Investigator:
Yuanfang Guan, PhD
Name of Institution:
University of Michigan, Ann Arbor, Michigan
Project Title:
Digital Biomarkers for Monitoring Parkinson’s Disease (PD)
Investigator Bio:
Dr. Guan’s PhD and postdoctoral training were in the fields of functional genomics and machine learning at Princeton University. After she started a faculty position at the University of Michigan, she became increasingly interested in developing machine learning methodologies to mine massive amounts of data. She has become particularly intrigued by the power of deep learning, which is capable of identifying intricate patterns in data that cannot be easily captured through classic statistics.
Objective:
This research will focus on developing models using voice records on a mobile phone to detect PD in preclinical populations.
Background:
Early diagnosis and close monitoring are the keys to effective management of PD. However, the current diagnosis of PD can be subjective and monitoring can be insufficient, as both rely on clinical manifestation of symptoms at infrequent clinical visits and unreliable patient recall. Recent popularization of mobile devices provides an unprecedented opportunity to allow at-home day-to-day monitoring of PD conditions, by analyzing the digital biomarkers of PD.
Methods/Design:
This study will generate a state-of-the-art algorithm for using voice information generated from mobile devices to predict and follow PD. Diagnosis by a movement disorders physician will be the gold standard to which the deep learning model will be compared.
Relevance to Diagnosis/Treatment of Parkinson’s disease:
With the growing popularity of mobile apps, we envision this model will become an indispensable tool to diagnose PD and follow progression.