Journal of Affective Disorders 173 (2015) 45–52

Contents lists available at ScienceDirect

Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad

Research report

Cortico-limbic network abnormalities in individuals with current and past major depressive disorder Paul Klauser a,b,n, Alex Fornito a,b, Valentina Lorenzetti a,b, Christopher G. Davey b,c, Dominic B. Dwyer b, Nicholas B. Allen c,d, Murat Yücel a,b a

Monash Clinical and Imaging Neuroscience, School of Psychological Sciences & Monash Biomedical Imaging, Monash University, Clayton, Australia Melbourne Neuropsychiatry Centre, Department of Psychiatry, The University of Melbourne and Melbourne Health, Australia c Orygen Youth Health Research Centre, Centre for Youth Mental Health, The University of Melbourne, Australia d Melbourne School of Psychological Sciences, The University of Melbourne, Australia b

art ic l e i nf o

a b s t r a c t

Article history: Received 25 June 2014 Received in revised form 17 October 2014 Accepted 20 October 2014 Available online 4 November 2014

Background: Brain abnormalities in fronto-temporal structures have been implicated in major depressive disorder (MDD). This study aims to identify their anatomical distribution and their relation to the time course of the disease. Methods: A whole-brain voxel based morphometry analysis was conducted to assess gray and white matter alterations in 56 participants with a lifetime history of MDD, including currently depressed (cMDD) and remitted patients (rMDD), and 33 matched healthy controls (HC). Results: Compared to HC, MDD participants showed increased white matter volume (WMV) in the uncinate fasciculus (UF) and decreased gray matter density (GMD) on the ventromedial prefrontal cortex (vmPFC). The increased WMV in UF was driven by both cMDD and rMDD groups and positively correlated with depression scores. The GMD decrease in the vmPFC resulted mainly from abnormalities in rMDD and was not correlated with depression scores. Finally, temporal UF and vmPFC white matter showed strong structural covariance suggesting functional interactions between these two brain regions. Limitations: The retrospective and cross-sectional design of the study limits the generalizability of the results. Information concerning ongoing treatment did not allow the exploration of interactions between medication and observed abnormalities. The duration of the remission period could have influenced abnormalities in the subgroup of remitted patients. Conclusions: Fronto-temporal alterations in MDD consist of alterations in a cortico-limbic network involving the ventromedial prefrontal cortex and temporal white matter tracts. State-like abnormalities in the UF survive remission and persist as trait-like abnormalities together with alteration in the vmPFC. & 2014 Elsevier B.V. All rights reserved.

Keywords: Magnetic resonance imaging Voxel-based morphometry Major depressive disorder Uncinate fasciculus Ventromedial prefrontal cortex

1. Introduction Although the biological underpinnings of major depressive disorder (MDD) are not well understood, there is converging evidence in favor of an involvement of frontal and medial temporal brain regions in the pathogenesis of the disease (Price and Drevets, 2012). Neuroimaging studies have played an important role in the identification of structural and functional brain abnormalities in MDD and helped to guide the development of novel therapeutics such as transcranial magnetic stimulation (George et al., 2013) and deep brain stimulation (Mayberg et al., 2005). n Correspondence to: Monash Clinical and Imaging Neuroscience, School of Psychological Sciences & Monash Biomedical Imaging, Monash University, 770 Blackburn Road, Clayton, Victoria 3168, Australia. Fax: þ61 3 9348 0469. E-mail address: [email protected] (P. Klauser).

http://dx.doi.org/10.1016/j.jad.2014.10.041 0165-0327/& 2014 Elsevier B.V. All rights reserved.

The frontal lobe is consistently involved in findings from structural neuroimaging studies, and there is robust evidence of gray matter loss in the anterior cingulate cortex (ACC) from three independent metaanalyses (Bora et al., 2012a; Du et al., 2012; Lai, 2013). The ACC is part of the medial prefrontal network that has been shown to be central in emotional processing. This medial prefrontal network encompasses the medial part of the orbito-frontal cortex, also known as ventromedial prefrontal cortex (vmPFC), and can also be extended to limbic structures as well as to basal ganglia and thalamus to form an extended medial prefrontal network (Price and Drevets, 2010). While findings concerning structural abnormalities in these extended cortical and subcortical components are inconsistent (Bora et al., 2012b), most functional neuroimaging studies in MDD report abnormal activation during tasks assessing emotional processing, and increased connectivity in task-free, so-called resting state conditions (Kerestes et al., 2013; Stuhrmann et al., 2011; Wang et al., 2012).

46

P. Klauser et al. / Journal of Affective Disorders 173 (2015) 45–52

Together, these findings raise the idea that MDD could result from abnormalities in an emotion-processing network implicating the medial prefrontal cortex (e.g., vmPFC) and related limbic structures (e.g., amygdala). This theoretical model of abnormalities in the extended medial prefrontal network in MDD is supported by a large body of structural and functional neuroimaging studies using different modalities, including magnetic resonance imaging (MRI) (Drevets et al., 2008), positron emission tomography (PET) (Monkul et al., 2012), single-photon emission computed tomography (SPECT) (Galynker et al., 1998; Nagafusa et al., 2012) and magnetoencephalography (MEG) (Lu et al., 2012). In some brain regions, these abnormalities are also associated with histopathological changes such as reduced cortical thickness and cellular alterations in frontal gray matter (Rajkowska et al., 1999) as all well as alterations of frontal white matter composition (Tham et al., 2011). A large number of neuroimaging findings from our group (Lorenzetti et al., 2010; Takahashi et al., 2009, 2010a), and others (for review, see Koolschijn et al. (2009)), as well as post-mortem observations, which support the involvement of the cortico-limbic network have used region of interest (ROI) analyses. ROI studies are limited towards areas that can be easily anatomically delimited and manually traced (e.g., amygdala, hippocampus) or regions of theoretical importance, which intrinsically depends on results from previous studies (Bora et al., 2012a). These limitations could be addressed using a broader investigation that is not regionally biased and in which gray and white matter changes associated with the cortico-limbic system (and possibly other networks) are identified via a whole-brain analysis. In addition, it is important to evaluate whether cortico-limbic abnormalities reflect an acute change associated with the experience of a depressive episode, or whether they reflect a more enduring, trait-like characteristic. Here we first hypothesized that a whole-brain comparison between participants with a lifetime history of MDD and healthy controls would reveal structural abnormalities in core regions of the extended medial prefrontal network, namely, the medial prefrontal cortex and the limbic system. Second, we predicted that the severity of these abnormalities would correlate with the level of depressive symptoms. Finally, we anticipated that the distinction between current (cMDD) and remitted (rMDD) depression would allow us to further distinguish between abnormalities present only during acute illness (state markers) and those that persist throughout periods of remission, representing putative trait markers. This study aimed to comprehensively investigate brain abnormalities underlying depression and their relation to symptom expression, as well as, to the time course of the illness. Accordingly, we conducted a whole-brain voxel-based morphometry analysis of the gray and white matter in individuals with a lifetime history of major depressive disorder, including those who are currently depressed and those in remission.

MDD episode and 33 were healthy controls. Six participants were excluded from the analysis due to the presence of gross brain abnormalities (two were currently depressed and 4 were remitted patients), leaving a final sample of 29 currently depressed participants, 27 participants with remitted depression and 33 healthy controls. Inclusion criteria were age between 18 and 50, English as a main language, IQ 470 and normal vision (or corrected to normal). General exclusion criteria were history of head trauma, impaired neuroendocrine function or steroid use, neurological condition or electro-convulsive therapy within the past 6 months. Patients with any current Axis I psychiatric disorder other than an anxiety disorder were also excluded. Healthy controls did not have any current or lifetime history of psychiatric disorder. Among the 56 MDD patients (29 currently depressed and 27 remitted), 33 had been prescribed an antidepressant and 19 were medication-naive for the 6 months preceding their assessment. 17 received selective serotonin reuptake inhibitors (SSRIs), 4 serotonin-norepinephrine reuptake inhibitors (SSNRIs), 3 noradrenergic and specific serotonergic antidepressants (NaSSAs), 2 tricyclic antidepressants (TCAs), 2 monoamines oxidase inhibitors (MAOIs), one lithium and one norepinephrine reuptake inhibitor (NRIs). Participants with a lifetime history of MDD and healthy controls were comparable concerning age, gender, intelligence, alcohol use and intracranial volume (Table 1). As expected, depression scores were the highest in the MDD group, and their positive affect score was the lowest. Within the MDD group, currently depressed participants had the highest depression scores (except for high positive affect), the earliest age of onset, and the highest rates of current anxiety disorder and medication (Table 1). The depression scores were higher in the remitted group than in the control group, suggesting residual depressive symptoms. Five ROI manual-tracing studies from this cohort have already been published (Takahashi et al., 2009, 2010a, 2010b; Lorenzetti et al., 2010; Lorenzetti et al., 2009), but the present study is the first whole-brain analysis on this sample. 2.2. Clinical and neuropsychological data All participants underwent general screening with the Structured Clinical Interview for DSM-IV-TR (SCID-IV-TR) and the Alcohol Use Disorders Identification Test (AUDIT) (Bush et al., 1998). Depressive symptoms were measured using a series of questionnaires including the Beck's Depression Inventory-II (BDI-II) (Beck et al., 1996) and the Mood and Anxiety Symptom Questionnaire (MASQ) (Watson et al., 1995b, 1995a). Premorbid and current intelligence were also measured using The Wechsler Test of Adult Reading (WTAR) (Wechsler, 2001) and the Wechsler Abbreviated Scale of Intelligence (WASI) (Wechsler, 1999) respectively. 2.3. MRI data acquisition

This study was approved by the Mental Health Research and Ethics Committee of Melbourne Health, Melbourne, Australia.

T1-weighted structural MRI data were obtained from a 1.5 T Siemens MAGNETOM Avanto scanner (Siemens, Erlangen, Germany) at St. Vincent's Hospital, Melbourne. Images were acquired with the following parameters: time to echo ¼ 2.3 ms, time repetition ¼2.1 ms, flip angle ¼151, matrix size ¼256  256, giving an isometric voxel dimension of 1 mm3.

2.1. Participants

2.4. MRI data preprocessing

Ninety-five participants were recruited in Melbourne from mental health clinics or from the general community through advertisement in the local media. They received reimbursement for their participation in the study. Among all participants, 31 had a current diagnosis of MDD, 31 were in remission from a previous

Each scan was visually checked to exclude the presence of artefacts or gross anatomical abnormalities that could impact image preprocessing. Voxel-wise analysis of brain gray and white matter volume or density differences was conducted using the DARTEL (Diffeomorphic Anatomical Registration Through

2. Material and methods

P. Klauser et al. / Journal of Affective Disorders 173 (2015) 45–52

47

Table 1 Demographic, clinical and anatomical characteristics of participants. MDD (n¼ 56)

cMDD (n ¼29)

rMDD (n ¼27)

HC (n¼ 33)

HC vs MDD

HC vs cMDD vs rMDD

cMDD vs rMDD

Age

34.02 7 8.96

33.097 8.25

35.02 79.72

34.717 9.93

0.337

F(2,86) ¼0.355, p ¼ 0.702

–

Gender Male Female

16 40

7 22

9 18

12 21

C2 ¼ 0.58, p ¼0.440 – –

C2 ¼ 1.13, p ¼0.568 – –

– – –

23.46 7 8.98

10 / 3 7 / 22 21.0777.96

– – 26.047 9.44

– – –

– – –

– – –

Illness duration

10.29 78.08

11.45 79.26

9.047 6.53

–

–

Number of episodes

3.357 2.96

3.677 3.36

3.09 7 2.64

–

–

21 / 6

12 / 13

–

–

– – t(54)¼  2.135, p ¼ 0.037 t(50.402) ¼ 1.133, p ¼ 0.263 t(31.875)¼ 0.593, p ¼ 0.557 –

18 / 10

4 / 23

–

–

–

t(86)¼ 1.404, p ¼ 0.164

Clinical data Melancholic/atypical First episode/recurrent Age of onset

Medication last 6 m: yes/no Current anxiety disorder: yes/no Current IQ

107.96 79.81

104.867 8.75

111.42 79.93

111.127 10.86

Premorbid IQ

109.54 710.42

107.48 711.39

111.747 8.94

111.647 12.29

BDI

25.36 7 15.80

36.83 78.94

13.04 711.73

3.55 7 4.09

MASQ GD mixed

45.75 7 10.310

50.50 77.78

40.447 10.32

27.87 7 8.28

MASQ GD depression

41.53 7 12.05

47.32 7 9.19

35.04 7 11.68

19.47 7 7.23

MASQ GD anxiety

28.667 8.99

32.187 8.73

24.727 7.67

16.41 76.39

MASQ anxious arousal

36.08 7 12.211

42 712.16

28.87 7 7.65

21.977 4.45

MASQ high positive affect

53.46 7 16.774

43.577 13.48

65.007 12.35

81.107 14.27

MASQ loss of interest

27.7777.68

31.617 6.36

23.487 6.79

14.727 5.04

AUDIT

5.53 7 5.51

5.377 6.23

5.69 7 4.77

4.617 3

Global brain volume ICV (ml)

F(2,85) ¼4.034, p ¼ 0.021 t(87)¼ 0.859, p ¼0.393 F(2,86) ¼1.407, p ¼ 0.250 t(66.718) ¼  9.792, F(2,86) ¼120.572, po 0.001 p o 0.001 t(82)¼  8.224, F(2,81) ¼49.209, po 0.001 p o 0.001 t(83)¼  10.548, F(2,82) ¼66.848, po 0.001 p o 0.001 t(80.682)¼  7.322, F(2,82) ¼32.307, po 0.001 p o 0.001 t(68.737) ¼  7.477, F(2,79) ¼40.472, po 0.001 p o 0.001 t(81)¼ 7.663, p o 0.001 F(2, 80) ¼ 57.194, p o 0.001 t(82.350)¼  9.450, F(2,82) ¼58.682, po 0.001 p o 0.001 t(82.796) ¼  1.003, F(2,83) ¼0.415, p¼ 0.319 p ¼ 0.662

1420.497 131.87 1423.677 129.30 1417.08 7136.96 1436.357 143.39 t(87)¼ 0.531, p ¼ 0.597 F(2,86) ¼0.155, p ¼ 0.856

– – C2 ¼4.96, p ¼ 0.026 C2 ¼14.02, p o 0.001 – – – – – – – – – –

–

If not otherwise specified, the values represent the mean7 SD. AUDIT, Alcohol Use Disorder Identification Test; cMDD, currently depressed participants; GD, General Distress; HC, Healthy Control participants; ICV, intracranial volume; MASQ, Mood and Anxiety Symptom Questionnaire; rMDD, remitted depressed participants. Degrees of Freedom vary due to missing data or differences in variance equality assumption.

Exponentiated Lie Algebra) procedure (Ashburner, 2007) implemented in SPM8 (http://www.fil.ion.ucl.ac.uk/spm/software/ spm8/) running in Matlab 2009b (http://www.mathworks.com. au/products/matlab/). Briefly, each participant's T1-weighted anatomical scan was segmented into distinct tissue compartments using the VBM8 toolbox (http://dbm.neuro.uni-jena.de/vbm) with default parameters. A study-specific template was then generated by normalizing each participant's segmented gray or white matter image to a common DARTEL-MNI space. Native-space gray or white matter images were then spatially normalized to this template to generate gray or white matter densities (i.e., GMD or WMD), optimal for the detection of “mesoscopic” changes like cortical thinning (Radua et al., 2014). In addition, we also generated gray or white matter volumes (i.e., GMV or WMV) that have demonstrated greater sensitivity for the detection of “macroscopic” changes like hippocampal atrophy (Bergouignan et al., 2009; Mevel et al., 2011) or white matter lesions (Li et al., 2013). For the generation of GMV and WMV, Jacobian determinants were employed to modulate gray or white voxel intensities with nonlinear warping only in order to preserve original gray or white matter volumes while discarding initial differences in brain sizes.

The images were then smoothed with an 8 mm full-width-halfmaximum Gaussian kernel prior to statistical analysis. 2.5. Statistical analysis General Linear Models (GLM) were used to test for group differences in volume (modulated data) or density (unmodulated data) at each voxel, as implemented in Randomise (http://fsl.fmrib. ox.ac.uk). All results were corrected for multiple comparison Type I error with a non-parametric cluster-size based procedure (Friston et al., 1993; Nichols and Holmes, 2002). A voxel-wise threshold was initially set to po0.001 as a compromise between sensitivity to spatially extended versus focal and intense differences (Woo et al., 2014). Then, a cluster-size threshold was calculated from a permutation test (10,000 permutations) and only clusters large enough to have an associated cluster-wise p valueo0.05 at the whole brain level were reported if not otherwise specified (i.e., small volume correction for pairwise analyses between HC and cMDD or rMDD). Age and gender were used as covariates in the GLM. Tests for group differences in demographics or neuropsychological and clinical characteristics were performed using SPSS version 21 (IBM

48

P. Klauser et al. / Journal of Affective Disorders 173 (2015) 45–52

Table 2 Volume and density differences between MDD and HC groups.

Modulated data WMV MDD4HC Unmodulated data GMD HC 4MDD

Cluster origin

Hemisphere

Extension

MNI

t

Voxels

Corrected p

Uncinate fasciculus Ventromedial prefrontal Cortex

L R and L

Inferior longitudinal Fasciculus –

 35, 6,  29 0, 54,  18

4.05 4.72

562 564

0.0434 0.0318

Two clusters of voxels result from the comparison between subjects that have been diagnosed with a major depressive disorder (MDD), including current or remitted depressed, and healthy controls (HC). The first cluster results from the comparison of modulated white matter volumes (WMV). The second cluster results from the comparison of unmodulated gray matter densities (GMD). Anatomical origin, extension, MNI coordinates (x, y, z) and size (in voxels) are specified for each cluster, as well as t score of the peak intensity voxel and the cluster-wise p values corrected at the whole brain level.

-28

L

-21

-14

-6

2

R

Fig. 1. Comparison between participants with a lifetime history of major depressive disorder and healthy controls. The red and blue clusters of voxels result from the comparison between MDD participants and healthy controls (HC): the red cluster located on the temporal part of the left uncinate fasciculus (UF) represents areas of white matter volume (WMV) increase in MDD (contrast: MDD 4HC); the blue cluster located bilaterally on the ventromedial prefrontal cortex (vmPFC) represents areas of gray matter density (GMD) decrease in MDD (contrast: HC 4MDD). The yellow clusters of voxels represent WMV areas showing significant covariance with first eigenvariate values extracted from the left UF cluster (red). The uncinate tract extracted from the JHU probabilistic atlas (Hua et al., 2008) is represented in green. MNI coordinates on the Z-axis are given on the top of each slice. All clusters are significant at the whole brain level (FWE p o 0.05). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Corp, Armonk, USA). Correlational analyses between cluster-level brain tissue measurements and clinical scores were performed using GraphPad Prism version 6.0d (GraphPad Software, San Diego, USA).

participants who had not (n ¼19). We found no significant differences in WMV or GMD between the two groups. 3.2. Correlation of UF and vmPFC changes with clinical scales of depression

3. Results 3.1. Comparison between individuals with a lifetime history of MDD and HC We first compared all individuals with a lifetime history of MDD (current and remitted) to HC based on two analyses of modulated volumes and unmodulated densities. Only two clusters survived correction for multiple comparisons at the whole brain level. In the analysis of volumes, depressed patients showed one cluster of increased WMV in the temporal segment of left uncinate fasciculus (UF). In the analysis of densities, depressed patients showed one cluster of decreased GMD in the ventromedial prefrontal cortex (vmPFC) extending bilaterally (Table 2 and Fig. 1). To determine whether these differences were driven by cMDD or rMDD patients, we compared each patient group separately with controls. Both cMDD and rMDD showed significantly increased WMV when compared to HC in the UF cluster, suggesting that it is an abnormality common to both groups. However, only the rMDD group showed significant reductions in vmPFC GMD (Fig. 2). To investigate possible medication effects in the MDD group, we ran an additional comparison between MDD participants who had been treated with medication (regardless of their status as current or remitted) during the last 6 months (n ¼33) and MDD

To determine whether the gray and white matter changes identified in MDD participants were related to symptom severity, we extracted first eigenvariate values from the UF WM and vmPFC GM clusters and correlated these with depression scores (i.e., BDI and MASQ) within the whole sample of MDD and HC. There was a significant correlation for UF WMV but not vmPFC GMD in the whole sample. Specifically, variations in WMV of the UF were positively correlated with BDI scores (r ¼0.32, n ¼88, p¼ 0.0023) and all MASQ categories except for the “High Positive Affect” scale, which was negatively correlated (Fig. 3). 3.3. WMV covariance analysis Our findings indicate that when compared to HC, participants with a history of MDD showed WM abnormalities in the temporal region of the left UF as well as GM changes in the vmPFC. The UF represents a major white matter bundle linking orbitofrontal regions with amygdala, thus representing critical components of the frontotemporal network that is central for affective processing and which is thought to be dysfunctional in MDD (Paillere Martinot et al., 2011; Price and Drevets, 2012; Von Der Heide et al., 2013). Prior ROI work in this sample has shown that rMDD patients have increased left amygdala volume when compared to HC (non-significant trend for

P. Klauser et al. / Journal of Affective Disorders 173 (2015) 45–52

UF Cluster

vmPFC Cluster HC rMDD cMDD

-29 0.8

HC rMDD cMDD

-18 0.7

Eigenvariates

0.7

Eigenvariates

L

49

0.6 0.5

0.6

R

Fig. 2. Pairwise comparisons between healthy controls and participants currently depressed or remitted from a major depressive episode. When compared to healthy controls (HC), the two groups of currently depressed (cMDD) and remitted depressed (rMDD) participants showed significant white matter volume (WMV) increase (red and blue clusters) in the uncinate fasciculus (UF) cluster (green), but only the rMDD group (blue) showed gray matter density (GMD) decrease in the ventromedial prefrontal cortex (vmPFC) cluster. Eigenvariates extracted from the initial green clusters in UF and vmPFC illustrate the distribution of data for each subgroup. MNI coordinates on the Zaxis are given on the top of each slice. Red and blue clusters are significant at the level of the green cluster (FWE p o0.05). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

r = 0.374, n = 84, p = 0.0005

60 40 20

r = 0.399, n = 85, p = 0.0002

80

MASQ GD depression

MASQ GD mixed

80

0

60 40 20 0

0.4

0.5

0.6

0.7

0.8

0.4

Eigenvariates UF

40

20

MASQ anxious arousal

MASQ GD anxiety

0.7

0.8

r = 0.401, n = 82, p = 0.0002

80 60 40 20 0

0 0.4

0.5

0.6

0.7

0.8

0.4

Eigenvariates UF

30 20 10 0 0.5

0.6

0.7

0.8

Eigenvariates UF

MASQ high positive affect

40

0.4

0.5

0.6

0.7

0.8

Eigenvariates UF

r = 0.454, n = 85, p < 0.0001

50

MASQ loss of interest

0.6

Eigenvariates UF

r = 0.348, n = 85, p = 0.0011

60

0.5

r = -0.306, n = 83, p = 0.0049

150

100

50

0 0.4

0.5

0.6

0.7

0.8

Eigenvariates UF

Fig. 3. White matter volumes in the UF cluster are correlated with MASQ scores. Spearman correlation coefficients were computed to assess the relationship between first eigenvariate values extracted from the uncinate fasciculus (UF) cluster and 6 MASQ categories in the whole sample of participants (n slightly varies due to missing data).

cMDD) (Lorenzetti et al., 2010). Consistent with these findings, the current voxel-based analysis revealed a non-significant trend for increased amygdala GMV bilaterally in MDD (k¼146 voxels on the right side and 166 on the left, voxel-wise po0.05 uncorrected).

To further examine the relationship between these abnormalities, we examined the structural covariance between UF and vmPFC white matter. For each individual, we extracted WMV first eigenvariate values from the UF cluster showing group differences

50

P. Klauser et al. / Journal of Affective Disorders 173 (2015) 45–52

between MDD and HC and correlated this measure with WMV variations elsewhere in the brain (age and gender were included as nuisance factors, results were thresholded at po0.05 whole-brain, cluster-wise corrected). This analysis revealed a robust structural covariance between the temporal portion of the UF and vmPFC white matter, including the frontal segment of UF, implying either strong anatomical connectivity or mutually trophic influences (Lerch et al., 2006; Mechelli et al., 2005) between these regions (Fig. 1).

4. Discussion In this study, we comprehensively evaluated structural brain changes in gray and white matter in patients with a lifetime history of depression. When compared to HC, all participants with a history of MDD, irrespective of their clinical status (i.e., current or remitted), showed increased WMV in the left UF. In addition we found evidence of a bilateral decrease in vmPFC GMD which was primarily driven by rMDD patients. The extent of the UF WMV abnormalities, but not vmPFC GMD, was associated with more severe depressive symptoms. Notably, UF WMV was significantly correlated with orbitofrontal white matter, implying a direct relationship between these abnormalities. These findings therefore lend to support to the idea of cortico-limbic disturbances as underlying the pathophysiology of MDD. Previous ROI studies in this cohort identified structural abnormalities associated with depression in selected brain regions including the thalamus (i.e., shorter interthalamic adhesion in cMDD as compared to HC) (Takahashi et al., 2009), the amygdala (i.e., increased left amygdala volume in rMDD as compared to HC) (Lorenzetti et al., 2010), the superior temporal regions (i.e., decreased volume in the superior temporal gyri bilaterally and in the left planum temporale in both cMDD and rMDD as compared to HC) (Takahashi et al., 2010b) and insula (i.e., decreased volume in the left insula of both cMDD and rMDD as compared to HC) (Takahashi et al., 2010a), but not in the pituitary gland (Lorenzetti et al., 2009). Together, these studies identified morphometric differences that could represent either state-related brain changes (e.g., shorter interthalamic adhesion) or vulnerability trait markers (e.g., decreased insular volume) in a set of regions distributed in the so called extended medial prefrontal network. Our inability to reproduce our previous ROI-based findings could be explained by the combination of two factors: the higher sensitivity of manual tracing methods to detect volumetric changes in brain areas like the medial temporal structures, including the amygdala (Morey et al., 2009), and the more stringent statistical inference associated with whole brain analysis in VBM (notably, there was a trend-level increase in amygdala GMV of MDD patients). The amygdala volume in rMDD (Lorenzetti et al., 2010) is particularly interesting at the light of our findings because interactions between amygdala and vmPFC in the context of depression have been suggested by several functional (Lee et al., 2007; Kim and Whalen, 2009) and structural (Cremers et al., 2011; Singh et al., 2013) MRI studies. Moreover, these interactions are likely to occur through the UF, which connects the prefrontal cortex to the temporal pole (Von Der Heide et al., 2013). In this context, our two main findings of increased WMV in the left UF and decreased GMD in the vmPFC, as well as the structural covariance between UF and vmPFC WMV, are very pertinent. Structural abnormalities in vmPFC are a common finding in MDD (Konarski et al., 2008) and have been confirmed via metaanalysis of region-of-interest studies (Koolschijn et al., 2009), though they are less common in whole-brain, voxel-based research (Bora et al., 2012a). One recent whole-brain analysis in a large sample of MDD patients combining voxel-based and surface-based approaches reported vmPFC as the only area in which a reduction of cortical thickness was significant (Grieve

et al., 2013). Our finding also echoes a post-mortem study that reported reduced cortical thickness, neuronal sizes, as well as neuronal and glial densities in the upper cortical layers of the rostral orbito-frontal cortex in MDD (Rajkowska et al., 1999). In addition to reduced GMD in vmPFC in the MDD group as a whole, we further reported that this effect was mainly driven by the subgroup of remitted patients, but not currently depressed patients. This difference may reflect a normalization of the deficit by medication, given recent reports of increased GMV in vmPFC after 8 weeks of fluoxetine treatment for MDD (Kong et al., 2014). However, nearly half of the remitted participants were still on medication and we did not find any GMD difference driven by the medication status within the MDD group. Together, it is unlikely that antidepressant medication was a confounding factor in our sample, but the paucity of information concerning the treatment in the MDD groups does not allow ruling out this hypothesis completely. From a mechanistic point of view, vmPFC is a key in the regulation of emotions (Motzkin et al., 2014). Reduced GMV in vmPFC has been correlated with hyperactivation in other regions implicated in emotion processing in several studies investigating both structural and functional changes (Scheuerecker et al., 2010; Wagner et al., 2008; Paillere Martinot et al., 2011). The putative connectivity between vmPFC and temporal lobes, including amygdala is suggested by our covariance analysis in WMV but is also supported by a recent study demonstrating a significant covariance of GMV between amygdala and vmPFC (Singh et al., 2013). Together, this supports the concept of dysregulated interactions between vmPFC and amygdala in MDD as suggested by previous functional MRI studies (Lee et al., 2007; Singh et al., 2013; Townsend et al., 2010). Accordingly, fronto-temporal white matter abnormalities are frequently reported in MDD and UF is probably the white matter tract that is the most frequently implicated in diffusion tensor imaging (DTI) studies (Cullen et al., 2010; Taylor et al., 2007; Dalby et al., 2010; Hettema et al., 2012; Zhang et al., 2012). The correlations that we have reported here between volumetric changes in the left UF cluster and levels of depression measured by BDI and MASQ are comparable to findings from diffusion tensor imaging (DTI) studies that found inverse correlation between fractional anisotropy in UF and depression scores (McIntosh et al., 2013; Zhang et al., 2012). Some overlap between white matter volume and diffusivity changes has been reported in some brain regions during healthy ageing (Fjell et al., 2008) as well as during the course of several mental disorders like Alzheimer's disease (Canu et al., 2010) or autism (Ke et al., 2009). Nevertheless, correlations between volumetric and diffusion properties are generally weak and certainly represent complementary markers of white matter anatomy at a macro and microscopic scales respectively (Abe et al., 2008; Wozniak and Lim, 2006). Accordingly, increased WMV could reflect a higher number of healthy connecting white fibers as well as an inflammatory process involving perivascular swelling in the context of axonal and/or myelin damage. These distinct conditions could theoretically have two opposite effects: an increase or a decrease of fronto-temporal connectivity respectively (Wozniak and Lim, 2006). Several studies support the notion that abnormalities in UF should be considered as vulnerability trait markers for depression because they are associated with a family history of MDD (Huang et al., 2011), as well as with several genetic polymorphisms including BDNF polymorphism (Carballedo et al., 2012), BDNF receptor polymorphism (Murphy et al., 2012) or 5-HTTPLR polymorphism (Pacheco et al., 2009), although these finding are also controversial (Keedwell et al., 2012; Jonassen et al., 2012). In our analysis, abnormalities in UF seem to represent a combination of both trait and state markers. On the one hand, the increased WMV in the UF was mediated by both the currently depressed and the

P. Klauser et al. / Journal of Affective Disorders 173 (2015) 45–52

remitted group of patients, supporting the notion of a trait marker. On the other hand, the significant correlations between WMV in UF and all the depression scores argues in favor of a state marker. This view is supported by a recent study on a very large sample of more than 600 individuals and showing that the association between UF integrity and the state of depressive symptoms was mediated by neuroticism and extraversion traits (McIntosh et al., 2013). In contrast to abnormalities in vmPFC that were detected by the analysis of GMD but not GMV, alterations in UF were significant for volume but not for density. This observation echoes previous studies reporting a strong influence of the modulation process on the sensitivity of the VBM analysis for different type of brain changes and gives clues concerning the nature of the anatomical differences in our sample. Recent experimental evidence indicates that the analysis of volumes generated by the modulation process favors the detection of gross shape and size changes like hippocampal hypertrophy (Bergouignan et al., 2009; Mevel et al., 2011) while the analysis of “densities” improves the sensitivity for smaller changes like gray matter atrophy (Radua et al., 2014). Taken together, this suggests that decreased GMD observed in vmPFC of participants with a history of MDD is likely to be secondary to gray matter loss and cortical atrophy. On the other hand, WMV increase in UF is probably a consequence of remodeling of temporal pole white matter, which would be consistent with amygdala hypertrophy (Lorenzetti et al., 2010). Our interpretation of the results is limited by the lack of details concerning the antidepressant medication taken by currently depressed and remitted participants. Complete information about the ongoing treatment (e.g., dosage, duration, compliance) could have allowed a finer testing of its possible interactions with gray and white matter changes in vmPFC and UF respectively. In addition the length of the remission period in the rMDD group would have been a useful variable that could have potentially given us some cues for the interpretation of the GMV differences between rMDD and cMDD in the vmPFC. In conclusion, we have highlighted that individuals with a lifetime history of MDD show alterations in a putative network of functionally interconnected regions implicating the temporal limbic structures and the ventromedial prefrontal cortex. The higher involvement of the remitted subgroup in changes observed in the prefrontal cortex suggests that abnormalities in this area could represent trait markers of MDD.

Role of funding source Paul Klauser was supported by the Swiss National Science Foundation and the Swiss Society for Medicine and Biology Scholarships (ID: 148384). Alex Fornito was supported by an National Health and Medical Research Council (NHMRC) Project Grant (ID: 1050504) and Australian Research Council (ARC) Future Fellowship (ID: FT130100589). Valentina Lorenzetti was supported by the Monash Bridging Postdoctoral Fellowship. Christopher Davey was supported by an NHMRC Career Development Fellowship (ID: 628922). Murat Yücel was supported by an NHMRC Fellowship Grant (ID: 1021973).

Conflict of interest No conflict declared.

Acknowledgments We thank Drs Orli Schwartz and Diana Maud, who kindly recruited and performed the clinical and neuropsychological assessment of the sample.

References Abe, O., Yamasue, H., Aoki, S., Suga, M., Yamada, H., Kasai, K., et al., 2008. Aging in the CNS: comparison of gray/white matter volume and diffusion tensor data. Neurobiol. Aging 29, 102–116.

51

Ashburner, J., 2007. A fast diffeomorphic image registration algorithm. NeuroImage 38, 95–113. Beck, A.T., Steer, R.A., Brown, G.K., 1996. Manual for the Beck Depression InventoryII.. The Psychological Corporation, San Antonio, TX. Bergouignan, L., Chupin, M., Czechowska, Y., Kinkingnehun, S., Lemogne, C., Le Bastard, G., et al., 2009. Can voxel based morphometry, manual segmentation and automated segmentation equally detect hippocampal volume differences in acute depression? NeuroImage 45, 29–37. Bora, E., Fornito, A., Pantelis, C., Yucel, M., 2012a. Gray matter abnormalities in Major Depressive Disorder: a meta-analysis of voxel based morphometry studies. J. Affect. Disord. 138, 9–18. Bora, E., Harrison, B.J., Davey, C.G., Yucel, M., Pantelis, C., 2012b. Meta-analysis of volumetric abnormalities in cortico–striatal–pallidal–thalamic circuits in major depressive disorder. Psychol. Med. 42, 671–681. Bush, K., Kivlahan, D.R., McDonell, M.B., Fihn, S.D., Bradley, K.A., 1998. The AUDIT alcohol consumption questions (AUDIT-C): an effective brief screening test for problem drinking. Ambulatory Care Quality Improvement Project (ACQUIP). Alcohol Use Disorders Identification Test. Arch. Intern. Med. 158, 1789–1795. Canu, E., McLaren, D.G., Fitzgerald, M.E., Bendlin, B.B., Zoccatelli, G., Alessandrini, F., et al., 2010. Microstructural diffusion changes are independent of macrostructural volume loss in moderate to severe Alzheimer's disease. J. Alzheimers Dis. 19, 963–976. Carballedo, A., Amico, F., Ugwu, I., Fagan, A.J., Fahey, C., Morris, D., et al., 2012. Reduced fractional anisotropy in the uncinate fasciculus in patients with major depression carrying the met-allele of the Val66Met brain-derived neurotrophic factor genotype. Am. J. Med. Genet. B Neuropsychiatr. Genet. 159B, 537–548. Cremers, H., van Tol, M.J., Roelofs, K., Aleman, A., Zitman, F.G., van Buchem, M.A., et al., 2011. Extraversion is linked to volume of the orbitofrontal cortex and amygdala. PLoS One 6, e28421. Cullen, K.R., Klimes-Dougan, B., Muetzel, R., Mueller, B.A., Camchong, J., Houri, A. et al. 2010. Altered white matter microstructure in adolescents with major depression: a preliminary study. J. Am. Acad. Child Adolesc. Psychiatry. 49, 173. e1-83.e1. Dalby, R.B., Frandsen, J., Chakravarty, M.M., Ahdidan, J., Sorensen, L., Rosenberg, R., et al., 2010. Depression severity is correlated to the integrity of white matter fiber tracts in late-onset major depression. Psychiatry Res. 184, 38–48. Drevets, W.C., Price, J.L., Furey, M.L., 2008. Brain structural and functional abnormalities in mood disorders: implications for neurocircuitry models of depression. Brain Struct. Funct. 213, 93–118. Du, M.Y., Wu, Q.Z., Yue, Q., Li, J., Liao, Y., Kuang, W.H., et al., 2012. Voxelwise metaanalysis of gray matter reduction in major depressive disorder. Prog. Neuropsychopharmacol. Biol. Psychiatry 36, 11–16. Fjell, A.M., Westlye, L.T., Greve, D.N., Fischl, B., Benner, T., van der Kouwe, A.J., et al., 2008. The relationship between diffusion tensor imaging and volumetry as measures of white matter properties. NeuroImage 42, 1654–1668. Friston, K.J., Worsley, K.J., Frackowiak, R.S.J., Mazziotta, J.C., Evans, A.C., 1993. Assessing the significance of focal activations using their spatial extent. Hum. Brain Mapp. 1, 210–220. Galynker, I.I., Cai, J., Ongseng, F., Finestone, H., Dutta, E., Serseni, D., 1998. Hypofrontality and negative symptoms in major depressive disorder. J. Nucl. Med. 39, 608–612. George, M.S., Taylor, J.J., Short, E.B., 2013. The expanding evidence base for rTMS treatment of depression. Curr. Opin. Psychiatry 26, 13–18. Grieve, S.M., Korgaonkar, M.S., Koslow, S.H., Gordon, E., Williams, L.M., 2013. Widespread reductions in gray matter volume in depression. Neuroimage Clin. 3, 332–339. Hettema, J.M., Kettenmann, B., Ahluwalia, V., McCarthy, C., Kates, W.R., Schmitt, J.E., et al., 2012. Pilot multimodal twin imaging study of generalized anxiety disorder. Depress. Anxiety 29, 202–209. Hua, K., Zhang, J., Wakana, S., Jiang, H., Li, X., Reich, D.S., et al., 2008. Tract probability maps in stereotaxic spaces: analyses of white matter anatomy and tract-specific quantification. NeuroImage 39, 336–347. Huang, H., Fan, X., Williamson, D.E., Rao, U., 2011. White matter changes in healthy adolescents at familial risk for unipolar depression: a diffusion tensor imaging study. Neuropsychopharmacology 36, 684–691. Jonassen, R., Endestad, T., Neumeister, A., Foss Haug, K.B., Berg, J.P., Landro, N.I., 2012. The effects of the serotonin transporter polymorphism and age on frontal white matter integrity in healthy adult women. Front. Hum. Neurosci. 6, 19. Ke, X., Tang, T., Hong, S., Hang, Y., Zou, B., Li, H., et al., 2009. White matter impairments in autism, evidence from voxel-based morphometry and diffusion tensor imaging. Brain Res. 1265, 171–177. Keedwell, P.A., Chapman, R., Christiansen, K., Richardson, H., Evans, J., Jones, D.K., 2012. Cingulum white matter in young women at risk of depression: the effect of family history and anhedonia. Biol. Psychiatry 72, 296–302. Kerestes, R., Davey, C.G., Stephanou, K., Whittle, S., Harrison, B.J., 2013. Functional brain imaging studies of youth depression: a systematic review. Neuroimage Clin. 4, 209–231. Kim, M.J., Whalen, P.J., 2009. The structural integrity of an amygdala–prefrontal pathway predicts trait anxiety. J. Neurosci. 29, 11614–11618. Konarski, J.Z., McIntyre, R.S., Kennedy, S.H., Rafi-Tari, S., Soczynska, J.K., Ketter, T.A., 2008. Volumetric neuroimaging investigations in mood disorders: bipolar disorder versus major depressive disorder. Bipolar Disord. 10, 1–37. Kong, L., Wu, F., Tang, Y., Ren, L., Kong, D., Liu, Y., et al., 2014. Frontal-subcortical volumetric deficits in single episode, medication-naive depressed patients and the effects of 8 weeks fluoxetine treatment: a VBM-DARTEL study. PLoS One 9, e79055.

52

P. Klauser et al. / Journal of Affective Disorders 173 (2015) 45–52

Koolschijn, P.C., van Haren, N.E., Lensvelt-Mulders, G.J., Hulshoff Pol, H.E., Kahn, R.S., 2009. Brain volume abnormalities in major depressive disorder: a metaanalysis of magnetic resonance imaging studies. Hum. Brain Mapp. 30, 3719–3735. Lai, C.H., 2013. Gray matter volume in major depressive disorder: a meta-analysis of voxel-based morphometry studies. Psychiatry Res. 211, 37–46. Lee, B.T., Seong, W.C., Hyung, S.K., Lee, B.C., Choi, I.G., Lyoo, I.K., et al., 2007. The neural substrates of affective processing toward positive and negative affective pictures in patients with major depressive disorder. Prog. Neuropsychopharmacol. Biol. Psychiatry 31, 1487–1492. Lerch, J.P., Worsley, K., Shaw, W.P., Greenstein, D.K., Lenroot, R.K., Giedd, J., et al., 2006. Mapping anatomical correlations across cerebral cortex (MACACC) using cortical thickness from MRI. NeuroImage 31, 993–1003. Li, W., He, H., Lu, J., Lv, B., Li, M., Jin, Z., 2013. Evaluation of multiple voxel-based morphometry approaches and applications in the analysis of white matter changes in temporal lobe epilepsy. Lect. Notes Comput. Sci. 8090, 268–276. Lorenzetti, V., Allen, N.B., Fornito, A., Pantelis, C., De Plato, G., Ang, A., et al., 2009. Pituitary gland volume in currently depressed and remitted depressed patients. Psychiatry Res. 172, 55–60. Lorenzetti, V., Allen, N.B., Whittle, S., Yucel, M., 2010. Amygdala volumes in a sample of current depressed and remitted depressed patients and healthy controls. J. Affect. Disord. 120, 112–119. Lu, Q., Li, H., Luo, G., Wang, Y., Tang, H., Han, L., et al., 2012. Impaired prefrontal– amygdala effective connectivity is responsible for the dysfunction of emotion process in major depressive disorder: a dynamic causal modeling study on MEG. Neurosci. Lett. 523, 125–130. Mayberg, H.S., Lozano, A.M., Voon, V., McNeely, H.E., Seminowicz, D., Hamani, C., et al., 2005. Deep brain stimulation for treatment-resistant depression. Neuron 45, 651–660. McIntosh, A.M., Bastin, M.E., Luciano, M., Maniega, S.M., Del, C.V.H.M., Royle, N.A., et al., 2013. Neuroticism, depressive symptoms and white-matter integrity in the Lothian Birth Cohort 1936. Psychol. Med. 43, 1197–1206. Mechelli, A., Friston, K.J., Frackowiak, R.S., Price, C.J., 2005. Structural covariance in the human cortex. J. Neurosci. 25, 8303–8310. Mevel, K., Desgranges, B., Baron, J.C., Landeau, B., de La Sayette, V., Viader, F., et al., 2011. Which SPM method should be used to extract hippocampal measures in early Alzheimer's disease? J. Neuroimaging 21, 310–316. Monkul, E.S., Silva, L.A., Narayana, S., Peluso, M.A., Zamarripa, F., Nery, F.G., et al., 2012. Abnormal resting state corticolimbic blood flow in depressed unmedicated patients with major depression: a (15)O–H(2)O PET study. Hum. Brain Mapp. 33, 272–279. Morey, R.A., Petty, C.M., Xu, Y., Hayes, J.P., Wagner, H.R., Lewis, D.V., et al., 2009. A comparison of automated segmentation and manual tracing for quantifying hippocampal and amygdala volumes. NeuroImage 45, 855–866. Motzkin, J.C., Philippi, C.L., Wolf, R.C., Baskaya, M.K., Koenigs, M., 2014. Ventromedial prefrontal cortex is critical for the regulation of amygdala activity in humans. Biol. Psychiatry, http://dx.doi.org/10.1016/j.biopsych.2014.02.014. Murphy, M.L., Carballedo, A., Fagan, A.J., Morris, D., Fahey, C., Meaney, J., et al., 2012. Neurotrophic tyrosine kinase polymorphism impacts white matter connections in patients with major depressive disorder. Biol. Psychiatry 72, 663–670. Nagafusa, Y., Okamoto, N., Sakamoto, K., Yamashita, F., Kawaguchi, A., Higuchi, T., et al., 2012. Assessment of cerebral blood flow findings using 99mTc-ECD single-photon emission computed tomography in patients diagnosed with major depressive disorder. J. Affect. Disord. 140, 296–299. Nichols, T.E., Holmes, A.P., 2002. Nonparametric permutation tests for functional neuroimaging: a primer with examples. Hum. Brain Mapp. 15, 1–25. Pacheco, J., Beevers, C.G., Benavides, C., McGeary, J., Stice, E., Schnyer, D.M., 2009. Frontal-limbic white matter pathway associations with the serotonin transporter gene promoter region (5-HTTLPR) polymorphism. J. Neurosci. 29, 6229–6233. Paillere Martinot, M.L., Martinot, J.L., Ringuenet, D., Galinowski, A., Gallarda, T., Bellivier, F., et al., 2011. Baseline brain metabolism in resistant depression and response to transcranial magnetic stimulation. Neuropsychopharmacology 36, 2710–2719. Price, J.L., Drevets, W.C., 2010. Neurocircuitry of mood disorders. Neuropsychopharmacology 35, 192–216. Price, J.L., Drevets, W.C., 2012. Neural circuits underlying the pathophysiology of mood disorders. Trends Cogn. Sci. 16, 61–71.

Radua, J., Canales-Rodriguez, E.J., Pomarol-Clotet, E., Salvador, R., 2014. Validity of modulation and optimal settings for advanced voxel-based morphometry. Neuroimage 86, 81–90. Rajkowska, G., Miguel-Hidalgo, J.J., Wei, J., Dilley, G., Pittman, S.D., Meltzer, H.Y., et al., 1999. Morphometric evidence for neuronal and glial prefrontal cell pathology in major depression. Biol. Psychiatry 45, 1085–1098. Scheuerecker, J., Meisenzahl, E.M., Koutsouleris, N., Roesner, M., Schopf, V., Linn, J., et al., 2010. Orbitofrontal volume reductions during emotion recognition in patients with major depression. J. Psychiatry Neurosci. 35, 311–320. Singh, M.K., Kesler, S.R., Hadi Hosseini, S.M., Kelley, R.G., Amatya, D., Hamilton, J.P., et al., 2013. Anomalous gray matter structural networks in major depressive disorder. Biol. Psychiatry 74, 777–785. Stuhrmann, A., Suslow, T., Dannlowski, U., 2011. Facial emotion processing in major depression: a systematic review of neuroimaging findings. Biol. Mood Anxiety Disord. 1, 10. Takahashi, T., Yucel, M., Lorenzetti, V., Nakamura, K., Whittle, S., Walterfang, M., et al., 2009. Midline brain structures in patients with current and remitted major depression. Prog. Neuropsychopharmacol. Biol. Psychiatry 33, 1058–1063. Takahashi, T., Yucel, M., Lorenzetti, V., Tanino, R., Whittle, S., Suzuki, M., et al., 2010a. Volumetric MRI study of the insular cortex in individuals with current and past major depression. J. Affect. Disord. 121, 231–238. Takahashi, T., Yucel, M., Lorenzetti, V., Walterfang, M., Kawasaki, Y., Whittle, S., et al., 2010b. An MRI study of the superior temporal subregions in patients with current and past major depression. Prog. Neuropsychopharmacol. Biol. Psychiatry 34, 98–103. Taylor, W.D., MacFall, J.R., Gerig, G., Krishnan, R.R., 2007. Structural integrity of the uncinate fasciculus in geriatric depression: relationship with age of onset. Neuropsychiatr. Dis. Treat. 3, 669–674. Tham, M.W., Woon, P.S., Sum, M.Y., Lee, T.S., Sim, K., 2011. White matter abnormalities in major depression: evidence from post-mortem, neuroimaging and genetic studies. J. Affect. Disord. 132, 26–36. Townsend, J.D., Eberhart, N.K., Bookheimer, S.Y., Eisenberger, N.I., Foland-Ross, L.C., Cook, I.A., et al., 2010. fMRI activation in the amygdala and the orbitofrontal cortex in unmedicated subjects with major depressive disorder. Psychiatry Res. 183, 209–217. Von Der Heide, R.J., Skipper, L.M., Klobusicky, E., Olson, I.R., 2013. Dissecting the uncinate fasciculus: disorders, controversies and a hypothesis. Brain 136, 1692–1707. Wagner, G., Koch, K., Schachtzabel, C., Reichenbach, J.R., Sauer, H., Schlosser Md, R. G., 2008. Enhanced rostral anterior cingulate cortex activation during cognitive control is related to orbitofrontal volume reduction in unipolar depression. J. Psychiatry Neurosci. 33, 199–208. Wang, L., Hermens, D.F., Hickie, I.B., Lagopoulos, J., 2012. A systematic review of resting-state functional-MRI studies in major depression. J. Affect. Disord. 142, 6–12. Watson, D., Clark, L.A., Weber, K., Assenheimer, J.S., Strauss, M.E., McCormick, R.A., 1995a. Testing a tripartite model: II. Exploring the symptom structure of anxiety and depression in student, adult, and patient samples. J. Abnorm. Psychol. 104, 15–25. Watson, D., Weber, K., Assenheimer, J.S., Clark, L.A., Strauss, M.E., McCormick, R.A., 1995b. Testing a tripartite model: I. Evaluating the convergent and discriminant validity of anxiety and depression symptom scales. J. Abnorm. Psychol. 104, 3–14. Wechsler, D., 1999. Manual for the Wechsler Abbreviated Scale of Intelli- gence. The Psychological Corporation, San Antonio, TX. Wechsler, D., 2001. Manual for the Wechsler Test of Adult Reading. The Psychological Corporation, San Antonio, TX. Woo, C.W., Krishnan, A., Wager, T.D., 2014. Cluster-extent based thresholding in fMRI analyses: pitfalls and recommendations. NeuroImage 91, 412–419. Wozniak, J.R., Lim, K.O., 2006. Advances in white matter imaging: a review of in vivo magnetic resonance methodologies and their applicability to the study of development and aging. Neurosci. Biobehav. Rev. 30, 762–774. Zhang, A., Leow, A., Ajilore, O., Lamar, M., Yang, S., Joseph, J., et al., 2012. Quantitative tract-specific measures of uncinate and cingulum in major depression using diffusion tensor imaging. Neuropsychopharmacology 37, 959–967.