Hi Jenny,

it's probably the topology correction that is causing your problem. What version are you running? Look for a topological error in that region (i.e. a hole or a handle). You can usually see the defects on the ?h.inflated.nofix surface. Load it with:

tksurfer -orig orig.nofix $subject $hemi inflated.nofix

this will let you use save point in tksurfer and goto point in tkmedit to find and fix the location of the defect (in the wm.mgz volume)

cheers, Bruce

On Thu, 9 Nov 2006, Jenny Kirchner wrote:

Hi Bruce,

thanks for the quick response. I am still not quite sure though where to do the edits. On the brainmask? I originally thought that I have to do the edits of the white matter segmentation on the wm.mgz. However, on the wm.mgz the white matter segmentation looks fine (i.e., the superior temporal gyrus is correctly classified as tissue). Thanks for your help.

Best, Jenny

On 11/9/06, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:

...

Hi Jenny,

the ?h.white is pretty much our final say one where the gray/white boundary is, and is usually substantially more accurate than the wm.mgz. Your dataset has a lot of motion artifact, but mostly looks okay except for the superior temporal gyrus, which has been chopped. You'll need to edit this back in. What version are you running? As for editing, an incorrect white surface can lead to an incorrect pial one, so you're probably better off rerunning things before trying to edit the pial. Also, the rerunning after pial editing is fast relative to the whole process, just a couple of hours I think.

cheers, Bruce

On Wed, 8 Nov 2006, Jenny Kirchner wrote:

...

Hi, I am working on a case for which I have run autorecon 2. Edits seem necessary on both the pial surface and the white matter segmentation. I am now wondering if it is better to first do the white matter edits and then run autorecon 2 again to see if the pial surface errors improved or if it is more economic to do the pial surface edits in the same step.

I also have a question concerning the differences in segmentation between the wm.mgz and the ?h.white. I am attaching two screenshots for you to be able to see it. The wm.mgz is showing white matter, which in the lh.white (loaded on top of the brainmask.mgz) is not defined as white matter as shown by the yellow line. (See temporal pole of the left hemisphere). I am wondering which of the both segmentations is correct, the wm.mgz version or the lh.white version and where exactely I have to do the edits.

Thanks a lot for your help! Best,

Jenny

Hi Jenny,

It would depend on what the real cause of the problem is, whether you would edit the wm.mgz or something else. It seems, from your emails, that you think the wm.mgz looks ok, but the ?h.white surface (which appears as a yellow line when displayed in tkmedit) does not follow it. When you load the ?h.orig or the ?h.orig.nofix into tkmedit, does this line follow the white matter boundary better, worse, or the same as your ?h.white surface.

I might have an easier time helping you figure this out if I could take a look at this particular subject. Could you upload the entire subject (with all outputs and sub dirs) on our file drop. https://www.nmr.mgh.harvard.edu/facility/filedrop/index.html

Thanks, Jenni

-----Original Message----- From: freesurfer-bounces@nmr.mgh.harvard.edu [mailto:freesurfer-bounces@nmr.mgh.harvard.edu] On Behalf Of Jenny Kirchner Sent: Thursday, November 09, 2006 1:52 PM To: Bruce Fischl Cc: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] white matter segmentation

Hi Bruce, I am running the latest version of Freesurfer. I am sorry but I am still not really sure as to how to fix the wrong segmentation. Even if I see that there was an error before the topology correction (which basically means that the correction made it worse?), how and where do I correct it. As I said, the wm.mgz looks fine, it's just that the ?h.white segmentation does not correspond to the wm.mgz. So the basic question is, if I do the edits in the brainmask or in the wm.mgz. Thanks again! Jenny

On 11/9/06, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:

Hi Jenny,

it's probably the topology correction that is causing your problem. What version are you running? Look for a topological error in that region (i.e. a hole or a handle). You can usually see the defects on the ?h.inflated.nofix surface. Load it with:

tksurfer -orig orig.nofix $subject $hemi inflated.nofix

this will let you use save point in tksurfer and goto point in tkmedit to find and fix the location of the defect (in the wm.mgz volume)

cheers, Bruce

On Thu, 9 Nov 2006, Jenny Kirchner wrote:

...

Hi Bruce,

thanks for the quick response. I am still not quite sure though where to do the edits. On the brainmask? I originally thought that I have to do the edits of the white matter segmentation on the wm.mgz. However, on the wm.mgz the white matter segmentation looks fine (i.e., the superior temporal gyrus is correctly classified as tissue). Thanks for your help.

Best, Jenny

On 11/9/06, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:

...

Hi Jenny,

the ?h.white is pretty much our final say one where the gray/white boundary is, and is usually substantially more accurate than the

wm.mgz.

...

...

Your dataset has a lot of motion artifact, but mostly looks okay except

for

...

...

the superior temporal gyrus, which has been chopped. You'll need to

edit

...

...

this back in. What version are you running? As for editing, an

incorrect

...

...

white surface can lead to an incorrect pial one, so you're probably

better

...

...

off rerunning things before trying to edit the pial. Also, the

rerunning

...

...

after pial editing is fast relative to the whole process, just a couple

of

...

...

hours I think.

cheers, Bruce

On Wed, 8 Nov 2006, Jenny Kirchner wrote:

...

Hi, I am working on a case for which I have run autorecon 2. Edits seem necessary on both the pial surface and the white matter segmentation. I am now wondering if it is better to first do the white matter edits and then run autorecon 2 again to see if the pial surface errors improved or if it is more economic to do the pial surface edits in

the

...

...

...

same step.

I also have a question concerning the differences in segmentation between the wm.mgz and the ?h.white. I am attaching two screenshots for you to be able to see it. The wm.mgz is showing white matter, which in the lh.white (loaded on top of the brainmask.mgz) is not defined as white matter as shown by the yellow line. (See temporal pole of the left hemisphere). I am wondering which of the both segmentations is correct, the wm.mgz version or the lh.white version and where exactely I have to do the edits.

Thanks a lot for your help! Best,

Jenny

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