The matrix that you have will fail because Col1+Col2=Col3 (and same for the other columns). To do what you want, use the DODS that mri_glmfit generates along with a contrast of [0 0 1 -1]. The constant term is modelled by the first two columns. doug

On 02/12/2013 03:19 PM, jm wrote:

Hi fs experts,

I need to execute *mri_glmfit* with -X option to probe interactions between Diagnosis and Age,

Statistic Software help suggest: The*homogeneity of slopes designs* can be used to test whether the continuous and categorical predictors interact in influencing responses, and thus, whether the traditional ANCOVA design or the separate slope file:///textbook/general-linear-models/#between_subject design is appropriate for modeling the effects of the predictor.

for 2 groups

thickness = b_0 + b_1 X_1 + b_2 X_2 + b_3 * Age + b4 X1*Age + b5 x2*Age

the Matrix Design Xm _corresponding

1 1 0 56 56 0 1 1 0 67 67 0 1 1 0 67 67 0 1 1 0 78 78 0 1 1 0 67 67 0 1 1 0 87 87 0 1 0 1 67 0 67 1 0 1 66 0 66 1 0 1 65 0 65 1 0 1 87 0 87 1 0 1 45 0 45 1 0 1 77 0 77

In matrix design generated by freesurfer _*file Xg.dat*_ for both DODS and DOSS the first column don't appears. Constant term "is not explicitly considered" my question is:

is correct to use *mri_glmfit* with -X Xm as above specified ? or I need to eliminate the first column?

1 0 56 56 0 1 0 67 67 0 1 0 67 67 0 1 0 78 78 0 1 0 67 67 0 1 0 87 87 0 0 1 67 0 67 0 1 66 0 66 0 1 65 0 65 0 1 87 0 87 0 1 45 0 45 0 1 77 0 77

thanks!!

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer