Subcortical volumes differentiate Major Depressive Disorder, Bipolar Disorder, and remitted Major Depressive Disorder
Introduction
Mood disorders are among the most prevalent and severe of all psychiatric disorders (Kessler et al., 2005, World Health Organization, 2012). Whereas both Major Depressive Disorder (MDD) and Bipolar Disorder (BD) are characterized by the presence of depressive episodes, BD is also associated with manic or hypomanic episodes. Because BD often presents clinically as a depressive episode, patients experiencing this disorder can be misdiagnosed as MDD, leading to inappropriate treatment and prolonged distress (Singh and Rajput, 2006). We know little about neurobiological differences between BD and MDD (de Almeida and Phillips, 2013), which limits effective prevention, diagnosis, and treatment of these disorders.
Subcortical gray matter structures are involved in cognitive processing and emotion generation and regulation (Lindquist et al., 2012, Ochsner et al., 2012); not surprisingly, therefore, investigators have implicated anomalies in these structures in mood disorders (Savitz and Drevets, 2009). More specifically, individuals diagnosed with mood disorders have been found to be characterized by structural and functional abnormalities in the amygdala, hippocampus, caudate and putamen, pallidum, nucleus accumbens, and thalamus (Savitz and Drevets, 2009, Hamilton et al., 2012).
Using meta-analytic methods, Kempton et al. compared regional brain volumes in MDD and BD participants and found that the caudate, corpus callosum cross-sectional area, putamen, pallidum, and hippocampus are smaller in MDD than in BD (Kempton et al., 2011). Importantly, these results were limited to common brain regions previously studied in both MDD and BD, and are susceptible to biases resulting from a wide range of participant inclusion criteria and neuroimaging and statistical methods across studies. Moreover, the broad comparisons of MDD versus BD did not account for heterogeneous disease states, including influences from BD I and BD II, euthymic, manic, hypomanic and depressed BD, and current and remitted MDD. Finally, Kempton et al.'s results may be confounded by differences in illness severity between MDD and BD. It is important, therefore, that investigators directly compare MDD and BD individuals in different states with comparable illness history.
To date, few studies have examined differences in brain structure between individuals diagnosed with MDD and BD. The results of these studies indicate that, compared to MDD, BD is associated with greater deep white matter hyperintensities (Dupont et al., 1995, Silverstone et al., 2003), reduced fractional anisotropy of the left superior longitudinal fasciculus (Versace et al., 2010), decreased habenula volume (Savitz et al., 2011), reduced cortical thickness in caudal middle frontal cortex, inferior parietal cortex, and precuneus (Lan et al., 2014), and increased thalamic volume (Dupont et al., 1995). In addition, Redlich et al. found clusters of reduced gray matter that spanned the hippocampal formation, amygdala, putamen, insula, and temporal pole in depressed BD compared to MDD individuals, and a cluster in anterior cingulate that was reduced in MDD compared to depressed BD individuals (Redlich et al., 2014).
Recently, investigators have begun to examine characteristics of MDD and BD that may persist beyond the clinical episode of depression or mania. For example, researchers have found that individuals with BD who are currently in remission exhibit impairment on tests of visuospatial recognition memory (Rubinsztein et al., 2000). Similarly, in a review of studies of cognitive impairment in individuals who had recovered from MDD, Hasselbach et al. (2011) found that in 9 of 11 of these studies remitted depressed participants exhibited impaired performance on at least one neuropsychological test (Hasselbalch et al., 2011). Researchers have also found that individuals continue to experience impairment in social and occupational functioning following remission of MDD or BD (e.g., Fagiolini et al., 2005, Romera et al., 2010). Importantly, investigators have documented abnormalities in regional brain volumes in individuals who have remitted from MDD and BD. For example, individuals with euthymic BD have lower metabolic rates than do healthy controls and depressed BD individuals (Yildiz et al., 2001). Similarly, individuals with remitted MDD have smaller total and posterior hippocampal volumes than do healthy controls (Neumeister et al., 2006; for review see Lorenzetti et al., 2009). Understanding temporary (i.e., state) vs. enduring (i.e., trait) characteristics of affective disorders will facilitate the identification of targets for prevention and treatment. This is particularly important for MDD and BD, given that improved characterization of remitted MDD and euthymic BD may allow for greater differentiation of these topographically similar states and help to avoid maladaptive consequences of misdiagnosis (Singh and Rajput, 2006).
In this study we directly compare, for the first time, subcortical volumetric differences between individuals diagnosed with BD who are currently euthymic and individuals diagnosed with MDD. In addition, to examine the state versus trait nature of volumetric anomalies in mood disorders, we included a sample of individuals with remitted Major Depression (RMD), in addition to a group of healthy (CTL) individuals. We used FreeSurfer's automated segmentation to assess regional subcortical gray matter volumes of the accumbens area, amygdala, caudate, hippocampus, pallidum, putamen, thalamus, and ventral diencephalon (VD; including hypothalamus). To assess the relation of volumetric abnormalities to the severity of impairment in data-to-day functioning across disorders, we related these volumes to individuals' level of global functioning (Global Assessment of Functioning [GAF]; Endicott et al., 1976). Finally, we used support vector machines (SVMs) to examine whether identified abnormal volumes can be used to classify participants on an individual-by-individual basis (Cortes and Vapnik, 1995).
We hypothesized that MDD individuals would have smaller volumes than would BD and CTL individuals in the regions identified in Kempton et al.'s meta-analysis, including caudate, pallidum, putamen and hippocampus. In addition, based on Redlich et al.’s findings with currently depressed BD individuals (Redlich et al., 2014), we hypothesized MDD-related reductions in amygdala relative to BD individuals. Although Kempton et al. did not find significant differences between BD and CTL participants in these regions, Redlich et al. found BD-related abnormalities that spanned hippocampus, amygdala, caudate, putamen, and thalamus; thus, we hypothesized that BD individuals would be distinguishable from CTLs in these regions. We also hypothesized that volumes of the RMD participants would fall between those of MDD and BD, and MDD and CTL participants. Finally, we hypothesized that using SVMs, the identified abnormal regions would successfully classify the MDD versus BD and both the MDD and BD versus CTL groups.
Section snippets
Participants and clinical information
Participants were 193 individuals: 40 diagnosed with BD (Bipolar I Disorder, all currently euthymic); 57 diagnosed with MDD; 35 diagnosed with past but not current MDD (RMD); and 61 CTLs. All individuals participated in studies at Stanford University in which MRI data were acquired. The Structured Clinical Interview for DSM was administered by trained interviewers to all participants in order to obtain DSM-IV-TR Axis I diagnoses (First et al., 2004). Our team of interviewers have demonstrated
Demographic and clinical characteristics
Demographic and clinical characteristics of the participants in the four groups are presented in Table 1, medications are presented in Supplement Table S2, and comorbidity information is presented in Supplement Table S3. The RMD group had significantly more female (χ2(3) = 12.186, p = 0.007) and older (F(3,189) = 3.076, p = 0.029) participants than did the other three groups, who did not differ significantly from each other (LSD post-hoc tests and two-group chi-square tests p > 0.05). One CTL
Discussion
The goal of this study was to identify subcortical volumes that differentiate participants diagnosed with MDD and BD, RMD participants, and healthy CTL participants, and to use the identified subcortical brain volumes to classify disorder status on an individual-by-individual basis. We found that the BD and MDD groups had smaller caudate volumes than did the CTL group, and that the VD was larger in the MDD group than in the BD and CTL groups, and marginally larger in the RMD than in the BD
Role of the funding source
This study was supported by NIMH grant MH59259 to Dr. Gotlib, a grant from the Gipuzkoako Foru Aldundia (Fellows Gipuzkoa Program) to Dr. Iglesias. The funding sources had no involvement in the study design; data collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.
Contributors
Mr. Sacchet conceived of the work that led to the submission. Mr. Sacchet and Dr. Gotlib designed the work that led to the submission, and drafted the manuscript. Dr. Gotlib and Ms. Livermore acquired the data. Mr. Sacchet, Ms. Livermore, and Drs. Iglesias, Glover, and Gotlib played an important role in interpreting the results, revising the manuscript, and approving the final version. No other individuals meet criteria for authorship. All authors have approved the final article.
Conflict of interest
This study was supported by a National Science Foundation Integrative Graduate Education and Research Traineeship (NSF IGERT) Recipient Award 0801700 to MDS, National Science Foundation Graduate Research Fellowship Program (NSF GRFP) DGE-1147470 to MDS, National Institute of Mental Health (NIMH) Neuroscience Research Training award T32 MH020016 to MDS, NIMH grant MH59259 to Dr. Gotlib, and a grant from the Gipuzkoako Foru Aldundia (Fellows Gipuzkoa Program) to Dr. Iglesias. The authors report
Acknowledgments
The authors are grateful to Brian Knutson and Sheri Johnson for their contribution to data acquisition, to consultants in the Stanford Department of Statistics, Karen Larocque, and Gautam Prasad for insights related to machine learning analyses, to Anderson Winkler for help regarding data visualization, to Bob Dougherty and Tom Brosnan for their help with the assessment of scan parameters, and to David Kennedy for insights regarding neuroanatomy.
References (63)
- et al.
Diagnostic conversion from depression to bipolar disorders: results of a long-term prospective study of hospital admissions
J Affect Disord
(2005) - et al.
Anatomical MRI study of basal ganglia in bipolar disorder patients
Psychiatry Res
(2001) - et al.
Hypothalamic-pituitary-adrenal Axis and bipolar disorder
Psychiatr Clin North Am
(2005) - et al.
Distinguishing between unipolar depression and bipolar depression: current and future clinical and neuroimaging perspectives
Biol Psychiatry
(2013) - et al.
Combining multivariate voxel selection and support vector machines for mapping and classification of fMRI spatial patterns
Neuroimage
(2008) - et al.
Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain
Neuron
(2002) - et al.
High dimensional endophenotype ranking in the search for major depression risk genes
Biol Psychiatry
(2012) Linking reward expectation to behavior in the basal ganglia
Trends Neurosci
(2003)- et al.
Reliability of MRI-derived measurements of human cerebral cortical thickness: the effects of field strength, scanner upgrade and manufacturer
Neuroimage
(2006) - et al.
Cognitive impairment in the remitted state of unipolar depressive disorder: a systematic review
J Affect Disord
(2011)
Reduced caudate gray matter volume in women with major depressive disorder
Psychiatry Res Neuroimaging
Structural brain abnormalities in major depressive disorder: a selective review of recent MRI studies
J Affect Disord
Decreased volume of the brain reward system in alcoholism
Biol Psychiatry
Lithium-induced increase in human brain grey matter
J Lancet
A comparison of automated segmentation and manual tracing for quantifying hippocampal and amygdala volumes
Neuroimage
Neurobiology of depression
Neuron
A volumetric magnetic resonance imaging study of monozygotic twins discordant for bipolar disorder
Psychiatry Res Neuroimaging
A quantitative magnetic resonance imaging study of caudate and lenticular nucleus gray matter volume in primary unipolar major depression: relationship to treatment response and clinical severity
Psychiatry Res
Neural circuits underlying the pathophysiology of mood disorders
Trends Cogn Sci
Social and occupational functioning impairment in patients in partial versus complete remission of a major depressive disorder episode. A six-month prospective epidemiological study
Eur Psychiatry
The dexamethasone/corticotropin-releasing hormone test in depression in bipolar and unipolar affective illness
J Psychiatr Res
Decreased pituitary volume in patients with bipolar disorder
Biol Psychiatry
Bipolar and major depressive disorder: neuroimaging the developmental-degenerative divide
Neurosci Biobehav Rev
Habenula volume in bipolar disorder and major depressive disorder: a high-resolution magnetic resonance imaging study
Biol Psychiatry
Right orbitofrontal corticolimbic and left corticocortical White matter connectivity differentiate bipolar and unipolar depression
Biol Psychiatry
31 P nuclear magnetic resonance spectroscopy findings in bipolar illness: a meta-analysis
Psychiatry Res
Basal ganglia volumes and white matter hyperintensities in patients with bipolar disorder
Am J Psychiatry
Caudate volume measurement in older adults with bipolar disorder
Int J Geriat Psychiatry
Cerebellar atrophy in essential tremor using an automated segmentation method
AJNR Am J Neuroradiol
Support-vector networks
Mach Learn
Magnetic resonance imaging and mood disorders: localization of White matter and other subcortical abnormalities
Arch Gen Psychiatry. Am Med Assoc
Cited by (61)
A structural equation modeling approach using behavioral and neuroimaging markers in major depressive disorder
2024, Journal of Psychiatric ResearchDecreased dorsal attention network homogeneity as a potential neuroimaging biomarker for major depressive disorder
2023, Journal of Affective DisordersUnsupervised cross-domain functional MRI adaptation for automated major depressive disorder identification
2023, Medical Image Analysis