Neuromodulation-induced normalization of cortical metastable dynamics signatures in Parkinson's disease

Deep brain stimulation (DBS) of the subthalamic nucleus (STN) alleviates motor symptoms in Parkinson's disease (PD), but how it modulates whole-brain dynamics to drive therapeutic effects remains unclear. We hypothesized that STN-DBS restores signatures of cortical metastability-a dynamic balance between integration and segregation of neural networks-critical for adaptive behavior. Using a novel Weighted Eigenvector Dynamics Analysis (WEiDA) on resting-state fMRI data from PD patients, we identified four recurring probabilistic metastable substates (PMS). PD patients exhibited aberrantly low occupancy of a state characterized by decoupling between visual (VN) and somatomotor/ventral attention (SMN/VAN) networks (State V-VA/SM), which normalized with DBS and correlated with motor improvement. This rebalancing was mediated by decreased local metastability and enhanced functional segregation between VN and SMN/VAN modules, linked to cholinergic gene expression (CHRNA10). Notably, analogous normalization of metastable dynamics and symptom relief emerged during noninvasive STN-targeted temporal interference stimulation (tTIS), indicating converged network-level mechanisms across neuromodulation modalities. Our findings highlight cortical metastability signatures as a potential substrate for DBS efficacy in PD, bridging invasive and noninvasive therapeutic strategies through dynamic brain state modulation.