O-EG004. Comparing the accuracy of sleep staging data from a new wearable sleep electroencephalography device versus the Fitbit Charge 3 with polysomnography as reference

Abstract

Introduction . Although sleep disorders are a common social problem, polysomnography (PSG), the gold standard in sleep diagnostics, can be burdensome and may interfere with natural sleep. We have developed a high-precision wearable electroencephalography (EEG) device for objectively measuring sleep at home. Actigraphy devices with the capability of measuring heart rate variability can also be used to evaluate sleep stages at home (e.g., Fitbit Charge 3). However, the accuracy of these inhome devices has not yet been compared. We hypothesized that the accuracy of our wearable sleep EEG device would be higher than that of the Fitbit Charge 3. Methods . We simultaneously recorded PSG, wearable sleep EEG, and data from the Fitbit Charge 3 in 42 (mean ± SD: 23.5 ± 7.8 years, 14 women) healthy participants. A registered polysomnographic technologist scored the PSG and wearable sleep EEG data, and we evaluated the sleep stages captured by the Fitbit Charge 3 using the Fitbit automatic sleep staging algorithm (stage W, N1+N2, N3, and R). We measured and compared the accuracy of the wearable EEG device and Fitbit Charge 3 data with respect to the PSG data. Results . The accuracy of the wearable sleep EEG device and Fitbit Charge 3 was 90.7 ± 2.8% and 69.5 ± 8.3 %, respectively (mean ± SD). The wearable sleep EEG device had significantly higher accuracy than the Fitbit (p < 0.001). Conclusion . Compared with the Fitbit Charge 3 data, data from the wearable sleep EEG device more closely mirrored the PSG data. Thus, the wearable sleep EEG device may be superior to the Fitbit Charge 3 as an alternative to PSG. Our wearable sleep EEG device has potential for use in sleep evaluations for multiple consecutive nights in an in-home setting.