Thanks Alfonso, that's very helpful!
Just to clarify, looking through the forum, I notice that, in some of your answers to posts asking how to do an MVPA analysis using a behavioral measure, you suggested to select All_Subjects and Behavior with the contrast [0 1]. What would lead me to do this rather than the contrast described below (Patients, Controls, Behavior Patients, Behavior Controls with the contrast [0 0 1 -1]? (both analyses give me completely different results)