BIOCLITE: Smartwatch Dataset for Parkinson's Disease in Supervised and Unsupervised Contexts

The dataset collected during the experimental campaign consists of triaxial acceleration and triaxial gyroscope data recorded from a smartwatch, gathered from 24 patients with Parkinson’s disease and 16 healthy participants. Data was recorded at a sampling rate of 50 Hz and expressed in m/s² and rad/s.

Each participant performed 8 different exercises in both supervised and unsupervised contexts. The complete set of exercises was performed once per day, typically at a consistent time. Each participant began with an initial supervised session, followed by seven consecutive days of unsupervised sessions, and concluded with a final supervised session.

Check the " Bioclite_README.txt" file for specific details about the dataset structure.

If you use this dataset, please cite:

Polvorinos-Fernández, C.; Sigcha, L.; Valero, M.M.; Grande, M.; de Arcas, G.; Pavón, I. Machine Learning Assessment of Parkinson’s Disease Using a Novel Free-Living Egg-Beating Motor Task. Technologies 2026, 14, 345. https://doi.org/10.3390/technologies14060345 

Polvorinos-Fernández, C., Sigcha, L., Centeno-Cerrato, M., de Arcas, G., Grande, M., Marín, M., Pareés, I., Martínez-Castrillo, J.C., Pavón, I. “Evaluation of Free-Living Motor Symptoms in Patients With Parkinson Disease Through Smartwatches: Protocol for Defining Digital Biomarkers”. JMIR Res Protoc 2025, 14, e72820, https://doi.org/10.2196/72820 

Sigcha, L., Polvorinos-Fernández, C., Costa, N., Costa, S., Arezes, P., Gago, M., ... & Pavón, I. "Monipar: Movement data collection tool to monitor motor symptoms in Parkinson's disease using smartwatches and smartphones". Frontiers in Neurology, 14, 1326640. https://doi.org/10.3389/fneur.2023.1326640