Patterns (Jun 2020)
Deep Learning Identifies Digital Biomarkers for Self-Reported Parkinson's Disease
Abstract
Summary: Large-scale population screening and in-home monitoring for patients with Parkinson's disease (PD) has so far been mainly carried out by traditional healthcare methods and systems. Development of mobile health may provide an independent, future method to detect PD. Current PD detection algorithms will benefit from better generalizability with data collected in real-world situations. In this paper, we report the top-performing smartphone-based method in the recent DREAM Parkinson's Disease Digital Biomarker Challenge for digital diagnosis of PD. Utilizing real-world accelerometer records, this approach differentiated PD from control subjects with an area under the receiver-operating characteristic curve of 0.87 by 3D augmentation of accelerometer records, a significant improvement over other state-of-the-art methods. This study paves the way for future at-home screening of PD and other neurodegenerative conditions affecting movement. The Bigger Picture: In-home digital surveillance has been proposed as the future for chronic, neurodegenerative disease such as Parkinson's disease (PD), which can be monitored by wearable devices from its motor-related symptoms. However, the disparities between uncontrolled in-home environments have introduced obstacles to the population-level application of digital screening of PD. In this study, we developed the first-place solution in the recent DREAM Parkinson's Disease Digital Biomarker Challenge, which calls for optimal algorithms to extract digital biomarkers of PD from crowd-sourced movement records. To combat the spatial and temporal bias in different movement records, we applied a variety of data-augmentation methods, which significantly improves the performance of the deep-learning model. Besides PD, our method provides a path for large-scale population screening and in-home monitoring using wearable devices in other related neurodegenerative disorders.
