Dear Antoine,
You can compute the interaction term by hand:
(A-mean(A)) .* (B-mean(B))
and add it as an extra regressor in the design matrix, next to A and B.
Best regards,
Guillaume.
On 15/06/2021 12:38, Rogeau, Antoine wrote:
> Dear experts,
>
> I am doing a correlation VBM analysis.
> I tested for molecule A and radiotracer B levels as covariates in two different multiple regression analyses.
> Now, I wish to look for an interaction between them (possibly a synergistic effect), but I am stuck with how to build the model. I thought, perhaps, using a multiple regression - include both as covariates and set a +1 +1 contrast, but I am doubtful about this.
> How would you suggest to build my design matrix to look for GM regions that are correlated with molecule A - radiotracer B interactive effect?
>
> Best regards,
> Antoine Rogeau
> King’s College London
>
--
Guillaume Flandin, PhD
Wellcome Centre for Human Neuroimaging
UCL Queen Square Institute of Neurology
London WC1N 3BG
|