The “social brain” is highly sensitive to the mere presence of social information: An automated meta-analysis and an independent study
The ªsocial brainº is highly sensitive to the mere presence of social information: An automated meta-analysis and an independent study
Ivy F. Tso 0 1 2
Saige Rutherford 0 1 2
Yu Fang 0 1 2
Mike Angstadt 0 1 2
Stephan F. Taylor 0 1 2
0 Funding: This research was supported by the National Institute of Mental Health (R01 MH064148 to S.F.T.; K23 MH108823 to I.F.T.). There was no
1 Department of Psychiatry, University of Michigan Medical School , Ann Arbor , Michigan, United States of America, 2 Department of Psychology, University of Michigan , Ann Arbor, Michigan , United States of America
2 Editor: Satoru Hayasaka, University of Texas at Austin , UNITED STATES
How the human brain processes social information is an increasingly researched topic in psychology and neuroscience, advancing our understanding of basic human cognition and psychopathologies. Neuroimaging studies typically seek to isolate one specific aspect of social cognition when trying to map its neural substrates. It is unclear if brain activation elicited by different social cognitive processes and task instructions are also spontaneously elicited by general social information. In this study, we investigated whether these brain regions are evoked by the mere presence of social information using an automated meta-analysis and confirmatory data from an independent study of simple appraisal of social vs. non-social images. Results of 1,000 published fMRI studies containing the keyword of ªsocialº were subject to an automated meta-analysis (http://neurosynth.org). To confirm that significant brain regions in the meta-analysis were driven by a social effect, these brain regions were used as regions of interest (ROIs) to extract and compare BOLD fMRI signals of social vs. non-social conditions in the independent study. The NeuroSynth results indicated that the dorsal and ventral medial prefrontal cortex, posterior cingulate cortex, bilateral amygdala, bilateral occipito-temporal junction, right fusiform gyrus, bilateral temporal pole, and right inferior frontal gyrus are commonly engaged in studies with a prominent social element. The social±non-social contrast in the independent study showed a strong resemblance to the NeuroSynth map. ROI analyses revealed that a social effect was credible in 9 out of the 11 NeuroSynth regions in the independent dataset. The findings support the conclusion that the ªsocial brainº is highly sensitive to the mere presence of social information.
Social cognitionÐthe cognitive processes involved in processing information about self, other
people, interpersonal relationships, and social interactions [
]Ðis critical to social
development and adaptation. Understanding how the human brain processes social information has
been an increasingly important topic in psychology and neuroscience. Research in this area
not only has increased our knowledge of the functional specialization and organization of the
healthy brain, but also provides a promising avenue to uncover the pathogenesis of complex
neuropsychiatric disorders in which abnormal social information processing is a prominent
feature, such as schizophrenia [
] and autism [
Social cognition comprises many cognitive processes, including perception of socially
relevant cues (faces, eye gaze, facial expressions, prosody, body movements and gesture),
understanding and making inferences about others' mental state, forming judgments of others, and
reflection on the self and its relation to others. Several brain regions are often implicated in
social information processing, including the medial prefrontal cortex (mPFC), superior
temporal sulcus/gyrus (STS/STG), fusiform gyrus, temporo-parietal junction (TPJ), temporal pole,
precuneus/posterior cingulate cortex (PCC), and amygdala [4±6]. The wealth of research data
in this area has provided some clues about the roles of these brain regions in social information
processing, although delineating their fine-grained functional specializations is still an active
topic of investigation. Some conclusions about specialized social functions have emerged from
the literature. For example, the mPFC appears to be associated with forming
meta-representations of the self and the mental states of other people [
]; lesions in this region result in deficits
in interpreting nonverbal social information, recognizing social faux pas and sarcasm, and
showing empathic concern for others [8±11]. Activity in the posterior STS/STG, fusiform
gyrus, and anterior temporal cortex is elicited by face and eye gaze processing [
observation of biological motion , and inferring intentions from others' actions [
in the TPJ is associated with both mental and spatial perspective taking [
understanding false beliefs [
]. The PCC is thought to integrate emotional and autobiographical
memory in the personal context during self-referential information processing . The
amygdala is involved in making judgments about faces and shows increased activation to
untrustworthy relative to trustworthy faces .
In order to map neural substrates of social cognition, neuroimaging researchers typically
design a task which seeks to isolate one specific process, such as theory of mind [
], leading to
activation patterns thought to be specific to that process. In this study, we took a slightly
different approach, namely to address the question as to whether these brain regions are evoked by
the mere presence of social information. One possibility is that very specific tasks are required to
activate these regions; whereas the alternative possibility is that the mere presence of social
information, regardless of task instructions, is sufficient to activate these areas. One might
expect the latter case, given that social processing is so robust and deeply programmed into the
cognitive-perceptual machinery that they are easily `turned on.' To test the hypothesis that
most of the regions described above are more generally dedicated to social information, we
employed a two-step strategy. Using an automated meta-analysis, NeuroSynth (neurosynth.
org; 30), we first identified the brain regions found in published neuroimaging studies that
contained the word `social' in the text at a prominent frequency (> 1 in 1000 words),
regardless of specific tasks, instructions, and contrasts. This affords us an inclusive and
comprehensive picture of the ªsocial brain regionsº commonly appearing in the literature, but elicited by
a variety of tasks and paradigms (e.g., facial emotion discrimination, social and moral
judgment, theory of mind, social exclusion, gambling task, rewarding processing, response
inhibition) and stimuli (e.g., faces, geometric shapes, cartoon strips, vocal sounds, speech). In a
second step, we conducted an independent neuroimaging study in which subjects viewed
complex visual images depicting social and non-social scenes of varying emotional valence.
Subjects performed a simple valence appraisal task in which the socialness of the stimuli was
manipulated while holding constant other stimuli characteristics (e.g., visual complexity,
valence, arousal). Although this task involved specific cognitive processes related to valence
2 / 13
appraisal, by contrasting the social and non-social conditions, we were able to isolate brain
activation specific to the mere presence of information that is social in nature. We
hypothesized that a majority of the brain regions identified using the NeuroSynth results would show
preferential activation for social (as opposed to non-social) images in the independent study,
providing support for the hypothesis that ªthe social brainº is very sensitive to the mere
presence of social information.
Materials and methods
The first part of the analyses of this study aimed to identify the brain regions that have shown
significant activation in published fMRI studies with a prominent social element in the
literature. Using the keyword ªsocialº yielded 1,000 published fMRI studies to include in an
automated meta-analysis on neurosynth.org (http://neurosynth.org/analyses/terms/social/) [
We used the reverse inference map of the result of the automated meta-analysis, which
represents z-scores corresponding to the likelihood that the term ªsocialº is used in a study given
the presence of reported activation (i.e., P[Social|Activation]). It is obtained by comparing all
the studies in the Neurosynth database that contain ªsocialº and those that do not. The
significant brain regions showing up in the reverse inference map represent those that are more
likely to be reported in ªsocialº studies than in ªnon-socialº studies. In contrast, the forward
inference map (P[Activation|Social]) does not consider the base-rate activation of the regions,
and the result may very well include regions that are involved in almost every task. As such,
the reverse inference map is a better indicator of how specific the activated regions are to social
information processing. The activation map was thresholded at FDR-corrected p < 0.01 by
default, yielding 135 significant clusters. The majority of the clusters were very small in terms
of voxel size (< 10 voxels). Eleven clusters were 100 voxels in size and were selected to
represent ªsocial brain regions in the literature.º Since the resolution of the NeuroSynth map was
higher than that of the independent dataset, the NeuroSynth results were downsampled from a
voxel size of 2 × 2 × 2 mm3 to 3 × 3 × 3 mm3 to facilitate later comparisons. Masks derived
from the 11 regions served as the regions of interest (ROIs) for beta extraction for the
Participants. Fifteen healthy participants were recruited from community advertisements
and completed the study. All participants were free of Axis I psychiatric disorders as established
with the Structured Clinical Interview for Diagnosis, non-patient version (SCID-NP)  and
were not taking any medications. The risks of the study were explained to all participants prior
to obtaining their written, informed consent to participate. The study was conducted in
accordance to the study protocol with ethical standards in line with the Declaration of Helsinki and
approved by the University of Michigan Medical School Institutional Review Board (IRBMED),
IRBMED# 2001±0283. One participant's fMRI data were lost due to archival errors; data of the
remaining 14 (4 female) participants, aged from 23 to 50 years (mean = 38.6, SD = 10.1), were
included in the analyses of this report. A previous peer-reviewed publication reported on
different aspects of this sample using this paradigm .
Visual stimuli. One hundred and twenty (including 60 social and 60 non-social) complex
images were selected from the International Affective Picture System (IAPS) . Image
selection began with identifying images that contained the presence of human and/or interactions
between social animals as candidate social images, and those that contained landscapes or
physical objects only as candidate non-social images. Candidate images were then classified as
3 / 13
negative, neutral, and positive based on their normative valence ratings. Finally, 20 negative,
20 neutral, and 20 positive images were selected for each of the social and non-social
conditions such that the two conditions were matched on valence and arousal based on the
normative ratings associated with each image. Some examples of these images included: a gory face
(social, negative), a soiled toilet (non-social, negative), a man facing a computer monitor
(social, neutral), a bus (nonsocial, neutral), two children playing with cats (social, positive),
and a colorful flower field (nonsocial, positive). Equivalent valence and arousal of the social
and non-social images was later confirmed using the subjective ratings by the participants in
the independent study (see Fig 1A and more details of procedure below). A complete list of the
IAPS images used in the independent study can be found in supporting information S1 Table.
In addition to the IAPS images, ªblankº (BL) images were included as a baseline condition.
They were 4 unique kinds of images composed of a colored polygon against a lightly textured,
gray-toned background of varying shades. The contrast and brightness of each set of the
images were adjusted to match on total luminance using Photoshop 4.0 (Adobe Systems).
Appraisal task: Design and procedure. The images were presented in 20-second blocks;
each block consisted of 4 images and each image was presented for 5 seconds. For each image,
participants were instructed to form a judgment as to whether it was pleasant, neutral, or
unpleasant, and to press a button to signal that they had formed a judgment. Appraisal
duration for social and non-social images did not differ significantly (Fig 1B). The task consisted of
a total of 30 blocks of IAPS images divided into 5 runs. Blocks of IAPS images alternated with
blocks of BL images. Participants completed a practice session before the fMRI scanning to
ensure comprehension of the task. Participants' attention was monitored using an eye tracker
in the scanner.
Immediately after the fMRI session, participants viewed all of the IAPS images again on a
computer outside of the scanner, presented in a randomized order, and rated each image for
valence and arousal on a 7-point scale. For valence, the prompt question was ªHow pleasant or
unpleasant does this picture make you feel?º and participants chose a number between 1 and
7, with 1 = ªExtremely unpleasant,º 2 = ªVery unpleasant,º 3 = ªMildly unpleasant,º 4 =
ªNeither,º 5 = ªMildly pleasant,º 6 = ªVery pleasant,º and 7 = ªExtremely pleasant.º For arousal,
the prompt question was ªHow calm or excited/aroused does this picture make you feel?º and
participants chose a number between 1 and 7, which was accompanied by a description
ªCalm, not aroused/excited ! A little ! Moderately ! Very ! Extremely aroused/excited.º
Social and non-social images were similar in both valence and arousal ratings, Fig 1A.
fMRI acquisition and processing. MRI scanning occurred on a GE 3T Signa scanner
(LX [8.3] release, General Electric Healthcare, Buckinghamshire, United Kingdom). A
T1-weighted image was acquired in the same prescription as the functional images to facilitate
co-registration. Functional images were acquired with a T2 -weighted, reverse spiral
acquisition sequence (gradient recalled echo, TR = 2000 ms, TE = 30 ms, FA = 90 degrees, field of
view = 20 cm, 40 slice, 3.0mm thick/0mm skip, equivalent to 64 x 64 voxel grid) sensitive to
signal in ventral medial frontal regions . Subjects underwent 5 runs (6 blocks/runs), each
consisting of 120 volumes, plus 4 initial, discarded volumes to allow for equilibration of
scanner signal, with isotropic voxels 3 mm after normalization. After acquisition of functional
volumes, a high resolution T1 scan (3D SPGR, field of view = 24 cm, TR = 25 ms, TE = 3 ms,
256 × 160 matrix, 100 slices, 1.5 mm interleaved with no skip) was obtained for anatomic
fMRI data were preprocessed with the Statistical Parametric Mapping (SPM8) package
(Wellcome Institute of Cognitive Neurology, London) and FSL (FMRIB, Oxford, UK) and
standard routines. Slice time was corrected using sinc-interpolation, weighted by a Hanning
kernel in time. Then all scans were realigned to the 10th volume acquired during each scan
4 / 13
Fig 1. Characteristics of the social and non-social images used in this study. a) Subjective valance (p = .647) and arousal
ratings (p = .464) by the participants did not differ significantly between the social and non-social conditions. b) Image
appraisal time in the scanner did not differ significantly between the social and non-social conditions (p = .856). Vertical
lines represent standard errors of mean.
5 / 13
("mcflirt") . Runs with movement exceeding either 1 voxel or 2 degrees rotation within a
scan were discarded; only 1 run of 1 subject was discarded as a result. The time series of
functional volumes were then co-registered with the high resolution T1 image, spatially normalized
to the MNI152 brain, and then spatially smoothed with a 6 mm isotropic Gaussian kernel.
Statistical analyses. fMRI data analyses were performed with SPM12. First-level analysis
began with applying a high pass filter (128 s) to the anatomically normalized time series, and
regressed on 2 regressors (social, non-social) convolved with a canonical hemodynamic
response function, along with 24 motion regressors (6 for each translation/rotation direction,
their first derivative, and quadratic terms for each direction and derivative). BL blocks were
modeled as implicit baseline.
Second-level analyses involved both whole-brain and ROI analyses. The former informs the
brain regions preferentially responding to social vs. non-social stimuli; the latter reveals the
extent to which ªsocial brain regionsº seen in the literature are engaged in processing the social
nature of stimuli.
For the whole-brain analysis, the t statistics map of the social±nonsocial contrast was
examined. Initial clusters were defined by a voxel threshold of uncorrected p < .005; ªsignificantº
clusters were determined by a threshold of false discovery error (FDR) corrected p < .05 based
on the Gaussian random field theory . Subsequently, a conjunction analysis was performed
to show the overlap between social networks identified in the NeuroSynth result and our data,
by first binarizing supra-threshold voxels of the NeuroSynth map and the social±nonsocial
contrast map of our data, and then finding the voxels that were above threshold in both maps.
The ROI analyses involved extracting beta estimates of the social and non-social conditions
in the independent study from the 11 NeuroSynth-informed ROIs. This was done by saving
each of the Neurosynth ROIs into separate masks, and then applying the masks to the
firstlevel results of the social (vs. baseline) and non-social (vs. baseline) contrasts in the
independent study. The first eigenvector of beta estimates from these ROIs was extracted and subject
to Bayesian inference. Specifically, we used the anovaBF command of the R package
ªBayesFactorº  to compare evidence of two competing modelsÐa model containing Socialness as
a fixed factor and a null modelÐfor each brain region given the data; subjects were modeled as
a random factor in both models. Relative evidence strength of the two models was expressed in
Bayes factor, such that a value < 1 indicates evidence favoring the denominator model (null
model) over the numerator model (Social effect model), whereas a value > 1 indicates evidence
favoring the numerator model over the denominator model. Further, interpretation of strength
of evidence followed guidelines by Jeffreys [
], where Bayes factors between 3 and 10 indicate
that the support for the Social effect model is ªsubstantial,º values between 10 and 30 ªstrong,º
values between 30 and 100 ªvery strong,º and values > 100 ªdecisiveº; similarly, values
between 0.10 and 0.33 indicate ªsubstantialº support for the null model, values between 0.033
and 0.10 ªstrong,º values between 0.01 and 0.033 ªstrongº support, and values < 0.01
From the NeuroSynth data, 11 clusters were identified as ªsocialº brain regions in the
literature: dorsomedial PFC (dmPFC), ventromedial PFC (vmPFC), PCC, bilateral amygdala, right
fusiform gyrus, bilateral OTJ, bilateral anterior temporal cortex/temporal pole, and right
inferior frontal gyrus (IFG) (Table 1). Please note that results are in reduced resolution of voxel
size 3 × 3 × 3 mm3 for easier comparison with the results of the independent dataset in
6 / 13
Center-of-mass coordinate (x, y, z)
-1.6, 54.2, 27.6
1.9, 47.2, -14.9
7 / 13
Fig 2. ªSocialº brain regions. Areas identified in the NeuroSynth meta-analysis result (yellow) and brain regions
showing preferential activation to social stimuli in our data (blue) showed remarkable overlap (green). Regions
significant in both our data and the NeuroSynth results are labeled in white, those significant only in NeuroSynth are
labelled in yellow. mPFC = medial prefrontal cortex; PCC = posterior cingulate cortex; L.Amyg = left amygdala; R.
Amgy = right amygdala; R.Fusi = right fusiform gyrus; L.OTJ = left occipito-temporal junction; R.OTJ = right
occipitotemporal junction; L.TempP = left temporal pole; R.TempP = right temporal pole; R.IFG = right inferior frontal gyrus.
Results of the social±nonsocial contrast of our data revealed that social images, compared
with nonsocial images, elicited significantly higher activation in a number of brain areas,
including mPFC extending across ventral and dorsal areas, PCC, bilateral amygdala extending
to hippocampus and anterior temporal cortex/temporal pole, bilateral OTJ extending to
fusiform gyri, right superior parietal cortex, and left superior/middle frontal gyrus (Table 2).
Overall, the social brain regions identified using the NeuroSynth data and our data showed
a strong resemblance. See Fig 2 for the results of the conjunction analysis, showing
simultaneously the NeuroSynth map and the social±nonsocial contrast of the independent study, as
well as their overlap.
Beta estimates of the social and non-social conditions in our data extracted from the
NeuroSynth-informed ROIs, and the results of statistical tests of a Social effect in these ROIs, are
displayed in Fig 3. In all 11 ROIs, social images elicited higher activation than non-social images.
Bayesian evidence favored the presence of a Social effect (Bayes factor > 1) in 9 out of the 11
regions (i.e., all but R IFG and L temporal pole). The evidence for a Social effect was
ªsubstantialº or stronger (Bayes factor > 3) in all of these 9 regionsÐdmPFC, vmPFC, PCC, R fusiform
gyrus, bilateral OTJ, R temporal pole, and bilateral amygdala.
This study investigated if there are sensitive, task-general modules for processing social
information in the human brain. We first examined the brain regions commonly activated in a
large number (N = 1,000) of published fMRI studies involving a prominent social element
using the automated meta-analysis method provided by NeuroSynth [
]. The NeuroSynth
map revealed distributed neural substrates related to social processing, including the ventral
and dorsal areas of the mPFC, precuneus/PCC, bilateral amygdala, bilateral OTJ (extending to
fusiform gyrus), bilateral anterior temporal cortex/temporal pole, and inferior frontal gyrus
extending to orbitofrontal cortex. This map is highly consistent with brain regions often
implicated in socio-emotional processing in the literature. Then we evaluated if these brain regions
are representative of social processing by conducting confirmatory analyses on an independent
dataset that specifically compared the socialness of the stimuli. By carefully matching the
8 / 13
Fig 3. BOLD signals in social and non-social conditions in the independent dataset in the 11 NeuroSynth ªsocialº brain regions. Bars (left y-axis) represent beta
estimates and vertical lines represent standard errors of mean. The line (right y-axis) indicates Bayes factor values comparing a model with Socialness as a fixed effect
(numerator) against a null model (denominator). dmPFC = dorsomedial prefrontal cortex; vmPFC = ventromedial prefrontal cortex; R.IFG = right inferior frontal
gyrus; PCC = posterior cingulate cortex; R.fusiform = right fusiform gyrus; L.OTJ = left occipito-temporal junction; R.OTJ = right occipito-temporal junction; L.
TempPole = left temporal pole; R.TempPole = right temporal pole; L.amyg = left amygdala; and R.amyg = right amygdala.
affective valence and levels of arousal of the images used in the social and non-social
conditions, we isolated social processing from other cognitive processes on brain activation.
Overlaying the results of the independent study on the NeuroSynth ªsocialº map showed a strong
correspondence of the two maps. ROI analyses examining brain activation in these
NeuroSynth regions in our data showed that a credible social effect (social > non-social) was present
in 9 out of 11 of these regions. Taken together, the results of this study provided convincing
support that a number of brain regions in the human brain are robustly and preferentially
activated when processing social information.
The similarities between the NeuroSynth map and the social±non-social contrast of the
independent dataset are remarkable given the differences in methods used to generate the two
maps. The NeuroSynth methods elicit very crude ªcontrastsº±the studies included were those
in which the term `social' appear in the article text at a ªhighº frequency (defined as > 1 in
every 1,000 words), and the coordinates that went into the meta-analysis were automatically
9 / 13
extracted from all tables reported in these studies, regardless of contrasts or (sub)groups. In
the independent study, participants were only given a vague task (to ªform a judgmentº of the
pleasantness of each of the images), rather than told explicitly to attend to the social aspect of
the images or to perform a specific social cognitive task. Additionally, the use of the social±
nonsocial contrast theoretically canceled out common cognitive processes (particularly,
valence appraisal) involved in the two conditions, making it reasonable to assume that the
result reflects brain activation associated with the sociality of the stimuli only. The results
showed that most of the brain regions from the NeuroSynth map were preferentially engaged
in response to the mere presence of sociality in stationary scenes of humans and social
interactions, consistent with the assertion that most social signals are processed nearly automatically
]. The strong correspondence between the NeuroSynth and the independent study suggests
that regardless of tasks and methods, certain cognitive processes are easily involved in
processing information social in nature: analysis of postures and biological motion (OTJ) ,
accessing social knowledge (temporal pole) , autobiographical recollection (PCC) , and
reflection on feelings and self-reference (mPFC) [
]. Further, the conjunction analysis showed
extensive overlap between the NeuroSynth map and the whole-brain analysis of the
independent dataset. Such overlapping regions may indicate subregions of the general social brain
areas that are sensitive to the degree of sociality.
Some brain regions from the NeuroSynth map did not show a credible social effect in the
independent data, such as the left temporal pole and right IFG. Additionally, some brain
regions that are often implicated in social cognitive processes (e.g., TPJ as involved in theory
of mind) did not show up in either the NeuroSynth map or our data. As noted in a review of
the social brain [
], brain regions involved in social cognition are modulated by the task
context and individual factors such as volitional regulation. The lack of a credible social effect in
the left temporal pole and the right IFG in the independent data could be due to that cognitive
processes recruiting the left temporal pole (e.g., semantic representation of sounds or objects)
and the IFG (e.g., response inhibition) may be prevalent among studies included in the
Neurosynth map but not required in the appraisal task in the independent study. Similarly, areas
such as TPJ did not appear in both the NeuroSynth and the independent study maps may be
because mental state attribution was not explicitly required in many of the studies included in
the automated meta-analysis. Therefore, the brain regions revealed in our dataset and the
NeuroSynth map should not be considered the complete ªsocial circuitryº given the simple task
used in this study and the variable representation of different social cognitive processes in the
literature. While the findings provide strong support for social brain modules such as mPFC,
PCC, right anterior temporal cortex, amygdala, and OTJ (extending to fusiform gyrus),
negative findings in TPJ and other regions do not mean that they are not involved in social
information processing. In a similar vein, the ªsocialº brain regions identified in this study should
not be interpreted as responsible for solely social information processing, as many (if not all)
brain regions are involved in multiple cognitive processes.
This study is limited by the small sample size of the independent dataset. Although we used
social and non-social stimuli carefully matched for valence and arousal, literature-informed
ROIs, and Bayesian statistics to increase the scientific rigor and the interpretability of the
results, we acknowledge that the results may be different with a larger or different sample. The
small sample also precludes the exploration of other important questions such as differences
between gender and diverse populations in the social brain network. Future investigations in
larger and cross-cultural studies to reveal critical biological and social factors in human social
cognition are warranted.
To conclude, this study provided support that core regions of the human social brain are
highly sensitive to the mere presence of social information, including the medial prefrontal
10 / 13
cortex, posterior cingulate cortex, temporal pole, and occipito-temporal junction extending to
fusiform gyrus. This knowledge may help guide future developmental and psychopathology
research. For example, tracking the qualitative and quantitative changes in this ªautomaticº
social brain over developmental or illness stages would inform whether such a neural
sensitivity to social information is innate, how it is associated with other important developmental
milestones and functional markers, how it is influenced by environmental and social factors
(e.g., poverty, abuse), and how its alterations may be responsible for the development and
symptom manifestations of different psychopathologies. Further, investigations of
high-resolution brain specialization as well as anatomical, functional, and effective brain connectivity
will help us gain a fuller understanding of the neural mechanisms of social information
processing and how the social brain network interacts with other brain systems to guide complex
social behavior in normal development and in psychopathologies with prominent social
deficits. Finally, the results of this study lend support to the usefulness of NeuroSynth in
neuroscience research, as it provides a relatively accurate picture of the neural substrates of a variety of
broadly conceived cognitive processes.
S1 Table. International Affective Picture System (IAPS) images used in the independent study.
Conceptualization: Ivy F. Tso, Stephan F. Taylor.
Data curation: Ivy F. Tso, Saige Rutherford, Yu Fang.
Formal analysis: Ivy F. Tso, Saige Rutherford, Yu Fang, Mike Angstadt.
Funding acquisition: Ivy F. Tso, Stephan F. Taylor.
Investigation: Stephan F. Taylor.
Methodology: Ivy F. Tso, Mike Angstadt, Stephan F. Taylor.
Project administration: Stephan F. Taylor.
Resources: Stephan F. Taylor.
Software: Ivy F. Tso, Saige Rutherford, Yu Fang, Mike Angstadt.
Supervision: Stephan F. Taylor.
Visualization: Ivy F. Tso, Saige Rutherford.
Writing ± original draft: Ivy F. Tso.
Writing ± review & editing: Ivy F. Tso, Saige Rutherford, Yu Fang, Mike Angstadt, Stephan F.
11 / 13
12 / 13
First MB, Spitzer RL, Gibbon M, Williams JBW. Structured Clinical Interview for DSM-IV Axis I
Disorders, Non-patient Edition (SCID-I/NP). New York: Biometrics Research, New York State Psychiatric
1. Frith CD . Social cognition . Philos Trans R Soc L B Biol Sci . 2008 ; 363 ( 1499 ): 2033 ± 2039 .
2. Green MF , Horan WP , Lee J . Social cognition in schizophrenia . Nat Rev Neurosci . 2015 ; 16 ( 10 ): 620 ± 31 . https://doi.org/10.1038/nrn4005 PMID: 26373471
3. Frith U . Mind blindness and the brain in autism . Neuron . 2001 ; 32 ( 6 ): 969 ± 79 . PMID: 11754830
4. Adolphs R. The social brain: neural basis of social knowledge . Annu Rev Psychol . 2009 ; 60 : 693 ± 716 . https://doi.org/10.1146/annurev.psych. 60 .110707.163514 PMID: 18771388
5. Brunet-Gouet E , Decety J . Social brain dysfunctions in schizophrenia: a review of neuroimaging studies . Psychiatry Res . 2006 ; 148 ( 2 ±3): 75 ± 92 . https://doi.org/10.1016/j.pscychresns. 2006 . 05 .001 PMID: 17088049
6. Frith CD , Frith U . Social cognition in humans . Curr Biol . 2007 ; 17 ( 16 ):R724± 32 . https://doi.org/10.1016/ j.cub. 2007 . 05 .068 PMID: 17714666
7. Amodio DM , Frith CD . Meeting of minds: the medial frontal cortex and social cognition . Nat Rev Neurosci . 2006 ; 7 ( 4 ): 268 ± 77 . https://doi.org/10.1038/nrn1884 PMID: 16552413
8. Beer JS , Heerey EA , Keltner D , Scabini D , Knight RT . The regulatory function of self-conscious emotion: insights from patients with orbitofrontal damage . J Pers Soc Psychol . 2003 ; 85 ( 4 ): 594 ± 604 . https:// doi.org/10.1037/ 0022 - 3514 . 85 .4.594 PMID: 14561114
9. Hornak J , Rolls ET , Wade D . Face and voice expression identification in patients with emotional and behavioural changes following ventral frontal lobe damage . Neuropsychologia . 1996 ; 34 ( 4 ): 247 ± 61 . PMID: 8657356
10. Koenigs M , Young L , Adolphs R , Tranel D , Cushman F , Hauser M , et al. Damage to the prefrontal cortex increases utilitarian moral judgements . Nature . 2007 ; 446 ( 7138 ): 908 ± 11 . https://doi.org/10.1038/ nature05631 PMID: 17377536
11. Shamay-Tsoory SG , Tomer R , Berger BD , Aharon-Peretz J . Characterization of empathy deficits following prefrontal brain damage: the role of the right ventromedial prefrontal cortex . J Cogn Neurosci . 2003 ; 15 ( 3 ): 324 ± 37 . https://doi.org/10.1162/089892903321593063 PMID: 12729486
12. Haxby J V, Hoffman EA , Gobbini MI . The distributed human neural system for face perception . Trends Cogn Sci . 2000 ; 4 ( 6 ): 223 ± 33 . PMID: 10827445
13. Hoffman E a , Haxby J V. Distinct representations of eye gaze and identity in the distributed human neural system for face perception . Nat Neurosci . 2000 Jan 1 ; 3 ( 1 ): 80 ±4. https://doi.org/10.1038/71152 PMID: 10607399
14. Puce A , Perrett D . Electrophysiology and brain imaging of biological motion . Philos Trans R Soc L B Biol Sci . 2003 ; 358 ( 1431 ): 435 ± 45 .
15. Pelphrey KA , Morris JP , McCarthy G . Grasping the intentions of others: the perceived intentionality of an action influences activity in the superior temporal sulcus during social perception . J Cogn Neurosci . 2004 ; 16 ( 10 ): 1706 ± 16 . https://doi.org/10.1162/0898929042947900 PMID: 15701223
16. Saxe R , Xiao DK , Kovacs G , Perrett DI , Kanwisher N. A region of right posterior superior temporal sulcus responds to observed intentional actions . Neuropsychologia . 2004 ; 42 ( 11 ): 1435 ± 46 . https://doi.org/ 10.1016/j.neuropsychologia. 2004 . 04 .015 PMID: 15246282
17. Aichhorn M , Perner J , Kronbichler M , Staffen W , Ladurner G. Do visual perspective tasks need theory of mind? Neuroimage . 2006 ; 30 ( 3 ): 1059 ± 68 . https://doi.org/10.1016/j.neuroimage. 2005 . 10 .026 PMID: 16337813
18. Blanke O , Mohr C , Michel CM , Pascual-Leone A , Brugger P , Seeck M , et al. Linking out-of-body experience and self processing to mental own-body imagery at the temporoparietal junction . J Neurosci . 2005 ; 25 ( 3 ): 550 ±7. https://doi.org/10.1523/JNEUROSCI.2612- 04 . 2005 PMID: 15659590
19. Apperly IA , Samson D , Chiavarino C , Humphreys GW . Frontal and temporo-parietal lobe contributions to theory of mind: neuropsychological evidence from a false-belief task with reduced language and executive demands . J Cogn Neurosci . 2004 ; 16 ( 10 ): 1773 ± 84 . https://doi.org/10.1162/ 0898929042947928 PMID: 15701227
20. Saxe R , Kanwisher N. People thinking about thinking people. The role of the temporo-parietal junction in ªtheory of mind . º Neuroimage . 2003 ; 19 ( 4 ): 1835 ± 42 . PMID: 12948738
21. Northoff G , Bermpohl F . Cortical midline structures and the self . Trends Cogn Sci . 2004 ; 8(3):102±7 . https://doi.org/10.1016/j.tics. 2004 . 01 .004 PMID: 15301749
23. Winston JS , Strange BA , O'Doherty J , Dolan RJ . Automatic and intentional brain responses during evaluation of trustworthiness of faces . Nat Neurosci . 2002 ; 5 ( 3 ): 277 ± 83 . https://doi.org/10.1038/nn816 PMID: 11850635
White SJ , Frith U , Rellecke J , Al-Noor Z , Gilbert SJ . Autistic adolescents show atypical activation of the brain's mentalizing system even without a prior history of mentalizing problems . Neuropsychologia . 2014 ; 56 ( 1 ): 17 ± 25 .
24. Yarkoni T , Poldrack RA , Nichols TE , Van Essen DC , Wager TD . Large-scale automated synthesis of human functional neuroimaging data . Nat Methods . 2011 ; 8 ( 8 ): 665 ± 70 . https://doi.org/10.1038/nmeth. 1635 PMID: 21706013
Taylor SF , Welsh RC , Chen AC , Velander AJ , Liberzon I. Medial frontal hyperactivity in reality distortion . Biol Psychiatry . 2007 ; 61 ( 10 ): 1171 ±8. https://doi.org/10.1016/j.biopsych. 2006 . 11 .029 PMID: 17434455 Lang PJ , Ohman A , Vaitl D. The international affective picture system [photographic slides] . Gainesville, FL: Center for Research in Psychophysiology, University of Florida; 1988 .
Yang Y , Gu H , Zhan W , Xu S , Silbersweig DA , Stern E. Simultaneous perfusion and BOLD imaging using reverse spiral scanning at 3T: Characterization of functional contrast and susceptibility artifacts . Magn Reson Med . 2002 Aug; 48 ( 2 ): 278 ± 89 . https://doi.org/10.1002/mrm.10196 PMID: 12210936 29 . Jenkinson M , Bannister P , Brady M , Smith S . Improved Optimization for the Robust and Accurate Linear Registration and Motion Correction of Brain Images . Neuroimage . 2002 Oct; 17 ( 2 ): 825 ± 41 . PMID: 12377157
Worsley KJ , Friston KJ . Analysis of fMRI Time-Series RevisitedÐAgain . Neuroimage. 1995 Sep; 2 ( 3 ): 173 ± 81 . https://doi.org/10.1006/nimg. 1995 .1023 PMID: 9343600
Morey RD , Rouder JN . BayesFactor: Computation of Bayes Factors for Common Designs . R package version 0 .9. 12 ± 2 . [Internet]. 2015 .
32. Jeffreys H . Theory of probability . Oxford: Oxford University Press, Clarendon Press; 1961 .
Peigneux P , Salmon E , van der Linden M , Garraux G , Aerts J , Delfiore G , et al. The role of lateral occipitotemporal junction and area MT/V5 in the visual analysis of upper-limb postures . Neuroimage . 2000 ; 11 ( 6 Pt 1 ): 644 ± 55 .
Olson IR , McCoy D , Klobusicky E , Ross LA . Social cognition and the anterior temporal lobes: a review and theoretical framework . Soc Cogn Affect Neurosci . 2013 ; 8 ( 2 ): 123 ± 33 . https://doi.org/10.1093/scan/ nss119 PMID: 23051902
Sugiura M , Shah NJ , Zilles K , Fink GR . Cortical representations of personally familiar objects and places: functional organization of the human posterior cingulate cortex . J Cogn Neurosci . 2005 ; 17 ( 2 ): 183 ± 98 . https://doi.org/10.1162/0898929053124956 PMID: 15811232