Multimodal Assessment

The high-dimensional, time-series nature of EEG data contains patterns that human researchers cannot perceive directly but that AI systems can reliably detect. The study combines chronic LFP sensing with EMG, EEG, motor inhibition testing, wearable monitoring, and patient-reported outcomes. Surface EMG is used as a core objective measure in the study. EMG is recorded simultaneously with LFP streaming during standardized motor tasks. The optional STAT-ON wearable sub-study records gait and bradykinesia before implantation and at 12 months. EEG is recorded over the motor strip at predefined visits to characterize cortical oscillations. EEG voltage signals are well-suited to machine learning analysis. Approximately four seconds of EEG brain signal was sufficient for the AI model to perform accurate classification.