> I'm trying to run randomise on a set of scans acquired in mice
> (manganese enhanced structural MRI scans, so just a normal scan not
> a skeleton)- the resolution in-plane is .086x.086mm, and the slices
> are .35mm. I have a couple of questions about how to set some of the
> parameters given that the voxels are not at all isotropic, and that
> the scale of everything is fairly small.
> First, I just want to make sure that there aren't any assumptions
> that any of the tests make about size/space that would cause a
> problem here?
> Second, I have a relatively small sample (14 per group), and want to
> use variance smoothing, but am wondering what size kernel you would
> recommend in this situation?
Does that mean you will be using the voxel-wise statistic? I am not
sure if anyone has looked at the effects of variance smoothing on the
cluster-size statistic, and what good values would be. And I am almost
certain no one has looked at it for TFCE.
If you use the voxel-wise stats I would choose a FWHM (for the
variance smoothing) of ~6-8 inplane voxels.
Good luck Jesper
> Katherine H. Karlsgodt, Ph.D.
> Center for Cognitive Neuroscience
> Department of Psychiatry
> University of California, Los Angeles
> email: [log in to unmask]
> phone: 310-206-3019
> fax: 310-794-9740