Thank you for your responses Peter and Rick.

Just to clarify, the processes would be the following:
1) Take the desired MNI template and run through freesurfer's autorecon process for generating surfaces and template sphere.
---- Do we need to (or should we ) use MapIcosahedron for anything at this step?
---- I assume this would be optional for MNI templates but then again I could be wrong!

2) [afni.nimh.nih.gov]

Based on the link above I assume that the standard template sphere is the one we created in step 1 and is output directly from freesurfer.
-----Is this correct?
---- Do we need to use MapIcosahedron for any reason at this stage?

3) We can then create a template surface in subject space (similar to inverse transform of the MNI template to subject space)
and run group analysis.


Additional Questions:
a) Does SUMA support CIFTI-2 files?

b) When smoothing data are we able to combine volumetric and surface smoothing in one step (similar to human connectome project) where surfaces are smoothed with one type of kernal (not across gyri) and volumetric as another type of kernel? I want to use the benefits of volumetric smoothing and surface constraints for localization.

c) Say I want to warp the subjects T1 image to the desired MNI template using 3dQwarp. Can I use that warp field in combination with information from my surface mesh (transforms from subject surface space to a template surface of the same MNI template as indicated in step 2 above) such that my surfaces are more accurately warped using the surface transform while I still use 3dQwarp for volumetric warping?




Thanks,
Ajay