Hi Siewmin,
you need the mean_FA_skeleton_mask to run your randomise analysis on MD.
So what you've run:
randomise -i all_MD_skeletonised -o tbss -m mean_FA_skeleton_mask -d
design.mat -t design.con -n 500 --T2 -V
is correct.
You simply need to change the name of the output (say to tbss_MD) so that it won't overwrite your FA results.
If both FA and MD results look similar, this is a good thing as it means FA decrease (presumably) co-localises with MD increase and it makes your interpretation more straigthforward.
Cheers,
Gwenaelle
--- En date de : Jeu 2.7.09, Siewmin Gan <[log in to unmask]> a écrit :
> De: Siewmin Gan <[log in to unmask]>
> Objet: [FSL] randomise on nonfa data
> À: [log in to unmask]
> Date: Jeudi 2 Juillet 2009, 12h56
> Hi, with the command tbss_nonfa MD, I
> get two results all_MD.nii.gz and
> MD_skeletonised.nii.gz in the stats folder. May I ask if it
> is the right
> step to do randomise with the command whereby I use the
> old
> mean_FA_skeleton_mask, as there is no
> mean_MD_skeleton_mask.nii.gz file that
> is generated?
> randomise -i all_MD_skeletonised -o tbss -m
> mean_FA_skeleton_mask -d
> design.mat -t design.con -n 500 --T2 -V
>
> The output files from doing randomise on the
> all_MD_skeletonised are in the
> exact tbss_tfce.. filenames that are generated from doing
> randomise on the
> all_FA_skeletonised data, and would replace the previous
> results from the FA
> data. When I look at the newly generated
> tbss_tfce_corrp_tstat1. It looks
> similiar to the results from running randomise on the FA
> files. Would this
> be right?
>
> Thanks
>
> Siewmin
>
|