rMVPA is an R library for multivariate pattern analysis of
neuroimaging data. The goal of this library is to make MVPA analyses
easy. It can be used both programmatically from within R or using a
command line interface. rMVPA provides a lightweight model registry
and efficient resampling methods for machine learning. What rMVPA
provides is the infrastructure for conducting machine learning analyses
on neuroimaging data.
Documentation and vignettes: https://bbuchsbaum.github.io/rMVPA/
The supported public workflow/API surface is now intentionally smaller and
explicit. Use rmvpa_api_lifecycle() to inspect lifecycle tiers and
rmvpa_stable_api() to list the stable entry points intended for scripts,
extensions, and downstream packages.
To install rMVPA from within R, use the devtools function
install_github. You will need the development version of neuroim2 as
well.
From within R:
library(devtools)
install_github("bbuchsbaum/neuroim2")
install_github("bbuchsbaum/rMVPA")
git clone git@github.com:bbuchsbaum/rMVPA.git
R CMD install rMVPA
rMVPA now ships packaged CLI wrappers with native command names:
rmvpa-searchlightrmvpa-regional
Install the package:
pak::pak("bbuchsbaum/rMVPA")Then copy the wrappers into a directory on your PATH:
rMVPA::install_cli("~/.local/bin", overwrite = TRUE)If needed, add that directory to PATH:
export PATH="$HOME/.local/bin:$PATH"On macOS/Linux, you can then run:
rmvpa-searchlight --help
rmvpa-regional --help
rmvpa-searchlight --versionThe unified CLI vignette is vignette("CommandLine"). The legacy source-tree
entrypoints remain available as compatibility shims, but the packaged commands
above are the supported interface.
If you use rMVPA in your research, please cite:
Buchsbaum, B. (2026). rMVPA: Multivoxel Pattern Analysis in R. R package version 0.1.2. https://github.com/bbuchsbaum/rMVPA