hmm, not sure what happened on the 2nd run, but you should be able to corect those problems by putting some control points in superior regions (in the white matter at the base of the strands that are lost). Make sure they are completely in the wm and not in partial-volumed voxels.

Bruce

On Thu, 25 Oct 2007 zheng.hui@singhealth.com.sg wrote:

Hi,

I did some editing, add in control points and edit pial surface on one set of our data on the left hand side of the screenshot attached, and re-run recon-all -autorecon2 -autorecon3 -s subjID

The result turned out to be wrong as shown in the right hand side of the screenshot. The algorithm worked well at the temporal lobe where control points were added, but it failed at the superior part. Can you suggest why it is so? Thanks

Zheng Hui