In order to normalize a brain with important bilateral lesions, I am trying to segmentate it with CAT12, and as seen in this topic thread, CAT12 - badTissueContrast - beginner tutorial, I also get this error.
CAT Preprocessing error for anat_minus_lesion: ------------------------------------------------------------------------
Bad tissue contrast (BG=0.02, CSF=0.35, GM=0.71, WM=0.52):
This can be the result of (i) an improper SPM segmentation caused by
failed affine registration, (ii) improper image properties with low
contrast-to-noise ratio, or (iii) by preprocessing error.
Please check image orientation and quality.
To summary what I did: I drew the lesions on MRIcron, then I substracted these lesions to my anatomical volume using ImCalc on SPM12 (so all voxels included in the lesion map are = 0), then I tried to segmentate using CAT12. I think the error comes from the fact that a large portion of the brain is now equal to 0, but I am wondering whether CAT12 is ok to work on lesioned brain? If so, should I put another value to voxels from the lesion?
My cat12 version seems to be CAT12.7 (seen in the spm_cat12.m script).
I saw in the other post that someone pinged CAT12 dev (@CGaser), so I do the same, please excuse me if it is not the right way to do,
My cat12 version seems to be CAT12.7 (seen in the spm_cat12.m script).
Please start by updating CAT to the latest version (CAT=12.8 r1932, SPM=r7771).
Brainstorm can now manage SPM and CAT as plugins. If you are running CAT from Brainstorm, please remove them from your Matlab path, then let Brainstorm install and update them automatically. https://neuroimage.usc.edu/brainstorm/Tutorials/Plugins
If you still get the same error, hopefully @CGaser will be able to address your questions.
I have updated my cat toolbox, and try again to segment and normalize my scan.
First with the original anat scan, and the result is not good at all, with suppression of the posterior cortex which is lesioned, and the ventricles joining the air outside the brain.
Second, I have tried with the anat scan minus the lesion (lesion voxels = 0), and i get the following warning :
