Douglas, 

I just realized that I was  using the wrong feeesurfer output. I had 3 processed MRIs for the same subject, and I just happened to pick the wrong one. The output of mri_surf2vol is indeed perfectly registered to the T1.mgz. So my problem is resolved.  It still bugs me that the --identity works even if you type anything as the subjectid, luckily this does not affect me at all.

Thanks for your help.

On another accord, I used the mri_morphology to erode the white matter and noticed that it erodes quite a lot in a single iteration, is there any other command which provides better controls in how much gets eroded? Also. is the erosion always happening inwards? 

Gabriel 

On Thu, May 18, 2017 at 12:29 PM, Gamaliz <gliza002@fiu.edu> wrote:

I thing I noticed is that if I change the "--identity" parameter it provides the same result. I am just typing anything and it still produces the same volume that gets created when the subject_id is used.

On Thu, May 18, 2017 at 12:16 PM, Gamaliz <gliza002@fiu.edu> wrote:
Douglas,

I am looking at them in freeview and mango. They are not aligned. I am able to align them using flirt, and they come out perfect. I have tried 5 cases. 3 of them they are aligned, two were not. I am loading below the wm and lh.wm. They are supposed to be aligned (wm is aligned with T1).

I got a surface by using Bruce's method, but it's much more irregular than the one obtained from mri_surf2vol. I wonder why the inconsistency. I guess I can create the surface and flirt using a solid transform to the T1. It takes longer, but it works.

Gabriel

--
gAbE

gAbE