The idea is to obtain a better spatial normalisation as it is based only on
grey matter. The arguments for using this approach are the same if you are
doing affine only spatial normalisation, or if you are also including
nonlinear warping. More accurate spatial normalisation will give you better
results, not only because homologous structures are better registered with
each other, but also because the prior probability maps should be overlayed
more accurately.
Best regards,
-John
> recently, Catriona Good (et al 2001, NeuroImage 14, 21-36) suggested to
> not simply segment normalized images, but rather to
> - segment in native space
> - normalize to the GM-template within SPM
> - apply these parameters to the original whole brain and
> - segment the "optimally" normalized whole brain again
> in order to base the normalization primarily on GM and to minimize
> non-brain tissue influence.
> I (think I) understand that the second normalization and segmentation is
> done with consideration of the more optimal non-linear normalization.
> Since I intend to do only an affine normalization anyway, I think I
> might as well just use the native GM-images and normalize them to the
> GM-template, omitting the "second round". Right or wrong ?
--
Dr John Ashburner.
Wellcome Department of Cognitive Neurology.
12 Queen Square, London WC1N 3BG, UK.
tel: +44 (0)20 78337491 or +44 (0)20 78373611 x4381
fax: +44 (0)20 78131420
http://www.fil.ion.ucl.ac.uk/~john
mail: [log in to unmask]
|