XCP-D summary report - misaligned surfaces

Software Support
1

Summary of what happened:

Hi everyone,

I’m using XCP-D on fMRIprep output. Looking at the executive summary of my subjects, I see that the first images from “BrainSprite Viewer: T1” show misaligned surfaces with respect to the anatomical data:

image
image1729×547 109 KB

I’m not sure what thes surfaces actually are.

In the same XCP-D summary report, the alignment between the atlas and the T1w images also looks fine:

image
image1117×414 81.4 KB

Any help in sorting this out would be greatly appreciated!
Best wishes,

Roey

Command used:

The command includes several environment variables, which I left as is for readability.

apptainer run --cleanenv --bind $FMRI_PATH,$LAB_SHARED "${XCPD_IMG}"  \
    ${FMRI_PATH}derivatives_${SUBJECT} \
 ${FMRI_PATH}/derivatives/xcp_d_070_acompcor_withtedana \
    participant \
    -w ${FMRI_PATH}/derivatives/xcpd_work \
    --participant_label "${SUBJECT}" \
    --nuisance-regressors "${REGRESSORS}" \
    --fs-license-file="${FREESURFER_LICENSE}" \
    --custom_confounds ${FMRI_PATH}/derivatives/custom_confounds_for_xcpd_ortho \
    --bids-filter-file "${LAB_SHARED}/CFG/code/Automated_Pipeline/functional_pipeline/xcpd_bids_filter.json" \
    --cifti

Version:

0.7.0 (a little outdated, I know)

Environment (Docker, Singularity / Apptainer, custom installation):

Singularity image running via Apptainer.

Data formatted according to a validatable standard? Please provide the output of the validator:

Data is in BIDS standard, but sue to possible misformatted directories, I decided to skip BIDS validation.

Relevant log outputs (up to 20 lines):

N/A

Screenshots / relevant information:

When I load the subject’s FreeSurfer output in freeview, the pial and white-matter surfaces look perfectly fine, suggesting that it’s not an issue of FreeSurfer’s surfaces:

image
image696×414 88.1 KB

Note that I’m using an older version of XCP-D since I followed this example for using custom confounds. Newer versions of XCP-D use a different syntax for that.

2

Hi @roeysc,

Please use the software support category in the future and edit your post to add your command and version to the post template.

Thank you,
Steven

1 Like
3

Hi @roeysc,

May you please use a more recent version of xcpd?

Best,
Steven

4

Of course. I repeated the analysis with XCP-D version 0.10.6 but the result stayed the same. The surfaces on top of the T1w images are still misaligned.

Any help would be greatly appreciated!

5

What fMRIPrep version was used for preprocessing?

6

It was fmriprep version 25.1.4.

7

I’m not sure what the problem could be. We did see this error in an older version of XCP-D, but that bug should be fixed in the current version. Are you able to share your fMRIPrep derivatives (or perhaps just the raw data from this run)?

8

Thanks, @tsalo. Certainly. It might be easier if I share with you a minimal portion of the raw data from this run. Perhaps sharing it privately with you by Drive is easiest?

In the meantime, I am rerunning fMRIprep and XCP-D from scratch to make sure no derivatives remain from previous attempts.

1 Like
9

Dear all,

While I thought I was running XCP-D 0.10.6, I in fact ran version 0.7.0.

I have now ran version 0.11.1 successfully.

In this version it seems that the summary report does not include any visualization of the anatomical surfaces, so I wasn’t able to verify that the original issue had been resolved. Still, all other outputs seem fine.

Many thanks for your help,
Roey

1 Like
10

The surface visualization is still available in 0.11.1. You just need to run with --file-format cifti and --abcc-qc enabled.

1 Like
11

I can confirm that after @tsalo has made some changes in the unstable version 0.11.1, the surfaces are now perfectly aligned. Many thanks for solving this bug, @tsalo!

2 Likes