6 Mar
2009
6 Mar
'09
2:27 a.m.
Hi all,
We used the longitudinal processing for our phase 2 data. For some subjects, I found very bad segmentation and surface reconstructions. Suspecting it's a registration problem, I loaded brainmask.mgz from tp1 and tp2 in tkmedit. I could see they are not aligned to each other (attached image).
My questions are: 1. Is the problem due to registration? If yes, how can I fix it? 2. If we are to edit the segmentation results from "-long" processing, do we follow the same steps as what we did for tp1?
Thanks, Zheng Hui
