Brain-OF: A Unified Foundation Model for fMRI, EEG, and MEG

Brain-OF is an omnifunctional brain foundation model jointly pretrained on fMRI, EEG, and MEG data. It handles both unimodal and multimodal inputs within a unified framework. The model introduces an Any-Resolution Neural Signal Sampler to reconcile heterogeneous spatiotemporal resolutions and projects diverse brain signals into a shared semantic space. Its backbone integrates DINT attention with a Sparse Mixture of Experts to manage semantic shifts. This addresses the limitation of existing models that are restricted to single functional modalities, enabling exploitation of complementary spatiotemporal dynamics and collective data scale across neuroimaging techniques.