Dear Luciana,
for any covariate, SPM expects one covariate value per input image. Thus, the age vector should have as many elements as you have images. My guess would be that you enter more than one image per subject, but only one age value per subject. The best way to avoid this would be to compute all within-subject contrasts on first level and only take one image per subject to the second level (maybe you need more than one second-level model then). This will also help you in separating within-subject and between-subject error variance terms correctly. If by any means you can't avoid a factorial design, you will have to replicate the age values such that each image of the same subject has the same age value.
Best,
Volkmar
|