Error in the coregistration of B0 to T1W

CARLOS_LEANDRO_SILVA · February 21, 2024, 12:57pm

Summary of what happened:

When using QSIPrep, there’s an issue with coregistering my DTI image and the T1w. What can I do to resolve this? I’ve already tried several combinations of the options available for dwi-to-Anatomical coregistration, and the bet is being done correctly.

Command used (and if a helper script was used, a link to the helper script or the command generated):

qsiprep-docker /home/pisa/Desktop/DTI/bids_AP /home/pisa/Desktop/DTI/derivatives_75_AP participant --participant-label 01 --recon_input /home/pisa/Desktop/DTI/derivatives_75_AP --recon_spec dsi_studio_gqi --fs-license-file /media/pisa/0ABA196EBA195811/Ubuntu/license.txt --use-syn-sdc --write-graph --output_resolution 1

Version:

0.19.1

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

Docker

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

Yes

Screenshots / relevant information:

Steven · February 21, 2024, 1:41pm

Hi @CARLOS_LEANDRO_SILVA,

Looking at that screenshot makes me wonder if there isn’t anything cropped off the raw DWI image. Do you have PA scans such that you can do better SDC?

Also --recon_input /home/pisa/Desktop/DTI/derivatives_75_AP is only needed if you are only doing recon (combined with --recon-only).

Best,
Steven

CARLOS_LEANDRO_SILVA · February 23, 2024, 5:01pm

Hi, @Steven!
I did what you suggested, but the error still persists. Upon examining the ‘Workdir’, the native space is perfectly fitted.

Steven · February 23, 2024, 5:08pm

Did you look at the raw DWI image? Provide a screenshot?

CARLOS_LEANDRO_SILVA · February 23, 2024, 8:22pm

Hi Steve!
Sorry for not being in the lab at the moment. As soon as I’m there, I can send the screenshot. But I wanted to mention that I took a look at the RAW data and the DTI fits perfectly into the T1w image in native space.
Additionally, the FA map has a part of the skull (bet was not done correctly), but apparently everything is okay with the DTI.
Best regards!

CARLOS_LEANDRO_SILVA · February 23, 2024, 8:35pm

Could this be a concatenation issue? Because I’m working with a subject having two DWI (multishell) files.

Steven · February 23, 2024, 8:43pm

Do you concatenate them in QSIPrep or afterwards? I see you have a BIDS AP, which implies you might be separating your data, which would not be recommended.

CARLOS_LEANDRO_SILVA · February 27, 2024, 5:26pm

Hi, @Steven !
This is my bids directory:

I thought QSIPrep automatically concatenated through the BIDS organization. What can I add to my code line so that QSIPrep concatenates?

Steven · February 27, 2024, 6:24pm

Do you have a bids_PA too? It sounded like you might have been separating your AP and PA scans.

CARLOS_LEANDRO_SILVA · February 27, 2024, 6:26pm

Yes, I have both AP and PA scans. With the PA scans, the processing also yielded a poor coregistration result.

Steven · February 27, 2024, 6:37pm

You should have a single bids directory containing both your AP and PA scans (with the dir-<> label added to the filenames).

Best,
Steven

CARLOS_LEANDRO_SILVA · February 27, 2024, 6:41pm

Thank you for letting me know. I’ll attempt to process it accordingly.

CARLOS_LEANDRO_SILVA · February 28, 2024, 5:03pm

I ran qsiprep again as instructed, but I’m still encountering an error with image registration. I also attempted to perform BET on the DWI images prior to execution, but all my attempts yielded the same outcome.

Command used

qsiprep-docker /home/pisa/Desktop/DTI/bids /home/pisa/Desktop/DTI/derivatives participant --participant-label 01 --recon_spec dsi_studio_gqi -w /home/pisa/Desktop/DTI/work --fs-license-file /media/pisa/0ABA196EBA195811/Ubuntu/license.txt --anatomical-template MNI152NLin2009cAsym --write-graph --output_resolution 1