To get Rudolph's curv stats in the current freesurfer version, you'll need to run this:
recon-all -s subjid -curvstats
(or add the -curvstats after -all if you havent processed your subjects yet).
It only takes a minute to run.
N.
On Fri, 2009-05-29 at 10:37 -0400, Rudolph Pienaar wrote:
The current version of Freesurfer calculates numerous curvature-based functions by default in the recon stream. If you look in the "surf" directory for a reconstructed surface, you'll see several "curvature" files that end in "*.crv".
By default, these are created for the gray/white junction surface (the *smoothwm* implies the smoothed white matter surface). The functions are all derived from the principal 'k1' and 'k2' curvatures that are calculated at each vertex point on the surface mesh, and are defined as
k1 maximum curvature k2 minimum curvature K Gaussian = k1*k2 H Mean = 0.5*(k1+k2) C Curvedness = sqrt(0.5*(k1*k1+k2*k2)) S Sharpness = (k1 - k2)^2 BE Bending Energy = k1*k1 + k2*k2 SI Shape Index = atan((k1+k2)/(k2-k1)) FI Folding Index = |k1|*(|k1| - |k2|)
The lh.curv and rh.curv files that are somewhat smoothed versions of the 'H' mean curvature, thus lh.curv is not equal to lh.smoothwm.H.crv, but close. Also, the lh.curv/rh.curv are calculated using a different algorithm than the curvature values for the above.
The main curvature calculating process is 'mris_curvature_stats'. If you want to generate all the curvature functions for the pial surface, you might try:
$>mris_curvature_stats --writeCurvatureFiles -G -F pial $SUBJECT $HEMI
where you should replace the "$SUBJET" with your subject, and "$HEMI" with either 'lh' or 'rh' for the hemisphere. When you run the command, your terminal will also report the results of several integral functions taken across the whole curvature function surface, and will also report the 'Folding Index', and several Intrinsic Curvature Indices: the positive (ICIp), negative (ICIn) and natural (INIn). The natural Intrinsic Curvature should be close to 1.0000 by definition.
For more information, I suggest you read the help/man page for 'mris_curvature_stats':
$>mris_curvature_stats -u | more
Zhangyuanchao wrote:
Dear Mr or Ms
Is there any way that we can obtain curvature value of the pial surface? I am not sure whether lh.curv is the curvature file that I want. How can I calculate the mean curvature, Gauss curvature of the pial surface and the GM/WM interface? Is there any built-in function in freesurfer to calculate the above index?
Is there any other measures of cortical foldings in freesurfer besides lGI and curvature?
Thanks a lot!
yczhang
好玩贺卡等你发，邮箱贺卡全新上线！ http://cn.rd.yahoo.com/mail_cn/tagline/card/*http://card.mail.cn.yahoo.com/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer