Scalp EEG Signatures of Pathological Basal Ganglia Sleep Activities in Parkinson's Disease

In this proposal, we will perform sleep recordings in individuals with Parkinson's disease already implanted with electrodes for deep brain stimulation (DBS). We will simultaneously record electrical activities from deep brain regions together with non-invasive EEG.

We will further focus on pathological activities in beta frequencies and use machine learning to train models to identify non-invasive signatures of these pathological beta activities. We predict that some beta activities, even when not identified visually in EEG, can still be identified by our algorithms.

After validating its performance, we will apply the detection tool on additional existing sleep data in patients without implanted electrodes, and test whether signatures of beta during sleep are associated with other disease markers such as the rate of symptom progression.

The proposed research has potential to promote a breakthrough by advancing a novel biomarker for early diagnosis, to track disease progression, and evaluate response to various treatments.