The video gaming industry is experiencing a significant surge in market capitalization, reaching unprecedented heights. In recent years, cognitive neuroscience has emerged as a powerful instrument for improving game development by leveraging insights into the mental processes of gamers. In this study, an experimental study is conducted to recognize five different game experience traits using electroencephalography (EEG) signals. EEG data from 27 participants for three trials of the (Traffic Racer) game is recorded using a commercially available four-channel Muse headband. The EEG data are labeled into two groups for flow, challenge, competence, tension, and negative affect game experience traits using the game experience questionnaire (GEQ) score. Significant bands of the EEG signal for each trait of the game experience are identified using a t-test. Three distinct groups of frequency domain features (mean power, rational asymmetry, and differential asymmetry) are extracted from statistically significant frequency bands of the EEG signal, which are employed to classify game experience traits using multiple machine learning classifiers. The accuracy of 92.59%, 91.36%, 86.42%, 87.65%, and 85.19% is achieved using a k-nearest neighbor (kNN) classifier for flow, challenge, competence, negative effect, and tension traits, respectively. The proposed game experience recognition algorithm performs better with features extracted from selected EEG bands with a reduced feature vector length (FVL).
Loading....