You are correct. This was a pain paradigm, and EV1 is the presence of heat pain (30s of 1s) and the absence of heat pain (30s of 0s). EV2 is a likert scale of pain perception- so that when EV1 was 'on', ratings were high, and when EV2 was 'off', ratings were low (or 0s). I think that this is why without orthogonalization, the activation basically went away entirely (because the GLM could not distinguish which EV to model).
|