Hello,
The number of subjects acts to limit the total number of possible permutations, however once the number of possible permutations exceeds ~5000 ( which is the case for most reasonable group analyses ) there are diminishing returns from increased permutations ( see https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/Randomise/Theory for more details )
Kindest Regards
Matthew
--------------------------------
Dr Matthew Webster
FMRIB Centre
John Radcliffe Hospital
University of Oxford
> On 5 Jul 2017, at 04:55, Kavous Salehzadeh Niksirat <[log in to unmask]> wrote:
>
> Dear FSLers,
>
> I am using FSLVBM for a single-group average with a behavioral measure, exactly as explained here:
>
> https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/GLM#Single-Group_Average_with_Additional_Covariate
>
> As I know default permutation number for randomise function is 5000. The point is that I deal with a large data including 1,000 subjects. I wonder whether is there any relationship between the number of subjects and permutation? Should I increase 'n' or not?
>
> Sincerely,
> Kavous
|