Many thanks for that. Am I correct in thinking that the Gauss_Bonnet scheme is given by the "--discrete" option in mris_curvature_stats, while the latter option is given by the "--continuous" flag?

L


On Fri, Feb 26, 2010 at 1:46 PM, Bruce Fischl <fischl@nmr.mgh.harvard.edu> wrote:
Hi Lisa,

we actually provide a discrete estimate using Gauss-Bonnet (Rudolph can give you the details), and a slightly larger-scale estimate by fitting a quadratic patch to each neighborhood (5mm on a side I think). The curvature then comes from the Hessian of the height function of that patch.

cheers,
Bruce


On Fri, 26 Feb 2010, Lisa Ronan wrote:

Hi All

I wonder could you tell me how Gaussian curvature values are computed at a
vertex, i.e. Gauss_Bonnet scheme? Is there any way to vary the neighborhood
over which the vertex curvature values are computed?

Thanks for your help,

Lisa Ronan

- - - - - - - - - -
Research Associate
Brain Mapping Unit
Sir William Hardy Building
Department of Psychiatry
University of Cambridge