v. 23.2.0a2 SDC TOPUP odd results

[manuegrx]

What happened?

Hi,

With the last version of fMRIPrep (v23.2.0a2) an "artifact" appear on our data (data from Philips scanner) after susceptibility distortion correction using TOPUP method. The ventricles seem to be duplicates:

It is not the case when using the v23.1.4, so I not sure if it is an issue with our data or with the last fMRIPrep version.

The "artifact" is present in both fMRIPrep version in the fieldmap image created by fMRIPrep (fmapid-pepolarfmap_desc-epi_fieldmap) (that why I am wondering if it is an issue with our data)

For the v23.2.0a2, the artifact is also present in the final image and in the "desc-coreg_boldref" image but not in the "desc-hmc_boldref" image.

For the v23.1.4, the final image is good.

Thank you,

What command did you use?

docker run -it --rm --user 2001:2001 -v /path/BIDS:/data -v /path/BIDS/derivatives/fmriprep:/out -v /path/work_tmp:/work nipreps/fmriprep:latest /data /out participant --participant-label 43 -w /work --fs-license-file /work/license.txt --output-spaces MNI152NLin2009cAsym func

What version of fMRIPrep are you running?

23.2.0a2

How are you running fMRIPrep?

Docker

Is your data BIDS valid?

Yes

Are you reusing any previously computed results?

No

Please copy and paste any relevant log output.

No response

Additional information / screenshots

No response

[effigies]

Hi, thanks for the bug report! I'm not sure what I'm supposed to be seeing.

What does it mean for ventricles to be duplicates? And can you show a comparison between the old and new results? It would be good to see a direct comparison with manually run topup, if you can.

[manuegrx]

Hi,

Thanks for your quick answer !

Here is a quick comparison between old and new results (the red arrows indicate what looks odd to me):

For both version the TOPUP estimation is the same (same results in /work/fmriprep_23_2_wf/sub_*_wf/fmap_preproc_wf/wf_pepolar_fmap1/topup ), but what seems different is the way the correction is applied.

I haven't yet had time to check t the workflow for this part in fMRIPrep (from what I've seen it's different between the 2 versions).

I can try to apply topup manually but I am not sure about which bold image I should use as an input if I want to reproduce what is done by fMRIPrep

[manuegrx]

Sorry for not closing this issue sooner ! It seems that it was an issue with our data (several runs were merged before to used fmriprep)