Hi List,
On 3. Nov 2006, at 19:48 Uhr, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 10 Date: Fri, 03 Nov 2006 13:48:18 -0500 From: Nick Schmansky nicks@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] parameters for N3 correction To: Chacko Cherian chacko_cherian@yahoo.com Cc: Freesurfer Mailing List freesurfer@nmr.mgh.harvard.edu Message-ID: 1162579698.27296.26.camel@minerva.nmr.mgh.harvard.edu Content-Type: text/plain
also, the command:
recon-all -s <subjid> -nuintensitycor
will run just that stage.
Something relted on the topic of N3. Using the default settings with surface coils (RX and TX through the same coil in our case) ends up looking ugly no matter the number of -nuiterations specified in recon-all. Hacking mri_nu_correct.mni to include "-iterations 150 -stop 0.0001" in the invocation of nu_correct, ends up with better results. Just as the N3 documentation recommends for surface coils ;). Now would it be possible to get a "-surfacecoil" argument for mri_nu_correct.mni (or even better recon_all via the experts options mechanism?) included into the distribution. If yes I would be willing to hack something up (at least including this properly into mri_nu_correct.mni). This might be helpful for other surface coil users out there...
ahoi & sorry
do you mean single channel surface coils? Can you get decent surfaces out of that kind of data??? On Fri, 3 Nov 2006, Sebastian Moeller wrote:
Hi List,
On 3. Nov 2006, at 19:48 Uhr, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 10 Date: Fri, 03 Nov 2006 13:48:18 -0500 From: Nick Schmansky nicks@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] parameters for N3 correction To: Chacko Cherian chacko_cherian@yahoo.com Cc: Freesurfer Mailing List freesurfer@nmr.mgh.harvard.edu Message-ID: 1162579698.27296.26.camel@minerva.nmr.mgh.harvard.edu Content-Type: text/plain
also, the command:
recon-all -s <subjid> -nuintensitycor
will run just that stage.
Something relted on the topic of N3. Using the default settings with surface coils (RX and TX through the same coil in our case) ends up looking ugly no matter the number of -nuiterations specified in recon-all. Hacking mri_nu_correct.mni to include "-iterations 150 -stop 0.0001" in the invocation of nu_correct, ends up with better results. Just as the N3 documentation recommends for surface coils ;). Now would it be possible to get a "-surfacecoil" argument for mri_nu_correct.mni (or even better recon_all via the experts options mechanism?) included into the distribution. If yes I would be willing to hack something up (at least including this properly into mri_nu_correct.mni). This might be helpful for other surface coil users out there...
ahoi & sorry
Hi Bruce,
On 3. Nov 2006, at 20:33 Uhr, Bruce Fischl wrote:
do you mean single channel surface coils? Can you get decent surfaces out of that kind of data???
Well, yes, and that depends on the definition of decency. Joking aside, it requires ship-loads of manual intervention (think manually edit all "slices" for brain-mask, manually deploy a ton of control points (high 4 digit counts), and last but not least edit the wm into compliance). Then again, it seems all monkey people relay on manual steps on the way to a good surface. I am intersted to send you one of my "specimens" to get any comments.
ahoi Sebastian
On Fri, 3 Nov 2006, Sebastian Moeller wrote:
Hi List,
On 3. Nov 2006, at 19:48 Uhr, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 10 Date: Fri, 03 Nov 2006 13:48:18 -0500 From: Nick Schmansky nicks@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] parameters for N3 correction To: Chacko Cherian chacko_cherian@yahoo.com Cc: Freesurfer Mailing List freesurfer@nmr.mgh.harvard.edu Message-ID: 1162579698.27296.26.camel@minerva.nmr.mgh.harvard.edu Content-Type: text/plain also, the command: recon-all -s <subjid> -nuintensitycor will run just that stage.
Something relted on the topic of N3. Using the default settings with surface coils (RX and TX through the same coil in our case) ends up looking ugly no matter the number of -nuiterations specified in recon-all. Hacking mri_nu_correct.mni to include "-iterations 150 -stop 0.0001" in the invocation of nu_correct, ends up with better results. Just as the N3 documentation recommends for surface coils ;). Now would it be possible to get a "-surfacecoil" argument for mri_nu_correct.mni (or even better recon_all via the experts options mechanism?) included into the distribution. If yes I would be willing to hack something up (at least including this properly into mri_nu_correct.mni). This might be helpful for other surface coil users out there...
ahoi & sorry
feel free to put one on our drop site On Fri, 3 Nov 2006, Sebastian Moeller wrote:
Hi Bruce,
On 3. Nov 2006, at 20:33 Uhr, Bruce Fischl wrote:
do you mean single channel surface coils? Can you get decent surfaces out of that kind of data???
Well, yes, and that depends on the definition of decency. Joking aside, it requires ship-loads of manual intervention (think manually edit all "slices" for brain-mask, manually deploy a ton of control points (high 4 digit counts), and last but not least edit the wm into compliance). Then again, it seems all monkey people relay on manual steps on the way to a good surface. I am intersted to send you one of my "specimens" to get any comments.
ahoi Sebastian
On Fri, 3 Nov 2006, Sebastian Moeller wrote:
Hi List,
On 3. Nov 2006, at 19:48 Uhr, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 10 Date: Fri, 03 Nov 2006 13:48:18 -0500 From: Nick Schmansky nicks@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] parameters for N3 correction To: Chacko Cherian chacko_cherian@yahoo.com Cc: Freesurfer Mailing List freesurfer@nmr.mgh.harvard.edu Message-ID: 1162579698.27296.26.camel@minerva.nmr.mgh.harvard.edu Content-Type: text/plain also, the command: recon-all -s <subjid> -nuintensitycor will run just that stage.
Something relted on the topic of N3. Using the default settings with surface coils (RX and TX through the same coil in our case) ends up looking ugly no matter the number of -nuiterations specified in recon-all. Hacking mri_nu_correct.mni to include "-iterations 150 -stop 0.0001" in the invocation of nu_correct, ends up with better results. Just as the N3 documentation recommends for surface coils ;). Now would it be possible to get a "-surfacecoil" argument for mri_nu_correct.mni (or even better recon_all via the experts options mechanism?) included into the distribution. If yes I would be willing to hack something up (at least including this properly into mri_nu_correct.mni). This might be helpful for other surface coil users out there...
ahoi & sorry
Hi Bruce,
On 3. Nov 2006, at 21:43 Uhr, Bruce Fischl wrote:
feel free to put one on our drop site
Great offer; I am uploading test.tar.bz2 (onto surfer.nmr.mgh.harvard.edu/transfer/incoming) which contains a T1.mgz (in test/mri) and all surfaces (in test/surf). I reduced the set to those files to keep the file size small. So far we only used the surfaces to display different functional statistical maps, the quality seems to be sufficient for that; I have not tried anything resembling morphometry so I have no idea about the qualification of our processing for that purpose ;). And pointers how to improve the quality are welcome.
ahoi Sebastian
On Fri, 3 Nov 2006, Sebastian Moeller wrote:
Hi Bruce,
On 3. Nov 2006, at 20:33 Uhr, Bruce Fischl wrote:
do you mean single channel surface coils? Can you get decent surfaces out of that kind of data???
Well, yes, and that depends on the definition of decency. Joking aside, it requires ship-loads of manual intervention (think manually edit all "slices" for brain-mask, manually deploy a ton of control points (high 4 digit counts), and last but not least edit the wm into compliance). Then again, it seems all monkey people relay on manual steps on the way to a good surface. I am intersted to send you one of my "specimens" to get any comments.
ahoi Sebastian
On Fri, 3 Nov 2006, Sebastian Moeller wrote:
Hi List, On 3. Nov 2006, at 19:48 Uhr, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 10 Date: Fri, 03 Nov 2006 13:48:18 -0500 From: Nick Schmansky nicks@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] parameters for N3 correction To: Chacko Cherian chacko_cherian@yahoo.com Cc: Freesurfer Mailing List freesurfer@nmr.mgh.harvard.edu Message-ID: 1162579698.27296.26.camel@minerva.nmr.mgh.harvard.edu Content-Type: text/plain also, the command: recon-all -s <subjid> -nuintensitycor will run just that stage.
Something relted on the topic of N3. Using the default settings with surface coils (RX and TX through the same coil in our case) ends up looking ugly no matter the number of -nuiterations specified in recon-all. Hacking mri_nu_correct.mni to include "-iterations 150 -stop 0.0001" in the invocation of nu_correct, ends up with better results. Just as the N3 documentation recommends for surface coils ;). Now would it be possible to get a "-surfacecoil" argument for mri_nu_correct.mni (or even better recon_all via the experts options mechanism?) included into the distribution. If yes I would be willing to hack something up (at least including this properly into mri_nu_correct.mni). This might be helpful for other surface coil users out there... ahoi & sorry
I just added the relevant stuff to recon-all and mri_nu_correct.mni. I've put the new versions at:
ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve
along with a sample xopts file.
Would you test this out for us? Make sure to make backups of your current versions of recon-all and mri_nu_correct.mni :).
doug
Sebastian Moeller wrote:
Hi List,
On 3. Nov 2006, at 19:48 Uhr, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 10 Date: Fri, 03 Nov 2006 13:48:18 -0500 From: Nick Schmansky nicks@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] parameters for N3 correction To: Chacko Cherian chacko_cherian@yahoo.com Cc: Freesurfer Mailing List freesurfer@nmr.mgh.harvard.edu Message-ID: 1162579698.27296.26.camel@minerva.nmr.mgh.harvard.edu Content-Type: text/plain
also, the command:
recon-all -s <subjid> -nuintensitycor
will run just that stage.
Something relted on the topic of N3. Using the default settingswith surface coils (RX and TX through the same coil in our case) ends up looking ugly no matter the number of -nuiterations specified in recon-all. Hacking mri_nu_correct.mni to include "-iterations 150 -stop 0.0001" in the invocation of nu_correct, ends up with better results. Just as the N3 documentation recommends for surface coils ;). Now would it be possible to get a "-surfacecoil" argument for mri_nu_correct.mni (or even better recon_all via the experts options mechanism?) included into the distribution. If yes I would be willing to hack something up (at least including this properly into mri_nu_correct.mni). This might be helpful for other surface coil users out there...
ahoi & sorry
Hi Doug, Hi List,
On 3. Nov 2006, at 21:03 Uhr, Doug Greve wrote:
I just added the relevant stuff to recon-all and mri_nu_correct.mni. I've put the new versions at:
ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve
along with a sample xopts file.
Would you test this out for us? Make sure to make backups of your current versions of recon-all and mri_nu_correct.mni :).
Wahoo, I am impressed, that was fast; that works for me and is much nicer than my unconditional hack to always set those parameters.
thanks & ahoi Sebastian
doug
Sebastian Moeller wrote:
Hi List,
On 3. Nov 2006, at 19:48 Uhr, freesurfer-request@nmr.mgh.harvard.edu wrote:
Message: 10 Date: Fri, 03 Nov 2006 13:48:18 -0500 From: Nick Schmansky nicks@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] parameters for N3 correction To: Chacko Cherian chacko_cherian@yahoo.com Cc: Freesurfer Mailing List freesurfer@nmr.mgh.harvard.edu Message-ID: 1162579698.27296.26.camel@minerva.nmr.mgh.harvard.edu Content-Type: text/plain
also, the command:
recon-all -s <subjid> -nuintensitycor
will run just that stage.
Something relted on the topic of N3. Using the default settingswith surface coils (RX and TX through the same coil in our case) ends up looking ugly no matter the number of -nuiterations specified in recon-all. Hacking mri_nu_correct.mni to include "-iterations 150 -stop 0.0001" in the invocation of nu_correct, ends up with better results. Just as the N3 documentation recommends for surface coils ;). Now would it be possible to get a "-surfacecoil" argument for mri_nu_correct.mni (or even better recon_all via the experts options mechanism?) included into the distribution. If yes I would be willing to hack something up (at least including this properly into mri_nu_correct.mni). This might be helpful for other surface coil users out there...
ahoi & sorry
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
In order to help us help you, please follow the steps in: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
freesurfer@nmr.mgh.harvard.edu
-
Bruce Fischl -
Doug Greve -
Sebastian Moeller