Dear Dimitri,
Well, I wouldn't be surprised to find *some* differences if all stimulus onset functions are shifted by one time bin, which should become more relevant with shorter durations compared to long block designs. In any case, the current code + 33 should be prefered, as this way, the "on" time bins correspond to the expectations (leaving aside the issue with the additional one time bin via the tof setting).
The different default microtime settings would only lead to differences if you accidentally compare models set up with different settings (which was the case in some previous messages, that's why I wanted to highlight that aspect), if you have specified the same values when setting up models with different SPM versions then this won't have any impact.
Concerning defaults.mask.thresh, are you sure the 0.05 is the original code and not a manual adjustment, as the current 0.8 dates back to versions even before SPM5. However, I don't know whether there has been some SPM release version in between in which it has accidentally been set to another value.
Best
Helmut
|