Hi, I’m having some issues generating surfaces for an individual and was hoping that a combination of tools might help. In this particular individual, the skull stripping went fine, but a fairly large proportion of the anterior temporal cortex (primarily inferior) volume is excluded from the data. My hunch is that it has to do with the fairly low contrast present in the images. Nonetheless, I had also worked on the data in SPM which did a fine job segmenting the wm/gm/other voxels when I inspect them as overlays in a program like MRICron.
I tried an experiment wherein I used the SPM segmented volumes to create a composite brainmask.mgz where the wm voxels were set to exactly 110, the gm voxels were set to a somewhat lower value (I used element-wise multiplication, so the voxels vary in value, but were around 70-100) and all other voxels were set to 1. The resulting volume looks like a meticulously cleaned brain mask. Unfortunately, recon-all -autorecon-pial didn’t work any better afterwards. So that experiment was a dud.
Assuming I can manipulate/combine the images in any manner, is there a way to take advantage of the segmented volumes that I was able to obtain from SPM? If I am going to be processing the data in SPM anyways, it would seem that having this information would make the process of tessellating the white and grey matter surfaces trivial.
/********************************************** * Chris McNorgan * Assistant Professor * Department of Psychology * University at Buffalo, * The State University of New York * http://ccnlab.buffalo.edu/ * Office: 716.645.0236 * Lab: 716.645.0222 **********************************************/
Hi Chris
did this get answered? I think you should be able to run autorecon1 and autorecon2, then paste your wm segmentation into wm.mgz and it will get preserved if the voxels you turn "on" are set to 255, and the ones you turn off are set to 1.
cheers Bruce
On Fri, 27 Oct 2017, Mcnorgan, Christopher wrote:
Hi, I’m having some issues generating surfaces for an individual and was hoping that a combination of tools might help. In this particular individual, the skull stripping went fine, but a fairly large proportion of the anterior temporal cortex (primarily inferior) volume is excluded from the data. My hunch is that it has to do with the fairly low contrast present in the images. Nonetheless, I had also worked on the data in SPM which did a fine job segmenting the wm/gm/other voxels when I inspect them as overlays in a program like MRICron. I tried an experiment wherein I used the SPM segmented volumes to create a composite brainmask.mgz where the wm voxels were set to exactly 110, the gm voxels were set to a somewhat lower value (I used element-wise multiplication, so the voxels vary in value, but were around 70-100) and all other voxels were set to 1. The resulting volume looks like a meticulously cleaned brain mask. Unfortunately, recon-all -autorecon-pial didn’t work any better afterwards. So that experiment was a dud.
Assuming I can manipulate/combine the images in any manner, is there a way to take advantage of the segmented volumes that I was able to obtain from SPM? If I am going to be processing the data in SPM anyways, it would seem that having this information would make the process of tessellating the white and grey matter surfaces trivial.
/**********************************************
- Chris McNorgan
- Assistant Professor
- Department of Psychology
- University at Buffalo,
- The State University of New York
- Office: 716.645.0236
- Lab: 716.645.0222
**********************************************/
freesurfer@nmr.mgh.harvard.edu
-
Bruce Fischl -
Mcnorgan, Christopher