Researchers propose a new framework called MD-SigLIP to improve brain-language decoding. This method aligns brain embeddings with text embeddings in a shared semantic space, allowing for more accurate retrieval-based decoding. The approach models the correspondence between neural representations and language semantics, addressing the ambiguity that limits interpretability in this area. Experiments show state-of-the-art performance under various evaluation settings. --- Why it matters: This matters because it could improve our understanding of how brains process language, which is crucial for developing more effective brain-computer interfaces and other applications. Source: https://arxiv.org/abs/2608.16975