You are on page 1of 11


Brian Knutson and Gregory R. Samanez-Larkin

OVERVIEW- region associated with anticipating losses (i.e.,

the anterior insula) precedes an increased
In this chapter, we summarize recent findings in tendency to avoid financial losses. By extension,
neuroeconomics suggesting that emotion individual differences in increased gain
(specifically, "anticipatory affect") can influence anticipation, decreased loss anticipation, or some
financial decisions, and then discuss how combination of the two (plus a third non-
individual differences in anticipatory affect may reflective factor) might promote proneness to
promote proneness to consumer debt. Thanks to debt. Ultimately, neuroeconomic advances may
improvements in spatial and temporal resolution, help individuals to optimize their investment
functional magnetic resonance imaging (FMRI) strategies, as well as empower institutions to
experiments have begun to suggest that minimize consumer debt.
activation of a brain region associated with Neuroeconomists seek to explain how brains
anticipating gains (i.e., the nucleus accumbens or choose. Thanks to technological advances,
NAcc) precedes an increased tendency to seek scientists can now “open the black box” of the
financial gains, while activation of another brain, moving below the surface mapping
between input and output to identification of
mediating neural and psychological processes.
Much of the research described in this chapter was Thus, neuroeconomic methods might allow
supported by funding from the National Institute on scientists to bridge gaps between neural,
Aging (R21-AG030778, P30-AG017253, P30-
psychological, and behavioral levels of analysis.
AG024957, F31-AG032804) and the FINRA Investor
Education Foundation. The views expressed here are
Below, we summarize ongoing attempts to forge
those of the authors and not of our funding agencies. links from affective neural circuits to affective
We thank Alan M. Garber and Camelia M. Kuhnen experience and, eventually, to decisions that can
for their ongoing contributions to our thinking about lead to debt.
the psychology of debt. Portions of this chapter were
adapted from Knutson and Greer (2008).

DEFINING DEBT AND POTENTIAL individual differences in cognitive and emotional

CAUSES function, while primarily economists focus on
life financial outcomes like debt (for exceptions,
To study whether individual difference variables see Lea, Mewse, and Wrapson, Chapter 7).
influence life financial outcomes, one must first Indebtedness implies one or more earlier
measure financial outcomes. Based on standard decisions to take on debt. From a psychological
accounting practices, life financial outcomes standpoint, the decision to take on debt involves
might broadly be divided into the two classes: choosing present gain at the cost of a greater
assets (related to savings) and debt (related to future loss. The decision to take on debt thus
outstanding expenditures). Although assets and involves two classes of decisions that have
debt undoubtedly fluctuate over time in response proven most difficult to explain with rational
to significant life events and the general choice models. First, taking on debt involves
economic climate, they may also show some weighing potential gains versus losses and, thus,
temporal stability both within and across may be related to risk preference. Second, taking
individuals. We focus on debt below, on debt also involves weighing potential present
operationally defined as money owed to any gains versus future losses, and so may be related
lender over an extended period of time— to time preference. In cases of both risk
although finer-grained analyses might preference and time preference, theorists have
distinguish home ownership (i.e., mortgage) debt sought to account for anomalies in choice (e.g.,
from other types of debts (e.g., revolving credit inconsistency) by invoking emotional
card debt). Even given such a rough index, mechanisms (Ainslie 1992; Loewenstein et al.
continuous measures of debt might allow 2001). If emotion influences immediate choices,
investigators to determine whether and which and does so repeatedly and consistently over
individual difference factors promote debt. time, it might have a significant cumulative
Eventually, measures of debt should ideally impact on life financial outcomes like debt.
demonstrate both test-retest reliability (e.g.,
similarity across instances of measurement) and
validity (e.g., self-report should agree with credit THE ANTICIPATORY AFFECT MODEL
Research has repeatedly linked both While evolutionary theorists have accorded
situational and personal factors to debt (Lea, emotion a central role in ancestral choices related
Webley, and Levine 1993; Stone and Maury to survival and procreation (e.g., approaching
2006). Economically, young people, people with sexual opportunities or avoiding predatory
lower incomes, or people who have suffered threats) (Darwin 1872), the importance of
recent financial hardship are more likely to be in emotion in choices related to abstract incentives
debt. Psychologically, more permissive attitudes (e.g., money) is less clear. In fact, a rational actor
towards debt and perceived control over finances might well rely solely upon abstract numerical
have been linked to debt in some, but not all, representations to make optimal financial
studies. Beyond these factors, over a century of choices. Accumulating brain imaging research,
research suggests that individuals reliably differ however, suggests that even complex financial
in terms of intelligence and socio-emotional choices recruit evolutionarily-preserved neural
capacities, and that these traits have a substantial circuits implicated in emotion (Knutson and
heritable component (> 50% of variation across Greer 2008).
individuals) (McGue & Bouchard 1998). Few Although popular sentiment implies that
studies, however, have examined the direct emotions can influence choice, physiological
influence of these factors on debt proneness. evidence for such an influence has remained
This gap in the literature may partially result elusive. Part of the difficulty in studying the
from the fact that primarily psychologists study

Figure 6–1. Anticipatory affect model (adapted from Knutson & Greer 2008).

influence of emotions on decisions has to do incentive cues resolves when uncertainty

with emotion’s dynamic and transient nature. collapses as the outcome either occurs or does
Researchers have traditionally focused on not. During anticipation, uncertainty increases
affective reactions to events only after they arousal, while potential gains increase valence
occur—“consequential” affect (Loewenstein et and potential losses decrease valence. Thus,
al. 2001). For instance, researchers might anticipation of uncertain gains should increase
measure the affect elicited by unexpected positive arousal (e.g., feelings such as
positive versus negative events, or by success excitement), while anticipation of uncertain
versus failure in achieving goals (Carver and losses should increase negative arousal (e.g.,
White 1994; Isen, Nygren, and Ashby 1988). feelings such as anxiety), which are
While more recent affective forecasting models psychometrically defined as independent rather
have focused on predicted affective responses to than opposite affective states (Watson and
events (Wilson and Gilbert 2003), these affective Tellegen 1985).
forecasting models are still “consequentialist,” In addition to generating affective
since they refer to peoples’ predictions about experience, positive arousal should promote
their affective responses to outcomes rather than approach behavior, while negative arousal
how people will feel during anticipation of those should promote avoidance behavior. The
outcomes. Affect that occurs during anticipation anticipatory affect model can thus forecast the
(“anticipatory” affect), however, is best situated effect of incentive cues on risk taking.
in time to influence upcoming decisions. Specifically, positive incentive cues should elicit
Accordingly, we have proposed an nucleus accumbens (NAcc) activation and
“anticipatory affect model” in which anticipation positive arousal, which should facilitate risk
of significant outcomes alters both affective taking, while negative incentive cues should
arousal and valence (Knutson and Greer 2008; elicit anterior insular activation and negative
Wundt 1897). The model assumes that all future arousal, which should diminish risk taking
outcomes are subjectively uncertain (i.e., (Figure 6–1).
probability < 1 and > 0), and all uncertain By extending prior accounts that focused
outcomes potentially evoke anticipation of both primarily on arousal but not valence during
gains and losses. The anticipation elicited by anticipation (Loewenstein et al. 2001), the

anticipatory affect model generates a number of in motivation to higher regions implicated in

novel predictions about how emotion might movement (Draganski et al. 2008; Haber, Fudge,
influence subsequent financial choices (Knutson and McFarland 2000; Lehéricy et al. 2004).
& Greer 2008). First, neural circuits that generate Thus, stimulation and connectivity literatures
positive arousal and negative arousal should both converge to implicate NAcc (and interconnected
show increased activation during anticipation of MPFC) activation as a promising potential neural
uncertain outcomes, but differential activation in marker for positive arousal (Figure 6–1, orange
response to anticipated gain versus loss. Second, circles).
significant activation of these circuits should The connections of circuitry in which
correlate with the self-reported experience of electrical stimulation elicits avoidance
positive arousal and negative arousal, behavior—descending from the insula (Figure 6–
respectively. Third, significant activation of 1, blue circles) and amygdala to the medial
these circuits should have consequences for hypothalamus and periaqueductal grey of the
immediately subsequent choice. As elaborated brainstem—have received less attention in the
below, emerging neural evidence broadly literature. In this circuit, the anterior insula lies
supports these predictions for immediate choices. closest to and shares prominent connections with
By temporal extension, the anticipatory the prefrontal cortex, particularly with the lateral
affect model also might yield predictions about prefrontal cortex, but also with the MPFC
life financial outcomes. To promote the decision (Mesulam and Mufson 1985). Thus, the anterior
to take on debt, the promise of immediate insula (and interconnected amygdala) might
monetary gain might elicit increased positive provide candidates for neural markers of
arousal, the promise of delayed monetary loss negative arousal. The distinctness of these
might not elicit sufficient negative arousal, or regions not only implies that positive arousal and
some combination of the two. Researchers are negative arousal are subserved by distinct
just beginning to turn their attention towards circuits, but also that the output of these circuits
these implications for distant or cumulative may converge in the MPFC (and the
choices. interconnected medial caudate) to influence
The first prediction of the anticipatory affect
NEURAL CIRCUITS FOR ANTICIPATORY model that anticipation of gain and anticipation
AFFECT of loss recruit activation in distinct neural
circuits can be addressed with a judicious
Which brain regions might index anticipatory combinination of brain imaging and incentive
affect in humans? Animal research provides tasks. The development of FMRI in the early
some leads (Panksepp 1998). Electrical 1990s provided the necessary spatial and
stimulation of mesolimbic circuitry elicits temporal resolution (in millimeters and seconds)
approach behavior in all mammalian species to allow researchers to visualize transient
studied (Olds & Fobes 1981). The mesolimbic changes in activation of these subcortical
circuit receives dopamine projections from structures in behaving humans. Initial studies
midbrain neurons (in the ventral tegmental area) (around year 2000) manipulated anticipation of
and includes both subcortical (i.e., the lateral gains and losses in the absence of choice (usually
hypothalamus and the ventral striatum including in the context of delayed response or gambling
the NAcc) and cortical (i.e., the medial prefrontal tasks). More recent studies included choice and
cortex or MPFC) components. Further, used brain activation from previously identified
anatomical studies of both monkeys and humans regions to predict choice. In both types of
indicate that striatal and prefrontal cortical studies, monetary incentives provided a powerful
regions interconnect in an ‘ascending spiral’ experimental tool, since experimenters could
fashion, running from lower regions implicated control anticipation versus outcome, gain versus

Figure 6–2. Changes in affect during anticipation of monetary gains and losses relative to
nonmonetary outcomes (adapted from Samanez-Larkin et al. 2007).

loss, magnitude, probability, and other aspects of performance). Initial findings suggested that
anticipation (Knutson and Cooper 2005). anticipation of monetary gain proportionally
A prototypical example of a task that elicits increased NAcc activation (Knutson et al. 2001).
anticipation of gain and anticipation of loss is In contrast, gain versus nongain outcomes
called the “monetary incentive delay” (MID) increased activation in a part of the MPFC and
task (Knutson et al. 2000). The MID task’s the posterior cingulate, after controlling for
design was inspired by the historic observation anticipation (Knutson et al. 2003). A recent
that in addition to food taste, food cues can elicit meta-analysis of a decade of these types of
salivation in dogs (Pavlov 1927). More recent studies verified the strength and reproducibility
electrophysiological evidence similarly suggests of this pattern of findings (Knutson & Greer
that juice cues elicit increased firing of dopamine 2008). As in initial reports, gain anticipation
neurons in monkeys (Schultz 1998). In a typical elicits greater activation in the NAcc, while loss
MID task trial, subjects initially see a cue anticipation elicits greater activation in some
indicating that they will have an opportunity to (but not all) regions of the anterior insula.
either gain or avoid losing a certain amount of Together, these findings are consistent with the
money, followed by a fixation cross. Next, a prediction that anticipation of gain and
target briefly appears on the screen, and subjects anticipation of loss recruit distinct neural
attempt to press a button before the target is circuits.
replaced by a fixation cross. Finally, subjects see The second prediction of the anticipatory
the outcome of their performance on that trial as affect model can be assessed by correlating brain
well as their cumulative earnings. activation with self-reported affective experience
The structure of the MID task allows separate (assessed either retrospectively or online). In the
visualization of brain responses during MID task, anticipation of gains increases
anticipation of incentives and their outcomes. positive arousal, while anticipation of losses
Separation of gain and loss trials enables increases negative arousal. Further, anticipatory
investigators to directly compare neural affect increases proportional to the magnitude of
responses to gains versus losses and to control anticipated gain or loss (Samanez-Larkin et al.
for potential confounds (related to sensory input, 2007) (Figure 6–2). Peripheral indices of arousal
motor output, arousal or salience, and (e.g., skin conductance) also increase when

subjects anticipate gains and losses (Nielsen et insula activation increased when subjects viewed
al. 2004). These findings suggest that in addition excessive prices (i.e., the displayed price was
to altering brain activation, anticipation of higher than subjects were willing to pay).
incentives reliably changes self-reported Importantly, NAcc activation during product
affective experience within subjects. presentation predicted that subjects would be
While anticipation of incentives influences more likely to buy a product, whereas insula
affect overall, do individual differences in neural activation during price presentation predicted
responses correlate with individual differences in that subjects would be less likely to buy a
affective response? Studies that explored this product (Knutson et al. 2007). After entering
association found that NAcc activation correlates only the brain activation variables into a logistic
with gain-cue-elicited positive arousal but not regression, trial-to-trial purchases could be
negative arousal (Bjork et al. 2004; Knutson et predicted at approximately 60% (versus 50%
al. 2001; Knutson et al. 2005). The specificity of chance, confirmed by cross-validation). New
insular activation to negative arousal, however, analytic techniques that can account for
is less clear. For instance, one study found that multivariate correlations, moreover, increase this
insular activation during anticipation of losses prediction rate to 67% (Grosenick, Greer, and
correlated with both negative arousal and Knutson 2008), and continuing statistical
positive arousal (Samanez-Larkin et al. 2007). refinements that incorporate information from
These findings mostly support the prediction that the whole brain may increase the prediction rate
neural activity should correlate with affective further.
experience during anticipation. Although most Other studies have used brain activation to
peripheral physiological measures (e.g., skin predict choice in the context of investing. For
conductance, pupillary dilation) index arousal, instance, the first FMRI study to use brain
current findings suggest that brain activity activity to predict choice on a trial-to-trial basis
(especially in the NAcc) also indexes valence, did so during an investing task (Kuhnen and
which should provide critical information for Knutson 2005). Although earlier studies had
predicting choice. associated NAcc activation with risk seeking and
To test the third prediction of the anticipatory anterior insula activation with risk aversion, they
affect model that brain activation can predict lacked the temporal resolution to establish
choice, investigators must reverse the traditional whether correlated activation had occurred
logic of brain imaging studies. Instead of before or after choice (Matthews et al. 2004;
examining the effects of input (e.g., cues) on Paulus et al. 2003). In a study designed to mimic
neural responses, investigators focus on whether financial investing, investigators examined
neural activation predicts subsequent output subjects’ anticipatory activation before they
(e.g., the choice to approach or avoid). This made high-risk (stock) or low-risk (bond)
additional constraint potentially focuses investment choices. Further, the investigators
predictions, since anticipation of incentives may determined whether subjects’ choices matched
activate many regions, but only a subset of those those of a risk-neutral rational (Bayesian
regions might influence upcoming choice. updating) actor. After controlling for
Investigators have used brain activation to econometric variables (uncertainty, overall
predict choice in the context of financial wealth, previous actual earnings, and previous
decisions that include purchasing and counterfactual earnings), findings indicated that
investment. anticipatory NAcc activation preceded both
With respect to purchasing, an initial FMRI optimal and suboptimal risk-seeking (stock)
study investigated subjects’ neural responses to choices, while anticipatory anterior insula
products and associated prices before choosing activation preceded both optimal and suboptimal
whether or not to purchase. Findings indicated risk-averse (bond) choices. These effects were
that while NAcc activation increased when most prominent before investors switched choice
subjects viewed preferred products, right anterior strategies, implicating these brain circuits to a

Figure 6–3. Right insular activation during loss anticipation predicts ability to avoid loss months later
(adapted from Samanez-Larkin et al. 2008).

greater extent in choices involving uncertainty towards gain seeking, while individual
than in habitual responses. Additionally, subjects differences in insula function might bias people
with greater insula activation overall tended to towards loss avoidance. For instance, in the
make more risk-averse choices (Kuhnen and investment task described above, when
Knutson 2005). individuals’ NAcc activation matched the
Together, these findings support key expected value of available risky choices, they
implications of the anticipatory affect model— tended to make more rational risk-seeking
anticipation of incentives elicits brain activation, choices (Samanez-Larkin, Wagner, and Knutson
which correlates with anticipatory affect, and can 2009). Additionally, these individuals reported
be used to predict choice. While consistent with greater real-life assets on average (Samanez-
a causal story, however, this evidence is Larkin et al. 2009). In another study, differences
correlational. One could test the causal effect of in insular activity during loss anticipation
activation in these circuits on financial choice by predicted individuals’ abilities to learn to avoid
increasing their activity prior to choice monetary loss in a separate task months after
opportunities. Moreover, such an intervention scanning (Samanez-Larkin et al. 2008) (Figure
need not necessitate electrodes and invasive 6–3). Based on these findings, an obvious
surgery, since incidental affective stimuli can direction for future research involves assessing
also increase activation in some of these circuits. neural and behavioral responses to incentives
Indeed, presentation of erotic pictures (versus and correlating these with real-life financial
frightening or neutral pictures) to heterosexual outcomes, including debt.
males increases their tendency to take financial Overall, brain activity associated with
risks, and this behavioral effect is partially anticipatory affect can be used to predict
mediated by increases in NAcc activation surprisingly diverse financial choices.
(Knutson et al. 2008). Specifically, while NAcc activation predicts
While the above findings focus on approaching gains (e.g., purchasing desirable
immediately upcoming choice, emerging products and approaching risky investments),
evidence is beginning to suggest that individual anterior insular activation predicts avoidance of
differences in NAcc function might bias people losses (e.g., not purchasing overpriced products

and avoiding risky investments). Stimuli or prediction of distal choices, as well as to the
events that incidentally increase activation in detection of chronic biases that might cumulate
these regions can also alter immediately and alter life financial outcomes. Some of the
subsequent financial choices. Although we have existing evidence already elucidates phenomena
focused here on findings from our laboratory relevant to debt. For instance, individual
(Knutson and Cooper 2005), many others have differences in anterior insula activation can
reported corroborating evidence (O'Doherty account for differential risk aversion in an
2004). For instance, activation in the NAcc plays investment task (Kuhnen and Knutson 2005),
a key role in learning to seek monetary gains, and individual differences in anterior insula
while activation in the insula plays a key role in activation in a cued response task can account
learning to avoid monetary losses (Pessiglione et for differential loss avoidance in a separate
al. 2006). laboratory task (Samanez-Larkin et al. 2008).
These and related clinical findings (Paulus and
Stein 2006) suggest that increased insular
CONCLUSIONS AND IMPLICATIONS sensitivity may bias individuals towards
avoiding loss in general, which may extend to
Improvements in brain imaging technology are the specific realm of finance. If financial
revealing a new view of human financial choice. laboratory tasks generalize to real-world
This new view goes below the cortex, changes outcomes, individuals who are sensitive to loss
dynamically on a second-to-second basis, and anticipation may repeatedly avoid debt. Current
implicates evolutionarily ancient circuits research in our laboratory is examining this
associated with affect alongside more recently- prediction, using both self-report and more
evolved circuits associated with deliberation. objective measures of debt.
Emerging findings suggest that incentive cues Neuroeconomic studies thus enable
activate distinct circuits, that this activation investigators to decompose apparently unitary
correlates with affective experience, and can be phenomena (like choice) into subcomponents
used to predict subsequent choices. Specifically, (like anticipatory affect). Successful
NAcc activation precedes approach towards decomposition might imply targeted
potential gains, while anterior insula activation applications. For instance, if a lack of sensitivity
precedes avoidance of potential losses. to future loss plays a more powerful role in
Additional findings (not reviewed here) suggest promoting debt than the attractiveness of present
that prefrontal regions may integrate these gain gains, then personal interventions for reducing
and loss evaluations and allow people to project debt might involve creative ways of making the
themselves into the future, facilitating integration loss obvious and bringing it into the present, or
of more abstract properties of incentives (e.g., other means of recruiting the anterior insula.
probability and delay). Together, these results Beyond enhancing personal control, institutions
have begun to support a nascent model of the might use such neuroeconomic findings to
influence of anticipatory affect on financial and implement wise choice architecture or craft
other choices. policy (as in the case of setting organ donation as
While remarkable progress has occurred in the applicable default rule so that people need
the decade-and-a-half since FMRI’s inception, not confront an unpleasant decision) (Johnson
the current literature only provides a handful of and Goldstein 2003; Thaler and Sunstein 2008).
preliminary demonstrations. Technically, brain Of course, mechanistic knowledge of the
imaging hardware and software improve each underpinnings of choice could be used for
year, but neither has yet been fully optimized for nefarious as well as benevolent purposes.
utilizing brain activity to predict choice. Theorists have argued that some institutions
Conceptually, existing studies have been able to actively encourage debt, ranging from credit
use brain activation to predict immediate choice, cards (which substitute and defer costs) to credit
yet the same frameworks could be extended to default swaps (which disperse and hide risk)

(Prelec and Loewenstein 1998). These Haber, Suzanne N., Julie L. Fudge, and Nikolaus
institutions probably evolved to achieve desired R. McFarland. 2000. “Striatonigrostriatal
effects, however, rather than from any deep Pathways in Primates Form an Ascending
understanding of the neuropsychological Spiral from the Shell to the Dorsolateral
mechanisms that support choice. Thus, Striatum.” Journal of Neuroscience 20:
individuals may derive greater benefit from an 2369-2382.
explicit understanding of these mechanisms, Isen, Alice M., Thomas E. Nygren, and F.
since understanding rather than ignorance will Gregory Ashby. 1988. “Influence of Positive
more likely confer control. Affect on the Subjective Utility of Gains and
Losses: It Is Just Not Worth the Risk.”
Journal of Personality and Social
REFERENCES Psychology 55: 710-717.
Johnson, Eric J., and Daniel Goldstein. 2003.
Ainslie, George. 1992. Picoeconomics: The “Do Defaults Save Lives?” Science 302:
Strategic Interaction of Successive 1338-1339.
Motivational States Within the Person. Knutson, Brian, Charles M. Adams, Grace W.
Cambridge and New York: Cambridge Fong, and Daniel Hommer. 2001.
University Press. “Anticipation of Increasing Monetary
Bjork, James M., Brian Knutson, Grace W. Reward Selectively Recruits Nucleus
Fong, Daniel M. Caggiano, Shannon M. Accumbens.” Journal of Neuroscience
Bennett, and Daniel W. Hommer. 2004. 21(RC159): 1-5.
“Incentive-Elicited Brain Activation in Knutson, Brian, and Jeffrey C. Cooper. 2005.
Adolescents: Similarities and Differences “Functional Magnetic Resonance Imaging of
from Young Adults.” Journal of Reward Prediction.” Current Opinion in
Neuroscience 24: 1793-1802. Neurology 18: 411-417.
Carver, Charles S., and Teri L. White. 1994. Knutson, Brian, Grace W. Fong, Shannon M.
“Behavioral Inhibition, Behavioral Bennett, Charles M. Adams, and Daniel
Activation, and Affective Responses to Hommer. 2003. “A Region of Mesial
Impending Reward and Punishment.” Prefrontal Cortex Tracks Monetarily
Journal of Personality and Social Rewarding Outcomes: Characterization with
Psychology 67: 319-333. Rapid Event-Related fMRI.” NeuroImage
Darwin, Charles. 1872. The Expression of the 18: 263-272.
Emotions in Man and Animals. London: J. Knutson, Brian, and Stephanie M. Greer. 2008.
Murray. “Anticipatory Affect: Neural Correlates and
Draganski, Bogdan, Ferath Kherif, Stefan Consequences for Choice.” Philisophical
Klöppel, Philip A. Cook, Daniel C. Transactions of the Royal Society B:
Alexander, Geoff J.M. Parker, Ralf Biological Sciences 363: 3771-3786.
Deichmann, John Ashburner, and Richard Knutson, Brian, Scott Rick, G. Elliott Wimmer,
S.J. Frackowiak. 2008. “Evidence for Drazen Prelec, and George Loewenstein.
Segregated and Integrative Connectivity 2007. “Neural Predictors of Purchases.”
Patterns in the Human Basal Ganglia.” Neuron 53: 147-156.
Journal of Neuroscience 28: 7143-7152. Knutson, Brian, Jonathan Taylor, Matthew
Grosenick, Logan, Stephanie M. Greer, and Kaufman, Richard Peterson, and Gary
Brian Knutson. 2008. “Interpretable Glover. 2005. “Distributed Neural
Classifiers for fMRI Improve Prediction of Representation of Expected Value.” Journal
Purchases.” IEEE Transactions on Neural of Neuroscience 25: 4806-4812.
Systems and Rehabilitation Engineering 16: Knutson, Brian, Andrew Westdorp, Erica Kaiser,
539-548. and Daniel Hommer. 2000. “FMRI
Visualization of Brain Activity During a

Monetary Incentive Delay Task.” Nielsen, Lisbeth, Brian Knutson, Matt Kaufman,
NeuroImage 12: 20-27. Lauren Weinstein, and Laura L. Carstensen.
Knutson, Brian, G. Elliott Wimmer, Camelia M. 2004. “Facial EMG Discriminates Gain and
Kuhnen, and Piotr Winkielman. 2008. Loss Anticipation and Outcome in a
“Nucleus Accumbens Activation Mediates Monetary Incentive Delay Task.”
the Influence of Reward Cues on Financial Psychophysiology 41: S80.
Risk Taking.” NeuroReport 19: 509-513. O’Doherty, John P. 2004. “Reward
Kuhnen, Camelia M., and Brian Knutson. 2005. Representations and Reward-Related
“The Neural Basis of Financial Risk Learning in the Human Brain: Insights from
Taking.” Neuron 47: 763-770. Neuroimaging.” Current Opinion in
Lea, Stephen E.G., Avril J. Mewse, and Wendy Neurobiology 14: 769-776.
Wrapson. 2010. “The Psychology of Debt in Olds, M.E., and J.L. Fobes. 1981. “The Central
Poor Households in Britain.” In A Debtor Basis of Motivation: Intracranial Self-
World: Interdisciplinary Perspectives on an Stimulation Studies.” Annual Review of
Indebted Global Society, edited by Ralph Psychology 32: 523-574.
Brubaker, Robert M. Lawless, and Charles J. Panksepp, Jaak. 1998. Affective Neuroscience:
Tabb. New York: Oxford University Press. The Foundations of Human and Animal
Lea, Stephen E.G., Paul Webley, and R. Mark Emotions. New York: Oxford University
Levine. 1993. “The Economic Psychology Press.
of Consumer Debt.” Journal of Economic Paulus, Martin P., Corianne Rogalsky, Alan
Psychology 14: 85-119. Simmons, Justin S. Feinstein, and Murray B.
Lehéricy, Stéphanie, Mathieu Ducros, Pierre- Stein. 2003. “Increased Activation in the
Francois Van de Moortele, Chantal Francois, Right Insula During Risk-Taking Decision
Lionel Thivard, Cyril Poupon, Nick Making Is Related to Harm Avoidance and
Swindale, Kamil Ugurbil, and Dae-Shik Neuroticism.” NeuroImage 19: 1439–1448.
Kim. 2004. “Diffusion Tensor Fiber Paulus, Martin P., and Murray B. Stein. 2006.
Tracking Shows Distinct Corticostriatal “An Insular View of Anxiety.” Biological
Circuits in Humans.” Annals of Neurology Psychiatry 60: 383-387.
55: 522-529. Pavlov, Ivan Petrovich. 1927. Conditioned
Loewenstein, George F., Elke U. Weber, Reflexes: An Investigation of the
Christopher K. Hsee, and Ned Welch. 2001. Physiological Activity of the Cerebral
“Risk as Feelings.” Psychological Bulletin Cortex. London: Oxford University Press.
127: 267–286. Pessiglione, Mathias, Ben Seymour, Guillaume
Matthews, Scott C., Alan N. Simmons, Scott D. Flandin, Raymond J. Dolan, and Chris D.
Lane, and Martin P. Paulus. 2004. “Selective Frith. 2006. “Dopamine-Dependent
Activation of the Nucleus Accumbens Prediction Errors Underpin Reward-Seeking
During Risk-Taking Decision Making.” Behaviour in Humans.” Nature 442: 1042-
NeuroReport 15: 2123-2127. 1045.
McGue, Matt, and Thomas J. Bouchard, Jr. Prelec, Drazen, and George Loewenstein. 1998.
1998. “Genetic and Environmental “The Red and the Black: Mental Accounting
Influences on Human Behavioral of Savings And Debt.” Marketing Science
Differences.” Annual Review of 17: 4-28.
Neuroscience 21: 1-24. Samanez-Larkin, Gregory R., Sasha E.B. Gibbs,
Mesulam, M.-Marsel, and Elliott J. Mufson. Kabir Khanna, Lisbeth Nielsen, Laura L.
1985. “The Insula of Reil in Man and Carstensen, and Brian Knutson. 2007.
Monkey: Architectonics, Connectivity, and “Anticipation of Monetary Gain But Not
Function.” In Vol. 4, Cerebral Cortex, Loss in Healthy Older Adults.” Nature
edited by Alan Peters and Edward G. Jones, Neuroscience 10: 787-791.
179-226. New York: Plenum Press.

Samanez-Larkin, Gregory R., Nick G. Hollon, Stone, Brice, and Rosalinda Vasquez Maury.
Laura L. Carstensen, and Brian Knutson. 2006. “Indicators of Personal Financial Debt
2008. “Individual Differences in Insular Using a Multi-Disciplinary Behavioral
Sensitivity During Loss Anticipation Predict Model.” Journal of Economic Psychology
Avoidance Learning.” Psychological Science 27: 543-556.
19: 320-323. Thaler, Richard H., and Cass R. Sunstein. 2008.
Samanez-Larkin, Gregory R., Camelia M. Nudge: Improving Decisions About Health,
Kuhnen, Daniel J. Yoo, and Brian Knutson. Wealth, and Happiness. New Haven: Yale
2009. “Variability in Nucleus Accumbens University Press.
Activity Mediates Age-Related Financial Watson, David, and Auke Tellegen. 1985.
Mistakes.” Working Paper. Stanford, CA: “Toward a Consensual Structure of Mood.”
Stanford University Department of Psychological Bulletin 98: 219-235.
Psychology. Wilson, Timothy D., and Daniel T. Gilbert.
Samanez-Larkin, G. R., Anthony D. Wagner, 2003. “Affective Forecasting.” In Vol. 35,
and Brian Knutson. 2009. “Integrated Advances in Experimental Social
Incentive Information Improves Risky Psychology, edited by Mark P. Zanna, 345-
Decision Making in Older Adults.” Working 411. New York: Academic Press.
Paper. Stanford, CA: Stanford University Wundt, Wilhelm Max. 1897. Outlines of
Department of Psychology. Psychology. Translated by Charles Hubbard
Schultz, Wolfram. 1998. “Predictive Reward Judd. Leipzig: Wilhelm Engelmann; New
Signal of Dopamine Neurons.” Journal of York: G.E. Stechert.
Neurophysiology 80: 1-27.