[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.159.197.114. Please contact the publisher to request reinstatement.
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Download PDF
Figure 1.
Regions Showing Functional Connectivity Group Differences With the Right Superior Anterior Temporal Lobe (RSATL) Seed Region for Self-blaming vs Other-Blaming Emotions Between the Recurring Episode Major Depressive Disorder (MDD) Group and the Stable Remission MDD Group
Regions Showing Functional Connectivity Group Differences With the Right Superior Anterior Temporal Lobe (RSATL) Seed Region for Self-blaming vs Other-Blaming Emotions Between the Recurring Episode Major Depressive Disorder (MDD) Group and the Stable Remission MDD Group

Cropped images are displayed at an uncorrected voxel-level threshold of P = .005, with no cluster-size threshold. A predictive linear discriminant analysis31 using cluster average regression coefficients from the subgenual cingulate cortex and adjacent septal region (SCSR), right temporal junction (RTPJ), and right ventral putamen and claustrum (RVP), including cases with lower magnetic resonance imaging quality (n = 64), resulted in 79% correctly classified patients into recurring vs stable groups, which could be cross-validated using the well-established leave-one-out method in 75% of the cases (positive predictive value of 74%, negative predictive value of 76%, Wilks λ = .681, P < .001, estimating prior probabilities from subgroup sizes, 1000 bootstrap samples). In contrast, repeating this analysis only based on clinical variables (number of previous episodes, time in remission, time since stopping antidepressant medication, Beck Depression Inventory, Montgomery-Åsberg Depression Rating Scale, and Global Assessment of Functioning scores), prediction accuracy was at chance levels (61% cross-validated, Wilks λ = .855, P = .16). The cluster size of the SCSR effect is 56 (Table), which is large compared with the relatively small size of the anatomical region. Our statistical inference for this finding is based on the peak voxel effect size that survives voxel-based familywise error correction at P = .05 over our a priori region of interest. The other clusters survive familywise error correction over the whole brain. Note that some functional magnetic resonance imaging findings appear to fall on white matter; this is at least partly because of individual anatomical variability. We used a gray matter mask for inclusive masking to only retain voxels with gray matter density values on averaged segmentations of greater than 0.10. Therefore, gray matter is present in at least some participants in all the voxels of each cluster, even if the projection onto the template gives the impression of falling on white matter. The peak voxel (Cohen d = 0.98, corresponding to t54 = 3.59) and cluster average (d = 1.07) effect sizes for the subgenual region are large. They are even larger for the RTPJ (d = 1.22) and RVP (d = 1.63) cluster averages. On a cautionary note, these effect size estimates are post hoc and may therefore overestimate the true effect size, which needs to be determined in future independent studies. In contrast, the large multivariate effect size (Wilks λ = 0.681) of the linear discriminant analysis, which was independently cross-validated, provides a robust estimate that can inform power calculations for future studies. Blue line indicates relatively higher connectivity; yellow dashed line, relatively lower connectivity.

Figure 2.
Connectivity Coefficients for Posterior Right Superior Anterior Temporal Lobe–Subgenual Cingulate Cortex and Adjacent Septal Region (RSATL-SCSR) for Self-blaming and Other-Blaming Emotions vs Baseline
Connectivity Coefficients for Posterior Right Superior Anterior Temporal Lobe–Subgenual Cingulate Cortex and Adjacent Septal Region (RSATL-SCSR) for Self-blaming and Other-Blaming Emotions vs Baseline

Patients with stable major depressive disorder (MDD) and recurring MDD are shown. The RSATL-SCSR connectivity differences were driven by an interaction between emotion (self-blaming vs other-blaming) and group (recurring episode MDD [n = 25] vs stable MDD [n = 31]) (F1,54 = 16.23, P < .001). As can be seen, the interaction arose by higher connectivity during self-blame and by lower connectivity during other-blame in the recurring episode MDD group and the opposite pattern in the stable remission MDD group. There were no significant main effects of emotion (F1,54 = 2.303, P = .14) or group (F1,54 = 0.016, P = .90).

Figure 3.
Connectivity Coefficients for Posterior Right Superior Anterior Temporal Lobe–Subgenual Cingulate Cortex and Adjacent Septal Region (RSATL-SCSR) for Self-blaming vs Other-Blaming Emotions
Connectivity Coefficients for Posterior Right Superior Anterior Temporal Lobe–Subgenual Cingulate Cortex and Adjacent Septal Region (RSATL-SCSR) for Self-blaming vs Other-Blaming Emotions

Patients with stable major depressive disorder (MDD), subthreshold MDD, and recurring MDD are shown. A secondary data analysis on the extracted SCSR regression coefficients (cluster averages) in the larger data set, including the subthreshold MDD group and patients with nonoptimal quality of functional magnetic resonance imaging data, confirmed a connectivity difference for self-blaming vs other-blaming emotions between the MDD groups (F2,74 = 6.39, P = .003). Post hoc pairwise comparisons showed increased RSATL-SCSR connectivity for self-blaming vs other-blaming emotions in the recurring episode MDD group compared with both the stable remission MDD group (mean [SE] difference, 1.65 [0.48]; 95% CI, 0.69-2.61; P = .001) and the subthreshold MDD group (mean [SE] difference, 1.59 [0.68]; 95% CI, 0.24-2.95; P = .001), with no difference between the stable remission and subthreshold MDD groups (mean [SE] difference, −0.05 [0.65]; 95% CI, −1.25 to 1.36; P = .93).

Table.  
RSATL Psychophysiological Interaction Effects for the Recurring Episode MDD Group vs the Stable Remission MDD Group (Self-blame vs Other-Blame Emotions)a
RSATL Psychophysiological Interaction Effects for the Recurring Episode MDD Group vs the Stable Remission MDD Group (Self-blame vs Other-Blame Emotions)a
1.
Eaton  WW, Shao  H, Nestadt  G, Lee  HB, Bienvenu  OJ, Zandi  P.  Population-based study of first onset and chronicity in major depressive disorder. Arch Gen Psychiatry. 2008;65(5):513-520.
PubMedArticle
2.
Zahn  R.  The role of neuroimaging in translational cognitive neuroscience. Top Magn Reson Imaging. 2009;20(5):279-289.
PubMedArticle
3.
Meyer-Lindenberg  A, Weinberger  DR.  Intermediate phenotypes and genetic mechanisms of psychiatric disorders. Nat Rev Neurosci. 2006;7(10):818-827.
PubMedArticle
4.
Abramson  LY, Seligman  ME, Teasdale  JD.  Learned helplessness in humans: critique and reformulation. J Abnorm Psychol. 1978;87(1):49-74.
PubMedArticle
5.
Beck  AT.  Thinking and depression, I: idiosyncratic content and cognitive distortions. Arch Gen Psychiatry. 1963;9:324-333.
PubMedArticle
6.
Green  S, Moll  J, Deakin  JF, Hulleman  J, Zahn  R.  Proneness to decreased negative emotions in major depressive disorder when blaming others rather than oneself. Psychopathology. 2013;46(1):34-44.
PubMedArticle
7.
Zahn  R, Lythe  KE, Gethin  JA,  et al.  Negative emotions towards others are diminished in remitted major depression. Eur Psychiatry. 2015;30(4):448-453.
PubMedArticle
8.
Segal  ZV, Kennedy  S, Gemar  M, Hood  K, Pedersen  R, Buis  T.  Cognitive reactivity to sad mood provocation and the prediction of depressive relapse. Arch Gen Psychiatry. 2006;63(7):749-755.
PubMedArticle
9.
Zahn  R, Moll  J, Paiva  M,  et al.  The neural basis of human social values: evidence from functional MRI. Cereb Cortex. 2009;19(2):276-283.
PubMedArticle
10.
Green  S, Lambon Ralph  MA, Moll  J, Deakin  JF, Zahn  R.  Guilt-selective functional disconnection of anterior temporal and subgenual cortices in major depressive disorder. Arch Gen Psychiatry. 2012;69(10):1014-1021.
PubMedArticle
11.
Zahn  R, de Oliveira-Souza  R, Bramati  I, Garrido  G, Moll  J.  Subgenual cingulate activity reflects individual differences in empathic concern. Neurosci Lett. 2009;457(2):107-110.
PubMedArticle
12.
Drevets  WC, Ongür  D, Price  JL.  Reduced glucose metabolism in the subgenual prefrontal cortex in unipolar depression. Mol Psychiatry. 1998;3(3):190-191.
PubMedArticle
13.
Ebert  D, Ebmeier  KP.  The role of the cingulate gyrus in depression: from functional anatomy to neurochemistry. Biol Psychiatry. 1996;39(12):1044-1050.
PubMedArticle
14.
Mayberg  HS, Lozano  AM, Voon  V,  et al.  Deep brain stimulation for treatment-resistant depression. Neuron. 2005;45(5):651-660.
PubMedArticle
15.
Ressler  KJ, Mayberg  HS.  Targeting abnormal neural circuits in mood and anxiety disorders: from the laboratory to the clinic. Nat Neurosci. 2007;10(9):1116-1124.
PubMedArticle
16.
Siegle  GJ, Thompson  WK, Collier  A,  et al.  Toward clinically useful neuroimaging in depression treatment: prognostic utility of subgenual cingulate activity for determining depression outcome in cognitive therapy across studies, scanners, and patient characteristics. Arch Gen Psychiatry. 2012;69(9):913-924.
PubMedArticle
17.
Green  S, Ralph  MA, Moll  J, Stamatakis  EA, Grafman  J, Zahn  R.  Selective functional integration between anterior temporal and distinct fronto-mesolimbic regions during guilt and indignation. Neuroimage. 2010;52(4):1720-1726.
PubMedArticle
18.
Zahn  R, Moll  J, Krueger  F, Huey  ED, Garrido  G, Grafman  J.  Social concepts are represented in the superior anterior temporal cortex. Proc Natl Acad Sci U S A. 2007;104(15):6430-6435.
PubMedArticle
19.
Zahn  R, Moll  J, Iyengar  V,  et al.  Social conceptual impairments in frontotemporal lobar degeneration with right anterior temporal hypometabolism. Brain. 2009;132(pt 3):604-616.
PubMedArticle
20.
Green  S, Lambon Ralph  MA, Moll  J,  et al.  The neural basis of conceptual-emotional integration and its role in major depressive disorder. Soc Neurosci. 2013;8(5):417-433.
PubMedArticle
21.
Atkinson  AJ, Colburn  WA, DeGruttola  VG,  et al; Biomarkers Definitions Working Group. Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin Pharmacol Ther. 2001;69(3):89-95.
PubMedArticle
22.
Savitz  JB, Rauch  SL, Drevets  WC.  Clinical application of brain imaging for the diagnosis of mood disorders: the current state of play. Mol Psychiatry. 2013;18(5):528-539.
PubMedArticle
23.
Friston  KJ, Buechel  C, Fink  GR, Morris  J, Rolls  E, Dolan  RJ.  Psychophysiological and modulatory interactions in neuroimaging. Neuroimage. 1997;6(3):218-229.
PubMedArticle
24.
ten Doesschate  MC, Bockting  CL, Koeter  MW, Schene  AH; DELTA Study Group.  Prediction of recurrence in recurrent depression: a 5.5-year prospective study. J Clin Psychiatry. 2010;71(8):984-991.
PubMedArticle
25.
Montgomery  SA, Åsberg  M.  A new depression scale designed to be sensitive to change. Br J Psychiatry. 1979;134:382-389.
PubMedArticle
26.
Wichers  M, Geschwind  N, van Os  J, Peeters  F.  Scars in depression: is a conceptual shift necessary to solve the puzzle? Psychol Med. 2010;40(3):359-365.
PubMedArticle
27.
American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders: DSM-IV-TR. Washington, DC: American Psychiatric Association; 2000.
28.
Weissman  MM, Wickramaratne  P, Adams  P, Wolk  S, Verdeli  H, Olfson  M.  Brief screening for family psychiatric history: the Family History Screen. Arch Gen Psychiatry. 2000;57(7):675-682.
PubMedArticle
29.
Keller  MB, Lavori  PW, Friedman  B,  et al.  The Longitudinal Interval Follow-up Evaluation: a comprehensive method for assessing outcome in prospective longitudinal studies. Arch Gen Psychiatry. 1987;44(6):540-548.
PubMedArticle
30.
Brett  M, Anton  JL, Valabregue  R, Poline  JB. Region of interest analysis using an SPM toolbox. Paper presented at: 8th International Conference on Functional Mapping of the Human Brain; June 2-6, 2002; Sendai, Japan.
31.
Stevens  J. Applied Multivariate Statistics for the Social Sciences.5th ed. New York, NY: Routledge, Taylor & Francis Group; 2009.
32.
Beck  AT, Steer  RA, Carbin  MG.  Psychometric properties of the Beck Depression Inventory: twenty-five years of evaluation. Clin Psychol Rev. 1988;8(1):77-100.Article
33.
Seminowicz  DA, Mayberg  HS, McIntosh  AR,  et al.  Limbic-frontal circuitry in major depression: a path modeling metanalysis. Neuroimage. 2004;22(1):409-418.
PubMedArticle
34.
Buckner  RL, Andrews-Hanna  JR, Schacter  DL.  The brain’s default network: anatomy, function, and relevance to disease. Ann N Y Acad Sci. 2008;1124(1):1-38.
PubMedArticle
35.
Farb  NA, Anderson  AK, Bloch  RT, Segal  ZV.  Mood-linked responses in medial prefrontal cortex predict relapse in patients with recurrent unipolar depression. Biol Psychiatry. 2011;70(4):366-372.
PubMedArticle
36.
Nixon  NL, Liddle  PF, Worwood  G, Liotti  M, Nixon  E.  Prefrontal cortex function in remitted major depressive disorder. Psychol Med. 2013;43(6):1219-1230.
PubMedArticle
37.
Foland-Ross  LC, Cooney  RE, Joormann  J, Henry  ML, Gotlib  IH.  Recalling happy memories in remitted depression: a neuroimaging investigation of the repair of sad mood. Cogn Affect Behav Neurosci. 2014;14(2):818-826.
PubMedArticle
38.
Beck  AT, Rush  AJ, Shaw  BF, Emery  G. Cognitive Therapy of Depression. New York, NY: Guilford Press; 1979.
39.
Moll  J, Zahn  R, de Oliveira-Souza  R, Krueger  F, Grafman  J.  Opinion: the neural basis of human moral cognition. Nat Rev Neurosci. 2005;6(10):799-809.
PubMedArticle
40.
Tavares  P, Lawrence  AD, Barnard  PJ.  Paying attention to social meaning: an fMRI study. Cereb Cortex. 2008;18(8):1876-1885.
PubMedArticle
41.
Ross  LA, Olson  IR.  Social cognition and the anterior temporal lobes. Neuroimage. 2010;49(4):3452-3462.
PubMedArticle
42.
Skipper  LM, Ross  LA, Olson  IR.  Sensory and semantic category subdivisions within the anterior temporal lobes. Neuropsychologia. 2011;49(12):3419-3429.
PubMedArticle
43.
Simmons  WK, Reddish  M, Bellgowan  PS, Martin  A.  The selectivity and functional connectivity of the anterior temporal lobes. Cereb Cortex. 2010;20(4):813-825.
PubMedArticle
44.
Patterson  K, Nestor  PJ, Rogers  TT.  Where do you know what you know? the representation of semantic knowledge in the human brain. Nat Rev Neurosci. 2007;8(12):976-987.
PubMedArticle
45.
Lambon Ralph  MA, Patterson  K.  Generalization and differentiation in semantic memory: insights from semantic dementia. Ann N Y Acad Sci. 2008;1124:61-76.
PubMedArticle
46.
Davey  CG, Yücel  M, Allen  NB, Harrison  BJ.  Task-related deactivation and functional connectivity of the subgenual cingulate cortex in major depressive disorder. Front Psychiatry. 2012;3:14.
PubMedArticle
47.
Grimm  S, Ernst  J, Boesiger  P,  et al.  Increased self-focus in major depressive disorder is related to neural abnormalities in subcortical-cortical midline structures. Hum Brain Mapp. 2009;30(8):2617-2627.
PubMedArticle
48.
Stringaris  A, Vidal-Ribas Belil  P, Artiges  E,  et al; IMAGEN Consortium.  The brain’s response to reward anticipation and depression in adolescence: dimensionality, specificity and longitudinal predictions in a community-based sample [published online June 18, 2015]. Am J Psychiatry.
PubMed
49.
Price  JL, Drevets  WC.  Neurocircuitry of mood disorders. Neuropsychopharmacology. 2010;35(1):192-216.
PubMedArticle
50.
Nieuwenhuys  R, Voogd  J, van Huijzen  C. The Human Central Nervous System, A Synopsis and Atlas. 4th ed. New York, NY: Springer-Verlag; 2007.
51.
Dunn  RT, Kimbrell  TA, Ketter  TA,  et al.  Principal components of the Beck Depression Inventory and regional cerebral metabolism in unipolar and bipolar depression. Biol Psychiatry. 2002;51(5):387-399.
PubMedArticle
52.
Decety  J, Lamm  C.  The role of the right temporoparietal junction in social interaction: how low-level computational processes contribute to meta-cognition. Neuroscientist. 2007;13(6):580-593.
PubMedArticle
53.
Zahn  R, Garrido  G, Moll  J, Grafman  J.  Individual differences in posterior cortical volume correlate with proneness to pride and gratitude. Soc Cogn Affect Neurosci. 2014;9(11):1676-1683.
PubMedArticle
54.
Greicius  MD, Flores  BH, Menon  V,  et al.  Resting-state functional connectivity in major depression: abnormally increased contributions from subgenual cingulate cortex and thalamus. Biol Psychiatry. 2007;62(5):429-437.
PubMedArticle
55.
Sheline  YI, Price  JL, Yan  Z, Mintun  MA.  Resting-state functional MRI in depression unmasks increased connectivity between networks via the dorsal nexus. Proc Natl Acad Sci U S A. 2010;107(24):11020-11025.
PubMedArticle
56.
Grimm  S, Ernst  J, Boesiger  P, Schuepbach  D, Boeker  H, Northoff  G.  Reduced negative BOLD responses in the default-mode network and increased self-focus in depression. World J Biol Psychiatry. 2011;12(8):627-637.
PubMedArticle
57.
Pulcu  E, Lythe  K, Elliott  R,  et al.  Increased amygdala response to shame in remitted major depressive disorder. PLoS One. 2014;9(1):e86900. doi:10.1371/journal.pone.0086900.
PubMedArticle
58.
Zeng  LL, Shen  H, Liu  L, Hu  D.  Unsupervised classification of major depression using functional connectivity MRI. Hum Brain Mapp. 2014;35(4):1630-1641.
PubMedArticle
59.
Marquand  AF, Mourão-Miranda  J, Brammer  MJ, Cleare  AJ, Fu  CH.  Neuroanatomy of verbal working memory as a diagnostic biomarker for depression. Neuroreport. 2008;19(15):1507-1511.
PubMedArticle
60.
Fu  CH, Mourao-Miranda  J, Costafreda  SG,  et al.  Pattern classification of sad facial processing: toward the development of neurobiological markers in depression. Biol Psychiatry. 2008;63(7):656-662.
PubMedArticle
61.
Schmaal  L, Marquand  AF, Rhebergen  D,  et al.  Predicting the naturalistic course of major depressive disorder using clinical and multimodal neuroimaging information: a multivariate pattern recognition study. Biol Psychiatry. 2015;78(4):278-286.
PubMedArticle
Original Investigation
November 2015

Self-blame–Selective Hyperconnectivity Between Anterior Temporal and Subgenual Cortices and Prediction of Recurrent Depressive Episodes

Author Affiliations
  • 1Neuroscience and Aphasia Research Unit, School of Psychological Sciences, The University of Manchester and Manchester Academic Health Sciences Centre, Manchester, England
  • 2Cognitive and Behavioral Neuroscience Unit, D’Or Institute for Research and Education, Rio de Janeiro, Brazil
  • 3doctoral student at The University of Manchester, Manchester, England
  • 4Neuroscience and Psychiatry Unit, Institute of Brain, Behaviour and Mental Health, The University of Manchester and Manchester Academic Health Sciences Centre, Manchester, England
  • 5Department of Psychological Medicine, Centre for Affective Disorders, Institute of Psychiatry, Psychology and Neuroscience, King’s College London, London, England
JAMA Psychiatry. 2015;72(11):1119-1126. doi:10.1001/jamapsychiatry.2015.1813
Abstract

Importance  Patients with remitted major depressive disorder (MDD) were previously found to display abnormal functional magnetic resonance imaging connectivity (fMRI) between the right superior anterior temporal lobe (RSATL) and the subgenual cingulate cortex and adjacent septal region (SCSR) when experiencing self-blaming emotions relative to emotions related to blaming others (eg, “indignation or anger toward others”). This finding provided the first neural signature of biases toward overgeneralized self-blaming emotions (eg, “feeling guilty for everything”), known to have a key role in cognitive vulnerability to MDD. It is unknown whether this neural signature predicts risk of recurrence, a crucial step in establishing its potential as a prognostic biomarker, which is urgently needed for stratification into pathophysiologically more homogeneous subgroups and for novel treatments.

Objective  To use fMRI in remitted MDD at baseline to test the hypothesis that RSATL-SCSR connectivity for self-blaming relative to other-blaming emotions predicts subsequent recurrence of depressive episodes.

Design, Setting, and Participants  A prospective cohort study from June 16, 2011, to October 10, 2014, in a clinical research facility completed by 75 psychotropic medication–free patients with remitted MDD and no relevant comorbidity. In total, 31 remained in stable remission, and 25 developed a recurring episode over the 14 months of clinical follow-up and were included in the primary analysis. Thirty-nine control participants with no personal or family history of MDD were recruited for further comparison.

Main Outcomes and Measures  Between-group difference (recurring vs stable MDD) in RSATL connectivity, with an a priori SCSR region of interest for self-blaming vs other-blaming emotions.

Results  We corroborated our hypothesis that during the experience of self-blaming vs other-blaming emotions, RSATL-SCSR connectivity predicted risk of subsequent recurrence. The recurring MDD group showed higher connectivity than the stable MDD group (familywise error–corrected P < .05 over the a priori SCSR region of interest) and the control group. In addition, the recurring MDD group also exhibited RSATL hyperconnectivity with the right ventral putamen and claustrum and the temporoparietal junction. Together, these regions predicted recurrence with 75% accuracy.

Conclusions and Relevance  To our knowledge, this study is the first to provide a robust demonstration of an fMRI signature of recurrence risk in remitted MDD. Additional studies are needed for its further optimization and validation as a prognostic biomarker.

Introduction

Patients with remitted major depressive disorder (MDD) are at increased risk of developing further episodes over their lifetime.1 Why some patients remain stable while others develop a recurrent episode, however, is elusive. Therefore, there is an urgent need to develop biomarkers of recurrence risk to stratify remitted MDD into pathophysiologically and prognostically more homogeneous subgroups. Mapping the neuroanatomical bases of cognitive and emotional functions using functional magnetic resonance imaging2 (fMRI) is the most promising approach to bridge the gap between clinical symptoms and psychosocial and molecular genetic bases of MDD.3 Such imaging biomarkers serve the development of refined disease models and of novel treatments.

One central feature of cognitive models of vulnerability to MDD is a tendency to overgenerally blame oneself for negative events occurring in one’s personal life4,5 (eg, “My relationship failed; therefore, I am a total failure”). In support of these models, self-blaming emotional biases remained detectable in remitted MDD,6,7 and dormant self-critical attitudes are associated with recurrence risk.8 Proneness to experience self-blaming emotions such as guilt was reproducibly associated with activation of the subgenual cingulate cortex and adjacent septal region (SCSR) in healthy individuals.911 Furthermore, the SCSR exhibited abnormal metabolism in patients with current MDD,12,13 and its normalization and its deep brain electrode-based modulation14 were associated with remission,15 underscoring its central pathophysiological importance. Moreover, SCSR activation predicts outcomes of cognitive therapy,16 which tackles overgeneralized self-blame as central to depressive thinking.5

The SCSR, however, is only part of a brain network relevant for self-blaming emotions and MDD. Using fMRI, our group demonstrated that proneness toward self-blaming emotions in healthy individuals was associated with increased functional connectivity between the SCSR and the right superior anterior temporal cortex (RSATL),17 which we had previously demonstrated to enable differentiated interpretations of the meaning of social behavior18,19 (eg, differentiating actions as “impolite,” or “absent-minded” rather than just overgeneralized as “bad”). A subsequent study10 confirmed our group’s hypothesis17 that patients with remitted MDD exhibit lower functional connectivity between the RSATL and SCSR when experiencing self-blaming emotions (eg, “guilt”) relative to other-blaming emotions (eg, “indignation or anger” toward others) during fMRI. These results provided a specific neural mechanism that can account for biases toward overgeneralized self-blaming emotions20 (eg, feeling “self-disgust” or “guilty for everything”) and the resulting feelings of worthlessness4 in MDD. It is unknown, however, whether these abnormalities prospectively predict risk of recurrence. Prospective prediction of clinical outcomes from the presence of an imaging abnormality is a crucial step in establishing its potential causal role in the pathophysiology of MDD and its promise as a prognostic biomarker that could be used as a novel treatment target.21,22

Herein, we addressed this question by using our group’s previous fMRI paradigm to investigate functional connectivity of temporo-fronto-subcortical networks at baseline in psychotropic medication–free individuals with remitted MDD to predict subsequent recurrence over 14 months of clinical follow-up. Participants were asked to make emotional judgments about sentences evocative of self-blaming emotions (eg, “Tom [participant’s name] acts greedily toward Sam [best friend’s name]”) and emotions related to blaming others (ie, other-blame) (eg, “Sam acts greedily toward Tom”). We also recruited a closely matched control group with no personal or family history of psychiatric disorders to determine whether fMRI signatures predictive of recurrence also differed from those of the control group.

We used psychophysiological interaction (PPI) analysis, an established measure of functional connectivity,23 to test the hypothesis that RSATL-SCSR connectivity for self-blaming relative to other-blaming emotions would predict risk of recurrence in MDD. Based on our group’s previous cross-sectional study10 in remitted MDD, our more specific prediction was to find lower connectivity in the MDD group with a recurring episode compared with the stable remission group and the control group. We further hypothesized that self-blame–selective RSATL-SCSR connectivity would show a predictive effect independent of established clinical predictors24 such as residual symptoms as measured on the Montgomery-Åsberg Depression Rating Scale25 (MADRS) and the number of previous episodes. This hypothesis was based on the expectation that the neural basis of vulnerability to recurrence is at least partly independent of incomplete remission and scarring effects of previous episodes.26

Methods
Participants

This prospective cohort study from June 16, 2011, to October 10, 2014, in a clinical research facility was approved by the South Manchester National Health Service Research Ethics Committee. All participants gave informed consent (verbal for telephone prescreening and written for all other stages) and were compensated for their time and travel costs (eMethods, eTable 1, and eTable 2 in the Supplement).

Inclusion criteria for the MDD group were MDD, according to DSM-IV-TR,27 in remission for at least 6 months (eTable 3 in the Supplement). Main exclusion criteria were current Axis I disorders, including a history of substance or alcohol abuse, and past comorbid Axis I disorders being the likely cause of depressive symptoms (eMethods in the Supplement). The healthy control group had no current or past Axis I disorders and no first-degree family history28 of MDD, bipolar disorder, or schizophrenia. Both groups were psychotropic medication free, right-handed, and native English speaking, with normal vision or vision corrected to normal.

After the initial clinical assessment (eTable 4 and eTable 5 in the Supplement), 138 eligible and available control subjects and participants with MDD underwent fMRI. Participants with MDD were subsequently followed up clinically at 3, 6, and 14 months in person or over the telephone using the well-validated Longitudinal Interval Follow-up Evaluation interview for DSM-IV29 (LIFE-IV) (eMethods in the Supplement). Raters (K.E.L., C.I.W., and R.Z.) were blinded to the fMRI results and had received training by the developers of the LIFE-IV. Interrater reliability was excellent (eTable 6 in the Supplement).

The fMRI data from 12 participants (7 MDD and 5 control) had to be excluded before analysis because of excessive head movement or excessive signal loss (eMethods in the Supplement). Six participants with MDD were lost to follow-up. Of the included 75 psychotropic medication–free patients with remitted MDD, 37 remained in remission (ie, stable MDD group), 27 developed a recurrent major depressive episode (MDE) (ie, recurring MDD group), and 11 developed significant symptoms not meeting MDE criteria (ie, subthreshold MDD group in eMethods in the Supplement) over the 14-month clinical follow-up period. For the primary imaging analysis, we focused on the 31 stable participants, 25 recurring participants, and 39 control subjects meeting the strictest imaging quality control threshold (eMethods, eFigure, and eTable 2 in the Supplement). To probe generalization of our results to the whole sample, we extracted the SCSR cluster averages,30 including those additional 10 MDD participants and 2 control group participants whose imaging data did not pass the strictest quality control threshold (ie, exhibiting greater movement or signal dropout outside the SCSR) and the subthreshold MDD group.

fMRI Acquisition and Paradigm

We used the same fMRI protocol (3-T Achieva; Philips) (eMethods in the Supplement) optimized for detection of ventral brain regions as described previously.10 The T1-weighted 3-dimensional MRIs were acquired for coregistration and axial T2-weighted images to rule out vascular and inflammatory abnormalities (eMethods in the Supplement).

As in our group’s previous study,10 participants saw sentences containing social concepts (eg, “stingy” or “impatient”) describing actions counter to sociomoral values. The agent was the participant (self-agency condition [n = 90]) or his or her best friend (other-agency condition [n = 90]). Self-agency and other-agency conditions contained the same social concepts. Participants were required to report how unpleasant they would feel (“mildly” or “very”) by pressing a button within 5 seconds, followed by a jittered intertrial interval with a mean duration of 4 seconds. A low-level resting-state baseline condition (null condition) requiring no response (n = 90) was pseudorandomly interspersed across 3 runs whose order was counterbalanced across participants.

After the imaging session, participants rated the degree of unpleasantness on a 7-point Likert-type scale (1 is not unpleasant, and 7 is extremely unpleasant) associated with each stimulus. In addition, they were asked to “choose the feeling that they would feel most strongly” from different self-blaming and other-blaming emotions as previously reported.7 Self-blaming and other-blaming emotion trials for the fMRI analysis were defined as those that were perceived as highly unpleasant (those rated at the individual median or above) in the respective self-agency and other-agency conditions.

Behavioral Data Analysis

Behavioral data were analyzed using 2-sided P = .05. A statistical software program was used (SPSS Statistics 20; IBM).

Image Analysis

Functional images were realigned, unwarped, coregistered to the participant’s T1-weighted images, and normalized to the SPM8 (http://www.fil.ion.ucl.ac.uk/spm/) template using the transformation parameters for the T1-weighted image. A smoothing kernel of 6-mm full-width at half maximum was then applied.

To measure functional connectivity, we used the well-established PPI analysis,23 which requires the extraction of the signal from a seed region (in this case, the RSATL) and the creation of an interaction term for the psychological variable (main effect of condition) with the physiological variable (the RSATL signal time course irrespective of condition). As shown previously,10 we used the RSATL seed region coordinates (Montreal Neurological Institute peak coordinates, 58, 0, −12; 6-mm sphere) shown to be equally activated for self-blaming and other-blaming emotions,10,17 which is ideal for a PPI seed region by avoiding confounding coactivation and connectivity differences between conditions to be expected in the SCSR. A PPI effect is a change in the slope of the regression effect of the RSATL on another brain area for one condition (eg, self-blame) relative to another (eg, other-blame).

To investigate our main hypothesis, between-group differences on the contrast self-blaming vs other-blaming emotions were thresholded at P = .005 (uncorrected voxel level). They were then corrected for familywise error at cluster level or voxel level at P = .05 over the a priori SCSR ROI (Montreal Neurological Institute peak coordinates, −4, 23, −5; 6-mm sphere) (eMethods in the Supplement) or the whole brain.

All analyses were inclusively masked with a gray matter mask, and only regions that survived inclusive masking vs the control group at uncorrected P = .005 are reported to ensure the results reflected abnormalities in connectivity. Regression coefficients for the cluster averages of regions resulting from the comparisons between the recurring and stable groups were entered into a predictive linear discriminant analysis31 in SPSS Statistics 20 (eMethods in the Supplement).

Results
Subgroup Characteristics

There were no group differences in the percentages of trials included in the self-blaming and other-blaming emotion conditions, their unpleasantness ratings, response times, or the degree of movement during fMRI (eTable 2 in the Supplement). There were also no differences in age, years of education, and sex between the recurring and stable MDD groups or the recurring MDD group and control group (eTable 7 in the Supplement). The recurring (n = 25) and stable (n = 31) MDD groups did not differ on the number of previous episodes, average length of the last MDE, or average time in remission (t < 0.19, P > .85) (eTable 4 in the Supplement). There was no difference in the MADRS scores at baseline, while the Beck Depression Inventory32 (BDI) scores were higher and the Global Assessment of Functioning (GAF) scores were lower in the recurring MDD group (eTable 7 in the Supplement).

fMRI Findings

Standard blood oxygenation level–dependent (BOLD) effect analyses for self-blaming vs other-blaming emotions revealed no differences between the recurring and stable MDD groups. In contrast, when investigating our main hypothesis using the PPI analysis for self-blaming vs other-blaming emotions, patients with recurring MDD exhibited increased RSATL connectivity with the posterior SCSR, the right ventrolateral putamen (extending into the claustrum), and the right temporoparietal junction compared with patients with stable MDD (Table and Figure 1). All these regions also showed increased connectivity in the recurring MDD group relative to the control group. The reverse comparison of stable vs recurring MDD revealed no areas of increased connectivity.

The RSATL-SCSR connectivity group differences were driven by patients with recurring MDD showing higher connectivity in the self-blaming emotion condition and lower connectivity in the other-blaming emotion condition compared with patients with stable MDD. These results are shown in Figure 2 and Figure 3.

The number of previous MDEs did not correlate with the RSATL-SCSR coupling coefficients in the MDD group (Spearman ρ = 0.111, P = .38), whereas there was a weak correlation with residual symptoms as measured on the MADRS (Spearman ρ = 0.285, P = .02). There were no correlations between the RSATL-SCSR coupling coefficients and the BDI or GAF scores in participants with MDD (ρ ≤ −0.202, P ≥ .11). The RSATL-SCSR connectivity group differences also remained unchanged when using these variables as covariates of no interest (eTable 8 in the Supplement). Finally, when comparing the physiological connectivity of the RSATL, there was no difference between the recurring and stable groups (t54 = −0.90, P = .37) within the SCSR cluster that showed a PPI effect.

Discussion

We found that during the experience of self-blaming relative to other-blaming emotions, connectivity of the RSATL with the SCSR predicted risk of subsequent recurring depressive episodes, as predicted. Intriguingly, contrary to our more specific hypothesis, patients with recurring MDD showed higher rather than lower connectivity compared with the stable group and the control group. In addition to RSATL hyperconnectivity with the SCSR, we also revealed similar effects with the right ventral putamen, claustrum, and the temporoparietal junction as distinctive of recurring compared with stable MDD. While residual depressive symptoms were associated with neural signatures of recurrence risk, most of the variance in RSATL-SCSR connectivity was independent of residual symptoms. Furthermore, there were no associations between RSATL-SCSR connectivity and the number of previous episodes. Together with our finding of 75% accuracy of predicting recurrence from our fMRI measures with no significant predictive value of our clinical measures (Figure 1), this result shows that the fMRI measures add information to clinical predictors and are not solely accounted for by incomplete remission or scarring effects of previous episodes.26

The RSATL-SCSR hyperconnectivity for self-blaming vs other-blaming emotions in the recurring MDD group relative to the other groups was confirmed when extending the analysis to patients with lower-quality MRI acquisitions, corroborating the feasibility of our fMRI measure for wider use. Furthermore, using full MDE criteria as a categorical outcome was supported in that RSATL-SCSR connectivity was comparable between stable and subthreshold MDD.

In contrast to the clear predictive effects of functional connectivity, standard BOLD analyses revealed no differences between recurring and stable MDD. This finding further corroborates the notion that functional connectivity has an important pathophysiological role in MDD33 and may be more informative than regional BOLD10 because it reflects the functional integration of information within networks23 such as the “default mode network” to which both the RSATL and SCSR are contributing.34 Two previous pilot studies35,36 comparing recurring episode patients (N = 10 and N = 7, respectively) and stable remission patients (N = 6 and N = 11, respectively), however, found lower36 and higher35 BOLD effects in medial frontal areas to be predictive of subsequent recurrence. Another study37 found lower ventrolateral frontal BOLD to correlate with subsequent worsening on the BDI. Although interesting, difficulties in adequately controlling for confounding effects of antidepressant medication status36,37 and randomization to different treatments,35 together with small sample sizes, may limit the generalizability of these findings.

This study corroborates the pathophysiological importance of the RSATL for MDD.10 The finding that self-blame–selective changes in RSATL connectivity are associated with vulnerability to MDD recurrence is in keeping with the hypothesis that deficient integration of conceptual social knowledge detail (eg, what it means to act “stingily”) increases proneness to overgeneralized self-blame (eg, “I acted badly”),17,20 described as a central cognitive feature of MDD.4,38 This result is compatible with the view that the RSATL may implicitly enrich moral feelings such as guilt with detailed implicit social meaning (ie, social conceptual representations) even in the absence of verbalization.39 Activation of the anterior temporal lobes for tasks probing social meaning has been corroborated independently.4043 Overall, this evidence is in agreement with a more general view of anterior temporal lobe function as a “hub” representing context-independent aspects of concepts.44,45

Our finding of increased self-blame–selective RSATL connectivity with the ventral putamen supports previous reports of abnormal ventral striatal functional connectivity46 and activation in response to self-negative attribution47 in current MDD. Notably, reduced reward-related ventral striatal BOLD prospectively predicted first-onset depression in adolescents.48 The putamen is part of a core frontal-subcortical circuit that has been implicated in hedonic abnormalities in mood disorders.49 Dysfunction of the adjacent claustrum, which is closely connected with the lateral amygdala,50 has also been associated with anhedonia and psychomotor symptoms in current MDD.51

Our result of self-blame–selective increases in RSATL connectivity with the right temporoparietal junction in the recurring MDD group relative to the stable MDD group and control group is in keeping with its proposed role in enabling internal predictions about external sensory events.52 Such a role is needed for mental models of social agency (self vs other52) probed on our task.53

Although the direction of effects (namely, self-blame–selective increases rather than decreases in RSATL-SCSR connectivity in our recurring MDD group) is at odds with the expectation based on our group’s previous cross-sectional study,10 it is in keeping with converging findings from resting-state fMRI-based connectivity analyses in current MDD showing subgenual cingulate hyperconnectivity with the default mode network,”54 particularly dorsomedial frontal regions,55 previously implicated in negative self-focus in MDD.47,56 Overall risk of recurrence in MDD samples differed between the present study (23% [13 of 56] with only one MDE) and our group’s previous study10 (56% [14 of 25] with only one MDE) and may explain the discrepancy in the direction of the results. This difference is because the MDD sample in our group’s previous study may have been biased toward patients with relatively stable remission, thus preventing detection of the self-blame–selective RSATL-SCSR hyperconnectivity effects that only occurred in the recurring group of the present study. Future studies are needed to determine whether lower self-blame–selective RSATL connectivity observed in patients with stable remission relative to the control group in the present study and in our group’s previous study reflects correlates of compensation mechanisms, rendering these individuals more resilient against recurrence, rather than correlates of vulnerability as previously surmised.

The following limitations of this study need to be discussed. First, we used a broader definition of self-blaming emotions in the present study compared with previous studies10,17 specifically investigating guilt, which makes the results less comparable, and may have included negative emotions that did not entail blame. This approach, however, increased the simplicity and power of our analysis for future applications and was justified by our finding that 2 important self-blaming emotions (shame and guilt) showed no BOLD activation differences in the SCSR.57 Second, although we demonstrated robust cross-validated positive and negative predictive values of self-blame–selective RSATL hyperconnectivity around 75% (Figure 1), this threshold falls short of the 80% benchmark suggested for clinically useful biomarkers.22 Rather than using a standard approach as chosen herein to investigate regional hypotheses, this benchmark could be achieved in further analyses by using machine-learning algorithms that capture multivariate information across the whole brain and have been successfully used for predicting treatment outcomes in current MDD.5861

Conclusions

We demonstrated that recurrence risk in MDD is predicted by a self-blame–selective increase in RSATL connectivity with the SCSR, right ventral putamen, claustrum, and right temporoparietal junction. Our finding supports the hypothesis that self-blame–selective changes in connectivity with the RSATL have a causal role in the pathophysiology of MDD.10 A definitive proof of causality, however, will require showing that modulation of this neural signature by specific interventions has effects on clinical outcomes.

Back to top
Article Information

Submitted for Publication: June 2, 2015; final revision received August 7, 2015; accepted August 9, 2015.

Corresponding Author: Roland Zahn, MD, Department of Psychological Medicine, Centre for Affective Disorders, Institute of Psychiatry, Psychology and Neuroscience, King’s College London, De Crespigny Park, Main Bldg London, London SE5 8AF, England (roland.zahn@kcl.ac.uk).

Published Online: October 7, 2015. doi:10.1001/jamapsychiatry.2015.1813.

Author Contributions: Drs Lythe and Zahn had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Moll, Green, Lambon Ralph, Deakin, Zahn.

Acquisition, analysis, or interpretation of data: Lythe, Moll, Gethin, Workman, Lambon Ralph, Zahn.

Drafting of the manuscript: Lythe, Deakin, Zahn.

Critical revision of the manuscript for important intellectual content: Moll, Gethin, Workman, Green, Lambon Ralph, Zahn.

Statistical analysis: Lythe, Zahn.

Obtained funding: Lambon Ralph, Deakin, Zahn.

Administrative, technical, or material support: Moll, Gethin, Workman, Lambon Ralph.

Study supervision: Lambon Ralph, Deakin, Zahn.

Conflict of Interest Disclosures: Dr Deakin reported providing consultancy and speaking engagements for Bristol-Myers Squibb, AstraZeneca, Eli Lilly, Schering-Plough, Janssen-Cilag, and Servier (all fees are paid to The University of Manchester as reimbursement for time taken). Dr Deakin also reported having share options in P1vital Limited. No other disclosures were reported.

Funding/Support: This study was funded by Medical Research Council Clinician Scientist Fellowship G0902304 (Dr Zahn). Dr Moll was supported by the LABS-D’Or Hospital Network, Rio de Janeiro, Brazil. Ms Gethin was funded by an Engineering and Physical Sciences Research Council United Kingdom PhD Studentship.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: Argyris Stringaris, PhD, and Paul Stokes, PhD (Institute of Psychiatry, Psychology and Neuroscience, King’s College London) provided helpful comments on the manuscript. No compensation was provided. We thank the participants of this study for their support.

Correction: This article was corrected on October 15, 2015, to add an inadvertently omitted word in the Abstract.

References
1.
Eaton  WW, Shao  H, Nestadt  G, Lee  HB, Bienvenu  OJ, Zandi  P.  Population-based study of first onset and chronicity in major depressive disorder. Arch Gen Psychiatry. 2008;65(5):513-520.
PubMedArticle
2.
Zahn  R.  The role of neuroimaging in translational cognitive neuroscience. Top Magn Reson Imaging. 2009;20(5):279-289.
PubMedArticle
3.
Meyer-Lindenberg  A, Weinberger  DR.  Intermediate phenotypes and genetic mechanisms of psychiatric disorders. Nat Rev Neurosci. 2006;7(10):818-827.
PubMedArticle
4.
Abramson  LY, Seligman  ME, Teasdale  JD.  Learned helplessness in humans: critique and reformulation. J Abnorm Psychol. 1978;87(1):49-74.
PubMedArticle
5.
Beck  AT.  Thinking and depression, I: idiosyncratic content and cognitive distortions. Arch Gen Psychiatry. 1963;9:324-333.
PubMedArticle
6.
Green  S, Moll  J, Deakin  JF, Hulleman  J, Zahn  R.  Proneness to decreased negative emotions in major depressive disorder when blaming others rather than oneself. Psychopathology. 2013;46(1):34-44.
PubMedArticle
7.
Zahn  R, Lythe  KE, Gethin  JA,  et al.  Negative emotions towards others are diminished in remitted major depression. Eur Psychiatry. 2015;30(4):448-453.
PubMedArticle
8.
Segal  ZV, Kennedy  S, Gemar  M, Hood  K, Pedersen  R, Buis  T.  Cognitive reactivity to sad mood provocation and the prediction of depressive relapse. Arch Gen Psychiatry. 2006;63(7):749-755.
PubMedArticle
9.
Zahn  R, Moll  J, Paiva  M,  et al.  The neural basis of human social values: evidence from functional MRI. Cereb Cortex. 2009;19(2):276-283.
PubMedArticle
10.
Green  S, Lambon Ralph  MA, Moll  J, Deakin  JF, Zahn  R.  Guilt-selective functional disconnection of anterior temporal and subgenual cortices in major depressive disorder. Arch Gen Psychiatry. 2012;69(10):1014-1021.
PubMedArticle
11.
Zahn  R, de Oliveira-Souza  R, Bramati  I, Garrido  G, Moll  J.  Subgenual cingulate activity reflects individual differences in empathic concern. Neurosci Lett. 2009;457(2):107-110.
PubMedArticle
12.
Drevets  WC, Ongür  D, Price  JL.  Reduced glucose metabolism in the subgenual prefrontal cortex in unipolar depression. Mol Psychiatry. 1998;3(3):190-191.
PubMedArticle
13.
Ebert  D, Ebmeier  KP.  The role of the cingulate gyrus in depression: from functional anatomy to neurochemistry. Biol Psychiatry. 1996;39(12):1044-1050.
PubMedArticle
14.
Mayberg  HS, Lozano  AM, Voon  V,  et al.  Deep brain stimulation for treatment-resistant depression. Neuron. 2005;45(5):651-660.
PubMedArticle
15.
Ressler  KJ, Mayberg  HS.  Targeting abnormal neural circuits in mood and anxiety disorders: from the laboratory to the clinic. Nat Neurosci. 2007;10(9):1116-1124.
PubMedArticle
16.
Siegle  GJ, Thompson  WK, Collier  A,  et al.  Toward clinically useful neuroimaging in depression treatment: prognostic utility of subgenual cingulate activity for determining depression outcome in cognitive therapy across studies, scanners, and patient characteristics. Arch Gen Psychiatry. 2012;69(9):913-924.
PubMedArticle
17.
Green  S, Ralph  MA, Moll  J, Stamatakis  EA, Grafman  J, Zahn  R.  Selective functional integration between anterior temporal and distinct fronto-mesolimbic regions during guilt and indignation. Neuroimage. 2010;52(4):1720-1726.
PubMedArticle
18.
Zahn  R, Moll  J, Krueger  F, Huey  ED, Garrido  G, Grafman  J.  Social concepts are represented in the superior anterior temporal cortex. Proc Natl Acad Sci U S A. 2007;104(15):6430-6435.
PubMedArticle
19.
Zahn  R, Moll  J, Iyengar  V,  et al.  Social conceptual impairments in frontotemporal lobar degeneration with right anterior temporal hypometabolism. Brain. 2009;132(pt 3):604-616.
PubMedArticle
20.
Green  S, Lambon Ralph  MA, Moll  J,  et al.  The neural basis of conceptual-emotional integration and its role in major depressive disorder. Soc Neurosci. 2013;8(5):417-433.
PubMedArticle
21.
Atkinson  AJ, Colburn  WA, DeGruttola  VG,  et al; Biomarkers Definitions Working Group. Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin Pharmacol Ther. 2001;69(3):89-95.
PubMedArticle
22.
Savitz  JB, Rauch  SL, Drevets  WC.  Clinical application of brain imaging for the diagnosis of mood disorders: the current state of play. Mol Psychiatry. 2013;18(5):528-539.
PubMedArticle
23.
Friston  KJ, Buechel  C, Fink  GR, Morris  J, Rolls  E, Dolan  RJ.  Psychophysiological and modulatory interactions in neuroimaging. Neuroimage. 1997;6(3):218-229.
PubMedArticle
24.
ten Doesschate  MC, Bockting  CL, Koeter  MW, Schene  AH; DELTA Study Group.  Prediction of recurrence in recurrent depression: a 5.5-year prospective study. J Clin Psychiatry. 2010;71(8):984-991.
PubMedArticle
25.
Montgomery  SA, Åsberg  M.  A new depression scale designed to be sensitive to change. Br J Psychiatry. 1979;134:382-389.
PubMedArticle
26.
Wichers  M, Geschwind  N, van Os  J, Peeters  F.  Scars in depression: is a conceptual shift necessary to solve the puzzle? Psychol Med. 2010;40(3):359-365.
PubMedArticle
27.
American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders: DSM-IV-TR. Washington, DC: American Psychiatric Association; 2000.
28.
Weissman  MM, Wickramaratne  P, Adams  P, Wolk  S, Verdeli  H, Olfson  M.  Brief screening for family psychiatric history: the Family History Screen. Arch Gen Psychiatry. 2000;57(7):675-682.
PubMedArticle
29.
Keller  MB, Lavori  PW, Friedman  B,  et al.  The Longitudinal Interval Follow-up Evaluation: a comprehensive method for assessing outcome in prospective longitudinal studies. Arch Gen Psychiatry. 1987;44(6):540-548.
PubMedArticle
30.
Brett  M, Anton  JL, Valabregue  R, Poline  JB. Region of interest analysis using an SPM toolbox. Paper presented at: 8th International Conference on Functional Mapping of the Human Brain; June 2-6, 2002; Sendai, Japan.
31.
Stevens  J. Applied Multivariate Statistics for the Social Sciences.5th ed. New York, NY: Routledge, Taylor & Francis Group; 2009.
32.
Beck  AT, Steer  RA, Carbin  MG.  Psychometric properties of the Beck Depression Inventory: twenty-five years of evaluation. Clin Psychol Rev. 1988;8(1):77-100.Article
33.
Seminowicz  DA, Mayberg  HS, McIntosh  AR,  et al.  Limbic-frontal circuitry in major depression: a path modeling metanalysis. Neuroimage. 2004;22(1):409-418.
PubMedArticle
34.
Buckner  RL, Andrews-Hanna  JR, Schacter  DL.  The brain’s default network: anatomy, function, and relevance to disease. Ann N Y Acad Sci. 2008;1124(1):1-38.
PubMedArticle
35.
Farb  NA, Anderson  AK, Bloch  RT, Segal  ZV.  Mood-linked responses in medial prefrontal cortex predict relapse in patients with recurrent unipolar depression. Biol Psychiatry. 2011;70(4):366-372.
PubMedArticle
36.
Nixon  NL, Liddle  PF, Worwood  G, Liotti  M, Nixon  E.  Prefrontal cortex function in remitted major depressive disorder. Psychol Med. 2013;43(6):1219-1230.
PubMedArticle
37.
Foland-Ross  LC, Cooney  RE, Joormann  J, Henry  ML, Gotlib  IH.  Recalling happy memories in remitted depression: a neuroimaging investigation of the repair of sad mood. Cogn Affect Behav Neurosci. 2014;14(2):818-826.
PubMedArticle
38.
Beck  AT, Rush  AJ, Shaw  BF, Emery  G. Cognitive Therapy of Depression. New York, NY: Guilford Press; 1979.
39.
Moll  J, Zahn  R, de Oliveira-Souza  R, Krueger  F, Grafman  J.  Opinion: the neural basis of human moral cognition. Nat Rev Neurosci. 2005;6(10):799-809.
PubMedArticle
40.
Tavares  P, Lawrence  AD, Barnard  PJ.  Paying attention to social meaning: an fMRI study. Cereb Cortex. 2008;18(8):1876-1885.
PubMedArticle
41.
Ross  LA, Olson  IR.  Social cognition and the anterior temporal lobes. Neuroimage. 2010;49(4):3452-3462.
PubMedArticle
42.
Skipper  LM, Ross  LA, Olson  IR.  Sensory and semantic category subdivisions within the anterior temporal lobes. Neuropsychologia. 2011;49(12):3419-3429.
PubMedArticle
43.
Simmons  WK, Reddish  M, Bellgowan  PS, Martin  A.  The selectivity and functional connectivity of the anterior temporal lobes. Cereb Cortex. 2010;20(4):813-825.
PubMedArticle
44.
Patterson  K, Nestor  PJ, Rogers  TT.  Where do you know what you know? the representation of semantic knowledge in the human brain. Nat Rev Neurosci. 2007;8(12):976-987.
PubMedArticle
45.
Lambon Ralph  MA, Patterson  K.  Generalization and differentiation in semantic memory: insights from semantic dementia. Ann N Y Acad Sci. 2008;1124:61-76.
PubMedArticle
46.
Davey  CG, Yücel  M, Allen  NB, Harrison  BJ.  Task-related deactivation and functional connectivity of the subgenual cingulate cortex in major depressive disorder. Front Psychiatry. 2012;3:14.
PubMedArticle
47.
Grimm  S, Ernst  J, Boesiger  P,  et al.  Increased self-focus in major depressive disorder is related to neural abnormalities in subcortical-cortical midline structures. Hum Brain Mapp. 2009;30(8):2617-2627.
PubMedArticle
48.
Stringaris  A, Vidal-Ribas Belil  P, Artiges  E,  et al; IMAGEN Consortium.  The brain’s response to reward anticipation and depression in adolescence: dimensionality, specificity and longitudinal predictions in a community-based sample [published online June 18, 2015]. Am J Psychiatry.
PubMed
49.
Price  JL, Drevets  WC.  Neurocircuitry of mood disorders. Neuropsychopharmacology. 2010;35(1):192-216.
PubMedArticle
50.
Nieuwenhuys  R, Voogd  J, van Huijzen  C. The Human Central Nervous System, A Synopsis and Atlas. 4th ed. New York, NY: Springer-Verlag; 2007.
51.
Dunn  RT, Kimbrell  TA, Ketter  TA,  et al.  Principal components of the Beck Depression Inventory and regional cerebral metabolism in unipolar and bipolar depression. Biol Psychiatry. 2002;51(5):387-399.
PubMedArticle
52.
Decety  J, Lamm  C.  The role of the right temporoparietal junction in social interaction: how low-level computational processes contribute to meta-cognition. Neuroscientist. 2007;13(6):580-593.
PubMedArticle
53.
Zahn  R, Garrido  G, Moll  J, Grafman  J.  Individual differences in posterior cortical volume correlate with proneness to pride and gratitude. Soc Cogn Affect Neurosci. 2014;9(11):1676-1683.
PubMedArticle
54.
Greicius  MD, Flores  BH, Menon  V,  et al.  Resting-state functional connectivity in major depression: abnormally increased contributions from subgenual cingulate cortex and thalamus. Biol Psychiatry. 2007;62(5):429-437.
PubMedArticle
55.
Sheline  YI, Price  JL, Yan  Z, Mintun  MA.  Resting-state functional MRI in depression unmasks increased connectivity between networks via the dorsal nexus. Proc Natl Acad Sci U S A. 2010;107(24):11020-11025.
PubMedArticle
56.
Grimm  S, Ernst  J, Boesiger  P, Schuepbach  D, Boeker  H, Northoff  G.  Reduced negative BOLD responses in the default-mode network and increased self-focus in depression. World J Biol Psychiatry. 2011;12(8):627-637.
PubMedArticle
57.
Pulcu  E, Lythe  K, Elliott  R,  et al.  Increased amygdala response to shame in remitted major depressive disorder. PLoS One. 2014;9(1):e86900. doi:10.1371/journal.pone.0086900.
PubMedArticle
58.
Zeng  LL, Shen  H, Liu  L, Hu  D.  Unsupervised classification of major depression using functional connectivity MRI. Hum Brain Mapp. 2014;35(4):1630-1641.
PubMedArticle
59.
Marquand  AF, Mourão-Miranda  J, Brammer  MJ, Cleare  AJ, Fu  CH.  Neuroanatomy of verbal working memory as a diagnostic biomarker for depression. Neuroreport. 2008;19(15):1507-1511.
PubMedArticle
60.
Fu  CH, Mourao-Miranda  J, Costafreda  SG,  et al.  Pattern classification of sad facial processing: toward the development of neurobiological markers in depression. Biol Psychiatry. 2008;63(7):656-662.
PubMedArticle
61.
Schmaal  L, Marquand  AF, Rhebergen  D,  et al.  Predicting the naturalistic course of major depressive disorder using clinical and multimodal neuroimaging information: a multivariate pattern recognition study. Biol Psychiatry. 2015;78(4):278-286.
PubMedArticle
×