Parkinson's disease (PD) is a prevalent neurodegenerative disorder characterized by both motor and non-motor symptoms. Subthalamic nucleus (STN) deep brain stimulation (DBS) effectively alleviates motor symptoms; however, non-motor symptoms such as sleep disorders significantly impair patients' quality of life. While preliminary evidence suggests DBS may improve sleep, the underlying neural mechanisms and long-term effects on sleep regulation remain poorly elucidated. This study employs a wearable polysomnography (PSG) device to investigate how STN-DBS modulates sleep architecture, local field potentials (LFPs), and clinical outcomes in PD patients. We hypothesize that STN-DBS contributes to both the regulation and disruption of normal sleep behavior. By synchronizing PSG with STN-LFP recordings under DBS-ON and DBS-OFF conditions, we aim to decode sleep stages using STN-LFPs, characterize PD-specific sleep abnormalities (e.g., sleep architecture, atonia), and analyze DBS-induced LFP changes and their correlations with clinical efficacy. According to the above data, the sleep stage characteristics were decoded, and the algorithm was used to determine the optimal clinical threshold current amplitude to improve sleep. Then, the patients will be given adaptive stimulation, and the sleep quality was monitored to verify its efficacy.This study will provide a neurophysiological foundation for developing closed-loop stimulation strategies targeting sleep dysfunction in PD.