Hi all

a hopefully simple question:

if I use 3dTshift with the default interpolation to the middle time point, and then I estimate an iresp with 3dDeconvolve, the time zero of the resulting iresp waveform should be interpreted as lying half a scan past the actual stimulus onset (assuming the stimulus was delivered at the onset of the scan)?

Or, conversely, to have the iresp correctly time-locked to the stimulus onset one should enter a stim_file vector where the stimuli occur half a scan earlier than their original scan number, and use stim_nptr=2?

I hope this is not too confusing, thanks for any comment!

giuseppe