Abstract Introduction Accurate inpatient sleep measurement is challenging: polysomnography (PSG), the gold standard, is impractical for routine use due to cost, complexity, and poor tolerability. Actigraphy, while feasible, tends to overestimate sleep in patients confined to bed. EEG-based wearable devices may provide objective data with minimal disruption, but their feasibility and tolerability in general inpatient settings remain unclear. We hypothesized that patients with pre-existing sleep disorders would be less likely to tolerate an EEG wearable device. Methods We conducted a feasibility study using an EEG wearable headband (Sleep Profiler, Advanced Brain Monitoring) in inpatient adults. Participants wore the device overnight to record EEG, electrooculogram, and electromyogram signals for automated sleep staging. Tolerance was defined as wearing the device for the entire night. Baseline sleep measures included the Insomnia Severity Index (ISI), STOP-Bang, and baseline Epworth Sleepiness Scale (ESS). We compared sleep outcomes between those who tolerated and those who did not using t-tests and two-proportion z-tests. Results Of 140 patients approached, 61 consented and 35 (57%) successfully wore the device overnight. Common reasons for intolerance included discomfort (36%) and headband displacement (21%). ISI scores were similar between tolerant and intolerant groups (11.79 ± 7.30 vs. 11.88 ± 7.56, p = 0.96), as were ESS scores (8.0 ± 4.95 vs. 8.73 ± 4.98, p = 0.58) and STOP-Bang scores (3.04 vs. 3.58, p = 0.20). Sleep metrics did not differ significantly, suggesting the device did not disrupt sleep. Sleep metrics were also comparable: total sleep time (347 ± 172 vs. 420 ± 185 min, p = 0.23), sleep onset latency (28 ± 28 vs. 60 ± 100 min, p = 0.10), and wake after sleep onset (50 vs. 40, p = 0.50). Conclusion EEG-based wearable devices were tolerated by over half of participants, and tolerance was not associated with pre-existing sleep disorders. Although differences in sleep measures were not statistically significant, some gaps (e.g., 70 minutes in total sleep time) could be clinically meaningful. Further research should determine whether these differences impact patient recovery or comfort and explore strategies to improve device wearability. Support (if any) This work was supported by the AASM Foundation Physician Scientist Training Grant.
Mathukumilli et al. (2026) studied this question.