Dear Helka
> At a DCM/PEB analysis, I have designed the matrix (100 100 110 110 101 101) to compare three groups (one between-subject factor) and I have got the following result: attached
I assume there's a line break after each group of three numbers, so in Matlab format: [100; 100; 110; 110; 101; 101]. If so, that's fine.
> As the groups B and C show additive effects (relative to A), is it correct to say that:
> 1. Group B has the same conections that group A, less the conection 1 to 2 (+0.07 –0.07 = 0)?
> 2. Group C has the same conections that group A, less the conection 2 to 3?
Yes that's correct. However, note these effect sizes are very small, assuming that this is DCM of neuroimaging data. You might like to check the explained variance of your models. If you're using DCM for fMRI, you can do this using the command: spm_dcm_fmri_check(GCM);
> If I had compared only the groups A and B using a design matrix (10 10 11 11), what could I say about the comparison between them? Attached
Sorry, I don't understand this second question?
Thanks a lot.
Helka
|