Dear Rajeev,
The idea behind "shuffle-corrected cross-correlation" is to dissociate
intrinsic neuronal dynamics (connectivity) from stimulus driven dynamics.
You correlate stimulus locked neuronal time-series, and compare the
correlation with that obtained when correlating stim locked time series that
have been obtained at different time points. The rationale behind this is,
that the stimulus induced locking should always be there, ie also in the
shuffled correlation. The neuronal dynamics should only be present, if you
take time series that have been acquired at the same time.
Obviously this could also be performed in single trial fMRI analyses with
long SOAs. You can look at the variable xSDM.Sess{nsess}.ons{nevent}, where
nsess is the session you are interested in and nevent is the event you want
to get the onsets of your stimuli.
I hope this helps
Christian
Dr. Christian Büchel
Neurologische Universitätsklinik, Haus B
Universitäts-Krankenhaus Eppendorf
Martinistr. 52
D-20246 Hamburg
Germany
Tel.: +49-40-42803-4726 Fax.: +49-40-42803-9955
[log in to unmask]
www.uke.uni-hamburg.de/kliniken/neurologie/pages/mitarbeiter/buechel_c.htm
----- Original Message -----
From: "Rajeev Raizada" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Thursday, June 27, 2002 9:26 PM
Subject: Q: fMRI equivalent of cross-correlation with shuffle-correction?
> Dear SPM list,
>
> I'm interested in how to do correlational analysis /
> functional connectivity in fMRI. In neurophysiology,
> the method of shuffle-corrected cross-correlation seems
> to be a very widely accepted standard, with some refinements
> suggested by Brody etc.
>
> But it's not clear to me what the fMRI analogue of this is.
> There are various correlation approaches floating around,
> e.g. using a given voxel's time-course as a regressor.
> This seems to bear some family resemblance to the
> neurophys shuffle-correction method, but I'm wondering
> whether there might be a different approach that's more
> closely equivalent.
>
> There's also the structural equation modeling approach,
> but that's further removed yet from kind of correlational
> analysis that the neurophysiologists do.
>
> Is there an fMRI equivalent of shuffle-corrected cross-correlation?
> Or does the difference in time-scales make a analogue to
> the neurophysiology approach impossible?
> I'd be very interested to hear opinions on this issue.
>
> Many thanks,
>
> Raj
> ----------------------------------------
> Rajeev Raizada
> Postdoctoral Research Fellow
> MGH-NMR Center / Harvard Medical School
> Building 149, 13th Street,
> Charlestown, MA 02129
> E.mail: [log in to unmask]
> Tel: 617 726 8790
> Fax: 617 726 7422
> ----------------------------------------
>
>
|