Hi,
--- Jiansong Xu <[log in to unmask]> a écrit :
> Dear Friends:
>
> After a little bit struggle and with lot of help
> from the list,
> finally I can look at stats results.
Well done (the second assumption was the good one:
swap MNI152_T1_1mm for the FMRIB58_FA_1mm).
> I have one group 16 subjects. I assessed the
> correlations between age
> and FA with following procedures.
>
> 1. Glu_gui, one EV, type in demeaned age,
> 2. set up two contrast [1] and [-1].
> 3. randomise
>
> I have following questions
>
> 1. -c 2, this option allow me to assess p at
> cluster level. What
> does 2 mean here? Is it t threshold of each voxel?
Yes
> If so, am I right
> by saying that voxles form a cluster if all of them
> has t > 2 and also
> contiguous with each other.
Yes
> The p in maxc_tstat is
> dependent on both
> the t value and number of voxels (extent) in the
> cluster, so this p is
> the p value of each cluster.
Yes
> 2. vox_tstat contains the p of each voxel. the p in
> this file is not
> very useful for reporting result, because not
> corrected for multiple
> comparison
That's right
> 2. what the mean of p in max_tstat?
Voxel-based corrected with the family-wise error
(FWE). Very conservative. Please have a look at:
http://www.fmrib.ox.ac.uk/fsl/randomise/index.html
You might also want to give a try to the TFCE...
Cheers,
Gwenaëlle
>
> Best
>
> Jiansong Xu, M.D., Ph.D.
> Associate Research Scientist
> Dept. of Psychiatry
> Yale School of Medicine
>
> Connecticut Mental Health Center, Room S-100
> 34 Park St.
> New Haven, CT 06519
>
_____________________________________________________________________________
Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail http://mail.yahoo.fr
|