Hi Meike, 

Maybe I can help you. If you have two groups, and two sessions, you can specify your contrasts as described in the CONN manual. 
For instance, if you are interested in the effect of drug across your groups, you can select 
group1, group2 (0.5 0.5) and placebo, agent (-1 1).

If you are interested in the between-group difference after the agent, you could select
group1, group2 (1 -1) and agent (1);

typically, however, you'd want to investigate possible group (group1, group2) x drug (placebo, agent) interactions first, by selecting the contrast
1 -1, -1 1, 

then test for main effects (of group and of drug, e.g. as the drug effect specified above ), and then test for further post hoc tests (e.g. as the second contrast specified above). 

Hope this helps!