Comparative Study of EEG Feature Representations for Visual Decoding with EEGNet and EEG Conformer

Electroencephalography (EEG)-based visual decoding is important for brain-computer interfaces, yet the relative effectiveness of feature extraction methods for sustained visual paradigms remains unclear. This study systematically compares eight feature extraction methods (coherence, Granger causality, directed transfer function, partially directed coherence, transfer entropy, discrete wavelet transform, empirical wavelet transform (EWT), and wavelet scattering transform) across three public EEG datasets: MindBigData MNIST, MindBigData MNIST-8B, and MSS. Two deep learning architectures, EEGNet and EEG Conformer, were trained under two pre-processing pipelines, with and without artifact removal. EWT achieved the highest accuracy, reaching 97.83% for digit-vs-blank and 77.10% for within-session natural image classification; connectivity-based methods consistently underperformed (best 91.67%), indicating that spectral power information is more discriminative than inter-channel relationships. Cross-subject generalization remained challenging (best near 68%). The findings establish wavelet-based adaptive spectral decomposition as a strong baseline for EEG visual decoding and highlight the need for domain adaptation techniques.