Independent Component Analysis for sleep-spindles detection using an Extended Infomax Algorithm and Fixed-point Algorithm

Abstract

We investigated the possibility to use the Independent Component Analysis (ICA) as a method for preprocessing the sleep EEG data with the aim to improve detection of sleep spindles - specific phenomena of sleep EEG recordings prevailingly occurring during the stage 2 of the sleep. We projected the strengths of individual Independent Components (ICs) onto the scalp sensors to detect potential spatial localization of sleep-spindles sources. We used two different algorithms for ICs separation with aim to compare the fitness of the algorithms in sleep-spindle detection problem.


Go back