Naming convention for fmriprep output and further analysis

1

Hello,

I have made some progress on my pipeline for analyzing data preprocessed using fmriprep. I am primarily using surface data, and would like to follow the conventions of fmriprep and BIDS as closely as possible. Do such conventions or recommendations exist, when it comes to naming (and potentially folder structure) for files that have been submitted for further analysis after fmriprep?

Also, a small comment: Why are the surface outputs from fmriprep named

*bold_space-fsnative.[LR].func.gii?

Should it not be:

*bold_space-fsnative_preproc.[LR].gii

to match the nifti naming convention:

*bold_space-T1w_preproc.nii.gz

My current strategy is to add named analysis steps after the “preproc” field, so that a file might be named:

bold_space-fsnative_preproc-detrend-smooth3fwhm.[LR].gii

But that could change if you recommend something else.

/Peter

2

Hi Peter,

We are currently discussing the specification of BIDS-Derivatives. As a result, release candidate 1 was posted last week (https://docs.google.com/document/d/17ebopupQxuRwp7U7TFvS6BH03ALJOgGHufxK8ToAvyI).

That document is the main reference for outputs in BIDS, and the place for adding questions/comments/suggestions.

On fMRIPrep’s side, BIDS-Derivatives RC1 meant that we are currently refactoring how derivatives are written to comply with the current draft of the spec.

1 Like