Abstract
Objective: Diagnostic differentiation between depressed patients with bipolar disorder or major depressive disorder is a critical issue, as misdiagnosis may result in inappropriate treatment and poor outcome. Specific biomarkers of each disorder that would allow differential diagnosis have not been identified. The aim of study is to determine whether depressed patients with bipolar disorder or major depressive disorder show distinct morphometric abnormalities.
Methods: 569 depressed patients with major depressive disorder, 140 depressed patients with bipolar disorder, and 717 age-matched healthy participants were studied. We examined gray matter volume using voxel-based morphometry. Group classification was performed using a support vector machine algorithm.
Results: Significant differences were found between depressed patients and healthy subjects in gray matter volumes in the left and right anterior cingulate cortex, left and right middle frontal cortex, right dorsolateral frontal cortex, left insula, and left and right temporal poles. Gray matter volumes in each of these regions, with the exception of the left middle frontal cortex and insula, were significantly smaller in depressed patients with bipolar disorder than those with major depressive disorder. A support vector machine model incorporating age, sex, and gray matter volumes in each brain region distinguished patients from healthy subjects with 69.9% accuracy and classified bipolar and major depressive disorder patients with 82.9% accuracy.
Conclusions: Reduced gray matter volume in limbic and paralimbic structures is a shared pathophysiological feature of bipolar disorder and major depressive disorder, whereas severe abnormalities allow differentiation between the two disorders. Our findings identify morphometric biomarkers of two neuropsychiatric disorders that may allow imaging-aided differential diagnosis.