Heterogeneous digital biomarker integration out-performs patient self-reports in predicting Parkinson\u2019s disease

Kaiwen Deng,Yueming Li,Roger L. Albin,Yuanfang Guan,Jian Wang,Hanrui Zhang

doi:10.1038/s42003-022-03002-x

Abstract

Parkinson’s disease (PD) is one of the first diseases where digital biomarkers demonstrated excellent performance in differentiating disease from healthy individuals. However, no study has systematically compared and leveraged multiple types of digital biomarkers to predict PD. Particularly, machine learning works on the fine-motor skills of PD are limited. Here, we developed deep learning methods that achieved an AUC (Area Under the receiver operator characteristic Curve) of 0.933 in identifying PD patients on 6418 individuals using 75048 tapping accelerometer and position records. Performance of tapping is superior to gait/rest and voice-based models obtained from the same benchmark population. Assembling the three models achieved a higher AUC of 0.944. Notably, the models not only correlated strongly to, but also performed better than patient self-reported symptom scores in diagnosing PD. This study demonstrates the complementary predictive power of tapping, gait/rest and voice data and establishes integrative deep learning-based models for identifying PD.

Highlights

Parkinson’s disease (PD) is one of the first diseases where digital biomarkers demonstrated excellent performance in differentiating disease from healthy individuals
The goal of this study is to explore and integrate digital biomarkers of movement beyond the gross motor skills captured by gait and rest evaluations
In this study, we proved the ability of finger-tapping positions on identifying the person with Parkinson (PwP) and the reported performances, an average Area Under Receiver Operating Characteristic Curve (AUC) of 0.935, support that digital biomarkers for diagnosing PD can be developed beyond gross motor skills such as gait/rest, the primary focus of prior literature[33,34,35]

Summary

Introduction

Parkinson’s disease (PD) is one of the first diseases where digital biomarkers demonstrated excellent performance in differentiating disease from healthy individuals. Performance of tapping is superior to gait/rest and voice-based models obtained from the same benchmark population. This study demonstrates the complementary predictive power of tapping, gait/ rest and voice data and establishes integrative deep learning-based models for identifying PD. A community-based challenge benchmarked algorithms using 30-s rest and gait data from cell phones to differentiate self-reported PwP from healthy subjects[14,15]. Despite a considerable number of studies using digital biomarkers for PwP detection, prior studies on the movement data have mainly focused on the evaluation of gross motor skills such as walking and rest with medium sample sizes[4]. There have been few digital biomarker studies focusing on fine-motor skills, and the applications of state-of-the-field machine learning techniques such as deep learning approaches only report a moderate performance[13]. There are no fair comparisons or integration of algorithms across different types of motor assessments aimed at differentiating PwP from healthy individuals

Objectives

Methods

Results

Conclusion