Question on group design matrix for FSL Feat

Hi, all,

I have some voxel-wise whole brain neural measures for each subject. My subjects were split into four groups. Each of them also came with a behavioural score (continuous variable). I am wondering how to set up the group-level design in FSL Feat to study the main effect of four groups, the main effect of behavioural scores, and the interaction between groups and behavioural scores?

I looked at the Feat manual, but get confused of how they set the design and contrasts (e.g. FEAT/UserGuide - FslWiki their dummy variable setting looks confusing). In addition, they only have examples for interactions between categorical regressors. In my case, should I standardize all my regressors before I multiply them for interaction terms?

Thank you very much!
Liwei