afni.nimh.nih.gov

Hello Zhihao,

The program catenates all the input to create a feature vector for each voxel. What that feature vector is depends on what you have in F1, F2, etc. The catenation of the input is simply a convenience feature, if your feature vector is already in one dataset (a time series perhaps) then you will have just that one dataset as input.

cheers,
Ziad