Sleep spindle detection in sleep EEG signal using sparse bump modeling

Abstract

Sleep spindle is the hallmark of second stage of sleep in human being, which is defined as a rhythmic sequence with waxing and waning waves, whose frequency is approximately between 8 to 14 Hz, and its time duration is between 0.5 to 2 seconds. Bump modeling is a method for extracting regions with higher amounts of energy in a related time-frequency map. The bump model of the sleep spindle consists of a group of high energy bumps concentrating in approximately 8 to 14 Hz frequency band. In this study, it will be shown that the power of bumps of EEG can be used in automated detection of sleep spindle. The presented method sensitivity is 99.41% which shows high correctly detection rate, and its error detection ratio is 14.51%, which demonstrates the low dependency of the presented algorithm to the subjects, and its low false detection ratio.