Hi Ajay
I have found the problem that messed up the surfaces and fixed it. We can
get you an updated mris_fix_topology that will not generate corrupted
surfaces. However, the underlying problem is huge topological defects
caused by low gray/white contrast in this data. For example in the left
hemi, there is on in the insula (e.g. voxels around 362, 213, 200 and 364,
229, 207 and in the temporal lobe (351, 286, 282, 283, 309, 258).
Basically the gray/white contrast isn't high enough for us to accurately
segment the wm with our default procedures. Control points may help a bit,
but you may also need to manually edit the wm.mgz to remove the largest
defects. The easiest way to do this is to visualize them on the
lh.inflated.nofix surface with the lh.defect_labels overlaid to show you
where the defects are, then go into the volume in the wm.mgz and either add
control points or erase wm to fix things.
Let us know if you want a new mris_fix_topology. I'm not sure it is worth
the trouble since while it will now generate topologically correct surfaces
for this dataset, they will not be geometrically acccurate in the vicinity
of the huge (>200,000 vertices!) defects.
cheers
Bruce
On Wed, 2 Dec 2015, Ajay Kurani wrote:
Hi Bruce,
The template is the ICBM 2009b nonlinear template (0.5mm isotropic) which I
downloaded from the
McGills site (http://www.bic.mni.mcgill.ca/ServicesAtlases/ICBM152NLin2009).
The template has a
range of values from 0-100. Then I combined the csf/wm/gm masks (derived from
the 2009c template
which is 1mm isotropic) and resampled to 0.5 mm to mask the brain.
Subsequently I rescaled the
data to have a range of 0-255 and used this output as my input T1 for
Freesurfer. Thank you very
much for looking into this matter, it is much appreciated! I will look at the
file as you
suggested.
Sincerely,
Ajay
On Tue, Dec 1, 2015 at 7:55 AM, Bruce Fischl <fis...@nmr.mgh.harvard.edu> wrote:
Hi Ajay
can you explain your entire processing stream, including the details of
the
acquisition? What sequence/coil/scanner was used? You could fix your
problem by looking
at the defects on the inflated.nofix, overlaying
lh.defect_labels on it to see where they are, then figuring out what
caused them in the
volume. However, there is a bug in mris_fix_topology that only comes up
when it
encounters a large enough defect (with nvertices>200,000). The code tries
to avoid
running it's normal genetic algorithm as it would take days (or weeks) to
finish, since
it scales with the square of the number of vertices. I haven't figured
out what is
going wrong yet, but hopefully will have a code fix soon.
cheers
Bruce
On Tue, 1 Dec 2015, Ajay Kurani wrote:
Hi Bruce, When looking at my top slices in brain ask I noticed
that the
gray/white matter shows but the darker areas surrounding it are not
present
(this usually goes to the edge of the brain). Could the large
defects be a
result of having masked this out on the input file? I multiplied
the
template with the GM/WM/CSF masks they provided to get the brain
extracted
so I only see this missing in the very top few slices of the brain.
Thanks,
Ajay
On Tuesday, December 1, 2015, Bruce Fischl
<fis...@nmr.mgh.harvard.edu>
wrote:
Hi Falk
I don't think you need to rerun the qsphere. I'm pretty sure
this is a bug in mris_fix_topology for giant defects, but
haven't finished tracking it down yet
cheers
Bruce
On Tue, 1 Dec 2015, Falk Lüsebrink wrote:
Hi Ajay,
you should
cp ?h.orig.nofix ?h.orig
cp ?h.inflated.nofix ?h.inflated
with ? being l or r depending your hemisphere.
After the fixing stage the
surfaces got damaged somehow, therefore you have to
recreate it them. I’m
not entirely sure about the qsphere. Probably you
have to run the -qsphere
stage again also.
Best,
Falk
Von: Ajay Kurani [mailto:dr.ajay.kur...@gmail.com]
Gesendet: Montag, 30. November 2015 21:50
An: Falk Lüsebrink
Cc: Freesurfer support list; Bruce Fischl
Betreff: Re: [Freesurfer] Freesurfer 6.0 -
Autorecon2 failure
(mris_euler_number) using high res (0.5mm
isotropic) T1 image
Hi Falk,
I used the command and got the following error:
Command: ris_topo_fixer -mgz -warning -seed 1234
ICBM lh
mris_topo_fixer -mgz -warning -seed 1234 ICBM lh
INFO: assuming .mgz format
setting seed for random number genererator to 1234
reading
inputsurface/home/imuser/Downloads/mni_icbm152_nlin_sym_09b_nifti/mni_icbm152_n
lin_sym
_09b/ICBM/surf/lh.orig...
mrisFindNeighbors:/home/imuser/Downloads/mni_icbm152_nlin_sym_09b_nifti/mn
i_icbm152_nlin_sym
_09b/ICBM/surf/lh.orig: face[46196].v[0] = 24215,
but face 46196 not in
vertex 24215 face list
mris_topo_fixer -mgz -warning -seed 1234 ICBM rh
INFO: assuming .mgz format
setting seed for random number genererator to 1234
reading
inputsurface/home/imuser/Downloads/mni_icbm152_nlin_sym_09b_nifti/mni_icbm152_n
lin_sym
_09b/ICBM/surf/rh.orig...
mrisFindNeighbors:/home/imuser/Downloads/mni_icbm152_nlin_sym_09b_nifti/mn
i_icbm152_nlin_sym
_09b/ICBM/surf/rh.orig: face[35822].v[2] = 18903,
but face 35822 not in
vertex 18903 face list
Any suggestions would be appreciated.
Thanks,
Ajay
On Mon, Nov 30, 2015 at 2:41 AM, Falk Lüsebrink
<falk.luesebr...@ovgu.de>
wrote:
Hi Ajay,
I ran into the same error processing hires data a
while ago using a nightly
build of centos 6 with the hires flag only. Disk
space or alike wasn't an
issue. I ran mris_topo_fixer instead of
mris_fix_topology to get working
surfaces.
Best,
Falk
-----Ursprüngliche Nachricht-----
Von: freesurfer-boun...@nmr.mgh.harvard.edu
[mailto:freesurfer-boun...@nmr.mgh.harvard.edu] Im
Auftrag von Bruce Fischl
Gesendet: Sonntag, 29. November 2015 22:48
An: Freesurfer support list
Betreff: Re: [Freesurfer] Freesurfer 6.0 -
Autorecon2 failure
(mris_euler_number) using high res (0.5mm
isotropic) T1 image
Hi Ajay
are you sure you didn't run out of disk space? That
error is a sanity check
on the surfaces that should never occur
cheers
Bruce
On Sun, 29 Nov 2015, Ajay
Kurani wrote:
> Hello Freesurfer experts,
> I am trying to run Freesurfer 6.0 beta
(downloaded 10/15/15
> version) with the -hires flag to process an MNI
template image which
> is 0.5mm isotropic. Stage 1 ran well and on
stage 2 there is an error
> during
> pocessing:
>
> Command run:
> recon-all -s ICBM -autorecon2 -openmp 8 -3T
-hires -expert expert.opts
>
> I supplied a skull stripped brain which was nu
corrected and so I used
> the T1.mgz and copied it to brainmask.mgz and
nu.mgz in autorecon1
> stage. I then ran autorecon2 and it was working
well until I was at
> the -fix stage and ran into the following error:
>
>
> Error:
> mris_euler_number ./surf/lh.orig
> mrisFindNeighbors: ./surf/lh.orig:
face[46196].v[0] = 24215, but face
> 46196 not in vertex 24215 face list
>
>
>
> I am not sure if this is due to the fact that we
have a very high
> resolution image (template) and if there are any
modifications needed
> aside from the hires flag. Are there any
suggestions you have?
>
> Thanks,
> Ajay
>
>
The information in this e-mail is intended only for the person
to whom it is
addressed. If you believe this e-mail was sent to you in error
and the e-mail
contains patient information, please contact the Partners
Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was
sent
to you in error
but does not contain patient information, please contact the
sender and properly
dispose of the e-mail.
_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is
addressed. If you believe this e-mail was sent to you in error and the e-mail
contains patient information, please contact the Partners Compliance HelpLine at
http://www.partners.org/complianceline . If the e-mail was sent to you in error
but does not contain patient information, please contact the sender and properly
dispose of the e-mail.
