Is there a particular way 3dANOVA2 calculates a mean that we are not aware of?
-amean k in 3dANOVA2 should be basically the average of the k-
th coefficients from all 12 subjects. I can't see anything special here. However, I don't understand why you had to use 3dROIstats to get those coefficients from those 12 subjects for each of the 4 voxels. Am I missing something?
Gang