Reproduce “Fast shared response model for fMRI data” 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
Hugo Richard, Lucas Martin, Ana Luısa Pinho, Jonathan Pillow, Bertrand Thirion
FastSRM makes shared-response analysis of fMRI data more practical by projecting voxel activity onto atlas-defined parcels, fitting deterministic SRM there, and recovering full-brain subject-specific maps. Across four datasets and tested atlases, it matched ProbSRM and DetSRM reconstruction performance. On CamCAN, it predicted age with 7.5-year mean absolute error; fitting used 1.1 GB and 15 minutes under tested conditions.
Key insights
- An intermediate projection onto a coarse brain atlas can dramatically improve the scalability and efficiency of the Shared Response Model without compromising fMRI data reconstruction accuracy.
- The optimized FastSRM enables practical application of SRM to large-scale fMRI datasets, facilitating the extraction of meaningful individual-specific traits such as age from naturalistic viewing paradigms.
- Spatial components derived from FastSRM are interpretable, revealing specific brain networks (e.g., visual dorsal pathway, precuneus) that exhibit age-related changes during movie-watching.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.