Hi Vitaly,
The confusion is perhaps caused by drawing parallels between SurfClust and 3dmerge. The more appropriate comparison would be to 3dclust.
At any rate, I have modified the help and added the options -out_clusterdset to have the program apply cluster-based thresholding on the input data. That is in addition to outputing a clusters table and an ROI type dataset that assigns to each node in a cluster the rank of that cluster.
I have also modified the help and I hope made it clearer. The option -out_fulllist has been the source of some confusion, see if the new help text clarifies its use.
Let me know how it goes.
Binaries are not ready yet, check for archives postdating this posting.
cheers,
-z