Sleep insights from the finger tip: How photoplethysmography can help quantify sleep

Abstract

Sleep is essential for a healthy and productive life, yet its importance is largely overlooked, allowing populations to sleep less and to develop sleep disturbances. This trend results into an epidemic of poor quality and insufficient sleep that is turn jeopardizes health, performance, mood, memory, social relationships and productivity. A first step to overcome this epidemic relies on uncovering it at the individual and societal levels. The availability of different wearable devices that can track physiological signals represents a great opportunity to define and quantify the problem. Many such devices incorporate a photoplethysmography (PPG) sensor. This triggers interest in studies aiming to get insight into human sleep structure based on information obtained from PPG sensors. This study aimed to validate a new automated sleep analysis which is simply based on the inter-beat-interval series obtained from PPG, and that uses features of heart rate variability. The candidate algorithm was tested against gold standard scoring of whole night state-of-the-art sleep studies. The PPG-based sleep scoring performs very well in differentiating sleep stages, however, the sleep/wake separation is not sufficient and requires improvement. This last task is facilitated by the fact that the majority of devices with PPG capabilities, are equipped with accelerometers providing additional information for better separation. Combining accelerometers and PPG signals from wearable devices in a sleep analyser is likely to provide a reliable and accurate automated detection of sleep and wakefulness, including sleep macro- and microarchitecture.