Hi Garth,

sorry, I don't understand. You can try running mri_normalize without the tal stuff by giving it -no1d (you can find the full command line in the scripts/recon-all.log file). If the tal is too far off it will cause the first step (the 1d part) of the normalization to fail.

cheers, Bruce

On Tue, 18 Aug 2009, Garth Coombs wrote:

When I check the talairach.xfm using tkregister2 --mgz --s amcg_anat --fstal --surf orig, it does not extend into the "darker" portion of the cortex and cuts short as if the cortex has been shaved off, but is only slightly off. In other areas, it looks very misaligned. It looks like possibly since it's not recognizing the cortex as brain, it's attempting to shift the whole alignment and aligning medial slices with more lateral slices so that the overall "height" of the cortex it's recognizing matches. I hope that makes sense. I would take a couple of pictures of it, but I'm not sure how to do so in tkregister2.

On Aug 17, 2009, at 5:40 PM, Bruce Fischl wrote:

...

wow, I haven't seen that before. Can you check the talairach.xfm and see if it's correct? On Mon, 17 Aug 2009, Garth Coombs wrote:

...

Hi,

I am running a patient of ours with some cortical degeneration through the recon process and have run across an interesting problem. It appears as if the intensity normalization process is missing a large portion of the cortex, resulting in an almost straight line through the middle of his head, with the center portion looking like the intensity normalization worked, but the cortex portion is very dark. The white matter in the cortex hasn't been normalized, and some of the gray matter looks as if it has been washed out or stripped away (that may just be due to the degeneration though). I have been going through and attempting to pick out control points to run -autorecon2-cp, but we wanted to check in and see if that was the best course of action for this particular subject before I got too far into it.

The data can be found in /autofs/space/dinah_001/users/balaji/ SUBJECTS_DIR/subjects/ and the subject ID is amcg_anat

Any advice would be greatly appreciated.

Thanks!

-Garth

Hi Nick and Bruce,

I followed those instructions and am rerunning the recon, so hopefully that will solve the problem. I'll let you know if any other issues arise with this patient. Thanks for your help!

-Garth

On Aug 18, 2009, at 11:03 AM, Nick Schmansky wrote:

Garth,

I had a look at the subject in tkregister, and it is badly misaligned, in a way that explains the intensity normalization problem. Follow these instructions to correct it:

http://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/Talairach

Also, I'll take a look at why our tal alignment code was not able to handle this image.

Nick

On Tue, 2009-08-18 at 10:05 -0400, Garth Coombs wrote:

...

When I check the talairach.xfm using tkregister2 --mgz --s amcg_anat -- fstal --surf orig, it does not extend into the "darker" portion of the cortex and cuts short as if the cortex has been shaved off, but is only slightly off. In other areas, it looks very misaligned. It looks like possibly since it's not recognizing the cortex as brain, it's attempting to shift the whole alignment and aligning medial slices with more lateral slices so that the overall "height" of the cortex it's recognizing matches. I hope that makes sense. I would take a couple of pictures of it, but I'm not sure how to do so in tkregister2.

On Aug 17, 2009, at 5:40 PM, Bruce Fischl wrote:

...

wow, I haven't seen that before. Can you check the talairach.xfm and see if it's correct? On Mon, 17 Aug 2009, Garth Coombs wrote:

...

Hi,

I am running a patient of ours with some cortical degeneration through the recon process and have run across an interesting problem. It appears as if the intensity normalization process is missing a large portion of the cortex, resulting in an almost straight line through the middle of his head, with the center portion looking like the intensity normalization worked, but the cortex portion is very dark. The white matter in the cortex hasn't been normalized, and some of the gray matter looks as if it has been washed out or stripped away (that may just be due to the degeneration though). I have been going through and attempting to pick out control points to run -autorecon2-cp, but we wanted to check in and see if that was the best course of action for this particular subject before I got too far into it.

The data can be found in /autofs/space/dinah_001/users/balaji/ SUBJECTS_DIR/subjects/ and the subject ID is amcg_anat

Any advice would be greatly appreciated.

Thanks!

-Garth

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

Garth,

I spoke too soon. Looking at the files in your mri/orig directory, 001.mgz is a structural, but 002.mgz appears to be an epi (partial brain), so the orig.mgz is a mix of the two (with the wacky intensity profile).

You should delete 002.mgz and either just run with 001.mgz (which looks good), or hunt down the other structural if there is one and replace 002.mgz.

Nick

On Tue, 2009-08-18 at 11:31 -0400, Garth Coombs wrote:

Hi Nick and Bruce,

I followed those instructions and am rerunning the recon, so hopefully that will solve the problem. I'll let you know if any other issues arise with this patient. Thanks for your help!

-Garth

On Aug 18, 2009, at 11:03 AM, Nick Schmansky wrote:

...

Garth,

I had a look at the subject in tkregister, and it is badly misaligned, in a way that explains the intensity normalization problem. Follow these instructions to correct it:

http://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/Talairach

Also, I'll take a look at why our tal alignment code was not able to handle this image.

Nick

On Tue, 2009-08-18 at 10:05 -0400, Garth Coombs wrote:

...

When I check the talairach.xfm using tkregister2 --mgz --s amcg_anat -- fstal --surf orig, it does not extend into the "darker" portion of the cortex and cuts short as if the cortex has been shaved off, but is only slightly off. In other areas, it looks very misaligned. It looks like possibly since it's not recognizing the cortex as brain, it's attempting to shift the whole alignment and aligning medial slices with more lateral slices so that the overall "height" of the cortex it's recognizing matches. I hope that makes sense. I would take a couple of pictures of it, but I'm not sure how to do so in tkregister2.

On Aug 17, 2009, at 5:40 PM, Bruce Fischl wrote:

...

wow, I haven't seen that before. Can you check the talairach.xfm and see if it's correct? On Mon, 17 Aug 2009, Garth Coombs wrote:

...

Hi,

I am running a patient of ours with some cortical degeneration through the recon process and have run across an interesting problem. It appears as if the intensity normalization process is missing a large portion of the cortex, resulting in an almost straight line through the middle of his head, with the center portion looking like the intensity normalization worked, but the cortex portion is very dark. The white matter in the cortex hasn't been normalized, and some of the gray matter looks as if it has been washed out or stripped away (that may just be due to the degeneration though). I have been going through and attempting to pick out control points to run -autorecon2-cp, but we wanted to check in and see if that was the best course of action for this particular subject before I got too far into it.

The data can be found in /autofs/space/dinah_001/users/balaji/ SUBJECTS_DIR/subjects/ and the subject ID is amcg_anat

Any advice would be greatly appreciated.

Thanks!

-Garth

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer