Dear Avihay,
With two levels for each of your within-subject factors, all
within-subject effects can be summarised with a single contrast image
per subject. I would go ahead with option 1) and use a one-sample t-test
or one-way ANOVA between-subject at the second level.
Best regards,
Guillaume.
On 15/01/2019 14:05, avihay cohen wrote:
> Hello SPM experts!
>
> I have the following study fMRI design which I wish to analyse:
> - 3 treatment groups (between)
> - 2 time points pre and post treatment (within)
> - 2 conditions (within)
>
> I'm interested in examining interaction between group, time and
> condition and was considering three options for model specification:
> 1. Modelling the within factors (time, condition) in the 1st level;
> modelling the between factor (group) in the 2nd level using a simple
> test (e.g. paired t-test, one-way ANOVA)
> 2. Modelling only one within factor (condition) in the 1st level. Then,
> using a /flexible /factorial model for modelling the other two within
> factors (time and subjects) and between factor (group); In the design
> matrix including Time*Group interaction and also the subject constants.
> 3. Similar to option 2 but without including the subject constants in
> the design matrix (only the time*group interaction).
>
> Would appreciate any feedback,
> Cohen Avihay
>
--
Guillaume Flandin, PhD
Wellcome Centre for Human Neuroimaging
UCL Queen Square Institute of Neurology
London WC1N 3BG
|