Dear all i have a question about setting contrast

Is it statistically fine to contrast 1 rest scan before a training with 2 averaged rest scans after the training?

I have an additional question on covariates.

I have 2 groups (A and B) with pre and post sessions, i m interested in the post>pre. I inserted the score of a behavioural task (BT) for the pre and for the post and, then, I calculated the difference (difference_BT). Then I did the interaction between groups (A and B) and the difference_BT by selecting the 2 groups plus the difference_BT covariate and i selected the interaction option. In this way i obtained 3 covariates A*difference_BT, B*difference_BT and A*B*difference_BT.

Now, i m not interested in comparing the 2 groups and i m considering only within group analysis. To analyse the covariate, should I select all subject plus the covariate that is the interaction of my group of interest and the score, let's say all subjects+A*difference_BT or should I select the group of interest, let's say A plus the covariate of the same group and the score of the behav, i.e. A*difference_BT?

thank you