[VoxBo] rfx with multiple timepoints

[Ling Wong]

On 1/16/07, Daniel Y Kimberg <kimberg at mail.med.upenn.edu> wrote:
>
> If I understand this correctly, it sounds like you have a
> within-subject treatment (session) and you'd like to do a paired
> t-test.  If that's the case, it's pretty easy, and you don't need to
> combine the sessions from each subject.  You can treat each session
> independently in your rfx script, so that you end up with two maps per
> subject (one for the cue effect in session 1, and one for the cue
> effect in session 2).  Then create a GLM with a session covariate, an
> intercept, and n-1 subject covariates.  The subject covariates would
> look like 0011000000000... (if the third and fourth volumes in the RFX
> were from the same subject), if you include an intercept you can
> include one for all but one of your subjects.
>
>
Thanks Dan,
On another note, I had a more conceptual question.  Based on Geoff's advice
earlier, we had planned on subtracting session maps from each other, for
each subject.  Then, these difference maps would be used to run an rfx.  Is
there a functional difference between doing it that way, and the way you
proposed (essentially, rfx first and then subtracting sessions in a
contrast)?  I just want to make sure I'm not missing something important
here...

Ling
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.voxbo.org/pipermail/voxbo-general/attachments/20070205/78dc328a/attachment.html