Self-Affirmation Activates the Ventral Striatum: A Possible Reward-Related Mechanism for Self-Affirmation

Abstract

Self-affirmation (reflecting on important personal values) has been shown to have a range of positive effects; however, the neural basis of self-affirmation is not known. Building on studies showing that thinking about self-preferences activates neural reward pathways, we hypothesized that self-affirmation would activate brain reward circuitry during functional MRI (fMRI) studies. In Study 1, with college students, making judgments about important personal values during self-affirmation activated neural reward regions (i.e., ventral striatum), whereas making preference judgments that were not self-relevant did not. Study 2 replicated these results in a community sample, again showing that self-affirmation activated the ventral striatum. These are among the first fMRI studies to identify neural processes during self-affirmation. The findings extend theory by showing that self-affirmation may be rewarding and may provide a first step toward identifying a neural mechanism by which self-affirmation may produce a wide range of beneficial effects.

Full text

Psychological Science
2016, Vol. 27(4) 455 –466
© The Author(s) 2016
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0956797615625989
pss.sagepub.com
Research Article
Self-affirmation—the process of reflecting on important
personal values or personal characteristics and
strengths—has been shown to have a broad range of
benefits in hundreds of studies (for reviews, see Cohen
& Sherman, 2014; Sherman & Cohen, 2006). For exam-
ple, self-affirmation has been shown to reduce defen-
siveness (Sherman, Nelson, & Steele, 2000) and stress
(Creswell et al., 2005) and to improve academic out-
comes (e.g., grade point average, problem-solving per-
formance; Cohen, Garcia, Apfel, & Master, 2006;
Creswell, Dutcher, Klein, Harris, & Levine, 2013). Self-
affirmation has also been shown to have a range of posi-
tive effects on social and affective behavior, including
improved self-control after rejection (Burson, Crocker, &
Mischkowski, 2012), increased well-being (Nelson, Fuller,
Choi, & Lyubomirsky, 2014), reduced rumination (Koole,
Smeets, Van Knippenberg, & Dijksterhuis, 1999), and
enhanced feelings of relational security (Stinson, Logel,
Shepherd, & Zanna, 2011). However, we know little
about the underlying neural mechanisms.
Recent research has focused on building mechanistic
accounts of self-affirmation; studies have suggested that
625989PSSXXX10.1177/0956797615625989Dutcher et al.A Possible Reward-Related Mechanism for Self-Affirmation
research-article2016
Corresponding Author:
Janine M. Dutcher, Department of Psychology, University of
California, Los Angeles, 1285 Franz Hall, Box 951563, Los Angeles, CA
90095-1563
E-mail: jdutcher@ucla.edu
Self-Affirmation Activates the Ventral
Striatum: A Possible Reward-Related
Mechanism for Self-Affirmation
Janine M. Dutcher1, J. David Creswell2, Laura E. Pacilio2,
Peter R. Harris3, William M. P. Klein4, John M. Levine4,
Julienne E. Bower1,5,6, Keely A. Muscatell7, and
Naomi I. Eisenberger1
1Department of Psychology, University of California, Los Angeles; 2Department of Psychology, Carnegie Mellon
University; 3School of Psychology, University of Sussex; 4Department of Psychology, University of Pittsburgh;
5Department of Psychiatry and Biobehavioral Science, University of California, Los Angeles; 6Semel Institute for
Neuroscience and Human Behavior, Cousins Center for Psychoneuroimmunology, University of California, Los
Angeles; and 7Robert Wood Johnson Foundation Health and Society Scholars Program, University of California,
San Francisco, and University of California, Berkeley
Abstract
Self-affirmation (reflecting on important personal values) has been shown to have a range of positive effects; however,
the neural basis of self-affirmation is not known. Building on studies showing that thinking about self-preferences
activates neural reward pathways, we hypothesized that self-affirmation would activate brain reward circuitry during
functional MRI (fMRI) studies. In Study 1, with college students, making judgments about important personal values
during self-affirmation activated neural reward regions (i.e., ventral striatum), whereas making preference judgments
that were not self-relevant did not. Study 2 replicated these results in a community sample, again showing that self-
affirmation activated the ventral striatum. These are among the first fMRI studies to identify neural processes during
self-affirmation. The findings extend theory by showing that self-affirmation may be rewarding and may provide a first
step toward identifying a neural mechanism by which self-affirmation may produce a wide range of beneficial effects.
Keywords
self-affirmation, ventral striatum, neural reward regions
Received 3/25/15; Revision accepted 12/15/15
at UCLA on April 13, 2016pss.sagepub.comDownloaded from

456 Dutcher et al.
self-affirmation triggers a psychological cascade of effects,
such as increasing attention to threat, broadening per-
spective, increasing feelings of social connection, and
enhancing coping resources (e.g., Correll, Spencer, &
Zanna, 2004; Crocker, Niiya, & Mischkowski, 2008; Klein
& Harris, 2009; Sherman & Hartson, 2011). In a recent
review, Cohen and Sherman (2014) noted there could be
multiple mechanisms for self-affirmation and using a
variety of methods could illuminate these processes.
However, most accounts have not identified the basic
neural and cognitive processes leading to these psycho-
logical effects. In particular, no published research to
date has considered which neural regions are involved
during self-affirmation (although some work has exam-
ined the neural consequences of self-affirmation; Falk
etal., 2015; Legault, Al-Khindi, & Inzlicht, 2012). Using
neuroimaging as a tool to investigate the self-affirmation
process could help advance self-affirmation theory, given
that neuroimaging does not rely on self-report.
In two studies, we proposed and tested a novel self-
affirmation neural-reward account. Specifically, we pos-
ited that engaging in self-affirmation activates not only
self-related neural regions (medial prefrontal cortex, pre-
cuneus) but also neural reward pathways in the brain’s
mesolimbic dopamine system. Our proposed account
was informed by studies showing that neural reward
regions, such as the ventral striatum (VS) and the ventral
tegmental area (VTA), are activated when people disclose
self-traits or beliefs (Tamir & Mitchell, 2012). The VS is
also more active when participants think about positive,
compared with neutral, autobiographical memories
(Speer, Bhanji, & Delgado, 2014). These studies provide
evidence that thinking about positive aspects of the self
may activate neural reward pathways.
This reward account may provide a mechanistic expla-
nation for self-affirmation’s effects on threat and stress
responding, given that previous research has found that
rewarding stimuli (e.g., sexual stimuli, sucrose) decrease
physiological stress responses in humans (Creswell,
Pacilio, Denson, & Satyshur, 2013) and rats (Ulrich-Lai
etal., 2010). It is possible that, when participants perform
a self-affirmation task, their reward-related neural activity
increases, which diminishes their neural responses
tothreat, allowing them to be more resilient and open
to self-related threats compared with nonaffirming
participants.
We conducted two neuroimaging studies to explore
neural activity specific to the process of self-affirmation;
in particular, we examined whether self-affirmation led
to increased activity in neural reward regions (VS and
VTA) relative to nonaffirmation (in Study 1) or control
(in Study 2).
In Study 1, college-age participants were randomly
assigned to either a self-affirmation group, in which they
made decisions about important personal values (experi-
mental condition) and decisions about less important
personal values (control condition), or a nonaffirmation
group, in which they made decisions about attributes
other people might prefer in toasters (experimental con-
dition) and decisions about less important toaster attri-
butes (control condition). Thus, this study used a 2 × 2
mixed design, with group (self-affirmation, nonaffirma-
tion) manipulated between subjects and ranking of
value or attribute (high, low) manipulated within sub-
jects. We included a nonaffirmation group to evaluate
whether neural reward activity was due specifically to
self-affirmation or to the difference in value between the
experimental conditions (important personal values or
toaster attributes) and control conditions (less important
personal values or toaster attributes). In Study 2, we
tested whether these effects extended to a community
sample of older adult women.
Study 1
Method
Participants. Forty university students (18 female;
mean age = 24.13 years, SD = 5.72) completed study pro-
cedures. Data collection was stopped after each group
reached a minimum of 20 participants, but data from 2
participants in the nonaffirmation group were not saved
after the completion of the imaging session, so that group
had 18 participants with usable data. All participants met
eligibility criteria for functional MRI (fMRI) studies (i.e.,
they were right-handed, not claustrophobic, free of
implanted metal, and not pregnant). Fifty-nine percent
were White, 3% were Hispanic, 11% were Black, 11%
were Asian American or Asian, and 16% were of “other”
race. The Carnegie Mellon University institutional review
board approved all study procedures.
Procedure. Before the scanning session, participants
were randomly assigned to either the self-affirmation
group (n = 20) or the nonaffirmation group (n = 18). We
used a standard self-affirmation decision-making task
(Steele, 1988; Steele & Liu, 1983); participants were given
a series of paired personal-value statements and were
asked to indicate their relative preference (adapted from
Vernon & Allport, 1931; see Fig. 1). In the self-affirmation
group, participants ranked five personal values (art, reli-
gion, science, social issues, politics) in order of impor-
tance. We used this list to create a scanner task that was
specific to each participant’s own personal values. In the
nonaffirmation group, participants were asked to rank a
list of five toaster attributes (e.g., slice capacity, color, size)
in the order of the importance that they believed an aver-
age college student would rank them (for frequency tables
at UCLA on April 13, 2016pss.sagepub.comDownloaded from

A Possible Reward-Related Mechanism for Self-Affirmation 457
reporting the personal values and toaster attributes selected
by participants, see the Supplemental Material available
online). We used this list to create a scanner task that was
specific to each participant’s beliefs about what other peo-
ple prefer. Before the scanning session, participants were
trained on the tasks assigned to them. In these training
sessions, participants were familiarized with the pictures
that represented each personal value or toaster attribute,
so the pictures were not novel to participants at test.
Although many self-affirmation studies use a nonaffir-
mation control condition in which participants focus on
a value of lower personal relevance and why it might be
important to someone else (Sherman & Cohen, 2006), the
design for this study was slightly different to accommo-
date standard imaging techniques. Specifically, fMRI con-
trasts require a within-subjects design, using comparisons
between two conditions to isolate neural activity to the
specific psychological differences between the condi-
tions. Therefore, both groups in the current study
included an experimental condition (self-affirmation
group: high personal-value ratings; nonaffirmation group:
high toaster attribute ratings) and a control condition
(self-affirmation group: low personal-value ratings; non-
affirmation group: low toaster attribute ratings).
Consequently, our within-subjects self-affirmation
conditions differed only in how the personal values dis-
played had previously been ranked by the participant,
which controlled for any self-related processes. The
within-subjects nonaffirmation conditions also differed
only in how the toaster attributes displayed had previ-
ously been ranked by the participant. Including both
within- and between-subjects levels of comparison
offered a more specific test of whether self-affirmation
relies on reward activity inherent in the self-affirmation
process itself or just relies on self- or value-related reflec-
tion. This design also closely approximated the original
paradigms used in behavioral self-affirmation studies,
ensuring that we manipulated self-affirmation.
Imaging procedures for the self-affirmation and non-
affirmation tasks. During each task, participants viewed
instructions, pictures, and words via a high-resolution
projector and were asked to make responses (when
appropriate) using a five-button data glove.
The self-affirmation task used a block design. Each
block included three decision-making trials, and each
trial lasted for 8 s, for a total of 24 s per block. In the self-
affirmation trials, participants were shown pictures asso-
ciated with two personal values, one of which was always
the top-ranked value; the values’ labels appeared beneath
the pictures (Fig. 1a). Participants were asked to think
about the role of these two personal values in their lives
and then indicate which of the two displayed values was
more important to them. Participants responded using a
5-point scale (1 = strongly prefer [the value on the left],
2 = slightly prefer [the value on the left], 3 = no
ab
Social Issues Politics
1234 5
Strongly
Prefer
No Preference Strongly
Prefer
123
45
Strongly
Prefer
No Preference Strongly
Prefer
Slice Capacity Color
Fig. 1. Examples of trials in the (a) self-affirmation and (b) nonaffirmation tasks in Study 1. In each trial, participants in the self-affirmation group (a)
saw pictures associated with two personal values, with the values’ labels below the pictures. Participants indicated which value they preferred, on a
scale from 1 to 5 (1 = strongly prefer [the value on the left], 2 = slightly prefer [the value on the left], 3 = no preference, 4 = slightly prefer [the value on
the right], 5 = strongly prefer [the value on the right]). In the experimental condition, one of the values was always the given participant’s top-ranked
value; in the yoked control condition, the values were always the given participant’s two bottom-ranked values. Participants in the nonaffirmation
group (b) saw pictures illustrating two toaster attributes, with the attributes’ labels below the pictures. Participants indicated which attribute they
thought the average college student would prefer, on the same scale from 1 to 5. In the experimental condition, one of the attributes was always
the given participant’s top-ranked attribute; in the yoked control condition, the attributes were always the given participant’s two bottom-ranked
attributes. The same pictures could represent either the experimental condition or the control condition, depending on a given participant’s rankings.
at UCLA on April 13, 2016pss.sagepub.comDownloaded from

458 Dutcher et al.
preference, 4 = slightly prefer [the value on the right], 5 =
strongly prefer [the value on the right]).
The control blocks had the same timing and instruc-
tions; however, participants were shown their fourth- and
fifth-ranked personal values on each trial. Thus, the only
difference between the experimental and control trials
was whether subjects were able to affirm important
values during the trials. Participants completed four
experimental blocks in one run and four control blocks
in another run. Run order was counterbalanced across
participants. The two conditions were completed in sepa-
rate runs to reduce carryover effects, given that self-
affirmation has been shown to have sustained benefits
over time (Cohen etal., 2006).
We wanted to have a comparison group to test whether
the observed neural correlates of self-affirmation were
due to the difference between making preference judg-
ments regarding more important characteristics and mak-
ing preference judgments regarding less important
characteristics. To this end, we created a task that was
similar to the self-affirmation task in design and demands
but did not lead participants to self-affirm. As in the self-
affirmation task, each nonaffirmation block included
three 8-s decision-making trials, for a total of 24 s per
block. In these trials, participants were shown pictures
associated with toaster attributes with the attribute label
beneath each picture (see Fig. 1b). Each experimental
block contained three experimental trials. In the experi-
mental trials, participants were asked to indicate which of
the two displayed toaster attributes, one of which was
always the top-ranked attribute, was more important to
the average college student. Participants responded using
a 5-point scale (1 = strongly prefer [the attribute on the
left], 2 = slightly prefer [the attribute on the left], 3 = no
preference, 4 = slightly prefer [the attribute on the right],
5= strongly prefer [the attribute on the right]). Each con-
trol block included three trials. In the control trials, par-
ticipants were given the same instructions, but were
shown the fourth- and fifth-ranked attributes. Each of the
trials and blocks was the same length as in the experi-
mental condition. Participants completed four experi-
mental blocks in one run and four control blocks in
another run to match the self-affirmation group. Runs
were counterbalanced.
Image acquisition. Data were acquired using a 3-T
MRI scanner (Verio; Siemens Medical Solutions USA,
Malvern, PA). Foam padding surrounded each parti-
cipant’shead to reduce head movement. For each par-
ticipant, we first acquired a high-resolution structural
magnetization-prepared rapid gradient-echo (MPRAGE)
imaging volume—repetition time (TR) = 1,700 ms, echo
time (TE) = 2.48, matrix size = 128 × 128, resolution =
1 × 1 × 1 mm, field of view (FOV) = 256 mm, 176 slices (1
mm thick), flip angle = 9°, and bandwidth = 170 Hz/pixel.
For the tasks, two functional gradient-echo scans (3 min
for self-affirmation/toasters and 3 min for control) were
acquired—TR = 2,000 ms, TE = 29 ms, flip angle = 79°,
matrix size = 64 × 64, resolution = 3 × 3 × 3 mm, FOV=
192 mm, 36 axial slices (3 mm thick), and bandwidth =
2232 Hz/pixel. These tasks were conducted in separate
runs to give participants a break and to minimize carry-
over effects.
fMRI data analysis. Imaging data were analyzed
using Statistical Parametric Mapping software (SPM8;
Wellcome Department of Cognitive Neurology, Institute
of Neurology, London, England). For preprocessing, we
first manually reoriented the echoplanar images to align
brains along a horizontal anterior commissure-posterior
commissure, with an image origin at the anterior com-
missure. For functional images, the first run’s first-image
parameters were applied to each subsequent volume in
the respective run to correct for head motion. Structural
MPRAGE images were normalized to Montreal Neuro-
logical Institute (MNI) space using diffeomorphic ana-
tomical registration through exponentiated lie (DARTEL)
algorithms and were then smoothed using an 8-mm
Gaussian kernel, full-width at half maximum. Before first-
level analyses, images were visually inspected for accu-
rate normalization. The 24 s of trials for each condition
(self-affirmation experimental condition, self-affirmation
control condition, nonaffirmation experimental condi-
tion, nonaffirmation control condition) were modeled as
blocks. Rest periods, when participants viewed a fixation
cross between blocks, comprised the implicit baseline.
We computed linear contrasts for the self-affirmation
experimental condition compared with its control condi-
tion (high-rated personal values vs. low-rated personal
values) for each participant. These individual contrast
images were then used in group-level analyses. For the
nonaffirmation group, we computed linear contrasts for
the experimental condition compared with its control
condition (high-rated toaster attributes vs. low-rated
toaster attributes) for each participant. These individual
contrast images were then used in group-level analyses.
In addition, to determine whether there were differences
in neural activation between the two groups, an indepen-
dent (two-sample) t test was computed comparing the
self-affirmation group (self-affirmation experimental
minus yoked control contrast) to the nonaffirmation
group (experimental minus yoked control contrast).
On the basis of a priori predictions that self-affirmation
would activate reward-related regions, group-level results
were examined using regions of interest (ROIs) of the left
and right VS and the VTA. VS ROIs were structurally
defined using the automated anatomical labeling atlas
(Tzourio-Mazoyer et al., 2002); caudate nucleus and
at UCLA on April 13, 2016pss.sagepub.comDownloaded from

A Possible Reward-Related Mechanism for Self-Affirmation 459
putamen from the atlas were combined and constrained
at x between 0 and –24, y between 4 and 18, and z
between 0 and –12 for the left ROI and x between 0 and
24, y between 4 and 18, and z between 0 and –12 for the
right ROI (based on ROIs from Inagaki & Eisenberger,
2012). Thus, we constrained the ROI to the ventral parts
of the caudate nucleus and putamen to create this VS
ROI. The VTA ROI was created in the MarsBar Toolbox
for SPM (http://marsbar.sourceforge.net; Brett, Anton,
Valabregue, & Poline, 2002) and centered at (x = 2, y =
−12, z = −8) within a 3-mm radius on the basis of previ-
ous work investigating social reward (Aron etal., 2005).
We examined activity within each of these ROIs for the
self-affirmation and nonaffirmation conditions relative to
their yoked control conditions. Parameter estimates rep-
resenting the average activity across all voxels in the ROI
were extracted and averaged. ROI analyses were run in
MarsBar, which reports an fMRI standard statistical thresh-
old of p < .05, one-tailed. Confidence intervals (CIs) for
these tests were estimated using the bias-corrected and
accelerated-percentile method (10,000 random samples
with replacement), implemented using the BOOTCI func-
tion in MATLAB (The MathWorks, Natick, MA).
Because this was the first study to explore neural
activity during self-affirmation, we also conducted explor-
atory whole-brain analyses to provide a complete picture
of the neural regions involved in this process. Thus, we
performed whole-brain analyses contrasting activity in
the self-affirmation experimental trials (relative to activity
in the yoked control trials) with activity in the nonaffir-
mation experimental trials (relative to activity in the
yoked control trials). We used an exploratory threshold
of p < .005 and 20 voxels (Lieberman & Cunningham,
2009). We then explored the post hoc simple effects. To
do so, we created ROIs based on the clusters of activity
in the whole-brain analysis and extracted and averaged
parameter estimates for the contrast between experimen-
tal and yoked control trials in the self-affirmation group
and for the contrast between experimental and yoked
control trials in the nonaffirmation group. Analyses were
run in MarsBar, which reports an fMRI standard statistical
threshold of p < .05, one-tailed.
Results
ROI analyses. To examine differences in neural activity
between the self-affirmation and nonaffirmation groups,
we investigated whether activity in the self-affirmation
experimental condition (measured as activity relative to
its yoked control condition) was greater than activity in
the nonaffirmation experimental condition (measured as
activity relative to its yoked control condition). We then
further examined these effects by examining neural activ-
ity in each group separately.
Results showed greater activity in the left VS in the
self-affirmation group compared with the nonaffirmation
group, t(36) = 2.04, p = .025 (Fig. 2). Consistent with our
hypotheses, within the self-affirmation group, there was
greater activity in the left VS during the experimental tri-
als than in the control trials, t(19) = 2.12, p = .024, mean
parameter estimate = 0.538, 95% CI = [0.022, 0.987],
whereas within the nonaffirmation group, there was no
difference in left VS activity during the experimental and
control trials, t(17) = −0.53, p > .250, mean parameter
estimate = −0.099, 95% CI = [−0.424, 0.280].
Results showed greater activity in the right VS in the
self-affirmation group than in the nonaffirmation group,
but the difference was not statistically significant, t(36) =
1.02, p = .157 (Fig. 2). As we found for the left VS, this
effect was driven by the self-affirmation group; there was
greater activity, albeit not significantly greater, in the right
VS during the experimental trials than during the control
trials, t(19) = 1.06, p = .152, mean parameter estimate =
0.306, 95% CI = [−0.257, 0.837], but there was no differ-
ence between the experimental and control trials in the
nonaffirmation group, t(17) = −0.32, p > .250, mean
parameter estimate = −0.072, 95% CI = [−0.483, 0.350].
Finally, there were no differences in VTA activity
between the self-affirmation and nonaffirmation groups,
t(36) = 0.13, p > .250 (Fig. 2). Specifically, in the self-
affirmation group, there was marginally greater (albeit
nonsignificant) VTA activity in the experimental trials
than in the control trials, t(19) = 1.34, p = .098, mean
parameter estimate = 0.731, 95% CI = [−0.054, 2.119], and
the same pattern was seen in the nonaffirmation group,
t(17) = 1.22, p = .119, mean parameter estimate = 0.632,
95% CI = [−.203, 1.806].
Whole-brain analyses. Like the ROI analyses, whole-
brain analyses revealed significantly greater activation in
the VS during the self-affirmation experimental trials (rel-
ative to activation in the yoked control) than in the non-
affirmation experimental trials (relative to activation in
the yoked control; Table 1). There was also significant
activation in clusters in the medial prefrontal cortex and
precuneus/posterior cingulate cortex, regions previously
shown to play a role in self-processing (Heatherton etal.,
2006; for a full list of activations, see Table 1). To further
investigate these findings, we used post hoc tests of sim-
ple effects to explore which trials drove the effect. As
expected for the VS, these tests revealed that there was
greater activity in the VS during self-affirmation experi-
mental trials than during the yoked control trials, t(19) =
2.26, p = .018, but there was no difference in activity
between the nonaffirmation experimental and yoked
control trials, t(17) = −2.90, p > .250. The other clusters
showed the same pattern: There was greater activity dur-
ing the self-affirmation experimental trials than during
at UCLA on April 13, 2016pss.sagepub.comDownloaded from

460 Dutcher et al.
the yoked control trials (ps < .024; medial prefrontal cor-
tex: p = .143), but there was no difference in activity
between the nonaffirmation experimental and yoked
control trials (ps > .559).
On the other hand, whole brain analyses revealed no
significant clusters of activity during the reverse contrast,
which tested for greater activity in the experimental con-
dition (relative to the yoked control condition) in the non-
affirmation group compared with the self-affirmation
group. In addition, within the self-affirmation group, there
were no clusters with significantly greater activity in the
control condition than in the experimental condition.
–0.4
–0.2
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Parameter Estimate
Self-Affirmation
Nonaffirmation
*
†
Left Ventral
Striatum
Right Ventral
Striatum
Ventral Tegmental
Area
*
Fig. 2. Results for Study 1: parameter estimates for activity in the left ventral striatum, the
right ventral striatum, and the ventral tegmental area during experimental trials are graphed
separately for each condition. Activation was measured relative to the corresponding control
trials. Error bars represent ±1 SE. Symbols indicate parameter estimates that are significantly or
marginally significantly different from 0 as well as a significant difference between conditions
(†p < .10, *p < .05). The coronal scan (top left) shows the location of the regions of interest in
the left ventral striatum and the right ventral striatum, and the transverse scan (top right) shows
the location of the region of interest in the ventral tegmental area.
Table 1. Results From Study 1: Brain Regions More Active During Self-Affirmation Trials Than During Nonaffirmation Trials
Anatomical region
Brodmann’s
area Hemisphere
MNI coordinates of peak voxel
t(36)
Number of
voxels (k) x y z
Ventral striatum — Left –3 15 –6 3.48 29
Medial prefrontal cortex 10 0 54 –6 4.15 330
Occipital cortex 18 Right 6 –78 –3 4.02 321
Subgenual anterior cingulate cortex 24 Left –18 24 –6 3.40 22
Thalamus — Right 3 –9 3 4.04 46
Middle temporal cortex 39 Left –36 –63 12 3.45 36
Rostrolateral prefrontal cortex 9 Left –24 30 36 4.41 40
Note: Activations were measured relative to activation in the yoked control trials. The table reports significant activations (p < .005) of
clusters with a minimum size of 20 voxels. The t tests were conducted at peak coordinates. MNI = Montreal Neurological Institute.
at UCLA on April 13, 2016pss.sagepub.comDownloaded from

A Possible Reward-Related Mechanism for Self-Affirmation 461
Interim Summary
Results indicated that the self-affirmation group showed
greater VS activity than did the nonaffirmation group.
This effect was driven by increased VS activity specific
to the self-affirmation condition. Whole-brain results
revealed that neural regions implicated in self-processing
(precuneus/posterior cingulate cortex) and reward pro-
cessing (VS) were more active in the self-affirmation
group than in the nonaffirmation group. Results from the
VTA were inconclusive. Most self-affirmation paradigms
do not include pictures, so we conducted a second study
without pictures to rule out the possibility that the VS
activity was due to viewing pictures. Moreover, to test the
generalizability of our findings, we conducted this study
with a community sample.
Study 2
Method
Participants. Twenty-one female participants (mean
age = 55.5 years) completed the study procedures. All
participants were deemed eligible for fMRI (i.e., right-
handed, not claustrophobic, free of implanted metal,
and not pregnant). Participants were recruited from a
larger study investigating the neurobiological pathways
linking psychological stress and inflammation in breast
cancer survivors and healthy control participants; thus,
6 of the participants were breast-cancer survivors.
Because there were no differences in neural activity
between the healthy participants and breast cancer sur-
vivors in any of the analyses for any of the ROIs (ps >
.20), we collapsed the data across participants for all
analyses reported here. We aimed to collect data from a
minimum of 20 participants, and data collection stopped
at the targeted enrollment for the larger study. Data
from 1 participant were excluded because she did not
follow task instructions. Seventy-six percent of partici-
pants were White, 14% were of “other” race, 5% were
Hispanic, and 5% were Asian American. The University
of California, Los Angeles, institutional review board
approved all procedures.
Procedure
Imaging procedures for the self-affirmation task. Before
the scanning session, participants were emailed a survey in
which they were asked to rank a list of 11 personal values
(e.g., art, religion, friends and family) in order of impor-
tance (for a frequency table reporting the personal values
selected by participants, see the Supplemental Materials).
From this, we were able to create tasks that were specific
to each participant’s most important personal value for the
scanning session. During the scan, participants viewed
instructions and words through scanner-compatible
goggles and were asked to make responses (when appro-
priate) using a four-button button box.
The self-affirmation task was similar to that used in
Study 1; however, for this study, all participants completed
the self-affirmation and control trials only (i.e., there was
no nonaffirmation group). For this experiment, partici-
pants selected which of the two personal values shown
onscreen was most important to them on each trial. How-
ever, in this study, participants were shown only the per-
sonal-value label, without a picture; we hoped to ensure
that the results observed in Study 1 were not being driven
by participants’ seeing pictures of important values. Each
self-affirmation block included three trials, lasting 7 s
each, separated by a 1-s fixation cross, for a total of 23 s
per block. During the self-affirmation trials, participants
were shown their top-ranked personal value and another
highly ranked value. They were asked to indicate which
of the two personal values displayed was more important
using a 4-point scale (1 = strongly prefer [the value on the
left], 2 = slightly prefer [the value on the left], 3 = slightly
prefer [the value on the right], 4 = strongly prefer [the
value on the right]). Each control block included three tri-
als. During the control trials, participants were given the
same instructions, but were shown two personal values
that they had rated as being less important to them, with
the same timing as the self-affirmation trials and blocks.
Participants completed three self-affirmation blocks and
three control blocks. Blocks were randomized once with
the constraint that there were never three consecutive
blocks of the same condition for any task.1 The blocks
could be ordered in seven different ways, but each par-
ticipant saw only one order.
Image acquisition. Data were acquired on a Sie-
mens Trio 3-T MRI scanner. Foam padding surrounded
each participant’s head to reduce head movement. For
each participant, we acquired a high-resolution struc-
tural matched-bandwidth scan—TR = 5,000 ms, TE =
34, matrix size = 128 × 128, resolution = 1.6 × 1.6 ×
3 mm, FOV = 200 mm, 36 slices (3 mm thick), flip angle =
90°, and bandwidth = 1302 Hz/pixel. The self-affirmation
task was completed in one functional scan lasting 436 s
(about 7 min, 16 s)—TR = 2,000 ms, TE = 25 ms, matrix
size = 64× 64, resolution = 3.1 × 3.1 × 4.0 mm, FOV =
200 mm, 33 axial slices (3 mm thick with 1-mm gap), flip
angle = 90°, and bandwidth = 2604 Hz/pixel.
fMRI data analysis. Imaging data were analyzed using
SPM8. For preprocessing, functional and anatomical
images were realigned, coregistered to the structural scan,
and normalized using the DARTEL procedure in SPM8.
For each participant, the 23 s of self-affirmation deci-
sion-making trials were modeled as the self-affirmation
blocks, and the 23 s of control trials were modeled as the
at UCLA on April 13, 2016pss.sagepub.comDownloaded from

462 Dutcher et al.
control blocks. Implicit baseline consisted of the rest peri-
ods (viewing a fixation cross).
We computed linear contrasts comparing the self-
affirmation trials with the control trials for each participant.
These individual contrast images were then used in group-
level analyses. We examined activity within each of the
ROIs used in Study 1 for the self-affirmation trials com-
pared with the control trials. Parameter estimates repre-
senting the average activity across all voxels in the ROI
were extracted and averaged. Analyses were run in Mars-
Bar. A standard statistical threshold of p < .05 was used
forthese ROI analyses. CIs for these tests were estimated
using the bias-corrected and accelerated-percentile method
(10,000 random samples with replacement; implemented
using the BOOTCI function in MATLAB).
To supplement the ROI analyses, we performed whole
brain analyses on the self-affirmation compared with
control contrast. Whole-brain analyses used an explor-
atory threshold (p < .005, k = 20; Lieberman & Cunning-
ham, 2009). All coordinates are reported in MNI space.
Results
ROI analyses. As in Study 1, compared with the control
trials, the self-affirmation trials produced significantly
more left VS activity, t(19) = 1.84, p = .041, mean param-
eter estimate = 0.261, 95% CI = [0.001, 0.548], and mar-
ginally more right VS activity, t(19) = 1.62, p = .061, mean
parameter estimate = 0.263, 95% CI = [−0.046, 0.560]
(Fig.3). As in Study 1, there was no difference in VTA
activity between the self-affirmation and control trials,
t(19) = 0.72, p = .240, mean parameter estimate = 0.190,
95% CI= [−0.111, 0.597] (Fig. 3).
Whole-brain analyses. As in Study 1, results from the
whole-brain analyses revealed a significant cluster in the
VS (putamen) as well as in the medial prefrontal cortex
(for a full list of activations, see Table 2). Results from
the posterior cingulate cortex at this threshold did not
reachstatistical significance. The control condition > self-
affirmation condition contrast revealed no significant
clusters of activation.
Discussion
Reflecting on important personal values during self-
affirmation activated neural reward pathways (VS) across
two studies with different age groups, using either
personal-value pictures with text or text only. VS activa-
tion was not due simply to making judgments about pref-
erences or personal values; rather, it was specific to
thinking about one’s most important personal value. This
is the first fMRI study to identify neural systems engaged
during self-affirmation, which extends self-affirmation
theory by suggesting that self-affirmation is rewarding.
This is a first step toward identifying the neural mecha-
nisms by which self-affirmation reduces threat and stress
responding, improves performance, reduces defensive-
ness, and alters social and health behaviors.
The VS is a key region in the mesolimbic dopamine
reward pathway, which suggests that affirming important
personal values is rewarding and may lead to a cascade
of effects associated with reward processing. Indeed,
when we investigated the term “reward” using Neuro-
synth (http://neurosynth.org/), a large-scale database of
neuroimaging studies that provides meta-analytic reverse-
inference analyses, the z score identified in the VS by
0.0
0.1
0.2
0.3
0.4
0.5
Left Ventral Striatum Right Ventral Striatum Ventral Tegmental Area
Parameter Estimate
†
*
Fig. 3. Results for Study 2: parameter estimates for activity in the left ventral striatum, the
right ventral striatum, and the ventral tegmental area during experimental trials. Error bars
represent ±1 SE. Symbols indicate the results of tests comparing the experimental trials with
the control trials (†p < .10, *p < .05).
at UCLA on April 13, 2016pss.sagepub.comDownloaded from

A Possible Reward-Related Mechanism for Self-Affirmation 463
Neurosynth suggested that neural activity reliably indi-
cated reward processing. In fact, the z score for inferring
that activity in the VS is indicative of reward (z = 26.11)
was higher than the z score for inferring that activity in
the occipital cortex was indicative of vision (z = 13.36) or
that activity in the amygdala was indicative of affect (z =
6.41), fear (z = 13.10), or emotion (z = 18.01). These find-
ings from Neurosynth provide further support for our
evidence that self-affirmation elicits reward-related
processing.
These two studies used different populations, which
suggests that the neural correlates of self-affirmation
extend from undergraduates to community-dwelling
adults. This is consistent with the wealth of previous
research on self-affirmation, which has found benefits for
a range of ages and ethnicities (Cohen & Sherman, 2014).
Study 2’s self-affirmation task differed from the one used
in Study 1 in that it used a different list of personal values
and no pictures, but it yielded similar effects. This pro-
vides converging evidence that increased VS activity dur-
ing self-affirmation is not specific to viewing personal-
value pictures. Rather, the VS is sensitive to thinking
about one’s most important value. Although the differ-
ence in VS activity between the self-affirmation and
yoked control conditions was statistically significant for
the left VS but not quite significant for the right VS, the
pattern was in the same direction for both regions.
Whole-brain results showed that, in addition to acti-
vating the VS, self-affirmation led to greater activity in
regions typically associated with thinking about the self
(the precuneus in Study 1; the medial prefrontal cortex in
Studies 1 and 2). Indeed, self-affirmation requires partici-
pants to reflect on their preferences, which makes these
findings unsurprising. However, self-affirmation also led
to greater activity in self-processing regions compared
with its yoked control, which still required participants to
think about their own preferences. The difference was
that the self-affirmation experimental condition had par-
ticipants think about their top-ranked personal value.
Self-affirmation may lead to a deeper processing of self-
related information, which may also be a key ingredient
in self-affirmation’s effects on behavior.
The present studies employed stringent control condi-
tions. For the self-affirmation task, participants received
the same instructions for both conditions; the only differ-
ence was the participant’s prior ratings of the personal
values shown. This allowed us to conclude that it was not
preference judgments about important values per se that
activated reward circuitry; rather, it was about focusing
on one’s most important personal values. Whereas many
standard self-affirmation manipulations use a control
condition in which participants think about why a less
important value might be important to someone else, our
design provided a more specific test of the neural activity
involved in the self-affirmation condition, providing
insight into a possible mechanism for self-affirmation.
One important question for future studies is whether
this neural-reward account of self-affirmation can explain
the subsequent cascade of neural and psychological
effects observed in previous studies (Creswell etal., 2005;
Falk et al., 2015; Legault et al., 2012; Sherman et al.,
2000). Falk etal. (2015) showed that activity in the ven-
tromedial prefrontal cortex in response to health mes-
sages was greater in participants who performed a
self-affirmation task than in participants who performed
a control task, and their findings may be consistent with
our findings. Specifically, the ventral striatum shows
functional connectivity with the ventromedial prefrontal
Table 2. Results From Study 2: Brain Regions More Active During Self-Affirmation Than During Control
Trials
Anatomical region
Brodmann’s
area Hemisphere
MNI coordinates
t(20)
Number of
voxels (k) x y z
Ventrolateral prefrontal cortex 47 Left –51 27 –3 4.14 135
Ventral striatum or putamenaLeft –21 21 –6 3.63
Ventrolateral prefrontal cortex 47 Right 33 27 –9 3.65 24
Medial prefrontal cortex 10 Right 9 60 18 3.74 27
Rostrolateral prefrontal cortex 10 Right 21 63 9 4.18 30
Dorsolateral prefrontal cortex 46 Right 45 30 21 3.82 62
Dorsomedial prefrontal cortex 9 Left –12 51 21 3.36 24
Supplementary motor area 6 — –6 24 54 4.13 276
Angular gyrus 40 Left –54 –66 39 3.74 24
Note: Activations were measured relative to activation in the yoked control trials. The table reports significant
activations (p < .005) of clusters with a minimum size of 20 voxels. The t tests were conducted at peak coordinates.
MNI = Montreal Neurological Institute.
aThis activation extended from the larger cluster listed in the previous row.
at UCLA on April 13, 2016pss.sagepub.comDownloaded from

464 Dutcher et al.
cortex (Di Martino etal., 2008). It is possible, then, that
VS activity during self-affirmation activates a cascade of
neural effects, including increased ventromedial prefron-
tal cortex activity, that in turn lead to psychological and
behavioral changes. Our findings could also be consis-
tent with the finding that self-affirmation leads to greater
neurophysiological error-related negativity during sub-
sequent tasks (Legault et al., 2012). Specifically, there
is evidence that reward and positive affect increase cor-
relates of error-related negativity event-related potentials,
which might relate to improved conflict adaptation
(Larson, Perlstein, Stigge-Kaufman, Kelly, & Dotson, 2006;
Stürmer, Nigbur, Schacht, & Sommer, 2011). Therefore, it
is plausible that self-affirmation activates reward process-
ing, which increases error-related negativity and causes a
shift in strategy that leads to improved performance
(Legault etal., 2012). Further research should investigate
the subsequent neural and behavioral processes that fol-
low self-affirmation’s reward activity.
There are a few limitations to these studies. In Study 1,
we chose to have participants think about the toaster attri-
butes most important to an average college student to
ensure this task was not inadvertently affirming. Future
work should have participants make decisions that are
important to them, but not in the values domain. In Study
1, the experimental and control conditions were in sepa-
rate runs, which is nonstandard for fMRI studies. Because
self-affirmation findings sometimes show lasting effects
(Cohen etal., 2006), this design was selected to provide
participants with a break to prevent carryover. Study 2
employed a more typical design with blocks of affirmation
and control randomly ordered in one run. However, the
results were the same for both studies. In these studies,
self-affirmation was manipulated via values affirmation.
Although this is the most common self-affirmation opera-
tionalization used, it is not the only one. Our results may
be specific to this values-affirmation procedure, and future
studies could determine whether all self-affirmation opera-
tionalizations rely on similar neural substrates.
These findings provide insight into the neural mecha-
nism by which self-affirmation reduces threat respond-
ing. Recent work has found that rewarding stimuli (e.g.,
sweet foods, sweet drinks, or sexual stimuli) lead to
reductions in stress responding (Creswell, Pacilio, etal.,
2013; Ulrich-Lai etal., 2010). This effect extends to social
rewards also, given that social support activates reward
regions, such as the VS (Inagaki & Eisenberger, 2013;
Strathearn, Fonagy, Amico, & Montague, 2009), and has
been shown to reduce threat-related neural activity
(Eisenberger etal., 2011; Younger, Aron, Parke, Chatter-
jee, & Mackey, 2010). It is possible that self-affirmation
relies on similar neural mechanisms to reduce threat
responding. In the present article, self-affirmation
(vs. control) led to greater VS activity, which could
correspond with activation decreases in neural threat
regions during subsequent tasks. In turn, this could con-
tribute to the array of threat reduction benefits that self-
affirmation has been shown to foster. Future studies
should assess whether this proposed reward-system
mechanism underlies the stress-buffering effects shown
in previous self-affirmation studies. Although we suggest
a reward-related mechanism, it is possible that the
reward associated with self-affirmation may be distinct
from nonself-related reward (e.g., food, winning money).
Future studies could investigate the possible distinct and
overlapping neural reward circuitry underlying different
reward processes.
Author Contributions
All authors contributed to the concept and to the design of the
studies. J. M. Dutcher, L. E. Pacilio, and K. A. Muscatell contrib-
uted to testing and data collection. J. M. Dutcher performed the
data analysis. J. M. Dutcher, J. D. Creswell, and N. I.
Eisenberger interpreted the data. J. M. Dutcher, J. D. Creswell,
P. R. Harris, W. M. P. Klein, J. M. Levine, and N. I. Eisenberger
drafted the manuscript. All authors approved the final version
of the manuscript for submission.
Acknowledgments
We thank Deborah Garet and Ivana Jevtic for recruiting partici-
pants, Jared Torre for assisting with data analysis, and UCLA’s
Staglin Center for Cognitive Neuroscience and Carnegie
Mellon’s Scientific Imaging & Brain Research Center for assist-
ing with data collection.
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with
respect to their authorship or the publication of this article.
Funding
This work was supported by National Science Foundation
Grant SES-0648044 (to W. M. P. Klein, P. R. Harris, J. D.
Creswell, and J. M. Levine), by Breast Cancer Research Founda-
tion funding (to J. E. Bower), and by funds supporting the
Wendell Jeffrey & Bernice Wenzel Term Chair in Behavioral
Neuroscience (to N. I. Eisenberger).
Supplemental Material
Additional supporting information can be found at http://pss
.sagepub.com/content/by/supplemental-data
Note
1. Participants also completed another self-affirmation task in
which they thought about important values for 23 s (this task
is similar to self-affirmation manipulations involving writing;
Cohen, Aronson, & Steele, 2000). However, because the neu-
roimaging results suggested this manipulation was not success-
ful, we do not focus on this task here. The blocks in this task
at UCLA on April 13, 2016pss.sagepub.comDownloaded from

A Possible Reward-Related Mechanism for Self-Affirmation 465
were modeled separately from the blocks in the task of interest.
Details of this task procedure and results are included in the
Supplemental Material.
References
Aron, A., Fisher, H., Mashek, D. J., Strong, G., Li, H., & Brown,
L. L. (2005). Reward, motivation, and emotion systems
associated with early-stage intense romantic love. Journal
of Neurophysiology, 94, 327–337.
Brett, M., Anton, J. L., Valabregue, R., & Poline, J. B. (2002).
Region of interest analysis using the MarsBar toolbox for
SPM 99. NeuroImage, 16, S497.
Burson, A., Crocker, J., & Mischkowski, D. (2012). Two types
of value-affirmation: Implications for self-control following
social exclusion. Social Psychological & Personality Science,
3, 510–516.
Cohen, G. L., Aronson, J., & Steele, C. M. (2000). When beliefs
yield to evidence: Reducing biased evaluation by affirm-
ing the self. Personality and Social Psychology Bulletin, 26,
1151–1164.
Cohen, G. L., Garcia, J., Apfel, N., & Master, A. (2006). Reducing
the racial achievement gap: A social-psychological inter-
vention. Science, 313, 1307–1310.
Cohen, G. L., & Sherman, D. K. (2014). The psychology of
change: Self-affirmation and social psychological interven-
tion. Annual Review of Psychology, 65, 333–371.
Correll, J., Spencer, S. J., & Zanna, M. P. (2004). An affirmed self
and an open mind: Self-affirmation and sensitivity to argu-
ment strength. Journal of Experimental Social Psychology,
40, 350–356.
Creswell, J. D., Dutcher, J. M., Klein, W. M. P., Harris, P. R.,
& Levine, J. L. (2013). Self-affirmation improves prob-
lem-solving under stress. PLoS ONE, 8(5), Article e62593.
doi:10.1371/journal.pone.0062593
Creswell, J. D., Pacilio, L. E., Denson, T. F., & Satyshur, M.
(2013). The effect of primary sexual reward manipula-
tion on cortisol responses to psychosocial stress in men.
Psychosomatic Medicine, 75, 397–403.
Creswell, J. D., Welch, W., Taylor, S. E., Sherman, D. K.,
Gruenewald, T., & Mann, T. (2005). Affirmation of personal
values buffers neuroendocrine and psychological stress
responses. Psychological Science, 16, 846–851.
Crocker, J., Niiya, Y., & Mischkowski, D. (2008). Why does
writing about important values reduce defensiveness? Self-
affirmation and the role of positive other-directed feel-
ings. Psychological Science, 19, 740–747.
Di Martino, A., Scheres, A., Margulies, D. S., Kelly, A. M.C.,
Uddin, L. Q., Shehzad, Z., . . . Milham, M. P. (2008).
Functional connectivity of human striatum: A resting state
fMRI study. Cerebral Cortex, 18, 2735–2747.
Eisenberger, N. I., Master, S. L., Inagaki, T. K., Taylor, S. E.,
Shirinyan, D., Lieberman, M. D., & Naliboff, B. D. (2011).
Attachment figures activate a safety signal-related neural
region and reduce pain experience. Proceedings of the
National Academy of Sciences, USA, 108, 11721–11726.
Falk, E. B., O’Donnell, M. B., Cascio, C. N., Tinney, F., Kang, Y.,
Lieberman, M. D., . . . Strecher, V. J. (2015). Self-affirmation
alters the brain’s response to health messages and
subsequent behavior change. Proceedings of the National
Academy of Sciences, USA, 112, 1977–1982. doi:10.1073/
pnas.1500247112
Heatherton, T. F., Wyland, C. L., Macrae, C. N., Demos, K. E.,
Denny, B. T., & Kelley, W. M. (2006). Medial prefrontal
activity differentiates self from close others. Social Cognitive
and Affective Neuroscience, 1, 18–25.
Inagaki, T. K., & Eisenberger, N. I. (2012). Neural correlates
of giving support to a loved one. Psychosomatic Medicine,
74, 3–7.
Klein, W. M. P., & Harris, P. R. (2009). Self-affirmation enhances
attentional bias toward threatening components of a per-
suasive message. Psychological Science, 20, 1463–1467.
Koole, S. L., Smeets, K., Van Knippenberg, A., & Dijksterhuis,
A. (1999). The cessation of rumination through self-affir-
mation. Journal of Personality and Social Psychology, 77,
111–125.
Larson, M. J., Perlstein, W. M., Stigge-Kaufman, D., Kelly, K. G., &
Dotson, V. M. (2006). Affective context-induced modulation
of the error-related negativity. NeuroReport, 17, 329–333.
Legault, L., Al-Khindi, T., & Inzlicht, M. (2012). Preserving
integrity in the face of performance threat: Self-affirmation
enhances neurophysiological responsiveness to errors.
Psychological Science, 23, 1353–1360.
Lieberman, M., & Cunningham, W. A. (2009). Type I and Type
II error concerns in fMRI research: Re-balancing the scale.
Social Cognitive and Affective Neuroscience, 4, 423–428.
Nelson, S. K., Fuller, J. A., Choi, I., & Lyubomirsky, S. (2014).
Beyond self-protection: Self-affirmation benefits hedonic
and eudaimonic well-being. Personality and Social
Psychology Bulletin, 40, 998–1011.
Sherman, D. A. K., Nelson, L. D., & Steele, C. M. (2000). Do
messages about health risks threaten the self? Increasing
the acceptance of threatening health messages via self-
affirmation. Personality and Social Psychology Bulletin, 26,
1046–1058.
Sherman, D. K., & Cohen, G. L. (2006). The psychology of
self-defense: Self-affirmation theory. In M. P. Zanna (Ed.),
Advances in experimental social psychology (Vol. 38, pp.
183–242). San Diego, CA: Academic Press.
Sherman, D. K., & Hartson, K. A. (2011). Reconciling self-
protection with self-improvement: Self-affirmation theory.
In M. Alicke & C. Sedikides (Eds.), The handbook of self-
enhancement and self-protection (pp. 128–151). New York,
NY: Guilford Press.
Speer, M. E., Bhanji, J. P., & Delgado, M. R. (2014). Savoring the
past: Positive memories evoke value representations in the
striatum. Neuron, 84, 847–856.
Steele, C. M. (1988). The psychology of self-affirmation:
Sustaining the integrity of the self. In L. Berkowitz (Ed.),
Advances in experimental social psychology (Vol. 21, pp.
261–302). San Diego, CA: Academic Press.
Steele, C. M., & Liu, T. J. (1983). Dissonance processes as self-
affirmation. Journal of Personality and Social Psychology,
45, 5–19.
Stinson, D. A., Logel, C., Shepherd, S., & Zanna, M. P. (2011).
Rewriting the self-fulfilling prophecy of social rejection: Self-
affirmation improves relational security and social behavior
up to 2 months later. Psychological Science, 22, 1145–1149.
at UCLA on April 13, 2016pss.sagepub.comDownloaded from

466 Dutcher et al.
Strathearn, L., Fonagy, P., Amico, J., & Montague, P. R. (2009).
Adult attachment predicts maternal brain and oxytocin
response to infant cues. Neuropsychopharmacology, 34,
2655–2666.
Stürmer, B., Nigbur, R., Schacht, A., & Sommer, W. (2011).
Reward and punishment effects on error processing and
conflict control. Frontiers in Psychology, 2, Article 335.
doi:10.3389/fpsyg.2011.00335
Tamir, D. I., & Mitchell, J. P. (2012). Disclosing information
about the self is intrinsically rewarding. Proceedings of the
National Academy of Sciences, USA, 109, 8038–8043.
Tzourio-Mazoyer, N., Landeau, B., Papathanassiou, D.,
Crivello, F., Etard, O., Delcroix, N., . . . Joliot, M. (2002).
Automated anatomical labeling of activations in SPM using
a macroscopic anatomical parcellation of the MNI MRI
single-subject brain. NeuroImage, 15, 273–289.
Ulrich-Lai, Y. M., Christiansen, A. M., Ostrander, M. M., Jones,
A. A., Jones, K. R., Choi, D. C., . . . Herman, J. P. (2010).
Pleasurable behaviors reduce stress via brain reward path-
ways. Proceedings of the National Academy of Sciences,
USA, 107, 20529–20534.
Vernon, P. E., & Allport, G. W. (1931). A test for personal val-
ues. The Journal of Abnormal and Social Psychology, 26,
231–248.
Younger, J., Aron, A., Parke, S., Chatterjee, N., & Mackey, S.
(2010). Viewing pictures of a romantic partner reduces
experimental pain: Involvement of neural reward systems.
PLoS ONE, 5(10), e13309. doi:10.1371/journal.pone.0013309
at UCLA on April 13, 2016pss.sagepub.comDownloaded from