PIGDAssess: Wearable Dual-Task Sensing for Self-Administered PIGD Assessment in Parkinson's Disease
Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies · 15.06.2026 · 10 (2) · 1-31
DOI: 10.1145/3810193 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, online erschienen am 15.06.2026. Verfasst von Yizhen Zhang und 6 weiteren Autor:innen. Die Arbeit nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Postural instability and gait difficulty (PIGD) are leading causes of falls in Parkinson's disease (PD), yet current clinical assessment is infrequent, subjective, and requires expert supervision. Such assessments cannot be safely performed at home. We introduce PIGDAssess, the first wearable system that enables fully self-administered at-home estimation of all four UPDRS-PIGD subitems, including postural stability, without clinician involvement. Patients wear three commodity IMUs (lumbar and both feet) and perform brief sit-to-stand, standing, and walking tasks under both single-task and dual-task (serial–3 subtraction) conditions. To robustly infer clinically meaningful scores from dual-task movement segments, we introduce three key modules: (1) a paired single/dual-task encoder that captures balance-sensitive signatures; (2) a multi-action, multi-task fusion module that jointly predicts all four PIGD subscores from one short protocol; and (3) a prior-guided adaptation stage that reduces subject-to-subject variability while preserving the ordinal 0-3 clinical severity structure. We evaluate PIGDAssess on data from 35 individuals with PD, which were collected in collaboration with a partner hospital in both clinical settings and patients' homes, and were annotated with clinician-scored UPDRS-PIGD labels. Averaged across all four PIGD subitems, the system achieves 0.83 accuracy, 0.87 macro-F1, and 0.25 MAE. For Postural Stability, which normally requires a hands-on pull test, PIGDAssess reaches 0.89 accuracy using only three IMUs in a self-administered protocol. These results suggest a path toward frequent, low-burden, clinician-free PD assessment at home. We further conducted structured interviews with four neurologists, supporting the clinical usefulness and deployment feasibility of PIGDAssess. We have released the first IMU dataset with item-level UPDRS-PIGD annotations to enable reproducible benchmarking and accelerate clinical translation.
Zitieren
APA
Zhang, Y., Wang, J., Xie, W., Hu, Q., Hu, H., Li, G., & Zhang, Q. (2026). PIGDAssess: Wearable Dual-Task Sensing for Self-Administered PIGD Assessment in Parkinson's Disease. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 10(2), 1-31. https://doi.org/10.1145/3810193
Vancouver
Zhang Y, Wang J, Xie W, Hu Q, Hu H, Li G, et al. PIGDAssess: Wearable Dual-Task Sensing for Self-Administered PIGD Assessment in Parkinson's Disease. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies. 2026;10(2):1-31. doi:10.1145/3810193
BibTeX
@article{Zhang2026,
author = {Zhang, Yizhen and Wang, Jinjian and Xie, Wentao and Hu, Qingyong and Hu, Haiyan and Li, Guihua and Zhang, Qian},
title = {PIGDAssess: Wearable Dual-Task Sensing for Self-Administered PIGD Assessment in Parkinson's Disease},
journal = {Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies},
year = {2026},
volume = {10},
number = {2},
pages = {1--31},
doi = {10.1145/3810193}
}
EUVIMED