EEG signals identification using neural network due to radiofrequency exposure

Abstract

Electroencephalogram (EEG) signals, alpha, beta, theta and delta sub bands were used as inputs to the signals identification system with three discrete outputs: left group, right group and control group. By identifying features in the EEG signals we want to distinguish the significant difference of the three groups of brainwaves and also between the sessions of exposure to the radiofrequency (RF). This article discusses a technique for analyzing EEG signals using asymmetry feature extraction and human brainwave signals identification using artificial neural network (ANN). Power asymmetry ratio (PAR) feature is particularly effective for representing brainwave dominance between left and right hemisphere. After proper processing of the data thru selected feature extraction, neural network system identification was obtained to classify the brainwave signals due to the exposure of mobile phone radiofrequency (RF). A unique and reliable classification model was developed through the combination of PAR as feature extraction and ANN as system identification. The emerging computationally powerful technique based on ANN was successful to identify the brainwave signals due to different groups of exposure with 100 percent accuracy during the exposure.