Single-subject gray matter networks in patients with early- to late-stage Parkinson's disease.

Although recent advances in network analysis have enabled individual assessment of brain networks using gray matter images, only a few studies have used gray matter networks in Parkinson's disease (PD) and were limited to a few patients with early-stage PD. We aimed to investigate gray matter network alterations and their correlation with clinical parameters, including disease severity and nigrostriatal degeneration, in patients with early- to late-stage PD. We recruited 258 patients with PD and 406 age- and sex-matched healthy controls (HC). The gray matter networks were constructed from three-dimensional T1-weighted MRI. Global and local network properties were analyzed using a graph-theoretical approach, and their correlations with disease severity, motor disability, and nigrostriatal degeneration were assessed. Globally, patients with PD exhibited decreased small-world properties compared to HC. The small-world properties decreased with increasing modified Hoehn and Yahr stage and negatively correlated with the Movement Disorder Society-sponsored revision of the Unified Parkinson's Disease Rating Scale Part III Score in patients with PD. Furthermore, they were positively correlated with the specific binding ratio derived from dopamine transporter imaging. Locally, patients with PD show regional network abnormalities in the putamen, thalamus, brainstem, precuneus, and anterior/posterior cingulate gyri, which are primarily involved in PD. Our findings suggest that the gray matter network in patients with PD moved toward a less optimal random network than that of HC as the disease progressed. Single-subject gray matter networks may enhance understanding of the pathophysiology of PD.