hi Agnieszka
> I have the following conditions in my design matrix:
>
> Words, Pseudowords, Fixation Cross (control condition),
>
> I assume that the t contrast: 1 0 -1 gives me the Word effect (relative to
> the control condition, this is with the activation due to low level visual
> processing subtracted out).
Yes.
> Is it possible to specify a t contrast: Words > Pseudowords (show me all the
> areas significantly more activated due to Words then to Pseudowords), but
> with the activation due to fixating on the Cross subtracted out? If yes,
> then how?
Actually, because the fixation condition would be the same, there's no
need—you would get the same answer as words > pseudowords (i.e.,
subtracting the same value from two conditions doesn't change how far
apart they are). Or:
words = [1 0 0]
pseudowords = [0 1 0]
fixation = [0 0 1]
words - fixation = [1 0 -1]
pseudowords - fixation = [0 1 -1]
(words - fixation) - (pseudowords - fixation) = [1 0 -1] - [0 1 -1] = [1 -1 0]
Hope this helps,
Jonathan
|