Matching anatomical to template before fmriprep

1

I have performed some basic transformations to the anatomical image prior to fmriprep to match a specific template and this is the image in my anat BIDS folder, is this somehow not advisable?

2

Hi @mri

Thank you for your question! In BIDS the anat folder assumes it is raw or unprocessed or minimally processed due to file format conversion. Applying a transform I think does not fall into this and becomes more of a derivative. fMRIPrep offers spaces if that may help your preprocessing

Thank you,
Franklin

3

Thanks @franklin. The idea is to feed fmriprep a more suitable version of the anatomical image (not the raw image). output-spaces will not do because the point is not to change the output space of the processed image, I simply want to reorient the anatomical slightly before starting with the preprocessing.
Assuming this BIDS structure

data/
   sub-01/
      anat/
        sub-01_T1w.nii.gz 
      func/
        sub-01_task-CAP_bold.nii.gz
   sub-02/
      anat/
        sub-02_T1w.nii.gz
      func/
        sub-02_task-CAP_bold.nii.gz
   ...
   derivatives/
      aligned/
         sub-01/
            anat/
              sub-01_T1w.nii.gz
	 sub-02/
            anat/
              szub-02_T1w.nii.gz

Can I somehow instruct fmriprep to use the anatomical image in the derivatives/aligned folder instead of the data folder?

4

Hi @mri

Thank you for the additional information! To clarify - when you want to reorient to the anatomical by chance are you referring to a tool such as reorient2std or performing a registration? I would caution in mixing between the raw and minimally touched reorientation. Perhaps denoting this in your README so it is documented for future use. It appears that fMRIPrep takes the raw directory as the principle input

Thank you,
Franklin