Hi - yes, your reply had made it onto the list - but not in the place on
the September listing that you expected because you sent it in reply to a
different thread :)
Yes, 1 and 2 make sense - though note that 2. should be a mixed-effects
(FLAME stage 1) analysis if you are going to feed the results up to 3.
3. sounds fine too, though if you are only interested in the A-B contrast
for each subject then you may as well just have all A and all B in the 2.
analysis for each subject, get the [1 -1] contrast from that, and just
feed that single value from each subject up into 3.
Cheers.
On Thu, 8 Sep 2005, N.M. van Strien wrote:
> Hi Steve,
>
> I replied before on your answer, but since I dont see it on the
> mailinglist I decided to post it again.
>
> First: thanks for your suggestions!
>
> I readily applied your fixed effects solution to my one subject problem.
> However, if I continue this study more subjects will be sampled and a
> mixed-effects model will be of interest to me. Is my approach (see below)
> then the most efficient or would you suggest a different approach for this
> analysis?
>
> 1. Single subject - per run analysis (creating 8 feat dirs)
> 2. Single subject - grouping data over A fix A
and over B fix B
>
runs (resulting in 2 gfeat dirs)
> 3. Multi subject analysis having one EV for the A B difference,
> plus an EV for each subject to model each subjects mean effect (similar to
> the Paired Two-Group Difference example).
>
> Thanks,
>
> Niels
>
--
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
|