the difference between 0 duration and dur>0 is the function convolved with the hrf (stick for 0, boxcoar if >0). For all durations greater than 0, the duration specifies the magnitude, assuming a linear summation of activity. You should use 2 trs.
I have fast ER design, where each trial consists of 2 TRs. Some time ago it
was suggested here that I should define my duration as "2" (TRs). However, I
am confused because in SPM help window it is written the follows: "Epoch and
event-related responses are modeled in exactly the same way but by
specifying their different durations. Events are specified with a duration
of 0." Do they implicitly assume, that in ER design the event is always one
TR and I am OK, in my two_TRs_per_trial_case, with defining it as "2"? Based
on this logic, defining "0" or "1" should in general be exactly the same?
Thanks for help,