Objective: Several neuroimaging studies have reported structural brain differences in bipolar disorder using automated methods. While these studies have several advantages over those using region of interest techniques, no study has yet estimated a summary effect size or tested for between-study heterogeneity. We sought to address this issue using meta-analytic techniques applied for the first time in bipolar disorder at the level of the individual voxel.
Methods: A systematic review identified 16 voxel-based morphometry (VBM) studies comparing individuals with bipolar disorder with unaffected controls, of which eight were included in the meta-analysis. In order to take account of heterogeneity, summary effect sizes were computed using a random-effects model with appropriate correction for multiple testing.
Results: Compared with controls, subjects with bipolar disorder had reduced grey matter in a single cluster encompassing the right ventral prefrontal cortex, insula, temporal cortex, and claustrum. Study heterogeneity was widespread throughout the brain; though the significant cluster of grey matter reduction remained once these extraneous voxels had been removed. We found no evidence of publication bias (Eggers p = 0.63).
Conclusions: Bipolar disorder is consistently associated with reductions in right prefrontal and temporal lobe grey matter. Reductions elsewhere may be obscured by clinical and methodological heterogeneity.
© 2012 John Wiley and Sons A/S.