Sorry to repost my question. I would appreciate any reply.
Best Regards, Gabor
2016-06-13 10:51 GMT+02:00 Gabor Perlaki petzinger.gabor@gmail.com:
Dear all,
I've a lot of patients and we are looking for the fastest solution to check the output of Freesurfer. On the troubleshooting page ( https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/TroubleshootingData) there are a lot of steps to check, but I think not all of that should be done to find out whether there is error in the final segmentations. I think, but I am not sure, that checking the surfaces (white and pial) and checking the subcortical segmentations (aseg) is enough to realize if something is wrong. After an error is detected based on the checked outputs (surfaces and aseg), we can find out the exact nature of the error and correct it based on the troubleshooting page. Is it a good solution for verifying the performance of Freesurfer by checking only these two things (surfaces and aseg) or I need to check all of the steps listed on the troubleshooting page anyway?
Best Regards, Gabor
Hi Gabor,
it depends on what you are doing. What you describe is sufficient for maps/structure volumes, but for example if you wanted to measure thickness within parcels you would have to check the accuracy of the cortical parcellation
cheers Bruce
On Tue, 28 Jun 2016, Gabor Perlaki wrote:
Sorry to repost my question. I would appreciate any reply.
Best Regards, Gabor
2016-06-13 10:51 GMT+02:00 Gabor Perlaki petzinger.gabor@gmail.com: Dear all,
I've a lot of patients and we are looking for the fastest solution to check the output of Freesurfer. On the troubleshooting page (https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/TroubleshootingData) there are a lot of steps to check, but I think not all of that should be done to find out whether there is error in the final segmentations. I think, but I am not sure, that checking the surfaces (white and pial) and checking the subcortical segmentations (aseg) is enough to realize if something is wrong. After an error is detected based on the checked outputs (surfaces and aseg), we can find out the exact nature of the error and correct it based on the troubleshooting page. Is it a good solution for verifying the performance of Freesurfer by checking only these two things (surfaces and aseg) or I need to check all of the steps listed on the troubleshooting page anyway?
Best Regards, Gabor
-- Gabor Perlaki research associate Diagnostic Center of Pécs H-7623 Pécs, Rét str. 2. Tel.: 0036-30-2084367 E-mail: petzinger.gabor@gmail.com
Dear Bruce,
Thank you for your answer. We will check the pial and white surfaces, but how to check the accuracy of the cortical parcellation beyond that? We don't find any suggestion on Freesurfer's troubleshooting page how to realize or correct such an error, where the pial and white surfaces are good but the cortical parcellation is not accurate. Manual intervention is only suggested for the aseg segmentations. Could you please suggest some more details about how to check/correct the possible errors in cases where both pial and white surfaces are accurate?
Best Regards, Gabor
Hi Gabor
the only way is to look at the parcellation over the white surface I guess. There are some not-very-well documented interventions, but maybe your best bet would be just to exclude subjects (since it almost never fails)
cheers Bruce
On Wed, 29 Jun 2016, Gabor Perlaki wrote:
Dear Bruce,
Thank you for your answer. We will check the pial and white surfaces, but how to check the accuracy of the cortical parcellation beyond that? We don't find any suggestion on Freesurfer's troubleshooting page how to realize or correct such an error, where the pial and white surfaces are good but the cortical parcellation is not accurate. Manual intervention is only suggested for the aseg segmentations. Could you please suggest some more details about how to check/correct the possible errors in cases where both pial and white surfaces are accurate?
Best Regards, Gabor
freesurfer@nmr.mgh.harvard.edu
-
Bruce Fischl -
Gabor Perlaki