BACKGROUND

Previous studies analyzed brain functional connectivity (FC) based on resting-state fMRI (RS-fMRI) data to reveal the neuropathology of bipolar disorder (BD) and suggested that their FC alterations are at widespread network-level. However, few studies have analyzed the dynamic functional network connectivity (dFNC) in BD. Thus, we aimed to reveal the dFNC properties of BD in this study.

METHODS

The RS-fMRI data were collected from 51 unmedicated depressed BD II patients and 50 healthy controls. We analyzed the dFNC properties by using an independent component analysis, sliding window correlation, k-means clustering, and graph theory methods.

RESULTS

The intrinsic brain FNC could be clustered into three configuration states, one with sparse connections between all functional networks (State 1), another with negative correlations between the salience network, cerebellum, basal ganglia and the sensory networks (State 2), and a third with negative correlations between the default mode network and the other functional networks (State 3). The BD patients had increased time in State 2, decreased time in State 3, and increased transition number between states. And the time spent in State 2 was positively correlated with the HDRS24 score in the BD patients. In addition, the BD patients had increased dynamic variance in the small-world properties of FNC.

LIMITATIONS

This study did not examine data from BD patients in other episodes and other BD types.

CONCLUSIONS

This study detected abnormal dFNC properties in BD, which indicated their FNC unstability and provided new insights into the neuropathology of their affective and cognitive deficits.