I noticed on the GLM wiki there is only an example of additional covariate with unpaired t-test. I am wondering can I add an additional covariate such as cortisol level to the paired-ttest. The variable is different for each subject and each session, so I believe it won't cause a rank deficiency problem. 

Can anyone tell me whether the repeated measure with covariate will be valid or not?

Thanks,

Wen