[Simnibs-discuss] charm head modelling problems

[Vincent Rochas]

Dear simNIBS users and programmers

I have used simNIBS previous versions with very good results and reasonable simulations for tCS. I was using headreco function with two MRIs T1 and T2, converted from Dicom to Nifti using SPM12.
I am now trying the new version of simNIBS (4.0) using the charm function and experiencing some troubles.
- The MRIs I was previously using are apparently not compliant with the needed standard in term of orientation (?) and there was a complain about sform and qform annotation, so I was advised to use –forceqform by the charm function which I did.
- After using the –forceqform option it runs the segmentation. However the resulting .msh file in the m2m folder corresponds to another face and head than my patient MRI!
I tried to re-run the ernie segmentation (without needing the –forceqform option). The command line was like: charm ernie ernie_T1.nii.gz ernie_T2.nii.gz The results is again this same face wrapped on ernie’s shape I guess. Or Ernie’s head wrapped differently I don’t know.  In the case of the patient it is surely not the head and face of the patient at all, and this same head though. Also the segmentation of the cortex doesn’t look very precise, sulci are not well delineated.
I then checked the log.html and I noticed a difference with the examples provided by simNIBS (the prototypic ernie segmentation). On the DEBUG line I have 'init_type': 'atlas' instead of 'init_type': mni' in the example segmentation.
I don’t know if it is the problem but might explain why it is differently sticking the atlas maybe on the patient’s T1 as a result.
That might be a trivial problem or an option that I missed. I would be happy to have your opinion.

Best,
Vincent

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.drcmr.dk/pipermail/simnibs-discuss/attachments/20221205/4023f00d/attachment-0001.html>