Impaired Inter-Hemispheric Integration in Bipolar Disorder Revealed with Brain Network Analyses

BackgroundThis represents the first graph theory-based brain network analysis study in bipolar disorder, a chronic and disabling psychiatric disorder characterized by severe mood swings. Many imaging studies have investigated white matter in bipolar disorder, with results suggesting abnormal white matter structural integrity, particularly in the fronto-limbic and callosal systems. However, many inconsistencies remain in the literature, and no study to date has conducted brain network analyses with a graph-theoretic approach.MethodsWe acquired 64-direction diffusion-weighted magnetic resonance imaging on 25 euthymic bipolar I disorder subjects and 24 gender- and age-equivalent healthy subjects. White matter integrity measures including fractional anisotropy and mean diffusivity were compared in the whole brain. Additionally, structural connectivity matrices based on whole-brain deterministic tractography were constructed, followed by the computation of both global and local brain network measures. We also designed novel metrics to further probe inter-hemispheric integration.ResultsNetwork analyses revealed that the bipolar brain networks exhibited significantly longer characteristic path length, lower clustering coefficient, and lower global efficiency relative to those of control subjects. Further analyses revealed impaired inter-hemispheric but relatively preserved intra-hemispheric integration. These findings were supported by whole-brain white matter analyses that revealed significantly lower integrity in the corpus callosum in bipolar subjects. There were also abnormalities in nodal network measures in structures within the limbic system, especially the left hippocampus, the left lateral orbitofrontal cortex, and the bilateral isthmus cingulate.ConclusionsThese results suggest abnormalities in structural network organization in bipolar disorder, particularly in inter-hemispheric integration and within the limbic system.