Dear Experts,
I am currently working with the non-linear FreeSurfer normalization. We have a small group of subjects N = 12 and in each brain we identified a single position with a fMRI localizer. The task is a group comparison of the individual results for this position. Therefore, the coordinates of each position in FSaverage was determined using mri_label2label. The radius of the region is an important quantity in my investigations. A smaller radius would help me a lot in my data processing. I tried to decrease the radius by generating a new group space template from my 12 subjects with make_average_subject. Then I used mri_label2label to determine the coordinates in the new group space template. Unfortunately, there was no difference in the radius. The intention behind this procedure was to achieve an improved normalization of the subjects. I have read much about this topic in the mailing list, but there are many details in the processing which I am not yet aware of. So, my question is if I proceeded the best possible way or if there is a better alternative to improve the normalization. Would it help to generate a new template as outlined in the second part on this web tutorial (chapter: Creating a registration template from scratch):
https://surfer.nmr.mgh.harvard.edu/fswiki/SurfaceRegAndTemplates
or do I need to use mris_register with my new template and my 12 subjects instead of mri_label2label in order to benfit from my new template as generated by make_average_subject (also outlined in the web tutorial in the chapter: Creating a registration template initialized with FreeSurfer template)?
Even a small decrease in the radius would help to improve my results.
Thanks in advance
pfannmoe