Hi Sonia,
> I was wondering - is there some way to do the non-parametric equivalent
> of a spm99 simple regression at the random effects level?
Yes. This has always been possible by selecting the 'Single
Subject, 1 covariate' option. But with the just-released SnPM2
(no beta), there now is a more-obvious option:
MultiSub: 1 covariate, 1 scan per subject
See http://www.sph.umich.edu/ni-stat/SnPM to get it.
> (or a non parametric comparison of activation associated with a
> particular contrast across 3+ groups).
This isn't possible currently. I need to create a new plug for
'1-way ANOVA between subjects F-test'
but if you did find an effect, you would be left doing separate t-tests
to characterize the difference. (I'll put this on the to-do list).
-Tom
-- Thomas Nichols -------------------- Department of Biostatistics
http://www.sph.umich.edu/~nichols University of Michigan
[log in to unmask] 1420 Washington Heights
-------------------------------------- Ann Arbor, MI 48109-2029
|