Long-term at-home sleep monitoring in late-onset epilepsy using wearable EEG: first insights
No Thumbnail Available
Author
Yilmaz, Gürkan
Hauser, Ramona
Martin, Noemí
Braun, Fabian
Lemay, Mathieu
Schindler, Kaspar
Seiler, Andrea
Jorge, João
DOI
10.1515/bmt-2025-1001
Abstract
Methods The present cohort includes 10 LOE patients (age 52–76, 3F) recruited at the Sleep-Wake-Epilepsy center, Bern University Hospital. Each patient received an ULTEEMNite headband (single-channel forehead-EEG, dry active electrodes, 256-Hz sampling) and docking station to recharge and automatically upload recordings to a secure server; the patients were taught how to use the device autonomously at home. Subsequently, each received recording was transformed to a spectrogram (30s-windows), and reviewed to identify segments of sleep EEG with acceptable quality, defined by: (i) RMS in 0.5–32.5 Hz band below 40 µV, and (ii) absence of a characteristic harmonic pattern induced when skin contact is lost.
Results We have currently received 14–28 recorded nights per patient, totalling 237 nights. In 7 patients, the recordings were of consistently high quality, yielding 6.6±1.1 to 7.8±1.9 hours of acceptable quality data per night, across patients – consistent with average human sleep duration in this age range (~7 hours). Sleep stage patterns could be clearly identified in the spectrograms. In contrast, despite comparable recording lengths, the other 3 patients exhibited frequent periods of elevated broadband power and poor skin contact – yielding 2.4±2.1 to 5.3±0.9 hours of acceptable data/night. This showed no relation to age.
Conclusion The current data indicate very favorable feasibility prospects for long-term, autonomous sleep EEG monitoring in LOE patients, with a predominance of consistently high-quality night recordings. We are investigating the less efficacious cases to identify opportunities for further improvement.
Publication Reference
BMT 2025, Muttenz (Switzerland)
Year
2025-09-11