Discrimination between schizophrenia and major depressive disorder by magnetic resonance imaging of the female brain
Introduction
Major depressive disorder (MDD) is a common disorder with a lifetime prevalence reported to range from 8% to 12% in almost every country worldwide (Andrade et al., 2003). Schizophrenia is also common and reported to be ranged from 0.16% to 1.21% (Saha et al., 2005). Depression manifested in 21% to 74% of acute patients with recent onset schizophrenia and in 13% to 50% of those with chronic schizophrenia, while depressive features were found in even greater rates, up to 80%, in schizophrenia (Kollias et al., 2008). These data indicate that discrimination between schizophrenia and MDD is often difficult in the clinical setting.
Many magnetic resonance imaging (MRI) and diffusion tensor imaging (DTI) studies have focused on structural brain abnormalities in schizophrenia and MDD. In schizophrenia, evidence has been obtained showing changes in the frontal and temporal lobes, thalamus, anterior cingulate cortex (ACC), and corpus callosum, and showing dilatation of the Sylvian fissure and the third ventricle (reviewed by Arnone et al., 2009; Glahn et al., 2008; Honea et al., 2005; White et al., 2008). In MDD, changes in the frontal and temporal lobes, cingulum, and the subcortical structures have been reported (Arnone et al., 2012; Bora et al., 2012; Sexton et al., 2009).
Some studies have attempted to discriminate between patients with schizophrenia and healthy subjects using brain anatomical structures obtained by MRI (Leonard et al., 1999; Nakamura et al., 2007; Takayanagi et al., 2010). Other studies including ours reported an unbiased, rater-independent technique known as the voxel-based morphometry (VBM)-based classification approach (Davatzikos et al., 2005; Kawasaki et al., 2007; Ota et al., 2012). Each of these studies showed a fair to excellent classification rate. Additionally, one study evaluated the effectiveness of the structural neuroanatomy derived from MRI images as a diagnostic marker of MDD, and showed a relatively low classification rate (Costafreda et al., 2009). However, to our knowledge, there has been no attempt to produce an MRI-based diagnostic tool to objectively discriminate between schizophrenia and MDD.
Functional MRI and DTI have revealed the fine neural networks in the central nervous system (CNS). Together the thalamus, insula, ACC, and corpus callosum are regarded as the central relay station in the brain, and they can be subdivided into functionally different clusters (Buchsbaum et al., 1996; Makris et al., 2006; McCormick et al., 2006; Witelson, 1989). Several studies have detected subdivided region-specific brain changes for schizophrenia and MDD (Buchsbaum et al., 1996; Coryell et al., 2005; Crespo-Facorro et al., 2000; Makris et al., 2006; Mitelman et al., 2009). However, previous neuroimaging studies conducted to discriminate between schizophrenia and control and between MDD and controls paid little attention to these divisions. Moreover, the robust change of the ventricle size was known to be useful to distinguish the bipolar patients from schizophrenia patients (reviewed by Arnone et al., 2009). Then, it would be suitable to add the ventricles for variables of discriminant analysis between MDD and schizophrenia.
We hypothesized that the characteristic distribution of regional brain changes, especially in the limbic system and ventricles, in schizophrenia patients could have diagnostic value in that it could be used to discriminate individuals with schizophrenia from those with MDD.
Section snippets
Participants
The analysis proceeded in two stages. The first analysis was conducted to produce a statistical model to classify subjects according to the current diagnostic system, and the second analysis was performed to validate the statistical model by classifying a new cohort.
Toward this end, subjects were assigned to two independent groups based on the timing of their participation. The first exploration sample consisted of 25 patients with schizophrenia and 25 with MDD. Consensus diagnosis by at least
Results
The demographic and clinical data of the subjects are shown in Table 1, and the mean regional volumes and FA values of each ROIs were described in Table 2. The differences between the patients with schizophrenia and MDD were analyzed by two sample t-test. There was no significant difference in age, years of education, or whole brain volume between the schizophrenia and MDD groups in either sample.
The stepwise discriminant analysis yielded a model in which three variables were selected. The
Discussion
To our knowledge, this is the first attempt to produce a diagnostic tool to discriminate between schizophrenia and MDD based on structural MRI of the brain. Among the 31 ROIs located in the thalami, ACC, insulae, corpus callosum, 3rd, 4th and lateral ventricles, three ROIs (i.e., the right subcallosal ACC, posterior subregion of the thalamus, and right medial portion of the anterior thalamus.) were selected to produce the discriminant model by the stepwise method. As for the correct
Role of funding sources
This study was supported by Health and Labor Sciences Research Grants (Comprehensive Research on Disability, Health, and Welfare) (M.O. and H.K.), Intramural Research Grant (24-11) for Neurological and Psychiatric Disorders of NCNP (M.O. and H.K.), "Understanding of molecular and environmental bases for brain health" carried out under the Strategic Research Program for Brain Sciences by the Ministry of Education, Culture, Sports, Science and Technology of Japan (H.K.).
Contributors
Miho Ota designed the study and wrote the first draft of the manuscript. Masanori Ishikawa collected data. Noriko Sato managed the analyses. Hiroaki Hori collected data. Daimei Sasayama collected data. Kotaro Hattori collected data. Toshiya Teraishi collected data. Takamasa Noda collected data. Satoko Obu collected data. Yasuhiro Nakata collected data. Teruhiko Higuchi managed the analyses. Hiroshi Kunugi managed the analyses. All authors contributed to and have approved the final manuscript.
Conflict of interest
All authors declare that they have no conflicts of interest.
References (54)
- et al.
Magnetic resonance imaging studies in unipolar depression: systematic review and meta-regression analyses
European Neuropsychopharmacology
(2012) A fast diffeomorphic image registration algorithm
Neuroimage
(2007)- et al.
Shaving diffusion tensor images in discriminant analysis: a study into schizophrenia
Medical Image Analysis
(2006) - et al.
Application of principal component analysis to distinguish patients with schizophrenia from healthy controls based on fractional anisotropy measurements
Neuroimage
(2008) - et al.
Insular cortex abnormalities in schizophrenia: a structural magnetic resonance imaging study of first-episode patients
Schizophrenia Research
(2000) - et al.
Classification of structural images via high-dimensional image warping, robust feature extraction, and SVM
Medical Image Computing and Computer-Assisted Intervention
(2005) - et al.
Meta-analysis of gray matter anomalies in schizophrenia: application of anatomic likelihood estimation and network analysis
Biological Psychiatry
(2008) - et al.
Cerebral asymmetry and the effect of sex and handedness on brain structure: a voxel-based morphometric analysis of 465 normal adult human brains
Neuroimage
(2001) - et al.
Hippocampal size in women but not men with schizophrenia relates to disorder duration
Psychiatry Research
(2011) - et al.
DtiStudio: resource program for diffusion tensor computation and fiber bundle tracking
Computer Methods and Programs in Biomedicine
(2006)