T2w-ACPC image not being output

[smeisler]

I have a BIDS-valid dataset with T1 and T2 images. The qsiprep boilerplate suggests that

The additional T2w reference image was registered to the T1w-ACPC reference image using an affine transformation in antsRegistration

However, I do not see any T2 images in my output directory:

/nese/mit/group/gablab/data/BRAID/bids/derivatives/qsiprep_0.20.0/sub-BRAID003P/anat/
├── sub-BRAID003P_acq-vNavNorm_from-orig_to-T1w_mode-image_xfm.txt
├── sub-BRAID003P_desc-aseg_dseg.nii.gz
├── sub-BRAID003P_desc-brain_mask.nii.gz
├── sub-BRAID003P_desc-preproc_T1w.nii.gz
├── sub-BRAID003P_dseg.nii.gz
├── sub-BRAID003P_from-MNI152NLin2009cAsym_to-T1w_mode-image_xfm.h5
├── sub-BRAID003P_from-T1wACPC_to-T1wNative_mode-image_xfm.mat
├── sub-BRAID003P_from-T1wNative_to-T1wACPC_mode-image_xfm.mat
└── sub-BRAID003P_from-T1w_to-MNI152NLin2009cAsym_mode-image_xfm.h5

QSIPrep completed without errors:
singularity run -e -B ${scratch},${cache_dir} ${qsiprep_img} $scratch/data $scratch/data/derivatives participant --participant_label ${subject:4} -w $scratch --fs-license-file ${scratch}/data/code/license.txt --denoise-method patch2self --pepolar-method TOPUP+DRBUDDI --unringing_method mrdegibbs --recon-spec $scratch/data/code/recon_spec.json --output-resolution 1.25 --freesurfer-input ${scratch}/data/derivatives/freesurfer/ --fs-license-file $scratch/license.txt --skip-odf-reports --notrack --mem-mb $mem_mb --nthreads 16 --omp-nthreads 8

I've noticed this with 0.19.0 and 0.20.0

Thanks for your attention,
Steven

[mattcieslak]

Ah! I forgot to add a datasink to the T2w. Thanks for opening a bug report.

Also, have you taken a look at the html report using the TOPUP+DRBUDDI? If you have "rpe_series" data then you'll see a new FA-alignment figure in there.

[smeisler]

I don't see the new FA-alignment figure in the HTML unfortunately - happy to share any data that may help

[mattcieslak]

And you're sure it's multiple dwi series in multiple phase encoding directions? The report only comes out in v0.20.0. The section header (I realize now it's not very descriptive) should be

Susceptibility distortion correction

Results of performing susceptibility distortion correction (SDC) on Fractional Anisotropy (FA) images.

[smeisler]

It's one DWI series and a RPE fmap.

[mattcieslak]

ahh ok then the FA SDC part won't run and there won't be a report. The T2w should still get sent to the derivatives directory though. do you see much movement in the DRBUDDI html report?

[smeisler]

Looks like the bulk of correction is handled by TOPUP, but there is some movement in DRBUDDI

TOPUP (lots of correction)

DRBUDDI (not as much, but still noticeable)

Before

After

[mattcieslak]

Very interesting. I haven't tried TOPUP+DRBUDDI with just an epi fmap before. It looks reasonable offhand, but it will be interesting to see how downstream things are affected.