Maristela

> We have examined the relationships between a continuous variable and 
> gray matter in two separate groups (one separate correlational 
> analysis for each group).
> Is it possible to perform a correlational analysis in SPM2 including 
> both groups as a predictor variable? Is it possible to compare such 
> relationships in one single model (group x variable interaction)?
> We have used the "single-subject: conditions and covariates" model 
> and we have specified our t-contrast as 00-11 and 001-1 (matrix 
> attached), but we are not sure whether it is the right direction.

I'm not sure to understand what youm mean by including both groups as a predictor .. the design matrix you have is fine and already code both groups - so a contrast 1 -1 will compare the two groups taking into account (regressing out) the other regressors you put in the model. The contrast you performed 0 0 1 -1 is fine and compare the correlation coef. between the two groups.

Another way to code this is to enter 1 single regressor for the groups (possibly what you want). In this case simnply enter 1 for group A and -1 for group B - note that with the same matrix as your but the group coding I proposed above, results could be slightly different because you gain 1 df :-). For any interaction simply .* the group regressor (11111-1-1-1-1-1) by the continous regressors (i.e. 1st part is positive and the second part negative).

Hope this helps

Cyril