to phonetic information. Journal of Experimental Psychol-
Pater, J., Stager, C. L., & Werker, J. F. (2004). The lexical acquisition ogy: Animal Behavior Processes.
of phonological contrasts. Language, 80(3), 361379. Ponton, C. W., Eggermont, J. J., Coupland, S. G., & Winkelaar, R.
Patterson, M. L., & Werker, J. F. (1999). Matching phonetic informa- (1992). Frequency-specific maturation of the eighth-nerve and
tion in lips and voice is robust in 41 2 -month-old infants. Infant Be- brain-stem auditory pathway: Evidence from derived Auditory
havior and Development, 22(2), 237247. Brain-Stem Responses (ABRs). Journal of the Acoustical Society
Patterson, M. L., & Werker, J. F. (2002). Infants ability to match dy- of America, 91, 15761587.
namic phonetic and gender information in the face and voice. Ponton, C. W., Eggermont, J. J., Don, M., Waring, M. D., Kwong, B.,
Journal of Experimental Child Psychology, 81(1), 93115. Cunningham, J., et al. (2000). Maturation of the mismatch nega-
Patterson, M. L., & Werker, J. F. (2003). Two-month-olds match tivity: Effects of profound deafness and cochlear implant use. Au-
vowel information in the face and voice. Developmental Science, diology and Neuro Otology, 5, 167185.
6(2), 191196. Ponton, C. W., Moore, J. K., & Eggermont, J. J. (1996). Auditory
Pegg, J. E., & Werker, J. F. (1997). Adult and infant perception of two brain stem response generation by parallel pathways: Differential
English phones. Journal of the Acoustical Society of America, maturation of axonal conduction time and synaptic transmission.
102(6), 37423753. Ear and Hearing, 17, 402410.
Pea, M., Maki, A., Kovacic, D., Dehaene-Lambertz, G., Koizumi, Price, C., Wise, R., Warburton, E., Moore, C., Howard, D., Patterson,
H., Bouquet, F., et al. (2003). Sounds and silence: An optical to- K., et al. (1996). Hearing and saying: The functional neuro-
pography study of language recognition at birth. Proceedings of anatomy of auditory word processing. Brain, 119(3), 919931.
the National Academy of Sciences, 100(20), 1170211705. Pujol, R., & Lavigne-Rebillard, M. (1992). Development of neurosen-
Phillips, C., Pellathy, T., & Marantz, A. (1999, October). Magnetic sory structures in the human cochlea. Acta Otolaryngologica,
mismatch field elicited by phonological feature contrast. Paper 112, 259264.
presented at the Cognitive Neuroscience Society Meeting, Wash- Ramus, F. (2002). Language discrimination by newborns: Teasing
ington, DC. apart phonotactic, rhythmic, and intonational cues. Annual Review
Pike, K. L. (1945). Step-by-step procedure for marking limited into- of Language Acquisition, 2, 85115.
nation with its related features of pause, stress and rhythm. In Ramus, F., Hauser, M. D., Miller, C., Morris, D., & Mehler, J. (2000).
C. C. Fries (Ed.), Teaching and learning English as a foreign lan- Language discrimination by human newborns and by cotton-top
guage (pp. 6274). Ann Arbor, MI: Publication of the English tamarin monkeys. Science, 288(5464), 349351.
Language Institute. Ramus, F., & Mehler, J. (1999). Language identification with
Pinker, S. (1984). Language learnability and language development. suprasegmental cues: A study based on speech resynthesis. Jour-
Cambridge, MA: Harvard University Press. nal of the Acoustical Society of America, 105(1), 512521.
Pinker, S. (1989) Learnability and cognition: The acquisition of argu- Ramus, F., Nespor, M., & Mehler, J. (1999). Correlates of linguistic
ment structure. Cambridge, MA: MIT Press. rhythm in the speech signal. Cognition, 73(3), 265292.
Pinker, S. (1994). The language instinct. New York: HarperCollins. Ratner, N. B. (1984). Patterns of vowel modification in mother-child
Pisoni, D. B. (1977). Identification and discrimination of the rela- speech. Journal of Child Language, 11, 557578.
tive onset of two component tones: Implications for voicing per- Ratner, N. B., & Luberoff, A. (1984). Cues to post-vocalic voicing in
ception in stops. Journal of the Acoustical Society of America, mother-child speech. Journal of Phonetics, 12, 285289.
61, 13521361. Remez, R. E., Rubin, P. E., Pisoni, D. B., & Carrell, T. D. (1981).
Plessinger, M. A., & Woods, J. R. (1987). Fetal auditory brain stem Speech perception without traditional speech cues. Science,
response: Effect of increasing stimulus rate during functional au- 212(4497), 947949.
ditory development. American Journal of Obstetrics and Gynecol- Repp, B. H. (1984). Against a role of chirp identification in duplex
ogy, 157, 1382. perception. Perception and Psychophysics, 35(1), 8993.
Polka, L., & Bohn, O. S. (1996). A cross-language comparison of vowel Rice, K., & Avery, P. (1995). Variability in a deterministic model of
perception in English-learning and German-learning infants. Jour- language acquisition: A theory of segmental acquisition. In J.
nal of the Acoustical Society of America, 100(1), 577592. Archibald (Ed.), Phonological acquisition and phonological theory
Polka, L., Colantonio, C., & Sundara, M. (2001). A cross-language (pp. 2342). Hillsdale, NJ: Erlbaum.
comparison of /d/-/th/ perception: Evidence for a new develop- Rivera-Gaxiola, M., Silva-Pereyra, J., & Kuhl, P. K. (2005). Brain po-
mental pattern. Journal of the Acoustical Society of America, tentials to native and non-native speech contrasts in 7- and 11-
109(Suppl. 5, Pt. 1), 21902201. month-old American infants. Developmental Science, 8, 162172.
Polka, L., Sundara, M., & Blue, S. (2002, June 37). The role of lan- Rosen, S., van der Lely, H., Adlard, A., & Manganari, E. (2000).
guage experience in word segmentation: A comparison of English, Backward masking in children with and without language disor-
French, and bilingual infants. Paper presented at the the 143rd ders [Abstract]. British Journal of Audiology, 34, 124.
References 105

Rosenblum, L. D., Schmuckler, M. A., & Johnson, J. A. (1997). Experimental Psychology: Human Perception and Performance,
The McGurk effect in infants. Perception and Psychophysics, 16, 642652.
59(3), 347357. Schneider, B. A., Trehub, S. E., Morrongiello, B. A., & Thorpe, L. A.
Rumelhart, D. E., & McClelland, J. L. (1986). Parallel distributed (1989). Developmental changes in masked thresholds. Journal of
processing. Cambridge, MA: MIT Press. the Acoustical Society of America, 86, 17331742.
Ruth, R. A., Horner, J. S., McCoy, G. S., & Chandler, C. R. (1983). Sebastin-Galls, N., & Bosch, L. (2002). Building phonotactic knowl-
Comparison of auditory brainstem response and behavioral au- edge in bilinguals: Role of early exposure. Journal of Experimental
diometry in infants. Audiology, Scandinavian, 17, 9498. Psychology: Human Perception and Performance, 28(4), 974989.
Saffran, J. R. (2001). The use of predictive dependencies in language Sebastin-Galls, N., Dupoux, E., Segui, J., & Mehler, J. (1992). Con-
learning. Journal of Memory and Language, 44(4), 493515. trasting syllabic effects in Catalan and Spanish. Journal of Mem-
ory and Language, 31, 1832.
Saffran, J. R. (2002). Constraints on statistical language learning.
Journal of Memory and Language, 47(1), 172196. Seidenberg, M., & Elman, J. L. (1999). Do infants learn grammar
with algebra or statistics? Science, 284, 433.
Saffran, J. R. (2003a). Absolute pitch in infancy and adulthood: The
Seidenberg, M. S., MacDonald, M. C., & Saffran, J. R. (2003). Are
role of tonal structure. Developmental Science, 6(1), 3543.
there limits to statistical learning? Science, 300, 5354.
Saffran, J. R. (2003b). Statistical language learning: Mechanisms
Shady, M. E., & Gerken, L. A. (1999). Grammatical and caregiver
and constraints. Current Directions in Psychological Science,
cues in early sentence comprehension. Journal of Child Lan-
12(4), 110114.
guage, 26, 113.
Saffran, J. R., Aslin, R. N., & Newport, E. L. (1996). Statistical Shafer, V. L., Shucard, D. W., Shucard, J. L., & Gerken, L. (1998). An
learning by 8-month-old infants. Science, 274(5294), 19261928. electrophysiological study of infants sensitivity to the sound pat-
Saffran, J. R., & Graf Estes, K. M. (2004, May). What are statistics terns of English speech. Journal of Speech, Language, and Hearing
for? Linking statistical learning to language acquisition in the Research, 41(4), 874886.
wild. Paper presented at the International Conference on Infant Shahidullah, S., & Hepper, P. G. (1994). Frequency discrimination by
Studies, Chicago, IL. the fetus. Early Human Development, 36(1), 1326.
Saffran, J. R., & Griepentrog, G. (2001). Absolute pitch in infant au- Shi, R., Morgan, J. L., & Allopenna, P. (1998). Phonological and acoustic
ditory learning: Evidence for developmental reorganization. De- bases for earliest grammatical category assignment: A cross-linguis-
velopmental Psychology, 37(1), 7485. tic perspective. Journal of Child Language, 25(1), 169201.
Saffran, J. R., Hauser, M., Seibel, R., Kapf hamer, J., Tsao, F., & Shi, R., & Werker, J. F. (2001). Six-month-old infants preference for
Cushman, F. (2005). Cross-species dif ferences in the capacity to lexical words. Psychological Science, 12(1), 7176.
acquire language: Grammatical pattern learning by human infants Shi, R., & Werker, J. F. (2003). The basis of preference for lexical words
and monkeys. Manuscript submitted for publication. in 6-month-old infants. Developmental Science, 6(5), 484488.
Saffran, J. R., Johnson, E. K., Aslin, R. N., & Newport, E. L. (1999). Shi, R., Werker, J. F., & Cutler, A. (2003). Function words in early
Statistical learning of tone sequences by human infants and speech perception. In Proceedings of the 15th International Con-
adults. Cognition, 70(1), 2752. ference of Phonetic Sciences (pp. 30093012). Adelaide, Aus-
Saffran, J. R., Loman, M. M., & Robertson, R. R. W. (2000). Infant tralia: Causal Productions.
memory for musical experiences. Cognition, 77, 1523. Shi, R., Werker, J. F., Cutler, A., & Cruickshank, M. (2004, April).
Saffran, J. R., Reeck, K., Niehbur, A., & Wilson, D. P. (in press). Facilitation ef fects of function words for word segmentation in in-
Changing the tune: Absolute and relative pitch processing by fants. Poster presented at the International Conference of Infant
adults and infants. Developmental Science, 7(7), 5371. Studies, Chicago.
Saffran, J. R., & Thiessen, E. D. (2003). Pattern induction by infant Shi, R., Werker, J. F., & Morgan, J. L. (1999). Newborn infants sensi-
language learners. Developmental Psychology, 39(3), 484494. tivity to perceptual cues to lexical and grammatical words. Cogni-
tion, 72(2), B11B21.
Saffran, J. R., & Wilson, D. P. (2003). From syllables to syntax:
Multilevel statistical learning by 12-month-old infants. Infancy, Shvachkin, N. K. (1948). The development of phonemic speech per-
4(2), 273284. ception in early childhood. In C. A. Ferguson & D. I. Slobin
(Eds.), Studies of child language development (pp. 91127). New
Sanes, D. H., & Walsh, E. J. (1998). The development of central audi- York: Holt, Rinehart and Winston.
tory function. In E. W. Rubel, R. R. Fay, & A. N. Popper (Eds.),
Singh, L., Bortfeld, H., & Morgan, J. (2002). Effects of variability on
Development of the auditory system (pp. 271314). New York:
infant word recognition. In A. H. J. Do, L. Domnguez, & A. Jo-
Springer Verlag.
hansen (Eds.), Proceedings of the 26th Annual Boston University
Santelmann, L. M., & Jusczyk, P. W. (1998). Sensitivity to discontin- Conference on Language Development (pp. 608619). Somerville,
uous dependencies in language learners: Evidence for limitations MA: Cascadilla Press.
in processing space. Cognition, 69(2), 105134.
Singh, L., Morgan, J., & White, K. (2004). Preference and processing:
Schafer, G., & Plunkett, K. (1998). Rapid word learning by 15- The role of speech affect in early spoken word recognition. Jour-
month-olds under tightly controlled conditions. Child Develop- nal of Memory and Language, 51, 173189.
ment, 69(2), 309320. Sininger, Y., & Abdala, C. (1996). Auditory brainstem response
Schneider, B. A., Bull, D., & Trehub, S. E. (1988). Binaural un- thresholds of newborns based on ear canal levels. Ear and Hear-
masking in infants. Journal of the Acoustical Society of America, ing, 17, 395401.
83, 11241132. Sininger, Y. S., Abdala, C., & Cone-Wesson, B. (1997). Auditory
Schneider, B. A., Morrongiello, B. A., & Trehub, S. E. (1990). The threshold sensitivity of the human neonate as measured by the au-
size of the critical band in infants, children, and adults. Journal of ditory brainstem response. Hearing Research, 104, 122.
106 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

Sinnott, J. M. (1989). Detection and discrimination of synthetic En- B. Blachman (Ed.), Foundations of reading acquisition and
glish vowels by Old World monkeys (Cercopithecus, Macaca) and dyslexia (pp. 4966). Mahwah, NJ: Erlbaum.
humans. Journal of the Acoustical Society of America, 86, 557565. Tallal, P., & Piercy, M. (1973). Defects of non-verbal auditory percep-
Sinnott, J. M., & Aslin, R. N. (1985). Frequency and intensity dis- tion in children with developmental aphasia. Nature, 241, 468469.
crimination in human infants and adults. Journal of the Acoustical Tallal, P., & Piercy, M. (1974). Developmental aphasia: Rate of audi-
Society of America, 78, 19861992. tory processing and selective impairment of consonant percep-
Sinnott, J. M., Pisoni, D. B., & Aslin, R. M. (1983). A comparison of tion. Neuropsychologia, 12, 8393.
pure tone auditory thresholds in human infants and adults. Infant
Tanenhaus, M. K., Spivey-Knowlton, M. J., Eberhard, K. M., & Se-
Behavior and Development, 6, 317.
divy, J. E. (1995). Integration of visual and linguistic information
Slobin, D. I. (1973). Cognitive prerequisites for the development of in spoken language comprehension. Science, 268, 16321634.
grammar. In C. A. Ferguson & D. I. Slobin (Eds.), Studies of child
Tarquinio, N., Zelazo, P. R., & Weiss, M. J. (1990). Recovery of
language development (pp. 175208). New York: Holt, Rinehart
neonatal head turning to decreased sound pressure level. Develop-
and Winston.
mental Psychology, 26, 752758.
Smith, S. L., Gerhardt, K. J., Griffiths, S. K., & Huang, X. (2003).
Intelligibility of sentences recorded from the uterus of a preg- Tees, R. C. (1976). Perceptual development in mammals. In G. Got-
nant ewe and from the fetal inner ear. Audiology and Neuro Otol- tlieb (Ed.),Studies on development of behavior and the nervous
ogy, 8, 347353. system (pp. 281326). New York: Academic Press.
Soderstrom, M., Seidl, A., Kemler Nelson, D. G., & Jusczyk, Teller, D. Y. (1979). The forced-choice preferential looking proce-
P. W. (2003). The prosodic bootstrapping of phrases: Evidence dure: A psychophysical technique for use with human infants. In-
from prelinguistic infants. Journal of Memory and Language, fant Behavior and Development, 2, 135153.
49(2), 249267. Tharpe, A. M., & Ashmead, D. H. (2001). A longitudinal investiga-
Spetner, N. B., & Olsho, L. W. (1990). Auditory frequency resolution tion of infant auditory sensitivity. American Journal of Audiology,
in human infancy. Child Development, 61, 632652. 10, 104112.
Stager, C. L., & Werker, J. F. (1997). Infants listen for more phonetic Thiessen, E. D. (2004). The role of distributional information in in-
detail in speech perception than in word-learning tasks. Nature, fants use of phonemic contrast. Unpublished doctoral disserta-
388(6640), 381382. tion, University of Wisconsin, Madison.
Stellmack, M. A., Willihnganz, M. S., Wightman, F. L., & Lutfi, Thiessen, E. D., & Saffran, J. R. (2003). When cues collide: Use of
R. A. (1997). Spectral weights in level discrimination by pre- stress and statistical cues to word boundaries by 7- to 9-month-
school children: Analytic listening conditions. Journal of the old infants. Developmental Psychology, 39(4), 706716.
Acoustical Society of America, 101, 28112821. Thiessen, E. D., & Saffran, J. R. (2004). Infants acquisition of
Storkel, H. L. (2001). Learning new words: Phonotactic probability in stress-based word segmentation strategies. In A. Brugos, L. Mic-
language development. Journal of Speech, Language, and Hearing ciulla, C. Smith (Eds.), Proceedings of the 28th annual Boston
Research, 44(6), 13211337. University Conference on Language Development, 2, 608619.
Strange, W., & Jenkins, J. J. (1978). Role of linguistic experience in Thompson, N. C., Cranford, J. L., & Hoyer, E. (1999). Brief-tone fre-
the perception of speech. In R. D. Walk & H. L. Pick (Eds.), Per- quency discrimination by children. Journal of Speech Language
ception and experience (pp. 125169). New York: Plenum Press. and Hearing Research, 42, 10611068.
Streeter, L. A. (1976). Language perception of 2-month-old infants Thorpe, L. A., & Schneider, B. A. (1987, April). Temporal integration
shows effects of both innate mechanisms and experience. Nature, in infant audition. Paper presented at the Society for Research in
259, 3941. Child Development, Baltimore, MD.
Studdert-Kennedy, M., & Shankweiler, D. P. (1970). Hemispheric Tincoff, R., & Jusczyk, P. W. (1999). Some beginnings of word com-
specialization for speech perception. Journal of the Acoustical So- prehension in 6-month-olds. Psychological Science, 10, 172175.
ciety of America, 48(2, Pt. 2), 579595.
Toro, J. M., & Trobaln, J. B. (2004). Statistical computations over a
Swingley, D. (2005). Statistical clustering and the contents of the in-
speech stream in a rodent. Manuscript submitted for publication.
fant vocabulary. Cognitive Psychology, 50, 86132.
Trainor, L. J., Wu, L., & Tsang, C. D. (2004). Long-term memory for
Swingley, D., & Aslin, R. N. (2000). Spoken word recognition and lexical
music: Infants remember tempo and timbre. Developmental Sci-
representation in very young children. Cognition, 76(2), 147166.
ence, 7, 289296.
Swingley, D., & Aslin, R. N. (2002). Lexical neighborhoods and the
word-form representations of 14-month-olds. Psychological Sci- Trehub, S. E. (1973). Infants sensitivity to vowel and tonal contrasts.
ence, 13(5), 480484. Developmental Psychology, 9(1), 9196.
Swingley, D., Pinto, J. P., & Fernald, A. (1999). Continuous process- Trehub, S. E. (1976). The discrimination of foreign speech contrasts
ing in word recognition at 24 months. Cognition, 71(2), 73108. by infants and adults. Child Development, 47, 466472.
Swoboda, P., Kass, J., Morse, P. A., & Leavitt, L. A. (1978). Memory Trehub, S. E. (2003). Musical predispositions in infancy: An update.
factors in infant vowel discrimination of normal and at-risk in- In I. Peretz & R. J. Zatorre (Eds.), The cognitive neuroscience of
fants. Child Development, 49, 332339. music (pp. 320). Oxford, England: Oxford University Press.
Swoboda, P. J., Morse, P. A., & Leavitt, L. A. (1976). Continuous Trehub, S. E., Endman, M. W., & Thorpe, L. A. (1990). Infants per-
vowel discrimination in normal and at risk infants. Child Develop- ception of timbre: Classification of complex tones by spectral
ment, 47(2), 459465. structure. Journal of Experimental Child Psychology, 49, 300313.
Tallal, P., Miller, S. L., Jenkins, W. M., & Merzenich, M. M. (1997). Trehub, S. E., Schneider, B. A., & Edman, M. (1980). Developmental
The role of temporal processing in the developmental language- changes in infants sensitivity to octave-band noises. Journal of
based learning disorders: Research and clinical implications. In Experimental Child Psychology, 29, 282293.
References 107

Trehub, S. E., Schneider, B. A., & Henderson, J. (1995). Gap detec- Vroomen, J., van Zon, M., & de Gelder, B. (1996). Cues to speech
tion in infants, children, and adults. Journal of the Acoustical So- segmentation: Evidence from juncture misperceptions and word
ciety of America, 98, 25322541. spotting. Memory and Cognition, 24(6), 744755.
Trehub, S. E., Schneider, B. A., Morrengiello, B. A., & Thorpe, L. A. Walker-Andrews, A. S., Bahrick, L. E., Raglioni, S. S., & Diaz, I.
(1988). Auditory sensitivity in school-age children. Journal of Ex- (1991). Infants bimodal perception of gender. Ecological Psy-
perimental Child Psychology, 46, 273285. chology, 3(2), 5575.
Trehub, S. E., Schneider, B. A., Thorpe, L. A., & Judge, P. (1991). Walton, G. E., & Bower, T. G. R. (1993). Amodal representation of
Observational measures of auditory sensitivity in early infancy. speech in infants. Infant Behavior and Development, 16, 233243.
Developmental Psychology, 27, 4049. Waters, R. A., & Wilson, W. A., Jr. (1976). Speech perception by rhe-
Tsao, F.-M., Liu, H.-M., & Kuhl, P. K. (2004). Speech Perception in sus monkeys: The voicing distinction in synthesized labial and
Infancy Predicts Language Development in the Second Year of velar stop consonants. Perception and Psychophysics, 19, 285289.
Life: A longitudinal study. Child Development, 75, 10671084. Weir, C. (1976). Auditory frequency sensitivity in the neonate: A sig-
Tsushima, T. T. O., Sasaki, M., Shiraki, S., Nishi, K., Kohno, M., nal detection analysis. Journal of Experimental Child Psychology,
Menyuk, P., et al. (1994). Discrimination of English /r-l / and /w- 21, 219225.
y/ by Japanese infants at 612 months: Language-specific devel- Weir, C. (1979). Auditory frequency sensitivity of human newborns:
opmental changes in speech perception abilities. Proceedings of Some data with improved acoustic and behavioral controls. Per-
International Conference of Spoken Language Processing, Yoko- ception and Psychophysics, 26, 287294.
hama, 94, 16951698. Werker, J. F., Burns, T., & Moon, E. (in press). Bilingual exposed new-
Turk, A. E., Jusczyk, P. W., & Gerken, L. (1995). Do English-learn- borns prefer to listen to both of their languages. Manuscript in
ing infants use syllable weight to determine stress? Language and preparation.
Speech, 38(2), 143158. Werker, J. F., Cohen, L. B., Lloyd, V. L., Casasola, M., & Stager,
Vargha-Khadem, F., & Corballis, M. C. (1979). Cerebral asymmetry C. L. (1998). Acquisition of word-object associations by 14-
in infants. Brain and Language, 8(1), 19. month-old infants. Developmental Psychology, 34(6), 12891309.
Viemeister, N. F. (1979). Temporal modulation transfer functions Werker, J. F., & Curtin, S. (2005). PRIMIR: A developmental frame-
based upon modulation thresholds. Journal of the Acoustical Soci- work of infant speech processing. Language Learning and Devel-
ety of America, 66, 13801564. opment, 1(2), 197234.
Werker, J. F., Fennell, C. T., Corcoran, K. M., & Stager, C. L. (2002).
Viemeister, N. F. (1996). Auditory temporal integration: What is
Infants ability to learn phonetically similar words: Effects of age
being accumulated? Current Directions in Psychological Sci-
and vocabulary size. Infancy, 3(1), 130.
ence, 5, 2832.
Werker, J. F., Frost, P. E., & McGurk, H. (1992). La langue et les lev-
Viemeister, N. F., & Plack, C. J. (1993). Time analysis. In W. A. Yost,
res: Cross-language inf luences on bimodal speech perception.
A. N. Popper, & R. R. Fay (Eds.), Human psychophysics (Vol. 3,
Canadian Journal of Psychology, 46(4), 551568.
pp. 116154). New York: Springer-Verlag.
Werker, J. F., Gilbert, J. H., Humphrey, K., & Tees, R. C. (1981). De-
Viemeister, N. F., & Schlauch, R. S. (1992). Issues in infant psychoa- velopmental aspects of cross-language speech perception. Child
coustics. In L. A. Werner & E. W. Rubel (Eds.), Developmental Development, 52(1), 349355.
psychoacoustics (pp. 191210). Washington, DC: American Psy-
Werker, J. F., & Lalonde, C. E. (1988). Cross-language speech per-
chological Association.
ception: Initial capabilities and developmental change. Develop-
Viemeister, N. F., & Wakefield, G. H. (1991). Temporal integration mental Psychology, 24(5), 672683.
and multiple looks. Journal of the Acoustical Society of America,
Werker, J. F., & Logan, J. S. (1985). Cross-language evidence for
90, 858865.
three factors in speech perception. Perception and Psychophysics,
Vihman, M., Nakai, S., dePaolis, R., & Hall, P. (2004). The role of 37(1), 3544.
accentual pattern in early lexical representation. Journal of Mem- Werker, J. F., & McLeod, P. J. (1989). Infant preference for both male
ory and Language, 50(3), 336353. and female infant-directed talk: A developmental study of atten-
Vitevitch, M. S., & Luce, P. A. (1998). Probabilistic phonotactics and tional and affective responsiveness. Canadian Journal of Psychol-
neighborhood activation in spoken word recognition. Journal of ogy, 43(2), 230246.
Memory and Language, 40, 374408. Werker, J. F., & Tees, R. C. (1983). Developmental changes across
Vitevitch, M. S., Luce, P. A., Charles-Luce, J., & Kemmerer, D. (1997). childhood in the perception of non-native speech sounds. Cana-
Phonotactics and syllable stress: Implications for the processing of dian Journal of Psychology, 37(2), 278286.
spoken nonsense words. Language and Speech, 40, 4762. Werker, J. F., & Tees, R. C. (1984). Cross-language speech percep-
Vouloumanos, A., Kiehl, K. A., Werker, J. F., & Liddle, P. F. (2001). tion: Evidence for perceptual reorganization during the first year
Detection of sounds in the auditory stream: Event-related fMRI of life. Infant Behavior and Development, 7(1), 4963.
evidence for differential activation to speech and nonspeech. Werker, J. F., Weikum, W. M., & Yoshida, K. A. (in press). Bilingual
Journal of Cognitive Neuroscience, 13(7), 9941005. speech processing. In W. Li (Series Ed.) & E. Hoff & P. McCardle
Vouloumanos, A., & Werker, J. F. (2002, April). Infants preference (Vol. Eds.), Multilingual Matters.
for speech: When does it emerge? Poster presented at the 13th bi- Werner, L. A. (1992). Interpreting developmental psychoacoustics. In
ennial International Conference of Infant Studies, Toronto, On- L. A. Werner & E. W. Rubel (Eds.), Developmental psychoa-
tario, Canada. coustics (pp. 4788). Washington, DC: American Psychological
Vouloumanos, A., & Werker, J. F. (2004). Tuned to the signal: The Association.
privileged status of speech for young infants. Developmental Sci- Werner, L. A. (1995). Observer-based approaches to human infant
ence, 7(3), 270276. psychoacoustics. In G. M. Klump, R. J. Dooling, R. R. Fay, &
108 The Infants Auditory World: Hearing, Speech, and the Beginnings of Language

W. C. Stebbins (Eds.), Methods in comparative psychoacoustics Wightman, F., Allen, P., Dolan, T., Kistler, D., & Jamieson, D.
(pp. 135146). Boston: Birkhauser Verlag. (1989). Temporal resolution in children. Child Development, 60,
Werner, L. A. (1999). Forward masking among infant and adult listen- 611624.
ers. Journal of the Acoustical Society of America, 105, 24452453. Wightman, F., Callahan, M. R., Lutfi, R. A., Kistler, D. J., & Oh, E.
Werner, L. A. (2003, March). Development of backward masking in (2003). Childrens detection of pure-tone signals: Informational
infants. Paper presented at the American Auditory Society, masking with contralateral maskers. Journal of the Acoustical So-
Scottsdale, AZ. ciety of America, 113, 32973305.
Werner, L. A., & Bargones, J. Y. (1991). Sources of auditory mask- Wilmington, D., Gray, L., & Jahrsdorfer, R. (1994). Binaural process-
ing in infants: Distraction effects. Perception and Psychophysics, ing after corrected congenital unilateral conductive hearing loss.
50, 405412. Hearing Research, 74, 99114.
Werner, L. A., & Boike, K. (2001). Infants sensitivity to broadband Wong, D., Miyamoto, R. T., Pisoni, D. B., Sehgal, M., & Hutchins, G.
noise. Journal of the Acoustical Society of America, 109, (1999). PET imaging of cochlear-implant and normal-hearing sub-
21012111. jects listening to speech and nonspeech stimuli. Hearing Research,
Werner, L. A., Folsom, R. C., & Mancl, L. R. (1993). The relation- 132, 3442.
ship between auditory brainstem response and behavioral Woodward, A. L., Markman, E. M., & Fitzsimmons, C. M. (1994).
thresholds in normal hearing infants and adults. Hearing Re- Rapid word learning in 13- and 18-month-olds. Developmental
search, 68, 131141. Psychology, 30, 553566.
Werner, L. A., Folsom, R. C., & Mancl, L. R. (1994). The relationship Wright, B. A., Lombardino, L. J., King, W. M., Puranik, C. S.,
between auditory brainstem response latencies and behavioral Leonard, C. M., & Merzenich, M. M. (1997). Deficits in auditory
thresholds in normal hearing infants and adults. Hearing Re- temporal and spectral resolution in language-impaired children.
search, 77, 8898. Nature, 387, 129130.
Werner, L. A., & Gillenwater, J. M. (1990). Pure-tone sensitivity of 2- to Yoshinaga-Itano, C., Sedey, A. L., Coulter, D. K., & Mehl, A. L.
5-week-old infants. Infant Behavior and Development, 13, 355375. (1998). Language of early- and later-identified children with
Werner, L. A., & Mancl, L. R. (1993). Pure-tone thresholds of 1- hearing loss. Pediatrics, 102, 11611171.
month-old human infants. Journal of the Acoustical Society of Zangl, R., & Fernald, A. (2003, October). Sensitivity to function mor-
America, 93, 2367. phemes in on-line sentence processing: Developmental changes
Werner, L. A., Marean, G. C., Halpin, C. F., Spetner, N. B., & Gillen- from 18 to 36 months. Paper presented at the biennial meeting of
water, J. M. (1992). Infant auditory temporal acuity: Gap detec- the Boston University Conference on Child Language Develop-
tion. Child Development, 63, 260272. ment, Boston, MA.
Whalen, D. H., & Liberman, A. M. (1987). Speech perception takes Zatorre, R. J., Evans, A. C., Meyer, E., & Gjedde, A. (1992). Lateral-
precedence over nonspeech perception. Science, 237, 169171. ization of phonetic and pitch discrimination in speech processing.
Wightman, F., & Allen, P. (1992). Individual differences in auditory Science, 256(5058), 846849.
capability among preschool children. In L. A. Werner & E. W. Zatorre, R. J., Meyer, E., Gjedde, A., & Evans, A. (1996). PET stud-
Rubel (Eds.), Developmental psychoacoustics (pp. 113133). ies of phonetic processing of speech: Review, replication, and re-
Washington, DC: American Psychological Association. analysis. Cerebral Cortex, 6(1), 2130.
CHAPTER 3

Infant Visual Perception


PHILIP J. KELLMAN and MARTHA E. ARTERBERRY

THEORIES OF PERCEPTUAL DEVELOPMENT 110 Edge Detection and Edge Classification 135
The Constructivist View 110 Detection and Classification of Contour Junctions 136
The Ecological View 112 Boundary Assignment 137
The Contemporary Situation in Perceptual Theory 113 Perception of Object Unity 137
BASIC VISUAL SENSITIVITIES IN INFANCY 114 Perception of Three-Dimensional Form 142
Visual Acuity 114 Perception of Size 145
Contrast Sensitivity 116 FACE PERCEPTION 146
Orientation Sensitivity 119 Preference for Facelike Stimuli 146
Pattern Discrimination 119 Perceiving Information about People through Faces 148
Color Vision 120 Mechanisms of Face Perception 149
Motion Perception 124 CONCLUSION 150
SPACE PERCEPTION 126 Levels of Analysis 150
Kinematic Information 127 Hardwiring versus Construction in
Stereoscopic Depth Perception 129 Visual Development 151
Pictorial Depth Perception 131 Future Directions 152
OBJECT PERCEPTION 134 REFERENCES 152
Multiple Tasks in Object Perception 134

How visual perception develops has long been a central less sensations. Coherent, meaningful, visual reality
question in understanding psychological development emerges only through a protracted learning process in
generally. During its emergence as a separate discipline which visual sensations become associated with each
in the late 1800s, psychology was focused primarily on other and with touch and action (Berkeley, 1709/1963).
how human knowledge originates (e.g., Titchener, 1910; Through most of the twentieth century, even as psy-
Wundt, 1862), an emphasis inherited from concerns in chology increasingly emphasized findings of empirical
philosophy. Much of the focus was on the relation be- research, this primarily philosophical view cast a long
tween sensation and perception, especially in vision. shadow. Its influence was so great as to be essentially a
The prevailing view, inherited from generations of em- consensus view of development. William James (1890)
piricist philosophers (e.g., Berkeley, 1709/1963; Hobbes, echoed its assumptions in his memorable pronouncement
1651/1974; Hume, 1758/1999; Locke, 1690/1971), was that the world of the newborn is a blooming, buzzing,
that at birth, a human being experiences only meaning- confusion. Modern developmental psychology, shaped
greatly by Piaget, incorporated the same ideas. Although
Preparation of this chapter was supported in part by research Piaget combined contributions of both maturation and
grants R01 EY13518-01 from the National Eye Institute and learning in his theories, his view of the starting points of
REC 0231826 from the National Science Foundation to PJK. perception was standard empiricist (e.g., Piaget, 1952,
We thank Heidi Vanyo for helpful assistance. 1954). He did place greater emphasis on action, rather

109
110 Infant Visual Perception

than mere sensory associations, as the means by which ment. These serve as valuable reference points in under-
meaningful reality emerges from initially meaningless standing how recent research has changed our concep-
sensations. tions of how perception begins.
This basic story about early perception and knowl-
edge persisted, in part, because researchers lacked
The Constructivist View
methods for investigating these topics scientifically. The
arguments of Berkeley and others were primarily logical The term constructivism here refers to the view that per-
ones. Claims about the origins of knowledge in the asso- ceptual reality must be constructed through extended
ciation of sensations initially came from theory and learning. Choosing one term to label this idea is effi-
thought experiments. Later, a few experiments with cient, but also unfortunate, as this set of ideas has many
adults were used to make inferences about aspects of names. In philosophy, this kind of account is most often
perception that might be based on learning (e.g., Wal- called empiricism, emphasizing the role of input from
lach, 1976) or not so based (e.g., Gottschaldt, 1926). experience in forming perception. If, as is usually the
Finding a more direct window into perception and case, associations among sensations are held to domi-
knowledge of a young infant seemed unlikely. As Riesen nate perceptual development, the position may also be
(1947, p. 107) put it: The study of innate visual organi- aptly labeled associationism. In the earliest days of psy-
zation in man is not open to direct observation in early chology as an independent discipline, the merging of
infancy, since a young baby is too helpless to respond current and remembered sensations to achieve objects in
differentially to visual excitation. the world was called structuralism (Titchener, 1910).
In the time since Riesens (1947) observation, the sci- Helmholtz (1885/1925) is often credited with applying
entific landscape in this area has changed entirely. Al- the label constructivism to the idea that sensations are
though the development of visual perception is among the combined with previously learned information using un-
most long-standing and fundamental concerns in the conscious inference to achieve perceptual reality. This
field, it is also an area that is conspicuous in terms of re- pedigree, along with Piagets later emphasis on inputs
cent and rapid progress. Beginning in the late 1950s, the from the learners actions in constructing reality, make
door to progress has been the development of methods for constructivism perhaps the best term to characterize
studying sensation, perception, and knowledge in human modern versions of this view. Unfortunately, the term
infants. The results of scientific efforts, continuing to the has been used elsewhere with different shades of mean-
present, have changed our conceptions of how perception ing. In considering issues in learning, developmental and
begins and develops. These changes, in turn, have gener- educational psychologists often contrast constructivism
ated important implications about the early foundations with associationism, where constructivism emphasizes
of cognitive, linguistic, and social development. the active contributions of the learner. Although a com-
In this chapter, we consider current knowledge of mon thread extends through the uses of constructivism,
early visual perception and its development. Besides de- its use here will be confined to the notion that percep-
scribing the origins and development of these perceptual tion is constructed from sensations and actions through
capabilities, we use them to illustrate general themes: learning. Our primary concern in addressing perception
the several levels of explanation required to understand is to consider, not particular modes of learning, but
perception; the roles of hardwired abilities, maturation, whether basic perceptual abilities are learned at all. For
and learning in perception; and some of the methods this reason, and others, the verdict on constructivism in
that allow assessment of early perception. These themes this domain may differ from the fates of constructivisms
all have broader relevance for cognitive and social in other studies of development.
development. The constructivist account of how perception devel-
ops is familiar to many. The key assumption is that at
birth, sensory systems function to produce only their
THEORIES OF characteristic sensations. Stimulation of the visual sys-
PERCEPTUAL DEVELOPMENT tem yields sensations of brightness and color, along with
some quality (a local sign) correlating with a position
As a backdrop for considering research in early vision, on the retina. Stimulation of the auditory system pro-
we describe two general theories of perceptual develop- duces loudnesses and pitches, and so on. Of course, per-
Theories of Perceptual Development 111

ceptual reality consists not of disembodied colors and that perception must be learned. From an evolutionary
loudnesses, but of objects arranged in space, relations perspective, it might also be considered curious that hu-
among them, and events, characterized by motion and mans have been so disadvantaged, beset with a frail and
change within that space. On the constructivist view of complicated scheme for attaining what mountain goats
perceptual development, all these commonplace occu- possess from birth.
pants of adult perceptual realityany tangible, material The preceding questions are not meant to be critical
object existing in the external world and, indeed, the ex- of generations of serious thinkers who have held the con-
ternal spatial framework itselfare hard-won construc- structivist position. Asking these questions helps to
tions achieved by learning. What allows construction of highlight what the issues were and how things have now
external reality is associative processes. Experiences of changed. The key fact is that the constructivist position
visual sensations coupled with touch, according to was embraced nearly universally because the arguments
Berkeley (1709/1963), allow creation of the idea that for it were logical. If valid, these arguments admitted
seen objects have substance. Connecting the muscular few alternatives. We can better understand current views
sensations of reaching with visual sensations allows the if we briefly review these logical arguments, sometimes
creation of depth and space. Sensations obtained from described as the ambiguity argument and the capability
one view of an object at a given time are associated by argument (Kellman & Arterberry, 1998).
contiguity in space and time, and by similarity. Sensa- The ambiguity argument traces to Berkeley and his
tions obtained a moment later from another view may 1709/1963 book Essay toward a New Theory of Vision.
become associated with the previous ones. An object be- Analyzing the projection of light onto the retina of a sin-
comes a structure of associated sensations stored in gle eye, Berkeley pointed out that while the projection
memory. In John Stuart Mills memorable formulation, onto different retinal locations might carry information
for the mind, an object consists of all the sensations it into the image about the left-right and up-down relations
might give us under various circumstances: An object is of objects in the world, there was no direct information
the permanent possibilities of sensation (Mill, 1865). to indicate the distance to an object. A given retinal
For Piaget (1952, 1954), the account is similar, except image could be the product of an infinitely large set of
that voluntary actions, not just tactile and muscular sen- possible objects (or, more generally, scenes) in the
sations, become associated, making objects consist ini- world. Because of this ambiguity, vision cannot provide
tially of sensorimotor regularities. knowledge of the solid objects in the world or their
How did this basic story of perceptual development three-dimensional (3D) positions and relations. Since
attain such preeminent status in philosophy and psychol- vision is ambiguous, the seeming ability of adult per-
ogy? The question is puzzling because the account was ceivers to see objects and space must derive from asso-
not based on scientific study in any meaningful way. Just ciating visual sensations with extravisual sensations
to anticipate a different possibility, we might consider (such as those involved in eye-muscle adjustments, and
the life of a mountain goat. Unlike a human baby, a with touch and locomotion).
mountain goat is able to locomote soon after birth. Re- The capability argument drew more on physiology
markably, the newborn mountain goat appears to per- than philosophy. The history of progress in understand-
ceive solid surfaces on which to walk and precipices to ing the nervous system reflects a progression from the
avoid. When tested on a classic test apparatus for the outside in. Long before much was known about the vi-
study of depth perceptionthe visual cliff new- sual cortex of the brain, parts of the eye were somewhat
born mountain goats unfailingly avoid the side with the understood. Even in the nineteenth century, it was clear
apparent drop-off (Walk & Gibson, 1961). that the retina contained numerous tiny receptors and
This example puts a fine point on the issue. Although that information left the eye for the brain in a bundle of
mountain goats appear innately able to perceive solidity fibers (the optic nerve). It is not surprising that reason-
and depth, generations of philosophers and psycholo- ing about the capabilities of the visual system centered
gists have argued that, as a matter of logic, humans must on these known elements. Consider the world of a single
be born helpless and must construct space, substance, visual receptor, at some location on the retina. If it ab-
and objects through a long associative process. The sorbs light, this receptor can signal its activation at that
humble mountain goat, as well as many other species, point. Receiving only tiny points of light, the receptor
provides a stark contradiction to any logical argument can know nothing of objects and spatial layout and, as
112 Infant Visual Perception

Berkeley contended, certainly nothing about the third as a viable possibility, but as we will see, a perspective
dimension (depth). To understand the system in aggre- compatible with much of the scientific evidence about
gate, one need only think of many receptors in many lo- how perception develops.
cations, each capable of signaling locally activations We call this view ecological because it connects per-
that the visual system encodes as brightness and color. ceptual capabilities to information available in the world
Clusters of brightnesses and colors are not objects or of the perceiver. Crucial among this information are reg-
scenes; thus, perceiving objects and scenes requires ularities and constraints deeply connected to the basic
something beyond sensations generated by activity in structure and operation of the physical world. These reg-
these receptors. ularities have existed across evolutionary time, and have
To make matters worse, it was understood as a logical shaped the operation of perceptual mechanisms.
matter that such sensations existed not in the world but The emergence of ecological views of perception
in the mind. As Johannes Muller (1838/1965) had and perceptual development owes most to the work
emphasized in his famous doctrine of specific nerve en- of James J. and Eleanor J. Gibson (E. Gibson, 1969;
ergies, whether one presses on the eyeball or whether J. Gibson, 1966, 1979). Earlier influences included the
retinal receptors absorb light, the mind experiences work of the physiologist Hering (18611864), who de-
brightness and color. Similarly, pressure or shock to the scribed the operation of the two eyes in binocular depth
auditory system produces experiences of sound. It seems perception as an integrated, and likely innate, system,
that sensory qualities are specific to the separate and the Gestalt psychologists (e.g., Koffka, 1935;
senses, regardless of the energy used to evoke them. If Wertheimer, 1923/1958) who emphasized the impor-
the visual system can produce only its own characteris- tance of abstract form and pattern, rather than concrete
tic sensations, how can it be said to obtain knowledge of sensory elements, in perception. Important strands of
the world? This is the capability argument: The visual J. Gibsons theories of perception have since been ad-
system, as a system that generates its own characteristic vanced in computational approaches to perception, es-
sensations, usually on stimulation by light, is not capa- pecially that of Marr (1982).
ble of directly revealing the objects, layout, and events Numerous facts lead naturally to a consideration of
of the external world. ecological ideas in perceptual development. Perhaps the
These powerful logical arguments have two conse- simplest is the observation that some species exhibit ef-
quences. One is that the apparent direct contact that fectively functioning perceptual systems from birth, as
we have through vision with a structured, meaningful, in the case of the mountain goat. Historically, however,
external world must be a developmental achievement, the issue that raised the curtain for contemporary views
accomplished through learning to infer the meanings of is the nature of information in perception (J. Gibson,
our sensations. The other consequence is that perceptual 1966, 1979).
knowledge in general must be an inference. Different In a certain sense, this is the logical starting point. If
versions of this theoretical foundation have character- the constructivist view was deemed correct because of
ized perception as inference, hypothesis, results of past logical limits on information received by the senses,
experience, and imagination. In Helmholtzs classic then any alternative view would need to address the am-
statement: Those objects are imagined to be in the field biguity and capability arguments head-on. This is one
of view that have frequently given rise to similar sensa- way of summarizing a several-decades-long effort led by
tions in the past (Helmholtz, 1885/1925). Lest one J. Gibson, foreshadowed in his 1950 book, The Percep-
think that this section has only historical significance, it tion of the Visual World, and emerging fully in The
is not uncommon to encounter precisely these same ar- Senses Considered as Perceptual Systems (1966) and The
guments today (e.g., Purves & Lotto, 2003). Ecological Approach to Visual Perception (1979). Ac-
cording to Gibson, both the ambiguity and capability ar-
guments rest on misunderstandings of the information
available for perception.
The Ecological View
Ecology and Ambiguity
Until recently, students of development have been less
familiar with an alternative to constructivist views of Ambiguity claims about vision centered on analysis of
perceptual development. The view is important, not only static retinal images given to a single eye. If these con-
Theories of Perceptual Development 113

straints are admitted, the analyses by Berkeley and oth- researchers in perception and perceptual development
ers are correct: For any given retinal image, there are in- have been busy ever since exploring the computations
finitely many possible configurations in the world that and mechanisms that extract higher order information.
could give rise to it. The problem with the analysis, how-
ever, is that the inputs to human vision are not restricted
The Contemporary Situation in Perceptual Theory
to single, static retinal images. As Hering (18611864)
had already noted, the two eyes sample the world from Philosophers, most cognitive scientists, and psycholo-
two distinct vantage points. This arrangement makes gists embrace the notion that, in a formal sense, percep-
possible direct information about the third dimension tion has the character of an inference (specifically, an
(which lay at the heart of Berkeleys ambiguity con- ampliative inference, in which the conclusion is not
cerns). J. Gibson argued that another primary fact had guaranteed by the premises or data, Swoyer, 2003). As
been missed: Sophisticated visual systems are the prop- virtual reality systems show us (and as dreams and hal-
erty of mobile organisms. Motion and change provide lucinations impressed Descartes and others), the per-
important information for perception. Although a single ceptual experience of 3D space and certain objects and
retinal image is ambiguous, the transformations over events does not guarantee their objective existence.
time of the optic array as the perceiver moves are highly Such arguments have been elaborated in detail (Fodor
specific to the arrangement of objects, space, and & Pylyshyn, 1981; Ullman, 1980; but see Turvey, Shaw,
events. If one can assume that the world is not deforming & Reed, 1981) to attack J. Gibsons assertion that per-
contingent on the perceivers motion, this kind of infor- ception is direct (it does not require inference). If per-
mation specifies the layout. Evolution may well have ception is formally inferential, perhaps Berkeley and his
picked up on such sources of information, allowing per- intellectual descendants were correct after all about how
ceptual systems to deliver meaningful information not perception must develop. Is there a paradox in holding
derived from learning. Whereas the mountain goat pro- an ecological view while admitting that perception has a
vides an existence proof of functional perception with- formally inferential character?
out learning, the analysis given by J. Gibson explained Resolving this apparent paradox by separating the
how this might be possible. two issues is important to understanding perceptual de-
velopment. Perception has the formal character of an in-
Ecology and Capability
ference, but that does not imply that perception in
Ambiguity issues focus on the information in the world. humans must be learned, or that vision must be supple-
Corresponding to the arguments about information are mented by touch or action. Perceptual inferences may be
revised ideas about the capabilities of a perceptual sys- exactly the kinds of things that have been built into per-
tem (J. Gibson, 1966). The description of inputs to vi- ceptual systems by evolution. Rock (e.g., 1984), a per-
sion in terms of brightness and color responses at ceptual theorist who stressed the inferential nature of
individual locations is inadequate. Further along in the perception, and Marr (1982), who put computational ap-
system are mechanisms sensitive to higher order rela- proaches to perception on a clear footing, were among
tionships in stimulation. There were precedents to this the earliest to articulate that perception could be both
view. Corresponding to Herings point about triangula- inferential and innate.
tion (sampling from two positions) was his assessment The analyses by the Gibsons and later investigators
that the brain handled inputs to the two eyes as a system, influenced the debate about perceptual development
detecting disparities between the two eyes views to by altering conceptions of the information available for
perceive depth. Likewise, the Gestalt psychologists em- perception. For a moving, two-eyed observer with
phasized the contribution of brain mechanisms in pro- mechanisms sensitive to stimulus relations, the ambigu-
cessing relations in the input. J. Gibson pointed out the ities envisioned by Berkeleymany different ordinary
importance of higher order information and suggested scenes leading to the same retinal imagedo not exist.
that perceptual systems are naturally attuned to pick up For Berkeley, visual ambiguity is so expansive that vi-
such information. He did not deal much with neurophys- sion requires lots of outside help. For J. Gibson (1979),
iological or computational details, and he confused visual information specific to arrangements of scenes
some by saying that perceptual systems resonate to in- and events is available, and humans possess perceptual
formation. Gibsons views still evoke controversy, yet mechanisms attuned to such information. In Marr
114 Infant Visual Perception

(1982) may be found a synthesis of the two extremes: ciated and used by those in related fields. The particular
Visual ambiguity is intrinsic but can be handled by rela- topics reflect our areas of expertise and our views of
tively few, general constraints. The interpretation of areas that are rapidly advancing and in which important
optic flow patterns in terms of 3D spatial layout re- knowledge has been gained. Some parts of this chapter
quires the assumption that the scene (or whatever pro- are modestly updated from the previous edition of the
vides images to the two eyes) is not changing contingent Handbook of Child Psychology (Kellman & Banks,
on the observers movements. This assumption is rarely, 1998), whereas others are new. In what follows, we first
if ever, violated in ordinary perception, although it is consider basic visual sensitivities in the infant, includ-
exactly the assumption that is violated when an observer ing acuity and contrast sensitivity, sensitivity to color,
dons the viewing goggles or helmet in a virtual reality pattern, and motion. We then consider spatial percep-
system. Many researchers have suggested that certain tion, object perception, and face perception.
assumptions (e.g., the lack of observer-contingent scene
changes or the movement of objects on continuous
BASIC VISUAL SENSITIVITIES IN INFANCY
space-time paths) have come to be reflected in percep-
tual machinery through evolution (J. Gibson, 1966;
The function of visual perception is to provide the per-
Johansson, 1970; Kellman, 1993; Kellman & Arter-
ceiver with information about the objects, events, and
berry, 1998; Shepard, 1984).
spatial layout in which he or she must think and act.
This possibility has far-reaching consequences with
Starting from this concern, the study of basic visual sen-
the potential to overturn a persistent and dominant view
sitivity and the psychophysical methods used to study
of perception based on learning. Yet it is important
infants visual perception may seem arcane to the non-
to recognize that the mere possibility of innate percep-
specialist. Yet, all higher-level abilities to see the forms,
tual mechanisms (incorporating assumptions about the
sizes, textures, and positions of objects, as well as to
world) does not decide their reality. Unlike mountain
apprehend spatial relations of objects at rest and in mo-
goats, human infants are not mobile at birth, and until re-
tion, depend on basic visual capabilities to resolve infor-
cently, their perceptual abilities were mostly unknown.
mation about spatial position. For this reason, the
Our discussion of constructivist and ecological views of
development of spatial vision has been a topic of great
perceptual development culminates in the observation
concern to those interested in infant perception.
that the answer is a matter for empirical science. More-
We begin an examination of spatial vision by consid-
over, different perceptual abilities may have different
ering sensitivities to variation across changing locations
contributions from native endowments, maturation, and
in the optic array. Two of the most basic dimensions of
learning. Researchers must write the story of each per-
sensitivity in describing spatial vision are visual acuity
ceptual capacity based on experimental evidence.
and contrast sensitivity. Our discussion of these basic
This conclusion sets our agenda for the remainder
capacities leads naturally into an assessment of basic
of this chapter. We consider the emerging scientific pic-
pattern discrimination abilities. We then consider color
ture of development for the crucial components of visual
vision and motion perception.
perception. This picture indicates decisively that, al-
though learning may be involved in calibration and fine-
Visual Acuity
tuning, visual perception depends heavily on inborn and
early maturing mechanisms. This picture has begun to Acuity is a vague term, meaning something like preci-
strongly influence views in other areas of development, sion. A particular variety of acuity is so often used to
as well as conceptions of the nature of perception. More describe visual performance that the phrase visual
unsettling is the failure to attend to the evidence on in- acuity has become its common label. This type of acu-
fant perception in some recent trends in cognitive sci- ity is more technically known as minimum separable
ence and neuroscience. After considering the evidence, acuity or grating acuity. Object recognition and identifi-
we return to these issues at chapters end. cation depend on the ability to encode differences
Our review of the field is necessarily selective. A across positions in the retinal image in luminance or
goal of the present chapter is to place what has been spectral composition. Visual acuity thus refers to the re-
learned about infant vision in a more general historical solving capacity of the visual systemits ability to dis-
and philosophical context, so that it may be easily appre- tinguish fine details or differences in adjacent positions.
Basic Visual Sensitivities in Infancy 115

Measuring this type of visual acuity by various means is 50


by far the most common way of assessing ocular health
and suitability for specific visual tasks, such as operat-

Visual Acuity (Cycles/deg)


ing cars or aircraft.
To assess acuity, high-contrast, black-and-white VEP
10 OKN
patterns of various sizes are presented at a fixed dis-
tance. The smallest pattern or smallest critical pattern FPL
5
element that can be reliably detected or identified is
taken as the threshold value and is usually expressed in
angular units. Many different acuity measures have
been used with adults, but only two have been widely 1
used in developmental studies, grating acuity and
vernier acuity.
0 5 10 15 20 25 30 35
Grating acuity tasks require resolving the stripes in a
Age (Weeks)
repetitive pattern of stripes. The finest resolvable grat-
ing is taken as the measure of acuity and it is generally Figure 3.1 Visual acuity estimates at different ages. The
expressed in terms of spatial frequency, which is the highest detectable spatial frequency of a high-contrast grating
number of stripes per degree of visual angle. Adult stimulus is plotted as a function of age. Circles: Visual evoked
grating acuity under optimal conditions is 45 to 60 cy- potential estimates. Squares: Optokinetic nystagmus (OKN)
estimates. Triangles: Forced-choice preferential looking esti-
cles/degree, which corresponds to a stripe width of 12 to
2
mates. Sources: Measurement of Visual Acuity from Pattern
3 minutes of arc (Olzak & Thomas, 1986). By optimal Reversal Evoked Potentials, by S. Sokol, 1978, Vision Re-
conditions, we mean that the stimulus is brightly illumi- search, 18, pp. 3340. Reprinted with permission; Matura-
nated, high in contrast, presented for at least 12 second, tion of Pattern Vision in Infants during the First 6 Months,
and viewed foveally with a well-focused eye. Change in by R. L. Fantz, J. M. Ordy, and M. S. Udelf, 1962, Journal of
Comparative and Physiological Psychology, 55, pp. 907917.
any of these viewing parameters causes a reduction in
Reprinted with permission; Visual Acuity Development in
grating acuity. Human Infants up to 6 Months of Age, by J. Allen, 1978, un-
Vernier acuity is tested in tasks requiring discrimina- published masters thesis, University of Washington, Seattle,
tion of positional displacement of one small target rela- WA. Reprinted with permission.
tive to another. The most common variety involves
distinguishing whether a vertical line segment is dis-
placed to the left or right relative to a line segment just
below it. In adults, the just-noticeable offset under opti- trates two points. First, acuity is low at birth and devel-
mal conditions is 2 to 5 seconds of arc (Westheimer, ops steadily during the 1st year. Grating acuity during
1979). Because this distance is smaller than the diame- the neonatal period is so low that these infants could be
ter of a single photoreceptor in the human eye, this kind classified as legally blind. Second, the acuity estimates
of performance has been called hyperacuity (West- obtained with behavioral techniques such as FPL and
heimer, 1979). As with grating acuity, the lowest OKN are generally lower than those obtained using elec-
vernier acuity thresholds are obtained when the stimu- trophysiological techniques such as VEP. Grating acuity
lus is brightly illuminated, high in contrast, presented develops beyond the 1st year and reaches adult levels
for at least 12 second, and viewed foveally with a well- around 6 years of age (e.g., Skoczenski & Norcia, 2002).
focused eye. We discuss the optical, receptoral, and neural factors
There have been numerous measurements of grating that determine grating acuity as a function of age in the
acuity (the highest detectable spatial frequency at high section on contrast sensitivity.
contrast) in human infants. Figure 3.1 plots grating acu- There have been fewer measurements of vernier acu-
ity as a function of age for some representative studies. ity; nonetheless, some intriguing observations have been
The displayed results were obtained using three re- reported. Shimojo and colleagues (Shimojo, Birch,
sponse measurement techniques: Forced-choice prefer- Gwiazda, & Held, 1984; Shimojo & Held, 1987) and
ential looking (FPL), optokinetic nystagmus (OKN), Manny and Klein (1984, 1985) used FPL to measure the
and the visual evoked potential (VEP). This figure illus- smallest offset infants could respond to at different
116 Infant Visual Perception

ages. They found that vernier acuity was much poorer in Adult contrast sensitivity and grating acuity are
8- to 20-week-old infants than in adults. The ratio of limited by optical, receptoral, and neural factors. Sensi-
adult vernier acuity divided by 8-week olds vernier tivity is best with good lighting, foveal fixation, suffi-
acuity is significantly greater than the corresponding ciently long stimulus duration, and a well-focused eye.
ratio for grating acuity. A similar finding has emerged Decreased illumination reduces both contrast sensitiv-
from VEP measurements of vernier and grating acuity; ity and the high-frequency cutoff (van Nes & Bouman,
adult levels of hyperacuity were not reached until 10 to 1967). Similar changes in contrast sensitivity occur
14 years of age (Skoczenski & Norcia, 2002). This sug- when the stimulus is imaged on the peripheral retina
gests that the visual mechanisms that limit vernier acu- (Banks, Sekuler, & Anderson, 1991) or the eye is not
ity undergo greater change with age than do the well focused (Green & Campbell, 1965). Understanding
mechanisms limiting grating acuity. Different hypothe- limitations on adult vision has been aided by modeling
ses have been offered concerning the differing growth the early stages of vision as a series of filtering stages.
rates (Banks & Bennett, 1988; Shimojo & Held, 1987; Visual stimuli pass sequentially through the eyes op-
Skoczenski & Norcia, 2002); however, direct empirical tics, which are responsible for forming the retinal
tests are needed. image; the photoreceptors, which sample and transduce
the image into neural signals; and two to four retinal
Contrast Sensitivity neurons, which transform and transmit those signals
into the optic nerve and eventually to the central visual
Contrast sensitivity refers to the ability to detect varia- pathways. In these early stages of visual processing,
tions in luminance. Most acuity testing is done at high considerable information is lost. The high-frequency
contrast (e.g., black characters on a white background falloff observed in the adult CSF is determined, by and
or gratings varying from white to black). Testing for large, by the filtering properties of the eyes optics and
contrast sensitivity involves finding the least difference the photoreceptors (Banks, Geisler, & Bennett, 1987;
between luminances that allows detection of structure. Pelli, 1990; Sekiguchi, Williams, & Brainard, 1993).
The contrast sensitivity function (CSF) represents the The loss of high-frequency sensitivity with peripheral
visual systems sensitivity to sinusoidal gratings of var- viewing has been modeled successfully by examination
ious spatial frequencies. The CSF has generality as an of the optics, receptors, and retinal circuits of the pe-
index of visual sensitivity because any two-dimensional ripheral retina (Banks et al., 1991). The sensitivity loss
pattern can be represented by its spatial frequency con- that accompanies a reduction in illumination has also
tent and, consequently, one can use the CSF along with been modeled reasonably successfully, at least at high
linear systems analysis to predict visual sensitivity to a spatial frequencies (Banks et al., 1987; Pelli, 1990) as
wide range of spatial patterns (Banks & Salapatek, has the loss that accompanies errors in the eyes focus
1983; Cornsweet, 1970). Thus, measurements of con- (Green & Campbell, 1965). From the emerging under-
trast sensitivity as a function of age should allow the standing of the optical, receptoral, and neural mecha-
prediction of sensitivity to and even preference for nisms that determine contrast sensitivity in adults,
many visual stimuli (Banks & Ginsburg, 1985; Gayl, attempts have been made to use similar techniques to
Roberts, & Werner, 1983). understand the development of contrast sensitivity in
The adult CSF has a peak sensitivity at 3 to 5 cy- human infants.
cles/degree, so the lowest detectable contrasts occur for Figure 3.2 displays an adult CSF measured using a
gratings of medium spatial frequency. At those spatial psychophysical procedure, along with infant CSFs mea-
frequencies, the just-detectable grating has light stripes sured using forced-choice preferential looking (Atkin-
that are only 0.5% brighter than the dark stripes. At son, Braddick, & Moar, 1977; Banks & Salapatek, 1978)
progressively higher spatial frequencies, sensitivity falls and the visual evoked potential (Norcia, Tyler, & Allen,
monotonically to the so-called high-frequency cutoff at 1986; Pirchio, Spinelli, Fiorentini, & Maffei, 1978).
about 50 cycles/degree. This is the finest grating an These data illustrate two common observations. First,
adult can detect when the contrast is 100% and it corre- contrast sensitivity (and grating acuity) in young infants
sponds to the persons grating acuity. At low spatial is substantially lower than that of adults, with the differ-
frequencies, sensitivity falls as well, although the steep- ence diminishing rapidly during the 1st year. Second, as
ness of this falloff is highly dependent on the conditions we saw earlier in Figure 3.1, measurements with the vi-
of measurements. sual evoked potential typically yield higher sensitivity
Basic Visual Sensitivities in Infancy 117

1000 rapid increases in acuity suggesting that visual input is


necessary for visual functioning. Beyond knowing that
visual input is necessary, the specific causes, anatomi-
cal and physiological, of the striking functional deficits
Adult CSF observed during the first few months of life are still
Contrast Sensitivity

100 being debated. Some investigators have proposed that


one can explain the low contrast sensitivity and grating
1-month VEP acuity of neonates as due to information losses caused
by optical and retinal immaturities (Jacobs & Blake-
10 more, 1988; Wilson, 1988, 1993); others have argued
1-month FPL
that those immaturities are not the whole story (Banks
& Bennett, 1988; Banks & Crowell, 1993; Brown, Dob-
son, & Maier, 1987).
Development of the eye and retina are important fac-
1 tors. Large ocular and retinal changes occur in develop-
0.1 1 10 100 ment and they have profound effects on the ability to see
Spatial Frequency (cycles/deg) spatial patterns. The eye grows significantly from birth
Figure 3.2 Adult and 1-month-old infant contrast sensitiv- to adolescence, with most growth occurring in the 1st
ity functions (CSFs). Contrast sensitivity is plotted as a func- year. The distance from the cornea at the front of the
tion of spatial frequency (the numbers of grating cycles per eye to the retina at the back is 16 to 17 mm at birth, 20
degree of visual angle). The upper dotted curve is an adult to 21 mm at 1 year, and 23 to 25 mm in adolescence and
CSF that was measured psychophysically. The lower solid adulthood (Hirano, Yamamoto, Takayama, Sugata, &
curve is the average of 1-month CSFs, measured using
Matsuo, 1979; Larsen, 1971). Shorter eyes have smaller
forced-choice preferential looking. The middle dash curve is
the average of 1-month CSFs, measured using visual-evoked retinal images. So, for example, a 1-degree target sub-
potential. Sources: Acuity and Contrast Sensitivity in 1-, 2-, tends about 200 microns on the newborns retina and
and 3-Month-Old Human Infants, by M. S. Banks and P. 300 microns on the adults (Banks & Bennett, 1988;
Salapatek, 1978, Investigative Ophthalmology and Visual Sci- Brown et al., 1987; Wilson, 1988). Thus, if newborns
ence, 17, pp. 361365. Reprinted with permission and De-
had the retinae and visual brains of adults, one would
velopment of Contrast Sensitivity in the Human Infant, by
A. M. Norcia, C. W. Tyler, and R. D. Hammer, 1990, Vision expect their visual acuity to be about two-thirds that of
Research, 30, pp. 14751486. adults simply because they have smaller retinal images
to work with.
Another ocular factor relevant to visual sensitivity is
the relative transparency of the ocular media. Two as-
(and acuity) estimates than do behavioral techniques pects of ocular media transmittance are known to change
(see Mayer & Adrendt, 2001 for a review). The time with age: the optical density of the crystalline lens pig-
course differs depending on whether a behavioral or ment and that of the macular pigment. In both cases,
electrophysiological technique is used. With evoked po- transmittance is slightly higher in the young eye, particu-
tential measurements, peak sensitivity approaches adult larly at short wavelengths (Bone, Landrum, Fernandez, &
values by 6 months of age, whereas behavioral measure- Martinez, 1988; Werner, 1982). Thus, for a given amount
ments exhibit a slower developmental time course. Not of incident light, the newborns eye actually transmits
illustrated is the systematic variability in the CSF slightly more to the photoreceptors than does the mature
across infants (Peterzell, Werner, & Kaplan, 1995). Al- eye. This developmental difference ought to favor the
though group functions are smooth in shape, individual newborn compared with the adult, but only slightly.
functions are not. The ability of the eye to form a sharp retinal image is
What accounts for the development of acuity and con- yet another relevant ocular factor. This ability is typi-
trast sensitivity? Infants who experience visual depriva- cally quantified by the optical transfer function. There
tion early in life due to monocular or binocular cataracts have been no measurements of the human neonates op-
show newborn levels of acuity once the cataract(s) are tical transfer function, but the quality of the retinal
removed (Maurer & Lewis, 1999), despite being 1 to 9 image almost certainly surpasses the resolution perfor-
months of age. Longitudinal follow-up, however, shows mance of the young visual system (Banks & Bennett,
118 Infant Visual Perception

1988). Thus, it is commonly assumed that the optical visual signal (Banks & Bennett, 1988; Banks & Crowell,
transfer function of the young eye is adultlike (Banks 1993). They concluded that the adult foveal cone lattice
& Crowell, 1993; Wilson, 1988, 1993). Refractive er- is dramatically better at absorbing photons of light and
rors or accommodation errors diminish the sharpness converting them into visual signals. By their calcula-
of the retinal image and thereby decrease sensitivity tions, if identical patches of light were presented to new-
to high spatial frequencies (Green & Campbell, 1965). born and adult eyes, roughly 350 photons would be
Hyperopic and astigmatic refractive errors are com- effectively absorbed in adult foveal cones for every pho-
mon in infants (Banks, 1980a; Howland, 1982); they ton absorbed in newborn cones. Similar estimates were
tend not to accommodate accurately until 12 weeks obtained by Wilson (1988, 1993). The newborns fovea
(Banks, 1980b; Braddick, Atkinson, French, & How- is less able to use light entering the eye than is the ma-
land, 1979; Haynes, White, & Held, 1965). Nonethe- ture fovea.
less, it is widely believed that infants refractive and The cones of the immature fovea are also more
accommodative errors do not constrain sensitivity or widely spaced than those of the adult (Banks & Bennett,
visual acuity significantly (Banks, 1980a, 1980b; 1988; Banks & Crowell, 1993; Wilson, 1988, 1993).
Braddick et al., 1979; Howland, 1982). Cone-to-cone separation in the center of the fovea is
If optical imperfections do not contribute signifi- about 2.3, 1.7, and 0.58 minutes of arc in neonates, 15-
cantly to the visual deficits observed in young in- month-olds, and adults, respectively. These dimensions
fants, receptoral and postreceptoral processes must impose a physical limit (the so-called Nyquist limit) on
do so. The retina and central visual system all exhibit the highest spatial frequency that can be resolved with-
immaturities at birth (Banks & Salapatek, 1983; out distortion or aliasing (Williams, 1985). From the
Hendrickson, 1993; Hickey & Peduzzi, 1987; current estimates of cone spacing, the foveas of new-
Yuodelis & Hendrickson, 1986), but morphological borns, 15-month-olds, and adults should theoretically be
immaturities are evident in the fovea, particularly unable to resolve gratings with spatial frequencies above
among the photoreceptors. 15, 27, and 60 cycles/degree, respectively.
The development of the fovea is dramatic in the 1st Investigators have calculated the contrast sensitivity
year of life, but subtle morphological changes continue and visual acuity losses that ought to be observed if the
until at least 4 years of age (Yuodelis & Hendrickson, only difference between the spatial vision of newborns
1986). The fovea, defined as the part of the retina that and adults were the eyes optics and the properties of the
contains no rods, is much larger at birth than in adult- foveal cones (Banks & Bennett, 1988; Banks & Crowell,
hood: Its diameter decreases from roughly 5.4 degrees at 1993; Wilson, 1988, 1993). The expected losses are sub-
birth to 2.3 degrees at maturity. Moreover, the individual stantial: Contrast sensitivity to medium and high spatial
cells and their arrangements are very different at birth frequencies is predicted to be as much as 20-fold lower in
than they will be later on. The newborns fovea possesses neonates than in adults. Nonetheless, the observed con-
three discernible layers of neuronsthe photoreceptors, trast sensitivity and grating acuity deficits in human
the neurons of the outer nuclear layer, and the retinal newborns are even larger than predicted (e.g., Skoczen-
ganglion cellswhereas the mature fovea contains only ski & Aslin, 1995), so this analysis of information losses
one layer, which is composed of photoreceptors. The in the optics and receptors implies that there are other
most dramatic histological differences, however, are the immaturities, presumably among retinal neurons and
sizes and shapes of foveal cones. Neonatal cones have central visual circuits, that contribute to the observed
inner segments that are much broader and shorter. The loss of contrast sensitivity and grating acuity.
outer segments are distinctly immature, too, being much Another hypothesis concerning the contrast sensitiv-
shorter than their adult counterparts. These shape and ity and visual acuity of young infants has been offered.
size differences render the newborns foveal cones less Because of the obvious immaturity of the fovea, perhaps
sensitive than those of the adult (Banks & Bennett, infants use another part of the retina to process points of
1988; Brown et al., 1987). interest in the visual scene. Cones in the parafoveal and
To estimate the efficiency of the neonates lattice of peripheral retina are relatively more mature at birth
foveal cones, Banks and colleagues calculated the abil- than their foveal counterparts, but they, too, undergo
ity of the newborns cones to capture light in the inner postnatal development (Hendrickson, 1993). The data,
segment, funnel it to the outer segment, and produce a however, do not support this hypothesis: Young infants
Basic Visual Sensitivities in Infancy 119

best acuity and contrast sensitivity is obtained with Direct behavioral tests of orientation sensitivity have
foveal stimulation. Lewis, Maurer, and Kay (1978) revealed evidence that it is innate. Slater, Morison, and
found that newborns could best detect a narrow light bar Somers (1988) used habituation measures with high-
against a dark background when it was presented in cen- contrast striped patterns. They found dishabituation to
tral vision, and D. Allen, Tyler, and Norcia (1996) changed orientation in situations where other stimulus
showed that visual evoked potential (VEP) acuity and variables (such as whether a particular screen position
contrast sensitivity is higher in central than in peripheral was black or white) could be ruled out. Their results
vision in 8- to 39-week-olds, by an average factor of 2.3. were confirmed by Atkinson, Hood, and Wattam-Bell
An important question that will be pursued vigor- (1988). Orientation sensitivity appears to be innate in
ously in future research is what factors not considered in humans, although it improves in the early weeks of life.
the preceding analyses account for the unexplained por-
tion of the contrast sensitivity and grating acuity losses.
Pattern Discrimination
There are numerous candidates including internal neural
noise (such as random addition of action potentials at Assessing acuity and contrast sensitivity largely involve
central sites; Skoczenski & Norcia, 1998), inefficient comparing responses to something versus nothing. The
neural sampling, and poor motivation to respond. exquisite spatial resolution of vision, however, serves
functions beyond mere detection. Encoding and discrim-
inating patterns, surfaces, and objects are key tasks of
Orientation Sensitivity
visual processing. Thus, describing pattern-processing
Sensitivity to orientation is an important foundation of capabilities in infant vision is important. But how can
much of higher level vision, such as perception of edges, pattern-perception capabilities be assessed in a compre-
patterns, and objects. In monkeys, it is well established hensive way? As in studies of adult vision, linear systems
that orientation sensitivity is innately present (Wiesel & theory from mathematics and signal processing is useful.
Hubel, 1974), and in cats orientation sensitivity also ap- Any distribution of luminance ( light and dark) in an
pears soon after birth, with or without visual experience image can be described, using a 2D Fourier transform, as
(Hubel & Wiesel, 1963). Paradoxically, development of a set of sinusoidally varying luminance components hav-
orientation sensitivity has been the topic of numerous ing particular frequencies and amplitudes, in particular
learning simulations in recent years (Linsker, 1989; Ol- orientations. Because any image can be analyzed in this
shausen & Field, 1996; von der Malsburg, 1973). These way, the frequency components form an important char-
results suggest interesting relations between orienta- acterization of the pattern. If the spatial phase of each
tion-sensitive cortical units and the statistics of images component is also encoded, the pattern is completely de-
of natural scenes. Such studies are often interpreted as scribed. Researchers have made progress characterizing
showing how the visual brain gets wired up by experi- infant pattern discrimination using linear systems con-
ence after birth (e.g., Elman, Bates, Johnson, cepts. This work has used tests of infants abilities to
Karmiloff-Smith, Parisi, & Plunkett, 1996). distinguish simple, suprathreshold patterns that vary in
Yet the evidence suggests that basic orientation sensi- contrast or in phase.
tivity in humans, as in monkeys and cats, is present at Sensitivity to contrast differences is typically mea-
birth. Some maturation of orientation processing was sured by presenting two sine-wave gratings of the same
suggested by visual evoked potential (VEP) studies by spatial frequency and orientation but differing con-
Braddick, Atkinson, and Wattam-Bell (1986). Their re- trasts. In experiments with adults, a participant is asked
sults showed responses emerging at 2 to 3 weeks for to indicate the grating of higher contrast. The increment
slowly modulated orientation changes (3 reversals/ in contrast required to make the discrimination varies
second) and responses at 5 to 6 weeks for more rapid ori- depending on the common contrasts of the two stimuli;
entation changes. In an elegant analysis, these investiga- as the common contrast is increased, a successively
tors showed that the pace of these developments was larger increment is required (Legge & Foley, 1980). Six-
maturational, in that preterm infants of the same gesta- to 12-week-old infants require much larger contrast in-
tional age showed patterns of development similar to crements than adults when the common contrast is near
full-term infants. In other words, gestational age, not detection threshold. At high common contrasts, how-
weeks of visual experience was crucial. ever, infants discrimination thresholds resemble those
120 Infant Visual Perception

of adults (Brown, 1993; Stephens & Banks, 1987). technique in which the contrasts of the constituent spa-
These findings suggest that infants ability to distin- tial frequencies from one pattern were combined with
guish spatial patterns on the basis of contrast differ- the phases of the constituent frequencies from the other
ences is poor at low contrast and reasonably good at high pattern. The perceptual appearance of these hybrid pat-
contrast. Different explanations for infants perfor- terns is most closely associated with the pattern from
mance in this task have been offered, but none has been which the phases rather than the contrasts came (Oppen-
confirmed by empirical observation (Brown, 1993; heim & Lim, 1981; Piotrowski & Campbell, 1982);
Stephens & Banks, 1987). stated another way, the hybrid pattern that appears most
Studies have also addressed discrimination based on facelike is the one that contains the phases from the
spatial phase differences. Spatial phase refers to the rel- original schematic face. Not surprisingly, 8-week-olds
ative position of the spatial frequency components (the preferred to fixate the hybrid that contained the phases
sine-wave gratings) of which the pattern is composed of the face and the contrasts of the lattice. Newborns
(Piotrowski & Campbell, 1982). Phase information is preferences, however, were for the hybrid that contained
crucial for the features and relations that are involved in the phases of the lattice and the contrasts of the face.
object perception, such as edges, junctions, and shape. One interpretation of this finding is that newborns are
Altering phase information in a spatial pattern greatly relatively insensitive to spatial phase, but other interpre-
affects its appearance and perceived identity to adults tations have been suggested (e.g., Badcock, 1990).
(Oppenheim & Lim, 1981). In phase discrimination The observation that young infants seem relatively
tasks, the subject is asked to distinguish between two insensitive to variations in spatial phase is extremely
patternsusually gratingsthat differ only in the important. If valid, it suggests that young infants ability
phase relationships among their spatial frequency com- to discriminate spatial patterns has a significant defi-
ponents. Adults are able to distinguish patterns that dif- ciency that is at least qualitatively similar to the
fer only slightly in the phases of their components when deficiency observed in the peripheral visual field of nor-
the stimulus is presented to the fovea (Badcock, 1984). mal adults (Bennett & Banks, 1987; Rentschler &
The ability to discriminate phase can fall dramatically, Treutwein, 1985) and in the central visual field of am-
however, when the stimulus is presented in the periph- blyopic adults (Levi, Klein, & Aitsebaomo, 1985). In
eral visual field (Bennett & Banks, 1987; Rentschler & functional terms, infants processing of basic perceptual
Treutwein, 1985). properties of objects, such as unity, size, shape, texture,
Relatively little work has directly addressed infants and so on depend implicitly on processing of phase in-
ability to use phase differences to discriminate spatial formation. To the extent that it is poor in the earliest
patterns. Braddick et al. (1986) presented periodic pat- weeks of life, these abilities will be limited. Conversely,
terns composed of different spatial frequency compo- tests of certain of these perceptual abilities, to be dis-
nents. When the components were added in one phase cussed, indicate striking newborn perceptual competen-
relationship, the resultant was a square-wave grating (a cies (e.g., for seeing object size and faces). One of the
repeating pattern of sharp-edged light and dark stripes); challenges of infant vision research is reconciling cer-
when the components were added in another phase, the tain poor sensitivity to basic sensory properties, such as
resultant appeared to adults to be a very different, more phase, with evidence of higher order abilities, such as
complex pattern. Eight-week-olds were able to discrimi- face perception. The most likely resolution of the appar-
nate these patterns. Remarkably, however, 4-week-olds ent paradox is that infant sensory capacities for proper-
seemed unable to make the discrimination. ties such as phase and orientation are worse than adults
In a similar vein, Kleiner (1987) and Kleiner and but not completely lacking, even at birth (for further dis-
Banks (1987) examined visual preferences for patterns cussion, see Kellman & Arterberry, 1998).
in which the phases of the constituent components were
altered. Kleiner and colleagues found that newborns and Color Vision
8-week-olds exhibit reliable fixation preferences for a
schematic face over a rectangle lattice (Fantz & Nevis, The term color refers to the component of visual experi-
1967). To examine the influence of spatial phase on fix- ence characterized by the psychological attributes of
ation preference, Kleiner used an image-processing brightness, hue, and saturation. Two of thesehue and
Basic Visual Sensitivities in Infancy 121

saturationare chromatic attributes, and the other the relative activities of the three photoreceptor types to
brightnessis actually an achromatic attribute. Hue is distinguish different colors.
primarily correlated with the dominant wavelength of The subsequent stages of the visual process must uti-
the stimulus whereas brightness is primarily correlated, lize the outputs of the different receptor types in a com-
but not isomorphic, with stimulus intensity. Saturation is plex way to produce the conscious experience of color.
correlated with the distribution of wavelengths in a stim- Psychophysical evidence from adult humans and physio-
ulus; stimuli with more broad band light mixed in are logical evidence from adult monkeys indicate that the
seen as less saturated. We refer to visual discriminations signals of the three cone types undergo a major transfor-
on the basis of differences in hue or saturation as chro- mation in the retina. Signals from two or three kinds of
matic discriminations and discriminations on the basis of cones are combined additively to form achromatic chan-
differences in brightness as achromatic discriminations. nels (coding brightness primarily) and are combined
The functional importance of perceiving color has subtractively to form two kinds of chromatic channels
been a matter of debate. Humans readily perceive ob- (coding hue primarily). The subtractive, chromatic
jects and events from nonchromatic displays, such as channels (red/green and blue/yellow) have been called
those in black-and-white movies or television. Why, opponent processes because different wavelength bands
then, have we evolved elaborate color vision mecha- evoke different directions of neural response.
nisms? In ordinary seeing, chromatic information prob- Many of the characteristics of photoreceptors and
ably aids object segmentation and recognition. In cases subsequent neural stages were originally inferred from
in which an object and its background are equal or adult behavioral studies. Our discussion of color vision
nearly equal in luminance, the objects shape can be centers on two questions:
perceived from chromatic differences. Chromatic infor-
mation can also help distinguish one version of an object 1. What hues are infants sensitive to and when?
(a red apple) from another (a green apple). Less well un- 2. What mechanisms account for the development of
derstood, but important, are the obvious contributions color vision?
of color to our aesthetic experiences.
Origins of Hue Discrimination
The human visual system has four types of photore-
ceptors, one type of rod and three types of cones. The When can infants discriminate stimuli on the basis of
cones are active under daylight viewing conditions and hue alone? Before 1975, a large number of behavioral
subserve color vision; rods are active under quite dim il- studies attempted to answer this question, but they all
lumination. We consider only cones in our discussion of failed to eliminate the possibility that infants were bas-
color vision. ing their discriminations on brightness cues rather than
The three cone types are sensitive to different, but hue (or saturation) cues (Kessen, Haith, & Salapatek,
overlapping, bands of wavelength. The cone types are 1970). To demonstrate convincingly that infants can dis-
generally called short-wavelength-sensitive (S), medium- criminate on the basis of hue alone, researchers have
wavelength-sensitive (M), and long-wavelength-sensitive used two strategies to rule out brightness artifacts.
(L) cones. (We prefer this terminology to the terms blue, (Elsewhere, we describe in detail the importance and
green, and red cones because those terms imply that difficulty of separating hue from brightness responses;
each cone type is responsible for the perception of a par- Kellman & Arterberry, 1998; Kellman & Banks, 1997.)
ticular hue, and this is not the case.) Each type of pho- The methods involve presenting two stimuli differing
toreceptor responds in an untagged fashion; that is, only in hue (e.g., red and green) and looking for a systematic
response quantity, and nothing else, varies with changes response (e.g., directional eye movement, VEP, or FPL)
in the incident light. The consequences of untagged re- to one as evidence for hue discrimination. One strategy
sponding are profound. The output of any single photore- for eliminating brightness artifacts involves using the
ceptor type can be driven to a given level by virtually spectral sensitivity function to match the brightnesses
any wavelength of light simply by adjusting the lights of two stimuli to a first approximation and then by vary-
intensity. Thus, information about the wavelength of a ing the luminances (a measure of stimulus intensity)
stimulus cannot be extracted from the output of a single of the stimuli unsystematically from trial to trial over
photoreceptor type. Instead the visual system must use a wide-enough range to ensure that one is not always
122 Infant Visual Perception

brighter than the other. Systematic responding by the in- used a neutral-point test to examine 8-week-olds color
fant to one of the two chromatic stimuli, across lumi- vision. They examined both white-on-white luminance
nances, can therefore not be attributed to discrimination discrimination and discrimination of chromatic targets
on the basis of brightness. Using this strategy, Oster from white. The colors of the test targets and back-
(1975), and Schaller (1975) demonstrated hue discrimi- ground are represented in Figure 3.3, which is a chro-
nation in 8- and 12-week-old infants, respectively. maticity diagram. Eight-week-olds discriminated many
The second strategy for eliminating brightness cues colors from white: red, orange, some greens, blue, and
was developed by Peeples and Teller (1975); subse- some purples; these colors are represented by the filled
quently, many others have used this strategy, so we ex-
plain it in some detail. They also used spectral sensitivity
data to match approximately the brightnesses of their
stimuli. They then varied luminance systematically 520
around the estimate of the brightness match. Several lu- 530
0.8
minances were presented, bridging a 0.8 log unit range in 540
510
small steps. Consequently, at least one of the luminance 0.7 550
pairings must have been equivalent in brightness for each 560
G
of the infants. Peeples and Teller showed that 8-week-olds 0.6
570
could discriminate red from white for all luminance pair- 500
0.5 580
ings. They concluded that 8-week-olds make true hue dis-
Y W Y 590
criminations.
0.4 600
Thus, three reports in 1975, using different tech- 530
niques, provided the first convincing evidence that 8- to 630
0.3 490 R 650
B
16-week-olds can make chromatic discriminations. P
Today, the story has been further refined: M and L cones 0.2
appear to function by 8 weeks of age and possibly as 480
0.1
early as 4 weeks (e.g., Bieber, Knoblauch, & Werner,
470
1998; Kelly, Borchert, & Teller, 1997); however, S cone 450
0.0
functionality does not appear to emerge until at least 3 to 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
4 months of age (e.g., Crognale, Kelly, Weiss, & Teller, X
1998; Suttle, Banks, & Graf, 2002). At birth, infants may Figure 3.3 The stimuli used in neutral-point experiments.
have very limited color experience, and during the first 4 Participants in both experiments were 8-week-old infants.
months of life their world becomes increasingly filled The format of the figure is the CIE Chromaticity diagram,
with color. And by 4 months, infants have color prefer- which allows one to plot chromatic stimuli differing in hue
and saturation. Saturated colors are represented at the exte-
ences that mirror adults: Saturated colors (such as royal
rior of the diagram, and unsaturated colors toward the mid-
blue) are preferred over less saturated colors (such as dle. The right corner of the diagram (around 650) represents a
pale blue; Bornstein, 1975). hue of red, the top of the diagram represents a hue of bluish-
green ( labeled 520) and the lower left corner represents a hue
Assessing Color Vision of violet (near 400). Each circular symbol represents a color
Three sorts of hue discriminationsRayleigh, tritan, that was presented to infants in these two experiments. Open
symbols represent hues that all infants failed to discriminate
and neutral-pointare particularly interesting theoreti- from white (W). Half-filled symbols represent hues that
cally, and research on infants ability to make these dis- some, but not all, infants discriminated from white. Filled
criminations fills out the picture of early competencies symbols represent hues that all infants reliably discriminated
and deficits. from white. Sources: Color Vision and Brightness Discrim-
The neutral-point test is based on the observation that ination in Human Infants, by D. R. Peeples and D. Y. Teller,
1975, Science, 189, pp. 11021103. Reprinted with permis-
color-normal adults are able to distinguish all spectral
sion and Discrimination of Chromatic from White Light by
(single wavelength) lights from white; that is, they do 2-Month-Old Human Infants, by D. Y. Teller, D. R. Peeples,
not exhibit a neutral point in such a comparison. Peeples and M. Sekel, 1978, Vision Research, 18, pp. 4148.
and Teller (1975) and Teller, Peeples, and Sekel (1978) Reprinted with permission.
Basic Visual Sensitivities in Infancy 123

symbols in the figure. Eight-week-olds did not discrimi- Twelve-week-olds were able to make Rayleigh discrimi-
nate yellow, yellow-green, one green, and some purples nations with 4- and 8-degree targets, but not 1- and 2-
from white; these are represented by the open symbols degree targets. D. Allen et al. (1988) and Clavadetscher
in the figure. Thus, 8-week-old infants seemed to exhibit et al. (1988) confirmed the Rayleigh discrimination
a neutral zone running from short wavelengths to yellow finding. They reported that 3- to 4-week-olds could not
and green, resulting from deficient S cones (in color distinguish a red target on a green background; 7- to 8-
parlance, they have tritanopia or tritanomalous trichro- week-olds could make this discrimination reliably.
macy; Teller et al., 1978). Later, Adams, Courage, and In sum, there is little evidence that the majority of in-
Mercer (1994) reported that the majority of newborns fants 4 weeks of age or younger make hue discrimina-
were able to discriminate broadband red from white and tions with the exception of discriminating red from
the majority were unable to discriminate blue, green, white. The paucity of positive evidence is consistent
and yellow from white. These results are quite similar to with the hypothesis that human neonates are generally
the 8-week results reported by Teller et al. (1978). color deficient. By 4 months of age, infant color vision
A tritan test is designed to assess the function of S abilities approximate adult abilities, although there con-
cones. By presenting two lights that activate M and L tinue to be differences between infants and adults
cones equally, the test isolates the S cones. Varner, chromatic profiles throughout the 1st year of life (Crog-
Cook, Schneck, McDonald, and Teller (1985) asked nale et al., 1998). We now turn to the question of what
whether 4- to 8-week infants could distinguish two such mechanism(s) underlie this development.
lights. Specifically, they presented violet targets in a
How Does Early Color Vision Develop?
green background. Eight-week-olds distinguished the
two lights at all luminances, so they do not appear to Two explanations have been proposed to account for
have an S-cone deficiency. Four-week olds, on the other young infants hue discrimination failures. One possi-
hand, did not discriminate the two lights reliably, sug- bility is the absence or immaturity of different cone
gesting that they have an S-cone defect. D. Allen, types or immaturities among postreceptoral chromatic
Banks, and Schefrin (1988) and Clavadetscher, Brown, channels. Banks and Bennett (1988) have called this the
Ankrum, and Teller (1988) confirmed this finding: In chromatic deficiency hypothesis. There is, however,
their experiment, 3- to 4-week-olds could not distinguish another possibility, raised initially by Banks and Ben-
a violet target on a green background, but 7- to 8-week- nett (1988) and elaborated by Brown (1990), Banks and
olds could. More recently, Teller, Brooks, and Palmer Shannon (1993), Teller and Lindsey (1993), and D,
(1997) found that tritan stimuli did not drive direction- Allen, Banks, and Norcia (1993). Perhaps neonates
ally appropriate eye movements even at 16 weeks of age. have a full complement of functional cone types and the
Rayleigh discrimination tests involve distinguishing requisite neural machinery to preserve and compare
brightness-matched, long-wavelength lights such as red their signals, but overall visual sensitivity is so poor
and green. They are diagnostically important because that it does not allow them to demonstrate their chro-
adults with the most common color defectsdeutera- matic capabilities. On this account, older infants may
nopia ( lacking M cones) and protanopia ( lacking L exhibit reliable chromatic discrimination because of in-
cones)are unable to make such discriminations. creased visual sensitivity. In this context, visual sensi-
Hamer, Alexander, Teller (1982) and Packer, Hartmann, tivity might include discrimination performance of a
and Teller (1984) examined the ability of 4-, 8-, and 12- visual system limited by optical and photoreceptor
week-olds to make Rayleigh discriminations. Either a properties plus a general postreceptoral loss. This hy-
green or red target was presented at one of a variety of pothesis has been called the visual ef ficiency hypothesis
luminances on a yellow background. Most 8-week-olds (D. Allen et al., 1993) and the uniform loss hypothesis
and essentially all 12-week-olds made these discrimina- (Teller & Lindsey, 1993).
tions reliably, providing clear evidence that most infants There is an interesting way to compare the chromatic
do not exhibit deutan or protan defects by 8 weeks of efficiency and visual efficiency explanations experi-
age. In contrast, the majority of 4-week-olds did not ex- mentally. Consider measurements of hue discrimination
hibit the ability to make either discrimination. Packer threshold (e.g., the chromatic contrast required to medi-
et al. (1984) also found a significant effect of target size. ate the discrimination of two lights of equal brightness
124 Infant Visual Perception

but different wavelength compositionsthe chromatic stimuli, which has implications for the relative involve-
threshold) and a brightness discrimination threshold ment and development of the magnocelluar and parvo-
(e.g., the luminance contrast required to mediate the dis- celluar pathways, which are responsible for spatial and
crimination of two lights of the same wavelength com- temporal locations of chromatic changes and color iden-
position but different luminancesthe luminance tity, respectively (e.g., Dobkins & Anderson, 2002;
threshold). The chromatic deficiency hypothesis pre- Dobkins, Anderson, & Kelly, 2001; Dobkins, Lia, &
dicts that the ratio of luminance threshold divided by Teller, 1997; Teller, 1998; Thomasson & Teller, 2000).
chromatic threshold will decrease with increasing age.
That is, luminance and chromatic thresholds may both
Motion Perception
improve with age, but chromatic thresholds change
more. The visual efficiency or uniform loss hypothesis Moving and perceiving are deeply linked. Many of the
predicts that the ratio of luminance threshold divided by most significant features of an environment to be per-
chromatic threshold is constant with age. That is, lumi- ceived are moving objects and the events in which they
nance and chromatic thresholds decrease at the same participate. Motion of the observer is also crucial, in
rates with increasing age because they are both limited two ways. To locomote safely through space requires
by a common factor such as overall visual sensitivity. that our visual system be structured to deal with contin-
Banks and Bennett (1988) and Banks and Shannon uously changing views of the environment. Moreover, in-
(1993) showed that this hypothesis can in fact account formation given by transforming views of the world turn
for the poor Rayleigh and neutral-point discriminations out to be a rich indicator not only of events but of per-
of neonates. sisting properties of the world, such as spatial layout
Other investigators have tested the chromatic defi- (J. Gibson, 1966, 1979; Johansson, 1970). Later, in dis-
ciency and visual efficiency hypotheses empirically, cussing space perception, we consider ways in which
but no clear consensus has yet emerged. The challenge motions of objects and observers offer high-fidelity in-
has been to develop paradigms in which infants sensi- formation about spatial layout and object form.
tivity can be made high enough to distinguish the pre- Early research on infant visual motion perception
dictions of the two hypotheses. In particular, recent showed that motion strongly attracts infant attention
work has focused on determining which hypothesis pro- (Fantz & Nevis, 1967; Haith, 1983; Kremenitzer,
vides a better account of young infants ability to use M Vaughan, Kurtzberg, & Dowling, 1979; White, Castle,
and L cones to make Rayleigh discriminations (e.g., & Held, 1964). Progress has been made in analyzing the
Adams & Courage, 2002; D. Allen et al., 1993, 1988; limits and probable mechanisms of motion sensitivity,
Clavadetscher et al., 1988; Morrone, Burr, & Fiorentini, including directional sensitivity, velocity sensitivity,
1993; Teller & Lindsey, 1993; Teller & Palmer, 1996; and perception of motion and stability.
Varner et al., 1985). On balance, the discrimination
failures observed with the youngest children and, for Directional Selectivity
small targets, with older children do not necessarily The ability to detect motion direction is one of the most
imply deficiencies among chromatic mechanisms per basic and important perceptual capacities, but its devel-
se. Rather the ratio of chromatic divided by luminance opment has been poorly understood until the last decade
sensitivity may well remain constant across age, sug- or so. Using both behavioral and visual evoked potential
gesting that neonates apparent inability to make (VEP) measures, Wattam-Bell (1991, 1992) tested di-
Rayleigh and neutral-point discriminations is caused by rectional sensitivity in longitudinal studies. In the VEP
an overall loss in visual efficiency. The predictions of studies, it was expected that if infants detected direc-
the visual efficiency hypothesis, however, are inconsis- tion reversals in an oscillating checkerboard pattern, a
tent with the tritan discriminations. Therefore, young measurable electrical response should be found at the
infants may in fact possess some form of color anomaly frequency of the stimulus reversals. Reliable VEPs were
involving a deficiency among S cones. first found at a median age of 74 days for 5 degrees/sec-
Future work will be needed to illuminate infants loss ond patterns and 90 days for 20 degree/second patterns.
of visual efficiency and/or deficiency in S cones. Re- Behavioral studies (Wattam-Bell, 1992) employed a dif-
searchers are also taking an interest in the difference ferent type of display. In one condition, an array of
between processing moving versus static chromatic randomly changing dots was shown in which appeared a
Basic Visual Sensitivities in Infancy 125

vertical strip of coherently (vertically) moving dots. gratings, Aslin and Shea (1990) found velocity thresh-
In another condition, the vertical motion was shown olds of about 9 degrees/second at 6 weeks dropping to 4
against a background having opposite direction motion. degrees/second at 12 weeks. Thresholds for detecting a
A visual preference paradigm was used in which the tar- difference between two velocities were studied by Dan-
get display appeared adjacent to a control display having nemiller and Freedland (1991) using paired displays
random or uniform motion. If an infant detected the ver- with horizontal bars oscillating at different rates; their
tical target strip having unique, coherent motion, the in- 20-month-old subjects distinguished bars moving at 3.3
fant was expected to look longer at this display. The degrees/second from 2.0 degrees/second, but not from
element displacement per frame was manipulated to find 2.5 degrees/second.
the greatest displacement that supported motion detec- Much lower thresholds for motion detection were ob-
tion (dmax). This measure was found to increase tained by von Hofsten, Kellman, and Putaansuu (1992).
markedly from 8 to15 weeks of age. The younger infants In habituation studies of observer-contingent motion
(8 to 11 weeks) could tolerate only about a .25 degree of with 14-week-olds, von Hofsten et al. found sensitivity
visual angle displacement (frame duration was 20 mil- to a differential velocity of .32 degrees/second but not
lisecond), whereas 14 to 15-week-olds showed a dmax of .16 degrees/second. Infants were also found to be sensi-
about .65. (The value for adults is about 2 degrees in tive to the relation of the motion direction to their own
this task.) motion. Higher sensitivity in this paradigm might have
Poor performance in the earliest weeks may be due to two explanations. It is possible that visual preference
a lack of motion detectors sensitive to high velocities, paradigms understate infant capacities. As is true in
that is, large displacements in short time intervals. This general with preference measures, infants might detect a
interpretation is supported by additional data that difference (e.g., between moving and stationary pat-
showed an increase in dmax when the temporal interval terns) but have no differential interest or attention to the
between frames was lengthened (Wattam-Bell, 1992). two displays. A second possibility is that the key differ-
ence relates to observer motion contingency in the von
Velocity Sensitivity
Hofsten et al. study. It is plausible that small, observer-
Human adults perceive motion over a great range of ve- contingent motions are processed by the motion per-
locities. Under optimal conditions, a motion as slow as 1 spective system as specifiers of object depth, rather
to 2 minutes of visual angle per second may be detected than as moving objects. Thus, a depth-from-motion sys-
as motion, as may faster motions up to 15 to 30 de- tem may have greater sensitivity than a motion detec-
grees/second, at which blurring or streaking occurs tion system, and the former might be engaged only by
(Kaufman, 1974). Estimates of the slowest velocity to observer movement (von Hofsten et al., 1992).
which infants respond have varied. Volkmann and Dob-
son (1976) used checkerboard patterns (check size = 5.5 Mechanisms for Processing Moving Patterns:
degrees) and found a moving display was clearly pre- Velocity, Flicker, and Position
ferred to a stationary one by 2- and 3-month-olds for a A moving stimulus may be characterized in different
velocity as slow as 2 degrees/second. One-month-olds ways. Similarly, a response to a moving stimulus may be
showed a weaker preference. Using rotary motion dis- based on more than one kind of mechanism. Consider a
plays, Kaufmann, Stucki, and Kaufmann-Hayoz (1985) vertical sine-wave grating drifting horizontally. Each
estimated thresholds at about 1.4 degrees/second at 1 edge moves at a certain velocity. At a given point, alter-
month and 0.93 degrees/second at 3 months, also using a nating dark and light areas will pass at a certain rate,
visual preference technique. presenting a temporal frequency of modulation or
Later studies designed to distinguish various possible flicker rate. This flicker rate depends both on the veloc-
mechanisms by which moving patterns might be de- ity of the pattern and on its spatial frequency (cycles
tected have yielded higher threshold estimates. Danne- per degree). Now consider preferential attention to such
miller and Freedland (1989), using unidirectional linear a stimulus over a nonmoving grating or a blank field.
motion of a single bar, found no reliable motion prefer- The preference could be based on a direction-sensitive
ences at 8 weeks. They estimated thresholds at about mechanism, a velocity-sensitive mechanism, or a
5 degrees/second for 16-week-olds and about 2.3 de- flicker-sensitive mechanism. Sustained flicker could be
grees/second for 20-week-olds. For vertically moving avoided by use of a single object in motion as opposed to
126 Infant Visual Perception

a repetitive pattern, but then the possibility arises that pected from self-produced movements (von Holst, 1954;
the motion could be detected by noting the change in po- Wallach, 1987) or more global relationships among opti-
sition of some unique object feature, that is, a position- cal changes occurring at a given time (Duncker, 1929; J.
sensitive mechanism may operate. Some research on Gibson, 1966).
motion sensitivity has aimed to separate these possibili- In the case of passive (non-self-produced) observer
ties experimentally. motion, relations in optic flow or some contribution
Perhaps the first effort to disentangle velocity- from the vestibular system must be used in perceiving a
sensitive, position-sensitive, and flicker-sensitive stable world. There is some indication that young infants
mechanisms was carried out by Freedland and Danne- show position constancy under such conditions. Later,
miller (1987). Several combinations of temporal fre- we mention work in object perception (Kellman, Gleit-
quency and spatial displacement were presented with man, & Spelke, 1987) suggesting that moving infants
random black and white checkerboard displays. Infants discriminate moving from stationary objects and per-
preferences were affected by both of these factors and ceive object unity only from real object motion. More
were not a simple function of velocity. The role of direct studies of position constancy and motion percep-
flicker could not be directly assessed in these experi- tion by moving observers have also been carried out
ments. Sensitivity to flicker versus velocity was exam- (Kellman & von Hofsten, 1992). In these studies, in-
ined by Aslin and Shea (1990) with vertically moving, fants were moved laterally while viewing an array of ob-
square-wave gratings. Various combinations of spatial jects. On each trial, one object in the array, either on the
frequency and velocity were used to vary flicker inde- left or right, moved while others remained stationary.
pendent of velocity. For example, the flicker rate (tem- The object motion was parallel to the observers motion.
poral frequency) at any point in the display remains Whether the optical change given to the observer in this
constant if spatial frequency is doubled and velocity is situation comes from a moving or stationary object de-
cut in half. Aslin and Shea (1990) found that velocity, pends on the objects distance. Thus, a stationary object
not flicker, determines preferences in infants 6 and 12 placed on the opposite side of the array at a different
weeks of age. Converging evidence for velocity-sensi- distance matched the optical displacement of the mov-
tive mechanisms was reported by Dannemiller and ing object. Infants were expected to look more at the
Freedland (1991). By using a display with motion of a moving object if its motion was detected. Both 8- and
single bar flanked by stationary reference bars, they ex- 16-week-olds showed this pattern when the object and
cluded ongoing flicker in any spatial position. More- observer motions were opposite in phase, but only 16-
over, manipulating extent of displacement allowed them week-olds appeared to detect the motion when object
to test the possibility that infants responses were deter- and observer moved in phase (Kellman & von Hofsten,
mined by the extent of positional displacement. Results 1992). It is not clear why the younger infants showed de-
were consistent with velocity-sensitive mechanisms. tection of the moving object only in the opposite phase
condition. Further study indicated that motion detection
Perceiving Motion and Stability
was eliminated in monocular viewing. It appears that
Perceiving moving objects is inextricably tied to its con- some ability to distinguish moving and stationary ob-
verse: perceiving nonmoving objects and surfaces as sta- jects during observer motion is in place as early as 8
tionary. The latter ability is less straightforward than it weeks of age and that binocular convergence may pro-
might at first appear. Neural models of motion detectors vide the distance information needed in this task (Kell-
suggest that these should respond to image features, such man & von Hofsten, 1992).
as edges, that change position on the retina over time. Yet
such retinal displacement occurs in perfectly stationary
environments whenever perceivers make eye, head, or
body movements. Perception of objects remaining at rest SPACE PERCEPTION
during observer motion, called position constancy, re-
quires use of information beyond that available to indi- In considering how we obtain knowledge through percep-
vidual motion-sensing units. Such information might tion, the philosopher Kant (1781/1902) concluded that
involve comparison of retinal changes with those ex- the mind must contain built-in (a priori) categories of
Space Perception 127

space and time into which experience is organized. Psy- evolved mechanisms to extract such information. On
chologically, understanding the origins and development this ecological view of development (E. Gibson, 1979;
of spatial perception has more nuances. Whether we ap- Shepard, 1984), the rudiments of 3D perception might
proach perception from the perspective of the philoso- be present even in the newborn, and their refinement
pher, cognitive scientist, psychologist, or engineer, might depend on sensory maturation and attentional
however, we will rediscover Kants insight that space is skill, rather than on associative learning.
fundamental. Our earlier treatment of basic spatial vi- Research on spatial perception has gone a consider-
sion set out the sensory limitationsin acuity, contrast able distance toward answering this question of the
sensitivity, and sensitivity to pattern variationthat constructivist versus ecological origins of the third
constrain the pickup of information. As we explore space dimension. Moreover, the emerging picture of early abil-
perception here, our main concern is the acquisition of ities provides important insights about functionally dis-
knowledge of positions and arrangements of objects and tinct classes of information and their neurophysiological
surfaces in the three-dimensional environment. underpinnings. Anticipating some of these distinctions,
Theoretical controversy about the development of vi- we divide spatial perception abilities into four cate-
sual space perception has centered on depth perception. gories: kinematic, oculomotor, stereoscopic, and picto-
When we examine the human visual apparatus, it is rela- rial. The classification reflects both differences in the
tively easy to see how we acquire information about two nature of information and in the perceptual mechanisms
of three spatial dimensions. The optics of the eye en- at work in extracting information (Kellman, 1995;
sure, to a high degree, that light originating from points Kellman & Arterberry, 1998; Yonas & Owsley, 1987).
in different directions from the observer will be mapped
onto distinct points on the retina. The result is a map Kinematic Information
that preserves information about adjacency in two spa-
tial dimensions (up-down and left-right). The apparent For guiding action and furnishing information about the
problem lies in the third (depth) dimension. Nothing in 3D environment, kinematic or motion-carried informa-
this map immediately indicates how far a ray of light has tion may be the most important class of visual informa-
traveled to get from an object to the eye. tion for adult humans. One reason for its centrality is
Traditionally, it has most often been claimed that it overcomes the ambiguity problems present with
that perception of three-dimensional (3D) space is a some other kinds of information, such as pictorial cues to
product of learning (Berkeley, 1709/1963; Helmholtz, depth. A stationary image given to one eye may be a cud-
1885/1925). Before the invention of methods to study dly kitten or a gigantic tiger further off, as Berkeley
infants perception, the basis for this view was the log- noted, or even a flat, 2D cutout of a cat or tiger. To the
ical problem of recovering three dimensions from a moving observer, the transforming optic array reveals
projection of the world onto a surface of two dimen- whether the object is planar or 3D and furnishes infor-
sions (the retina). Learning might overcome the limita- mation about relative distance and size. The mapping be-
tion through the associating and storing of sensations tween the optical transformations and the 3D scene is
of vision and touch, allowing relevant information governed by projective geometry, and under reasonable
about tactile correlates of visual sensations; these in constraints, it allows recovery of many properties of the
turn could be retrieved when familiar visual input re- layout (Koenderink, 1986; Lee, 1974; Ullman, 1979).
curred (Berkeley, 1709/1963; Helmholtz, 1885/1925; Among the residual ambiguities is a problem analogous
Titchener, 1910). Piaget went a step further in arguing to the one Berkeley raised about a single image. If ob-
that self-initiated action and its consequences provide jects and surfaces in the scene deform (alter their
the necessary learning. shapes) contingent on the observers motion, a unique
Modern analyses of the information available for vi- 3D scene is not recoverable. Now the problem is recover-
sion have raised a radically different possibility for the ing four dimensions (spatial layout plus change over
origins of spatial perception. Transforming optical input time) from three (two spatial dimensions of the input
given to a moving organism carries information specific plus time). In ordinary perception, simulation of the
to the particular 3D layout (J. Gibson, 1966, 1979; Jo- exact projective changes consistent with a particular, but
hansson, 1970), and humans and animals may well have not present, layout, would almost never occur by chance.
128 Infant Visual Perception

It does, however, make possible the realistic depiction of has been shown to be used in adult visual perception, to
3D space in television, motion pictures, and in virtual re- establish both depth order and shape, even when no
ality setups. Because kinematic information about space other sources of information are available (Andersen &
depends on geometry, not on knowledge of what particu- Cortese, 1989; Kaplan, 1969; Shipley & Kellman, 1994).
lar spatial layouts exist in the world, it is imaginable that Infants shape perception from accretion /deletion of
perceptual mechanisms have evolved to make use of it. texture was studied by Kaufmann-Hayoz, Kaufman, and
An additional reason to suspect that sensitivity to this Stucki (1986). They habituated 3-month-olds to one
kind of information might appear early is that early shape specified by accretion /deletion and tested recov-
learning about the environment may be optimized by re- ery from habituation to the same and a novel shape. In-
lying on sources of information that are most accurate fants dishabituated more to the novel shape. Although
(Kellman, 1993; Kellman & Arterberry, 1998). On the this result suggests that accretion /deletion specifies
other hand, adults acquire much kinematic information edges and shape at 3 months, we cannot tell much
from their own movements through the environment. The about perceived depth order from this study. That accre-
human infant does not self-locomote until the second tion /deletion specifies depth order at 5 to 7 months is
half-year of life although kinematic information could suggested by a different study (Granrud, Yonas, et al.,
still be made available from moving objects, from the in- 1985). These investigators assumed that infants would
fant being carried through the environment, or from self- reach preferentially to a surface perceived as nearer
produced head movements. than another. Computer generated, random dot, kine-
Motion-carried or kinematic information is often matic displays were shown in which a vertical boundary
divided into subcategories, of which we consider three. was specified by only accretion /deletion information.
Relative depths of surfaces can be specified by Infants of 5 and 7 months of age were tested, and both
accretion/deletion of texture. Relative motion between groups showed modestly greater reaching to areas spec-
an object and observer may be given by optical ified as nearer by accretion /deletion than to areas
expansion/contraction. Relative depth, and under some specified as farther. More recently Johnson and Mason
conditions perhaps metric information about distance, (2002) provided evidence that 2-month-olds are able to
can be provided by motion parallax or motion perspec- use accretion /deletion of texture for perceiving depth
tive. Another important kinematically based spatial relations.
ability, recovery of object shape from transforming op- Craton and Yonas (1990) suggested that ordinary ac-
tical projections (structure-from-motion), is discussed cretion /deletion displays actually contain two kinds of
in connection with object perception. information. In addition to the disappearance and ap-
pearance of texture elements, there are relationships of
Accretion/Deletion of Texture
individual elements to the location of the boundary be-
In the late 1960s, Kaplan, Gibson, and their colleagues tween surfaces. A visible element on one side of a
discovered a new kind of depth information, a striking boundary remains in a fixed relation to it, whereas an
achievement given that depth perception had at that element on the other side (the more distant surface)
point been systematically studied for over 200 years changes its separation from the boundary over time.
(J. Gibson, Kaplan, Reynolds, & Wheeler, 1969; Kaplan, This separate information, termed boundary f low, ap-
1969). Most surfaces have visible texturevariations of pears to be usable by adults in the absence of element
luminance and color across their surfaces. The new type accretion /deletion (Craton & Yonas, 1990) and possibly
of depth information involves what happens to visible by 5-month-old infants (Craton & Yonas, 1988).
points of texture (texture elements) when an observer or
Optical Expansion/Contraction
object moves. When the observer moves while viewing a
nearer and more distant object, the elements on the When an object approaches an observer on a collision
nearer surface remain visible whereas those on the more course, its optical projection expands symmetrically. It
distant surface gradually pass out of sight along one side can be shown mathematically that a ratio of an object
(deletion) of the nearer object and come into view along points retinal eccentricity and its retinal velocity gives
the other side (accretion). The same kind of transforma- its time to contact, that is, the time until it will hit the
tion occurs when the motion is given by a moving object observer. Newborns of other species show defensive re-
instead of a moving observer. This kind of information sponses to this kind of information (Schiff, 1965).
Space Perception 129

When presented with optical expansion patterns, functions quite early, but these suggestions have been
human infants of 1 to 2 months of age were reported to based on indirect evidence (Walk & Gibson, 1961;
retract their heads, raise their arms, and blink (Ball & Yonas & Owsley, 1987). Walk and Gibson (1961) stud-
Tronick, 1971; Bower, Broughton, & Moore, 1970). Not ied newborns of various species on the visual cliff and
all of these responses, however, may indicate perception noted that some species made lateral head movements
of an approaching object (Yonas et al., 1977). Head before choosing the shallow side of the cliff over the
movement may result from infants tracking visually the deep side. It is difficult to make a similar inference
top contour of the pattern, and relatively undifferenti- about human infants, because they do not self-locomote
ated motor behavior may cause the arms to rise in con- until around 6 months of age.
cert. Yonas et al. tested this hypothesis using a display in Some results relevant to the development of motion
which only the top contour moved. This optical change perspective in 4-month-old infants were reported by
is not consistent with approach of an object. Infants from von Hofsten et al. (1992). Infants moved back and forth
1 to 4 months displayed similar head and arm move- while viewing an array of three vertical bars. The mid-
ments to this new display as to an optical expansion dis- dle bar was moved in concert with the infants chair,
play. The result supports the hypothesis that tracking the giving it an optical displacement that would have been
top contour, rather than defensive responding, accounts consistent with a stationary rod placed somewhat fur-
for the behavior infants show to expansion displays. ther away. If motion perspective operates, the observer
It turns out, however, that both the tracking hypothe- contingent motion should indicate that the middle rod is
sis and the hypothesis of defensive responding appear to furthest from the subject (see Figure 3.4). After habitu-
be correct. When eye blink was used as the dependent ation to such an array, moving infants looked more at a
measure, reliably more responding was observed to the stationary array consisting of three aligned, stationary
approach display than to the moving top contour display. rods than to another stationary array with the middle
It appears that blinking may best access infant percep- rod 15 cm further away than the others. (The latter dis-
tion of object approach and does so reliably from about play produced identical motion perspective as the habit-
1 month of age (Nanez, 1988; Nanez & Yonas, 1994; uation display.) Two other experiments showed that the
Yonas, 1981; Yonas, Pettersen, & Lockman, 1979). effect disappeared if the contingent motion was reduced
from the original .32 degrees/second to .16 degrees/
Motion Perspective
second and that infants were sensitive to the contin-
Motion perspective is an important source of spatial lay- gency between the optical changes and their own move-
out information. When an observer moves and looks ment. These results are consistent with infants early
perpendicular to the movement direction, the visual di- use of motion perspective. They might also be ex-
rection of a nearer object changes at a faster velocity plained, however, by infants responding to particular
than that of a more distant object. Comparing two such optical changes and the contingency of these optical
objects or points defines the classical depth cue of mo- changes on the observers movement. The results do not
tion parallax. J. Gibson (1950, 1966) argued that per- include any test to verify that the optical changes were
ceptual systems might use relative velocities of many taken to indicate depth. An interesting possibility is
points, that is, gradients of relative motion provide more that the perceptual process that uses motion perspec-
information than a pair of points. To express this con- tive to assign depth is far more sensitive to optical dis-
cept, he coined the term motion perspective. Some ex- placement than processes used to see moving objects.
perimental evidence indicates that gradients are in fact
used by human perceivers (e.g., E. Gibson, Gibson,
Smith, & Flock, 1959). Stereoscopic Depth Perception
Motion perspective is virtually always available to a
moving observer in a lighted environment, and it ordi- Stereoscopic depth perception refers to the use of dif-
narily provides unambiguous indication of depth order. ferences in the optical projections at the two retinas to
Given these considerations, one might expect that neural determine depth. This ability is among the most precise
mechanisms have evolved to exploit this kind of infor- in adult visual perception. Under optimal conditions,
mation, and that accordingly, it might appear early in an adult observer may detect depth when the angular
development. Several investigators have suggested that it difference in a viewed points location at the two eyes
130 Infant Visual Perception

more distant than the fixation point will have uncrossed


disparity. The visual direction of such a point will be
more to the left in the visual field of the left eye than in
the right eye. Crossed disparity characterizes points
nearer than the fixated point. The visual direction of
these points will be more leftward in the right eye than
in the left.
Observations from other species suggest the existence
of innate brain mechanisms subserving stereoscopic
depth perception, specifically, cortical cells tuned to
particular disparities at birth or soon after (Hubel &
Wiesel, 1970; Pettigrew, 1974; Ramachandran, Clarke,
& Whitteridge, 1977). Such single-cell recording studies
are not possible in human infants; moreover, they do not
directly address functional operation of stereoscopic
depth perception. Evidence about human infants comes
mostly from behavioral studies and suggests that stereo-
scopic depth perception arises around 4 months of age as
a result of maturational processes.
A number of studies have used stationary displays
and preferential looking as the dependent variable. One
of two adjacently presented displays contains binocular
disparities that might specify depth differences within
the pattern. Infants are expected to look longer at a dis-
Figure 3.4 Top views of displays used in motion parallax ex-
play containing detectable depth differences than at a
periment. Top: Moving observers were habituated to a linear
array of rods in which the center rod moved in phase with the similar one having no depth variation (Atkinson &
observer. The dotted line indicates the virtual object specified Braddick, 1976; Held, Birch, & Gwiazda, 1980). A dif-
by motion parallax. Bottom: The two test arrays pictured were ferent method eliminates any possible monocular cues.
shown after habituation. Source: Young Infants Sensitivity Using random dot kinematograms, Fox, Aslin, Shea, and
to Motion Parallax, by C. von Hofsten, P. Kellman, and Dumais (1980) presented disparity information that
J. Putaansuu, 1992, Infant Behavior and Development, 15(2),
would, if detected, specify a moving square. Using the
pp. 245264. Reprinted with permission.
forced-choice preferential looking method, adult ob-
servers judged the direction of motion on each trial
solely by watching the infants responses.
( binocular disparity) is only 5 to 15 seconds of arc Estimates of the age of onset of disparity sensitivity
(Westheimer & McKee, 1980). A 5-second disparity from these methods show reasonable agreement. In lon-
would translate into detection of a 1.4 mm depth differ- gitudinal studies by Held and his colleagues (Birch,
ence between two objects at a distance of one meter. We Gwiazda, & Held, 1982; Held et al., 1980), reliable pref-
can distinguish two types of binocular disparity, crossed erences for a vertical grating pattern with disparity vari-
and uncrossed. A prerequisite for precise computation of ation appeared at 12 weeks for crossed disparities and 17
disparity between the two eyes is fixation by the two weeks for uncrossed. Fox et al. (1980) found that 3- to 5-
eyes on a common environmental point. We can measure month-olds reliably oriented to a moving square speci-
the disparities of other imaged points by comparison fied by disparity, but infants younger than 3 months did
to this zero disparity fixation point. Other points at not. Petrig, Julesz, Kropfl, and Baumgartner (1981)
roughly the same distance from the observer as the fix- found a similar onset of sensitivity using recordings of
ated point will project to corresponding retinal loca- visual evoked potentials.
tions, that is, having the same angular separation and A thorny issue in the interpretation of these studies is
direction from the fovea on each of the two eyes. Points whether the observed behavioral responses index depth
Space Perception 131

perception from binocular disparity or merely sensitivity to approximate those present at 2 months of age. Under
to disparity itself. It is hard to settle this issue with cer- these conditions, stereoacuity was reduced substan-
tainty; however, some observations suggest that depth is tially, but not sufficiently to explain infants inability to
perceived. Held et al. (1980), for example, found that in- respond to large disparities before 3 to 4 months of age.
fants who showed clear preferences for vertical line dis- Developmental changes in convergence also appear un-
plays containing horizontal disparity showed no such likely to explain the onset of stereoacuity. Evidence
preferences when the displays were rotated 90 degrees to on the development of convergence (Hainline, Riddell,
give 34 minutes of vertical disparity (a condition that Grose-Fifer & Abramov, 1992) indicates that it may be
produces rivalry for adults). Fox et al. (1980) observed nearly adultlike at 1 to 2 months of age. Also, conver-
that infants did not track a moving object specified gence changes would not explain differences in the onset
by very large disparities that do not signal depth to of crossed and uncrossed disparity (Held et al., 1980).
adults. They found instead that infants reliably looked Given these considerations, most investigators believe
away from such displays. This result is double-edged: the explanation for the onset of stereoscopic vision is
Although it shows different reactions by infants to dif- some maturational change in cortical disparity-sensitive
ferent magnitudes of disparity as might be expected if units. Such a mechanism underlies improvement of
only some disparities produce perceived depth, it also stereoscopic discrimination performance in kittens
shows that disparities per se can affect infants fixation. (Pettigrew, 1974; Timney, 1981). In humans, it has been
From these studies, it is plausible but not certain that suggested that the particular change in disparity-
infants responses in these studies indicate functional sensitive cells may be segregation of ocular dominance
stereoscopic depth perception. Other studies have shown columns in layer 4 of the visual cortex (Held, 1985,
that disparity-sensitive infants outperform disparity- 1988). At birth, cells in layer 4 generally receive projec-
insensitive infants on tasks involving depth and three- tions from both eyes. Between birth and 6 months, inputs
dimensional shape perception (Granrud, 1986; Yonas, from the two eyes separate into alternating columns
Arterberry, & Granrud, 1987a). receiving input from the right and left eyes (Hickey &
What mechanisms are responsible for the onset of Peduzzi, 1987). Eye-of-origin information is needed to
stereoscopic sensitivity after several months of life? An extract disparity information, so this neurological devel-
argument for maturational causes is that sensitivity very opment is a plausible candidate for the onset of stereo-
quickly attains adultlike precision. Held et al. (1980) re- scopic function.
ported that thresholds change over 3 to 4 weeks from
greater than 60 minutes to less than 1 minute of dispar- Pictorial Depth Perception
ity, with the latter measured value limited by the appara-
tus; even so, this value is comparable to adult sensitivity The pictorial cues are so named because they allow
under some conditions. depth to be portrayed in a flat, two-dimensional picture.
What mechanisms might be maturing at this time? Sometimes these are called the classical depth cues, be-
One possibility is that disparity-sensitive cortical cells cause they have been discussed and used by artists and
are coming online. Another is that improvements in the students of perception for centuries. Theoretically, they
mechanisms of convergence or visual acuity that are have been central to classical arguments about the need
prerequisites to fine stereopsis might explain the ob- for learning in spatial perception. The fact that the same
served onset of disparity sensitivity. Some evidence information can be displayed in a flat picture or a real
suggests that the onset of stereopsis is not dependent on 3D scene immediately points to their ambiguity as signi-
improvements in visual acuity (grating acuity). When fiers of reality. It is a short step to the classical perspec-
both acuity and disparity sensitivity are measured lon- tive on the acquisition of such cues: If these cues are not
gitudinally in the same infants, little or no change in unequivocally tied to particular spatial arrangements,
grating acuity is found during the period in which stere- our perception of depth from these cues must derive
opsis appears (Held, 1993). A different method pointing from learning about what tends to be the case in our par-
toward the same conclusion comes from a study by Wes- ticular environment. (The environment, until recently,
theimer and McKee (1980). Adults were given artifi- had many more 3D scenes offering information than 2D
cially reduced acuity and contrast sensitivity designed representations.)
132 Infant Visual Perception

Ecologically, the pictorial cues to depth are diverse, leagues (see Yonas, Arterberry, & Granrud, 1987b;
but many of them rest on similar foundations. The laws Yonas & Owsley, 1987 for reviews). For brevity, we con-
of projection ensure that a given physical magnitude sider only two examples: interposition and familiar size.
projects an image of decreasing extent at the retina with The development of other pictorial cues that have been
increasing distance from the observer. Applying this studied, such as linear perspective and shading, appears
geometry in reverse, if two physical extents are known to be similar.
or assumed to have the same physical (real) size, then
differences in their projected size can be used to estab- Interposition
lish their depth order. This information comprises the The depth cue of interposition, sometimes called over-
depth cue of relative size. Very similar is linear perspec- lap, specifies relative depth of surfaces based on con-
tive. If two lines in the world are known or assumed to be tour junction information. When surface edges form a
parallel, then their convergence in the optical projection T junction in the optical projection, the edge that
may be taken to indicate their extending away from the comes to an end at the intersection point (the vertical
observer in depth. Generalizing this notion to whole edge in the letter T; see Figure 3.5A) belongs to a sur-
fields of visible elements comprises the rich source of face passing behind the surface bounded by the other
information in natural scenes known as texture gradi- edge (the horizontal edge in the letter T). Interposition
ents (J. Gibson, 1950). If a surface is assumed to be is a powerful depth cue in human vision (Kellman &
made up of physically uniform or stochastically regular Shipley, 1991). Infant use of interposition information
tokens (pebbles, plants, floor tiles, etc.), then the de- was tested by Granrud and Yonas (1984). They used
creasing projective size of texture elements indicates in- three similar displays made of three parts each but dif-
creasing depth. A different kind of assumed equality is fering in the presence of interposition information. In
illustrated by the