I've just uploaded a new Python script that does random permutation analyses using AFNI. It can handle 3dttest as well as various flavors of 3dANOVA*. I haven't tested it very much on the more complicated multi-factor analyses though (this is version 1.0), so let me know if you have any problems. The page is here: [
kurage.nimh.nih.gov]. It was originally designed for MEG data (specifically SAM volumes) but should be general enough to be of wider interest.