Can finger twitches be used to detect REM sleep?

Introduction: Abnormal rapid eye movement (REM) sleep is often symptomatic of chronic disorders. However polysomnography, the only validated method to measure REM sleep, is often impractical. This study considered a novel approach to detect REM sleep by measuring finger twitching, based on elevated levels of distal muscle twitching observed during REM sleep in previous studies. We predicted two key effects: that there would be more finger twitching in REM sleep than in non‐REM sleep, and there would be more finger twitching in late REM sleep than in early REM sleep. Method: An observational, within‐subjects design was used. One night of sleep data was collected from each of 18 participants using standard polysomnography. Finger movement was detected using a piezo‐electric limb sensor on the index finger of each participant, allowing finger twitch densities to be calculated. Results: Finger twitch density was found to be greater in REM sleep than in non‐REM sleep. Each sleep stage was also found to have a unique finger twitch density, with the exception of REM and stage N1 sleep, which had similar densities. Finger twitch density was greater in late REM sleep than in early REM sleep, and there was also a time‐state interaction: the difference between finger twitch densities in REM sleep and non‐REM sleep was greater in late sleep than in early sleep. Discussion: The results provide evidence for the hypothesised effects, and suggest that finger twitching in REM sleep is distinguishable from finger twitching in most other sleep stages, and increases and becomes more distinguishable from finger twitching in non‐REM as sleep progresses. A simple algorithm was developed to demonstrate that finger twitch density can be used to determine epoch‐by‐epoch sleep states (wakefulness, REM sleep or non‐REM sleep) suggesting further investigation of this approach to REM sleep measurement is warranted.