Dear Arun,
Ambitious idea. I wonder about your basic assumption though:
>The assumption from this point on is that there is a linear
>relationship between the PET data and my fmri results on a voxel by
>voxel level. In the preprocessing, I also took care that the final
>smoothing of the data is the same in both modalities (i.e. the fmri data
>was smoothed more than the pet data).
I think this may be a large assumption to make. The most obvious fact is
that BOLD and rCBF are not the same thing, even though they are clearly
related, and thus their response characteristics probably do not meet this
assumption. In fact, there is evidence from some parametric studies done
in London showing that rCBF has a much more linear response profile than
BOLD (see Rees et al. Neuroimage. 1997 Nov;6(4):270-8). In addition,
voxel-by-voxel there are differences due to the distortions in the B0 field
so that in the BOLD case, you get a shift towards draining veins and also
macroscopic susceptibility effects such as distortion and signal loss, none
of which are present in PET images. I would expect these effects to be
rather significant in any cross-modality voxel-by-voxel comparison.
I'm sorry this doesn't answer your question!
Joe
--------------------
Joseph T. Devlin, Ph. D.
FMRIB Centre, Dept. of Clinical Neurology
University of Oxford
John Radcliffe Hospital
Headley Way, Headington
Oxford OX3 9DU
Phone: 01865 222 738
Email: [log in to unmask]
|