Dear Alexa,
the first columns in your design matrix are not correct if you have two groups and two time points each. Then the first 4 columns should look like:
1 0 0 0
0 1 0 0
1 0 0 0
0 1 0 0
...
0 0 1 0
0 0 0 1
0 0 1 0
0 0 0 1
...
However, in your design matrix I only see 3 columns that code the time points. Check that you have coded the time points and groups correctly with with "1" and "2".
Best,
Christian
On Sun, 29 May 2022 14:10:40 +0200, Alexa Häger <[log in to unmask]> wrote:
>Dear SPM-community,
>
>
>
>I am currently performing a longitudinal VBM analysis on an intervention
>study on patients in CAT12. Therefore, I am applying a flexible factorial
>design with factor group (two groups A and B, i.e. control and intervention
>group) and time (two time points of acquisition for each subject) with no
>covariates. I performed this according to the description in the CAT12
>manual (p.43 ff. “longitudinal data in two groups”,
>http://www.neuro.uni-jena.de/cat12/CAT12-Manual.pdf), also following the
>preprocessing pipeline of the T1 data for longitudinal study designs. I
>have attached the design matrix for reference.
>
>
>
>However, when I am building the t-contrast for Time 1 > Time 2 in Group B
>(i.e., the intervention group), which is (0 0 1 -1) in the contrast
>manager, an error message “invalid contrast” is displayed. In contrast, for
>group A (control group) I can create the same contrast without problems.
>
>Does anyone have advice why the contrast could possibly not be created? I
>get the same error message when creating F-contrasts including the
>regressors 3 and 4. Might this be due to colinearity?
>
>
>
>Thank you very much in advance,
>
>Alexa
>
|