Hi
So this is a repeated measures 2x2 anova... makes for a heck of a design matrix. One simple way is indeed to combine individual subjects at an intermediate level. An wasy way of doing this is to average the three versions of dr_stage2 per subject into a single estimate and then run only these 48 estimates in a 2x2 anova design. This then does amount to fixed-effects between sessions.
hth
Christian
On 11 Oct 2011, at 16:26, Monica Wey wrote:
> Dear Experts,
>
> I am working on some resting states fMRI data analysis, and would like to perform dual regression for group comparisons.
> However, I have some questions regarding the design.
>
> We have 48 subjects randomly assigned into four groups (treatment, sham, learning, not learning, 12 subjects in each group), and each subjects have 3 runs (~ 6min) of resting state bold scans.
> I would like do a group comparison to find the treatment and learning effects as well as model the interaction.
> I am thinking to use the 2 factors, 2 levels design as shown on the website.
> But, I am not sure if I should do any additional steps for the 3 runs of each subject??
> By running ICA, we will concatenate all runs, so I am thinking it might not be necessary (e.g. to use a fixed effect model for each subject)?
>
> I would appreciate any input and advice.
> Thanks!
>
> Monica
|