Generalizing Parkinson’s disease detection using keystroke dynamics: a self-supervised approach
Journal of the American Medical Informatics Association · 20.05.2024 · 31 (6) · 1239-1246
DOI: 10.1093/jamia/ocae050 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of the American Medical Informatics Association, online erschienen am 18.03.2024. Verfasst von Shikha Tripathi und 4 weiteren Autor:innen. Beteiligte Einrichtung: D. Bradley McWilliams School of Biomedical Informatics, University of Texas Health Science Center at Houston, Houston, TX 77030, United States. Hauptthemen laut MEDLINE: Parkinson Disease, Supervised Machine Learning. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 21 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NIH HHS, NINDS NIH HHS, NIBIB NIH HHS.
Abstract (Originalsprache)
Objective
Passive monitoring of touchscreen interactions generates keystroke dynamic signals that can be used to detect and track neurological conditions such as Parkinson’s disease (PD) and psychomotor impairment with minimal burden on the user. However, this typically requires datasets with clinically confirmed labels collected in standardized environments, which is challenging, especially for a large subject pool. This study validates the efficacy of a self-supervised learning method in reducing the reliance on labels and evaluates its generalizability. Materials and Methods We propose a new type of self-supervised loss combining Barlow Twins loss, which attempts to create similar feature representations with reduced feature redundancy for samples coming from the same subject, and a Dissimilarity loss, which promotes uncorrelated features for samples generated by different subjects. An encoder is first pre-trained using this loss on unlabeled data from an uncontrolled setting, then fine-tuned with clinically validated data. Our experiments test the model generalizability with controls and subjects with PD on 2 independent datasets.
Results
Our approach showed better generalization compared to previous methods, including a feature engineering strategy, a deep learning model pre-trained on Parkinsonian signs, and a traditional supervised model.
Discussion
The absence of standardized data acquisition protocols and the limited availability of annotated datasets compromise the generalizability of supervised models. In these contexts, self-supervised models offer the advantage of learning more robust patterns from the data, bypassing the need for ground truth labels.
Conclusion
This approach has the potential to accelerate the clinical validation of touchscreen typing software for neurodegenerative diseases.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- D. Bradley McWilliams School of Biomedical Informatics, University of Texas Health Science Center at Houston, Houston, TX 77030, United States.
- nQ Medical, MA 02142, United States.
ORCID-Profile: Luca Giancardo
Förderung
- NIH HHS
- NINDS NIH HHS
- NIBIB NIH HHS
- NIA NIH HHS
- nQ-Medical Inc.
- NIH
Zitieren
APA
Tripathi, S., Acien, A., Holmes, A. A., Arroyo-Gallego, T., & Giancardo, L. (2024). Generalizing Parkinson’s disease detection using keystroke dynamics: a self-supervised approach. Journal of the American Medical Informatics Association, 31(6), 1239-1246. https://doi.org/10.1093/jamia/ocae050
Vancouver
Tripathi S, Acien A, Holmes AA, Arroyo-Gallego T, Giancardo L. Generalizing Parkinson’s disease detection using keystroke dynamics: a self-supervised approach. Journal of the American Medical Informatics Association. 2024;31(6):1239-1246. doi:10.1093/jamia/ocae050
BibTeX
@article{Tripathi2024,
author = {Tripathi, Shikha and Acien, Alejandro and Holmes, Ashley A and Arroyo-Gallego, Teresa and Giancardo, Luca},
title = {Generalizing Parkinson’s disease detection using keystroke dynamics: a self-supervised approach},
journal = {Journal of the American Medical Informatics Association},
year = {2024},
volume = {31},
number = {6},
pages = {1239--1246},
doi = {10.1093/jamia/ocae050}
}
EUVIMED