MENTAL WORKLOAD ASSESSMENT OF GAMERS’ EEG WITH MULTI-DOMAIN FEATURE-BASED COGNITIVE MODEL AND ITS VALIDATION

Abstract

Electroencephalography (EEG) has been recognized as a competent mode for cognitivists to study brain activity potential variations with imposed mental workload. The elementary focus of a game design is to provide entertainment to the players and be a source of recreation and relaxation at the same time. However, it is observed that with increasing difficulty levels the game may start imposing a mental workload beyond a critical threshold. To work on such an issue, this paper presents a multi-domain EEG features-based cognitive model that classifies the mental workload levels on a difficulty basis. It investigated the brain activity of gamers while spotting the differences in similar-looking pictures at three levels of difficulty. The collected brain signatures developed a mental workload model using a Support Vector Machine (SVM) within an accuracy range of 84% to 98% ([Formula: see text]). After the machine learning-based modeling, the model was subjectively validated with NASA’s Task Load Index (NASA-TLX) rating scale. Every level was rated by the players for different mental workload factors and a workload score was obtained for every level. A correlation study showed the association of different EEG features with the workload score within a correlation coefficient range of 0.306 to 0.748 ([Formula: see text]-2-tailed). Findings of this type may lead to the scholarly and professional measurement and improvement of cognitive aptitude.