Hierarchical reorganization of cortical and cerebellar functional gradients in Parkinson's disease

Parkinson's disease (PD) is a common movement disorder with heterogeneous clinical manifestations, and its tremor-predominant presentations may overlap with essential tremor (ET), complicating clinical differentiation in some patients. Functional gradient analysis provides a continuous framework for characterizing macroscale brain hierarchical organization and may offer candidate imaging features for disease characterization.

We first investigated surface-based resting-state functional measures and cerebral perfusion in 50 patients with idiopathic PD and 33 healthy controls. We subsequently incorporated functional MRI data from 17 patients with ET to examine alterations in cortical and cerebellar functional gradients in PD and ET and to evaluate their potential utility for distinguishing the two disorders. Group comparisons, gradient range and variance analyses, network-level analyses, and multiclass support vector machine classification were performed.

Patients with PD showed widespread abnormalities in both cortical and cerebellar Gradient 1 and Gradient 2, whereas patients with ET exhibited fewer cortical alterations and no significant cerebellar gradient abnormalities. At the group level, both PD and ET showed broader cortical gradient ranges; however, only PD demonstrated significantly increased individual-level gradient range and variance. Network-level analyses further revealed pronounced hierarchical redistribution between sensorimotor and higher-order associative systems in PD, while ET displayed relatively limited alterations. The combined use of cortical and cerebellar Gradient 1 and Gradient 2 features showed discriminatory performance within the present dataset, with area under the receiver operating characteristic curve values of 1.000 for healthy controls, 0.960 for PD, and 0.889 for ET. Shapley additive explanations (SHAP) analysis indicated that both cerebral and cerebellar gradient features contributed to classification, with broader feature involvement in PD than in ET.

These findings suggest that PD is characterized by extensive disruption of cortical-cerebellar functional hierarchy, exceeding the relatively limited alterations observed in ET. Functional gradients may therefore provide candidate imaging features for characterizing PD-related hierarchical reorganization and for exploratory differentiation of PD from ET.