Estimation of Affective States in Virtual Reality Environments using EEG

Recent developments in accessible virtual reality (VR) technologies have inspired research into increasing the user's sense of immersion via monitoring and feedback of physiological and cognitive states. This work presents a pilot study that models the user's affective state from electroencephalographic (EEG) measurements during observation of immersive VR videos. Participants passively viewed a series of short VR video clips and subjectively rated the level of valence, arousal, liking, and dominance. Separate regression and classification models were developed to estimate the subjective ratings for each category using EEG spectral features. The results indicate that both models are capable of capturing relevant EEG features for estimating the user's affective state ratings.