I am working on an analysis for a study design on 'effort'. We are hypothesizing that neural activity related to effort will increase over time for our 'effort' events.
I am attempting to model this using parametric modulators, which I believe is the correct approach. But I have no direct experience using parametric modulators and thus would prefer to check my work with the list before showing results to co-authors.
We have 6 sessions per person. I am not interested in session x effort interactions at this time (due to the nature of the experiment), but wish to identify voxels where the neural activity to our effort events increases over time. So, I added a 1-st order time modulation. We actually have 2 event types, and I added a modulator for each, and ran the GLM using an HRF only model for now.
So my design matrix has 4 columns per run [type1, type1*time, type2, type2*time].
if I want to find voxels where activity increases over time for event type2, can I just run a contrast of 0 0 0 1? I think this is explicitly effect of time after removing task effects, and thus doesn't have the sort of issues (task-default switch, visual activation, etc.) with a main effects contrast of a task event (like 1 0 0 0).
Any advice and/or confirmation that I am on the right track would be appreciated.