Hello,
I've seen a few threads on how to set up a group * time interaction in the GLM, but I think I have a slightly different question: How would one go about finding a specific form of an interaction (e.g., ordinal vs. disordinal)? Or is that even possible within the framework of the GLM?
In my case, I have two groups that should not differ during a "pre-treatment" baseline condition but then one group should change during a "post-treatment" condition. I've considered excluding voxels that show group differences during the "pre-treatment" condition -- but I don't want to look at voxels based on the absence of an effect.
Thanks,
David
|