Research on dynamic functional connectivity and graph theoretic features of different motor subtypes in Parkinson’s disease based on ICA

Abstract

Objective This study combined independent component analysis (ICA)-derived dynamic functional connectivity (dFNC) with dynamic graph theory analysis methods to analyze the changes in dFNC and dynamic graph theoretic indicators in different motor subtypes of Parkinson’s disease (PD)—tremor-dominant (TD) subtype and postural instability/gait difficulty (PIGD) subtype—and healthy controls (HC). We aimed to explore the neural circuitry abnormal mechanisms and differences between the two subtypes. Methods Using a prospective observational study design, from January 2021 to May 2024 at Northern Jiangsu People’s Hospital, we enrolled 36 diagnosed PD patients according to inclusion criteria—19 patients with PIGD subtype and 17 patients with TD subtype—along with 19 healthy subjects matched for age, sex, and education level. Resting-state functional magnetic resonance imaging (rs-fMRI), demographic information, and clinical motor function assessment data were collected from all participants. The ICA and dynamic functional network connectivity (dFNC) analysis were performed using the GIFT toolbox platform based on MATLAB. Combined with dynamic graph theory analysis, features such as characteristic path length, clustering coefficient, and node degree were extracted. The dFNC and dynamic graph theoretic indicators of the PIGD group, TD group, and HC group were compared horizontally to clarify the common and specific dynamic abnormalities of brain networks between the two subtypes. Results Nine functional brain networks were identified through ICA. Significant differences among the three groups were observed in five brain networks: auditory network, left frontoparietal network, right frontoparietal network, ventral attention network, and sensorimotor network. Based on network-to-network functional connectivity analysis using these nine functional brain networks, after strict FDR correction, only inter-network coupling strength between the auditory network and sensorimotor network showed significant group differences ( T = 7.548, FDR-corrected p = 0.00137). Four recurring functional connectivity states were identified through dFNC analysis. Across all participants, State 2 and State 3 accounted for 38 and 23% of total time, respectively. PIGD patients tended to spend more time in these two low-connectivity states. In terms of temporal properties, only the occurrence frequency of state 4 showed significant differences among the three groups ( F = 4.855, FDR-corrected p = 0.047). Post-hoc test results showed that HC groups were significantly higher than PIGD groups and presented a trend higher than TD groups, while there was no statistically significant difference between the two PD subtypes. One-way analysis of variance was performed on whole-brain node-to-node functional connectivity and graph theoretic attributes of the three groups; no significant differences were found after FDR correction. Conclusion Compared with the HC group, PD patients (both TD and PIGD groups) exhibited spatial distribution abnormalities in multiple brain networks. Nominally significant alterations suggestive of compensatory activation of the motor cortex were observed at the uncorrected statistical level, alongside weakened functional connectivity in specific networks.