EEG brain signal analysis II

Abstract

Electroencephalogram (EEG) is a powerful tool in the study of human brain and emotions. In this project, an EEG analysis is carried out to study the electrical activity of the brain and the happy and sad emotions. EEG data is collected from a group of 23 subjects during an audio-visual screening intended to stimulate the desired emotions. Images from the International Affective Picture System (IAPS) were compiled with sound clips into a slideshow and while it was being screened, EEG data is collected and extracted using MATLAB. An emotion detection algorithm written on MATLAB is implemented to analyze the results obtained. The algorithm employs a Butterworth band pass filter to extract the alpha brainwave signal and carries out Feature Extraction in the Time domain and Frequency domain for the useful data obtained. Thereafter, Classification based on Support Vector Machine (SVM) is carried out. It was found that the Time domain analysis derived average results of higher accuracy in the range of 70% to 80% compared to Frequency domain analysis which gave only average accuracy of about 50%. The emotion detection algorithm implemented using a combination of seven features in Time domain analysis obtained the highest average classification accuracy of 86.6% and maximum accuracy of 90% which can be considered to be a good classification. Future students may extract different features to further improve on the findings.