Are fmaps redundant if 2 DWI volumes with opposite phase encoding directions are acquired?

1

I have a dataset with the following input files

sub-001/
├── anat
│   ├── sub-001_T1w.json
│   └── sub-001_T1w.nii.gz
├── dwi
│   ├── sub-001_dir-AP_dwi.bval
│   ├── sub-001_dir-AP_dwi.bvec
│   ├── sub-001_dir-AP_dwi.json
│   ├── sub-001_dir-AP_dwi.nii.gz
│   ├── sub-001_dir-PA_dwi.bval
│   ├── sub-001_dir-PA_dwi.bvec
│   ├── sub-001_dir-PA_dwi.json
│   └── sub-001_dir-PA_dwi.nii.gz
└── fmap
    ├── sub-001_dir-AP_epi.json
    ├── sub-001_dir-AP_epi.nii.gz
    ├── sub-001_dir-PA_epi.json
    └── sub-001_dir-PA_epi.nii.gz

i notice that qsiprep does not make use of any of the fmaps, even though their IntendedFor parameters are correctly specified in the .json files.

IntendedFor": [
    "dwi/sub-001_dir-AP_dwi.nii.gz",
    "dwi/sub-001_dir-PA_dwi.nii.gz"]

I’m wondering if the fmaps are redundant in this case, since according to DWI distortion correction using dwifslpreproc — MRtrix3 3.0 documentation (Reversed phase encoding for all DWIs) ‘This allows for the combination of the two volumes corresponding to each unique diffusion gradient direction & strength into a single volume, where the relative compression / expansion of signal between the two volumes is exploited.’

2

Hi @gerardyu,

Yes, QSIPrep prefers to use b0s from the DWI for SDC over fmaps if available.

Best,
Steven