Automatic Onset Detection of Rapid Eye Movements in REM Sleep EEG Data

摘要

Rapid eye movements (REM) during sleep is a descriptor of the REM sleep stage. Parameters associated with REM sleep, such as REM sleep numbers, REM density, REM latency, and pre-REM negativity have been associated with the functional role of REM sleep. The temporal properties of these parameters appear to play an essential role in REM sleep, so precise knowledge of these temporal properties, particularly REM onset, can help elucidate the temporal dynamics of neural activity related to REM sleep. However, manual detection of this event is a time-consuming and subjective process that can be facilitated by an automatic detection tool. We developed an automatic REM onset detection algorithm based on features describing rapid eye movements and compared the results obtained with human detection by a sleep expert.