Typically I carry out contrasts with 3dDeconvolve's with the gltsym option. I know I can conduct operations like 1 sampled t test and paired sample test with 3dttest but what if I need to combine sets? To illustrate what I mean by this lets say I have two tasks each with fast and slow trials. For each condition (TaskA_Fast, TaskA_Slow, TaskB_Fast, & TaskB_Slow) I have beta series correlation maps (fisher z transformed) for each of my subjects. One of my goals is to contrast TaskA and TaskB collapsing across speed (+1 AF +1 AS -1 BF -1 BS). What AFNI function can I use to do this? Also, I figure I can use one of AFNI's ANOVA (in the case described 3dANOVA2 since it has two factors) functions to get at this question (simple look at main effect of task).