Dear,
I have a doubt regarding the input T1 image used for the Hippocampal Subfield segmentation in Freesurfer 6.0 - https://surfer.nmr.mgh.harvard.edu/fswiki/HippocampalSubfields
In this documentation it says that I can use a standard 1mm T1 image or the same standard 1mm T1 image with an additional scan. However, I want to use a 0.7mm T1 image as input, with an additional T2 scan. When I run the program with it, the output images (nu.mgz and orig.mgz) are downsampled to 1mm voxel size. So, I don't know if this happens just because of the recon-all program or if the segmentation is being done in the downsampled T1 image or in the original 0.7mm.
By reading the article "A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI" I saw that, in the methods section, the algorithm is used to segment 0.6mm T1 images. But is this possible to do in the Freesurfer package? Obviously, doing the segmentation directly on 0.7mm data would be better than a downsampled version to 1mm...
Thanks,
Fernando Pasquini Santos PhD student in Dynamic Systems fernandosantos@pitt.edu fernando.pasquini.santos@usp.br
I would be interested to know how well 0.7mm isotropic resolution data performs at hippocampal segmentation. It would be better if folks could do this then needing to acquire specialized anisotropic scans for hippocampal segmentation.
Peace,
Matt.
From: freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Fernando Pasquini Santos fernandopasq@gmail.com Reply-To: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Date: Thursday, June 9, 2016 at 1:23 PM To: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] Hippocampal subfields segmentation - what is the voxel size of the input?
Dear,
I have a doubt regarding the input T1 image used for the Hippocampal Subfield segmentation in Freesurfer 6.0 - https://surfer.nmr.mgh.harvard.edu/fswiki/HippocampalSubfields
In this documentation it says that I can use a standard 1mm T1 image or the same standard 1mm T1 image with an additional scan. However, I want to use a 0.7mm T1 image as input, with an additional T2 scan. When I run the program with it, the output images (nu.mgz and orig.mgz) are downsampled to 1mm voxel size. So, I don't know if this happens just because of the recon-all program or if the segmentation is being done in the downsampled T1 image or in the original 0.7mm.
By reading the article "A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI" I saw that, in the methods section, the algorithm is used to segment 0.6mm T1 images. But is this possible to do in the Freesurfer package? Obviously, doing the segmentation directly on 0.7mm data would be better than a downsampled version to 1mm...
Thanks,
Fernando Pasquini Santos PhD student in Dynamic Systems fernandosantos@pitt.edu fernando.pasquini.santos@usp.br _______________________________________________ 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.
Hi Matt, you can use strategy 3 of my previous email ;-) Cheers, Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 19:31, Matt Glasser <matt@ma-tea.commailto:matt@ma-tea.com> wrote:
I would be interested to know how well 0.7mm isotropic resolution data performs at hippocampal segmentation. It would be better if folks could do this then needing to acquire specialized anisotropic scans for hippocampal segmentation.
Peace,
Matt.
From: <freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu> on behalf of Fernando Pasquini Santos <fernandopasq@gmail.commailto:fernandopasq@gmail.com> Reply-To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edumailto:freesurfer@nmr.mgh.harvard.edu> Date: Thursday, June 9, 2016 at 1:23 PM To: "freesurfer@nmr.mgh.harvard.edumailto:freesurfer@nmr.mgh.harvard.edu" <freesurfer@nmr.mgh.harvard.edumailto:freesurfer@nmr.mgh.harvard.edu> Subject: [Freesurfer] Hippocampal subfields segmentation - what is the voxel size of the input?
Dear,
I have a doubt regarding the input T1 image used for the Hippocampal Subfield segmentation in Freesurfer 6.0 - https://surfer.nmr.mgh.harvard.edu/fswiki/HippocampalSubfields
In this documentation it says that I can use a standard 1mm T1 image or the same standard 1mm T1 image with an additional scan. However, I want to use a 0.7mm T1 image as input, with an additional T2 scan. When I run the program with it, the output images (nu.mgz and orig.mgz) are downsampled to 1mm voxel size. So, I don't know if this happens just because of the recon-all program or if the segmentation is being done in the downsampled T1 image or in the original 0.7mm.
By reading the article "A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI" I saw that, in the methods section, the algorithm is used to segment 0.6mm T1 images. But is this possible to do in the Freesurfer package? Obviously, doing the segmentation directly on 0.7mm data would be better than a downsampled version to 1mm...
Thanks,
Fernando Pasquini Santos PhD student in Dynamic Systems fernandosantos@pitt.edumailto:fernandosantos@pitt.edu fernando.pasquini.santos@usp.brmailto:fernando.pasquini.santos@usp.br _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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. _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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.
Hi Eugenio,
Did you already do this? If so, a published reference would be more useful to me than instructions for how to do it, as I¹m more interested in making an argument for an acquisition protocol modification.
Matt.
From: freesurfer-bounces@nmr.mgh.harvard.edu on behalf of "Iglesias, Eugenio" e.iglesias@ucl.ac.uk Reply-To: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Date: Thursday, June 9, 2016 at 1:37 PM To: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Hippocampal subfields segmentation - what is the voxel size of the input?
Hi Matt, you can use strategy 3 of my previous email ;-) Cheers, Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 19:31, Matt Glasser matt@ma-tea.com wrote:
I would be interested to know how well 0.7mm isotropic resolution data performs at hippocampal segmentation. It would be better if folks could do this then needing to acquire specialized anisotropic scans for hippocampal segmentation.
Peace,
Matt.
From: freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Fernando Pasquini Santos fernandopasq@gmail.com Reply-To: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Date: Thursday, June 9, 2016 at 1:23 PM To: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] Hippocampal subfields segmentation - what is the voxel size of the input?
Dear,
I have a doubt regarding the input T1 image used for the Hippocampal Subfield segmentation in Freesurfer 6.0 - https://surfer.nmr.mgh.harvard.edu/fswiki/HippocampalSubfields
In this documentation it says that I can use a standard 1mm T1 image or the same standard 1mm T1 image with an additional scan. However, I want to use a 0.7mm T1 image as input, with an additional T2 scan. When I run the program with it, the output images (nu.mgz and orig.mgz) are downsampled to 1mm voxel size. So, I don't know if this happens just because of the recon-all program or if the segmentation is being done in the downsampled T1 image or in the original 0.7mm.
By reading the article "A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI" I saw that, in the methods section, the algorithm is used to segment 0.6mm T1 images. But is this possible to do in the Freesurfer package? Obviously, doing the segmentation directly on 0.7mm data would be better than a downsampled version to 1mm...
Thanks,
Fernando Pasquini Santos PhD student in Dynamic Systems fernandosantos@pitt.edu fernando.pasquini.santos@usp.br _______________________________________________ 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. _______________________________________________ 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.
_______________________________________________ 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.
Hi Fernando,
Unfortunately, FreeSurfer does downsample the T1s to 1mm, and the subfield code assumes that the T1 comes from FreeSurfer… For the paper, I hacked the code to take advantage of multiple images with higher resolution, but it hasn’t been thoroughly tested and it’s not publicly available at this point. That said, there are different things you can do at this point: 1. Use the T1+T2 as is (disadvantage: T1 will be resampled to 1 mm). 2. Use only the T2. 3. Use only the T1. To do this, you would use the T1 as additional scan on its own (i.e., the main and additional scans are the same, but the additional one won’t be resampled). 4. There might be a way (hack) of using both scans at full resolution. If you change the header of the T1 scan and specify that the voxel size is 1mm, no resampling will happen; FreeSurfer will believe that you’re dealing with a subject with a very large head. If you do this, you’ll also have to enlarge the voxel size of the T2 by a factor 1/0.7. Also, the resulting volumes will have to be divided by (1/0.7)^3 in order to correct for the “wrong” voxel size.
I hope this helps!
Cheers,
Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 19:23, Fernando Pasquini Santos <fernandopasq@gmail.commailto:fernandopasq@gmail.com> wrote:
Dear,
I have a doubt regarding the input T1 image used for the Hippocampal Subfield segmentation in Freesurfer 6.0 - https://surfer.nmr.mgh.harvard.edu/fswiki/HippocampalSubfields
In this documentation it says that I can use a standard 1mm T1 image or the same standard 1mm T1 image with an additional scan. However, I want to use a 0.7mm T1 image as input, with an additional T2 scan. When I run the program with it, the output images (nu.mgz and orig.mgz) are downsampled to 1mm voxel size. So, I don't know if this happens just because of the recon-all program or if the segmentation is being done in the downsampled T1 image or in the original 0.7mm.
By reading the article "A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI" I saw that, in the methods section, the algorithm is used to segment 0.6mm T1 images. But is this possible to do in the Freesurfer package? Obviously, doing the segmentation directly on 0.7mm data would be better than a downsampled version to 1mm...
Thanks,
Fernando Pasquini Santos PhD student in Dynamic Systems fernandosantos@pitt.edumailto:fernandosantos@pitt.edu fernando.pasquini.santos@usp.brmailto:fernando.pasquini.santos@usp.br _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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.
Hi Eugenio, Would it also be possible to use the -hires flag in the initial recon-all command to prevent the T1 from being downsampled and maintain the voxel size that the T1 was acquired at, and use that resulting recon for the hippocampal subfield segmentation? I've tested this with MEMPRAGE scans acquired at 0.75mm isotropic and coronal T2 anisotropic scans acquired at 0.33mm in plane with 1mm thick slices, and the results look reasonable. The "1mm" label files are correctly downsampled to 0.75mm instead of 1mm, and the higher resolution label files remain at 0.33mm resolution.
Thanks, Matt
Stufflebeam Lab Athinoula A. Martinos Center Massachusetts General Hospital 149 13th Street, Room 1115S Charlestown, MA 02129 ________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu [freesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Iglesias, Eugenio [e.iglesias@ucl.ac.uk] Sent: Thursday, June 09, 2016 2:36 PM To: Freesurfer support list Subject: Re: [Freesurfer] Hippocampal subfields segmentation - what is the voxel size of the input?
Hi Fernando,
Unfortunately, FreeSurfer does downsample the T1s to 1mm, and the subfield code assumes that the T1 comes from FreeSurfer… For the paper, I hacked the code to take advantage of multiple images with higher resolution, but it hasn’t been thoroughly tested and it’s not publicly available at this point. That said, there are different things you can do at this point: 1. Use the T1+T2 as is (disadvantage: T1 will be resampled to 1 mm). 2. Use only the T2. 3. Use only the T1. To do this, you would use the T1 as additional scan on its own (i.e., the main and additional scans are the same, but the additional one won’t be resampled). 4. There might be a way (hack) of using both scans at full resolution. If you change the header of the T1 scan and specify that the voxel size is 1mm, no resampling will happen; FreeSurfer will believe that you’re dealing with a subject with a very large head. If you do this, you’ll also have to enlarge the voxel size of the T2 by a factor 1/0.7. Also, the resulting volumes will have to be divided by (1/0.7)^3 in order to correct for the “wrong” voxel size.
I hope this helps!
Cheers,
Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 19:23, Fernando Pasquini Santos <fernandopasq@gmail.commailto:fernandopasq@gmail.com> wrote:
Dear,
I have a doubt regarding the input T1 image used for the Hippocampal Subfield segmentation in Freesurfer 6.0 - https://surfer.nmr.mgh.harvard.edu/fswiki/HippocampalSubfields
In this documentation it says that I can use a standard 1mm T1 image or the same standard 1mm T1 image with an additional scan. However, I want to use a 0.7mm T1 image as input, with an additional T2 scan. When I run the program with it, the output images (nu.mgz and orig.mgz) are downsampled to 1mm voxel size. So, I don't know if this happens just because of the recon-all program or if the segmentation is being done in the downsampled T1 image or in the original 0.7mm.
By reading the article "A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI" I saw that, in the methods section, the algorithm is used to segment 0.6mm T1 images. But is this possible to do in the Freesurfer package? Obviously, doing the segmentation directly on 0.7mm data would be better than a downsampled version to 1mm...
Thanks,
Fernando Pasquini Santos PhD student in Dynamic Systems fernandosantos@pitt.edumailto:fernandosantos@pitt.edu fernando.pasquini.santos@usp.brmailto:fernando.pasquini.santos@usp.br _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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.
Hi Matt, Bruce actually mentioned this possibility yesterday. Good to hear it doesn’t crash when you run it! That said, I need to go through the code, and make sure that data are handled the way they are supposed to: the fact that it doesn’t crash doesn’t directly imply that it is working properly. Hopefully I’ll find time to do this next week… Cheers, Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 21:51, Hibert, Matthew Louis <MHIBERT@mgh.harvard.edumailto:MHIBERT@mgh.harvard.edu> wrote:
Hi Eugenio, Would it also be possible to use the -hires flag in the initial recon-all command to prevent the T1 from being downsampled and maintain the voxel size that the T1 was acquired at, and use that resulting recon for the hippocampal subfield segmentation? I've tested this with MEMPRAGE scans acquired at 0.75mm isotropic and coronal T2 anisotropic scans acquired at 0.33mm in plane with 1mm thick slices, and the results look reasonable. The "1mm" label files are correctly downsampled to 0.75mm instead of 1mm, and the higher resolution label files remain at 0.33mm resolution.
Thanks, Matt
Stufflebeam Lab Athinoula A. Martinos Center Massachusetts General Hospital 149 13th Street, Room 1115S Charlestown, MA 02129 ________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu [freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Iglesias, Eugenio [e.iglesias@ucl.ac.ukmailto:e.iglesias@ucl.ac.uk] Sent: Thursday, June 09, 2016 2:36 PM To: Freesurfer support list Subject: Re: [Freesurfer] Hippocampal subfields segmentation - what is the voxel size of the input?
Hi Fernando,
Unfortunately, FreeSurfer does downsample the T1s to 1mm, and the subfield code assumes that the T1 comes from FreeSurfer… For the paper, I hacked the code to take advantage of multiple images with higher resolution, but it hasn’t been thoroughly tested and it’s not publicly available at this point. That said, there are different things you can do at this point: 1. Use the T1+T2 as is (disadvantage: T1 will be resampled to 1 mm). 2. Use only the T2. 3. Use only the T1. To do this, you would use the T1 as additional scan on its own (i.e., the main and additional scans are the same, but the additional one won’t be resampled). 4. There might be a way (hack) of using both scans at full resolution. If you change the header of the T1 scan and specify that the voxel size is 1mm, no resampling will happen; FreeSurfer will believe that you’re dealing with a subject with a very large head. If you do this, you’ll also have to enlarge the voxel size of the T2 by a factor 1/0.7. Also, the resulting volumes will have to be divided by (1/0.7)^3 in order to correct for the “wrong” voxel size.
I hope this helps!
Cheers,
Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.comhttp://www.jeiglesias.com/ http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 19:23, Fernando Pasquini Santos <fernandopasq@gmail.commailto:fernandopasq@gmail.com> wrote:
Dear,
I have a doubt regarding the input T1 image used for the Hippocampal Subfield segmentation in Freesurfer 6.0 - https://surfer.nmr.mgh.harvard.edu/fswiki/HippocampalSubfields
In this documentation it says that I can use a standard 1mm T1 image or the same standard 1mm T1 image with an additional scan. However, I want to use a 0.7mm T1 image as input, with an additional T2 scan. When I run the program with it, the output images (nu.mgz and orig.mgz) are downsampled to 1mm voxel size. So, I don't know if this happens just because of the recon-all program or if the segmentation is being done in the downsampled T1 image or in the original 0.7mm.
By reading the article "A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI" I saw that, in the methods section, the algorithm is used to segment 0.6mm T1 images. But is this possible to do in the Freesurfer package? Obviously, doing the segmentation directly on 0.7mm data would be better than a downsampled version to 1mm...
Thanks,
Fernando Pasquini Santos PhD student in Dynamic Systems fernandosantos@pitt.edumailto:fernandosantos@pitt.edu fernando.pasquini.santos@usp.brmailto:fernando.pasquini.santos@usp.br _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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.
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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.
Hi all,
I'll give some feedback about the two methods you've mentioned. I also tried using a 0.5 T1 image and then:
*Changing the headers of the T1 and T2 images:* It worked for the 0.7mm T1, however, in the 0.5mm the tailarach aligment failed.
In both cases I had the warning "MRIsetRas2VoxFromMatrix(): voxels sizes are inconsistent"
Output:
talairach_afd -T 0.005 -xfm transforms/talairach.xfm
ERROR: talairach_afd: Talairach Transform: transforms/talairach.xfm ***FAILED*** (p=0.0434, pval=0.0034 < threshold=0.0050)
Manual Talairach alignment may be necessary, or include the -notal-check flag to skip this test, making sure the -notal-check flag follows -all or -autorecon1 in the command string. See: http://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/Talairach
*Using the -hires flag:*Although it is not sure if this is handling the original resolution, it also worked only for the 0.7mm, and for the 0.5mm it had a segmentation fault in the topology correction step.
Em sex, 10 de jun de 2016 às 03:37, Iglesias, Eugenio e.iglesias@ucl.ac.uk escreveu:
Hi Matt, Bruce actually mentioned this possibility yesterday. Good to hear it doesn’t crash when you run it! That said, I need to go through the code, and make sure that data are handled the way they are supposed to: the fact that it doesn’t crash doesn’t directly imply that it is working properly. Hopefully I’ll find time to do this next week… Cheers, Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 21:51, Hibert, Matthew Louis MHIBERT@mgh.harvard.edu wrote:
Hi Eugenio, Would it also be possible to use the -hires flag in the initial recon-all command to prevent the T1 from being downsampled and maintain the voxel size that the T1 was acquired at, and use that resulting recon for the hippocampal subfield segmentation? I've tested this with MEMPRAGE scans acquired at 0.75mm isotropic and coronal T2 anisotropic scans acquired at 0.33mm in plane with 1mm thick slices, and the results look reasonable. The "1mm" label files are correctly downsampled to 0.75mm instead of 1mm, and the higher resolution label files remain at 0.33mm resolution.
Thanks, Matt
Stufflebeam Lab Athinoula A. Martinos Center Massachusetts General Hospital 149 13th Street, Room 1115S Charlestown, MA 02129
*From:* freesurfer-bounces@nmr.mgh.harvard.edu [ freesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Iglesias, Eugenio [ e.iglesias@ucl.ac.uk] *Sent:* Thursday, June 09, 2016 2:36 PM *To:* Freesurfer support list *Subject:* Re: [Freesurfer] Hippocampal subfields segmentation - what is the voxel size of the input?
Hi Fernando,
Unfortunately, FreeSurfer does downsample the T1s to 1mm, and the subfield code assumes that the T1 comes from FreeSurfer… For the paper, I hacked the code to take advantage of multiple images with higher resolution, but it hasn’t been thoroughly tested and it’s not publicly available at this point. That said, there are different things you can do at this point:
- Use the T1+T2 as is (disadvantage: T1 will be resampled to 1 mm).
- Use only the T2.
- Use only the T1. To do this, you would use the T1 as additional scan on
its own (i.e., the main and additional scans are the same, but the additional one won’t be resampled). 4. There might be a way (hack) of using both scans at full resolution. If you change the header of the T1 scan and specify that the voxel size is 1mm, no resampling will happen; FreeSurfer will believe that you’re dealing with a subject with a very large head. If you do this, you’ll also have to enlarge the voxel size of the T2 by a factor 1/0.7. Also, the resulting volumes will have to be divided by (1/0.7)^3 in order to correct for the “wrong” voxel size.
I hope this helps!
Cheers,
Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 19:23, Fernando Pasquini Santos fernandopasq@gmail.com wrote:
Dear,
I have a doubt regarding the input T1 image used for the Hippocampal Subfield segmentation in Freesurfer 6.0 - https://surfer.nmr.mgh.harvard.edu/fswiki/HippocampalSubfields
In this documentation it says that I can use a standard 1mm T1 image or the same standard 1mm T1 image with an additional scan. However, I want to use a 0.7mm T1 image as input, with an additional T2 scan. When I run the program with it, the output images (nu.mgz and orig.mgz) are downsampled to 1mm voxel size. So, I don't know if this happens just because of the recon-all program or if the segmentation is being done in the downsampled T1 image or in the original 0.7mm.
By reading the article "A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI" I saw that, in the methods section, the algorithm is used to segment 0.6mm T1 images. But is this possible to do in the Freesurfer package? Obviously, doing the segmentation directly on 0.7mm data would be better than a downsampled version to 1mm...
Thanks,
Fernando Pasquini Santos PhD student in Dynamic Systems fernandosantos@pitt.edu fernando.pasquini.santos@usp.br _______________________________________________ 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.
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.
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.
Hi Eugenio, Have you had a chance to look into the possibility of combining the -hires flag with the hippocampal subfield segmentation?
Thanks, Matt ________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edu [freesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Iglesias, Eugenio [e.iglesias@ucl.ac.uk] Sent: Friday, June 10, 2016 3:36 AM To: Freesurfer support list Subject: Re: [Freesurfer] Hippocampal subfields segmentation - what is the voxel size of the input?
Hi Matt, Bruce actually mentioned this possibility yesterday. Good to hear it doesn’t crash when you run it! That said, I need to go through the code, and make sure that data are handled the way they are supposed to: the fact that it doesn’t crash doesn’t directly imply that it is working properly. Hopefully I’ll find time to do this next week… Cheers, Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 21:51, Hibert, Matthew Louis <MHIBERT@mgh.harvard.edumailto:MHIBERT@mgh.harvard.edu> wrote:
Hi Eugenio, Would it also be possible to use the -hires flag in the initial recon-all command to prevent the T1 from being downsampled and maintain the voxel size that the T1 was acquired at, and use that resulting recon for the hippocampal subfield segmentation? I've tested this with MEMPRAGE scans acquired at 0.75mm isotropic and coronal T2 anisotropic scans acquired at 0.33mm in plane with 1mm thick slices, and the results look reasonable. The "1mm" label files are correctly downsampled to 0.75mm instead of 1mm, and the higher resolution label files remain at 0.33mm resolution.
Thanks, Matt
Stufflebeam Lab Athinoula A. Martinos Center Massachusetts General Hospital 149 13th Street, Room 1115S Charlestown, MA 02129 ________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu [freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Iglesias, Eugenio [e.iglesias@ucl.ac.ukmailto:e.iglesias@ucl.ac.uk] Sent: Thursday, June 09, 2016 2:36 PM To: Freesurfer support list Subject: Re: [Freesurfer] Hippocampal subfields segmentation - what is the voxel size of the input?
Hi Fernando,
Unfortunately, FreeSurfer does downsample the T1s to 1mm, and the subfield code assumes that the T1 comes from FreeSurfer… For the paper, I hacked the code to take advantage of multiple images with higher resolution, but it hasn’t been thoroughly tested and it’s not publicly available at this point. That said, there are different things you can do at this point: 1. Use the T1+T2 as is (disadvantage: T1 will be resampled to 1 mm). 2. Use only the T2. 3. Use only the T1. To do this, you would use the T1 as additional scan on its own (i.e., the main and additional scans are the same, but the additional one won’t be resampled). 4. There might be a way (hack) of using both scans at full resolution. If you change the header of the T1 scan and specify that the voxel size is 1mm, no resampling will happen; FreeSurfer will believe that you’re dealing with a subject with a very large head. If you do this, you’ll also have to enlarge the voxel size of the T2 by a factor 1/0.7. Also, the resulting volumes will have to be divided by (1/0.7)^3 in order to correct for the “wrong” voxel size.
I hope this helps!
Cheers,
Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.comhttp://www.jeiglesias.com/ http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 19:23, Fernando Pasquini Santos <fernandopasq@gmail.commailto:fernandopasq@gmail.com> wrote:
Dear,
I have a doubt regarding the input T1 image used for the Hippocampal Subfield segmentation in Freesurfer 6.0 - https://surfer.nmr.mgh.harvard.edu/fswiki/HippocampalSubfields
In this documentation it says that I can use a standard 1mm T1 image or the same standard 1mm T1 image with an additional scan. However, I want to use a 0.7mm T1 image as input, with an additional T2 scan. When I run the program with it, the output images (nu.mgz and orig.mgz) are downsampled to 1mm voxel size. So, I don't know if this happens just because of the recon-all program or if the segmentation is being done in the downsampled T1 image or in the original 0.7mm.
By reading the article "A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI" I saw that, in the methods section, the algorithm is used to segment 0.6mm T1 images. But is this possible to do in the Freesurfer package? Obviously, doing the segmentation directly on 0.7mm data would be better than a downsampled version to 1mm...
Thanks,
Fernando Pasquini Santos PhD student in Dynamic Systems fernandosantos@pitt.edumailto:fernandosantos@pitt.edu fernando.pasquini.santos@usp.brmailto:fernando.pasquini.santos@usp.br _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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.
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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.
Not yet, sorry. But it’s on my todo list.
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/
On 6 Jul 2016, at 14:39, Hibert, Matthew Louis <MHIBERT@mgh.harvard.edumailto:MHIBERT@mgh.harvard.edu> wrote:
Hi Eugenio, Have you had a chance to look into the possibility of combining the -hires flag with the hippocampal subfield segmentation?
Thanks, Matt ________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu [freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Iglesias, Eugenio [e.iglesias@ucl.ac.ukmailto:e.iglesias@ucl.ac.uk] Sent: Friday, June 10, 2016 3:36 AM To: Freesurfer support list Subject: Re: [Freesurfer] Hippocampal subfields segmentation - what is the voxel size of the input?
Hi Matt, Bruce actually mentioned this possibility yesterday. Good to hear it doesn’t crash when you run it! That said, I need to go through the code, and make sure that data are handled the way they are supposed to: the fact that it doesn’t crash doesn’t directly imply that it is working properly. Hopefully I’ll find time to do this next week… Cheers, Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.comhttp://www.jeiglesias.com/ http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 21:51, Hibert, Matthew Louis <MHIBERT@mgh.harvard.edumailto:MHIBERT@mgh.harvard.edu> wrote:
Hi Eugenio, Would it also be possible to use the -hires flag in the initial recon-all command to prevent the T1 from being downsampled and maintain the voxel size that the T1 was acquired at, and use that resulting recon for the hippocampal subfield segmentation? I've tested this with MEMPRAGE scans acquired at 0.75mm isotropic and coronal T2 anisotropic scans acquired at 0.33mm in plane with 1mm thick slices, and the results look reasonable. The "1mm" label files are correctly downsampled to 0.75mm instead of 1mm, and the higher resolution label files remain at 0.33mm resolution.
Thanks, Matt
Stufflebeam Lab Athinoula A. Martinos Center Massachusetts General Hospital 149 13th Street, Room 1115S Charlestown, MA 02129 ________________________________ From: freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu [freesurfer-bounces@nmr.mgh.harvard.edumailto:freesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Iglesias, Eugenio [e.iglesias@ucl.ac.ukmailto:e.iglesias@ucl.ac.uk] Sent: Thursday, June 09, 2016 2:36 PM To: Freesurfer support list Subject: Re: [Freesurfer] Hippocampal subfields segmentation - what is the voxel size of the input?
Hi Fernando,
Unfortunately, FreeSurfer does downsample the T1s to 1mm, and the subfield code assumes that the T1 comes from FreeSurfer… For the paper, I hacked the code to take advantage of multiple images with higher resolution, but it hasn’t been thoroughly tested and it’s not publicly available at this point. That said, there are different things you can do at this point: 1. Use the T1+T2 as is (disadvantage: T1 will be resampled to 1 mm). 2. Use only the T2. 3. Use only the T1. To do this, you would use the T1 as additional scan on its own (i.e., the main and additional scans are the same, but the additional one won’t be resampled). 4. There might be a way (hack) of using both scans at full resolution. If you change the header of the T1 scan and specify that the voxel size is 1mm, no resampling will happen; FreeSurfer will believe that you’re dealing with a subject with a very large head. If you do this, you’ll also have to enlarge the voxel size of the T2 by a factor 1/0.7. Also, the resulting volumes will have to be divided by (1/0.7)^3 in order to correct for the “wrong” voxel size.
I hope this helps!
Cheers,
Eugenio
Juan Eugenio Iglesias Translational Imaging Group University College London http://www.jeiglesias.comhttp://www.jeiglesias.com/ http://cmictig.cs.ucl.ac.uk/
On 9 Jun 2016, at 19:23, Fernando Pasquini Santos <fernandopasq@gmail.commailto:fernandopasq@gmail.com> wrote:
Dear,
I have a doubt regarding the input T1 image used for the Hippocampal Subfield segmentation in Freesurfer 6.0 - https://surfer.nmr.mgh.harvard.edu/fswiki/HippocampalSubfields
In this documentation it says that I can use a standard 1mm T1 image or the same standard 1mm T1 image with an additional scan. However, I want to use a 0.7mm T1 image as input, with an additional T2 scan. When I run the program with it, the output images (nu.mgz and orig.mgz) are downsampled to 1mm voxel size. So, I don't know if this happens just because of the recon-all program or if the segmentation is being done in the downsampled T1 image or in the original 0.7mm.
By reading the article "A computational atlas of the hippocampal formation using ex vivo, ultra-high resolution MRI: Application to adaptive segmentation of in vivo MRI" I saw that, in the methods section, the algorithm is used to segment 0.6mm T1 images. But is this possible to do in the Freesurfer package? Obviously, doing the segmentation directly on 0.7mm data would be better than a downsampled version to 1mm...
Thanks,
Fernando Pasquini Santos PhD student in Dynamic Systems fernandosantos@pitt.edumailto:fernandosantos@pitt.edu fernando.pasquini.santos@usp.brmailto:fernando.pasquini.santos@usp.br _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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.
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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.
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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.
freesurfer@nmr.mgh.harvard.edu
-
Fernando Pasquini Santos -
Hibert, Matthew Louis -
Iglesias, Eugenio -
Matt Glasser