Reproduce “Brain2Vec: A Self-Supervised 3D Foundation Model for Structural Brain MRI” 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
Spiros Maggioros, Guray Erus, Gareth Harman, George Aidinis, Alexander Getka, Aristeidis Sotiras, Christos Davatzikos
Brain2Vec learns from unlabeled structural brain MRI by masking scan patches and reconstructing hidden content. Its encoder can be fine-tuned for a task or used to produce fixed features. On external IXI brain-age testing, it achieved mean absolute error of 5.1667 years versus 6.1846 with random initialization. This result applies in that setup and does not establish benefit across tasks or populations.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.