Altered Functional Connectivity between Emotional and
Cognitive Resting State Networks in Euthymic Bipolar I
Disorder Patients
Giannis Lois, Julia Linke, Michèle Wessa*
Department of Clinical Psychology and Neuropsychology, Institute of Psychology, Johannes Gutenberg-University Mainz, Mainz, Germany
Abstract
Bipolar disorder is characterized by a functional imbalance between hyperactive ventral/limbic areas and hypoactive dorsal/
cognitive brain regions potentially contributing to affective and cognitive symptoms. Resting-state studies in bipolar
disorder have identified abnormal functional connectivity between these brain regions. However, most of these studies
used a seed-based approach, thus restricting the number of regions that were analyzed. Using data-driven approaches,
researchers identified resting state networks whose spatial maps overlap with frontolimbic areas such as the default mode
network, the frontoparietal networks, the salient network, and the meso/paralimbic network. These networks are specifically
engaged during affective and cognitive tasks and preliminary evidence suggests that functional connectivity within and
between some of these networks is impaired in bipolar disorder. The present study used independent component analysis
and functional network connectivity approaches to investigate functional connectivity within and between these resting
state networks in bipolar disorder. We compared 30 euthymic bipolar I disorder patients and 35 age- and gender-matched
healthy controls. Inter-network connectivity analysis revealed increased functional connectivity between the meso/
paralimbic and the right frontoparietal network in bipolar disorder. This abnormal connectivity pattern did not correlate
with variables related to the clinical course of the disease. The present finding may reflect abnormal integration of affective
and cognitive information in ventral-emotional and dorsal-cognitive networks in euthymic bipolar patients. Furthermore,
the results provide novel insights into the role of the meso/paralimbic network in bipolar disorder.
Citation: Lois G, Linke J, Wessa M (2014) Altered Functional Connectivity between Emotional and Cognitive Resting State Networks in Euthymic Bipolar I Disorder
Patients. PLoS ONE 9(10): e107829. doi:10.1371/journal.pone.0107829
Editor: Andreas Reif, University of Wuerzburg, Germany
Received October 7, 2013; Accepted August 23, 2014; Published October 24, 2014
Copyright:  2014 Lois et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The study was funded by the Deutsche Forschungsgemeinschaft (SFB 636/C6). The funders had no role in study design, data collection and analysis,
decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* Email: wessa@uni-mainz.de
Introduction ICA is a data-driven approach that identifies temporally coherent
patterns of BOLD signal that are maximally independent from
Bipolar disorder (BD) is a heterogeneous disease characterized each other. ICA takes into account the BOLD signal from the
by acute dysfunctional mood states, alternating between mania whole brain to generate functional maps of different brain
(BD-I) or hypomania (BD-II) and depression, and related to networks [17].
dysfunctional emotion generation and regulation [1]. Further, BD Studies that employed ICA to investigate functional connectiv-
patients suffer from cognitive impairments such as impaired verbal ity during resting state were able to delineate a dozen of resting
memory, deficits in executive functions and attentional deficits state networks (RSNs) that are consistent across subjects and
which are also present during remission [2,3].
sessions [18,19] and show high concordance with measures of
It has been proposed that the neural mechanisms underlying structural connectivity [20]. Additionally, the recently developed
dysfunctional emotion regulation as well as cognitive impairments
functional network connectivity (FNC) technique provides the
in BD are related to hypoactive ventral prefrontal areas that exert
possibility to quantify functional interactions between RSNs.
diminished top-down control on limbic brain structures resulting
Interestingly, some of the identified RSNs are engaged during
in hyperactivity of these structures [1,4–7]. This disturbed
tasks that target affective and cognitive processes [17,21]. For
frontolimbic functional connectivity (FC) has been shown during
instance, the meso/paralimbic network (MPN) [22,23] or else
cognitive tasks [8], emotional tasks [9–11], and during resting state
described as the medial temporal lobe network [24], composed of
[12–16] in both symptomatic and euthymic BD patients.
amygdala, hippocampus, parahippocampal gyrus, and temporal
Two main approaches to investigate functional brain connec-
poles, is implicated in processing of emotional information and
tivity are the seed-based analysis (SBA) and independent
interoceptive awarenenss [21,24], whereas the left and right
component analysis (ICA). SBA is a hypothesis-driven approach
frontoparietal networks (FPNs) are implicated in cognitive control
that correlates the time-series of the blood-oxygen-level-dependent
and attention (right FPN) or language processes and working
(BOLD) signal of one brain region ‘‘seed’’ with the time-series of
memory (left FPN) [17,21] and are comprised of lateral prefrontal
all other brain regions, resulting in a map that defines the
regions and inferior parietal cortex [25,26]. Moreover, the salient
functional connections of the predefined brain region. In contrast,
PLOS ONE | www.plosone.org 1 October 2014 | Volume 9 | Issue 10 | e107829
Resting State Connectivity in Bipolar Disorder
network (SN) involved in the detection of salient stimuli [27] is between functional connectivity measures and some aspects of the
mainly composed of the anterior insula and the anterior cingulate history of the disease (i.e. onset of illness, time in remission,
cortex [28]. The SN and the FPNs are broadly engaged during a number of manic and depressive episodes, and history of psychotic
wide range of cognitively demanding tasks [28]. In contrast to symptoms) and medication. We focused on these variables as they
these networks, the default mode network (DMN), which is describe sufficiently the clinical course of BD and some of them
comprised of posterior cingulate cortex/precuneus, ventro-medial seem to correlate with behavioral variables [38,39] and with brain
prefrontal cortex (vmPFC) and bilateral angular gyri, is strongly activation and connectivity patterns [5,13] within euthymic BD
associated with self-referential activities and is deactivated during patients.
tasks that are directed towards external stimuli [29,30]. Despite the
differences between these RSNs, recent findings showed that the Materials and Methods
SN, the DMN and the FPNs functionally interact with each other
[28,31] and that deficits in engagement and disengagement of Ethical statement
these 3 networks may be relevant to cognitive and affective The ethics committee of the Medical Faculty Mannheim of the
dysfunctions [27,32]. Taking into account the association of these University Heidelberg approved the study and all participants
RSNs with affective and cognitive functions known to be impaired gave written informed consent before study participation. Partic-
in BD and given the overlap between the spatial maps of these ipants/patients having legal guardians or caretakers were excluded
RSNs and frontolimbic areas that are affected in BD, it can be from study participation. Only participants/patients, whose
argued that patterns of functional connectivity within or between capacity to consent was not compromised were investigated in
these networks may be impaired in BD. the present study. The capacity to consent was evaluated by senior
The majority of RS-fMRI studies in BD demonstrated clinicians as part of the diagnostic interview as well as by cognitive
abnormal ventral PFC connectivity with amygdala [14,33] or testing (forward and backward digit-span task, Culture Fair
other subcortical areas (i.e. thalamus, striatum) [12] using SBA. To Intelligence Test (CFT-20)), which was in the normal range for
date, only 4 RS-fMRI studies have employed ICA and investigat- all participants. All potential participants who declined to
ed within or between-networks connectivity in adult BD-I patients. participate or otherwise did not participate were not disadvan-
One study focused on the DMN and reported reduced engage- taged in any other way by not participating in the study. The study
ment of the medial PFC within the DMN in manic BD patients has been conducted according to the principles expressed in the
[34]. Another RS-fMRI study compared within-network func- Declaration of Helsinki.
tional connectivity between psychotic BD patients, schizophrenia
patients, their respective first degree relatives and healthy subjects Participants
using the ICA as a tool to explore potential psychosis We assessed 30 patients with bipolar I disorder (BD) who were
endophenotypes [22]. They identified networks that may be recruited through epidemiological studies at the Central Institute
implicated in BD (e.g. DMN, MPN, SN and FPNs) and they of Mental Health in Mannheim, Germany, or local support
reported aberrant connectivity in psychotic BD patients within the groups. Thirty five healthy controls (HC) were recruited from the
posterior DMN and the MPN. These abnormalities were shared registry office of the city of Mannheim. Demographic differences
with schizophrenia patients but not BD relatives suggesting that between groups were not statistically significant (Table 1).
these findings may be related to the common psychotic symptoms Exclusion criteria for all participants were age under 18,
between the two clinical groups. Using a largely overlapping neurological disorder or head trauma with unconsciousness, and
sample, the same research group investigated inter-networks common MRI exclusion criteria. Patients were also excluded if
connectivity in BD [23]. They found increased connectivity they fulfilled criteria of another Axis-I mental disorder as defined
between the MPN and the SN in psychotic BD patients. This by the DSM-IV within the last 6 months and if they had lifetime
increased connectivity uniquely distinguished this group from diagnosis of rapid cycling, schizoaffective disorder or schizophre-
healthy BD relatives and schizophrenia patients and correlated nia. To increase ecological validity of the patient sample, patients
positively with the negative mood symptoms of BD patients [23]. with lifetime (not current) Axis-I disorders were not excluded.
Finally, a recent RS-fMRI study investigating both within- and More specifically, 7 patients met the criteria for lifetime history of
between-network connectivity in BD and borderline personality substance abuse or dependence, and 3 patients met the criteria for
disorder patients revealed aberrant connectivity patterns in BD lifetime anxiety and eating disorders (i.e. agoraphobia, social
between networks implicated in self-referential processing [35]. anxiety disorder, and bulimia). HCs were excluded if they fulfilled
However, this study did not examine subcortical networks (i.e. the criteria for any lifetime or current DSM-IV axis I mental
MPN). disorder or took any psychotropic medication. The ethics
Given the sparse empirical evidence of RS-fMRI studies that committee of the Medical Faculty of Mannheim of the University
employed a data-driven approach in BD, the present study sought Heidelberg approved the study and all participants gave written
to investigate functional connectivity within and between RSNs in informed consent before study participation.
BD using ICA. Our study extends prior work that used similar
methodology by examining euthymic BD patients with minimum Diagnostic assessment
residual mood symptoms and no current psychotic symptoms. We Main and comorbid diagnoses (for patients) and exclusion
focused on the DMN, the FPNs (i.e. right and left), the SN, and the criteria (for HCs) were evaluated based on the Structured Clinical
MPN as these networks or brain regions within these networks Interview for DSM-IV Axis I Disorders. The Young Mania Rating
have been previously implicated in BD [5,22,23,34,36,37]. Based Scale, the Hamilton Rating Scale for Depression (HAM-D), and
on the neurobiological model of BD and previous findings from the Beck Depression Inventory (BDI) were administered to assess
RS-fMRI studies in BD, we expected to find altered functional residual mood symptoms. Based on the clinical interview, patients
connectivity within and between these RSNs that may reflect the were euthymic for at least two months before testing and had
underlying pathophysiology of the disorder. To further investigate minimum residual symptoms (Table 1).
whether abnormal functional connectivity patterns are affected by Variables describing the clinical course of the disease such as the
the clinical course of the disease, we examined the relationship number of past depressive and manic episodes, the age at illness
PLOS ONE | www.plosone.org 2 October 2014 | Volume 9 | Issue 10 | e107829
Resting State Connectivity in Bipolar Disorder
Table 1. Demographic and Clinical Characteristics of Participants.
Characteristics BD (N = 30) HC (N = 35) Statistics p-value
Gender (Female/Male) 17/13 20/15 x2 = 0.34 0.853
Age, years, mean (SD) 40.83 (9.43) 41.94 (8.36) T(65) = 0.5 0.62
Years of education, mean (SD) 14.97 (2.58) 14.89 (2.31) T(65) = 0.6 0.54
Current Symptoms, [range], mean (SD)
YMRSa Score [0–5] 1.3 (1.6) [0–2] 0.11 (0.40) T(65) = 3.9 0.00
BDIb Score [0–20] 6.7 (5.94) [0–12] 1.26 (2.44) T(65) = 5.5 0.00
HAM-Dc Score [0–5] 0.83 (1.17) [0–2] 0.14 (0.43) T(65) = 3.2 0.02
History of illness
No. of past depressive episodes, mean (SD) 3.63 (3.12)
No. of past manic episodes, mean (SD) 2.66 (2.48)
Age at illness onset, years, mean (SD) 23.73 (9.51)
Time in remission, years, mean (SD) 3.08 (3.44)
Medication load, mean (SD) 3.10 (2.35)
History of psychotic symptoms (Yes/No) 13/17
aYMRS = Young Mania Rating Scale,
bBDI = Beck Depression Inventory,
cHAM-D = Hamilton Depression Rating Scale.
doi:10.1371/journal.pone.0107829.t001
onset, the time in remission, and the presence of psychotic state, participants were instructed to lie still with their eyes closed
symptoms were acquired for every patient. We also verified that and not to fall asleep for 5 minutes. 120 whole-brain volumes were
the medication status of all patients had been stable during the past acquired with the initial 4 images being discarded to allow for T2
6 months. Among the patients 4 did not take any medication, 25 stabilization effects.
patients took mood stabilizers (lithium: n = 6, valproate: n = 11,
lamotrigin: n = 6, carbamazepin: n = 2), 11 patients took antide- Data Analysis
pressants (venlaflaxin: n = 3, citalopram: n = 2, sertralin: n = 2, Pre-Processing. Imaging data were preprocessed according
duloxetin: n = 1, clomipramin: n = 1, mirtazapin: n = 1, opipra- to standard procedures using SPM8 (http://www.fil.ion.ucl.ac.uk/
mol: n = 1) and 11 patients took antipsychotic medication spm). Preprocessing involved realignment to the mean image of
(quetiapin: n = 7, perazin: n = 1, risperidon: n = 1, olanzapin: each run, slice-timing, and normalization into Montreal Neuro-
n = 1, aripiprazol: n = 1) at the time of scanning. In accordance to logical Institute standardized space (http://www.mni.mcgill.ca).
previous studies [11], we coded the number and dosage of each During normalization, the images were resampled every 3 mm
medication and calculated a composite measure of total medica- using sinc interpolation and smoothed with a 96969 mm
tion load. For antidepressants and mood stabilizers, we categorized Gaussian kernel to decrease spatial noise. One BD patient and
each medication into low-dose (1 or 2 levels) or high-dose (3 or 4 one HC were excluded from the initial sample (i.e. 31 BD patients
levels) groupings as previously performed [40]. We added a no- and 36 HC) due to excessive movement during the fMRI scans
dose subtype for those not taking these medications. Using a (i.e. more than 3 mm translation or 3 degrees rotation and 2 mm
similar procedure, we estimated the clinical equivalency of the or degrees in the first derivatives of motion parameters). In order
antipsychotic medication to Chlorpromazine dose equivalent, to exclude the possible confound of different motion artefacts in
coding them as 0 (no medication), 1 (equal to or below the the compared groups, we additionally estimated the maximum
chlorpromazine dose equivalent), or 2 (above the chlorpromazine absolute displacement of each brain volume as compared to the
dose equivalent) [41]. A composite measure of medication load previous volume from the translation parameters in the x (left/
was generated by summing all individual medication codes for the right), y (anterior/posterior), and z (superior/inferior) directions
two categories for each individual participant. The demographic [42]. No group differences were present in the maximum motion
and clinical characteristics of the two groups of participants are (t = 1,32, df = 65, p = 0.19). This variable was used as nuisance
shown in Table 1. regressor in the group comparisons.
Group-ICA. The Group ICA of fMRI toolbox (GIFT;
Data Acquisition version 1.3i; http://icatb.sourceforge.net) [43] was used to carry
Data were acquired on a 3-T whole body scanner (Magnetom out a group ICA in the preprocessed and normalized data. Using
Trio, Siemens Medical Solutions, Erlangen, Germany). We the minimum description length criteria to determine the number
conducted one high-resolution T1-weighted 3 dimensional MRI of independent components (ICs), we were not able to identify all
sequence (slice thickness = 1.1 mm, field of view = 25662406 the RSNs of interest. It has been shown that increasing the
176 mm3, TR = 2.3 s, TE = 2.98 ms). We acquired 40 gradient- number of ICs yields refined components that correspond to
echo T2*-weighted slices (slice thickness = 2.3 mm) per volume known anatomical and functional segmentations [17,44,45].
with the following parameters: TR = 2.7 s, flip angle = 90u, Following previous procedures [44,45], we gradually increased
TE = 27 ms, field of view = 220 mm2, matrix = 96696, voxel the number of ICs to 40 until all networks of interest were
size = 2.3 mm62.3 mm62.3 mm. During acquisition of resting- identified. An initial data reduction step used principal component
PLOS ONE | www.plosone.org 3 October 2014 | Volume 9 | Issue 10 | e107829
Resting State Connectivity in Bipolar Disorder
analysis on the subject-specific data and was followed by an IC
estimation that produced 40 time courses and spatial maps with
the Infomax algorithm [46]. This algorithm was repeated 20 times
in Icasso, each time with different initial conditions. Resulting
components of different runs were clustered to estimate the
reliability of the decomposition [47]. The index Iq, which ranges
from 0 to 1, reflects the difference between intra-cluster and extra-
cluster similarity [47]. Each of the 40 components had a cluster
quality index greater than 0.8. Group-level spatial maps were
estimated as the centrotypes of component clusters to reduce
sensitivity to initial algorithm parameters. These components were
identified in the group of 65 subjects without differentiating
between HCs and BD patients to ensure that the same
components are identified in each group. Group-level ICs (both
spatial maps and time courses) were then back-reconstructed for
each subject using the GIGA method [43,48]. Therefore, the
subject-specific time course of each component represented a
pattern of synchronized brain activity, whose coherency pattern
across voxels was represented in the associated spatial map.
Component intensity values were then z-scaled to provide
normalized scores across subjects.
Based on previous ICA studies and to validate our results, we
repeated the Group ICA analysis with 75 ICs components; this
high model order yield refined components that correspond to
known anatomical and functional segmentation (see Analysis S1).
Identifying Valid RSNs. To identify valid RSNs, we first
estimated the voxel-wise spatial overlap of the ICs with SPM’s
standard tissue-type probability maps for gray matter, white
matter and cerebrospinal fluid (http://imaging.mrc-cbu.cam.ac.
uk/imaging/Templates) using Pearson’s correlation (Table S1).
Subsequently, group-level ICs were spatially sorted in GIFT
toolbox using templates of the RSNs of interest (i.e. DMN, right
and left FPN, SN, MPN; Table S1). These templates were based
on whole-brain task-based co-activation networks, which were
derived from ICA analyses on peak activation coordinates
archived in a large neuroimaging database (i.e. BrainMap
Database) [21]. On the basis of their spatial overlap with the
templates, 6 ICs were identified as RSNs of interest. The DMN
was divided into an anterior (aDMN) and a posterior component
(pDMN) (Figure 1A and B) and the FPNs comprised of two
lateralized components (i.e. right and left) (Figure 1C and D). The
MPN and SN were each identified as a single component
(Figure 1E and F). The highest voxel-wise spatial overlap between
the 6 ICs and the corresponding templates were similar to
previously reported values (r = 0.31–0.55; [49,50]). The last step in
the selection procedure was to examine the spectral characteristics
of the 6 selected ICs using the same procedure as in Allen and
colleagues [18]. For each group-level IC, we estimated the
difference between the peak spectral power and minimum spectral
power at frequencies to the right of the peak (i.e. dynamic range),
and the ratio of the integral of power below 0.10 Hz to the integral Figure 1. Resting state networks of interest. Illustration of one-
sample-t-test maps of the anterior and posterior default mode network,
of power between 0.15 and 0.25 Hz (i.e. low frequency to high right and left frontoparietal network, the salience network, and the
frequency power ratio). Visual inspection of the scatter plot of low meso/paralimbic network identified in the control (left column) and
frequency to high frequency power ratio versus dynamic range patient (right column) group. Maps are thresholded at P,0.05 (whole-
confirmed that all 6 ICs are dominated by frequency fluctuations brain FWE corrected). R, Right.
inside the 0.01–0.1 Hz window (Figure S1). doi:10.1371/journal.pone.0107829.g001
Within-network connectivity analysis. Statistical inference
was carried out on the subject-specific z-maps of the 6 RSNs of from the one-sample t test results, which was further masked to
interest using SPM8. For each group, we computed one-sample t- exclude the non-gray matter voxels (Table S2). Group compar-
test to create statistical maps of the 6 components of interest isons were restricted to voxels within these binary masks. Two-
(Figure 1). Threshold for the one-sample t-test was set at p,0.05 sample t-tests were performed to compare the 6 RSNs between the
(Family-wise error, FWE correction) to select only the core nodes two groups (p,0.05 FWE correction). In addition, to account for
of each RSN. Subsequently, we created a binary representation of type I error inflation for testing 6 RSNs, voxels were only
the conjunction of the statistical maps of the two groups, obtained considered significant if they survived a subsequent Bonferroni
PLOS ONE | www.plosone.org 4 October 2014 | Volume 9 | Issue 10 | e107829
Resting State Connectivity in Bipolar Disorder
Figure 2. Functional network connectivity (FNC) between components. Average Fisher’s z-transformed Pearson’s correlation coefficients
from FNC analysis reflecting connectivity between the 6 components of interest (15 possible pairs of components) for the healthy control (HC) and
patient group (BD). Error bars represent the standard error of the mean (SEM). Asterisks (**) depict FDR significant finding.
doi:10.1371/journal.pone.0107829.g002
correction for multiple testing at p,0.008 (0.05/6). Age, gender network connectivity and subsequently we conducted bivariate
and maximum motion were used as nuisance variables in the Pearson’s correlations between the extracted values and the illness-
analysis. related variables within the patient group. Similarly, bivariate
Between-networks connectivity analysis. The subject- Pearson’s correlations were conducted between correlation coef-
specific time courses of each RSN were entered in the FNC ficients estimating FNC between the MPN and right FPN
analysis (http://mialab.mrn.org/software). The FNC estimates the (aberrant pair of networks) and variables related to the clinical
Pearson’s correlation coefficient between pairs of time courses with course of the disease within the BD group.
a maximal lagged correlation approach (i.e. 23 to +3 s lag) [32]. To address the question whether the history of psychotic
Time course data were first band-pass filtered with a Butterworth symptoms influences abnormal connectivity patterns, we divided
filter with cutoff frequencies of.01–.1 Hz [51] and then pairwise the patient group into patients with (n = 13) and without (n = 17)
correlations were computed between the time courses of the 6 history of psychotic symptoms and we performed a two-sample
RSNs of interest, resulting in 15 total pairwise correlations for each T-test to examine between-group differences in functional
subject. Resulting Pearson’s correlation coefficients were trans- connectivity within- and between-networks (limited to aberrant
formed to Fisher’s z values and within-group correlations between connectivity).
networks as well as between-group differences in correlation
coefficients were estimated. To control for multiple comparisons, Results
p-values were thresholded according to a false discovery rate
(FDR) of 0.05. Age gender and maximum motion were used as Components of interest and statistical comparison of
nuisance variables in the analysis. RSNs
Correlations with variables related to the clinical course The 6 RSNs of interest were identified separately for the HC
of the disease. Additionally, we tested the effect of variables and BD group by performing one-sample-t-tests on the subject
related to the clinical course of the disease on abnormal within- specific z-maps (p,0.05, FWE-corrected, Figure 1). Consistent
and between-networks functional connectivity. Using the Mars- with previous studies that used a high order model, the DMN was
BaR toolbox (http://marsbar.sourceforge.net) we extracted the divided into an anterior part (aDMN) that covers the medial PFC/
mean subject-specific z-values of the clusters showing significant ACC (BA 10/32/24) and a posterior part (pDMN) that covers the
between-group differences (p,.05, FDR correction) in the within– precuneus/posterior cingulate cortex and angular gyri (BA 23/39)
PLOS ONE | www.plosone.org 5 October 2014 | Volume 9 | Issue 10 | e107829
Resting State Connectivity in Bipolar Disorder
Table 2. Pearson’s correlation coefficients and p values from FNC analysis.
Component combination HC BD Between groups p (FDR corrected)
MPN-SN 20.166* 20.132* 0.835
MPN- Left FPN 20.086 0.026 0.069
MPN- Right FPN 20.078 0.122* 0.0011
MPN- aDMN 20.032 20.005 0.890
MPN- pDMN 20.049 20.080 0.794
Left FPN-SN 20.114* 20.148* 0.579
Right FPN-SN 0.217* 0.160* 0.335
aDMN-SN 20.322* 20.235* 0.219
pDMN-SN 20.292* 20.160 0.122
Right FPN-Left FPN 0.318* 0.214* 0.098
Left FPN- aDMN 0.128* 0.110* 0.829
Left FPN- pDMN 0.282* 0.350* 0.216
Right FPN- aDMN 20.014 20.019 0.838
Right FPN- pDMN 0.184* 0.163* 0.566
aDMN- pDMN 0.400* 0.290* 0.205
* Denotes significant (p,0.05 FDR corrected) Pearson’s coefficient correlations between components within each group.
1Denotes significant (p,0.05 FDR corrected) group differences in the magnitude of the correlation between components.
doi:10.1371/journal.pone.0107829.t002
(Figure 1A and 1B). The right and left fronto-parietal networks variables related to the clinical course of the disease or medication
mainly covered the inferior parietal cortex (BA 39/40) dorsolateral load within the BD group. There was no correlation between
PFC (BA 45/9) and ventrolateral PFC (BA 46/47) (Figure 1C and illness-related variables or medication load and abnormal
1D). The SN encompassed the anterior insula/lateral orbitofrontal between-networks connectivity patterns even when no correction
cortex (BA 38/45/47) and the supramarginal gyrus (BA 40) for multiple testing was applied. Furthermore, the two-sample T-
(Figure 1E). The MPN included the amygdala, hippocampus, test revealed no differences in functional connectivity between
parahippocampal gyrus, temporal poles, and a part of insular patients with and without history of psychotic symptoms.
cortex (BA 28/30/34/38) (Figure 1F). The location of the FC
peaks and corresponding z-scores of every network are presented Discussion
in Table S3.
There were no significant between-group differences within the The present study extends previous knowledge by investigating
6 RSN of interest. An exploratory analysis without Bonferroni functional connectivity within and between RSNs that have not
correction for testing multiple components also yielded no been examined in detail despite their potential relevance for both
significant between-group differences. affective and cognitive dysfunctions in BD. Compared to previous
RS-fMRI studies in BD that employed an exploratory multivariate
Between network connectivity method, the present study comprised a larger sample of euthymic
The 15 possible pairwise network combinations were tested for BD-I patients without current psychotic symptoms. Analysis of
significant maximal-lagged correlations in each group (Figure 2 between-network connectivity revealed that in contrast to HC
and Table 2). As expected, both groups showed strong positive subjects where activity of the executive network does not correlate
connectivity between the right and left FPN and between the with activity of the MPN, BD patients display an increased
pDMN and aDMN. The SN showed positive FNC with the right interaction between the MPN and the right FPN.
FPN and negative FNC with the left FPN and the two DMNs. To better interpret this finding, we drew upon evidence from
Both groups showed positive connectivity between the pDMN and meta-analytic studies that compared activation maps of different
the two lateralized fronto-parietal networks as well as between the behavioral paradigms with ICA-derived spatial maps of RSNs
aDMN and the left fronto-parietal network. We observed [17,21]. In this context, the positive FNC between the MPN and
significant group differences in FNC between the MPN and the the right FPN in the BD group may reflect abnormal communi-
right FPN (p = .001, FDR-corrected; Figure 2 and 3). Whereas this cation between a limbic-paralimbic network implicated in
combination of networks showed a tendency towards negative emotional processing and a right-lateralized dorsal network
FNC in the HC group (r =20.078), there was significant positive involved in executive functions and cognitive control [21]. In line
FNC between the two networks in the BD group (r = 0.122). Mean with this assumption, neurobiological models of BD proposed that
lag times did not differ between groups for any of the 15 impaired emotion regulation in BD may result from a functional
combinations. imbalance between a dorsal-cognitive network (i.e. dorsal ACC,
dorsal PFC, inferior parietal cortex), and a ventral-emotional
Relationship with variables related to the clinical course network (i.e. amygdala, parahippocampal gyrus, striatum, sub-
of the disease genual cingulate cortex, orbitofrontal cortex) [1,4,52]. In this
We conducted bivariate Pearson’s correlations to investigate the respect, it is noteworthy that a trend-level group-related effect was
relationship between abnormal functional connectivity and also observed for the MPN-left FPN pair (p = .069) in the same
PLOS ONE | www.plosone.org 6 October 2014 | Volume 9 | Issue 10 | e107829
Resting State Connectivity in Bipolar Disorder
Figure 3. Functional network connectivity (FNC) between MPN-right FPN in the individual level. Fisher’s z-transformed Pearson’s
correlation coefficients depicting individual FNC values for the MPN-right FPN pair. Both samples are normally distributed and have equal variance.
The line connects the average FNC values of the two groups.
doi:10.1371/journal.pone.0107829.g003
direction as the main MPN-right FPN finding. Both findings characterizing the clinical course of the disease, namely, the
suggest a general functional imbalance of the limbic network with number of manic and depressive episodes, the age of illness onset,
fronto-parietal/cognitive networks in BD. the time in remission, and the history of psychotic symptoms.
The present finding may be interpreted as impaired top-down Previous research has shown that some of these clinical
control or abnormally increased bottom-up interference between characteristics of BD patients correlate with behavioral variables
these networks through the reciprocal connections between limbic- [38,39] and with brain activation and connectivity patterns [5,13].
paralimbic structures and lateral PFC regions [53]. Independent of The absence of significant correlations in the present sample of
the direction of the interaction (i.e. top-down or bottom-up), this euthymic BD patients with minimum residual mood symptoms
result supports previous evidence from seed-based studies showing and no current psychotic symptoms indicates that aberrant FNC
abnormal FC between limbic structures, such as amygdala, and between the MPN and the right FPN may constitute a trait marker
right PFC at rest [13,16] or during emotional tasks [9,54,55]. Our of BD. Emerging questions, for future studies to address, focus on
results are also in line with a recent meta-analysis showing whether this FC abnormality reflects a developmental failure to
increased ventral-limbic activation and decreased right-lateralized establish healthy prefrontal–limbic modulation early in life which
PFC activation during emotional processing in BD patients [37]. may later result in the onset of the disease [5] and may be present
Aberrant FNC between the MPN and the right FPN in the BD in healthy populations at high risk for developing BD or represents
group did not correlate with medication load or variables
PLOS ONE | www.plosone.org 7 October 2014 | Volume 9 | Issue 10 | e107829
Resting State Connectivity in Bipolar Disorder
common state- and mood-independent effects of previous BD synergistic and independent from each other. Therefore, catego-
episodes. ry-specific medication effects may still confound the present
Contrary to our hypothesis, we did not observe abnormal findings. However, due to possible pharmacological interactions
within-network FC in any of the 6 RSNs of interest also when no between different types of medication the independent impact of
correction for multiple comparisons was applied. In line with the every class of medication on connectivity cannot be directly
lack of significant within-network FC findings in the present study, assessed in the present sample. Future studies using medication-
a recent RS-fMRI study showed significant BD-related impair- based selection criteria for the patient group and a drug-free
ments in between- but not in within-network FC [35]. This baseline are more suitable to address questions about potential
evidence leads to the assumption that BD like schizophrenia may pharmacological effects on brain connectivity.
be mainly characterized by dysconnectivity in large-scale networks Finally, the interpretation of resting state data in light of
[56,57]. previous task-related neuroimaging studies has to be made with
Based on our hypotheses, abnormal FC within and between the caution. It still remains an open question to what extent task-based
RSNs of interest may underlie affective and cognitive symptoms of differences in specific brain regions in BD are reflected in different
the disorder. Therefore, these abnormalities would more likely connectivity patterns of their respective brain networks at rest.
characterize symptomatic rather than euthymic patients. In line Furthermore, to date, there is very sparse evidence linking RSNs
with this idea, previous ICA studies that investigated functional with behavior and cognitive or emotional processes which could
connectivity patterns in BD at rest reported altered connectivity provide functional interpretations of widely observed RSNs.
within the DMN in manic [34] or psychotic [22,23] BD patients. In conclusion, the present study substantially extends prior work
The assumption that these alterations are state-dependent is in BD by employing a data-driven multivariate approach to
further supported by evidence that these abnormal FC patterns are examine euthymic BD-I patients without current psychotic
either shared by schizophrenic patients [22] or correlate with symptoms. We showed abnormal interactions between the MPN
negative mood symptoms [23]. and the right FPN in BD at rest. This abnormality may underlie
Another plausible explanation for the lack of BD-related impaired integration of affective and cognitive processes leading to
abnormal FC in the present study may be the context-dependent dysfunctional emotion regulation. Furthermore, the present results
nature of these abnormalities, which may only be present or more highlight the importance of the MPN in this disorder.
pronounced during cognitive and emotional tasks that engage
these networks and not at rest. Future ICA studies comparing rest Supporting Information
and task conditions should investigate this possibility.
The present findings have to be interpreted in the light of some Figure S1 Scatter plot of low frequency (LF) to high
limitations. First, inherent to the nature of the ICA, there is no frequency (HF) power ratio versus dynamic range for all
optimal way to estimate the number of ICs (problem of components. Red squares represent the 6 components of
dimensionality). Yet, the number of ICs has a significant impact interest selected in the present study.
on the spatial characteristics of the RSNs [58]. Most studies use (TIF)
the Minimum Description Length criteria as standard criteria to
Table S1 Voxel-wise spatial overlap (Pearson’s r) of
determine the number of ICs. However, using these criteria, we
RSNs of interest with gray matter, white matter,
were not able to identify all common RSNs that were of interest
cerebrospinal fluid masks and templates of RSNs.
for the present study. As previous studies showed that detection of
some RSNs requires higher dimensionality models [44], we (XLSX)
decided to gradually increase the dimensionality until the point Table S2 Number of voxels before and after excluding
that all the networks of interest were identified [44]. The aim was non-gray matter voxels.
to find the minimum number of components in which our (XLSX)
hypotheses could be tested. Additionally, we performed the same
analysis with a standard high model order ICA (i.e. 75 ICs), as Table S3 Peak within-network functional connectivity
previous studies have demonstrated that this model order yield for both groups.
refined components that correspond to known anatomical and (XLSX)
functional segmentations [17,44,45]. Importantly, we found Analysis S1 Group Independent component analysis
similar results concerning the MPN with both 40 and 75 ICs with 75 ICs.
(see Analysis S1). (DOCX)
A second limitation is related to the correlation analysis between
abnormal FC patterns and clinical characteristics. The absence of Acknowledgments
correlations may also be related to the lack of power as the sample
was relatively small for this type of analysis and patients The authors thank Johanna Forneck, Janine Heissler, Manuela Glas-
investigated here were relatively healthy leading to little variance brenner, Katja Nitsche, Kristina Schaffer, Claudia Stief, and Birgül Sarun
with regard to symptomatic variables. for their assistance in data acquisition.
Furthermore, medication load was assessed as a composite
measure assuming synergistic rather than antagonistic effects of Author Contributions
the different types of medication namely antidepressants, mood Conceived and designed the experiments: JL MW. Performed the
stabilizers, and antipsychotics. It is rather unlikely that the effects experiments: JL. Analyzed the data: GL. Contributed reagents/materi-
of different classes of medication on brain connectivity are als/analysis tools: GL JL MW. Wrote the paper: GL JL MW.
References
1. Phillips ML, Ladouceur CD, Drevets WC (2008) A neural model of voluntary pathophysiology and neurodevelopment of bipolar disorder. Mol Psychiatry
and automatic emotion regulation: implications for understanding the 13: 829, 833–857.
PLOS ONE | www.plosone.org 8 October 2014 | Volume 9 | Issue 10 | e107829
Resting State Connectivity in Bipolar Disorder
2. Mann-Wrobel MC, Carreno JT, Dickinson D (2011) Meta-analysis of 31. Sridharan D, Levitin DJ, Menon V (2008) A critical role for the right fronto-
neuropsychological functioning in euthymic bipolar disorder: an update and insular cortex in switching between central-executive and default-mode
investigation of moderator variables. Bipolar Disord 13: 334–342. networks. Proc Natl Acad Sci U S A 105: 12569–12574.
3. Sole B, Bonnin CM, Torrent C, Martinez-Aran A, Popovic D, et al. (2012) 32. Jafri MJ, Pearlson GD, Stevens M, Calhoun VD (2008) A method for functional
Neurocognitive impairment across the bipolar spectrum. CNS Neurosci Ther network connectivity among spatially independent resting-state components in
18: 194–200. schizophrenia. Neuroimage 39: 1666–1681.
4. Phillips ML, Drevets WC, Rauch SL, Lane R (2003) Neurobiology of emotion 33. Chepenik LG, Raffo M, Hampson M, Lacadie C, Wang F, et al. (2010)
perception II: Implications for major psychiatric disorders. Biol Psychiatry 54: Functional connectivity between ventral prefrontal cortex and amygdala at low
515–528. frequency in the resting state in bipolar disorder. Psychiatry Res 182: 207–210.
5. Strakowski SM, Adler CM, Almeida J, Altshuler LL, Blumberg HP, et al. (2012) 34. Ongur D, Lundy M, Greenhouse I, Shinn AK, Menon V, et al. (2010) Default
The functional neuroanatomy of bipolar disorder: a consensus model. Bipolar mode network abnormalities in bipolar disorder and schizophrenia. Psychiatry
Disord 14: 313–325. Res 183: 59–68.
6. Wessa M, Kanske P, Linke J (2013) Bipolar disorder: A neural network 35. Das P, Calhoun V, Malhi GS (2014) Bipolar and borderline patients display
perspective on a disorder of emotion and motivation. Restor Neurol Neurosci. differential patterns of functional connectivity among resting state networks.
7. Wessa M, Linke J (2009) Emotional processing in bipolar disorder: behavioural Neuroimage.
and neuroimaging findings. Int Rev Psychiatry 21: 357–367. 36. Chen CH, Suckling J, Lennox BR, Ooi C, Bullmore ET (2011) A quantitative
8. Pompei F, Dima D, Rubia K, Kumari V, Frangou S (2011) Dissociable
meta-analysis of fMRI studies in bipolar disorder. Bipolar Disord 13: 1–15.
functional connectivity changes during the Stroop task relating to risk, resilience
37. Houenou J, Frommberger J, Carde S, Glasbrenner M, Diener C, et al. (2011)
and disease expression in bipolar disorder. Neuroimage 57: 576–582.
Neuroimaging-based markers of bipolar disorder: evidence from two meta-
9. Foland LC, Altshuler LL, Bookheimer SY, Eisenberger N, Townsend J, et al.
analyses. J Affect Disord 132: 344–355.
(2008) Evidence for deficient modulation of amygdala response by prefrontal
38. Linke J, Sonnekes C, Wessa M (2011) Sensitivity to positive and negative
cortex in bipolar mania. Psychiatry Res 162: 27–37.
feedback in euthymic patients with bipolar I disorder: the last episode makes the
10. Townsend JD, Torrisi SJ, Lieberman MD, Sugar CA, Bookheimer SY, et al.
difference. Bipolar Disord 13: 638–650.
(2013) Frontal-amygdala connectivity alterations during emotion downregula-
tion in bipolar I disorder. Biol Psychiatry 73: 127–135. 39. Wright KA, Lam D, Brown RG (2008) Dysregulation of the behavioral
11. Versace A, Thompson WK, Zhou D, Almeida JR, Hassel S, et al. (2010) activation system in remitted bipolar I disorder. J Abnorm Psychol 117: 838–
Abnormal left and right amygdala-orbitofrontal cortical functional connectivity 848.
to emotional faces: state versus trait vulnerability markers of depression in 40. Sackeim HA (2001) The definition and meaning of treatment-resistant
bipolar disorder. Biol Psychiatry 67: 422–431. depression. J Clin Psychiatry 62 Suppl 16: 10–17.
12. Anand A, Li Y, Wang Y, Lowe MJ, Dzemidzic M (2009) Resting state 41. Gardner DM, Murphy AL, O’Donnell H, Centorrino F, Baldessarini RJ (2010)
corticolimbic connectivity abnormalities in unmedicated bipolar disorder and International consensus study of antipsychotic dosing. Am J Psychiatry 167:
unipolar depression. Psychiatry Res 171: 189–198. 686–693.
13. Anticevic A, Brumbaugh MS, Winkler AM, Lombardo LE, Barrett J, et al. 42. Van Dijk KR, Sabuncu MR, Buckner RL (2012) The influence of head motion
(2013) Global prefrontal and fronto-amygdala dysconnectivity in bipolar I on intrinsic functional connectivity MRI. Neuroimage 59: 431–438.
disorder with psychosis history. Biol Psychiatry 73: 565–573. 43. Calhoun VD, Adali T, Pearlson GD, Pekar JJ (2001) A method for making
14. Chai XJ, Whitfield-Gabrieli S, Shinn AK, Gabrieli JD, Nieto Castanon A, et al. group inferences from functional MRI data using independent component
(2011) Abnormal medial prefrontal cortex resting-state connectivity in bipolar analysis. Hum Brain Mapp 14: 140–151.
disorder and schizophrenia. Neuropsychopharmacology 36: 2009–2017. 44. Abou-Elseoud A, Starck T, Remes J, Nikkinen J, Tervonen O, et al. (2010) The
15. Mamah D, Barch DM, Repovs G (2013) Resting state functional connectivity of effect of model order selection in group PICA. Hum Brain Mapp 31: 1207–
five neural networks in bipolar disorder and schizophrenia. J Affect Disord. 1216.
16. Torrisi S, Moody TD, Vizueta N, Thomason ME, Monti MM, et al. (2013) 45. Kiviniemi V, Starck T, Remes J, Long X, Nikkinen J, et al. (2009) Functional
Differences in resting corticolimbic functional connectivity in bipolar I euthymia. segmentation of the brain cortex using high model order group PICA. Hum
Bipolar Disord 15: 156–166. Brain Mapp 30: 3865–3886.
17. Smith SM, Fox PT, Miller KL, Glahn DC, Fox PM, et al. (2009) 46. Bell AJ, Sejnowski TJ (1995) An information-maximization approach to blind
Correspondence of the brain’s functional architecture during activation and separation and blind deconvolution. Neural Comput 7: 1129–1159.
rest. Proc Natl Acad Sci U S A 106: 13040–13045. 47. Himberg J, Hyvarinen A, Esposito F (2004) Validating the independent
18. Allen EA, Erhardt EB, Damaraju E, Gruner W, Segall JM, et al. (2011) A components of neuroimaging time series via clustering and visualization.
baseline for the multivariate comparison of resting-state networks. Front Syst Neuroimage 22: 1214–1222.
Neurosci 5: 2. 48. Erhardt EB, Rachakonda S, Bedrick EJ, Allen EA, Adali T, et al. (2011)
19. Damoiseaux JS, Rombouts SA, Barkhof F, Scheltens P, Stam CJ, et al. (2006) Comparison of multi-subject ICA methods for analysis of fMRI data. Hum
Consistent resting-state networks across healthy subjects. Proc Natl Acad Brain Mapp 32: 2075–2095.
Sci U S A 103: 13848–13853. 49. Assaf M, Jagannathan K, Calhoun VD, Miller L, Stevens MC, et al. (2010)
20. Greicius MD, Supekar K, Menon V, Dougherty RF (2009) Resting-state Abnormal functional connectivity of default mode sub-networks in autism
functional connectivity reflects structural connectivity in the default mode spectrum disorder patients. Neuroimage 53: 247–256.
network. Cereb Cortex 19: 72–78. 50. von dem Hagen EA, Stoyanova RS, Baron-Cohen S, Calder AJ (2013) Reduced
21. Laird AR, Fox PM, Eickhoff SB, Turner JA, Ray KL, et al. (2011) Behavioral functional connectivity within and between ‘social’ resting state networks in
interpretations of intrinsic connectivity networks. J Cogn Neurosci 23: 4022–
autism spectrum conditions. Soc Cogn Affect Neurosci 8: 694–701.
4037.
51. Cordes D, Haughton VM, Arfanakis K, Carew JD, Turski PA, et al. (2001)
22. Khadka S, Meda SA, Stevens MC, Glahn DC, Calhoun VD, et al. (2013) Is
Frequencies contributing to functional connectivity in the cerebral cortex in
Aberrant Functional Connectivity A Psychosis Endophenotype? A Resting State
‘‘resting-state’’ data. AJNR Am J Neuroradiol 22: 1326–1333.
Functional Magnetic Resonance Imaging Study. Biol Psychiatry.
52. Strakowski SM, Adler CM, Holland SK, Mills NP, DelBello MP, et al. (2005)
23. Meda SA, Gill A, Stevens MC, Lorenzoni RP, Glahn DC, et al. (2012)
Abnormal FMRI brain activation in euthymic bipolar disorder patients during a
Differences in resting-state functional magnetic resonance imaging functional
counting Stroop interference task. Am J Psychiatry 162: 1697–1705.
network connectivity between schizophrenia and psychotic bipolar probands
and their unaffected first-degree relatives. Biol Psychiatry 71: 881–889. 53. Mayberg HS, Liotti M, Brannan SK, McGinnis S, Mahurin RK, et al. (1999)
24. Damoiseaux JS, Beckmann CF, Arigita EJ, Barkhof F, Scheltens P, et al. (2008) Reciprocal limbic-cortical function and negative mood: converging PET findings
Reduced resting-state brain activity in the ‘‘default network’’ in normal aging. in depression and normal sadness. Am J Psychiatry 156: 675–682.
Cereb Cortex 18: 1856–1864. 54. Perlman SB, Almeida JR, Kronhaus DM, Versace A, Labarbara EJ, et al. (2012)
25. Corbetta M, Shulman GL (2002) Control of goal-directed and stimulus-driven Amygdala activity and prefrontal cortex-amygdala effective connectivity to
attention in the brain. Nat Rev Neurosci 3: 201–215. emerging emotional faces distinguish remitted and depressed mood states in
26. Vincent JL, Kahn I, Snyder AZ, Raichle ME, Buckner RL (2008) Evidence for a bipolar disorder. Bipolar Disord 14: 162–174.
frontoparietal control system revealed by intrinsic functional connectivity. 55. Vizueta N, Rudie JD, Townsend JD, Torrisi S, Moody TD, et al. (2012)
J Neurophysiol 100: 3328–3342. Regional fMRI hypoactivation and altered functional connectivity during
27. Menon V (2011) Large-scale brain networks and psychopathology: a unifying emotion processing in nonmedicated depressed patients with bipolar II disorder.
triple network model. Trends Cogn Sci 15: 483–506. Am J Psychiatry 169: 831–840.
28. Seeley WW, Menon V, Schatzberg AF, Keller J, Glover GH, et al. (2007) 56. Uhlhaas PJ, Singer W (2012) Neuronal dynamics and neuropsychiatric
Dissociable intrinsic connectivity networks for salience processing and executive disorders: toward a translational paradigm for dysfunctional large-scale
control. J Neurosci 27: 2349–2356. networks. Neuron 75: 963–980.
29. Greicius MD (2003) Neuroimaging in developmental disorders. Curr Opin 57. Uhlhaas PJ (2013) Dysconnectivity, large-scale networks and neuronal dynamics
Neurol 16: 143–146. in schizophrenia. Curr Opin Neurobiol 23: 283–290.
30. Raichle ME, MacLeod AM, Snyder AZ, Powers WJ, Gusnard DA, et al. (2001) 58. Cole DM, Smith SM, Beckmann CF (2010) Advances and pitfalls in the analysis
A default mode of brain function. Proc Natl Acad Sci U S A 98: 676–682. and interpretation of resting-state FMRI data. Front Syst Neurosci 4: 8.
PLOS ONE | www.plosone.org 9 October 2014 | Volume 9 | Issue 10 | e107829