search

MICA-MNI / micapipe

micapipe from the Multimodal imaging and connectome analysis lab (http://mica-mni.github.io) at the Montreal Neurological Institute. Read The Docs documentation below
http://micapipe.readthedocs.io
GNU General Public License v3.0
80 stars 27 forks source link

MPC surface ordering distorted #52

Closed rb643 closed 1 year ago

rb643 commented 1 year ago

When running the micapipe container with the MPC module it appears that for some subjects and sessions my surface are not ordered correctly resulting in wrong lateralised correlation matrices, see the attached screenshots of the intensity profiles and resulting correlation matrices for two sessions of the same subject. This does not always happen in the same manner (i.e., the misordering of surfaces isn't consistent in the same way across subjects).

My suspicion is that the surfaces are not created in the proper temporal order (or at least something is failing in line 158 of 03_MPC.sh). Unfortunately as I am running the containerised default version all individual surface files are removed so I can't check that assumption without running the code step by step.

If there is an easy fix that could be plugged into the already run data (i.e., all freesufer and other registrations look totally fine and I would hate to rerun) then that would be much appreciated.

Screenshot 2022-08-23 at 11 53 17

rcruces commented 1 year ago

I believe that this issue could be due to bad registrations, we also found this in some cases in some datasets. This issue was fix in the version 0.1.3. We are currently building a new container with the FIX.

Thank you for your comment

rcruces commented 1 year ago

I've seen something similar before and I think is an issue more related with the registration between the qMRI and the freesurfer space. could you please check the registration of sub-01_ses-02 between: ${subject_dir}/anat/sub-01_ses-02_space-fsnative_desc-micro.nii.gz and ${subject_dir}/anat/sub-01_ses-02_space-fsnative_t1w.nii.gz

where: subject_dir: is the path to the micapipe/sub-01/ses-02

rcruces commented 1 year ago

This issue was fixed in the v0.1.5 and registrations were improved and optimized for better performance in v0.2.0.

We invite you to use the newest version of the pipeline: https://github.com/MICA-MNI/micapipe/releases/tag/v0.2.0