It looks like you want (PF+PM)-(CF+CM) which would correspond to [0 0 0 0 1 1 -1 -1]/2. This just averages the slopes for F&M in each group, then takes the difference between the average slopes.

doug

Katherine Reiter wrote:

Hey freesurfers,

I'm interested in running an interaction between a continuous and categorical variable, covarying for another categorical variable. I understand that this analysis is not possible through the QDEC GUI so I am working through the command line group analysis and setting up the contrast matrix.

Discrete Factors: Group with 2 levels: Patient and Control Gender with 2 levels: Male and Female

Continuous variables: Age

Meaning, I have 8 Regressors: 1: PatientMale intercept 2: PatientFemale intercept 3. ControlMale intercept 4. ControlFemale intercept 5: PatientMale age 6: PatientFemale age 7. ControlMale age 8. ControlFemale age

In order to look at the interaction of group and age, covarying for gender, would the matrix be:

0 0 0 0 -.5 .5 -.5 .5

However, this matrix seems looks more like a difference in slopes comparison and less like an interaction. Any help would be much appreciated!

Thanks, Katie

---

---

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