Dear Tobias,
> I have just a little question concerning the design of a fMRI-study.
>
> We have four conditions (A,A',B,B') which are examined in one session per
> condition with alternating on/off epochs, 4 rest/avtivation cycles per session,
> 40 sec per epoch.
>
> We are interested in the differences between A / A' and B / B'.
>
> I am pretty new to the field, so I am not 100% sure about how to define the
> correct contrasts for those questions.
>
> I want to define one big design matrix consisting of all four conditions by
> modelling the boxcar-function convolved with the hrf for every single condition.
> My understanding is that (1 -1 0 0) and (0 0 1 -1) would answer the question
> about the differences.
>
> Is this correct ?
Yes, the contrast is fine, but it's a one-sided test, i.e. with [1 -1 0
0] you're asking, where A causes more effect than A'. Similarily,
contrast [0 0 1 -1] tests, where B larger than B' (given that the order
in the design matrix is A A' B B').
>
> And if so, what is the difference to (-1 1 0 0) and (0 0 -1 1), respectively ?
Here you're asking (as it follows now from above), where A' larger than
A. (B' larger than B).
Stefan
--
Stefan Kiebel
Functional Imaging Laboratory
Wellcome Dept. of Cognitive Neurology
12 Queen Square
WC1N 3BG London, UK
Tel.: +44-(0)20-7833-7478
FAX : -7813-1420
email: [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|