Network localization of regional homogeneity alterations in Parkinson's disease.

BACKGROUND

Resting-state functional MRI (rs-fMRI) studies using regional homogeneity (ReHo) have identified localized functional changes in Parkinson's disease (PD), but findings across studies exhibit considerable heterogeneity. The emerging network perspective suggests these disparate findings might reflect nodes within a single interconnected network. Functional Connectivity Network Mapping (FCNM) offers an approach to test this hypothesis.

METHODS

We conducted a systematic literature search (PubMed, Embase, Web of Science, CNKI, and Wanfang) for studies reporting whole-brain ReHo differences (PD vs. healthy controls). Resting-state fMRI data from the Human Connectome Project (HCP;  = 1,093) were analyzed using FCNM to map ReHo abnormalities in PD onto common functional brain networks. Robustness was assessed using 1 mm and 7 mm radii, and spatial overlap with canonical brain networks was quantified.

RESULTS

A total of 52 studies, comprising 72 datasets reporting ReHo differences between 2,052 PD patients and 1,401 healthy controls, were included in the analysis. The FCNM analysis identified a distributed PD-associated dysfunctional network. This network showed significant spatial overlap primarily with the visual (49.24%), somatomotor (32.35%), dorsal attention (44.49%), and ventral attention (67.97%) canonical networks. The network topography demonstrated high consistency across different seed radii (1 mm and 7 mm), confirming robustness.

CONCLUSION

By integrating heterogeneous ReHo findings via FCNM, this study delineates robust PD-associated dysfunctional networks involving key sensory, motor, and attentional systems. This network-centric view offers a unifying perspective on PD pathophysiology, highlighting large-scale systems disruption and potentially reconciling previous localization inconsistencies. This approach underscores the value of network neuroscience for understanding PD mechanisms.