Dear freesurfer experts,
I have a question about the conceptual use of mri_glmfit-sim. I wish to apply permutation based cluster-correction to my between-groups results (patients vs controls), and it is not clear to me which volume should be the input for this type of simulation.
For example, the acceptable input for mri_glmfit-sim appears to be only the ".ces" map, which is not present in my between-groups directory, but only in my lower-level one-sample group means (osgm). This raises the question of whether to run the permutation algorithm 1) one time on the .ces map that results from the total sample OSGM (i.e. patients + controls), or 2) separately on OSGM for patients and then controls. Probably I am mistaken but neither seem to be conceptually correct, as I am mainly interested in the cwp for clusters emerging in the between-groups contrast.
Any help would be greatly appreciated!