If I understood Sjin correctly there's only one group, so rather Within-subject ANOVA I'd say ;-)
Concerning rest, well, I'm not sure. What are the expectations? If the timecourse consisted of box-car functions only, then one would indeed overfit the data. But the boxcar functions are convolved with the HRF. Thus, an extra regressor for rest might be closer to the actual timecourse (?).
|