Hi Elijah,
Try following this example:
http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/GLM#ANOVA:_1-factor_4-levels__.281-way_between-subjects_ANOVA.29
Instead of 4, you have 3 groups.
When running randomise, you'll specify the design.mat (with the design
matrix), the design.con (with the t contrasts) and the design.fts (with
the F contrasts) and you'll automatically have the results of the ANOVA
and what would be the post hoc t-tests. The option --fonly will produce
only the ANOVA result.
All the best,
Anderson
Am 25.01.14 19:39, schrieb Elijah Mak:
> Dear FSL Experts,
>
> I am trying to produce the correct design matrix to perform randomize
> for voxelwise SIENA analyses.
>
> In total, I have 3 groups: A, B, and C. The 4D image has been built
> using fslmerge.
>
> If I understand the user guide correctly, I am supposed to built a
> FEAT-style design matrix and con? Is there a way to perform ANOVA and
> post-hoc pair-wise analyses? Or is it better to perform a series
> 2-sampled paired T-Tests (A vs B, B vs C, A vs C)?
> http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/Randomise/UserGuide
>
> Thank you very much!
>
> Best Regards,
> Elijah
>
|