On 10/17/2012 04:46 PM, Caroline Jiang wrote:
Hi,
I’m trying to set up my fsgd file for repeated measures ANOVA with two groups and 6 time points (0, 1, 2, 3, 4, 5) by following the example from here: http://surfer.nmr.mgh.harvard.edu/fswiki/RepeatedMeasuresAnova
Does this look correct for the fsgd file? Subject 2 is missing time points 2 and 5.
GroupDescriptorFile 1
Title
RepeatedMeasures
Class
Group1
Class
Group2
Variables
TP0-vs-TP1
TP0-vs-TP2
TP0-vs-TP3
TP0-vs-TP4
TP0-vs-TP5
Input
subject1_tp0
Group1
1
1
1
1
1
Input
subject1_tp1
Group1
-1
0
0
0
0
Input
subject1_tp2
Group1
0
-1
0
0
0
Input
subject1_tp3
Group1
0
0
-1
0
0
Input
subject1_tp4
Group1
0
0
0
-1
0
Input
subject1_tp5
Group1
0
0
0
0
-1
Input
subject2_tp0
Group1
1
1
1
1
1
Input
subject2_tp1
Group1
-1
0
0
0
0
Input
subject2_tp3
Group1
0
0
-1
0
0
Input
subject2_tp4
Group1
0
0
0
-1
0
Input
subject3_tp0
Group2
1
1
1
1
1
Input
subject3_tp1
Group2
-1
0
0
0
0
Input
subject3_tp2
Group2
0
-1
0
0
0
Input
subject3_tp3
Group2
0
0
-1
0
0
Input
subject3_tp4
Group2
0
0
0
-1
0
Input
subject3_tp5
Group2
0
0
0
0
-1
I’m interested in the overall group effect [1 -1 0 0 0 0 0 ], overall time effect, and group-x-time interaction.
Is this the correct contrast for the overall time effect?
0 0 1 0 0 0 0
0 0 0 1 0 0 0
0 0 0 0 1 0 0
0 0 0 0 0 1 0
0 0 0 0 0 0 1
yes
What would be the contrast for group-x-time interaction?
If you want to do a group-x-time, then you need to set up a different fsgd file where you would have 8 variables, the 4 you have above each one of them for each group. If the subject does not belong to the group, then set the columns for that group to 0. And the contrast matrix would then be
0 0 1 0 0 0 0 -1 0 0 0 0
0 0 0 1 0 0 0 0 -1 0 0 0
0 0 0 0 1 0 0 0 0 -1 0 0
0 0 0 0 0 1 0 ...
0 0 0 0 0 0 1 ...
doug
Thanks!
Caroline
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer