Advancing Visual Decoding in EEG: Enhancing Spatial Density in Surface EEG for Decoding Color Perception

Investigating the early visual cortex's sensitivity to fundamental visual characteristics such as orientation, contrast, spatial frequency, and colors is a broad field in neural engineering. Based on previous studies showing improved results with higher EEG electrode density, this work aims to advance our understanding of how the brain processes visual information by using ultra-high-density EEG (uHD EEG) surface scalp recordings to capture more detailed spatial and temporal information. In this study, the focus lies on decoding color and black-and-white stimuli by densely covering the occipital region of three healthy subjects with 512 channels, using flexible surface electrode grids. Visual evoked potentials (VEPs) were analyzed to investigate the temporal dynamics of neural activity, and topographical maps were generated on 3D reconstructed head models of all subjects for enhanced spatial visualization. Additionally, a 4-class classification model was employed to assess the discriminative power of extracted visual evoked potentials, focusing on different colors (red, green, blue) and black-and-white stimuli. The results indicate consistent VEP timing and amplitudes across all subjects, with discernible differences observed among colors and black-and-white stimuli. Additionally, distinct focal areas of activation for each stimulus type within the occipital region of the subjects have been identified. Classification results revealed a grand average classification accuracy of 87.2% for all subjects in the 4-class problem, achieved through averaging ten trials of VEPs.