Hi
Your contrasts all look OK - in the design matrix you then have to mdel all anova cells.
hth
christian
On 11 Oct 2011, at 16:06, Alberto Inuggi wrote:
> hi all.
>
> i performed dual regression analysis on resting state data. now i would like to correlate the strenght of my networks with behavioural data (reaction times) structured in more than one factor.
>
> that is, i would like to find correlations with the main factors and also investigate their interactions.
>
> assuming factor A with 2 levels and factor B with 3.
> is it correct such design ???
>
> A1B1 A1B2 A1B3 A2B1 A2B2 A2B3
>
> pos A 1 1 1 0 0 0
> neg A -1 -1 -1 0 0 0
> pos B 0 0 0 1 1 1
> neg B 0 0 0 -1 -1 -1
>
> pos A1B1 1 0 0 0 0 0
> neg A1B1 -1 0 0 0 0 0
>
> A>B 1 1 1 -1 -1 -1
>
> with the last contrast testing which voxels are correlated more (in terms of correlation slope) with A than with B ???
>
>
> thanks in advance.
> regards
> Alberto
>
|