[VoxBo] permutation testing question

[Kathy Tang]

On Wed, Aug 16, 2006 at 12:38:59PM -0400, Ranjani Prabhakaran wrote:
> 
> I thought I remembered hearing that you can't use permutation testing if you
> have a fixed order of events in each trial, but I'm not sure if that is true. 
> I looked for more information on the voxbo wiki pages, but I didn't see
> anything about the assumptions made in permutation testing.
> 

 possibly you are thinking of fixed effects rather than fixed order? 
Geoff Aguirre says

"Importantly, only data in which 
one can assume that the observations are independent can be analyzed 
using permutation techniques. The two primary examples of applications 
of this approach are to across-subject, random effects analyses and to 
the analysis of perfusion fMRI data."

in http://www.voxbo.org/pipermail/voxbo-general/2002-December/000068.html

kt