I have acquired data, and I obtain this kind of BIDS folder using dcm2bids:
anat = T1w
fmap = SE A>>P and P
func = 4D images
I am running fMRIprep on it, and everything is just fine.
I have just learned about defacing. From my understanding, one should deface anatomical images before running fMRIprep. My question in this regard is: how to organise and name data?
Will fMRIprep accept his structure & naming, and will *_defaced.nii.gz be recognized as the image to use? Or should I replace the original file with the defaced one to have only *.nii.gz, obtaining something like
dcm2bids + defacing + renaming & reorganizing:
anat
|_ sub-01_T1w.json
|_ sub-01_T1w.nii.gz
Hope I’m clear, thanks in advance to all of you folks!
You can use a rec-defaced label, sub-x_rec-defaced_T1w.nii.gz. You should do this before fmriprep. Also you’ll want to either delete or filter out the original non-defaced images (e.g., with bids filter file).
Hi @Steven , thank you very much for your reply! Quick and clear, as usual.
So fMRIprep will recognise a /anat/sub-x_rec-defaced_T1w.nii.gz automatically? Should the JSON be renamed also (/anat/sub-x_rec-defaced_T1w.json)?
Regarding the second point “Also you’ll want to either delete or filter out the original non-defaced images (e.g., with bids filter file).”, are you mentioning the .bidsignore file, with something like
sub-*/anat/*_T1w.nii.gz to ignore original T1
!sub-*/anat/*_rec-defaced_T1w.nii.gz to keep the defaced T1
All in, if I am never interested in reusing the original T1 but only the defaced, maybe keeping this structure is simpler?