A validation study of automatic sleep detection with wearable devices in children

In recent years an increasing body of evidence has confirmed the crucial role of sleep in the overall health of the individual and, more specifically, in the psychophysical development of the child. The result is a growing interest in the methods of determining the sleep/wake cycle: although the gold standard for sleep monitoring is still polysomnography, its invasiveness and costs restrict its use, especially in pediatric population. A reliable alternative is actigraphy, which has been confirmed as an accurate device for the measurement of sleep parameters, allowing the study of night and day sleep patterns for long period of time, in the most unobtrusive way. Actigraphy has lately become widely consumed through the diffusion of commercial activity trackers. In this study, we propose the use of a commercially available wearable smartband, Fitbit® Flex, for the assessment of sleep in children. We obtained data on motor activity on a heterogeneous population of children with and without neurodevelopmental disorders, then we analysed the collected data by applying an algorithm already electro-medically validated on adults, to determine sleep/wake patterns. We compared the resulting data with parental reports obtained through a validated and clinically relevant sleep diary and with sleep reports provided by Fitbit webapp platform itself. Our hypothesis that the data obtained with Fitbit® Flex may represent a valuable tool for the study of sleep in children, as well as in adults, has been confirmed, supporting the potential of this device and its promising role in pediatric sleep medicine. These results highlight the importance and need of thorough and validated algorithms for the actigraphic detection of sleep and waking in children