Dear Susanne,
if you have two experimental conditions (treatment vs. no treatment) then the model with two groups would fit very well:
https://neuro-jena.github.io/cat12-help/#long_two_interact
The group factor is treatment and you could also add another factor with randomization, but would probably start with the simpler model with just treatment as factor and the covariate.
Please note that if you have always two timepoints an alternative is to work with difference images and also with the difference for the covariate. The advantage is that you can use simple Anova. In CAT12 you will find the tools for calculating differences images and surfaces under "Volume Tools" and "Surface Tools".
Best,
Christian
On Tue, 29 Nov 2022 09:47:41 +0000, Susanne Meinert <[log in to unmask]> wrote:
>Dear all,
>
>we would like to conduct a longitudinal analysis with CAT12, however we are uncertain whether our model is correct, as most examples in the CAT12 manual are for 2 groups. We ran an experiment in 30 participants and would like to model the effect of mood (measured twice before and after experimental condition) on the changes of brain morphometry (also measured twice) via the experimental condition (treatment vs. no treatment). Further, we would like to include a factor of randomization (treatment at first or second time point).
>
>Is it correct to set the model as follows:
>
>Flexible factorial design
>
>Experimental condition
>- Independence: No
>- Variance: Equal
>- Grand mean scaling: No
>- ANCOVA: No
>
>Subject
>- Independence: Yes
>- Variance: Equal
>- Grand mean scaling: No
>- ANCOVA: No
>
>Main effects & Interactions
>- Main effect; Factor number 2
>- Main effect; Factor number 1
>
>Covariates
>- Name: Randomization (a vector of 0 and 1)
>- Interactions: None
>- Centering: Overall mean
>
>- Name: Mood
>- Interactions: With factor experimental condition
>- Centering: Overall mean
>
>The model looks similar to the example "Longitudinal data in two groups with interaction of covariate by group (example for two time points)" in the CAT12 manual.
>
>Now we set the contrast as follows
>treatment; no treatment; randomization; treatment*mood; no treatment*Mood
>0 0 0 1 -1
>0 0 0 -1 1
>as an F-Contrast.
>
>Our interpretation would be:
>If we find a significant Interaction, than the increase in mood via experimental condition is significantly correlated with the increase in grey matter volume via experimental condition.
>
>Is that correct? Thanks a lot in advance!
>
>Best
>Susanne
|