I have got good results with Melodic FSL for my two group, although it is a small sample (14)....but, I suspect about the age is a very strong variable I would take into account because, althought I don't find statistical differences between the two groups....the main group (n=10) is younger than the smoller group (n=4) and I think there aren't no differences because of the size of the two samples. So, I have thouhgt about covariate for age with melodic. I have introduced a GLM with 3 contrasts....but...I am doing something wrong because it appears a message of warning about my matrix design....I am not sure if the problem is in the matrix of contrast itself or because I have introduced the age directly in the third EVs. I have been reading about GLM in FSL web but there must be something I have missed or not understood well. Maybe the question is too basic, but I am naive, sorry. Could anyone help me? 1 (group one) 2 (group two) 3 (age) Contrasts Matrix: Contrast 1: 1 -1 2 Contrast 2: -1 1 -2 Contrast 3: 2 1 -1 For Events in the third columm I write the age directly...maybe this is the problem? Many thanks for advance. Rosalia.