Extraction of Entropy and Hurst Exponent Biomarkers of EEG from a Person Seeing 2D, 3D, and 4D Images Using a Virtual Reality Device

The human brain's capacity to process visual information in varying dimensions remains a subject of profound scientific inquiry. This paper investigates the neural correlates of perceiving 2D, 3D, and conceptual 4D visual stimuli presented through a Virtual Reality (VR) device, using Electroencephalography (EEG). Specifically, we extracted and analyzed non-linear EEG biomarkers: Entropy (Permutation Entropy) and the Hurst Exponent, to characterize brain activity patterns. A total of 25 healthy participants were exposed to a standardized set of visual stimuli, comprising static 2D images, stereoscopic 3D images, and dynamic, rotational projections of 4D hypercubes designed to induce a sense of higher-dimensional complexity within a 3D VR environment. EEG data were recorded, preprocessed, and segmented for feature extraction. Our findings indicate significant differences in both Permutation Entropy and Hurst Exponent across the three stimulus conditions, particularly in frontal and parietal regions. Subsequent classification analysis, employing a Support Vector Machine (SVM) classifier, revealed remarkably high accuracy in distinguishing brain states corresponding to the different dimensional perceptions. The classification accuracy for 4D stimuli reached an impressive 95%, significantly outperforming the classification of 2D and 3D stimuli (78% and 86% respectively). These results suggest that exposure to perceptually complex, higher-dimensional representations, even conceptually simulated, elicits unique and highly distinguishable neural responses, as captured by these non-linear EEG biomarkers. This study highlights the potential of VR as a tool for exploring advanced cognitive processing and the utility of non-linear EEG analysis in understanding the brain's engagement with novel perceptual experiences, with implications for brain-computer interfaces and neurorehabilitation.