Dear Russ,
> I'm not sure if this is a discussion that's already been had on the list, but > I worry about using the MNI-Talairach conversion.
> In particular, what's the logic for computing the converted coordinates?
Here's a little reply, for what it's worth.
I work in a pyschology unit; every now and again someone comes into my office
and asks me questions - sometime later there may appear a web page. I wrote
the MNI Talairach page because a) most of my colleagues feel impelled to
produce estimated Brodmann areas for the tables in their papers, and b) it
didn't seem to be generally known that the 'Talairach coordinates' from SPM do
not bear a straighforward relationship to the Talairach brain.
Now, one could take the position that Talairach Brodmann areas are nearly
meaningless, and personally I have a lot of sympathy with that point of view.
But regardless, it is customary to report some sort of estimation of Brodmann's
areas for activations, if only for the sake of comparison with other studies.
As far as I know, we have no information on BAs for the MNI brain, and the
Talairach, however flawed, remains the standard.
In a way of course, it is irrelevant whether it's a good idea or not to use the
Talairach Brodmann's areas, because people are going to do it anyway. In that
situation, it is surely better to do it in some principled way.
You are quite right of course that the transform on my page has not been
comprehensively validated (see http://www.mailbase.ac.uk/lists/spm/1999-10/
0136.html), but if you take a look at some sample points, I think it's clear
that the mni2tal transform is better than no transform. I am currently working
on a better transform - http://www.mrc-cbu.cam.ac.uk/~matthew.brett/MNITal/
mnital.html,
Anyway, more two pennies,
Best wishes,
Matthew
|