The ideal way to do this (most correct with respect to subject anatomy) might be to generate the ROIs on every subject (structural space or MNI space MSMAll midthickness, depending on how you want to deal with things like brain size variability), then average them together and threshold at 0.5. This way, it uses the exact distances on subject surfaces, rather than an approximation. You could also do this with a few dozen subjects instead of the entire set, to get a similar answer with less computation time.
We do have a way to approximate the original geodesic distances by using the average vertex areas, but that command doesn't currently have an option for it - it isn't that hard to add, though. Tim On Tue, Feb 14, 2017 at 4:12 PM, Chihuang Liu <[email protected]> wrote: > Hi all, > > I'm using FIX-Denoised MSMAll registered resting state fMRI data and would > like to define neighborhood structure of the surface vertices that's common > among a group of subjects (each vertex have the same neighbor vertices > across all subjects). I know that I can use wb_command > -surface-geodesic-rois to get neighbors of all vertices, but I'm not > exactly sure which surface mesh is most appropriate to use to compute > geodesic distances in this case. It appears to me that there are many > different meshes like sphere.32k_fs_LR, midthickness.32k_fs_LR, > midthickness_MSMAll.32k_fs_LR, etc. I'm not sure what's the difference > between them. > > Any help is greatly appreciated! > > Thanks, > Chihuang > > _______________________________________________ > HCP-Users mailing list > [email protected] > http://lists.humanconnectome.org/mailman/listinfo/hcp-users > _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
