Tracking Parkinson’s Disease with transformer models of everyday looking behaviors

Project Abstract

It is more common nowadays for people to have their own wearable devices to measure physiological signals like heart rate and respiration to keep track of physical diseases. However, monitoring decline in cognitive functions or development of neurodegenerative diseases, such as Parkinson’s (PD), is still complex and tricky. Traditional approaches require patients visit clinics, and hence, continuous assessments on degradation in cognitive functions is not plausible. The patient population of neurodegenerative disease is keep increasing. Therefore, it is necessary to build new methods with which people can easily measure their own cognitive or neurodegenerative disorders in daily life. One possible approach is assessing abnormalities of vision. Degradation of visual function is frequently observed in neurodegenerative disease patients. For instance, PD patients are known to have problems for perceiving colors, objects, and motion. We hypothesize that from behaviors of looking during the daily activities, like walking or doing house chores, we can detect abnormalities in neurodegenerative patient groups using machine learning algorithms. The project aims to build a foundation for minimally-invasive and sensitive measures for diagnosing and monitoring neurodegenerative diseases. We expect that the project output will make people easily track their own health status of cognitive function in the future.

Project Details

Funding Type:

Interdisciplinary Scholar Award

Award Year:


Lead Researcher(s):

Team Members:

Justin Gardner (Faculty Sponsor, Psychology)