Hello Paul:
I'm not sure what you mean by "collapse across individual subjects".
There are various 3dmerge options for clustering and counting active voxels
across subjects.
If you want to perform a statistical test across subjects at each voxel
location, you should perform this analysis BEFORE clustering. There are
numerous AFNI programs for across-subject statistical analysis, including
3dttest, 3dANOVA, 3dANOVA2, 3dANOVA3, 3dRegAna, 3dMannWhitney, 3dWilcoxon,
3dKruskalWallis, and 3dFriedman. Clustering can then be performed on this
final statistical result. Program AlphaSim may be helpful in setting the
combination of prob. threshold and cluster size threshold to obtain the
desired overall Alpha level.
Doug Ward
