Dear list,
I am about to start processing a largish longitudinal dataset and wondered whether I could save some processing time in the following ways:
1) I have seen written in a few places that only norm.mgz is used from the cross-sectional stream during base template creation at the next stage. Does that mean I can run "-autorecon1" only on each of the timepoints during the cross-sectional processing stage rather than "-all"? 2) As I am only interested in the subcortical output of the longitudinal processing, can I include the flags "-autorecon1 -subcortseg" in all remaining steps in the longitudinal stream to skip cortical processing, or will that cause problems?
Thanks for your help, Jordan
Jordan, The -subcortseg step is going to be looking for inputs that were created before it, so you can't really skip the steps before it. I suppose you can skip the surface steps afterwards as long as you understand that the aseg.mgz will not give you accurate wm and gm volumes.
In order to get the volume stats, you need to run the -segstats step in -autorecon3. I'm not sure if you will get any complaints for missing files from not running the surfaces.
I'm also not sure if there is anything in the longitudinal processing stream that would complain about this method - but hopefully Martin or Nick will chime in about that. Allison
--
On Thu, 15 Apr 2010, Jordan Poppenk wrote:
Dear list,
I am about to start processing a largish longitudinal dataset and wondered whether I could save some processing time in the following ways:
- I have seen written in a few places that only norm.mgz is used from the
cross-sectional stream during base template creation at the next stage. Does that mean I can run "-autorecon1" only on each of the timepoints during the cross-sectional processing stage rather than "-all"? 2) As I am only interested in the subcortical output of the longitudinal processing, can I include the flags "-autorecon1 -subcortseg" in all remaining steps in the longitudinal stream to skip cortical processing, or will that cause problems?
Thanks for your help, Jordan
Thank you Allison. Does anyone have info about the longitudinal aspect of this question? Imagine this general approach:
1) run autorecon1 on all time points (skip autorecon2-3 since only norm.mgz is needed at this stage)
2) create base template, running autorecon1 and autorecon2 (skip autorecon3 since I am only interested in subcortical parcellation)
3) use template to generate longitudinal timepoints with autorecon1 and autorecon2 (again skipping autorecon3)
Would that procedure produce accurate subcortical volumes for each timepoint in aseg.mgz? It could save a lot of processing time for a large dataset.
Cheers, Jordan
On 16-Apr-10, at 9:06 AM, Allison Stevens wrote:
Jordan, The -subcortseg step is going to be looking for inputs that were created before it, so you can't really skip the steps before it. I suppose you can skip the surface steps afterwards as long as you understand that the aseg.mgz will not give you accurate wm and gm volumes.
In order to get the volume stats, you need to run the -segstats step in -autorecon3. I'm not sure if you will get any complaints for missing files from not running the surfaces.
I'm also not sure if there is anything in the longitudinal processing stream that would complain about this method - but hopefully Martin or Nick will chime in about that. Allison
--
On Thu, 15 Apr 2010, Jordan Poppenk wrote:
Dear list,
I am about to start processing a largish longitudinal dataset and wondered whether I could save some processing time in the following ways:
- I have seen written in a few places that only norm.mgz is used
from the cross-sectional stream during base template creation at the next stage. Does that mean I can run "-autorecon1" only on each of the timepoints during the cross-sectional processing stage rather than "-all"? 2) As I am only interested in the subcortical output of the longitudinal processing, can I include the flags "-autorecon1 -subcortseg" in all remaining steps in the longitudinal stream to skip cortical processing, or will that cause problems?
Thanks for your help, Jordan
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.
yes, this approach will work if you are only interested in subcortical structures. but an improvement is to use -autorecon2-volonly instead of -autorecon2. the former stops prior to the surface stage.
n.
On Tue, 2010-04-27 at 15:56 -0400, Jordan Poppenk wrote:
Thank you Allison. Does anyone have info about the longitudinal aspect of this question? Imagine this general approach:
- run autorecon1 on all time points (skip autorecon2-3 since only
norm.mgz is needed at this stage)
- create base template, running autorecon1 and autorecon2 (skip
autorecon3 since I am only interested in subcortical parcellation)
- use template to generate longitudinal timepoints with autorecon1
and autorecon2 (again skipping autorecon3)
Would that procedure produce accurate subcortical volumes for each timepoint in aseg.mgz? It could save a lot of processing time for a large dataset.
Cheers, Jordan
On 16-Apr-10, at 9:06 AM, Allison Stevens wrote:
Jordan, The -subcortseg step is going to be looking for inputs that were created before it, so you can't really skip the steps before it. I suppose you can skip the surface steps afterwards as long as you understand that the aseg.mgz will not give you accurate wm and gm volumes.
In order to get the volume stats, you need to run the -segstats step in -autorecon3. I'm not sure if you will get any complaints for missing files from not running the surfaces.
I'm also not sure if there is anything in the longitudinal processing stream that would complain about this method - but hopefully Martin or Nick will chime in about that. Allison
--
On Thu, 15 Apr 2010, Jordan Poppenk wrote:
Dear list,
I am about to start processing a largish longitudinal dataset and wondered whether I could save some processing time in the following ways:
- I have seen written in a few places that only norm.mgz is used
from the cross-sectional stream during base template creation at the next stage. Does that mean I can run "-autorecon1" only on each of the timepoints during the cross-sectional processing stage rather than "-all"? 2) As I am only interested in the subcortical output of the longitudinal processing, can I include the flags "-autorecon1 -subcortseg" in all remaining steps in the longitudinal stream to skip cortical processing, or will that cause problems?
Thanks for your help, Jordan
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
Hi Jordan,
this won't work.
It is true, that you need only the norm.mgz for the base run, but you need the aseg.mgz and all corresponding transforms of each time point for the longitudinal runs.
There might still be a possibility to save time, by only doing the subcortical processing in each time point cross sectionally and in the base. And then do the longitudinals as described below. This has never been tested.
Best, Martin.
On Tue, 2010-04-27 at 16:05 -0400, Nick Schmansky wrote:
yes, this approach will work if you are only interested in subcortical structures. but an improvement is to use -autorecon2-volonly instead of -autorecon2. the former stops prior to the surface stage.
n.
On Tue, 2010-04-27 at 15:56 -0400, Jordan Poppenk wrote:
Thank you Allison. Does anyone have info about the longitudinal aspect of this question? Imagine this general approach:
- run autorecon1 on all time points (skip autorecon2-3 since only
norm.mgz is needed at this stage)
- create base template, running autorecon1 and autorecon2 (skip
autorecon3 since I am only interested in subcortical parcellation)
- use template to generate longitudinal timepoints with autorecon1
and autorecon2 (again skipping autorecon3)
Would that procedure produce accurate subcortical volumes for each timepoint in aseg.mgz? It could save a lot of processing time for a large dataset.
Cheers, Jordan
On 16-Apr-10, at 9:06 AM, Allison Stevens wrote:
Jordan, The -subcortseg step is going to be looking for inputs that were created before it, so you can't really skip the steps before it. I suppose you can skip the surface steps afterwards as long as you understand that the aseg.mgz will not give you accurate wm and gm volumes.
In order to get the volume stats, you need to run the -segstats step in -autorecon3. I'm not sure if you will get any complaints for missing files from not running the surfaces.
I'm also not sure if there is anything in the longitudinal processing stream that would complain about this method - but hopefully Martin or Nick will chime in about that. Allison
--
On Thu, 15 Apr 2010, Jordan Poppenk wrote:
Dear list,
I am about to start processing a largish longitudinal dataset and wondered whether I could save some processing time in the following ways:
- I have seen written in a few places that only norm.mgz is used
from the cross-sectional stream during base template creation at the next stage. Does that mean I can run "-autorecon1" only on each of the timepoints during the cross-sectional processing stage rather than "-all"? 2) As I am only interested in the subcortical output of the longitudinal processing, can I include the flags "-autorecon1 -subcortseg" in all remaining steps in the longitudinal stream to skip cortical processing, or will that cause problems?
Thanks for your help, Jordan
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
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Hi Nick and Martin,
In that case, in a few subjects, I'll compare the outcome of this limited approach to the results of the full processing stream prior to proceeding. Thank you for your help and suggestions!
Cheers, Jordan
On 27-Apr-10, at 4:17 PM, Martin Reuter wrote:
Hi Jordan,
this won't work.
It is true, that you need only the norm.mgz for the base run, but you need the aseg.mgz and all corresponding transforms of each time point for the longitudinal runs.
There might still be a possibility to save time, by only doing the subcortical processing in each time point cross sectionally and in the base. And then do the longitudinals as described below. This has never been tested.
Best, Martin.
On Tue, 2010-04-27 at 16:05 -0400, Nick Schmansky wrote:
yes, this approach will work if you are only interested in subcortical structures. but an improvement is to use -autorecon2-volonly instead of -autorecon2. the former stops prior to the surface stage.
n.
On Tue, 2010-04-27 at 15:56 -0400, Jordan Poppenk wrote:
Thank you Allison. Does anyone have info about the longitudinal aspect of this question? Imagine this general approach:
- run autorecon1 on all time points (skip autorecon2-3 since only
norm.mgz is needed at this stage)
- create base template, running autorecon1 and autorecon2 (skip
autorecon3 since I am only interested in subcortical parcellation)
- use template to generate longitudinal timepoints with autorecon1
and autorecon2 (again skipping autorecon3)
Would that procedure produce accurate subcortical volumes for each timepoint in aseg.mgz? It could save a lot of processing time for a large dataset.
Cheers, Jordan
On 16-Apr-10, at 9:06 AM, Allison Stevens wrote:
Jordan, The -subcortseg step is going to be looking for inputs that were created before it, so you can't really skip the steps before it. I suppose you can skip the surface steps afterwards as long as you understand that the aseg.mgz will not give you accurate wm and gm volumes.
In order to get the volume stats, you need to run the -segstats step in -autorecon3. I'm not sure if you will get any complaints for missing files from not running the surfaces.
I'm also not sure if there is anything in the longitudinal processing stream that would complain about this method - but hopefully Martin or Nick will chime in about that. Allison
--
On Thu, 15 Apr 2010, Jordan Poppenk wrote:
Dear list,
I am about to start processing a largish longitudinal dataset and wondered whether I could save some processing time in the following ways:
- I have seen written in a few places that only norm.mgz is used
from the cross-sectional stream during base template creation at the next stage. Does that mean I can run "-autorecon1" only on each of the timepoints during the cross-sectional processing stage rather than "-all"? 2) As I am only interested in the subcortical output of the longitudinal processing, can I include the flags "-autorecon1 -subcortseg" in all remaining steps in the longitudinal stream to skip cortical processing, or will that cause problems?
Thanks for your help, Jordan
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
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Allison Stevens -
Jordan Poppenk -
Martin Reuter -
Nick Schmansky