Primary Color Decoding Using Deep Learning on Source Reconstructed EEG Signal Responses

The exploration of utilizing the brain’s response to varied colors for controlling one’s surroundings via a brain-computer interface (BCI) presents a promising avenue. Enabling users to interact with their environment by focusing on specific colors could offer an intuitive interface, allowing for tasks like adjusting lighting or opening doors. In this study, the focus lies on developing an intra-subject classifier for red, green, and blue (RGB) visual evoked potentials (VEPs) recorded through electroencephalogram (EEG). By employing three distinct deep neural networks (DNNs), previous research propositions were tested, assessing information in both source and cathode space. Notably, analyses conducted in electrode space consistently outflanked those in source space. The best classifier accomplished a normal precision of 78% across subjects, with a singular subject arriving at an amazing 97% exactness. In terms of clinical significance, this research underscores the potential of leveraging deep learning techniques to effectively distinguish between VEPs corresponding to different colors within EEG recordings, offering insights into potential applications in assistive technologies and neurological research.