Hi Matthew
Thanks for the quick response - In this case the binary images are the dataset, they aren't a mask (though they are restricted to a small region). i.e. is this voxel classified on average across this population as x. I'm not using them to restrict later analyses.
I could simply convert to a population probability map instead, but I'd like to be able to threshold in a vaguely principled way, though I can't think of any specific non-parametric test to do this.
_________________
Hi,
There shouldn't be a problem using your ROI as a mask, as long as the ROI is derived from some prior belief about areas of interest ( e.g. I'm only interested in results for white-matter voxels ).
Many Regards
Matthew
> Hi FSL experts,
>
> I'm looking to perform a 1 sample t-test using binary mask output from another analysis (classification masks). Is this valid with randomise, the results look reasonable but I have a sneaking suspicion I'm breaking the rules a bit?
>
> All the best
> Jonathan
>
|