Attention’s Substates: Unlocking Adaptive VR through EEG Insights

To enable adaptive virtual reality (VR) based systems, this study explored to differentiate attention’s 3 substates – i.e., orienting (OR), alerting (AL), and cognitive maintenance (CM). Human participants performed a task within a VR-based cockpit to evoke these substates, while their brain activity was recorded as electroencephalographic data to extract features. Statistical analyses revealed significant differences of these features among the substates. Compared 5 machine-learning models, naïve bayes and K-nearest neighbors achieved similarly the best accuracy (77%~ 95%) to classify the features. These findings implied a potential of the features for informing design of adaptive VR-based systems.