EEG Based Stress Level Detection During Gameplay

Playing mobile games comes with the benefits of entertainment but many of these also introduce stress to the human brain. This mental stress can be identified by Brain-computer interface (BCI) techniques. It has various neuroimaging approaches which help in analyzing the brain signals and Electroencephalogram is the most popular approach among them, as it is inexpensive, portable, and non-invasive. Here, this paper investigates the pattern in EEG signals when introduced with mental stress. Ten healthy volunteers played the Word search game for three distinct levels, whose aim was to search hidden words from the grid. The levels are chosen randomly. The EEG signals during gameplay are recorded to investigate the impacts of stress with the changing levels from easy to medium and hard. A total of 16 features of EEG are analyzed in this experiment which includes Power band features with relative powers, the ratio of band powers, Event-Related Desynchronization, and statistical features. Support Vector Machine classifier is used which gave an accuracy of 77.2% for three-level based stress analysis. For two levels, the accuracies of 78% 83.7% and 90% are achieved.