Effects of visual differences between real and HMD-based virtual objects on brain activity: an EEG study

Head mounted displays (HMDs) have become popular since they can improve the quality of experience by enhancing immersion compared to ordinary 3D displays. It is important to understand the functionality of the brain while interacting with virtual and real objects, in order to design and optimize superior HMD multimedia processing techniques. EEG signals were exploited in this study to achieve two goals. Our first goal was to compare brain activity of subjects while watching objects through HMD and while observing the same objects in real world. This was carried out using topographic brain maps, based on TRPD/TRPI measure. The results suggested that parietal areas were more active (higher TRPD) for HMD session, whereas in real object session, occipital and frontal areas were so. Our second goal was to investigate the possibility of classifying these two classes, using relevance vector machines (RVM) and also finding the most significant features in this task. Wavelet features being the most significant, overall classification accuracy (ACC) of 96.6% was achieved.