1114 Evidence for Deep Sleep Homeostasis in Real World Sleep Tracking Data

Introduction Investigations of sleep homeostasis often involve tightly controlled experimental sleep deprivation in service of understanding mechanistic physiology. The extent to which the deep sleep response to recent sleep loss occurs in naturalistic settings remains under-studied. We tested the hypothesis that a homeostatic increase in deep sleep occurs on the first recovery night following occasional short duration nights that arise in naturalistic settings. Methods We analyzed sleep staging data in participants who provided informed consent to participate in the Apple Heart and Movement Study and elected to contribute sleep data. The analysis group included n=33,358 participants with at least 30 nights of sleep staging data from Apple Watch. Results Short nights of sleep that were >=2 hours shorter than each participant’s median sleep duration occurred at least once in 65.1% of the cohort, most often in isolation (< 5% of instances were consecutive short nights). Assessing the amount of deep sleep on the recovery night revealed that deep sleep increased in proportion to the amount of sleep loss on the preceding short night, for short night definitions ranging from 30 minutes to >3 hours below the within-participant median sleep duration. Focusing on short nights that were at least 2 hours below median, we found that 64.9% of participants showed any increase in deep sleep on the recovery nights, with a median increase of 23.9% (absolute increase: median 11 minutes, interquartile range 5 to 19). Among participants with at least three short nights at any time in a 6-month window (28.6%), the percentage of recovery nights having >=5 minutes extra deep sleep showed small but significant Spearman correlations with two measures of sleep consistency, the standard deviation of sleep duration (r = -0.16), and the standard deviation of sleep start time (r = -0.12). Conclusion The results provide evidence for homeostatic responses in a naturalistic setting. Although the deep sleep recovery amounts are modest, naturalistic short nights are a lesser perturbation compared to experimental conditions of greater restriction or total deprivation. The findings illustrate the utility of longitudinal sleep tracking to assess real-world correlates of sleep phenomenology established in controlled experimental settings. Support (if any) n/a