Article Text
PDF
Abstract
Background Sleep is affected by Huntington’s disease (HD), and sleep quality is linked to neurodegeneration. Wearable devices, such as Fitbit, use inertial sensors and photoplethysmography to estimate sleep stages during sleep, providing a low-cost solution for home-based sleep monitoring. While Fitbit devices have been validated in healthy individuals against the gold standard, polysomnography (PSG), they have not yet been validated in HD.
Aims To establish the accuracy of Fitbit sleep metrics against PSG, and to quantify sleep quality over 7 nights at home.
Methods Participants wore a Fitbit Charge 4 during overnight PSG, followed by 7 nights at home. PSG sleep stages were scored by an expert sleep physiologist. Fitbit sleep data were extracted every 30 s. Time in bed (TIB), total sleep time (TST), wake time (TWT), time in REM sleep (REM), stage N3 (Deep) and stages N1 and N2 (Light), sleep onset latency (SOL) and wake after sleep onset (WASO) were examined on each night. Fitbit sensitivity and specificity to each sleep stage was examined. Home sleep metrics were compared to PSG.
Results Data for 1 male with early-stage HD (UHDRS motor score 5) are reported. Using PSG, TIB was 474.5 min, TST was 373 min, TWT was 101.5 min, REM was 96 min, Deep was 100 min, Light was 177 min, SOL was 95 min, and WASO was 6.5 min. Compared to PSG, Fitbit overestimated TST by 23 mins, Light by 37.5 min, REM by 11.5 min, WASO by 38.5 min, and underestimated SOL by 67 min, TWT by 28.5 min, Deep by 12.5 min, TIB by 5.5 min. Fitbit sensitivity and specificity to wake was 98% and 54%, REM was 97% and 100%, and Deep sleep was 96% and 74%. PSG TST and sleep stage times were within the range observed at home, but PSG TWT was greater than observed at home.
(a) Hypnogram comparison for hospital-based overnight polysomnography (PSG) study: Fitbit Charge 4 sleep stages in blue, and PSG hypnogram in green. (b) Boxplots for sleep metrics during 7 nights at home using Fitbit Charge 4. To compare the quality of home-based sleep with hospital-based sleep during PSG, each sleep metric calculated during this subject’s PSG study are also plotted (Fitbit in blue, PSG in green).
Conclusion Initial results in one participant indicate that Fitbit Charge 4 is suitable to monitor sleep stages in HD, particularly REM. Issues distinguishing wake and light sleep were observed in this case. Data collection is ongoing.