Hi Victor,

I don't think we've seen that, at least I never have. If you have a subject that this happens to can you upload it for us with the first set of control points that result in missing a lot of surface (together with the poor surfaces)?

cheers Bruce On Fri, 26 Feb 2010, Laluz, Victor wrote:

Hello,

Throughout most of Freesurfer's version history, I have noticed this problem pop up every 50 or so cases I do:

I will have a pretty well-processed surface but with some inaccuracies in the WM surface, I will put control points down to correct it, and reprocess with autorecon2-cp as usual. The surface comes back with the areas I put CPs for accurate, but the rest of the cortex in a localized area will have a massive amount of the surface missing. An example is attached as a .png image.

Now, if I go back and put down control points throughout the part of the surface that is missing after the first reprocessing and repeat autorecon2-cp, the end result at that point is usually a perfectly accurate surface. However, I was wondering if the massive loss in surface after the first autorecon2-cp is attributable to some kind of bug that I might be able to avoid.

Any thoughts would be great, thanks!

Victor

Victor Laluz Imaging Coordinator UCSF Memory and Aging Center 350 Parnassus Avenue, Suite 905

Box 1207

San Francisco, CA 94143-1207 Phone: (415) 476.1680 Fax: (415) 476.1816 Email: vlaluz@memory.ucsf.edumailto:vlaluz@memory.ucsf.edu memory.ucsf.eduhttp://www.memory.ucsf.edu/ www.youtube.com/ucsfmemoryandaginghttp://www.youtube.com/ucsfmemoryandaging