Abstract
Sleep plays a critical role in stroke recovery. However, there are limited practices to measure sleep for individuals with stroke, thus inhibiting our ability to identify and treat poor sleep quality. Wireless, body-worn sensors offer a solution for continuous sleep monitoring. In this study, we explored the feasibility of (1) collecting overnight biophysical data from patients with subacute stroke using a simple sensor system and (2) constructing machine-learned algorithms to detect sleep stages. Ten individuals with stroke in an inpatient rehabilitation hospital wore two wireless sensors during a single night of sleep. Polysomnography served as ground truth to classify different sleep stages. A population model, trained on data from multiple patients and tested on data from a separate patient, performed poorly for this limited sample. Personal models trained on data from one patient and tested on separate data from the same patient demonstrated markedly improved performance over population models and research-grade wearable devices to detect sleep/wake. Ultimately, the heterogeneity of biophysical signals after stroke may present a challenge in building generalizable population models. Personal models offer a provisional method to capture high-resolution sleep metrics from simple wearable sensors by leveraging a single night of polysomnography data.
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
