Catalyst Awardee

Project Description

A multidomain biometric sensing platform that detects asymptomatic prodromal phase of Parkinson’s disease

Chin-Hsien Lin, MD, PhD; Li-Chen Fu, PhD; Jyh-Shing Roger Jang, PhD

Competition Sponsor: Academia Sinica

Parkinson’s disease (PD) is one of the most common neurodegenerative disorders with an estimated number will be nearly 9 million in the year 2030.The symptoms of PD include resting tremor, bradykinesia, rigidity and gait postural instability. There is currently no cure for PD patients. Given the likely entry of mechanism-targeted therapies into early human clinical trials, early detection of PD, especially in the asymptomatic early stage of the disease is crucial.
There are several clues of biometric soft signs that can be detected before the occurrence of the classical motor symptoms of PD, including decreased facial expression, linguistic changes, and changes of gait pattens. Although there are several apps that have been developed in the smartphone or smartwatch, these apps mainly focus on detection of the arm swing movement pattern while wearing the smartwatch or carrying the smartphone. This single-domain modality may not be PD-specific as patients with stroke or arthritis may manifest with decreased arm swinging. An integrated platform with multidomain biometric features with automatic analysis are more sensitive than single modality feature to precisely and accurately detect the prodromal phase of PD. Hence, in this study, we aim to establish a deep learning-based multidomain biometric sensing platform that the combine the biometric physical features, including speech, facial expression, gait and associated movements to detects asymptomatic prodromal phase of susceptible PD patients.
Our results will facilitate future large population screening of PD for early preclinical diagnosis and proper treatment, decreasing medical burden in an aging society.

Sign up for updates