EEGDM Uses Latent Diffusion to Capture Global EEG Patterns

EEGDM introduces a self-supervised framework that replaces masked reconstruction with a latent diffusion objective. This approach forces the model to capture holistic temporal patterns and cross-channel relationships in neural activity.