Hi Daniel,
> I'm struggling with the GLM GUI to
> set up a design matrix for a fsl_vbm 2 group
> analysis.
>
> In brief, I have 2 groups each containing N=15
> participants.
>
> As per the fsl_vbm instructions, I've labeled the
> "Patients" as "PAT_XXXX"
> and "Controls" are "CON_XXXX" (x's are numbers).
>
> Whereas the feat_gui for fMRI analyses allows users to
> paste the
> subdirectories for each subject, with fsl_vbm, all of the
> data (for all 30
> subjects) remains in the same folder.
>
> I tried going through the fsl_vbm tutorial/practical
> exercise on the fsl website,
> but it did not have enough detail.
>
> Any suggestions?
I'm afraid I'm not sure what your problem is here.
If you use the Glm gui, and switch in the glm setup window from timeseries design to higher level analysis/non timeseries design, you will be able to create your design.mat/con that need to be fed in randomise (either using the wizard button for a t-test, or the paste button for something more complex)...
Details of how to run randomise in your stats directory are then in the fslvbm and randomise manual.
Cheers,
Gwenaelle
|