This is an ongoing assignment and will be worked on each week. This is what is required this week. I also picked to focus on Borderline Personality for my topic.Select the topic for your Critical Revi

Bipolar and borderline patients display differential patterns of functional connectivity among resting state networks Pritha Das a,b,c , Vince Calhoun d,e , Gin S. Malhi a,b,c, ⁎ aDepartment of Psychiatry, Sydney Medical School-Northern Campus, University of Sydney, AustraliabAdvanced Research Clinical High-ﬁeld Imaging (ARCHI) Facility, Sydney Medical School-Northern Campus, University of Sydney, AustraliacCADE Clinic, Department of Psychiatry, Royal North Shore Hospital, St Leonards, Sydney, AustraliadThe Mind Research Network, Albuquerque, NM, USAeDept. of ECE, University of New Mexico, Albuquerque, NM, USA abstract article info Article history:

Accepted 23 April 2014 Available online 2 May 2014 Keywords:

Bipolar Borderline fMRI Resting state ICA Functional network connectivity Bipolar disorder (BD) and borderline personality (BPD) disorder share clinical features such as emotional lability and poor interpersonal functioning but the course of illness and treatment differs in these groups, which suggests that the underlying neurobiology of BD and BPD is likely to be different. Understanding the neural mechanisms behind the pathophysiology of BD and BPD will facilitate accurate diagnosis and inform the administration of targeted treatment.

Since deﬁcits in social cognition or emotion regulation or in the self-referential processing system can give rise to these clinical features, and impairment in these domains have been observed in both patient groups, functional connectivity within and between networks subserving these processes during resting was investigated using functional magnetic resonance imaging. Data were acquired from 16 patients with BD, 14 patients with BPD, and 13 healthy controls (HC) and functional connectivity strength was correlated with scores using the Difﬁcul- ties in Emotion Regulation Scale.

Functional network connectivity (FNC) patterns differentiated BD and BPD patients from HC. In BD, FNC was in- creased while in BPD it was decreased. In BD impaired FNC was evident primarily among networks involved in self-referential processing while in BPD it also involved the emotion regulatory network. Impaired FNC displayed an association with impulsivity in BPD and emotional clarity and emotional awareness in BD.

This study shows that BD and BPD can perhaps be differentiated using resting state FNC approach and that the neural mechanisms underpinning overlapping symptoms discernibly differ between the groups. Theseﬁndings provide a potential platform for elucidating the targeted effects of psychological interventions in both disorders.

© 2014 Elsevier Inc. All rights reserved. Introduction Bipolar disorder (BD) and borderline personality disorder (BPD) are serious mental illnesses, which present clinically with overlapping psy- chopathology. Typically this includes features of emotional lability and impulsivity, which are particularly pronounced in the context of inter- personal interactions. This makes cross-sectional distinction of the two disorders difﬁcult and often results in BPD being misdiagnosed as part of the bipolar spectrum (Coulston et al., 2012; Kuiper et al., 2013). In contrast, the longitudinal course of the two disorders is quite separate and usually the two disorders respond differentially to biological and psychological treatments. This suggests that BD, an affective illness, and BPD, a developmental construct of personality, though seemingly similar when viewed cross-sectionally, may have differentialneurobiological substrates (Coulston et al., 2012; Koenigsberg, 2010).

It is possible that overlapping symptoms in these disorders vary in their multi-dimensional construct which are governed by different neu- ral systems which will be reﬂected in interactions among different brain networks. This may partly explain why the two disorders respond to dif- ferent treatments. This conceptual approach is in keeping with research that has revealed unique neural underpinnings in disorders, which often manifest overlapping symptoms and on occasion cannot be differ- entiated clinically such as schizophrenia and bipolar disorders (Khadka et al., 2013; Meda et al., 2012a).

Emotional lability and poor interpersonal functioning often stem from poor emotion regulation, which in turn is a multi-dimensional construct. Emotion dysregulation can be a consequence of impairments in the social cognitive domain, or in the appraisal of emotional stimuli and the regulation of emotional responses (Phillips et al., 2008), or in the self-referential evaluation of emotional stimuli (Herbert et al., 2011). Therefore, an investigation of interactions among the networks involved in these processes in BD and BPD simultaneously may help in NeuroImage 98 (2014) 73–81 ⁎Corresponding author at: Department of Psychiatry, Level 3, Main Building, Royal North Shore Hospital, St Leonards, NSW 2065, Sydney, Australia. Fax: + 61 2 9926 4062.

E-mail address:[email protected](G.S. Malhi).

http://dx.doi.org/10.1016/j.neuroimage.2014.04.062 1053-8119/© 2014 Elsevier Inc. All rights reserved. Contents lists available atScienceDirect NeuroImage journal homepage: www.elsevier.com/locate/ynimg identifying the neural substrates of emotion dysregulation in these dis- orders. Studies mapping these functions onto the brain using neuroim- aging have shown that the prefrontal cortex (PFC) plays a pivotal role in these processes with some regional (both lateral-medial and ventro- dorsal) speciﬁcity. Social cognition involves the lateral or dorso- medial PFC, depending on the engagement of implicit/explicit mentalizing respectively. Emotional processing depends on the inter- play between the ventral and dorsal medial PFC involved in generating and regulating emotional responses (Etkin et al., 2011). Self-referential processing also engages medial PFC, and in particular the midline corti- cal structures of the“default mode”(DM) (Raichle et al., 2001)network encompassing the medial PFC and medial parietal brain regions. Studies using theory of mind have shown that the ability to mentalize is com- promised in both disorders (Malhi et al., 2008; Sharp et al., 2011) and ﬁndings from neuroimaging studies investigating cognitive and emo- tional functioning have shown both common and differential patterns of lateral and medial PFC engagement in BD and BPD (Coulston et al., 2012). Moreover, dysfunctions in the DM network have been reported for both BD and BPD (Calhoun et al., 2011; Liu et al., 2012b; Ongur et al., 2010; Whitﬁeld-Gabrieli and Ford, 2012; Wolf et al., 2011), but di- rect comparisons are limited because no study to date has concurrently employed the same experimental paradigms in both patient groups (Sripada and Silk, 2007). However, a recent publication from our group compared BD and BPD to a group of healthy subjects while performing an emotional Stroop task and found that impaired neural activity encompassing the lateral PFC was common to both groups but increased activity in the medial PFC was evident only in BD (Malhi et al., 2013). Theseﬁndings prompted our hypothesis that the neurobi- ological mechanisms of emotion dysregulation in these disorders have both shared and distinct components.

A recent and novel approach to understanding the neural processes within the brain is to take advantage of its innate organization into par- allel interacting networks, particularly at rest. Functional magnetic res- onance imaging (fMRI) data acquired at rest avoids explicit task-related confounds and permits the investigation of spontaneous neuronal activ- ity and functioning of temporally coherent networks. These neural sys- tems represent functional modules of the brain that may in the future hold valuable diagnostic and prognostic information (Fox and Raichle, 2007; Whitﬁeld-Gabrieli and Ford, 2012).

Two methodologies that are widely used for evaluating functional connectivity in fMRI are: (i)seed-based correlation(Biswal et al., 1995), and (ii)independent component analysis(ICA) (Calhoun et al., 2001). The former method reports connectivity between any two syn- chronous voxels/regions independent of their allegiance to intrinsically connected networks but requires prior assumptions regarding brain function. This limits the number of brain regions that can be concurrent- ly investigated by this method. In contrast ICA, which is increasingly being used in neuroimaging research, does not require any such a priori assumptions and is particularly useful for evaluating the hidden spatio- temporal structure representing the human functional connectome (Calhoun and Adali, 2012; Van Dijk et al., 2010) Thus far only a handful of studies have investigated resting state functional connectivity in BD (Anand et al., 2009; Anticevic et al., 2013; Chai et al., 2011; Chepenik et al., 2010; Khadka et al., 2013; Liu et al., 2012a, 2012b; Mamah et al., 2013; Meda et al., 2012b; Ongur et al., 2010; Torrisi et al., 2013). The majority of these studies have used seed-based approaches to examine connections between the PFC and limbic regions (Anand et al., 2009; Anticevic et al., 2013; Chai et al., 2011; Chepenik et al., 2010; Mamah et al., 2013; Torrisi et al., 2013 ) because aberrant frontal-limbic functioning has been repeatedly implicated in the pathophysiology of BD, and is thought to drive the resulting emotion dysregulation (Townsend and Altshuler, 2012). How- ever, a handful of studies have used ICA to investigate the spatial and temporal aspects of functional connectivity within (Khadka et al., 2013; Ongur et al., 2010) and between networks (Meda et al., 2012b) in BD, and besides functional connectivity, other measures such asmeasures of regional homogeneity (Liu et al., 2012b) and amplitude of low frequencyﬂuctuations (Liu et al., 2012a) have also been exam- ined. However, the populations examined have been extremely heterogeneous with respect to phenotype, mood state, medication sta- tus, and treatment outcome, and the variety of imaging analysis tech- niques employed makes it difﬁcult to synthesize theﬁndings meaningfully.

In BPD patients there have been even fewer neuroimaging studies and only one group has examined resting state fMRI (Wolf et al., 2011). Using ICA, Wolf et al. investigated functional connectivity among four key frontal networks in BPD, and found a signiﬁcant corre- lation between functional connectivity strength in the DM network and BPD symptoms.

Interestingly, despite this emerging interest and the use of a variety of techniques to understand BD and BPD, no study to date has speciﬁcal- ly investigated whether BD and BPD can be differentiated on the basis of functional communicationbetweennetworks involved in social cogni- tion, emotion regulation and self-referential processing at rest and whether impairment in the functional connectivity underlie emotion dysregulation in these patients. Very recently investigation of functional connectivity between networks has shown to partition the neural underpinnings in disorders showing overlapping symptoms (Meda et al., 2012a). Therefore, in the present study, we investigated functional connectivity within and between networks subserving social cognition or emotion regulation or self-referential processing using ICA and functional network connectivity (FNC) approach (Jafri et al., 2008) and correlated impaired FNC with clinical measures of emotion dysregulation.

We hypothesized that BD and BPD willdisplay a differential pattern of functional connectivity that will distinguish the two disorders. Further- more we hypothesized that impaired functional connectivity within each group will correlate with key clinical measures of emotion dysregulation.

Material and methods Participants Forty-three female adults participated in the study. Among them, 16 were euthymic BD patients, 14 were BPD patients and 13 were healthy controls (HC). Participants were recruited through the CADE Clinic (www.cadeclinic.com) at Royal North Shore Hospital and via advertise- ments. All participants underwent a structured clinical interview by a psy- chiatrist (GSM) to determine whether they met the DSM-IV criteria for BD or BPD and were currently euthymic. Healthy controls were also assessed to ensure the absence of unusual symptoms and any psychiatric history.

Exclusion criteria included non-euthymic state, current hospitaliza- tion, substance abuse, a history of traumatic head injury, neurological illness (e.g., seizures or stroke), learning or developmental disorders, or poor English proﬁciency. Subjects were excluded from analysis if their functional scans showed extreme motion (translationN3 mm, roughly two voxels and rotationN1°). All participants were instructed to refrain from alcohol or recreational drug use 24 h prior to the brain scan and from caffeine or nicotine within 1 h of the scan.

At the time of scanning three patients from each group were not tak- ing any medication. Both groups of patients were taking antidepressants (venlafaxine: BD (2), BPD (2); citalopram: BD (1), BPD (2); sertraline hydrochloride: BD (1), BPD (2);ﬂ uoxetine hydrochloride: BD (2), BPD (1); tricyclics: BPD (1); reboxetine: BD (1)) and an equal number of pa- tients from each group (6) were taking antipsychotic medications. With the exception of one BPD patient these were predominantly atypical an- tipsychotics. Predictably, mood stabilizers were being used more by BD (8 BD and 4 BPD) and the mood stabilizers being used included lithium carbonate (BD (5), BPD (3)), sodium valproate (BPD (1)), lamotrigine (BD (2)) and carbamazepine (BD (1)). 74P. Das et al. / NeuroImage 98 (2014) 73–81 Questionnaires A series of questionnaires were administered, including The Child- hood Trauma Questionnaire (CTQ), Barratt Impulsiveness Scale (BIS), Beck Depression Inventory (BDI), The Depression Anxiety Stress Scales (DASS) and The Difﬁculties in Emotion Regulation Scale (DERS). The DERS has 6 subscales, Non-Acceptance of Emotional Responses, Difﬁcul- ties Engaging in Goal-Directed Behavior, Impulse Control Difﬁculties, Lack of Emotional Awareness, Limited Access to Emotion Regulation Strategies, and Lack of Emotional Clarity.

MRI data acquisition Imaging datasets (structural and functional) were acquired on a Uni- versity of Sydney 3 T Siemens Magnetom Trio Scanner based at the Ad- vanced Research and Clinical High-ﬁeld Imaging (ARCHI) facility. A T2*- weighted gradient echo echo-planner imaging (EPI) sequence (29 axial slices, slice thickness 4 mm with 1 mm gap, TR = 2000 ms, TE = 35 ms, ﬂip = 70 o, 64 × 64 matrix) was used to acquire 155 whole-brain vol- umes of functional data. A high-resolution T1-weighted structural image was also acquired for precise localization of brain activity using a magnetization prepared rapid gradient echo (MPRAGE) sequence (TR = 1570 ms, TE = 3.22 ms, Flip = 15 o,matrix256×256,192slices).

Subjects were instructed to keep their eyes open during the functional scan and stare passively at a foveally presentedﬁxation cross, as this is suggested to facilitate network delineation compared to eyes-closed conditions (Van Dijk et al., 2010). Head motion during scanning was re- strained using foam pads inserted on each side. All participants were judged as awake and alert at the start and conclusion.

Clinical data analysis Analyses of Variance (ANOVA) and SPSS Statistics Version 19 were used to determine differences between groups in age, years of educa- tion, premorbid IQ, childhood trauma and also in the self-report mea- sures of mood, impulsivity, and emotional dysregulation. Where signiﬁcant results from the ANOVA were found, post-hoc analyses were conducted and the Scheffé procedure was applied for multiple comparisons. An initial alpha level of 0.05 was used for all statistical tests.

fMRI data analysis Pre-processing was done using SPM5 software (http://www.ﬁl.ion.

ucl.ac.uk/spm/software/spm5). Each subject's functional and structural images wereﬁrst inspected visually for scanner artifacts and gross ana- tomical abnormalities, and then re-oriented so that the origin of the image lay within 3 cm of the anterior commissure. The initial 5 images were discarded to remove T1 equilibration effects. Images were then realigned using INRIAlign—a motion correction algorithm unbiased by local signal changes and corrected for slice time variation using the middle slice as the reference frame. Then these images were spatially normalized to a common stereotactic space using the Montreal Neuro- logical Institute (MNI) EPI template and spatially smoothed with a Gaussian kernel of 8 mm 3full width at half maximum.

Group spatial ICA was used to identify networks from the pre- processed fMRI data of all participants using the GIFT software (version 2.0e;http://mialab.mrn.org/software/gift/). A complete description of the methods implemented in GIFT has been published (Calhoun et al., 2001; Erhardt et al., 2011) but brieﬂy,ﬁrst“minimum description length”(MDL) criterion was used to estimate the number of indepen- dent“sources”/components present in each participant. MDL predicted 20 as a median number of components. Then, data from each participant was reduced from 150 time points to 30 time points using standard principal component analysis (PCA). Before reduction data was pre- processed to remove mean per time point. Finally, data from all subjectswere concatenated and this aggregate data set was then further reduced to 20 independent components (ICs) using subject speciﬁc PCA, follow- ed by an independent component estimation using an algorithm which attempts to minimize mutual information (Infomax) (Bell and Sejnowski, 1995). This algorithm was repeated 20 times in ICASSO (http://research.ics.tkk.ﬁ/ica/icasso/) and the resulting components were clustered to estimate the reliability of the decomposition. Follow- ing the group decomposition, single subject time courses (TCs) and spa- tial maps (SMs) were then back-reconstructed using GICA and calibrated using z scores. Components were visually inspected for arti- facts. Two masks representing prefrontal and midline cortical structures were created using ALL atlas within WFU_pickatlas tool (http://www.

rad.wfubmc.edu/fmri). The prefrontal mask was created following Wolf et al. (2011)which included superior, middle, and inferior frontal regions. The midline mask was created following meta-analysis of neu- roimaging studies focused on self-referential processing (Northoff and Bermpohl, 2004) which included dorsal and ventral areas of the medial prefrontal and anterior cingulate cortices, as well as the posterior cingu- late cortex and precuneus. Networks for analysis were chosen on the basis of three conditions, 1) they had to exhibit activations in gray mat- ter, low spatial overlap with known vascular, ventricular motion, and susceptibility artifacts, 2) to be dominated by low frequencyﬂuctua- tions i.e. the ratio of integral of spectral power below 0.10 Hz to power between 0.15 Hz and 0.25 Hz has to be greater than 5 (Allen et al., 2011), and 3) they should represent either prefrontal or cortical midline structures. Finally six components of interest were chosen for further analysis among which three components displayed highest spa- tial correlation with the midline structures and three components with the prefrontal mask.

To visualize the spatial maps of a component, all subjects' maps for that particular component were entered into a random-effect analysis model (1 samplet-test in SPM5). Brain regions were considered to be within each network if they met a height threshold of pb0.00001 corrected for multiple comparisons using the family-wise error (FWE) and an extent threshold of 50 voxels.

To investigate whether chosen networks were different between groups, functional connectivity within and between networks was in- vestigated. Within network connectivity was assessed using the network's SM and low frequency power ratio [PR, the ratio of spectral power between 0.01 and 0.10 Hz to the integral of power between 0.15 and 0.25 Hz (Allen et al., 2011)] and between network connectivity was determined using FNC.

First component's SM was thresholded based on the distribution of voxelwise t-statistics so that it represents voxels that have shown strong and consistent activation across subjects and include regions most associated with the component's TC. Second, FNC was calculated following the procedure described byJafri et al. (2008). The time course data associated with the selected components were detrended, despiked, andﬁ ltered using aﬁfth-order Butterworth low-passﬁlter with a high frequency cut-off of 0.15 Hz (Allen et al., 2011). For each subject, correlations between pairwise combinations were calculated using Pearson's correlations and these values were then transformed to z scores using Fisher's transformation to use as a measure of FNC.

Thenﬁnally, the MANCOVAN utility (Allen et al., 2011)withinGIFT was used to determine differences between groups in SM and FNC.

MANCOVAN uses a multivariate model selection strategy to reduce the number of statistical tests performed. Itﬁrst uses multivariate anal- ysis of covariance to identify factors that inﬂuence the response matrix and then perform univariate tests with a reduced design matrix and cor- rect for multiple comparisons using the false discovery rate (FDR). The design matrix included group membership (BPD, BD, HC) as covariates of interest, as well as clinical scores of depression, anxiety, stress, as measured using DASS, and motion (average scan-to-scan rotation and translation from INRIAlign motion estimates) as nuisance predictors.

Signiﬁcant differences (pb0.05) between groups in PR were deter- mined using a Multivariate General Linear Model approach in SPSS 75 P. Das et al. / NeuroImage 98 (2014) 73–81 where clinical scores of depression, anxiety, and stress, as measured using DASS, were controlled for.

Clinical scores of emotion dysregulation and their relation to FNC strengths in BD and BPD In order to understand how impaired FNC relates to emotion dysreg- ulationinbothBDandBPD,thesubscalescoresofDERSandthestrength of FNC which displayed impairment were correlated and this was done separately for each patient group. Clinical scores of depression, anxiety, and stress were controlled for within each correlation analysis.

Results There were no signiﬁcant differences between groups with respect to age, years of education, premorbid IQ and childhood trauma but BDI, BIS, and total and subscale scores of DASS and DERS differed across groups which have been summarized inTable 1.

Spatial maps of six selected components/networks are shown in Fig. 1. These networks were highly stable (reliability indexN0.86) as determined by ICASSO. Primary regions within each network are pro- vided inTable 2.

Among three networks which displayed strong spatial correlation with the frontal mask, two of them represented left and right fronto- parietal (LFP and RFP) networks encompassing frontal and lateral pari- etal regions associated with affect regulation. These networks have pre- viously been identiﬁed from resting state data (Allen et al., 2011; Wolf et al., 2011). The third network encompassed regions such as the dorsal anterior cingulate (dACC) and orbital fronto-insular cortices which are part of the broader“salience network”(Seeley et al., 2007). This net- work also contained regions such as the inferior frontal gyrus (IFG), pos- terior superior temporal sulcus (STS), inferior parietal lobule (IPL), and the temporoparietal junction (TPJ) associated with mentalizing. The IPL, STS and the TPJ regions have also been associated with externally fo- cused social processing (Sui et al., 2013). This network may therefore play a role in detecting social salience and therefore in this study it has been referred as the social salience (SS) network. One of the net- works among three that displayed strong spatial correlation with the midline mask was the DM network. This network has previously been identiﬁed from resting state data (Jafri et al., 2008; Wolf et al., 2011).

The other two were the ventral medial PFC (vmPFC) and the precuneus (Precuneus) networks. These networks are similar to the respective net- works IC 25 and IC 53, which have been identiﬁed previously in resting state data from a large cohort (Allen et al., 2011). Notably, theydescribed the IC53 which is equivalent to our Precuneus network as the“core”of the posterior DM network.

Power ratio and spatial maps did not differ between groups, indicat- ing no differences between groups with respect to functionality of each network. However, the functional connectivity pattern between net- works revealed signiﬁcant differences between BD and BPD. From 15 pairwise network combinations four sets of coupling were signiﬁcantly different between BD and BPD and in all of them BD had increased func- tional connectivity compared to BPD (Figs. 2 and 3). These four sets were: 1) SS-RFP, 2) SS-Precuneus, 3) SS-vmPFC, and 4) DM-Precuneus.

Although coupling between DM-Precuneus and SS-vmPFC were sig- niﬁcantly different between BD and HC, and BD had signiﬁcantly in- creased functional connectivity compared to HC, these differences did not survive multiple comparisons.

Again, differences in coupling between SS-Precuneus and SS-RFP were observed between BPD and HC where BPD had reduced functional connectivity compared to HC but these differences also did not survive multiple comparisons.

Relation between FNC strengths and emotion dysregulation In BPD, functional connectivity strength between SS-RFP, which was signiﬁcantly reduced compared to both HC and BP, was signiﬁcantly negatively correlated with the scores of impulse control difﬁculties (r = 0.639, p = 0.034) (Fig. 2) suggesting that in patients with BPD when functional connectivity between these network de- creases impulse control difﬁculties increase.

In BD, functional connectivity strength between SS-vmPFC, which was signiﬁcantly increased compared to both HC and BPD, was signiﬁ- cantly positively correlated with lack of emotional clarity (r = 0.605, p = 0.029) (Fig. 2), suggesting that in patients with BD when functional connectivity between these networks increases emotional clarity de- creases. Interestingly, the connectivity strength between DM- Precuneus, which was also signiﬁcantly increased in BD compared to HC and BPD, was signiﬁcantly negatively correlated with lack of emo- tional awareness scores (r = 0.574, p = 0.040) (Fig. 2), indicating that in patients with BD an increased functional connection between these networks enhances emotional awareness.

In HC no signiﬁcant correlation was observed between these func- tional connectivity strengths and subscale scores of DERS.

Discussion Supporting our hypothesis BD andBPD patients displayed a dif- ferential pattern of functional network connectivity (FNC) among Table 1 Means (and standard deviations) of demographic and clinical characteristics of subjects who participated in the study.

The Following abbreviations have been used: HC = healthy controls; BD = patients with bipolar disorder; BPD = patients with borderline personality disorder; BDI = Beck Depression Inventory; BIS = Barratt Impulsiveness Scale; DASS = Depression, Anxiety Stress Scales; DERS = Difﬁculties in Emotion Regulation Scale.

BD BPD HCF Age 35.63 (10.71) a 32.00 (7.86) a 31.15 (11.07) a F(2,40) = .84 Education 15.87 (1.51) a 14.64 (2.41) a 15.38 (1.85) a F(2,39) = 1.44 BDI 12.13 (11.68) a 27.93 (11.38) b 5.54 (4.71) a F(2,40) = 18.19⁎⁎ BIS 70.06 (11.12) a 76.36 (8.42) a 59.77 (10.94) b F(2,40) = 8.94⁎ DASS depression 4.38 (4.77) a 9.71 (6.73) b 2.08 (2.72) a F(2,42) = 8.26⁎ DASS anxiety 3.13 (3.24) a 6.43 (4.54) b 0.92 (1.44) a F(2,42) = 9.27⁎⁎ DASS stress 7.50 (5.11) a,b 10.36 (4.53) a 3.77 (2.71) b F(2,42) = 7.85⁎ DASS total score 15.00 (9.71) a 26.50 (12.38) b 6.77 (5.10) a F(2,40) = 26.63⁎⁎ DERS non-acceptance of emotional responses 14.00 (6.47) a 20.93 (6.04) b 11.07 (2.50) a F(2,42) = 11.92⁎⁎ DERS difﬁculty engaging in goal-directed behavior 17.44 (4.16) a,b 20.29 (3.29) a 13.38 (5.56) b F(2,42) = 8.38⁎⁎ DERS impulse control difﬁculties 14.88 (6.44) a 22.14 (4.56) b 9.54 (3.45) c F(2,42) = 20.94⁎⁎ DERS lack of emotional awareness 14.00 (5.57) a 19.86 (5.26) b 13.46 (5.70) a F(2,42) = 5.82⁎ DERS limited access to emotion regulation strategies 18.31 (5.45) a 29.71 (6.03) b 12.15 (3.33) c F(2,42) = 41.18⁎⁎ DERS lack of emotional clarity 11.00 (3.83) a 17.07 (3.79) b 9.23 (2.28) a F(2,42) = 19.92⁎⁎ DERS total score 89.63 (19.59) a 130.00 (18.60) b 68.85 (17.49) c F(2,40) = 37.94⁎⁎ For each scale, means with the same superscript indicate that they do not signiﬁcantly differ from each other; Superscript a,b indicates it is not different from both a and b, *pb0.01; **pb0.001. 76P. Das et al. / NeuroImage 98 (2014) 73–81 resting state networks involved in detecting social salience, affect regulation, and self-referential processing. Though differences com- pared to HC did not survive correction for multiple comparisons it is clearly evident from ourﬁndings (Fig. 3) that connectivity is in- creased in BD, as compared to both BPD and HC, involving speciﬁcal- ly SS-vmPFC and DM-Precuneus coupling, and that it is reduced in BPD, compared to both HC and BD, involving SS-Precuneus and SS- RFP coupling. This suggests that the resting state FNC technique can be used to understand the neural underpinnings of BD and BPD and ultimately this approach may provideﬁndings that are of assis- tance in differentiating these disorders. Moreover, ourﬁnding of an association between the strength of impaired FNCs with different constructs of emotion dysregulation is in line with our suggestion that the neural underpinning of emotion dysregulation in the two disorders is likely to be different.

Abnormal self-referential thought processes, that are thought to un- derpin emotional dysregulation in mood disorders, are known to en- gage cortical midline brain structures (Marchand, 2012) and this brain region is subdivided on the basis of the functions the various compo- nents perform. For example, ventral anterior regions (including ventralmedial PFC and ACC) are responsible for self-referencing information whereas posterior midline regions (including the precuneus and poste- rior cingulate) are more concerned with placing one's‘self in context’, and drawing upon autobiographical memory. Another distinction that separates anterior and posterior components is that the former is in- volved in more inward-focused thought processing whereas the latter is engaged more so by outward-directed, social, and contextual focused thought processing. In both BD and BPD, the functional connectivity be- tween the SS network and midline structures (vmPFC and Precuneus) is altered both in terms of direction of connectivity and with respect to re- gional speciﬁcity. In BPD, for instance, it is the posterior components (Precuneus) of midline structures that are affected, whereas in BD it is the anterior (vmPFC) components that are involved. Moreover, the ﬁnding of an association between measures of emotion regulation and FNC strengths suggests that emotion dysregulation in both disorders may in part be a consequence of communication being compromised between social salience detection and self-referential processing, and that the differences between the disorders may hinge on the focus of thought processes, which culminates in incorrect and/or disproportion- ate salience being assigned to social stimuli. Fig. 1.Six chosen components/networks are shown.77 P. Das et al. / NeuroImage 98 (2014) 73–81 Impulsivity is a key feature of BPD psychopathology (Links et al., 1999), which among other behaviors often manifests as self-harm. Gen- erally, impulsive behavior reﬂects a deﬁcit in the ability to inhibit prepo- tent responses and to reappraise emotional responses. Impulsivity in BPD may be due to impaired functional connectivity between the SS and RFP networks, which subsume regions that have been implicated in both response inhibition and reappraisal of emotional responses (Aron et al., 2004; Buhle et al., 2013). Interestingly, supporting this pos- sibility, the functional connectivity strength between the SS and RFP networks displayed a negative correlation with a clinical measure of im- pulsivity (measured using DERS) in BPD.

An additional role of RFP is to inform decision making by integrating information from the external environment and stored internal repre- sentations (Vincent et al., 2008). Notably, the functional connectivity strength between the SS-RFP and SS-Precuneus in BPD were signiﬁcant- ly positively correlated, suggesting further that the resultant impulsivity may be related to a failure of integration of information from the social salience network with any stored internal representation from the Precuneus network. Failure to integrate information from the environ- ment and internal representation may explain the high degree of reac- tivity to environmental stimuli that BPD patients often manifest and why self-validation plays a key role in maladaptive behaviors. This may also explain why components of dialectical behavior therapy (Lynch et al., 2007) are effective in this patient group.

Similarly in BD, increased functional connectivity, compared to BPD and HC, between the SS and vmPFC networks also points to impairment in the system pivotal to social understanding. This is because the vmPFCis principally associated with internally focused self-processing, which involves self-reﬂection and the retrieval of knowledge pertaining to the self. In contrast, the IPL, STS and the TPJ regions of the SS network are primarily thought to be involved in externally focused social pro- cessing (Sui et al., 2013). Interestingly, it has been found that assign- ment of personal signiﬁcance can modulate social evaluation by increasing activity in the vmPFC and the left posterior STS and function- al coupling strength between these regions correlates with behavioral self-bias (Sui et al., 2013). Therefore impairment in communication be- tween these networks may explain the subtle deﬁcits in social function- ing often observed in BD even when euthymic (Malhi et al., 2008).

In psychological terms, the ability to understand one's own emotions and correctly label them is termed emotional clarity. It speciﬁcally in- volves the ability to discriminate between different emotions and have‘knowledge’of this. In BD emotional clarity displayed negative as- sociation with the FNC strength between the vmPFC and SS networks.

This inverse relation has also been observed in depressive rumination where a decrease in emotion clarity is associated with an increase in ac- tivity in the vmPFC (Cooney et al., 2010).

Compared to BPD and HC, BD patients also displayed increased func- tional connectivity between the DM and Precuneus networks. Con- scious self-representation, which is fundamental to meaningfulness in life, is thought to be dependent upon the successful retrieval of episodic memory. Functional imaging studies have shown differential activity in the medial prefrontal and medial parietal cortices during episodic mem- ory retrieval, and the medial parietal cortices, in conjunction with medi- al prefrontal and lateral parietal cortices are also implicated in episodic Table 2 Primary brain regions within each network identiﬁed using 1 samplet-test in SPM 5 are shown here.

Independent component (network) No of voxels t max Coordinate Left-fronto-parietal (LFP) Inferior/middle frontal gyrus 1794 21.75 42, 45, 0 Inferior/superior parietal lobule (left) 1534 20.11 48, 42, 51 Inferior/superior parietal lobule (right) 320 16.11 33, 60, 45 Inferior/middle temporal gyrus 242 15.32 54, 60, 12 Cerebellum391 15.57 12, 81, 30 Right fronto-parietal (RFP) Inferior/middle/superior frontal gyrus 1953 19.10 48, 36, 27 Medial frontal gyrus 99 12.01 6, 30, 42 Inferior parietal lobule (right) 1360 24.90 51, 48, 48 Inferior parietal lobule (left) 338 15.31 39, 54, 48 Cingulate gyrus190 14.57 6, 36, 36 Precuneus73 13.60 6, 69, 51 Cerebellum410 19.25 12, 78, 30 Default mode (DM) Ventro-medial Prefrontal/anterior cingulate cortex 1719 25.90 3, 51, 3 Precuneus/posterior cingulate cortex 2966 40.37 3, 54, 27 Parahippocampal gyrus 116 13.75 27, 36, 21 Inferior parietal lobule (right) 463 28.36 42, 66, 45 Inferior parietal lobule (left) 719 25.48 39, 72, 45 Inferior temporal gyrus (left) 141 17.14 60, 15, 21 Inferior temporal gyrus (right) 101 12.61 60, 12, 21 Cerebellum72 13.21 42, 72, 42 Precuneus Precuneus2294 27.70 9, 69, 51 Supramarginal gyrus 58 12.54 57, 42, 30 Cingulate gyrus63 11.11 6, 24, 33 Superior Frontal Gyrus 70 9.75 30, 45, 18 Social salience (SS) Orbital fronto-insular cortex (right) 577 21.41 45, 27, 9 Orbital fronto-insular cortex (left) 637 19.69 42, 21, 9 Dorsal anterior cingulate/medial prefrontal cortex 1455 20.71 9, 45, 36 Superior temporal gyrus/sulcus and inferior parietal lobule 931 21.29 51, 57, 33 Superior/middle temporal gyri and inferior parietal lobule 917 20.96 63, 30, 12 Cingulate gyrus73 13.26 0, 15, 39 Cerebellum121 13.22 15, 81, 27 Ventro-medial PFC (vmPFC) Subgenual anterior cingulate/orbitofrontal cortex 1171 17.22 3, 39, 3 78P. Das et al. / NeuroImage 98 (2014) 73–81 memory retrieval in the context of self-representation (Lou et al., 2004).

Impairment in functional connectivity between these networks i.e. the DM (encompassing MPFC and lateral parietal cortices) and Precuneus (encompassing medial parietal cortices), may therefore reﬂect a com- promise in the conscious self-monitoring system in BD. Self-monitoring is the ability to consciously observe and regulate one's own behavior, and in order to execute this successfully, it is essential to be able to recognize one's own feelings and those of others (emotion- al awareness). Noticeably, in the current study increased FNC between DM and Precuneus networks in BD patients displayed a signiﬁcant Fig. 2.Functional network connectivity (FNC) patterns differentiated bipolar (BD) and borderline (BPD) patients. BD displayed increased network connectivity, compared to both BPD and healthy controls (HC), in coupling of SS-vmPFC and DM-Precuneus networks, whereas BPD displayed decreased connectivity, compared to both HC and BD,in coupling of SS-Precuneus and SS-RFP networks. Interestingly, these impairments in FNC showed an association with different aspects of emotion dysregulation in BPD and BD. Speciﬁcally, in BPD it was associated with impulsivity whereas in BD it was associated with emotion clarity and awareness.

Fig. 3.Mean correlation coefﬁcient of the four sets of coupling that were signiﬁcantly different bipolar (BD) and borderline (BPD) patients. In general, compared to healthy controls (HC), BD displayed increased functional connectivity and BPD decreased functional connectivity. ** denotes signiﬁcant differences which survived multiple comparison correction. * denotes signiﬁcant differences which did not survive multiple comparison.79 P. Das et al. / NeuroImage 98 (2014) 73–81 positive correlation with their clinical measure of emotional awareness, suggesting that BD patients have heightened awareness and possibly in- creased sensitivity and receptivity to social inputs. However, because of diminished emotional clarity, it is likely that they are unable to process the meaning of these inputs, in particular, with respect to their internal emotional milieu. This may explain why patients oftenﬁnd it difﬁcult to express their feelings but are aware that their ideas and behaviors have a negative impact on those around them and for whom they care, and this may also help explain why psychotherapies which aim to transform internalized representations into more realistic models of one's self, such as mindfulness, are effective in bipolar disorder (Stange et al., 2011).

Before drawing anyﬁrm conclusions it is important to consider sev- eral limitations of this study. First, ourﬁndings cannot be generalized to male patients although the investigation of females in this study does have the advantage of minimizing the potential confound of gender dif- ferences in resting state FNC (Filippi et al., 2013). Second, there are other networks that have not been considered in this study, yet may play an equally important role in the pathophysiology of BD and BPD.

Third, both patient groups were medicated and differences in medica- tions across the groups could have inﬂuenced theﬁndings. Finally, the relatively small sample sizes mean that theseﬁndings, though informa- tive, remain preliminary and require replication.

Conclusions Findings of this study indicate that BD and BPD can be differentiated on the basis of resting state functional connectivity among networks in- volved in the determination of social salience, self-referential processing, and the regulation of emotion. The symptoms of BD patients perhaps re- ﬂect an impaired interaction of the social salience detection system with the self-referential processing, whereas in BPD disorder, a similar clinical picture arises primarily due to impaired interaction with the emotion reg- ulatory system. This study is theﬁrst to dissect the two disorders in this manner and having done so we believe that it provides a platform for conducting further research that partitions the disorders and elucidate targeted clinical interventions in both disorders.

References Allen, E.A., Erhardt, E.B., Damaraju, E., Gruner, W., Segall, J.M., Silva, R.F., Havlicek, M., Rachakonda, S., Fries, J., Kalyanam, R., Michael, A.M., Caprihan, A., Turner, J.A., Eichele, T., Adelsheim, S., Bryan, A.D., Bustillo, J., Clark, V.P., Feldstein Ewing, S.W., Filbey, F., Ford, C.C., Hutchison, K., Jung, R.E., Kiehl, K.A., Kodituwakku, P., Komesu, Y.M., Mayer, A.R., Pearlson, G.D., Phillips, J.P., Sadek, J.R., Stevens, M., Teuscher, U., Thoma, R.J., Calhoun, V.D., 2011.A baseline for the multivariate comparison of resting-state networks. Front. Syst. Neurosci. 5, 00002.

Anand, A., Li, Y., Wang, Y., Lowe, M.J., Dzemidzic, M., 2009.Resting state corticolimbic con- nectivity abnormalities in unmedicated bipolar disorder and unipolar depression.

Psychiatry Res. 171, 189–198.

Anticevic, A., Brumbaugh, M.S., Winkler, A.M., Lombardo, L.E., Barrett, J., Corlett, P.R., Kober, H., Gruber, J., Repovs, G., Cole, M.W., Krystal, J.H., Pearlson, G.D., Glahn, D.C., 2013.Global prefrontal and fronto-amygdala dysconnectivity in bipolar I disorder with psychosis history. Biol. Psychiatry 73, 565–573.

Aron, A.R., Robbins, T.W., Poldrack, R.A., 2004.Inhibition and the right inferior frontal cor- tex. Trends Cogn. Sci. 8, 170–177.

Bell, A.J., Sejnowski, T.J., 1995.An information-maximization approach to blind separation and blind deconvolution. Neural Comput. 7, 1129–1159.

Biswal, B., Yetkin, F.Z., Haughton, V.M., Hyde, J.S., 1995.Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magn. Reson. Med. 34, 537–541.

Buhle, J.T., Silvers, J.A., Wager, T.D., Lopez, R., Onyemekwu, C., Kober, H., Weber, J., Ochsner, K.N., 2013.Cognitive reappraisal of emotion: a meta-analysis of human neu- roimaging studies. Cereb. Cortex 13, 13.

Calhoun, V.D., Adali, T., 2012.Multisubject independent component analysis of fMRI: a decade of intrinsic networks, default mode, and neurodiagnostic discovery. IEEE Rev. Biomed. Eng. 5, 60–73.

Calhoun, V.D., Adali, T., Pearlson, G.D., Pekar, J.J., 2001.A method for making group infer- ences from functional MRI data using independent component analysis. Hum. Brain Mapp. 14, 140–151.

Calhoun, V.D., Sui, J., Kiehl, K., Turner, J., Allen, E., Pearlson, G., 2011.Exploring the psycho- sis functional connectome: aberrant intrinsic networks in schizophrenia and bipolar disorder. Front. Psychiatry 2, 00075.Chai, X.J., Whitﬁeld-Gabrieli, S., Shinn, A.K., Gabrieli, J.D., Nieto Castanon, A., McCarthy, J.M., Cohen, B.M., Ongur, D., 2011.Abnormal medial prefrontal cortex resting-state connectivity in bipolar disorder and schizophrenia. Neuropsychopharmacology 36, 2009–2017.

Chepenik, L.G., Raffo, M., Hampson, M., Lacadie, C., Wang, F., Jones, M.M., Pittman, B., Skudlarski, P., Blumberg, H.P., 2010.Functional connectivity between ventral prefron- tal cortex and amygdala at low frequency in the resting state in bipolar disorder. Psy- chiatry Res. 182, 207–210.

Cooney, R.E., Joormann, J., Eugene, F., Dennis, E.L., Gotlib, I.H., 2010.Neural correlates of rumination in depression. Cogn. Affect. Behav. Neurosci. 10, 470–478.

Coulston, C.M., Tanious, M., Mulder, R.T., Porter, R.J., Malhi, G.S., 2012.Bordering on bipo- lar: the overlap between borderline personality and bipolarity. Aust. N. Z. J. Psychiatry 46, 506–521.

Erhardt, E.B., Rachakonda, S., Bedrick, E.J., Allen, E.A., Adali, T., Calhoun, V.D., 2011.Com- parison of multi-subject ICA methods for analysis of fMRI data. Hum. Brain Mapp.

32, 2075–2095.

Etkin, A., Egner, T., Kalisch, R., 2011.Emotional processing in anterior cingulate and medi- al prefrontal cortex. Trends Cogn. Sci. 15, 85–93.

Filippi, M., Valsasina, P., Misci, P., Falini, A., Comi, G., Rocca, M.A., 2013.The organization of intrinsic brain activity differs between genders: a resting-state fMRI study in a large cohort of young healthy subjects. Hum. Brain Mapp. 34, 1330–1343.

Fox, M.D., Raichle, M.E., 2007.Spontaneousﬂ uctuations in brain activity observed with functional magnetic resonance imaging. Nat. Rev. Neurosci. 8, 700–711.

Herbert, C., Pauli, P., Herbert, B.M., 2011.Self-reference modulates the processing of emo- tional stimuli in the absence of explicit self-referential appraisal instructions. Soc.

Cogn. Affect. Neurosci. 6, 653–661.

Jafri, M.J., Pearlson, G.D., Stevens, M., Calhoun, V.D., 2008.A method for functional net- work connectivity among spatially independent resting-state components in schizo- phrenia. Neuroimage 39, 1666–1681.

Khadka, S., Meda, S.A., Stevens, M.C., Glahn, D.C., Calhoun, V.D., Sweeney, J.A., Tamminga, C.A., Keshavan, M.S., O'Neil, K., Schretlen, D., Pearlson, G.D., 2013.Is aberrant function- al connectivity a psychosis endophenotype? A resting state functional magnetic res- onance imaging study. Biol. Psychiatry 74, 458–466.

Koenigsberg, H.W., 2010.Affective instability: toward an integration of neuroscience and psychological perspectives. J. Pers. Disord. 24 (1), 60–82.

Kuiper, S., McLean, L., Malhi, G.S., 2013.To BD or not to BD: functional neuroimaging and the boundaries of bipolarity. Expert. Rev. Neurother. 13, 75–86.

Links, P.S., Heslegrave, R., van Reekum, R., 1999.Impulsivity: core aspect of borderline personality disorder. J. Pers. Disord. 13, 1–9.

Liu, C.H., Li, F., Li, S.F., Wang, Y.J., Tie, C.L., Wu, H.Y., Zhou, Z., Zhang, D., Dong, J., Yang, Z., Wang, C.Y., 2012a.Abnormal baseline brain activity in bipolar depression: a resting state functional magnetic resonance imaging study. Psychiatry Res. 203, 175–179.

Liu, C.H., Ma, X., Li, F., Wang, Y.J., Tie, C.L., Li, S.F., Chen, T.L., Fan, T.T., Zhang, Y., Dong, J., Yao, L., Wu, X., Wang, C.Y., 2012b.Regional homogeneity within the default mode network in bipolar depression: a resting-state functional magnetic resonance imag- ing study. PLoS One 7, 2.

Lou, H.C., Luber, B., Crupain, M., Keenan, J.P., Nowak, M., Kjaer, T.W., Sackeim, H.A., Lisanby, S.H., 2004.Parietal cortex and representation of the mental Self. Proc. Natl.

Acad. Sci. U. S. A. 101, 6827–6832.

Lynch, T.R., Trost, W.T., Salsman, N., Linehan, M.M., 2007.Dialectical behavior therapy for borderline personality disorder. Annu. Rev. Clin. Psychol. 3, 181–205.

Malhi, G.S., Lagopoulos, J., Das, P., Moss, K., Berk, M., Coulston, C.M., 2008.AfunctionalMRI study of Theory of Mind in euthymic bipolar disorder patients. Bipolar Disord. 10, 943–956.

Malhi, G.S., Tanious, M., Fritz, K., Coulston, C.M., Bargh, D.M., Phan, K.L., Calhoun, V., Das, P., 2013.Differential engagement of the fronto-limbic network during emotion process- ing distinguishes bipolar and borderline personality disorder. Mol. Psychiatry 26, 22.

Mamah, D., Barch, D.M., Repovs, G., 2013.Resting state functional connectivity ofﬁve neu- ral networks in bipolar disorder and schizophrenia. J. Affect. Disord. 150, 601–609.

Marchand, W.R., 2012.Self-referential thinking, suicide, and function of the cortical mid- line structures and striatum in mood disorders: possible implications for treatment studies of mindfulness-based interventions for bipolar depression. Depress Res.

Treat. 246725, 25.

Meda, S.A., Gill, A., Stevens, M.C., Lorenzoni, R.P., Glahn, D.C., Calhoun, V.D., Sweeney, J.A., Tamminga, C.A., Keshavan, M.S., Thaker, G., Pearlson, G.D., 2012a.Differences in resting-state functional magnetic resonance imaging functional network connectivity between schizophrenia and psychotic bipolar probands and their unaffectedﬁrst- degree relatives. Biol. Psychiatry 71, 881–889.

Meda, S.A., Gill, A., Stevens, M.C., Lorenzoni, R.P., Glahn, D.C., Calhoun, V.D., Sweeney, J.A., Tamminga, C.A., Keshavan, M.S., Thaker, G., Pearlson, G.D., 2012b.Differences in resting-state functional magnetic resonance imaging functional network connectivity between schizophrenia and psychotic bipolar probands and their unaffectedﬁrst- degree relatives. Biol. Psychiatry 71, 881–889.

Northoff, G., Bermpohl, F., 2004.Cortical midline structures and the self. Trends Cogn. Sci. 8, 102–107.

Ongur, D., Lundy, M., Greenhouse, I., Shinn, A.K., Menon, V., Cohen, B.M., Renshaw, P.F., 2010.Default mode network abnormalities in bipolar disorder and schizophrenia.

Psychiatry Res. 183, 59–68.

Phillips, M.L., Ladouceur, C.D., Drevets, W.C., 2008.A neural model of voluntary and auto- matic emotion regulation: implications for understanding the pathophysiology and neurodevelopment of bipolar disorder. Mol. Psychiatry 13, 833–857.

Raichle, M.E., MacLeod, A.M., Snyder, A.Z., Powers, W.J., Gusnard, D.A., Shulman, G.L., 2001.A default mode of brain function. Proc. Natl. Acad. Sci. U. S. A. 98, 676–682.

Seeley, W.W., Menon, V., Schatzberg, A.F., Keller, J., Glover, G.H., Kenna, H., Reiss, A.L., Greicius, M.D., 2007.Dissociable intrinsic connectivity networks for salience process- ing and executive control. J. Neurosci. 27, 2349–2356. 80P. Das et al. / NeuroImage 98 (2014) 73–81 Sharp, C., Pane, H., Ha, C., Venta, A., Patel, A.B., Sturek, J., Fonagy, P., 2011.Theory of mind and emotion regulation difﬁculties in adolescents with borderline traits. J. Am. Acad.

Child Adolesc. Psychiatry 50, 563–573.

Sripada, C.S., Silk, K.R., 2007.The role of functional neuroimaging in exploring the overlap be- tween borderline personality disorder and bipolar disorder. Curr. Psychiatry Rep. 9, 40–45.

Stange, J.P., Eisner, L.R., Holzel, B.K., Peckham, A.D., Dougherty, D.D., Rauch, S.L., Nierenberg, A.A., Lazar, S., Deckersbach, T., 2011.Mindfulness-based cognitive therapy for bipolar disorder: effects on cognitive functioning. J. Psychiatr. Pract. 17, 410–419.

Sui, J., Rotshtein, P., Humphreys, G.W., 2013.Coupling social attention to the self forms a network for personal signiﬁcance. Proc. Natl. Acad. Sci. U. S. A. 110, 7607–7612.

Torrisi, S., Moody, T.D., Vizueta, N., Thomason, M.E., Monti, M.M., Townsend, J.D., Bookheimer, S.Y., Altshuler, L.L., 2013.Differences in resting corticolimbic functional connectivity in bipolar I euthymia. Bipolar Disord. 15, 156–166.Townsend, J., Altshuler, L.L., 2012.Emotion processing and regulation in bipolar disorder:

a review. Bipolar Disord. 14, 326–339.

Van Dijk, K.R., Hedden, T., Venkataraman, A., Evans, K.C., Lazar, S.W., Buckner, R.L., 2010.

Intrinsic functional connectivity as a tool for human connectomics: theory, proper- ties, and optimization. J. Neurophysiol. 103, 297–321.

Vincent, J.L., Kahn, I., Snyder, A.Z., Raichle, M.E., Buckner, R.L., 2008.Evidence for a frontoparietal control system revealed by intrinsic functional connectivity. J.

Neurophysiol. 100, 3328–3342.

Whitﬁeld-Gabrieli, S., Ford, J.M., 2012.Default mode network activity and connectivity in psychopathology. Annu. Rev. Clin. Psychol. 8, 49–76.

Wolf, R.C., Sambataro, F., Vasic, N., Schmid, M., Thomann, P.A., Bienentreu, S.D., Wolf, N.D., 2011.Aberrant connectivity of resting-state networks in borderline personality disor- der. J. Psychiatry Neurosci. 36, 402–411.81 P. Das et al. / NeuroImage 98 (2014) 73–81 R epro duce d w ith p erm is sio n o f th e c o pyrig ht o w ner. F urth er r e pro ductio n p ro hib ite d w ith out p erm is sio n.