It sounds like at least a couple of people on here are using the pial surface in their surfclust commands. However, I've noticed the interactive clustering in suma uses the smoothed wm surface. This seems to make a significant difference for the size that the clusters are estimated at, thus changing which ones get kept.
I'm not seeing a slow_surf_clustsim.py option that specifies which surface the estimates are for. Is there a preferred surface to use in SurfClust?
Thanks!
Walker