Hi Gang,
I just saw your message, sorry!
Using 3dWilcoxon worked rather well, i.e., the results make sense (and I may feel inclined to say that they are beautiful) with respect to location and direction. Therefore, I believe the sporadic but extreme outliers were addressed while the results are robust to show the expected patterns.
Do you have more specific questions? I would gladly provide more information!
In contrast, using 3dMann-Whitney for the between-group comparison seems a bit conservative (or the effect is just not there though 3dMW also did not show a significant effect in another study where 3dttest++ and 3dMEMA showed significant effects). Are there other non-parametric tests for between-group comparisons in AFNI? I could not find anything suitable.
Thanks!
Nic
