fMRIPrep 23.0.0

We have just released fMRIPrep 23.0.0, which includes a number of improvements.

There are some new features highlighted in the release notes below, but please do test this out if susceptibility distortion correction (SDC) has been causing troubles for you in recent releases.

If you are still seeing issues, please use the new Software Support - Neurostars category and make sure to tag with fmriprep, sdc and include “23.0” in the title.

Release Notes

This release adds improvements for workflows targeting the fsLR grayordinate space. Namely, morphometric (curvature, sulcal depth and cortical thickness) measures are output as .dscalar.nii files and high-variance voxels can be excluded from the resampling step using --project-goodvoxels.

Additionally, T2w images are now resampled to the T1w-defined subject space if FreeSurfer reconstruction is used. If multiple T2w images are provided, they are merged into a single image first.

PEPolar fieldmaps with R/L phase-encoding directions or in non-standard orientations are now better supported. We continue to work toward better support for more SDC configurations.

23.0.0 supports FreeSurfer 7.3.2, which is now bundled in the Docker image.

ICA-AROMA support will be removed in 23.1.0.

Full changelog


Do you know why support for ICA-AROMA will be removed?

Thanks in advance,

Hi @haya and welcome to Neurostars!

AROMA is no longer actively maintained and is losing comparability with recent FSL versions. See here: Remove ICA-AROMA · Issue #2936 · nipreps/fmriprep · GitHub


One thing I’ve considered is pulling the old AROMA code out into a small BIDS app that will take an fMRIPrep output dataset that has been normalized to MNI152NLin6Asym and generate AROMA derivatives. That app could be distributed in a Docker container that never needs to upgrade FSL.

1 Like