Hi, you can try the concatenation approach hth christian On 14 Jan 2008, at 21:33, lingling pu wrote: > Dear FSLer, > > It is said that for multisession tensor ICA, each subject needs to > have consistent stimulus paradigm. Does that mean that if I have > randomized stimulus design, the group > data I obtained cannot be ana > lyzed using tensor ICA? If th > at is the case, how should I do group > analysis? I can think of a way that > is doing single session ICA on individual data, then cross > correlation the IC maps between subjects. Is that the only way to > find common (group) effects? > > Thanks a lot! > Lingling > > Never miss a thing. Make Yahoo your homepage.