Re: [Freesurfer] incorrect surfaces and aseg labeling in subject with large occipital horns of lateral ventricle
Dear Doug,
I have uploaded them as tar.gzipped file
occipital_horn_ventricle_aseg_and_surf_error.tar.gz
to
anonymous@surfer.nmr.mgh.harvard.edu/transfer/incoming
I have uploaded it again in case the transfer failed.
I also appologize to forgot to attach the screenshots.
Sending them this time.
Regards,
Antonin
Hi Antonin, where did you upload them to? On 03/13/2017 05:51 PM, Antonin Skoch wrote:
Dear experts,
in my longitudinal study with 2 timepoints I encountered issue with aseg.presurf labeling and white/pial surface estimation of one subject in region of occipital horns of lateral ventricle. The voxels in horns are mislabeled as either white or gray matter or contain zero values. I am still in the stage 1 of longitudinal stream, so this is classical cross-sectional recon-all stream.
I tried to reprocess this subject by using recon-all -s subj -bigventricles -FLAIRpial (despite the fact that the subject does not have generally enlarged ventricles I thought that this could help nonlinear registration by mri_ca_register ). In case of timepoint 2 this helped to correctly label left occipital horn and corrected error with surfaces, but right occipital horn remained labeled as white matter.
In contrast, in case of timepoint 1 the -bigventricles did not help to label left occipital horn and additionally new massive error with white surface placement appeared in the region of occipital horn of left lateral ventricle. See the attached screenshots to ilustrate this issue:
tp1_nobigventricles_surface_OK tp1_bigventricles_surface_error tp2_nobigventricles_surface_error tp2_bigventricles_surface_OK
I have uploaded file occipital_horn_ventricle_aseg_and_surf_error.tar.gz containing 4 recons (timepoint 1 and 2 with and withoug -bigventricles) to your server.
I suppose that the root of the problem is incorrect nonlinear registration. Could you please advise how to diagnose and solve the error?
I am using development version from beginning of february 2017.
Regards,
Antonin Skoch
I have the data, it might take a few days to get back to you on my conclusions ...
On 3/13/17 6:30 PM, Antonin Skoch wrote:
Dear Doug,
I have uploaded them as tar.gzipped file
occipital_horn_ventricle_aseg_and_surf_error.tar.gz
to
anonymous@surfer.nmr.mgh.harvard.edu ftp://%3Ca%20href=/transfer/incoming
I have uploaded it again in case the transfer failed.
I also appologize to forgot to attach the screenshots.
Sending them this time.
Regards,
Antonin
Hi Antonin, where did you upload them to? On 03/13/2017 05:51 PM, Antonin Skoch wrote:
Dear experts,
in my longitudinal study with 2 timepoints I encountered issue with aseg.presurf labeling and white/pial surface estimation of one subject in region of occipital horns of lateral ventricle. The voxels in horns are mislabeled as either white or gray matter or contain zero values. I am still in the stage 1 of longitudinal stream, so this is classical cross-sectional recon-all stream.
I tried to reprocess this subject by using recon-all -s subj -bigventricles -FLAIRpial (despite the fact that the subject does not have generally enlarged ventricles I thought that this could help nonlinear registration by mri_ca_register ). In case of timepoint 2 this helped to correctly label left occipital horn and corrected error with surfaces, but right occipital horn remained labeled as white matter.
In contrast, in case of timepoint 1 the -bigventricles did not help to label left occipital horn and additionally new massive error with white surface placement appeared in the region of occipital horn of left lateral ventricle. See the attached screenshots to ilustrate this issue:
tp1_nobigventricles_surface_OK tp1_bigventricles_surface_error tp2_nobigventricles_surface_error tp2_bigventricles_surface_OK
I have uploaded file occipital_horn_ventricle_aseg_and_surf_error.tar.gz containing 4 recons (timepoint 1 and 2 with and withoug -bigventricles) to your server.
I suppose that the root of the problem is incorrect nonlinear registration. Could you please advise how to diagnose and solve the error?
I am using development version from beginning of february 2017.
Regards,
Antonin Skoch
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Antonin Skoch -
Douglas Greve