2nd level analysis

Dear brain reseachers

Please let me know about the 2nd level analysis.

I have two groups, patients and the controls.
I have done between group comparison by ROI-to-ROI analysis including age as a nuisance covariate.
This yielded multiple supra-threshold connections.  

I have neuropsychological tests results in the patients.  I want to know which connections are significantly associated with the test- score.
When I want to restrict analysis within the connections that showed significant changes in the group comparison, how contrast should look like?  Test -scores of controls are lacking, so I put 0 in test-score of the controls.

Covariate was defined as follows;

patient  controls  age  test-score   

Is contrast [ 1  -1  0  1] correct to identify connections particularly associated with the test score?

I would appreciate answer to this question.

Regards,

Ken Kazumata, Japan.