Portal de Periódicos da CAPES

Crowdsourcing digital health measures to predict Parkinson’s disease severity: the Parkinson’s Disease Digital Biomarker DREAM Challenge

2021; Nature Portfolio; Volume: 4; Issue: 1 Linguagem: Inglês

10.1038/s41746-021-00414-7

2398-6352

Solveig K. Sieberts, Jennifer Schaff, Marlena Duda, Bálint Pataki, Ming Sun, Phil Snyder, Jean‐François Daneault, Federico Parisi, Gianluca Costante, Udi Rubin, Peter Banda, Yooree Chae, Elias Chaibub Neto, E. Ray Dorsey, Zafer Aydın, Aipeng Chen, Laura L. Elo, Carlos Lodeiro, Enrico Glaab, Ethan Goan, Fatemeh Noushin Golabchi, Yasin Görmez, Maria K. Jaakkola, Jitendra Jonnagaddala, Riku Klén, Dongmei Li, Christian McDaniel, Dimitri Perrin, Thanneer M. Perumal, Nastaran Mohammadian Rad, Erin Rainaldi, Stefano Sapienza, Patrick Schwab, Nikolai Shokhirev, Mikko S. Venäläinen, Gloria Vergara‐Diaz, Yuqian Zhang, Avner Abrami, Aditya Adhikary, Carla Agurto, Sherry Bhalla, Halil Bilgin, Vittorio Caggiano, Jun Cheng, Eden Deng, Qiwei Gan, Rajan Girsa, Zhi Han, Stephen Heisig, Kun Huang, Samad Jahandideh, Wolfgang Kopp, Christoph Kurz, Gregor Lichtner, Raquel Norel, Gajendra P. S. Raghava, Tavpritesh Sethi, Nicholas Shawen, Vaibhav Tripathi, Matthew Tsai, Tongxin Wang, Yi Wu, Jie Zhang, Xinyu Zhang, Yuanjia Wang, Yuanfang Guan, Daniela Brunner, Paolo Bonato, Lara M. Mangravite, Larsson Omberg,

Voice and Speech Disorders

Abstract Consumer wearables and sensors are a rich source of data about patients’ daily disease and symptom burden, particularly in the case of movement disorders like Parkinson’s disease (PD). However, interpreting these complex data into so-called digital biomarkers requires complicated analytical approaches, and validating these biomarkers requires sufficient data and unbiased evaluation methods. Here we describe the use of crowdsourcing to specifically evaluate and benchmark features derived from accelerometer and gyroscope data in two different datasets to predict the presence of PD and severity of three PD symptoms: tremor, dyskinesia, and bradykinesia. Forty teams from around the world submitted features, and achieved drastically improved predictive performance for PD status (best AUROC = 0.87), as well as tremor- (best AUPR = 0.75), dyskinesia- (best AUPR = 0.48) and bradykinesia-severity (best AUPR = 0.95).