Dear Eileen,

the error comes from the fact that you did not enter any second level covariate/variable. You need to at least specifiy one variable like 'All Subjects' an array containing as many ones as you have subjects so if you have 10 subjects you enter [1 1 1 1 1 1 1 1 1 1]. If not conn does not use the first level data.
Else your between-condition contrast looks fine to me.

Maybe also see in Alfonso's FAQ here: http://www.alfnie.com/software/conn about more details how to specifiy second level contrasts.

Best,
Julia