Dear all,
We are running a longitudinal VBM study. I would like to be sure that the contrast weights are correct, especially for the group*time interactions. The results seem a bit strange to me in comparison with the time effect results in each group when it is performed separately (Time effect in group A, time effect in groupe B). All comments are welcome.
For the flexible design we followed cat12 manual instructions: Factor 1 is subject, factor 2 is group (2 levels A and B, 60 subjects per group), factor 3 is time (2 timepoints).
Specify subjects, Scans & Factors:
Scans:
T0 scan
T1 scan
Conditions for the first group and two timepoints:
1 1
1 2
Conditions for the second group and two timepoints:
2 1
2 2
Contrast weight:
T1 > T0 in Group A = -1 1
T1 > T0 in Group B = 0 0 -1 1
T1 > T0 in Group A & B = -1 1 -1 1
T1 > T0 & Group A > B = -1 1 1 -1
T1 > T0 & Group B > A = 1 -1 -1 1
...
Is the reasoning OK?
Thank you and best regards,
Damien
|