Dear SPM Users,
I know variants of this question appear in the archives. However I still don't understand what I can practically *do* to help this problem, so I ask a new version of an old question.
In a design where trials making up two to-be-compared conditions by nature have different durations, the assumption is that the longer trials will have overall a higher HRF amplitude (although why / when this happens depends on use of impluse vs. epoch functions). If I compare the short trials to the long trials (the two trial types happen in separate blocks), is it likely that differences in signal change are an artifact of modeling assumptions regarding HRF amplitude? If so, which direction should I expect artifactual differences (short < long or long < short)? And finally, what strategy can I use to combat this possible problem.