I’m trying to use slightly untraditional pepolar SDC inputs. I have a functional scan with matching sbref image (automatically generated using the cmrr c2p sequence). I also have a single reverse phase encode ge-epi sequence with a matching sbref (similarly acquired with the cmrr c2p sequence). For clarity:
fmap:
sub-14771_ses-01_acq-c2p_dir-PA_run-01_epi.nii.gz
sub-14771_ses-01_acq-c2p_dir-PA_run-01_sbref.nii.gz
func:
sub-14771_ses-01_task-audio_acq-c2p_run-01_bold.nii.gz
sub-14771_ses-01_task-audio_acq-c2p_run-01_sbref.nii.gz
I’m trying to use the two ge-epi sbref images (with higher signal and better contrast) to estimate the fieldmap, and apply it to the main functional run. The trouble I’ve run into after setting up the BIDS dataset and trying to run fmriprep, is that the dataset does not pass BIDS validation, I believe because the pepolar fmaps don’t allow for the sbref suffix:
"Evidence: sub-14771_ses-01_acq-c2p_dir-PA_run-01_sbref.json"
One could argue that the sbref image should be acquired directly, and not as a byproduct of acquiring a parameter matched pepolar epi, but I think this method is simpler for most users to set up (at least with the cmrr c2p sequence), and having both the sbref and non-sbref images presents two possible options for SDC inputs.
I could delete the non-sbref image (since it doesn’t have an intended use) and use its naming convention for the sbref images, but that feels like bad practice, where the nature of the acquisition is ignored and opaque. In theory it’s nice to have both in case there is a problem with the sbref images or some other reason to fall back to using the non-sbref images for SDC.
Is there some other convention I should follow? Should the sbref suffix be added as an option for pepolar fieldmaps?