Magnetic resonance imaging studies in bipolar disorder and schizophrenia: meta-analysis.

BACKGROUND Several magnetic resonance imaging (MRI) studies have identified structural abnormalities in association with bipolar disorder. The literature is, however, heterogeneous and there is remaining uncertainty about which brain areas are pivotal to the pathogenesis of the condition. AIMS To identify, appraise and summarise volumetric MRI studies of brain regions comparing bipolar disorder with an unrelated control group and individuals with schizophrenia. METHOD A systematic review and random-effects meta-analysis was carried out to identify key areas of structural abnormality in bipolar disorder and whether the pattern of affected areas separated bipolar disorder from schizophrenia. Significant heterogeneity was explored using meta-regression. RESULTS Participants with bipolar disorder are characterised by whole brain and prefrontal lobe volume reductions, and also by increases in the volume of the globus pallidus and lateral ventricles. In comparison with schizophrenia, bipolar disorder is associated with smaller lateral ventricular volume and enlarged amygdala volume. Heterogeneity was widespread and could be partly explained by clinical variables and year of publication, but generally not by differences in image acquisition. CONCLUSIONS There appear to be robust changes in brain volume in bipolar disorder compared with healthy volunteers, although most changes do not seem to be diagnostically specific. Age and duration of illness appear to be key issues in determining the magnitude of observed effect sizes.