Hey Wenjing-
I've gotten similar problems before.
It looks like the segmentation didn't run properly.  This can occur if there are anomolies in the contrast of the images to begin with.
1- did you normalize to a T1- template?  If so you should check all your images there and make sure nothing weird happened.
2- check all of your subjects and see if this is just an issue with a few of them.
Also it helps to go through the scans to determine if there is a brightness contrast difference between weird subjects and not-weird subjects.  If the subjects are run on different days, random anomolies can lead to issues with contrast which is what VBM uses to perform segmentation.

On Fri, Oct 31, 2008 at 10:26 AM, Wenjing Li <[log in to unmask]> wrote:
Dear SPMers,

       I encountered  some problems during my use of vbm toolbox.
       I have reoriented each image and put the origin close to AC before using vbm toolbox, but when I used the vbm5 toolbox in SPM5 on these images , I got bad results(see attached figure).
       I have run the unified segmentation in SPM5, the results were fine. I don't know how I should solve this problem.

       Thank you very much!
Wenjing Li

Megan Kay Walsh
Graduate Student
Johns Hopkins University
Psychological & Brain Sciences
204 Ames Hall
3400 N. Charles St.
Baltimore, MD  21218