Hi guys,
I did an ANOVA with GroupAna and found a nice significant cluster showing a Group ME. I created a mask which looks nice and noted the Olay cluster # for my planned extraction. I used the mask to extract with 3dmaskave this cluster from each subject's GLM outputs (normalized same as ANOVA results). However, when I generate the average 1d files for each subject and throw the numbers into an outside stats program, the group difference is nowhere close to significant, even though the p-value from the AFNI GUI shows a sig cluster at p=.0049.
I double checked my subbriks that went into GroupAna and the ones in my individual subject data. I double checked the contrasts that I set up in GroupAna.
Maybe there is still something I haven't checked? I would be grateful for any advice.
Thanks much,
Desmond