Hi, I used new regressors from EEG features to build new GLM for every subject. I have got some common activation areas from each subject. But when I tried to use 3dttest+ (GUI) to do group analysis on these subjects, the broad common activation area were not shown and there were lots of sparse activation spread like noise. Was anyone faced with this problem before?
It’s hard for me to pinpoint the problem without access to the data. The simplest approach is to pick one interesting voxel, and extract the values at that voxel from all the subjects, and compute the t-statistic. Then you can check if 3dttest++ did the correct calculation.