Inverse Reinforcement Learning for Interpretable Keystroke Biomarkers in Parkinson's Disease
arXiv:2606.25270v1 Announce Type: new Abstract: Keystroke dynamics have been explored extensively as a passive digital biomarker for Parkinson's disease (PD), typically by extracting summary statistics from typing timing and training a classifier to discriminate PD from healthy controls. We instead apply inverse reinforcement learning (IRL) to keystroke data, modeling each keystroke as a discrete choice over typing speed and recovering, per subject, an interpretable reward function that explains their observed timing behavior. To our knowledge this is the first application of IRL to keystroke
The proliferation of digital data streams (like keystrokes) and advancements in AI techniques like Inverse Reinforcement Learning are enabling new analytical approaches to health monitoring.
This research introduces a novel, interpretable method for biomarker detection in Parkinson's disease, showing how AI can move beyond simple classification to provide deeper insights into disease progression.
Traditional summary statistics are augmented by AI-driven interpretable reward functions, offering a more nuanced understanding of individual neurological states through passive data collection.
- · Neurology researchers
- · Digital health platforms
- · AI/ML developers
- · Patients with Parkinson's disease
- · Traditional biomarker extraction methods (potentially)
Individual keystroke patterns can be used to generate personalized, real-time insights into neurological conditions.
This methodology could extend to other neurological or even psychological conditions, enabling early detection and personalized interventions.
The development of a new class of 'interpretable AI biomarkers' could transform remote patient monitoring and preventative medicine.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG