fMRIPrep Preprocessing time

Software Support
1

Summary of what happened:

I need to preprocess structural (T1 and T2) and functional (rs-fMRI) data of 1475 subjects within the OASIS3 dataset with fmrprep. I intend to perform some surface analyses, but I have realized that performing freesurfer reconstruction takes up quite a long time. So as to save processing time, would it make sense not add the --cifti-output 91k flag, use the anatomical surface models relative to the MNI template, and to obtain the cifti 91k output after fmriprep with the preprocessed volumetric functional data?

Command used (and if a helper script was used, a link to the helper script or the command generated):

fmriprep \
  /home/jovyan/Desktop/DatasetOASIS3/raw \
  /home/jovyan/Desktop/DatasetOASIS3/derivatives \
  participant \
  --skip-bids-validation \
  --n-cpus 16 \
  --mem 240000 \
  --work-dir /home/jovyan/Desktop/DatasetOASIS3/work \
  --output-spaces MNI152NLin2009cAsym:res-2 \
  --bold2anat-init t2w \
  --cifti-output 91k \
  --fs-license-file /home/jovyan/Desktop/PBL_Neuro/Freesurfer/license.txt \
  --random-seed 34 \
  --notrack

Version:

25.1.3

Environment (Docker, Singularity / Apptainer, custom installation):

Neurodesktop

Data formatted according to a validatable standard? Please provide the output of the validator:

PASTE VALIDATOR OUTPUT HERE

Relevant log outputs (up to 20 lines):

PASTE LOG OUTPUT HERE

Screenshots / relevant information:

2

Hi @Ibai_Azpeitia_Loiti,

I don’t think that would make sense. Removing the cifti-output flag would still have freesurfer running. If you disable freesurfer, it not only takes away your subject specific surface, but also disables helpful BOLD preprocessing steps (such as boundary based registration). Also, your step would add an extra layer of interpolation, which should be avoided if possible.

Best,
Steven