Spatiotemporal Organization of Sleep
We perform comprehensive spatiotemporal analyses of sleep patterns in a sample of over 100,000 individuals across Germany spanning almost 3 years. As part of the Robert Koch Institute’s Data Donation Project for the early detection of COVID-19, sleep data, in conjunction with heart rate and activity data, was collected in a privacy-preserving fashion by passive sensors in consumer wearable devices from April 2020 until December 2022, yielding more than 45 million nights of sleep observations overall. Sleep tracking studies for the first time enable a means of large-scale, real-time, non-invasive, continuous, longitudinal objective sleep measurements at a relatively low-cost with high accessibility. This combination of long observation period, and high temporal and spatial resolution is ideally suited to track sleep behavior over time, enabling the identification of robust seasonal trends as well as fine-grained analyses of short-term changes. The large number of regular participants from all over Germany allows for spatial analyses in a highly-resolved manner to discern complex patterns along geographical position or population density.
Spatiotemporal analyses of sleep behavior have great informative value for public health. Mapping sleep patterns is essential to understanding which external conditions pose risks for sleep, which ones may be protective, and how these conditions interact. The two main conditions that we focus on here are in constant conflict with each other in our modern lives: - sun time (or solar time) refers to the natural time structure on Earth based on the position of the sun in the sky, which gives rise to the timing and duration of day and night as well as the seasons. - social time (or societal time) refers to the human-constructed timing framework, shaped by policy and social forces, that determines weekly structure, work and school schedules, time zone allocation and adherence to daylight saving time.
Individual aspects of this interaction of sun and social time on sleep and circadian rhythms have been investigated in numerous studies to date. These studies range from small diary studies and detailed physiological studies in small but high-quality samples to large questionnaire-based studies, with studies based on wearable sensor data in larger samples a most recent addition. Their seemingly conflicting results on the relative contribution of social and sun time and the seasonality of human sleep highlight the complexity of the matter at hand. Moreover, it indicates the biological and epidemiological insights that may be gleaned from assessing systematic patterns in sleep with season, weekly structure and geographical position to pinpoint vulnerable populations and effective countermeasures where sun and social time conflicts may lead to unfavorable sleep outcomes and risks for acute and long-term health and safety may arise.
With the Corona Data Donation project we were able to assess many of the social and solar attributes of sleep for the first time together in one large, longitudinal data set of objective sleep phenotypes. Our results emphasize the effectiveness of passive sensing for monitoring and understanding trends in public health.
