Structural magnetic resonance imaging studies in bipolar disorder: A meta-analysis

Bipolar disorder is frequently described in reviews as being associated with subtle structural brain abnormalities, but it is also acknowledged that the existing literature is sparse. 1, 2 The most consistently reported brain abnormalities are increased rates of white matter hyperintensities and mild lateral ventricular enlargement, with studies often disagreeing on other regional volume deviations. This disagreement is contributed to by the small numbers of subjects studied, sample heterogeneity and a dependence on tests of significance with inconsistent ‘positive’ and ‘negative’ findings. Meta-analysis is a powerful technique for integrating quantitative data from several studies; the resultant increase in sample size provides more statistical power to detect subtle volume deviations and provide a more accurate estimate of the effect size. Previous meta-analyses have found that bipolar disorder is associated with preservation of cerebral size 3 and increased rates of white matter hyperintensities. 4, 5 However, these prior meta-analyses were mostly based on computed tomography (CT) studies or qualitative ratings of magnetic resonance imaging (MRI) studies. In this chapter we present results of a meta-analysis of regional brain volume abnormalities of subjects with bipolar disorder based on studies using high-resolution MRI and complete coverage of each brain region.