Altered Dynamic Functional Connectivity in Parkinson's Disease Patients With Depression.

Background: Depression is one of the most prevalent and disturbing non-motor symptoms in Parkinson's disease (PD), with few dynamic functional connectivity (dFC) features measured in previous studies. Our aim was to investigate the alterations of the dynamics in patients with PD with depression (dPD).

Methods: We performed dFC analysis on the data of resting-state functional MRI from 21 dPD, 34 patients with PD without depression (ndPD), and 43 healthy controls (HCs). Group independent component analysis, a sliding window approach, followed by k-means clustering were conducted to assess functional connectivity states (which represented highly structured connectivity patterns reoccurring over time) and temporal properties for comparison between groups. We further performed dynamic graph-theoretical analysis to examine the variability of topological metrics.

Results: Four distinct functional connectivity states were clustered dFC analysis. Compared to patients with ndPD and HCs, patients with dPD showed increased fractional time and mean dwell time in state 2, characterized by default mode network (DMN)-dominated and cognitive executive network (CEN)-disconnected patterns. Besides, compared to HCs, patients with dPD and patients with ndPD both showed weaker dynamic connectivity within the sensorimotor network (SMN) in state 4, a regionally densely connected state. We additionally observed that patients with dPD presented less variability in the local efficiency of the network.

Conclusions: Our study demonstrated that altered network connection over time, mainly involving the DMN and CEN, with abnormal dynamic graph properties, may contribute to the presence of depression in patients with PD.