Content uploaded by Saskia B.J. Koch
Author content
All content in this area was uploaded by Saskia B.J. Koch on Feb 18, 2019
Content may be subject to copyright.
Acute stress alters the ‘default’brain processing
Wei Zhang
a
,
b
,
*
, Mahur M. Hashemi
a
,
b
, Reinoud Kaldewaij
a
,
b
, Saskia B.J. Koch
a
,
b
,
Christian Beckmann
a
,
c
,
d
, Floris Klumpers
a
,
b
,
1
, Karin Roelofs
a
,
b
,
1
a
Donders Institute, Centre for Cognitive Neuroimaging, Radboud University, Nijmegen, the Netherlands
b
Behavioural Science Institute, Radboud University, Nijmegen, the Netherlands
c
Department of Cognitive Neuroscience, Radboud University Medical Centre, Nijmegen, the Netherlands
d
Oxford Centre for Functional Magnetic Resonance Imaging of the Brain (FMRIB), University of Oxford, Oxford, UK
ARTICLE INFO
Keywords:
Stress
Resting-state fMRI
Functional connectivity
Resting-state networks
RSNs
Stress vulnerability
Stress reactivity
ABSTRACT
Active adaptation to acute stress is essential for coping with daily life challenges. The stress hormone cortisol, as
well as large scale re-allocations of brain resources have been implicated in this adaptation. Stress-induced shifts
between large-scale brain networks, including salience (SN), central executive (CEN) and default mode networks
(DMN), have however been demonstrated mainly under task-conditions. It remains unclear whether such network
shifts also occur in the absence of ongoing task-demands, and most critically, whether these network shifts are
predictive of individual variation in the magnitude of cortisol stress-responses.
In a sample of 335 healthy participants, we investigated stress-induced functional connectivity changes (delta-
FC) of the SN, CEN and DMN, using resting-state fMRI data acquired before and after a socially evaluated cold-
pressor test and a mental arithmetic task. To investigate which network changes are associated with acute stress,
we evaluated the association between cortisol increase and delta-FC of each network.
Stress-induced cortisol increase was associated with increased connectivity within the SN, but with decreased
coupling of DMN at both local (within network) and global (synchronization with brain regions also outside the
network) levels.
These findings indicate that acute stress prompts immediate connectivity changes in large-scale resting-state
networks, including the SN and DMN in the absence of explicit ongoing task-demands. Most interestingly, this
brain reorganization is coupled with individuals’cortisol stress-responsiveness. These results suggest that the
observed stress-induced network reorganization might function as a neural mechanism determining individual
stress reactivity and, therefore, it could serve as a promising marker for future studies on stress resilience and
vulnerability.
1. Introduction
Beyond traditional group-level analyses, recent investigations have
moved towards characterizing individual profiles of functional connec-
tivity (FC), which predict cognitive and behavioral performance at the
single-subject level (Finn et al., 2017;Marquand et al., 2017). FC fin-
gerprints, derived from resting-state fMRI (rs-fMRI) data in particular,
have been widely used in studies examining the abnormalities of con-
nectivity profiles in patients with stress-related psychiatric disorders
(Koch et al., 2016;Nicholson et al., 2015;Oathes et al., 2015). However,
it remains unclear how stress induction leads to changes in resting-state
FC (rs-FC) profiles and how those changes may be linked to central
stress-response systems, the major of which is the hypothalamic pituitary
adrenal (HPA) axis.
Stress-related disorders like post-traumatic stress disorder (PTSD)
have been suggested to be characterized by abnormal organization and
functioning of three major large-scale brain networks, namely the
salience network (SN), central executive network (CEN) and default
model network (DMN; Menon, 2011). The interpretability of these
findings however, hinges on whether they can be linked to quantitative
biological markers of acute stress states such as the HPA-axis activity and
its end product cortisol, which have extensively been linked to stress
adaptation (De Kloet et al, 2005;McEwen, 1998). To understand the
functional implications of neural network shifts in relation to stress
* Corresponding author. Kapittelweg 29, 6525 EN, Nijmegen, the Netherlands.
E-mail address: w.zhang@donders.ru.nl (W. Zhang).
1
Equal author contributions.
Contents lists available at ScienceDirect
NeuroImage
journal homepage: www.elsevier.com/locate/neuroimage
https://doi.org/10.1016/j.neuroimage.2019.01.063
Received 27 June 2018; Received in revised form 16 January 2019; Accepted 24 January 2019
Available online 28 January 2019
1053-8119/©2019 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
NeuroImage 189 (2019) 870–877
adaptation, a number of recent investigations directly manipulated acute
stress states and found that exposure to stress-induction helps reveal the
fundamental neural origins of individual stress responsiveness (Cousijn
et al., 2010;Henckens et al., 2012;van Oort et al., 2017). At the network
level, a small number of promising studies have revealed a stress-induced
re-allocation of neural resources entailing increases in SN connectivity at
the cost of decreases in CEN connectivity (Hermans et al. 2011,2014).
This dynamic re-prioritization can generally be beneficial as it allows for
adaptive responses to changing environmental conditions.
Importantly, these large-scale network shifts after stress induction
have mainly been identified under task conditions so far (Hermans et al.,
2011,2014;McMenamin and Pessoa, 2015;Young et al., 2016). Due to
the dominant roles of the SN and CEN to meet ongoing task-demands
(Dosenbach et al., 2007;Seeley et al., 2007), this potentially biases the
observed network shifts towards states involving SN and CEN func-
tioning, and reduces variations in internally-driven neural fluctuations
(i.e., resting-state DMN). It therefore remains unclear whether stress in-
duction could result in similar network shifts when external task de-
mands are absent, as in a resting-state. So far, there has been very little
investigation of system-level resting-state network connectivity changes
after stress induction. While limited evidence from studies using a
seed-based approach suggest a general increase in the SN connectivity
and mixed patterns in different DMN regions after stress induction (see
review by van Oort et al., 2017), observations from clinical populations
with stress-related disorders indicate the involvement of increased SN
and reduced DMN connectivity in psychopathology (Admon et al., 2013;
Koch et al., 2016). Accordingly, we predicted that with our network
approach, we would observe similar increases in the SN connectivity and
decreases in the DMN connectivity after stress induction while no
changes in CEN connectivity were expected.
Most critically, despite the well-known variation in individual stress-
responses, it remains unclear whether those network shifts are associ-
ated with individual stress response sensitivity, in part because most of
those studies were not adequately powered to detect individual differences
(van Oort et al., 2017). Until now, limited evidence has suggested that
stress-induced FC increases in the SN under task conditions might be
linked to individual variances in cortisol,
α
-amylase and subjective stress
responses (Hermans et al., 2011), as well as to instant heart-rate changes
(Young et al., 2016). This leaves the question open whether these obser-
vations are linked to specific task conditions or represent a shift in default
functioning of these neural networks. In the current study, we aimed to test
the hypothesis that acute stress-induced rs-FC changes in large-scale net-
works would occur as a function of individual differences in the cortisol
stress-responses. Specifically, we expected stronger cortisol increases to be
associated with increased SN and decreased DMN connectivity.
We tested our hypothesis in a well-powered sample of healthy in-
dividuals (N ¼335), who underwent a formal stress induction, preceded
and followed by rs-fMRI scans (i.e. without external stimuli input). In
specific, we investigated stress-induced network connectivity changes
within the SN, CEN and DMN (i.e. local connectivity changes), as well as
their synchronization with other brain regions (i.e., global connectivity
changes; Cole et al., 2012;Cole et al., 2011;Gonzalez-Castillo et al., 2015).
By assessing both local and global connectivity changes, the current study
aimed to capture a wider picture of the connectivity patterns following
stress induction, not only specifically within the restricted areas (i.e.,
within each network) but also in the areas extending beyond our network
definition. Further, to understand functional implications of these network
reorganizations, we tested if these connectivity changes would occur as a
function of the individual cortisol-stress reactivity.
2. Materials and methods
2.1. Participants
A total of 372 participants completed the current study. An additional
group of 23 participants were tested to generate independent resting-
state network templates (see details below). Exclusion criteria included
any current psychiatric or neurological disorder, history of, or current
endocrine or neurological treatment, current use of psychotropic medi-
cation, and current drug or alcohol abuse (full details in Koch et al.,
2017). After exclusion (see more details below), data from a total of 335
participants, including 276 police students who had recently started their
education at the police academy, were analysed. Sixty-one out of a total
of N ¼80 female participants in this sample reported hormonal contra-
ceptive uses.
As part of a larger project consisting of multiple tests including
approach-avoidance, reversal learning and emotional Go-NoGo tasks, the
current study was implemented as the last experiment in the late after-
noon (please refer to Koch et al., 2017 for a complete overview of the
project) and was conducted in accordance with the principles of the
Declaration of Helsinki and approved by the Independent Review Board
Nijmegen (IRBN), the Netherlands. All participants gave their written
informed consent before the study and all data were collected at the
Donders Institute for Brain, Cognition and Behavior in Nijmegen, The
Netherlands.
2.2. Experimental design and procedure
The experiment took place after 4PM, when cortisol levels are rela-
tively stable because of the diurnal rhythm, so reliable individual stress-
responses could be obtained (Miller et al., 2016). Two runs of fMRI
scanning were implemented, one before and one after stress induction.
This experiment was placed in the last imaging session of the experi-
mental day, i.e. participants were already acquainted with the scanning
procedure and thus not scanner naïve (full details of experiment pro-
tocols in Koch et al., 2017).
Stress responses were induced by sequential administration of a so-
cially evaluated cold pressor task (SECPT) and a mental arithmetic (MA)
task, a procedure that has been shown to successfully induce psycho-
physiological and subjective stress responses (Luo et al., 2018;Schwabe
et al., 2008). Following a similar procedure as in previous studies (Luo
et al., 2018;Vogel et al., 2015), participants were instructed to immerse
their right foot in icy-cold (0–3C) water for 3 minutes. Immediately
after SECPT, a 3-minute MA task was administered. Participants were
instructed to count back out loud from 2053 in steps of 17 as quickly and
accurately as possible. The full stress-induction procedure lasted
approximately 8 min, including instructions (full details in Supplemental
Methods and Materials).
2.3. Data acquisition and analysis
2.3.1. fMRI data acquisition
Each fMRI run involved one 6-minute long resting-state fMRI scan
(RS1 and RS2) and one 2-minute long field-mapping scan (not used for
the current analyses), leading to a total of 8 minutes per fMRI run. Par-
ticipants were instructed to lie still and to stare at a small white cross at
the screen center during both scanning runs, which has been suggested to
increase the reliability of within-network connectivity (Birn et al., 2013).
All images were collected using a 3T Siemens Magnetom Prisma
fit
MRI
scanner (Erlangen, Germany) with a 32-channel head coil. T2*-weighted
EPI BOLD-fMRI images were acquired for the resting-state scans, using a
multi-band 8 protocol with an interleaved slice acquisition sequence
(slice number ¼64, TR ¼735ms,TE¼39ms,flip angle ¼52, voxel
size ¼2.42.42.4mm
3
, slice gap ¼0mm, FOV ¼210 mm)that was
optimized from the standard recommended scanning protocol of the
Human Connectome Project (http://protocols.humanconnectome.org/
HCP/3T/imaging-protocols.html). High-resolution structural images
(111mm
3
) were also acquired, using a T1-weighted MP-RAGE
sequence (TR ¼2300ms,TE¼3.03ms,flip angle ¼8,
FOV ¼256256192 mm
3
).
W. Zhang et al. NeuroImage 189 (2019) 870–877
871
2.3.2. Stress measurement collection
In total, five salivary samples were taken using Salivettes
®
collection
tubes (Sarstedt, Germany) at 10, 0, þ10, þ20, and þ30 minutes with
respect to the onset time of stress induction (Fig. 1). In a group of 61 par-
ticipants, the last sample (i.e., at þ30 minutes) has not been obtained,
resulting in a sample of N ¼311 participantswith complete measurements.
Together with saliva sampling, self-reported ratings of positive and
negative affect (PANAS; Watson et al., 1988) were collected. Subjective
ratings on negative affect were based on the sum of the scores of the 10
negative affect items for each participant. Each rating took place on a
5-point likert scale, with the sum score consequently ranging between 10
and 50. The same subsample as mentioned above (N ¼311) was
measured with complete measurements at all five time points (see Fig. 1).
2.3.3. Analyses on stress measures
Statistical analyses on stress measures were carried out separately on
the sample of participants with complete data for all individual stress
measurements (N ¼311), and on the full sample (N ¼372). Only the
results from the sample with complete data are reported below. Results
from the full sample were highly similar and can be found in Supple-
mental Results.
Main effects of sampling time (subsequent time points) on salivary
cortisol,
α
-amylase levels and negative affect sum-scores were tested to
index acute stress effects, using a linear mixed model with a random
intercept for each individual. As salivary cortisol has been shown to be a
robust and reliable measure, frequently used as a biomarker of stress
responses (Bozovic et al., 2013;Hellhammer et al., 2009), cortisol level
increases (the difference between time 20 minutes and baseline 0 min-
ute) were used to investigate the association with imaging measures. To
evaluate the typical gender effect on cortisol (Kudielka and Kirschbaum,
2005;Reschke-Hern
andez et al., 2017), as well as potential group effects
(i.e., police students vs. remaining participants) in our sample, cortisol
increases were compared between those groups. In short, while we
observed typical gender (but not group) effects on cortisol responses
(males >females), the main resulting associations between cortisol and
neural responses were found to hold when taking into account gender.
Full details of these supplementary analyses can be found in Supple-
mental Methods and Materials.
2.4. fMRI preprocessing and analysis
2.4.1. Preprocessing
Imaging data from 10% (N ¼37) of the total participants were
excluded from analysis due to technical issues (N ¼4), motion (based on
the mean value of the relative displacement; top 5% participants from
each rs-fMRI scan leading to a total of 8.5%, with N ¼32 from the entire
sample; Pruim et al., 2015) and incidental neurological findings (N ¼1),
which resulted in a sample of 335 participants, including 276 police
students.
To allow for T2*equilibration effects, the first five images of each
resting-state scan were discarded. Analysis of fMRI data was performed
with FSL5.0.9 (FMRIB, Oxford, UK). Preprocessing included motion
correction by aligning all images to the first scan using rigid body
transformations, spatial smoothing with a 5 mm FWHM kernel, denoising
using ICA-AROMA (Pruim et al., 2015), and high-pass filtering with a
cut-off of 100 Hz. The preprocessed images were then fed into a general
linear model to regress out nuisance effects. Specifically, twenty-four
head motion parameters (i.e., the six realignment parameters, their
temporal derivatives and the quadratic terms of both the original pa-
rameters and derivatives; Caballero-Gaudes and Reynolds, 2017;Friston
et al., 1996;Zu Eulenburg et al., 2012) were included in the model to
minimize the motion artefacts. Additionally, each individual T1 image
Fig. 1. Stress measures. Stress induction took place in between
two runs of rs-fMRI scans (RS1 &RS2). Salivary samples and
subjective affect reports were collected in a total of five times
(i.e., 10m to þ30m), with time interval between each sample
being approximately 10 minutes. Increases in cortisol were
observed 20 and 30 minutes after the onset of stress induction
while increases in
α
-amylase and ratings of negative affect were
observed immediately after stressinduction (i.e., at time þ10m),
in comparison to baseline measurement at time 0. Of note, the
last sample of
α
-amylase (i.e.,at þ30 min) was removed from the
analysis dueto large increases resultingfrom physical movement
of the participants exiting the scanner room. Asterisks indicated
statistically significantdifferences relative to time 0 immediately
preceding stressinduction (***p <.0001, **p <.001, *p <.05).
W. Zhang et al. NeuroImage 189 (2019) 870–877
872
was segmented for subject-specific white matter and CSF masks that were
subsequently thresholded with a 95% probability and registered with
functional image. Mean signal intensities of white matter and CSF were
extracted and included in the GLM (Caballero-Gaudes and Reynolds,
2017;Satterthwaite et al., 2013).
The residual images from this linear model were normalized to the
Montreal Neurological Institute template (MNI152), using linear and
nonlinear transformations via boundary based registration (BBR; Greve
and Fischl, 2009), FLIRT (Jenkinson et al., 2002;Jenkinson and Smith,
2001) and FNIRT (Andersson et al., 2007). Consequently, each partici-
pant had two normalized residual images (i.e., cleaned rs-fMRI data) that
index BOLD signal fluctuations, before and after acute stress induction,
respectively.
2.4.2. Identifying delta-FC of RSNs
Group-level network templates based on data of 23 independently
tested non-stressed participants were produced, using group independent
component analysis (ICA) as implemented in MELODIC (Beckmann and
Smith, 2005). ICA components showing the highest cross-correlation of
mean time-series with pre-selected functional ROIs (i.e., anterior SN, left
CEN, right CEN and ventral DMN) from the Stanford FIND atlas (Shirer
et al., 2012) were identified as RSNs of interest. This approach allowed us
to select RSNs of interest that were not biased towards the data either
before or after stress induction. Importantly, the final selected RSN
templates involves all major nodes/areas that are typically considered as
hub regions in those networks (Fig. S1). For example, the selected SN
included the bilateral anterior insula, dorsal ACC and amygdala; the CEN
included dorsolateral prefrontal cortex and posterior parietal cortex
while the DMN included ventromedial prefrontal cortex, para-
hippocampal gyrus, posterior cingulate cortex and precuneus. Connec-
tivity changes after stress induction (i.e., delta-FC) were defined as the
differences in each network of interest before and after stress induction.
All individual increased (after >before) and decreased (before >after)
delta-FC images were then tested at the group-level, using permutation
tests via Randomise (Winkler et al., 2014), to examine significant
delta-FC after stress induction for each network at both local (i.e., within
network) and more global (i.e., synchronization with regions both within
and outside our network definitions) levels.
The results from these initial exploratory tests were considered sig-
nificant using a family-wise-error (FWE) corrected p-value of 0.00625,
derived from a threshold-free cluster enhancement approach (Smith and
Nichols, 2009) that accounts for the number of individual networks (i.e.,
SN, LCEN, RCEN, DMN), as well as the number of connectivity change
directions (i.e., increases and decreases) involved in the comparisons.
2.4.3. Linking delta-FC of RSNs to stress responses
Although the lack of a non-stressful control group of adequate size
precluded a direct group comparison for testing the specificity of the
observed stress effects in the initial group analysis, the large sample size
(N ¼335) of the current study allowed us to verify that the observed
connectivity changes in large-scale networks indeed covaried with indi-
vidual responsiveness of the HPA axis, and therefore linked to changes in
the stress response. To this end, increased cortisol level was added as a
covariate in the permutation tests to link the connectivity changes to
acute stress-responses. To examine significant stress-related changes
within each network (i.e., local delta-FC), we used our group ICA tem-
plates (see above) as the masks for small volume correction (SVC) to
directly test our a priori hypotheses. As we had no hypothesis to test the
CEN unilaterally, individual delta-FC of the left and right CEN for each
participant were combined for these, and all following hypothesis-testing
analyses. Results from group permutation tests with cortisol increase as
covariates were considered significant with a FWE corrected p-value of
0.0167 that takes into account the number of hypothesis tests for three
networks (i.e., SN, CEN and DMN), using Bonferroni correction.
In addition to the standard univariate voxel-wise approach described
above, we also investigated individual differences at the network level.
To this end, mean coefficients of delta-FC were extracted from the clus-
ters that showed significant connectivity changes after stress induction
(Fig. S2). These coefficients indicated the strength of delta-FC between
each RSN and all brain regions that showed changes in connectivity after
stress induction (i.e., widespread changes referred as global synchroni-
zation including both changes with brain regions within- and outside the
network), and were correlated with cortisol increase across participants.
To test whether any association between delta-FC coefficients and
cortisol increases existed specifically within each network (i.e., local
changes), mean coefficients of individual delta-FC maps were extracted
with masks of our group ICA templates. We firstly tested statistical sig-
nificance of mean coefficients, using one-sample t-test (see results in the
Supplemental Methods and Materials). Subsequently, Spearman rank
correlation analyses were used to test above associations, which mini-
mize the potential influences from extreme values in the variables. Re-
sults were considered significant with an adjusted p-value of p <.0167 to
account for the number of analyses that were carried out to test our a
priori hypotheses on three RSNs. Bootstrapped confidence intervals
(boot.CI) were calculated for these rank correlation analyses, using the
adjusted bootstrap percentile (BCa) method with n ¼1000 iterations.
3. Results
3.1. Stress measures
Stress-induction was successful as was indicated by significant in-
creases in all stress measures. Specifically, main effects of sampling time
were observed for salivary cortisol (F(4,1182.08) ¼145.76, p <.0001),
α
-amylase (F(3830.97) ¼9.78, p <.0001) and negative affect ratings
(F(4,1230.01) ¼99.24, p <.0001; Fig. 1).
In line with the delayed cortisol stress responses (Schwabe et al.,
2008;Kirschbaum and Hellhammer, 1994), cortisol levels significantly
increased at þ20 and þ30 minutes after the onset of stress induction
compared to pre-stress baseline (i.e., at time 0 minute; all p's <0.0001),
while no significant difference was observed between þ20
and þ30 minutes (t(1181.93) ¼-2.64, p ¼.06; Fig. 1). As expected,
α
-amylase and subjective stress levels peaked immediately after stress
induction (i.e., at time þ10 minutes; t
α
-amylase
(829.36) ¼-4.68,
p<.0001; t
affect
(1229.13) ¼-14.18, p<.0001). While
α
-amylase level
remained high (i.e., at time þ20 minutes; t(831.08) ¼-4.03, p <.001),
subjective scores of negative affect quickly declined again (i.e., at time
þ20 minutes; t(1229.13) ¼1.27, p ¼.71) and eventually ended below
the pre-stress baseline (i.e., time 0; t(1229.44) ¼3.74, p <.005).
3.2. Cortisol-related FC changes of RSNs
Following acute stress induction, whole-brain analyses revealed both
increased and decreased connectivity patterns for all four RSNs with
wide-spread regions (Supplemental Results; Fig. S2;Table S1). To inves-
tigate the functional implications of these connectivity changes, we
linked the observed delta-FC to the stress response marker cortisol. Our
voxel-wise analyses identified an increased overall connectivity of the SN
with a cluster in the right dACC predictive of individual cortisol increase
(Bonferroni correction adjusted p
FWE
<.0167; Fig. 2). Additional control
tests further confirmed that this effect was not associated with head
motion change, defined as the difference in the mean value of relative
frame-wise displacement between RS1 and RS2 (Rs ¼0.07, p ¼.22).
No other networks showed connectivity increases or decreases related to
the cortisol stress response in the voxel-wise analysis.
Thereafter, we calculated the average connectivity changes across all
regions showing significant increase or decrease, separately, for each of
three resting-state networks and correlated them with individual cortisol
increases. At this more global level, reductions in DMN connectivity with
the regions also outside the network was significantly correlated with
cortisol responses (Rs ¼0.16, Bonferroni correction adjusted
p<.0167, boot.CI¼(0.050, 0.266); Fig. 3A). Further investigation
W. Zhang et al. NeuroImage 189 (2019) 870–877
873
revealed a trend of correlation between this globally decreased DMN
connectivity and the head motion changes (Rs ¼0.10, p ¼.06). However,
results from a multiple regression model confirmed the association be-
tween DMN connectivity decrease and cortisol increase when effects of
head motion were controlled (t(315) ¼2.58, p <.05).
Concerning the connectivity changes within networks as a function of
individual differences in cortisol responses, we found an association
between the decreased delta-FC within the DMN and larger cortisol in-
creases (FIND atlas vDMN mask, Rs ¼0.16, Bonferroni correction
adjusted p <.0167, boot.CI¼(-0.27, 0.05); ICA-derived DMN group
template, Rs ¼0.11, p ¼.047, boot.CI¼(-0.22, 0.0007); Fig. 3B).
None of those RSN changes showed correlations with the head motion
changes (p's >0.11) and none of the other RSNs showed such a linkage
with cortisol increase (p's >0.05).
4. Discussion
The current study investigated rapid, stress-induced connectivity-
changes within major resting-state networks (SN, CEN, DMN), as well as
between these networks and other brain regions as a function of indi-
vidual stress-response magnitude, measured by the stress-hormone
cortisol. Specifically, cortisol stress response levels were associated
with increased connectivity within SN, the network that is critical for
detecting behaviorally relevant stimuli and for coordinating neural re-
sources in response to these stimuli. Interestingly, the results also show
that the reduction in local (i.e., within network) and global synchroni-
zation of DMN, known for its involvement in internal processing and
homeostasis, was linked to individual differences in stress-induced
cortisol levels. These findings match with the idea of a stress-induced
network reorganization and suggest that increased SN and decreased
DMN connectivity may function as relevant neural indicators for stress
responsiveness.
In line with previous findings, we observed connectivity changes of
large-scale networks following stress induction (Fig. S2;Table S1). Given
our interest in RSN connectivity changes in relation to the individual
stress-responses, we specifically linked the observed delta-FC to cortisol
increase at the individual level. Concerning within-network delta-FC
(i.e., local changes), our voxel-wise results demonstrate that cortisol-
stress responses were associated with connectivity increases within the
SN even when there is limited external input (i.e. during a resting-state
scan). This extends previous observations of increased SN connectivity
in response to acute stress induction during task-positive conditions
(Hermans et al., 2011;Young et al., 2016;van Oort et al., 2017). Spe-
cifically, we identified increased connectivity between the SN as a whole
and its subregion, the dACC in participants with high cortisol stress re-
sponses. As a crucial node of the SN, the dACC has been implicated in
diverse functions at the intersection of cognition and emotion including
interoceptive-autonomic processing (Craig, 2002;Critchley et al., 2004),
pain and negative affect processing (Rotge et al., 2015) and integrating
information relevant for cognitive control (Shenhav et al., 2013),
Fig. 2. Increased overall SN connectivity
with dorsal ACC (red-yellow, a core SN sub-
region) was associated with cortisol increase
in response to acute stress induction. Results
shown on the left panel are whole brain
corrected without additional correction for
the number of networks for visualization
purpose (P
fwe
<.05) and imposed on our ICA-
derived SN template that was used to restrict
the search space (green; z >3). Results
illustrated on the right panel are individual
cortisol increase against mean coefficients
extracted from individual increased SN,
using the cluster showing significant in-
creases as the mask (Bonferroni adjusted
P
fwe
<.0167). Follow-up tests confirmed that
this effect was not driven by extreme values:
the association remained significant also
when the data of N ¼3 participants with
relatively extreme values (i.e., >3std from
the mean) were removed (Rs ¼0.16,
p<.005).
Fig. 3. Stress-induced cortisol increases are
correlated with reduced synchronization at a
more global level between DMN and brain
regions also outside of the network, such as
frontal gyrus and temporal gyrus
a
(A) and at
a local level within the DMN (B). Stress-
induced connectivity changes are indexed
by mean coefficients at the x-axis, extracted
from each participant using a connectivity
map that contains all brain regions showing a
significantly reduced synchronization with
DMN (A), and using the ICA-derived DMN
group template (B), respectively. Brain im-
ages depicted the masks that were used to
extract aforementioned coefficients.
#p <.05.
*p <.0167 (Bonferroni corrected).
a
full list can be found in Supplementary
Table S1.
W. Zhang et al. NeuroImage 189 (2019) 870–877
874
suggesting potential involvement of this region in responding to chal-
lenging conditions and the appraisal and expression of anxiety (Etkin
et al., 2011). Abnormalities in dACC and more generally in SN connec-
tivity have been implicated as the neurobiological correlate of enhanced
salience or threat processing, a major characteristic of stress-related
psychiatric disorders (Etkin and Wager, 2007;Koch et al., 2016;Vais-
vaser et al., 2013).
With respect to changes in global synchronization, we found that
higher cortisol increases after stress induction were associated with
larger reductions in global synchronization of DMN (i.e., reductions in
the interaction between the whole DMN and widespread brain regions
also outside the DMN, as listed in Table S1 “decreased FC with DMN”).
Interestingly, a similar association was also identified for the overall
connectivity decrease within the DMN, indicated by the reduced mean
coefficient of delta-FC from the FIND atlas-defined network core regions.
These results suggests the involvement of the DMN in the processing of
acute stress induction without on-going task demands that was not
captured in the previous investigations (Young et al., 2016;Hermans
et al., 2011). The DMN has largely been linked to self-referential pro-
cesses (Andrews-Hanna et al., 2010;Buckner et al., 2008). Alterations in
the DMN connectivity have consistently been implicated in various
psychiatric disorders and particularly in stress-related disorders. For
example, reduced baseline DMN connectivity has been linked to PTSD
patients, while insufficient suppression of DMN has been implicated in
remitted major depression (Admon et al., 2013;Bartova et al., 2015;
Koch et al., 2016).
With enhanced SN connectivity on the one hand, yet reduced DMN
connectivity on the other, our findings are generally in line with previous
investigations that demonstrated a stress-induced network shift towards
the SN (Hermans et al., 2011,2014). In the current study, however, such
a reallocation of neural resources appears to occur between the SN and
DMN rather than CEN, when no external stimuli (i.e., ongoing stressors)
are present. Our findings of an opposite impact of stress on connectivity
of the DMN and SN appear compatible with theories of a neural resource
reassignment from the DMN to the SN in the interest of processing more
relevant information under a stressful state (Maron-Katz et al., 2016;
Quaedflieg et al., 2015;Vaisvaser et al., 2013,2016). Nevertheless, the
current study extends the literature by showing SN and DMN fluctuations
in the absence of external task demands that might be dependent on the
magnitude of the individual cortisol stress responses. Similar alterations
in the SN and DMN connectivity have been implicated in a wide range of
psychiatric disorders and particularly in stress-related disorders (Admon
et al., 2013;Bartova et al., 2015;Etkin and Wager, 2007;Koch et al.,
2016;Sripada et al., 2012;Vaisvaser et al., 2013). On the other hand,
however, studies in animals as well as in human suggest that the
HPA-axis is highly relevant for fast adaptation to stressful situations (De
Kloet et al., 2005;Jo€
els and Baram, 2009). It will be of interest for future
investigations to examine longitudinally whether the increased SN con-
nectivity and decreased DMN connectivity indicate individual adaptation
or vulnerability to acute stress induction, and whether those
stress-induced neural network responses can predict the development of
psychopathology after trauma exposure.
Several limitations of the present study should be mentioned. Firstly,
although we recruited an independent group to derive the group network
templates for imaging analysis in a non-biased fashion, it was of an
insufficient sample size (N ¼23) to serve as a direct control group for
validating stress-induced neural effects observed in a sample of N ¼335.
It is possible that the lack of such a control group could potentially
confound the observed stress effects at neural level with scanning order.
However, this concern is mitigated by the fact that our participants were
not scanner-naïve (i.e., had previously been tested in the same scanner
twice on the same testing day) and showed no cortisol increases before
the RS1 and stress induction. Most importantly, within our experimental
group, we confirmed that both enhanced SN connectivity and reduced
DMN connectivity correlated significantly with individual cortisol
responsiveness. These results together strongly suggest that the observed
neural effects are stress related. Secondly, the current acquisition length
of 6.5 minutes is shorter than the recommended acquisition length
(i.e., 9–12 min) of resting-state imaging data (Birn et al., 2013). How-
ever, together with our large sample size, the fast multi-band imaging
protocol (TR ¼735ms) enabled us to obtain a relatively large number of
scans (N ¼500) in each session, which increases the reliability of our
results. Thirdly, the effect size of the observed correlations could
be arguably considered small by traditional standards (i.e., coefficient
between 0.1 and 0.15). Recent meta-analyses however show that
traditional guidelines for interpreting correlation coefficients may have
been too stringent (Gignac and Szodorai, 2016;Hemphill, 2003) because
observed correlations are practically dampened by the imperfect
measurement reliability of two variables in the correlation almost in any
studies (Hedge et al., 2017;Vul et al., 2009). In the current study, it is
very well conceivable that factors beyond our experimental control (e.g.
sleep quality before the experiment day) might have influenced both the
cortisol and neural responses, and thus diluted correlation effect sizes.
Furthermore, evidence from simulations show that increasing sample
size is generally associated with decreasing correlation coefficients and
that a large sample size (e.g., N >250) entails more stable effect size
(Sch€
onbrodt and Perugini, 2013). The small effect size of the correlations
resulting from our large sample size therefore likely reflected a mean-
ingful and robust underlying association between neural network pro-
cessing and stress responses. Fourthly, it could be considered as a
limitation that physiological recordings (e.g., respiration) were not
included to further clean up the imaging data. However, acute stress
induction has been shown to influence physiological responses.
Regressing out physiological parameters further will enhance the risk of
only investigating the neural processes that are independent of
stress-induced sympathetic and parasympathetic activities (Murphy
et al., 2013). To control for potential non-neural physiology, we followed
the common practice in the literature to remove the mean intensity of the
WM and CSF from the imaging data (Henckens et al., 2012;Hermans
et al., 2011;Maron-Katz et al., 2016;Vaisvaser et al., 2013,2016). More
importantly, we went above and beyond this common practice by
acquiring imaging data with a fast sampling sequence (i.e., multiband 8),
which in combination of ICA has been shown to facilitate the identifi-
cation and elimination of physiological components (Boubela et al.,
2014;Parkes et al., 2018;Pruim et al., 2015). Finally, it is worth
mentioning that our current resting-state connectivity measurement in
the immediate aftermath of a stressor likely contains a mixture of both
acute stress reactions and stress recovery processes. For future studies,
including scans during stress recovery (i.e. after acute stress subsides)
would be of interest in order to more systematically study the temporal
dynamics of the stress-induced network changes observed here (Hermans
et al., 2014;Vaisvaser et al., 2013).
In conclusion, our results demonstrate distributed connectivity
changes in large-scale RSNs after stress induction. More importantly, the
strengthened coupling within the SN, as well as the degree of decoupling
within the DMN, and between the DMN and other brain regions, was
associated with individual cortisol stress-responsiveness. These results
suggest that acute stress induction alters default brain processing and
that such an alteration might potentially function as a neural mechanism
determining individual stress reactivity.
Acknowledgments
FK and KR contributed equally to this article. We are especially
thankful to Dutch Police Academy (Politieacademie) for their coopera-
tion and to Annika Smit for her valuable help with recruiting participants
and facilitating our study. We also thank our former and current col-
leagues Vanessa van Ast, Ingrid Kersten, Naomi de Valk, Geoffrey Bertou,
Leonore Bovy, Iris Hulzink, Tiele Dopp, Marijolein Hartgerink, Bart
Becker, Madine Zoet, Delphin van Benthem, Job de Brouwer, Lisanne
Nuijen, Pepijn van Houten, Klaas van Groesen and Nienke Flipsen for
their help in setting up the study, recruiting participants and acquiring
W. Zhang et al. NeuroImage 189 (2019) 870–877
875
data, to Maarten Mennes for his valuable suggestions to control motion
artefacts, to Zahra Fazal for sharing her thoughts on physiological con-
founds, and to Paul Gaalman for his technical assistance in fMRI data
acquisition.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.neuroimage.2019.01.063.
Funding and disclosure
This work was supported by a VICI grant (#453-12-001) from the
Netherlands Organization for Scientific Research (NWO) and a starting
grant from the European Research Council (ERC_StG2012_313749)
awarded to Karin Roelofs. The authors report no biomedical financial
interests or potential conflicts of interest.
References
Admon, R., Milad, M.R., Hendler, T., 2013. A causal model of post-traumatic stress
disorder: Disentangling predisposed from acquired neural abnormalities. Trends
Cognit. Sci. 17 (7), 337–347. https://doi.org/10.1016/j.tics.2013.05.005.
Andersson, J.L.R., Jenkinson, M., Smith, S., 2007. Non-linear registration, aka spatial
normalisation. FMRIB Technial Report TR07JA2. Oxford Centre for Functional
Magnetic Resonance Imaging of the Brain, Department of Clinical Neurology. Oxford
University, Oxford, UK.
Andrews-Hanna, J.R., Reidler, J.S., Sepulcre, J., Poulin, R., Buckner, R.L., 2010.
Functional-anatomic fractionation of the brain's default network. Neuron 65 (4),
550–562. https://doi.org/10.1016/j.neuron.2010.02.005.
Bartova, L., Meyer, B.M., Diers, K., Rabl, U., Scharinger, C., Popovic, A., et al., 2015.
Reduced default mode network suppression during a working memory task in
remitted major depression. J. Psychiatr. Res. 64, 9–18. https://doi.org/10.1016/j.
jpsychires.2015.02.025.
Beckmann, C.F., Smith, S.M., 2005. Tensorial extensions of independent component
analysis for multisubject FMRI analysis. Neuroimage 25 (1), 294–311. https://doi.
org/10.1016/j.neuroimage.2004.10.043.
Birn, R.M., Molloy, E.K., Patriat, R., Parker, T., Meier, T.B., Kirk, G.R., et al., 2013. The
effect of scan length on the reliability of resting-state fMRI connectivity estimates.
Neuroimage 83, 550–558. https://doi.org/10.1016/j.neuroimage.2013.05.099.
Boubela, R.N., Kalcher, K., Nasel, C., Moser, E., 2014. Scanning fast and slow: current
limitations of 3 Tesla functional MRI and future potential. Front. Phys. 2. https://doi.
org/10.3389/fphy.2014.00001.
Bozovic, D., Racic, M., Ivkovic, N., 2013. Salivary Cortisol Levels as a Biological Marker of
Stress Reaction. Med. Arch. 67, 374. https://doi.org/10.5455/medarh.2013.67.374-
377.
Buckner, R.L., Andrews-Hanna, J.R., Schacter, D.L., 2008. The brain's default network:
Anatomy, function, and relevance to disease. Ann. N. Y. Acad. Sci. 1124, 1–38.
https://doi.org/10.1196/annals.1440.011.
Caballero-Gaudes, C., Reynolds, R.C., 2017. Methods for cleaning the BOLD fMRI signal.
Neuroimage 154, 128–149. https://doi.org/10.1016/j.neuroimage.2016.12.018.
Cole, M.W., Anticevic, A., Repovs, G., Barch, D., 2011. Variable global dysconnectivity
and individual differences in schizophrenia. Biol. Psychiatry 70 (1), 43–50. https://
doi.org/10.1016/j.biopsych.2011.02.010.
Cole, M.W., Yarkoni, T., Repovs, G., Anticevic, A., Braver, T.S., 2012. Global connectivity
of prefrontal cortex predicts cognitive control and intelligence. J. Neurosci. 32 (26),
8988–8999. https://doi.org/10.1523/JNEUROSCI.0536-12.2012.
Cousijn, H., Rijpkema, M., Qin, S., van Marle, H.J.F., Franke, B., Hermans, E.J., et al.,
2010. Acute stress modulates genotype effects on amygdala processing in humans.
Proc. Natl. Acad. Sci. Unit. States Am. 107 (21), 9867–9872. https://doi.org/10.
1073/pnas.1003514107.
Craig, a. D., 2002. How do you feel? Interoception: the sense of the physiological
condition of the body. Nat. Rev. Neurosci. 3 (8), 655–666. https://doi.org/10.1038/
nrn894.
Critchley, H.D., Wiens, S., Rotshtein, P.,
€
Ohman, A., Dolan, R.J., 2004. Neural systems
supporting interoceptive awareness. Nat. Neurosci. 7 (2), 189–195. https://doi.org/
10.1038/nn1176.
De Kloet, E.R., Jo€
els, M., Holsboer, F., 2005. Stress and the brain: From adaptation to
disease. Nat. Rev. Neurosci. 6, 463–475. https://doi.org/10.1038/nrn1683.
Dosenbach, N.U.F., Fair, D.A., Miezin, F.M., Cohen, A.L., Wenger, K.K.,
Dosenbach, R.A.T., et al., 2007. Distinct brain networks for adaptive and stable task
control in humans. Proc. Natl. Acad. Sci. Unit. States Am. 104 (26), 11073–11078.
https://doi.org/10.1073/pnas.0704320104.
Etkin, A., Egner, T., Kalisch, R., 2011. Emotional processing in anterior cingulate and
medial prefrontal cortex. Trends Cognit. Sci. 15, 85–93. https://doi.org/10.1016/j.
tics.2010.11.004.
Etkin, A., Wager, T.D., 2007. Functional neuroimaging of anxiety: a meta-analysis of
emotional processing in PTSD, social anxiety disorder, and specific phobia. Am. J.
Psychiatry 164 (10), 1476–1488. https://doi.org/10.1176/appi.ajp.2007.07030504.
Finn, E.S., Scheinost, D., Finn, D.M., Shen, X., Papademetris, X., Constable, R.T., 2017.
Can brain state be manipulated to emphasize individual differences in functional
connectivity? Neuroimage 160, 140–151. https://doi.org/10.1016/j.neuroimage.
2017.03.064.
Friston, K.J., Williams, S., Howard, R., Frackowiak, R.S.J., Turner, R., 1996. Movement-
Related effects in fMRI time-series. Magn. Reson. Med. 35 (3), 346–355. https://doi.
org/10.1002/mrm.1910350312.
Gignac, G.E., Szodorai, E.T., 2016. Effect size guidelines for individual differences
researchers. Pers. Indiv. Differ. 102, 74–78. https://doi.org/10.1016/j.paid.2016.06.
069.
Gonzalez-Castillo, J., Hoy, C.W., Handwerker, D.A., Robinson, M.E., Buchanan, L.C.,
Saad, Z.S., Bandettini, P.A., 2015. Tracking ongoing cognition in individuals using
brief, whole-brain functional connectivity patterns. Proc. Natl. Acad. Sci. Unit. States
Am. 112 (28), 8762–8767. https://doi.org/10.1073/pnas.1501242112.
Greve, D.N., Fischl, B., 2009. Accurate and robust brain image alignment using boundary-
based registration. Neuroimage 48 (1), 63–72. https://doi.org/10.1016/j.
neuroimage.2009.06.060.
Hedge, C., Powell, G., Sumner, P., 2017. The reliability paradox: Why robust cognitive
tasks do not produce reliable individual differences. Behav. Res. Methods 1–21.
https://doi.org/10.3758/s13428-017-0935-1.
Hellhammer, D.H., Wüst, S., Kudielka, B.M., 2009. Salivary cortisol as a biomarker in
stress research. Psychoneuroendocrinology 34 (2), 163–171. https://doi.org/10.
1016/j.psyneuen.2008.10.026.
Hemphill, J.F., 2003. Interpreting the Magnitudes of Correlation Coefficients. Am.
Psychol. 58, 78–79. https://doi.org/10.1037/0003-066X.58.1.78.
Henckens, M. J. a G., van Wingen, G. a, Jo€
els, M., Fern
andez, G., 2012. Corticosteroid
induced decoupling of the amygdala in men. Cerebr. Cortex 22 (10), 2336–2345.
New York, N.Y. : 1991. https://doi.org/10.1093/cercor/bhr313.
Hermans, E.J., Henckens, M.J.A.G., Jo€
els, M., Fern
andez, G., 2014. Dynamic adaptation
of large-scale brain networks in response to acute stressors. Trends Neurosci. 37 (6),
304–314. https://doi.org/10.1016/j.tins.2014.03.006.
Hermans, E.J., van Marle, H.J.F., Ossewaarde, L., Henckens, M. J. a. G., Qin, S., van
Kesteren, M.T.R., et al., 2011. Stress-related noradrenergic activity prompts large-
scale neural network reconfiguration. Science 334 (6059), 1151–1153. https://doi.
org/10.1126/science.1209603.
Jenkinson, M., Bannister, P., Brady, M., Smith, S., 2002. Improved optimization for the
robust and accurate linear registration and motion correction of brain images.
Neuroimage 17 (2), 825–841. https://doi.org/10.1016/S1053-8119(02)91132-8.
Jenkinson, M., Smith, S., 2001. A global optimisation method for robust affine
registration of brain images. Med. Image Anal. 5 (2), 143–156. https://doi.org/10.
1016/S1361-8415(01)00036-6.
Jo€
els, M., Baram, T.Z., 2009. The neuro-symphony of stress. Nat. Rev. Neurosci. 10 (6),
459–466. https://doi.org/10.1038/nrn2632.
Kirschbaum, C., Hellhammer, D.H., 1994. Salivary cortisol in psychoneuroendocrine
research: recent developments and applications. Psychoneuroendocrinology 19 (4),
313–333. https://doi.org/10.1016/0306-4530(94)90013-2.
Koch, S.B.J., Klumpers, F., Zhang, W., Hashemi, M.M., Kaldewaij, R., van Ast, V.A., et al.,
2017. The role of automatic defensive responses in the development of posttraumatic
stress symptoms in police recruits: protocol of a prospective study. Eur. J.
Psychotraumatol. 8 (1), 1412226. https://doi.org/10.1080/20008198.2017.
1412226.
Koch, S.B.J., van Zuiden, M., Nawijn, L., Frijling, J.L., Veltman, D.J., Olff, M., 2016.
Aberrant resting-state brain activity in posttraumatic stress disorder: A meta-analysis
and systematic review. Depress. Anxiety 33 (7), 592–605. https://doi.org/10.1002/
da.22478.
Kudielka, B., Kirschbaum, C., 2005. Sex differences in HPA axis responses to stress: a
review. Biol. Psychol. 69, 113–132. https://doi.org/10.1016/j.biopsycho.2004.11.
009.
Luo, Y., Fern
andez, G., Hermans, E., Vogel, S., Zhang, Y., Li, H., Klumpers, F., 2018. How
acute stress may enhance subsequent memory for threat stimuli outside the focus of
attention: DLPFC-amygdala decoupling. Neuroimage 171, 311–322. https://doi.org/
10.1016/j.neuroimage.2018.01.010.
Maron-Katz, A., Vaisvaser, S., Lin, T., Hendler, T., Shamir, R., 2016. A large-scale
perspective on stress-induced alterations in resting-state networks. Sci. Rep. 6
(February), 21503. https://doi.org/10.1038/srep21503.
Marquand, A.F., Haak, K.V., Beckmann, C.F., 2017. Functional corticostriatal connection
topographies predict goal-directed behaviour in humans. Nat. Human Behav. 1 (8).
https://doi.org/10.1038/s41562-017-0146.
McEwen, B.S., 1998. Stress, adaptation, and disease: allostasis and allostatic load. Ann. N.
Y. Acad. Sci. 840 (1), 33–44. https://doi.org/10.1111/j.1749-6632.1998.tb09546.x.
McMenamin, B.W., Pessoa, L., 2015. Discovering networks altered by potential threat
(“anxiety”) using quadratic discriminant analysis. Neuroimage 116, 1–9. https://doi.
org/10.1016/j.neuroimage.2015.05.002.
Menon, V., 2011. Large-scale brain networks and psychopathology: A unifying triple
network model. Trends Cognit. Sci. 15 (10), 483–506. https://doi.org/10.1016/j.tics.
2011.08.003.
Miller, R., Stalder, T., Jarczok, M., Almeida, D.M., Badrick, E., Bartels, M., et al., 2016.
The CIRCORT database: Reference ranges and seasonal changes in diurnal salivary
cortisol derived from a meta-dataset comprised of 15 field studies.
Psychoneuroendocrinology 73 (8), 16–23. https://doi.org/10.1016/j.psyneuen.
2016.07.201.
Murphy, K., Birn, R.M., Bandettini, P.A., 2013. Resting-state fMRI confounds and cleanup.
Neuroimage 80, 349–359. https://doi.org/10.1016/j.neuroimage.2013.04.001.
Nicholson, A.A., Densmore, M., Frewen, P.A., Th
eberge, J., Neufeld, R.W.J.,
McKinnon, M.C., Lanius, R.A., 2015. The dissociative subtype of posttraumatic stress
disorder: unique resting-state functional connectivity of Basolateral and
W. Zhang et al. NeuroImage 189 (2019) 870–877
876
Centromedial Amygdala complexes. Neuropsychopharmacology 40 (10), 2317–2326.
https://doi.org/10.1038/npp.2015.79.
Oathes, D.J., Patenaude, B., Schatzberg, A.F., Etkin, A., 2015. Neurobiological signatures
of anxiety and depression in resting-state functional magnetic resonance imaging.
Biol. Psychiatry 77 (4), 385–393. https://doi.org/10.1016/j.biopsych.2014.08.006.
Parkes, L., Fulcher, B., Yücel, M., Fornito, A., 2018. NeuroImage An evaluation of the ef fi
cacy , reliability , and sensitivity of motion correction strategies for resting-state
functional MRI. Neuroimage 171, 415–436. July 2017. https://doi.org/10.1016/j.
neuroimage.2017.12.073.
Pruim, R.H.R., Mennes, M., van Rooij, D., Llera, A., Buitelaar, J.K., Beckmann, C.F., 2015.
ICA-AROMA: A robust ICA-based strategy for removing motion artifacts from fMRI
data. Neuroimage 112, 267–277. https://doi.org/10.1016/j.neuroimage.2015.02.
064.
Quaedflieg, C.W.E.M., van de Ven, V., Meyer, T., Siep, N., Merckelbach, H., Smeets, T.,
2015. Temporal dynamics of stress-induced alternations of intrinsic amygdala
connectivity and neuroendocrine levels. PLoS One 10 (5) e0124141. https://doi.org/
10.1371/journal.pone.0124141.
Reschke-Hern
andez, A.E., Okerstrom, K.L., Bowles Edwards, A., Tranel, D., 2017. Sex and
stress: Men and women show different cortisol responses to psychological stress
induced by the Trier social stress test and the Iowa singing social stress test.
J. Neurosci. Res. 95 (1–2), 106–114. https://doi.org/10.1002/jnr.23851.
Rotge, J.Y., Lemogne, C., Hinfray, S., Huguet, P., Grynszpan, O., Tartour, E., et al., 2015.
A meta-analysis of the anterior cingulate contribution to social pain. Soc. Cognit.
Affect Neurosci. 10 (1), 19–27. https://doi.org/10.1093/scan/nsu110.
Satterthwaite, T.D., Elliott, M.A., Gerraty, R.T., Ruparel, K., Loughead, J., Calkins, M.E.,
et al., 2013. An improved framework for confound regression and filtering for control
of motion artifact in the preprocessing of resting-state functional connectivity data.
Neuroimage 64 (1), 240–256. https://doi.org/10.1016/j.neuroimage.2012.08.052.
Sch€
onbrodt, F.D., Perugini, M., 2013. At what sample size do correlations stabilize?
J. Res. Pers. 47 (5), 609–612. https://doi.org/10.1016/j.jrp.2013.05.009.
Schwabe, L., Haddad, L., Schachinger, H., 2008. HPA axis activation by a socially
evaluated cold-pressor test. Psychoneuroendocrinology 33 (6), 890–895. https://doi.
org/10.1016/j.psyneuen.2008.03.001.
Seeley, W.W., Menon, V., Schatzberg, A.F., Keller, J., Glover, G.H., Kenna, H., et al., 2007.
Dissociable intrinsic connectivity networks for salience processing and executive
control. J. Neurosci. 27 (9), 2349–2356. https://doi.org/10.1523/JNEUROSCI.5587-
06.2007.
Shenhav, A., Botvinick, M.M., Cohen, J.D., 2013. The expected value of control: An
integrative theory of anterior cingulate cortex function. Neuron 79, 217–240. https://
doi.org/10.1016/j.neuron.2013.07.007.
Shirer, W.R., Ryali, S., Rykhlevskaia, E., Menon, V., Greicius, M.D., 2012. Decoding
subject-driven cognitive states with whole-brain connectivity patterns. Cerebr. Cortex
22 (1), 158–165. https://doi.org/10.1093/cercor/bhr099.
Smith, S.M., Nichols, T.E., 2009. Threshold-free cluster enhancement: Addressing
problems of smoothing, threshold dependence and localisation in cluster inference.
Neuroimage 44 (1), 83–98. https://doi.org/10.1016/j.neuroimage.2008.03.061.
Sripada, R.K., King, A.P., Welsh, R.C., Garfinkel, S.N., Wang, X., Sripada, C.S., Liberzon, I.,
2012. Neural dysregulation in posttraumatic stress disorder: evidence for disrupted
equilibrium between salience and default mode brain networks. Psychosom. Med. 74
(9), 904–911. https://doi.org/10.1097/PSY.0b013e318273bf33.
Vaisvaser, S., Lin, T., Admon, R., Podlipsky, I., Greenman, Y., Stern, N., et al., 2013.
Neural traces of stress: cortisol related sustained enhancement of amygdala-
hippocampal functional connectivity. Front. Hum. Neurosci. 7, 313. https://doi.org/
10.3389/fnhum.2013.00313.
Vaisvaser, S., Modai, S., Farberov, L., Lin, T., Sharon, H., Gilam, A., et al., 2016. Neuro-
epigenetic indications of acute stress response in humans: The case of microRNA-29c.
PLoS One 11 (1), 1–17. https://doi.org/10.1371/journal.pone.0146236.
van Oort, J., Tendolkar, I., Hermans, E.J., Mulders, P.C., Beckmann, C.F., Schene, A.H.,
et al., 2017. How the brain connects in response to acute stress: A review at the
human brain systems level. Neurosci. Biobehav. Rev. 83 (April), 281–297. https://
doi.org/10.1016/j.neubiorev.2017.10.015.
Vogel, S., Klumpers, F., Krugers, H.J., Fang, Z., Oplaat, K.T., Oitzl, M.S., et al., 2015.
Blocking the mineralocorticoid receptor in humans prevents the stress-induced
enhancement of centromedial amygdala connectivity with the Dorsal Striatum.
Neuropsychopharmacology 40 (4), 947–956. https://doi.org/10.1038/npp.2014.
271.
Vul, E., Harris, C., Winkielman, P., Pashler, H., 2009. Puzzlingly High Correlations in
fMRI Studies of Emotion, Personality, and Social Cognition1. Perspect. Psychol. Sci. 4
(3), 274–290. https://doi.org/10.1111/j.1745-6924.2009.01125.x.
Watson, D., Clark, L. a., Tellegan, a., 1988. Worksheet 3.1 the positive and negative affect
schedule (PANAS; Watson et al., 1988) PANAS questionnaire. J. Pers. Soc. Psychol.
54, 1063–1070. https://doi.org/10.1521/soco_2012_1006.
Winkler, A.M., Ridgway, G.R., Webster, M.A., Smith, S.M., Nichols, T.E., 2014.
Permutation inference for the general linear model. Neuroimage 92, 381–397.
https://doi.org/10.1016/j.neuroimage.2014.01.060.
Young, C.B., Raz, G., Everaerd, D., Beckmann, C.F., Tendolkar, I., Hendler, T., et al., 2016.
Dynamic shifts in large-scale brain network balance as a function of arousal.
J. Neurosci. 37, 281–290. https://doi.org/10.1523/JNEUROSCI.1759-16.2016.
Zu Eulenburg, P., Caspers, S., Roski, C., Eickhoff, S.B., 2012. Meta-analytical definition
and functional connectivity of the human vestibular cortex. Neuroimage 60 (1),
162–169. https://doi.org/10.1016/j.neuroimage.2011.12.032.
W. Zhang et al. NeuroImage 189 (2019) 870–877
877