Dear Adam,
there are probably several other ways as well, but if you're concered about memory load during the masking period I'd say "half trials" are a good solution. You would add trials in which the stimulus is presented, but after the mask only a fixation cross is shown, indicating that the trial is over / subjects don't have to decide. Half trials should thus require perceptual/attentional processes plus working memory ressources but no decision-related activity. You would then go with a regressor "stimulus" for each of the conditions, covering both half and complete trials, and another regressor "decision". Half/complete trials could also be combined with a jittered mask in between, but this has some drawbacks as you've already mentioned. Depending on the exact requirements for your study you could add some more control conditions (e.g. trials in which only the stimulus is but no mask, ...).
For ROI analyses you could also go with FIR and plot the timecourse for trials in which subjects have to decide minus trials in which they do not. In that case might want to use a regressor for half trials and another for complete trials, but no differentiation between stimulus and decision period on regressor level (as you would do so based on the two time courses for half/complete trials).
Best,
Helmut
|