Stochastic process modeling for waking during nocturnal sleep; based on EEG analysis using a corticothalamic model

Abstract

Waking during nocturnal sleep (WS) is gaining attention in the health science field. Although some findings, (e.g., the power-law distribution of the wake duration time) have been reported, the mechanism of it is still unclear. In this talk, we light on this issue. In a previous study, we showed that corticothalamic feedback loop strength (c2), estimated by an EEG analysis method, clearly tracks the wake-to-sleep transition just before nocturnal sleep. This time, we applied this method to nocturnal sleep. The c2-distribution (n = 52) during nocturnal sleep exhibited clear bimodal, which can be well approximated by superposition of two Gaussian distributions corresponding to sleep (Non-REM 1-4 and REM) and wake states, respectively. This observation suggests that the dynamical change of c2 obeys the Ornstein-Uhlenbeck process in each state. Based on these findings, we show an attempt for stochastic process modeling of WS.