Dynamic functional connectivity (dFC) based on resting-state fMRI has attracted interest in the field of bipolar disorder (BD), because dFC can better capture the evolving processes of emotion and cognition, which are typically impaired in BD. However, previous dFC studies of BD have typically focused on specific seed brain regions or specific functional brain networks, and they have ignored global dynamic information interaction in the whole brain. This study is aimed to reveal aberrant and interpretable whole-brain dFC patterns of BD.
The resting-state fMRI data collected from 35 euthymic BD patients and 30 healthy people. We developed a new dFC inference pipeline, including the sliding-window method, k-means clustering, a new permutation with zero-inflated Poisson regression method, and a similarity analysis for interpretable states, to examine the different patterns of dFC states between BD patients and healthy participants.
BD patients had significantly more frequent transitions between two specific dFC states, which were respectively close to high-level cognitive networks and low-level sensory networks, than healthy controls (p < 0.05, FDR).
The size of samples and other BD types need to be expanded to validate the results of this study. Possible confounding effect of medication.
This study detected aberrant dFC pattern of BD, which indicated the increased lability of the processes of cognition and emotion in BD, and this finding could improve our understanding of the neuropathological mechanism of BD.

Copyright © 2021. Published by Elsevier B.V.