Hello everyone,
I have a question about making task-fMRI based comparisons *across* runs:
I am interested in comparing BOLD response for correct versus incorrect trials between "early" and "late" learning. Early learning comes from run1 (and run 3) and late learning comes from run 2 (and run 4). Intuitively, I think I just need to set up my first level design files such that I would have onsets files that correspond to correct_early, correct_late, incorrect_early, incorrect_late, with the implication of course that there will be no time points for late conditions in runs 1 and 3 and vice versa for 2 and 4 (so I would set those to empty; all zeros when specifying onsets). I would then set up a contrast that would entail (early_correct-early_incorrect - late_correct-late_incorrect).
Is this correct or is there something fundamentally wrong with setting up contrasts in which a conditions are derived from different runs?
Thanks,
Dani
|