Abstract
INTRODUCTION: Deep brain stimulation (DBS) is a proven effective treatment for Parkinson's disease (PD). However, titrating DBS stimulation parameters is a labourious process and requires frequent hospital visits. Additionally, its current application uses continuous high-frequency stimulation at a constant intensity, which may reduce efficacy and cause side effects. The objective of the AI-DBS study is to identify patient-specific patterns of neuronal activity that are associated with the severity of motor symptoms of PD. This information is essential for the development of advanced responsive stimulation algorithms, which may improve the efficacy of DBS. METHODS AND ANALYSIS: This longitudinal prospective observational cohort study will enrol 100 patients with PD who are bilaterally implanted with a sensing-enabled DBS system (Percept PC, Medtronic) in the subthalamic nucleus as part of standard clinical care. Local neuronal activity, specifically local field potential (LFP) signals, will be recorded during the first 6 months after DBS implantation. Correlations will be tested between spectral features of LFP data and symptom severity, which will be assessed using (1) inertial sensor data from a wearable smartwatch, (2) clinical rating scales and (3) patient diaries and analysed using conventional descriptive statistics and artificial intelligence algorithms. The primary objective is to identify patient-specific profiles of neuronal activity that are associated with the presence and severity of motor symptoms, forming a 'neuronal fingerprint'. ETHICS AND DISSEMINATION: Ethical approval was granted by the local ethics committee of the Amsterdam UMC (registration number 2022.0368). Study findings will be disseminated through scientific journals and presented at national and international conferences.