Purification of Brain Signals Using Various Blind Source Separation Techniques

Abstract

There are lot of challenges when analyzing the brain signals that did not yet have a basic solution,
as there are electrical activities between neurons in the brain, related to all activities in the body. This
activity can be seen using a non-surgical technique that is called EEG (i.e. Electroencephalography) such
as the appearance of artifacts through the registration process, which increases the difficulty of analyzing
the signals of the brain, so the technique of blind source separation (BSS) has been used to overcome the
problem of artifacts and to separate the main sources (Mixed) without making noise around the original
sources.. Therefore, the system for rejecting all artifacts based on the algorithms for separating the blind
source has been proposed, by making a comparison between four separation algorithms and choosing the
best ones according to criteria. After passing a data set simulated through those criteria, the proposed
system can remove the artifacts including Electrocardiogram (ECG), Electrooculogram (EOG) as well as
a power line noise interference (LN)) and other EEG mixtures. The proposed method's influence is checked
by two performance indexes Interference to signal ratio (ISR) and (SNR) signal to noise ratio. The results
indicated that the BEFICA algorithm is the best and most efficient, as it achieved the highest ratio of VSNR
among four separation algorithms due to its developmental advantages