Detection and classification of Eye Blink Artifact in electroencephalogram through Discrete Wavelet Transform and Neural Network

Abstract

Electroencephalography (EEG) is the recording of electrical activity along the scalp of human brain. EEG is most often used to diagnose brain disorders i.e. epilepsy, sleep disorder, coma, brain death etc. EEG signals are frequently contaminated by Eye Blink Artifacts generated due to the opening and closing of eye lids during EEG recording. To analyse signal of EEG for diagnosis it is necessary that the EEG recording should be artifact free. This paper is based on the work to detect the presence of artifact and its actual position with extent in EEG recording. For the purpose of classification of artifact or non-artifact activity Artificial Neural Network (ANN) is used and for detection of contaminated zone the Discrete Wavelet Transform with level 6 Haar is used. The part of zone detection is necessary for further appropriate removal of artifactual activities from EEG recording without losing the background activity. The results demonstrated from the ANN classifier are very much promising such as- Sensitivity 98.21 %, Specificity 87.50 %, and Accuracy 95.83 %.