Whats the difference between an epoch HRF and an event HRF?
If you’re using AFNI/BrainVoyager/SPM2, there isn’t any. So don’t worry about it. But if you’re using SPM99 or earlier, the story above about the canonical HRF is actually oversimplified. In SPM99 (and earlier), epoch-related and event-related studies had the same underlying design matrix form – the onset of a trial was marked with a 1, but the actual trial itself was all 0’s. Event-related studies were simply modeled by convolving those with a canonical HRF, but epoch-related studies clearly needed to account for the length of the trial. So there was a separate, epoch-related, canonical HRF, that was also based off a mixture of gamma functions, but was specifically scaled to account for the length of the trial – so the HRF that was convolved with the design matrix was different for a 12-second epoch and a 30-second epoch. The epoch HRF generally looked like a wider, fatter canonical HRF, and represented a model of linearly summed HRFs over the course of the trial. With a re-vamping in
