Reproduce “iBrain: A Unified Foundation Model Reading the Brain from Surface to Spikes” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ying Chen, Tiou Wang, Zhifeng Yue
iBrain jointly pretrains on intracranial EEG (iEEG) and spike recordings for decoding. Separate input encoders feed a shared spatiotemporal Transformer, which learns across channels and time. In downstream tasks, joint pretraining improves several scores over single-signal pretraining, including Speech AUC (0.89 versus 0.86) and Perich T-RT R² (0.69 versus 0.64), while Pitch ties; training budgets were not matched.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.