Neuroanatomical dysconnectivity in bipolar disorder
MetadataShow full item record
This item's downloads: 573 (view details)
Bipolar Disorder (BD) is a major psychotic illness characterized by cyclic mood dysregulation. The dysconnectivity hypothesis suggests that psychotic illnesses arise not from regionally specific focal pathophysiology, but rather from impaired neuroanatomical integration across networks of brain regions. Decreased white matter organization has been hypothesized to be a feature of psychotic illnesses in general, and is supported by meta-analyses of Diffusion Tensor Imaging (DTI) studies in BD and schizophrenia. Although repeatedly associated with white matter microstructural alterations, predominantly among frontal-limbic and posterior parietal white matter, BD has been relatively unexplored through complex network analysis. This method combines structural and diffusion magnetic resonance imaging (MRI) to model the brain as a network and evaluate its topological properties. Few graph analyses investigations to date have probed neuroanatomical connectivity in BD and findings are inconsistent. This thesis utilizes graph analyses to examine features of structural integration and segregation in a moderate sized euthymic bipolar cohort, and then further in a large multi-centre collaboration. Three main methodological approaches were employed: global and regional graph analysis, sub-network analysis, and rich-club connectivity analysis. Through the investigation into the topology of brain networks this work aims to elucidate the extent of neuroanatomical dysconnectivity in BD. The statistical approaches included ANCOVA and independent t-tests in the moderate size cohort. In the multi-centre investigations linear mixed effects modeling was employed to account for the inherent differences due to research centre scanner hardware. Next, our sub-network and rich-club analyses implemented permutation testing to identify between group differences. Results of the current work demonstrate neuroanatomical dysconnectivity as a feature of BD, supported through measures of integration and segregation. Regional dysconnectivity was most prominent in fronto-limbic and posterior parietal nodes across cohorts with BD compared with healthy volunteers. Sub-networks were more weakly connected in left fronto-temporal connections in the larger cohort with BD compared with controls. Rich-club connectivity findings indicated weak deficits in BD across the samples. Furthermore, analysis of rich-club membership provided evidence for impaired organization of right frontal structures. Taken together, the findings suggest that neuroanatomical dysconnectivity in BD is present, diffuse and extends beyond fronto-limbic regions. Neuroanatomical dysconnectivity could be contributed to by impaired integration between neighboring brain regions. The application of graph theory to diffusion tensor imaging data can identify abnormalities of large scale networks associated with BD and help to elucidate the underlying neurobiology of the condition.