Novel Methods for Stress Features Identification using EEG Signals

This paper introduces new methods to extract stress features from electroencephalogram (EEG) signals during two cognitive states; Closed-Eyes (CE) and Open-Eyes (OE) using Relative Energy Ratio (RER), Shannon Entropy (SE) and Spectral Centroids (SC). The group with the stress features was identified and classified using k-Nearest Neighbor (k-NN). The RER in term of Energy Spectral Density (ESD) for each frequency band (delta, theta, alpha and beta) in four different groups consisted of 180 EEG data were calculated and analyzed. Then, the SE was used to confirm the pattern of stress features. Meanwhile, SC was applied to the RER of each group and then the results were selected as input features to k-Nearest Neighbor (k-NN) for the classification purposes. The training and testing of the classifier were evaluated at 50:50 ratios and 70:30 ratios. The proposed method showed promising results where the combination of RER, SE and SC techniques with the training and testing of k-NN set at 70:30 able to detect and classify the group with the unique stress features at 88.89% accuracy.