Altered static and dynamic functional network connectivity in Parkinson's disease: A multisite functional magnetic...

Aberrant static functional network connectivity (sFNC) and dynamic functional network connectivity (dFNC) have been inconsistently reported in Parkinson's disease (PD), reflecting the limitations of single-site studies with restricted sample sizes. To address these limitations, we leveraged pooled resting-state functional magnetic resonance imaging (rs-fMRI) data from a multisite study to explore FNC differences in PD. Data from 160 patients with PD and 75 age matched and sex-matched healthy controls (HCs) were collected from four repositories (PPMI, OpenfMRI, FCP/INDI). Independent component analysis combined with sliding-window and k-means clustering approach was applied to examine the FNC among seven resting-state networks: sensorimotor network (SMN), auditory network (AN), visual network (VN), cognitive control network (CCN), default mode network (DMN), subcortical network (SC), and cerebellar network (CB). The FNC and dynamic indices (fraction time, mean dwell time, transition number) were calculated. Static and dynamic measures were compared between the two groups. For sFNC, the PD group exhibited predominantly decreased connectivity between the VN and CCN/AN/SMN, and between the AN and SMN; alongside increased connectivity between the AN and SC/CB. For dFNC, four distinct recurrent states were identified. Significant group differences emerged in states 1, 3, and 4, characterized by widespread abnormalities in both intra- and inter-network connectivity (e.g., DMN, CCN, SC, CB, and sensory networks). Temporally, PD patients demonstrated altered dynamics, including reduced mean dwell time and fractional time in state 4, while showing increased mean dwell time in state 1 and fractional time in states 1 and 3. These multisite rs-fMRI findings demonstrate that PD is characterized by complex patterns of connectivity disruption and inflexible temporal network reconfigurations. Notably, dFNC metrics revealed extensive connectivity reductions that were not apparent in static analysis, highlighting their superior sensitivity and potential utility as biomarkers for tracking disease progression and therapeutic response.