Hello everybody,
I'm very new to VLSM analysis, so  sorry for the silly questions:

- Assuming that I have a sample of stroke patients with specific lesions (T1 AND Flair .nii sequences) do I have to normalize every segmented lesion from each patients (saved as .nii file) to a MNI space before to create a .mat file to be used by Niistat?
- If so, can I use normalize-segmentation function of clinicaltoolbox fir SPM12 to do such a normalization?

Thank you very much for your help.

Bye

Paul

Yes, I would use the Clinical Toolbox to warp each image to standard space

1. Use the clinical toolbox to warp the lesion to standard space
   https://github.com/neurolabusc/Clinical

2. You can then use the nii_nii2mat function to convert the NIfTI format lesion to the mat format
  https://github.com/neurolabusc/NiiStat/b...
e.g.
  nii_nii2mat('lesionP11.nii','lesion','P11.mat');

3. Note that NiiStat will use EITHER the NIfTI lesion file or the MAT file. So you can actually skip step 2. However, if you use the .mat file the software will cache region of interest values into the file, which makes subsequent analyses a bit faster.

4. If you do region of interest analyses, you should make sure that the files in the ROI folder are in the same space as your images. By default, these assume SPM12's clinical toolbox was used. This creates brains with a shape influenced by the average brains for older individuals (who have larger ventricles than young adults). You will want to use different ROI templates if you normalized to the default SPM12 template (which has smaller ventricles) or to an MNI template (the MNI template is larger than the average brain), as shown in Figure 1 of
  https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6050588/

5. For voxelwise analyses, you just need all brains warped to the SAME template, but you do not need to worry about matching the images in the ROI folder.