Gait-cycle-locked cortico-subthalamic signatures underlying freezing of gait and their modulation in Parkinson's disease.

OBJECTIVE: Research on freezing of gait (FOG) in Parkinson's disease (PD) has identified relevant electrophysiological markers. However, their brief temporal windows limit their utility for individualized deep brain stimulation (DBS). This study explored gait performance and neural features in freezing-susceptible walking to develop novel FOG-predictive biomarkers. METHODS: Gait kinematics and local field potentials (LFP) from the cortex and subthalamic nucleus (STN) were simultaneously acquired in FOG patients during walking. Using the gait cycle as the analytic unit, we compared freezing trials (FOGT) and non-freezing trials (nFOGT) under the no-intervention condition (OFF) to identify gait parameters and neural features associated with FOG risk. Subsequently, we assessed the modulatory effects of high-frequency (HFS) and low-frequency (LFS) STN-DBS on abnormal gait and cortical power. Finally, we analyzed changes in abnormal gait and cortico-STN coherence after levodopa administration. RESULTS: FOGT showed aberrant gait parameters compared to nFOGT, along with disrupted lowbeta oscillations in primary somatosensory cortex (S1) and superior parietal lobule (SPL). Both HFS and LFS mitigated gait impairment and freezing severity, with LFS exerting broader effects: HFS reversed pathological lowbeta power reduction in SPL during the double support phase, while LFS restored phase-dependent oscillations between the stance phase and swing phase in S1. Additionally, abnormal theta coherence between S1 and STN could be modulated by levodopa, accompanied by gait recovery. CONCLUSION: This study identifies gait-cycle-locked cortico-STN signatures for FOG, which have extended temporal windows and are modulable by DBS, suggesting the gait cycle as a promising intervention target.