Dynamic functional network connectivity impairment in bipolar disorder and its relationship with global functioning

Abnormal functional connectivity (FC) has been consistently associated with bipolar disorders (BD). Classical FC analyses assume stationarity of brain interactions, although connectivity actually varies over time. Here, we examined alterations in dynamic functional network connectivity (dFNC) in BD, their associations with symptom severity and global functioning, and potential differences between bipolar disorder type I (BD1) and type II (BD2). In this case-control study, we investigated dFNC in 57 patients with BD (29 BD1, 28 BD2) and 43 healthy controls (HCs). Most patients were euthymic at scanning (∼86 %), with only a minority showing residual depressive or hypomanic/mixed symptoms. Resting-state fMRI data were decomposed with spatially constrained independent component analysis and analyzed using a sliding-window approach. Meta-state metrics-number of meta-states, transitions, total distance, and span-were derived and compared across groups. Correlations with clinical measures and Global Assessment of Functioning (GAF) scores were tested. Dynamic metrics (number of meta-states, state transitions, and total distance) were reduced in BD relative to HCs, with the greatest reduction in BD1, followed by BD2. State span did not differ between groups. Across the BD sample, higher GAF scores were positively associated with greater dynamic fluidity, whereas no significant associations emerged with standard symptom scales. In conclusion, BD is characterized by a graded disruption of the spatio-temporal dynamics of large-scale brain networks, most pronounced in BD1. Reduced neural flexibility is linked to poorer global functioning, suggesting that dFNC meta-state metrics may provide clinically relevant markers of illness burden in bipolar disorder.