Thanks Gang!

> Are you looking for the interaction effect between
> conditions and days?

It's exactly the interaction we're interested in. We have reason to think that the variability of that estimate may vary on Day 2 (sleep deprivation) so 3dMEMA was idea.

> Day1( A - B ) - Day2 ( A - B )

So if we have stimulus files for Day1A Day1B Day2A Day2B do we just load them in separately and specify the contrast as something like +1_Day1A -1_Day1B -1_Day2A +1_Day2B ?

Final two clarifications:

1) Do we need to add any columns to 3dDeconvolve for Day 1's Runs vs Day 2's runs or will splitting the stimulus files out be sufficient?

2) Are contrasts for Day1 and Day2 individually still valid from this larger model? That is, if we wanted to extract independent beta-weights for the individual days, in addition to the interaction, do we need to revert to the original models with them modeled separately?

Thanks again!
Jared