Dear fsl community,
I have a very simple question for you, but unfortunately I wasn't able to solve my doubt by reading other threads. I have set up the tbss pipeline for my DTI experiment, which analized data from one group only, containing 35 partecipants. My main concern was to evaluate correlation between FA (and other indices) and some cognitive scores I have for the same subjects. Correlation is ok, with the simple Glm model having its only EV to be the cognitive score itself, demeaned, for each subject. I was now trying to introduce age as a further regressor, and I tried to add it as a second EV, demeaned as well. However, I get the following error message from the Glm GUI
"Problem with processing the model: Warning: at least one EV is (close to) a linear combination of the others. Need to alter your design. (Design matrix is rank deficient - ratio of min:max eigenvalues in SVD of matrix is 0) Contrasts involving these combinations will be set to zero."
I am not able to resolve among these possible causes for this error message:
-was I wrong to demean age EV to? It looks like I should from the Glm guide...
-since the age span of my group is really narrow, most of the participants show the exact same age...could this be the problem?
-should I interpret this warning message as: "age and your cognitive index are co-dependent"?
I would really appreciate any help.
Kindest regards,
Eleonora Mastrorilli
|