Cortical thickness differences between bipolar depression and major depressive disorder

Abstract

Objectives: Bipolar disorder (BD) is a psychiatric disorder with high morbidity and mortality that cannot be distinguished from major depressive disorder (MDD) until the first manic episode. A biomarker able to differentiate BD and MDD could help clinicians avoid risks of treating BD with antidepressants without mood stabilizers.

Methods: Cortical thickness differences were assessed using magnetic resonance imaging in BD depressed patients (n = 18), MDD depressed patients (n = 56), and healthy volunteers (HVs) (n = 54). A general linear model identified clusters of cortical thickness difference between diagnostic groups.

Results: Compared to the HV group, the BD group had decreased cortical thickness in six regions, after controlling for age and sex, located within the frontal and parietal lobes, and the posterior cingulate cortex. Mean cortical thickness changes in clusters ranged from 7.6 to 9.6% (cluster-wise p-values from 1.0 e-4 to 0.037). When compared to MDD, three clusters of lower cortical thickness in BD were identified that overlapped with clusters that differentiated the BD and HV groups. Mean cortical thickness changes in the clusters ranged from 7.5 to 8.2% (cluster-wise p-values from 1.0 e-4 to 0.023). The difference in cortical thickness was more pronounced when the subgroup of subjects with bipolar I disorder (BD-I) was compared to the MDD group.

Conclusions: Cortical thickness patterns were distinct between BD and MDD. These results are a step toward developing an imaging test to differentiate the two disorders.

Keywords: bipolar disorder; cortical thickness; magnetic resonance imaging; major depressive disorder; neuroimaging.

Fig. 1

Cortical thickness comparisons of subjects with bipolar disorder (BD), major depressive disorder (MDD), and healthy volunteers (HVs) in the right (A) and left hemispheres (B). Six regions were thinner in patients with BD when compared to HVs, including those in the left inferior parietal, the right caudal middle frontal, the left superior parietal, the right posterior cingulate, the right inferior parietal, and the right supramarginal. Three regions with significant spatial overlap with the regions of BD and HV comparison were found when subjects with BD were compared to subjects with MDD, including the right caudal middle frontal region, the left inferior parietal, and the right precuneus regions. No regions were found altered between the MDD and HV groups. Blue coloration indicates significance threshold of p-value < 0.05. Clusters are overlaid on average inflated images with sulci displayed as darker than gyri.

Fig. 2

Whole brain cluster-wise comparison of cortical thickness of subgroups of the primary comparisons with age and sex as covariates. Whole brain cluster-wise comparison of cortical thickness of subjects with bipolar I disorder (n = 12) when compared to subjects with major depressive disorder (n = 56). The pattern of cortical thickness decreases was more extensive in this subgroup of patients than the combined bipolar disorder population.

Fig. 3

Whole brain cluster-wise comparison of cortical thickness of subjects with major depressive disorder (MDD) with a first-degree relative with a history of mood disorder (n = 24) and healthy volunteers (n = 54). One cluster in the lateral orbital frontal region was found to be thicker in patients in the MDD group.

Fig. 3

Whole brain cluster-wise comparison of cortical thickness of subjects with major depressive disorder (MDD) with a first-degree relative with a history of mood disorder (n = 24) and healthy volunteers (n = 54). One cluster in the lateral orbital frontal region was found to be thicker in patients in the MDD group.