Hi Darren,
I guess that if there is indeed a global atrophy that would be compensated for by the affine scaling, then you should use the siena tools. VBM is essentially conceived to look at local/accelerated atrophy when a tool like siena, while giving a global measure, is very sensitive to subtle changes in brain size.
Having said that, it would seem that a global atrophy of the brain does not really impact on the affine scaling that much, hence correcting with the TIV as a covariate or excluding the affine component from the Jacobian modulation are very similar ways of accounting for the same thing.
Hope this helps,
Gwenaelle
--- En date de : Mer 5.5.10, D Gitelman <[log in to unmask]> a écrit :
De: D Gitelman <[log in to unmask]>
Objet: Re: [FSL] FSLVBM - question about modulation and design
À: [log in to unmask]
Date: Mercredi 5 mai 2010, 16h14
Hi Jesper:
I am well and I hope you are too.
Thanks for your answer. The whole issue of modulation and covarying for brain size is a difficult one. What happens if there is atrophy?
If we imagine that a brain is smaller because of added atrophy then wouldn't correcting by just affine scaling the brain negate this effect, i.e., it would make the brain bigger to fit the template but I don't think scaling alone could differentiate a normal small brain from a mildly shrunken brain.
Correcting by total intracranial volume (assuming that G+W+CSF is a good approximation to the original TIV) would not be affected by atrophy since it assumes that the brain should have originally filled out the intracranial space. Thus it should be more sensitive to atrophy in addition to any difference in brain size.
Darren.
