I am trying to examine the interaction of 2 factors (2 x 4 groups) and have difficulties to find the right contrast matrix for mri_glmfit.
We have patients and controls at four different aquisition sites and want to know if the changes (i.e. thickness) between patients and controls are equal in every site.
so we have 8 groups:
patientssite1
patientssite2
patientssite3
patientssite4
controlssite1
controlssite2
controlssite3
controlssite4
the contrast for differences between patients and controls (T-Test) would be:
1/4 1/4 1/4 1/4 -1/4 -1/4 -1/4 -1/4
the contrast to investigate if the observed measured value is equal in all sites (F-Test) would be:
1/2 -1/2 0 0 1/2 -1/2 0 0
0 1/2 -1/2 0 0 1/2 -1/2 0
0 0 1/2 -1/2 0 0 1/2 -1/2
How would the contrast matrix look like if I want to test if there is an overall interaction between site and diagnosis, i.e. does the effect of diagnosis vary by site (F-Test)?