Dear Peter,
We are conducting a spDCM analysis and we are now in the group comparison step. We want to compare the connectivity strength of 3 different groups. We have been using the "https://en.wikibooks.org/wiki/SPM/Parametric_Empirical_Bayes_(PEB)#First%20level%20analysis " web page. However, we still have a couple of questions:
1. How would be the correct design matrix if we want to compare:
(Group 1 vs. group 2) + (Group 1 vs. group 3) + (Group 2 vs. group 3)
2. So far, we compared group 1 vs. group 2 and group 1 vs. group 3 as follows:
[1 1
1 1
-1 0
-1 0
0 -1
0 -1]
Is that correct? When we swap the columns, the result remains the same. And if we change the positive values to negative values and vice versa, the results are the same, but inversed. So we are sure that so far, this is correct.
When we decided to change the second comparison for another one: instead of (group 1 vs. group 2) + (group 1 vs. group 3), we put (group 1 vs. group 2) + (group 2 vs. group 3):
[1 0
1 0
-1 1
-1 1
0 -1
0 -1]
Theoretically, the first regressor column is the same as before, however, by changing the second one, the results of the first comparison change. Why is this? Where is the error? Maybe you can only make a comparison of 3 groups, when at least 1 group remains the same in both comparisons? Therefore in this case, the group 2 should be defined in the same way?
3. Finally, I would like to ask you whether the order of the subjects is important. Should the subjects in each group be sequentially ordered? Or does it not matter as long as their group assignment is correct? An example:
[1 1
-1 0
0 -1
-1 0
1 1
0 -1]
Thank you so much for your help in advance.
Sofia,
|