bidser

BIDS in R – (it’s a start!)

The goal of bidser is to make working with the BIDS neuroimaging format convenient in R. Current support is strongest for MRI datasets, with explicit query helpers, metadata inheritance, derivative pipeline discovery, and compatibility-oriented support for fmriprep workflows.

Installation

Install the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("bbuchsbaum/bidser")

Example

See https://bbuchsbaum.github.io/bidser/articles/quickstart.html

fMRIPrep confounds

read_confounds() selects nuisance regressors from fMRIPrep confound tables. Rather than hand-listing version-specific column names, use the high-level, version-robust helpers:

# Named, version-robust sets (resolve to whatever columns your dataset has)
read_confounds(proj, cvars = confound_set("motion24"))
read_confounds(proj, cvars = confound_set("36p"))

# PCA + raw denoising strategies (recommended modern default)
read_confounds(proj, cvars = confound_strategy("pcabasic80"))

# Discover what is available
list_confound_sets()
list_confound_strategies()

Code that previously reached into the unexported bidser:::DEFAULT_CVARS2 should switch to the stable public handle confound_set("legacy_default"), which returns the identical 26-name set. See ?read_confounds, ?confound_set, and the confounds-and-variables vignette for details.