Hello, I am trying to run a cortical thickness analysis to look for an interaction between two continuous variables using mri_glmfit. I created an fsgd file including only the two original CVs to investigate main effects and obtained the output. When I added the interaction term (product of the two CVs) as a third continuous variable in my fsgd file and tried to run a dods analysis, I received the message: "ERROR: matrix is ill-conditioned or badly scaled." I don't understand how this could be the case. Is there a simple way to check whether my design matrix is ill-conditioned?
Susie
This can happen if the covariates have a very different scale (eg, age in years and something else in the range of 1-2). The product may have a different range as well. The product may be highly correlated with the first two CVs
doug
Susie Heo wrote:
Hello, I am trying to run a cortical thickness analysis to look for an interaction between two continuous variables using mri_glmfit. I created an fsgd file including only the two original CVs to investigate main effects and obtained the output. When I added the interaction term (product of the two CVs) as a third continuous variable in my fsgd file and tried to run a dods analysis, I received the message: "ERROR: matrix is ill-conditioned or badly scaled." I don't understand how this could be the case. Is there a simple way to check whether my design matrix is ill-conditioned?
Susie
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Douglas N Greve -
Susie Heo