[Freesurfer] skull stripping problem

[Marina Fernández]

External Email - Use Caution

Dear Freesurfeer experts,

After running the recon-all, a part of the temporal lobe of one subject was
removed.

I used the watershed algorithm to be less agressive in the skull stripping
step:

recon-all -skullstrip -wsthresh 35 -clean-bm -no-wsgcaatlas -subjid sub_001

And the problem was solved in the brainmask.mgz volume because I can see
the temporal lobe completely, but the surfaces are still displaced. In
order to correct the surfaces I used the white matter edit tool but there
is no change after correcting WM and running the recon-all -autorecon2-wm
-autorecon3. I also used control points and but the following error was
shown in the terminal when I run the recon-all -autorecon2-cp -autorecon3:

recon-all -s sub_001 existed with errors

I detected that there are very high values in the white matter of this part
of the temporal lobe (even 132), maybe it has something to do. What can I
do to solve the problem with this subject?


Thank you very much in advance.

Best regards,

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

Re: [Freesurfer] Skull Stripping Problem

[Brian T. Quinn]

iag iag-
you may want to try increasing the preflooding height and use the -more 
option. i've found the seed points don't work very well, if at all. if 
all else fails, you may have to manually fix it in tkmedit with the clone 
tool. open tkmedit with the brain and T1 volumes as main and aux, 
respectively. select edit voxels tool, then Tools-Configure Volume Brush. 
in the dialog box, select 'Clone' under the Mode section, and 'Aux Volume' 
under the Clone Source section. now when you make edits to the main 
volume, it will fill in with the values from the T1.

good luck,
brian t. quinn
On Wed, 2 Mar 2005, iag iag wrote:
Hello all.
I am having a problem with skull stripping that I can't figure out a way
around.  One subject out of 10 is having the problem.  The intensity
normalization works fine and all values hover around 110 just like in the
other 9 subjects.  However, during skull stripping the pons is very
neatly cut out in this one subject.  Of course this causes problems later
when WMFill can't find the pons.
I tried to change the flood height, but that didn't change anything
(actually it changed NOTHING which kind of suprised me).  I tried adding
a seed point, but that didn't change the segmentation either.  I even
tried to add a seed point in Volume Index and Volume RAS coords just in
case I was doing something wrong with the coord system expected by the
routine.  I also tried not using the watershed.  But no matter what I try
the pons is always neatly chopped out so I can't get past WMFill on this
subject.
Does anyone have any ideas on what I might be able to do to allow the
pons to be left during RmSkull?
Thanks for any ideas.

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