> From [log in to unmask] Mon May 17 17:19:28 1999
> Date: Mon, 17 May 1999 12:23:13 -0400 (EDT)
> Subject: Two temporal basis functions
Dear Irena and Walter,
> I have some questions regarding the two temporal basis function
> approach. We have two experimental conditions and one baseline
> condition presented in the following manner:
>
> Run1: experimental condition A and baseline C
> Run2: experimental condition B and baseline C'
>
> There are 6 images in each experimental block and 4 images in each
> baseline block.
>
> We would like to compare the response function between condition A and
> condition B. However, because each condition is in a separate run, we
> are not sure which contrast to use. Which of the following contrasts
> would be correct:
>
> A C B C'
> 1 -1 0 0 -1 1 0 0
> -1 1 0 0 1 -1 0 0
>
> or
>
> 1 -1 -1 1 -1 1 1 -1
> -1 1 1 -1 1 -1 -1 1
Both are 'correct'. The first contrast[s] is effectively an (early
vs. late) x (A vs. B) interaction whereas the second is the three way
interaction (early vs. late) x (C vs. not C) x (Run1 vs. Run2). I
think the second addresses your question more comprehensively because
it evaluates the differential reeponse profiles in A and B relative to
some reference condition (C). This may remove nonspecific differences
(due to differences between C in Run1 and Run2).
I hope this helps - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|