Fine-Grained Sleep-Wake Behaviour Analysis

Sleep stages are traditionally assessed by experts from polysomnography measurements following specific guidelines. Sleep stage behaviour is subsequently used to detect anomalies and diagnose diseases in a laboratory setting. Recently, with the development of Internet of Things, there is a trend to use everyday technologies for sleep behaviour analysis at home, having the potential to eliminate expensive in-hospital monitoring. We propose a fine-grained sleep-wake behaviour analysis approach, which takes into consideration the influences of various factors, such as gender, health status and race. In addition, we investigate the combination of multiple data sources, in particular, actigraphy and heart rate variability, for enhancing model accuracy. Initial results show the proposed approach is recognising sleep and wake stages accurately and is providing a flexible recognition approach towards personalised sleep-based health monitoring.