Thank you so much for this fast and elaborated answer!
The contrasts you specified are the contrasts I actually used in the first place. I used the results as a screening step (to identify the ROI-pairs that are significantly correlated with group or emotion), and then entered group+emotion+covariates into a single ancova model in spss using exported data from CONN (sorry for not explaining that in the original massage, I thought it was long and complicated as is), to compare the relative contribution of group and emotion (each after controlling for all the rest, i.e. group after controlling for covariates+emotion, and vice verca).
The issue is that in the "screening step", i.e. looking at CONN results, I use two different methods to identify the roi-pairs. I was advised to use the same method for both, even though the regression results should be the same as ancova results (i.e. regression of emotion and connectivity controlling for group and covariates should be the same as an ancova that takes into account emotion+group+covariates when looking the the contribution of emotion in predicting connectivity). Theoretically, I could have just entered all into a single ancova model in CONN and looked at the contribution of group and emotion separately (like I do in spss), but CONN gives a single value- the value for a specific covariate after controlling for all all others.

For this reason I would like, if possible, to find an ancova contrast that is equivalent to the regression contrast that you specified:
"b) contrast [0 0 0 1 0 0 0 0] to test for Emotion effects while controlling for all other effects (Group & covariates)"

Thank you very much for your help!

Liron.