Aberrant network topology of cortical-subcortical circuits in schizophrenia and bipolar disorder: a graph theory study.

BACKGROUND: Schizophrenia (SCH) or bipolar disorder (BD) patients exhibit a variety of abnormalities in brain network organization and function. More comprehensive characterization of both shared and disease-specific network features could provide essential clues to the underlying pathogenesis of these disorders, as well as potential therapeutic targets. OBJECTIVE: In this study, we used graph theory methods to examine the common and specific characteristics of brain network between schizophrenia and bipolar disorder. METHOD: The patients of schizophrenia (n = 89) , patients of bipolar disorder (n = 57) and healthy control (HC) subjects (n = 45) were recruited. Resting-state functional magnetic resonance images were analyzed with a group-level connectivity matrix (GCM) threshold varying from 0.10 to 0.40 in 0.02 steps. RESULTS: Network synchronization at the global level was significantly reduced in both SCH and BD groups compared to the HC group. Patients also exhibited impaired nodal parameters (nodal degree and betweenness centrality), especially SCH patients, in the hippocampus, temporal lobe, parietal lobe, and occipital lobe. The SCH group demonstrated sparser edges [fewer functional connectivity (FC) pathways] within somatosensory-motor Module I, denser edges (stronger FC) within limbic Module V, and reduced edge density between Module I and Module IV (including the default mode network), while both SCH and BD patients displayed denser edges between attention control Module III and Module V. CONCLUSIONS: These results revealed global-level network abnormalities in both SCH and BD, with more extensive nodal- and module-level abnormalities in SCH. These distinct topological characteristics could be useful biomarkers for differential diagnosis as well as treatment guidance and response evaluation.