This project explores an end-to-end masked prediction, a self-supervised foundation model for EEG that learns unified representations transferable to any downstream task without task-specific ...