We are looking for an enthusiastic Master student in our AI for Parkinson lab. The focus of this lab is the development of AI-based remote monitoring tools for people with Parkinson’s disease (PWP), to improve quality of care and to make clinical trial more efficient. The AI for Parkinson lab is a collaboration between Radboudumc, Radboud University and Verily. The lab is embedded in the Centre of Expertise for Parkinson’s and Movement Disorders at Radboudumc.
One of the research topics in the lab is investigating autonomic dysfunction in Parkinson's disease (PD) with a particular focus on utilizing photoplethysmography (PPG) data for heart rate (HR) analysis. PD is not solely characterized by motor symptoms such as tremor, rigidity, and bradykinesia; it also often involves autonomic nervous system abnormalities (caused by autonomic dysfunction), which significantly affect quality of life. Heart rate (HR) patterns reflect autonomic functioning, and could be a viable way to assess autonomic dysfunction in daily life.
Photoplethysmography (PPG) sensors allow the continuous monitoring of HR patterns to capture real-time autonomic changes. By systematically collecting and analyzing PPG data from PWP and using AI-based methods to extract HR-related features, we seek to identify and quantify autonomic dysregulation. A major challenge for using PPG in an ambulatory setting is the sensitivity of PPG signals to noise and motion artifacts. In this project, our primary objective is to develop a deep learning-based algorithm capable of accurately defining heart rate (HR) parameters, including heart rate variability (HRV), by leveraging PPG and accelerometer data from wearable devices. We will validate the accuracy of this algorithm by comparing its results to the gold standard of holter electrocardiography (ECG) measurements. The successful implementation of such a deep learning algorithm could significantly improve clinical trials by providing a more objective and patient-centric method compared to the current clinical rating scales.
Contact: Twan van Laarhoven and Kars Veldkamp