Hi there,
> Dear all,
> in my study I had the following design (simplified):
> baseline -> Task1 -> Task2 -> Task3
> After having performed the simple comparisons, I'd like to perform
> following comparison:
> (Task3 - baseline) - [(Task1 - baseline) + (Task2 - baseline)], which
> simplifies to the equation
> Task3 - Task1 - Task2 + baseline
> How do I set the contrasts for a one-sample t-test? I tried following:
> [1 -1 -1 1] and ended up with some weird activation patterns... Is my
> contrast correct and if not, what would be the correct one?
> Thanks alot already in advance!
your contrast is correct ; however if some voxels have negative values
relative to baseline that might influence the results in a way you don't
want.
ideally, you wouldn't model baseline see
http://www.frontiersin.org/Journal/10.3389/fnins.2014.00001/abstract
[Misconceptions in the use of the General Linear Model applied to
functional MRI: a tutorial for junior neuro-imagers] - this way you
would simply have to do [2 -1 -1] as contrast .. which could give a
different result (but not necessarily see ref )
Cyril
--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.
|