Hi - yes, I think Option 2 looks good, with, as you suggest, replacing the
grand mean EV with a separate mean EV for each subject. If you do an
f-contrast across EVs 4 and 5 then this should hopefully show you where
there is any interaction effect.
Cheers, Steve.
On Fri, 5 Aug 2005, No Name wrote:
> Dr. Smith,
>
> I apologize for the confusing email from the other day. I hope that this one asks my question more
> clearly. Essentially we have fMRI data from two groups across three runs. We are interested in looking
> at differences between groups and between runs (1- 3) and the interaction of group x run. I have
> attached a text file with two possible setups for the FLAME analysis that we are considering and I
> would appreciate your feedback. In a previous email to this list, you mentioned that it would be
> valuable to model each subject individually rather than the grand mean- should we do this here? If
> so, do you model every subject or n-1 subjects to avoid overmodeling?
>
> Thank you for your time,
>
> Amy L. Krain, Ph.D.
> Institute for Pediatric Neuroscience
> NYU Child Study Center
>
--
Stephen M. Smith DPhil
Associate Director, FMRIB and Analysis Research Coordinator
Oxford University Centre for Functional MRI of the Brain
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
|