Dear FS list,
I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from an analysis point of view, this still qualifies as longitudinal. I want to check for
- the main within-subjects effect of time point (pre&post) - the main between-subjects effect of group (treatment A, treatment B, control), - the time x group interaction
I intend to look at thickness, surface area, volume, and lGI.
I read on the wikihttp://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessingthat FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis to supplement a VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are:
1) Can QDEC be used for such an analysis, and if so, what would be different as compared to a cross-sectional (no temporal/within factor) study?
2) Also, is the pre-processing stage any different?
3) In FSL, for longitudinal designs you do stats on images obtained as the difference between consecutive time points. Does this have to be done in FreeSurfer as well, and if so, is it done at the recon-all level or only at the stats (QDEC) level?
Thanks!
Tudor
Sorry for the repeat, wasn't sure whether this was received the first time. Tudor
On 6 May 2014 19:55, Tudor Popescu tudor3@gmail.com wrote:
Dear FS list,
I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from an analysis point of view, this still qualifies as longitudinal. I want to check for
- the main within-subjects effect of time point (pre&post)
- the main between-subjects effect of group (treatment A, treatment B,
control),
- the time x group interaction
I intend to look at thickness, surface area, volume, and lGI.
I read on the wikihttp://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessingthat FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis to supplement a VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are:
- Can QDEC be used for such an analysis, and if so, what would be
different as compared to a cross-sectional (no temporal/within factor) study?
Also, is the pre-processing stage any different?
In FSL, for longitudinal designs you do stats on images obtained as the
difference between consecutive time points. Does this have to be done in FreeSurfer as well, and if so, is it done at the recon-all level or only at the stats (QDEC) level?
Thanks!
Tudor
Hi Tudor
I'll leave ths stats question to Martin or Jorge, but I'm not sure why the wiki would say out longitudinal stream is not optimal. It is certainly as optimal as we have had the resources to make it, although there are a long list of things we would still like to do.
cheers Bruce
On Thu, 8 May 2014, Tudor Popescu wrote:
Sorry for the repeat, wasn't sure whether this was received the first time. Tudor
On 6 May 2014 19:55, Tudor Popescu tudor3@gmail.com wrote: Dear FS list,
I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from an analysis point of view, this still qualifies as longitudinal. I want to check for
- the main within-subjects effect of time point (pre&post)
- the main between-subjects effect of group (treatment A, treatment B, control),
- the time x group interaction
I intend to look at thickness, surface area, volume, and lGI.
I read on the wiki that FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis to supplement a VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are:
- Can QDEC be used for such an analysis, and if so, what would be
different as compared to a cross-sectional (no temporal/within factor) study?
Also, is the pre-processing stage any different?
In FSL, for longitudinal designs you do stats on images obtained as
the difference between consecutive time points. Does this have to be done in FreeSurfer as well, and if so, is it done at the recon-all level or only at the stats (QDEC) level?
Thanks!
Tudor
Hi Tudor,
the longitudinal pipeline in FS is actually one of the best on the planet as far as I know :-). If there is any contradictory information on the wiki, can you point me to that so I can see what causes the misconception. Really: compared to independent processing, it significantly increases sensitivity. Also we have designed it to be unbiased with respect to a single time point or directionality. It is quite mature by now.
You should definitely use the longitudinal pipeline for the analysis of your data. Now to your questions
1. QDEC: I am not too familiar with qdec. You can definitely try the 2-stage approach described on the wiki. There you first compute a measure of change (e.g. hippocampal volume change during your week) and then compare that measure across groups similar to a cross sectional volume/thickness analysis. You can also use our tools to run a linear mixed effects model if you want to do that (it is more involved and requires you to use matlab tools). In your case, you probably have 2 time points for all subjects and the time distance is probably the same for all subjects, so the 2-stage approach should be fine.
2. The image processing is done via the longitudinal pipeline (three steps: cross, base, long), to prepare the data look at the description of the 2-stage model http://freesurfer.net/fswiki/LongitudinalTwoStageModel and also the longitudinal tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial
3. At the recon all level in FS you get (after the 3 steps) measurement for all time points. So you would compare those results across time in the stats.
Hope that helps, Martin
On 05/08/2014 08:14 AM, Tudor Popescu wrote:
Sorry for the repeat, wasn't sure whether this was received the first time. Tudor
On 6 May 2014 19:55, Tudor Popescu <tudor3@gmail.com mailto:tudor3@gmail.com> wrote:
Dear FS list, I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from an analysis point of view, this still qualifies as longitudinal. I want to check for * the main within-subjects effect of time point (pre&post) * the main between-subjects effect of group (treatment A, treatment B, control), * the time x group interaction I intend to look at thickness, surface area, volume, and lGI. I read on the wiki <http://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing> that FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis to supplement a VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are: 1) Can QDEC be used for such an analysis, and if so, what would be different as compared to a cross-sectional (no temporal/within factor) study? 2) Also, is the pre-processing stage any different? 3) In FSL, for longitudinal designs you do stats on images obtained as the difference between consecutive time points. Does this have to be done in FreeSurfer as well, and if so, is it done at the recon-all level or only at the stats (QDEC) level? Thanks! Tudor
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Thanks very much Martin and Bruce. I guess I'd misread the Wiki (my own fault, not the text's), and am glad to hear that the longitudinal pipeline is in fact perfectly suitable for my needs here.
Having run the first 2 steps (cross and base), I'm a bit unclear how the output so far has to be manually inspected. It says in the tutorialhttp://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorialthat you should load each subject's base volume + surfs in freeview and then "move back and forth a few slices". However, even just loading each base in this manner takes ~1 min on my PC (CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving with PgUp/PgDn between all coronal slices (starting from the default slice=128, going all the way posterior and then all the way anterior) is excruciatingly slow. All of this would have to be repeated for all my 72 subjects - is there any way to optimise this manual inspection?
Also, if the surfs turn out to not follow the volume correctly, presumably the thing to do is white surface correction + re-running recon. But what should one do if, due to an erroneous averaging between timepoints, you see blurs/ghosts in your base template?
Many thanks! Tudor
On 9 May 2014 21:33, Martin Reuter mreuter@nmr.mgh.harvard.edu wrote:
Hi Tudor,
the longitudinal pipeline in FS is actually one of the best on the planet as far as I know :-). If there is any contradictory information on the wiki, can you point me to that so I can see what causes the misconception. Really: compared to independent processing, it significantly increases sensitivity. Also we have designed it to be unbiased with respect to a single time point or directionality. It is quite mature by now.
You should definitely use the longitudinal pipeline for the analysis of your data. Now to your questions
- QDEC: I am not too familiar with qdec. You can definitely try the
2-stage approach described on the wiki. There you first compute a measure of change (e.g. hippocampal volume change during your week) and then compare that measure across groups similar to a cross sectional volume/thickness analysis. You can also use our tools to run a linear mixed effects model if you want to do that (it is more involved and requires you to use matlab tools). In your case, you probably have 2 time points for all subjects and the time distance is probably the same for all subjects, so the 2-stage approach should be fine.
- The image processing is done via the longitudinal pipeline (three
steps: cross, base, long), to prepare the data look at the description of the 2-stage model http://freesurfer.net/fswiki/LongitudinalTwoStageModel and also the longitudinal tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial
- At the recon all level in FS you get (after the 3 steps) measurement
for all time points. So you would compare those results across time in the stats.
Hope that helps, Martin
On 05/08/2014 08:14 AM, Tudor Popescu wrote:
Sorry for the repeat, wasn't sure whether this was received the first time. Tudor
On 6 May 2014 19:55, Tudor Popescu tudor3@gmail.com wrote:
Dear FS list,
I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from an analysis point of view, this still qualifies as longitudinal. I want to check for
- the main within-subjects effect of time point (pre&post)
- the main between-subjects effect of group (treatment A, treatment
B, control),
- the time x group interaction
I intend to look at thickness, surface area, volume, and lGI.
I read on the wikihttp://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessingthat FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis to supplement a VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are:
- Can QDEC be used for such an analysis, and if so, what would be
different as compared to a cross-sectional (no temporal/within factor) study?
Also, is the pre-processing stage any different?
In FSL, for longitudinal designs you do stats on images obtained as
the difference between consecutive time points. Does this have to be done in FreeSurfer as well, and if so, is it done at the recon-all level or only at the stats (QDEC) level?
Thanks!
Tudor
Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 Martin, Wasn't sure whether you'd seen my reply below, look forward to hear back your thoughts, thanks! Tudor On 25 May 2014 21:40, "Tudor Popescu" tudor3@gmail.com wrote:
Thanks very much Martin and Bruce. I guess I'd misread the Wiki (my own fault, not the text's), and am glad to hear that the longitudinal pipeline is in fact perfectly suitable for my needs here.
Having run the first 2 steps (cross and base), I'm a bit unclear how the output so far has to be manually inspected. It says in the tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial that you should load each subject's base volume + surfs in freeview and then "move back and forth a few slices". However, even just loading each base in this manner takes ~1 min on my PC (CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving with PgUp/PgDn between all coronal slices (starting from the default slice=128, going all the way posterior and then all the way anterior) is excruciatingly slow. All of this would have to be repeated for all my 72 subjects - is there any way to optimise this manual inspection?
Also, if the surfs turn out to not follow the volume correctly, presumably the thing to do is white surface correction + re-running recon. But what should one do if, due to an erroneous averaging between timepoints, you see blurs/ghosts in your base template?
Many thanks! Tudor
On 9 May 2014 21:33, Martin Reuter mreuter@nmr.mgh.harvard.edu wrote:
Hi Tudor,
the longitudinal pipeline in FS is actually one of the best on the planet as far as I know :-). If there is any contradictory information on the wiki, can you point me to that so I can see what causes the misconception. Really: compared to independent processing, it significantly increases sensitivity. Also we have designed it to be unbiased with respect to a single time point or directionality. It is quite mature by now.
You should definitely use the longitudinal pipeline for the analysis of your data. Now to your questions
- QDEC: I am not too familiar with qdec. You can definitely try the
2-stage approach described on the wiki. There you first compute a measure of change (e.g. hippocampal volume change during your week) and then compare that measure across groups similar to a cross sectional volume/thickness analysis. You can also use our tools to run a linear mixed effects model if you want to do that (it is more involved and requires you to use matlab tools). In your case, you probably have 2 time points for all subjects and the time distance is probably the same for all subjects, so the 2-stage approach should be fine.
- The image processing is done via the longitudinal pipeline (three
steps: cross, base, long), to prepare the data look at the description of the 2-stage model http://freesurfer.net/fswiki/LongitudinalTwoStageModel and also the longitudinal tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial
- At the recon all level in FS you get (after the 3 steps) measurement
for all time points. So you would compare those results across time in the stats.
Hope that helps, Martin
On 05/08/2014 08:14 AM, Tudor Popescu wrote:
Sorry for the repeat, wasn't sure whether this was received the first time. Tudor
On 6 May 2014 19:55, Tudor Popescu tudor3@gmail.com wrote:
Dear FS list,
I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from an analysis point of view, this still qualifies as longitudinal. I want to check for
- the main within-subjects effect of time point (pre&post)
- the main between-subjects effect of group (treatment A, treatment
B, control),
- the time x group interaction
I intend to look at thickness, surface area, volume, and lGI.
I read on the wiki http://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing that FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis to supplement a VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are:
- Can QDEC be used for such an analysis, and if so, what would be
different as compared to a cross-sectional (no temporal/within factor) study?
Also, is the pre-processing stage any different?
In FSL, for longitudinal designs you do stats on images obtained as
the difference between consecutive time points. Does this have to be done in FreeSurfer as well, and if so, is it done at the recon-all level or only at the stats (QDEC) level?
Thanks!
Tudor
Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 Tudor,
I don't think there is a way to speed things up. Let me know if you find a case where the template is blurry or has ghosts. It should not happen, but if it does it indicates a bad registratration, you'd have to run the mri_robust_template command with different parameters manually then.
Best, Martin
On 05/27/2014 06:13 PM, Tudor Popescu wrote:
Hi Martin, Wasn't sure whether you'd seen my reply below, look forward to hear back your thoughts, thanks! Tudor
On 25 May 2014 21:40, "Tudor Popescu" <tudor3@gmail.com mailto:tudor3@gmail.com> wrote:
Thanks very much Martin and Bruce. I guess I'd misread the Wiki (my own fault, not the text's), and am glad to hear that the longitudinal pipeline is in fact perfectly suitable for my needs here. Having run the first 2 steps (cross and base), I'm a bit unclear how the output so far has to be manually inspected. It says in the tutorial <http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial> that you should load each subject's base volume + surfs in freeview and then "move back and forth a few slices". However, even just loading each base in this manner takes ~1 min on my PC (CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving with PgUp/PgDn between all coronal slices (starting from the default slice=128, going all the way posterior and then all the way anterior) is excruciatingly slow. All of this would have to be repeated for all my 72 subjects - is there any way to optimise this manual inspection? Also, if the surfs turn out to not follow the volume correctly, presumably the thing to do is white surface correction + re-running recon. But what should one do if, due to an erroneous averaging between timepoints, you see blurs/ghosts in your base template? Many thanks! Tudor On 9 May 2014 21:33, Martin Reuter <mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu>> wrote: Hi Tudor, the longitudinal pipeline in FS is actually one of the best on the planet as far as I know :-). If there is any contradictory information on the wiki, can you point me to that so I can see what causes the misconception. Really: compared to independent processing, it significantly increases sensitivity. Also we have designed it to be unbiased with respect to a single time point or directionality. It is quite mature by now. You should definitely use the longitudinal pipeline for the analysis of your data. Now to your questions 1. QDEC: I am not too familiar with qdec. You can definitely try the 2-stage approach described on the wiki. There you first compute a measure of change (e.g. hippocampal volume change during your week) and then compare that measure across groups similar to a cross sectional volume/thickness analysis. You can also use our tools to run a linear mixed effects model if you want to do that (it is more involved and requires you to use matlab tools). In your case, you probably have 2 time points for all subjects and the time distance is probably the same for all subjects, so the 2-stage approach should be fine. 2. The image processing is done via the longitudinal pipeline (three steps: cross, base, long), to prepare the data look at the description of the 2-stage model http://freesurfer.net/fswiki/LongitudinalTwoStageModel and also the longitudinal tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial 3. At the recon all level in FS you get (after the 3 steps) measurement for all time points. So you would compare those results across time in the stats. Hope that helps, Martin On 05/08/2014 08:14 AM, Tudor Popescu wrote:Sorry for the repeat, wasn't sure whether this was received the first time. Tudor On 6 May 2014 19:55, Tudor Popescu <tudor3@gmail.com <mailto:tudor3@gmail.com>> wrote: Dear FS list, I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from an analysis point of view, this still qualifies as longitudinal. I want to check for * the main within-subjects effect of time point (pre&post) * the main between-subjects effect of group (treatment A, treatment B, control), * the time x group interaction I intend to look at thickness, surface area, volume, and lGI. I read on the wiki <http://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing> that FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis to supplement a VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are: 1) Can QDEC be used for such an analysis, and if so, what would be different as compared to a cross-sectional (no temporal/within factor) study? 2) Also, is the pre-processing stage any different? 3) In FSL, for longitudinal designs you do stats on images obtained as the difference between consecutive time points. Does this have to be done in FreeSurfer as well, and if so, is it done at the recon-all level or only at the stats (QDEC) level? Thanks! Tudor _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer-- Martin Reuter, Ph.D. Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129 Phone:+1-617-724-5652 <tel:%2B1-617-724-5652> Email: mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> reuter@mit.edu <mailto:reuter@mit.edu> Web :http://reuter.mit.edu _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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
Tudor,
in your virtual machine, make sure the 3D rendering option, or 'use hardware rendering' (or something like that) is enabled. this would be on the Windows VM config side of things. otherwise it will do software rendering which is painfully slow.
Nick
Hi Tudor,
I don't think there is a way to speed things up. Let me know if you find a case where the template is blurry or has ghosts. It should not happen, but if it does it indicates a bad registratration, you'd have to run the mri_robust_template command with different parameters manually then.
Best, Martin
On 05/27/2014 06:13 PM, Tudor Popescu wrote:
Hi Martin, Wasn't sure whether you'd seen my reply below, look forward to hear back your thoughts, thanks! Tudor
On 25 May 2014 21:40, "Tudor Popescu" <tudor3@gmail.com mailto:tudor3@gmail.com> wrote:
Thanks very much Martin and Bruce. I guess I'd misread the Wiki (my own fault, not the text's), and am glad to hear that the longitudinal pipeline is in fact perfectly suitable for my needs here. Having run the first 2 steps (cross and base), I'm a bit unclear how the output so far has to be manually inspected. It says in the tutorial <http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial> that you should load each subject's base volume + surfs in freeview and then "move back and forth a few slices". However, even just loading each base in this manner takes ~1 min on my PC (CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving with PgUp/PgDn between all coronal slices (starting from the default slice=128, going all the way posterior and then all the way anterior) is excruciatingly slow. All of this would have to be repeated for all my 72 subjects - is there any way to optimise this manual inspection? Also, if the surfs turn out to not follow the volume correctly, presumably the thing to do is white surface correction + re-running recon. But what should one do if, due to an erroneous averaging between timepoints, you see blurs/ghosts in your base template? Many thanks! Tudor On 9 May 2014 21:33, Martin Reuter <mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu>> wrote: Hi Tudor, the longitudinal pipeline in FS is actually one of the best on the planet as far as I know :-). If there is any contradictory information on the wiki, can you point me to that so I can see what causes the misconception. Really: compared to independent processing, it significantly increases sensitivity. Also we have designed it to be unbiased with respect to a single time point or directionality. It is quite mature by now. You should definitely use the longitudinal pipeline for the analysis of your data. Now to your questions 1. QDEC: I am not too familiar with qdec. You can definitely try the 2-stage approach described on the wiki. There you first compute a measure of change (e.g. hippocampal volume change during your week) and then compare that measure across groups similar to a cross sectional volume/thickness analysis. You can also use our tools to run a linear mixed effects model if you want to do that (it is more involved and requires you to use matlab tools). In your case, you probably have 2 time points for all subjects and the time distance is probably the same for all subjects, so the 2-stage approach should be fine. 2. The image processing is done via the longitudinal pipeline (three steps: cross, base, long), to prepare the data look at the description of the 2-stage model http://freesurfer.net/fswiki/LongitudinalTwoStageModel and also the longitudinal tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial 3. At the recon all level in FS you get (after the 3 steps) measurement for all time points. So you would compare those results across time in the stats. Hope that helps, Martin On 05/08/2014 08:14 AM, Tudor Popescu wrote:Sorry for the repeat, wasn't sure whether this was received the first time. Tudor On 6 May 2014 19:55, Tudor Popescu <tudor3@gmail.com <mailto:tudor3@gmail.com>> wrote: Dear FS list, I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from an analysis point of view, this still qualifies as longitudinal. I want to check for * the main within-subjects effect of time point(pre&post) * the main between-subjects effect of group (treatment A, treatment B, control), * the time x group interaction
I intend to look at thickness, surface area, volume, andlGI.
I read on the wiki <http://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing> that FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis to supplement a VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are: 1) Can QDEC be used for such an analysis, and if so, what would be different as compared to a cross-sectional (no temporal/within factor) study? 2) Also, is the pre-processing stage any different? 3) In FSL, for longitudinal designs you do stats on images obtained as the difference between consecutive time points. Does this have to be done in FreeSurfer as well, and if so, is it done at the recon-all level or only at the stats (QDEC) level? Thanks! Tudor _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Martin Reuter, Ph.D. Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129 Phone:+1-617-724-5652 <tel:%2B1-617-724-5652> Email: mreuter@nmr.mgh.harvard.edumailto:mreuter@nmr.mgh.harvard.edu reuter@mit.edu mailto:reuter@mit.edu Web :http://reuter.mit.edu
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Many thanks Martin and Nick! The rendering option was already checked, but the freeview inspection commands were still very very slow, so I used commands of the following form, which as far as I understand should be equivalent: *tkmedit subjID norm.mgz -surfs* For several subjects (maybe 20 out of the total of 72), I noticed that the surfaces don\t seem to quite follow the GM/WM demarcation line as expected - for instance, by having portions of white surface (yellow line, see attached screenshots) surrounded by pial surface (red line), or by having very large cortical thickness in some parts of the brain. Is it possible that these apparent artefacts make sense "in context" (by looking at the adjacent slides), or is it the case that I need to go back and do white surface correction, and then redo all 3 steps of the longitudinal process?
Thanks again for your help! Tudor
On 29 May 2014 17:39, Nick Schmansky nicks@nmr.mgh.harvard.edu wrote:
Tudor,
in your virtual machine, make sure the 3D rendering option, or 'use hardware rendering' (or something like that) is enabled. this would be on the Windows VM config side of things. otherwise it will do software rendering which is painfully slow.
Nick
Hi Tudor,
I don't think there is a way to speed things up. Let me know if you find a case where the template is blurry or has ghosts. It should not happen, but if it does it indicates a bad registratration, you'd have to run the mri_robust_template command with different parameters manually then.
Best, Martin
On 05/27/2014 06:13 PM, Tudor Popescu wrote:
Hi Martin, Wasn't sure whether you'd seen my reply below, look forward to hear back your thoughts, thanks! Tudor
On 25 May 2014 21:40, "Tudor Popescu" <tudor3@gmail.com mailto:tudor3@gmail.com> wrote:
Thanks very much Martin and Bruce. I guess I'd misread the Wiki (my own fault, not the text's), and am glad to hear that the longitudinal pipeline is in fact perfectly suitable for my needs here. Having run the first 2 steps (cross and base), I'm a bit unclear how the output so far has to be manually inspected. It says in the tutorial <http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial> that you should load each subject's base volume + surfs in freeview and then "move back and forth a few slices". However, even just loading each base in this manner takes ~1 min on my PC (CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving with PgUp/PgDn between all coronal slices (starting from the default slice=128, going all the way posterior and then all the way anterior) is excruciatingly slow. All of this would have to be repeated for all my 72 subjects - is there any way to optimise this manual inspection? Also, if the surfs turn out to not follow the volume correctly, presumably the thing to do is white surface correction + re-running recon. But what should one do if, due to an erroneous averaging between timepoints, you see blurs/ghosts in your base template? Many thanks! Tudor On 9 May 2014 21:33, Martin Reuter <mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu>> wrote: Hi Tudor, the longitudinal pipeline in FS is actually one of the best on the planet as far as I know :-). If there is any contradictory information on the wiki, can you point me to that so I can see what causes the misconception. Really: compared to independent processing, it significantly increases sensitivity. Also we have designed it to be unbiased with respect to a single time point or directionality. It is quite mature by now. You should definitely use the longitudinal pipeline for the analysis of your data. Now to your questions 1. QDEC: I am not too familiar with qdec. You can definitely try the 2-stage approach described on the wiki. There you first compute a measure of change (e.g. hippocampal volume change during your week) and then compare that measure across groups similar to a cross sectional volume/thickness analysis. You can also use our tools to run a linear mixed effects model if you want to do that (it is more involved and requires you to use matlab tools). In your case, you probably have 2 time points for all subjects and the time distance is probably the same for all subjects, so the 2-stage approach should be fine. 2. The image processing is done via the longitudinal pipeline (three steps: cross, base, long), to prepare the data look at the description of the 2-stage model http://freesurfer.net/fswiki/LongitudinalTwoStageModel and also the longitudinal tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial 3. At the recon all level in FS you get (after the 3 steps) measurement for all time points. So you would compare those results across time in the stats. Hope that helps, Martin On 05/08/2014 08:14 AM, Tudor Popescu wrote:Sorry for the repeat, wasn't sure whether this was received the first time. Tudor On 6 May 2014 19:55, Tudor Popescu <tudor3@gmail.com <mailto:tudor3@gmail.com>> wrote: Dear FS list, I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from an analysis point of view, this still qualifies as longitudinal. I want to check for * the main within-subjects effect of time point(pre&post) * the main between-subjects effect of group (treatment A, treatment B, control), * the time x group interaction
I intend to look at thickness, surface area, volume, andlGI.
I read on the wiki <http://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing%3E
that FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis to supplement a VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are: 1) Can QDEC be used for such an analysis, and if so, what would be different as compared to a cross-sectional (no temporal/within factor) study? 2) Also, is the pre-processing stage any different? 3) In FSL, for longitudinal designs you do stats on images obtained as the difference between consecutive time points. Does this have to be done in FreeSurfer as well, and if so, is it done at the recon-all level or only at the stats (QDEC) level? Thanks! Tudor _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Martin Reuter, Ph.D. Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129 Phone:+1-617-724-5652 <tel:%2B1-617-724-5652> Email: mreuter@nmr.mgh.harvard.edumailto:mreuter@nmr.mgh.harvard.edu reuter@mit.edu mailto:reuter@mit.edu Web :http://reuter.mit.edu
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 Tudor
are you assessing the thickness visually? If so, you are probably being misled by the orientation of the surface w.r.t. the viewing plane. Same fo the apparent bubbles. You need to look in a different orientation
cheers Bruce
On Sun, 1 Jun 2014, Tudor Popescu wrote:
Many thanks Martin and Nick! The rendering option was already checked, but the freeview inspection commands were still very very slow, so I used commands of the following form, which as far as I understand should be equivalent: tkmedit subjID norm.mgz -surfs For several subjects (maybe 20 out of the total of 72), I noticed that the surfaces don\t seem to quite follow the GM/WM demarcation line as expected - for instance, by having portions of white surface (yellow line, see attached screenshots) surrounded by pial surface (red line), or by having very large cortical thickness in some parts of the brain. Is it possible that these apparent artefacts make sense "in context" (by looking at the adjacent slides), or is it the case that I need to go back and do white surface correction, and then redo all 3 steps of the longitudinal process?
Thanks again for your help! Tudor
On 29 May 2014 17:39, Nick Schmansky nicks@nmr.mgh.harvard.edu wrote: Tudor,
in your virtual machine, make sure the 3D rendering option, or 'use hardware rendering' (or something like that) is enabled. this would be on the Windows VM config side of things. otherwise it will do software rendering which is painfully slow. Nick > Hi Tudor, > > I don't think there is a way to speed things up. > Let me know if you find a case where the template is blurry or has > ghosts. It should not happen, but if it does it indicates a bad > registratration, you'd have to run the mri_robust_template command with > different parameters manually then. > > Best, Martin > > On 05/27/2014 06:13 PM, Tudor Popescu wrote: >> >> Hi Martin, >> Wasn't sure whether you'd seen my reply below, look forward to hear >> back your thoughts, thanks! >> Tudor >> >> On 25 May 2014 21:40, "Tudor Popescu" <tudor3@gmail.commailto:tudor3@gmail.com> wrote:
Thanks very much Martin and Bruce. I guess I'd misread the Wiki (my own fault, not the text's), and am glad to hear that the longitudinal pipeline is in fact perfectly suitable for my
needs
here.
Having run the first 2 steps (cross and base), I'm a bit
unclear
how the output so far has to be manually inspected. It says in
the
tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial that you should load each subject's base volume + surfs in freeview and then "move back and forth a few slices". However, even just loading each base in this manner takes ~1 min on my
PC
(CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving with PgUp/PgDn between all coronal slices (starting from the default slice=128, going all the way posterior and then all the way anterior) is excruciatingly slow. All of this would have to
be
repeated for all my 72 subjects - is there any way to optimise this manual inspection?
Also, if the surfs turn out to not follow the volume correctly, presumably the thing to do is white surface correction + re-running recon. But what should one do if, due to an
erroneous
averaging between timepoints, you see blurs/ghosts in your base template?
Many thanks! Tudor
On 9 May 2014 21:33, Martin Reuter <mreuter@nmr.mgh.harvard.edu mailto:mreuter@nmr.mgh.harvard.edu> wrote:
Hi Tudor,
the longitudinal pipeline in FS is actually one of the best
on
the planet as far as I know :-). If there is any
contradictory
information on the wiki, can you point me to that so I can
see
what causes the misconception. Really: compared to
independent
processing, it significantly increases sensitivity. Also we have designed it to be unbiased with respect to a single
time
point or directionality. It is quite mature by now.
You should definitely use the longitudinal pipeline for the analysis of your data. Now to your questions
1. QDEC: I am not too familiar with qdec. You can
definitely
try the 2-stage approach described on the wiki. There you first compute a measure of change (e.g. hippocampal volume change during your week) and then compare that measure
across
groups similar to a cross sectional volume/thickness
analysis.
You can also use our tools to run a linear mixed effects
model
if you want to do that (it is more involved and requires
you
to use matlab tools). In your case, you probably have 2
time
points for all subjects and the time distance is probably
the
same for all subjects, so the 2-stage approach should be
fine.
2. The image processing is done via the longitudinal
pipeline
(three steps: cross, base, long), to prepare the data look
at
the description of the 2-stage model http://freesurfer.net/fswiki/LongitudinalTwoStageModel and also the longitudinal tutorial
http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial
3. At the recon all level in FS you get (after the 3 steps) measurement for all time points. So you would compare those results across time in the stats.
Hope that helps, Martin
On 05/08/2014 08:14 AM, Tudor Popescu wrote:
Sorry for the repeat, wasn't sure whether this was
received
the first time. Tudor
On 6 May 2014 19:55, Tudor Popescu <tudor3@gmail.com mailto:tudor3@gmail.com> wrote:
Dear FS list,
I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from an
analysis
point of view, this still qualifies as longitudinal. I want to check for
* the main within-subjects effect of time point (pre&post) * the main between-subjects effect of group
(treatment
A, treatment B, control), * the time x group interaction
I intend to look at thickness, surface area, volume,
and
lGI.
I read on the wiki
http://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing
that FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis to
supplement a
VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are:
1) Can QDEC be used for such an analysis, and if so,
what
would be different as compared to a cross-sectional
(no
temporal/within factor) study?
2) Also, is the pre-processing stage any different?
3) In FSL, for longitudinal designs you do stats on images obtained as the difference between consecutive time points. Does this have to be done in FreeSurfer
as
well, and if so, is it done at the recon-all level or only at the stats (QDEC) level?
Thanks!
Tudor
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu mailto:Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone:+1-617-724-5652 tel:%2B1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu mailto:mreuter@nmr.mgh.harvard.edu reuter@mit.edu mailto:reuter@mit.edu Web :http://reuter.mit.edu
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu mailto: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
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 Bruce
I suspected this might be the case, since assessing 3D surfaces from a succession of 2D slices can be tricky, which is why I was wondering whether there's a better way to do these visual inspections. I checked those apparent artefacts in the other 2 planes (axial, saggital), but did not find it helpful in deciding whether the surfs was as they should or not.
Something that I think is more likely to be a real problem (i.e., more than just a viewing plane disorientation), is that for many subjects, posterior slices only start to be followed by the surfaces quite late, e.g. from coronal slice 80 onward. This seems to be because those slices only appear in the brainmask.mgz, but not also in the wm.mgz (see screenshots attached).
Also, after having run the cross and base steps, and ended up with folders of the type subj1_scanA, subj1_scanB etc, for the "long" step I launched simultaneously (for all subjects&scans) commands of the type recon-all -long subj1_scanA subj1 -all recon-all -long subj1_scanB subj1 -all however these resulted in two types of error messages: * mghRead(/vols/Scratch/tpopescu/learning/FS/subj1_scanA.long.subj1/mri/aseg.mgz, -1): read error* (those particular aseg.mgz files had size 0kB), and * (standard_in) 2: Error: comparison in expression* I initially had disk space issues, but after making space I relaunched the same commands with the -no-isrunning flag, but with the same results. Is it the case that the long commands for the different scans of a subject have to be launched sequentially rather than in parallel, i.e. the command for scanB should be run after the command for scanA has finished? Otherwise, what could the problem be?
Thanks again for your help.
Tudor
On 1 June 2014 15:17, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Tudor
are you assessing the thickness visually? If so, you are probably being misled by the orientation of the surface w.r.t. the viewing plane. Same fo the apparent bubbles. You need to look in a different orientation
cheers Bruce
On Sun, 1 Jun 2014, Tudor Popescu wrote:
Many thanks Martin and Nick! The rendering option was already checked, but
the freeview inspection commands were still very very slow, so I used commands of the following form, which as far as I understand should be equivalent: tkmedit subjID norm.mgz -surfs For several subjects (maybe 20 out of the total of 72), I noticed that the surfaces don\t seem to quite follow the GM/WM demarcation line as expected - for instance, by having portions of white surface (yellow line, see attached screenshots) surrounded by pial surface (red line), or by having very large cortical thickness in some parts of the brain. Is it possible that these apparent artefacts make sense "in context" (by looking at the adjacent slides), or is it the case that I need to go back and do white surface correction, and then redo all 3 steps of the longitudinal process?
Thanks again for your help! Tudor
On 29 May 2014 17:39, Nick Schmansky nicks@nmr.mgh.harvard.edu wrote: Tudor,
in your virtual machine, make sure the 3D rendering option, or 'use hardware rendering' (or something like that) is enabled. this would be on the Windows VM config side of things. otherwise it will do software rendering which is painfully slow. Nick > Hi Tudor, > > I don't think there is a way to speed things up. > Let me know if you find a case where the template is blurry or has > ghosts. It should not happen, but if it does it indicates a bad > registratration, you'd have to run the mri_robust_template command with > different parameters manually then. > > Best, Martin > > On 05/27/2014 06:13 PM, Tudor Popescu wrote: >> >> Hi Martin, >> Wasn't sure whether you'd seen my reply below, look forward to hear >> back your thoughts, thanks! >> Tudor >> >> On 25 May 2014 21:40, "Tudor Popescu" <tudor3@gmail.commailto:tudor3@gmail.com> wrote:
Thanks very much Martin and Bruce. I guess I'd misread the Wiki (my own fault, not the text's), and am glad to hear that the longitudinal pipeline is in fact perfectly suitable for myneeds
here. Having run the first 2 steps (cross and base), I'm a bitunclear
how the output so far has to be manually inspected. It says inthe
tutorial <http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial> that you should load each subject's base volume + surfs in freeview and then "move back and forth a few slices". However, even just loading each base in this manner takes ~1 min on myPC
(CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving with PgUp/PgDn between all coronal slices (starting from the default slice=128, going all the way posterior and then all the way anterior) is excruciatingly slow. All of this would have tobe
repeated for all my 72 subjects - is there any way to optimise this manual inspection? Also, if the surfs turn out to not follow the volume correctly, presumably the thing to do is white surface correction + re-running recon. But what should one do if, due to anerroneous
averaging between timepoints, you see blurs/ghosts in your base template? Many thanks! Tudor On 9 May 2014 21:33, Martin Reuter <mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu>> wrote: Hi Tudor, the longitudinal pipeline in FS is actually one of the beston
the planet as far as I know :-). If there is anycontradictory
information on the wiki, can you point me to that so I cansee
what causes the misconception. Really: compared toindependent
processing, it significantly increases sensitivity. Also we have designed it to be unbiased with respect to a singletime
point or directionality. It is quite mature by now. You should definitely use the longitudinal pipeline for the analysis of your data. Now to your questions 1. QDEC: I am not too familiar with qdec. You candefinitely
try the 2-stage approach described on the wiki. There you first compute a measure of change (e.g. hippocampal volume change during your week) and then compare that measureacross
groups similar to a cross sectional volume/thicknessanalysis.
You can also use our tools to run a linear mixed effectsmodel
if you want to do that (it is more involved and requiresyou
to use matlab tools). In your case, you probably have 2time
points for all subjects and the time distance is probablythe
same for all subjects, so the 2-stage approach should befine.
2. The image processing is done via the longitudinalpipeline
(three steps: cross, base, long), to prepare the data lookat
the description of the 2-stage model http://freesurfer.net/fswiki/LongitudinalTwoStageModel and also the longitudinal tutorialhttp://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial
3. At the recon all level in FS you get (after the 3 steps) measurement for all time points. So you would compare those results across time in the stats. Hope that helps, Martin On 05/08/2014 08:14 AM, Tudor Popescu wrote:Sorry for the repeat, wasn't sure whether this wasreceived
the first time. Tudor On 6 May 2014 19:55, Tudor Popescu <tudor3@gmail.com <mailto:tudor3@gmail.com>> wrote: Dear FS list, I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from ananalysis
point of view, this still qualifies as longitudinal. I want to check for * the main within-subjects effect of time point(pre&post) * the main between-subjects effect of group
(treatment
A, treatment B, control), * the time x group interaction I intend to look at thickness, surface area, volume,and
lGI.
I read on the wikihttp://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing
that FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis tosupplement a
VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are: 1) Can QDEC be used for such an analysis, and if so,what
would be different as compared to a cross-sectional(no
temporal/within factor) study? 2) Also, is the pre-processing stage any different? 3) In FSL, for longitudinal designs you do stats on images obtained as the difference between consecutive time points. Does this have to be done in FreeSurferas
well, and if so, is it done at the recon-all level or only at the stats (QDEC) level? Thanks! Tudor _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Martin Reuter, Ph.D. Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129 Phone:+1-617-724-5652 <tel:%2B1-617-724-5652> Email: mreuter@nmr.mgh.harvard.edumailto:mreuter@nmr.mgh.harvard.edu reuter@mit.edu mailto:reuter@mit.edu Web :http://reuter.mit.edu
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu>https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for theperson
to whom it is addressed. If you believe this e-mail was sent to you inerror
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 contactthe
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
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 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 Tudor,
you can run the -long processes in parallel. Something else is wrong. Check if you have enough disk space. Re-run the long from scratch (to make sure everything is generated new).
Best, Martin
On Jun 2, 2014, at 5:34 PM, Tudor Popescu tudor3@gmail.com wrote:
Hi Bruce
I suspected this might be the case, since assessing 3D surfaces from a succession of 2D slices can be tricky, which is why I was wondering whether there's a better way to do these visual inspections. I checked those apparent artefacts in the other 2 planes (axial, saggital), but did not find it helpful in deciding whether the surfs was as they should or not.
Something that I think is more likely to be a real problem (i.e., more than just a viewing plane disorientation), is that for many subjects, posterior slices only start to be followed by the surfaces quite late, e.g. from coronal slice 80 onward. This seems to be because those slices only appear in the brainmask.mgz, but not also in the wm.mgz (see screenshots attached).
Also, after having run the cross and base steps, and ended up with folders of the type subj1_scanA, subj1_scanB etc, for the "long" step I launched simultaneously (for all subjects&scans) commands of the type recon-all -long subj1_scanA subj1 -all recon-all -long subj1_scanB subj1 -all however these resulted in two types of error messages: mghRead(/vols/Scratch/tpopescu/learning/FS/subj1_scanA.long.subj1/mri/aseg.mgz, -1): read error (those particular aseg.mgz files had size 0kB), and (standard_in) 2: Error: comparison in expression I initially had disk space issues, but after making space I relaunched the same commands with the -no-isrunning flag, but with the same results. Is it the case that the long commands for the different scans of a subject have to be launched sequentially rather than in parallel, i.e. the command for scanB should be run after the command for scanA has finished? Otherwise, what could the problem be?
Thanks again for your help.
Tudor
On 1 June 2014 15:17, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote: Hi Tudor
are you assessing the thickness visually? If so, you are probably being misled by the orientation of the surface w.r.t. the viewing plane. Same fo the apparent bubbles. You need to look in a different orientation
cheers Bruce
On Sun, 1 Jun 2014, Tudor Popescu wrote:
Many thanks Martin and Nick! The rendering option was already checked, but the freeview inspection commands were still very very slow, so I used commands of the following form, which as far as I understand should be equivalent: tkmedit subjID norm.mgz -surfs For several subjects (maybe 20 out of the total of 72), I noticed that the surfaces don\t seem to quite follow the GM/WM demarcation line as expected - for instance, by having portions of white surface (yellow line, see attached screenshots) surrounded by pial surface (red line), or by having very large cortical thickness in some parts of the brain. Is it possible that these apparent artefacts make sense "in context" (by looking at the adjacent slides), or is it the case that I need to go back and do white surface correction, and then redo all 3 steps of the longitudinal process?
Thanks again for your help! Tudor
On 29 May 2014 17:39, Nick Schmansky nicks@nmr.mgh.harvard.edu wrote: Tudor,
in your virtual machine, make sure the 3D rendering option, or 'use hardware rendering' (or something like that) is enabled. this would be on the Windows VM config side of things. otherwise it will do software rendering which is painfully slow. Nick > Hi Tudor, > > I don't think there is a way to speed things up. > Let me know if you find a case where the template is blurry or has > ghosts. It should not happen, but if it does it indicates a bad > registratration, you'd have to run the mri_robust_template command with > different parameters manually then. > > Best, Martin > > On 05/27/2014 06:13 PM, Tudor Popescu wrote: >> >> Hi Martin, >> Wasn't sure whether you'd seen my reply below, look forward to hear >> back your thoughts, thanks! >> Tudor >> >> On 25 May 2014 21:40, "Tudor Popescu" <tudor3@gmail.commailto:tudor3@gmail.com> wrote:
Thanks very much Martin and Bruce. I guess I'd misread the Wiki (my own fault, not the text's), and am glad to hear that the longitudinal pipeline is in fact perfectly suitable for myneeds
here. Having run the first 2 steps (cross and base), I'm a bitunclear
how the output so far has to be manually inspected. It says inthe
tutorial <http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial> that you should load each subject's base volume + surfs in freeview and then "move back and forth a few slices". However, even just loading each base in this manner takes ~1 min on myPC
(CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving with PgUp/PgDn between all coronal slices (starting from the default slice=128, going all the way posterior and then all the way anterior) is excruciatingly slow. All of this would have tobe
repeated for all my 72 subjects - is there any way to optimise this manual inspection? Also, if the surfs turn out to not follow the volume correctly, presumably the thing to do is white surface correction + re-running recon. But what should one do if, due to anerroneous
averaging between timepoints, you see blurs/ghosts in your base template? Many thanks! Tudor On 9 May 2014 21:33, Martin Reuter <mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu>> wrote: Hi Tudor, the longitudinal pipeline in FS is actually one of the beston
the planet as far as I know :-). If there is anycontradictory
information on the wiki, can you point me to that so I cansee
what causes the misconception. Really: compared toindependent
processing, it significantly increases sensitivity. Also we have designed it to be unbiased with respect to a singletime
point or directionality. It is quite mature by now. You should definitely use the longitudinal pipeline for the analysis of your data. Now to your questions 1. QDEC: I am not too familiar with qdec. You candefinitely
try the 2-stage approach described on the wiki. There you first compute a measure of change (e.g. hippocampal volume change during your week) and then compare that measureacross
groups similar to a cross sectional volume/thicknessanalysis.
You can also use our tools to run a linear mixed effectsmodel
if you want to do that (it is more involved and requiresyou
to use matlab tools). In your case, you probably have 2time
points for all subjects and the time distance is probablythe
same for all subjects, so the 2-stage approach should befine.
2. The image processing is done via the longitudinalpipeline
(three steps: cross, base, long), to prepare the data lookat
the description of the 2-stage model http://freesurfer.net/fswiki/LongitudinalTwoStageModel and also the longitudinal tutorialhttp://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial
3. At the recon all level in FS you get (after the 3 steps) measurement for all time points. So you would compare those results across time in the stats. Hope that helps, Martin On 05/08/2014 08:14 AM, Tudor Popescu wrote:Sorry for the repeat, wasn't sure whether this wasreceived
the first time. Tudor On 6 May 2014 19:55, Tudor Popescu <tudor3@gmail.com <mailto:tudor3@gmail.com>> wrote: Dear FS list, I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from ananalysis
point of view, this still qualifies as longitudinal. I want to check for * the main within-subjects effect of time point(pre&post) * the main between-subjects effect of group
(treatment
A, treatment B, control), * the time x group interaction I intend to look at thickness, surface area, volume,and
lGI.
I read on the wikihttp://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing
that FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis tosupplement a
VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are: 1) Can QDEC be used for such an analysis, and if so,what
would be different as compared to a cross-sectional(no
temporal/within factor) study? 2) Also, is the pre-processing stage any different? 3) In FSL, for longitudinal designs you do stats on images obtained as the difference between consecutive time points. Does this have to be done in FreeSurferas
well, and if so, is it done at the recon-all level or only at the stats (QDEC) level? Thanks! Tudor _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Martin Reuter, Ph.D. Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129 Phone:+1-617-724-5652 <tel:%2B1-617-724-5652> Email: mreuter@nmr.mgh.harvard.edumailto:mreuter@nmr.mgh.harvard.edu reuter@mit.edu mailto:reuter@mit.edu Web :http://reuter.mit.edu
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu>https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for theperson
to whom it is addressed. If you believe this e-mail was sent to you inerror
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 contactthe
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
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 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.
<02-06-2014, 22.17.45.png><02-06-2014, 22.17.55.png><only LH posterior.png><norm.png>_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
--------------------------------- Dr. Martin Reuter Assistant in Neuroscience - Massachusetts General Hospital Instructor in Neurology - Harvard Medical School MGH / HMS / MIT
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
Hi Martin,
I reran everything from scratch and this time the recons ended without error. Few more questions:
1) Is the brainmask.mgz vs wm.mgz problem that I described in my previous email (surfaces only starting to show rather late at the posterior end of some subjects) something that I need to correct (e.g., by editing wm.mgz), or is it fine as it is? 2) Given that I visually inspected all bases, Is it necessary to inspect *all *slices for the longitudinal overlaps (vols+surfs) of all subjects, or is it enough to just check the central coronal slice for each subject? I ask because it's not just *opening *freeview with all those layers that is very slow, but also navigating between slices 3) The tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorialcovers the creation of long.qdec.table.dat for a single-group scenario; I have 3 groups, with scans on the first and final (fifth) day of training, and was wondering whether the following table looks correct
*fsid* *fsid-base* *group* *day* *ageDemeaned * *gender* *score* subj1_day1 subj1 treatmentA 1 1.34 f 109 subj1_day5 subj1 treatmentA 5 1.34 f 109 subj2_day1 subj2 treatmentB 1 -5.85 m 102 subj2_day5 subj2 treatmentB 5 -5.85 m 102 subj3_day1 subj2 control 1 3.06 f 98 subj3_day5 subj2 control 5 3.06 f 98 4) Is this mixed design (3 groups and 2 time points, with both effect of group and of time being relevant) suitably analysed with only a simple 2 stage model, or is a linear mixed effects model necessary? Also, how are the stats different given there are 3 groups and not the usual 2 than most FS tutorials are based on? 5) Is the default long_mris_slopes command (which I understand does preliminary analyses, before going to QDEC) mentioned in the tutorial, i.e.
long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas thickness --hemi lh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time years --qcache fsaverage --sd $SUBJECTS_DIR
suitable for my particular case? I will also look at surface area and lGI, so I suspect it's at least "--meas area" and "--meas lGI" that I need to add to the above command?
Thanks so much! Tudor
On 3 June 2014 02:02, Martin Reuter mreuter@nmr.mgh.harvard.edu wrote:
HI Tudor,
you can run the -long processes in parallel. Something else is wrong. Check if you have enough disk space. Re-run the long from scratch (to make sure everything is generated new).
Best, Martin
On Jun 2, 2014, at 5:34 PM, Tudor Popescu tudor3@gmail.com wrote:
Hi Bruce
I suspected this might be the case, since assessing 3D surfaces from a succession of 2D slices can be tricky, which is why I was wondering whether there's a better way to do these visual inspections. I checked those apparent artefacts in the other 2 planes (axial, saggital), but did not find it helpful in deciding whether the surfs was as they should or not.
Something that I think is more likely to be a real problem (i.e., more than just a viewing plane disorientation), is that for many subjects, posterior slices only start to be followed by the surfaces quite late, e.g. from coronal slice 80 onward. This seems to be because those slices only appear in the brainmask.mgz, but not also in the wm.mgz (see screenshots attached).
Also, after having run the cross and base steps, and ended up with folders of the type subj1_scanA, subj1_scanB etc, for the "long" step I launched simultaneously (for all subjects&scans) commands of the type recon-all -long subj1_scanA subj1 -all recon-all -long subj1_scanB subj1 -all however these resulted in two types of error messages:
mghRead(/vols/Scratch/tpopescu/learning/FS/subj1_scanA.long.subj1/mri/aseg.mgz, -1): read error* (those particular aseg.mgz files had size 0kB), and
- (standard_in) 2: Error: comparison in expression*
I initially had disk space issues, but after making space I relaunched the same commands with the -no-isrunning flag, but with the same results. Is it the case that the long commands for the different scans of a subject have to be launched sequentially rather than in parallel, i.e. the command for scanB should be run after the command for scanA has finished? Otherwise, what could the problem be?
Thanks again for your help.
Tudor
On 1 June 2014 15:17, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Tudor
are you assessing the thickness visually? If so, you are probably being misled by the orientation of the surface w.r.t. the viewing plane. Same fo the apparent bubbles. You need to look in a different orientation
cheers Bruce
On Sun, 1 Jun 2014, Tudor Popescu wrote:
Many thanks Martin and Nick! The rendering option was already checked,
but the freeview inspection commands were still very very slow, so I used commands of the following form, which as far as I understand should be equivalent: tkmedit subjID norm.mgz -surfs For several subjects (maybe 20 out of the total of 72), I noticed that the surfaces don\t seem to quite follow the GM/WM demarcation line as expected - for instance, by having portions of white surface (yellow line, see attached screenshots) surrounded by pial surface (red line), or by having very large cortical thickness in some parts of the brain. Is it possible that these apparent artefacts make sense "in context" (by looking at the adjacent slides), or is it the case that I need to go back and do white surface correction, and then redo all 3 steps of the longitudinal process?
Thanks again for your help! Tudor
On 29 May 2014 17:39, Nick Schmansky nicks@nmr.mgh.harvard.edu wrote: Tudor,
in your virtual machine, make sure the 3D rendering option, or 'use hardware rendering' (or something like that) is enabled. this would be on the Windows VM config side of things. otherwise it will do software rendering which is painfully slow. Nick > Hi Tudor, > > I don't think there is a way to speed things up. > Let me know if you find a case where the template is blurry or has > ghosts. It should not happen, but if it does it indicates a bad > registratration, you'd have to run the mri_robust_template command with > different parameters manually then. > > Best, Martin > > On 05/27/2014 06:13 PM, Tudor Popescu wrote: >> >> Hi Martin, >> Wasn't sure whether you'd seen my reply below, look forward to hear >> back your thoughts, thanks! >> Tudor >> >> On 25 May 2014 21:40, "Tudor Popescu" <tudor3@gmail.commailto:tudor3@gmail.com> wrote:
Thanks very much Martin and Bruce. I guess I'd misread the Wiki (my own fault, not the text's), and am glad to hear that the longitudinal pipeline is in fact perfectly suitable for myneeds
here. Having run the first 2 steps (cross and base), I'm a bitunclear
how the output so far has to be manually inspected. It says inthe
tutorial <http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial> that you should load each subject's base volume + surfs in freeview and then "move back and forth a few slices". However, even just loading each base in this manner takes ~1 min on myPC
(CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving with PgUp/PgDn between all coronal slices (starting from the default slice=128, going all the way posterior and then all the way anterior) is excruciatingly slow. All of this would have tobe
repeated for all my 72 subjects - is there any way to optimise this manual inspection? Also, if the surfs turn out to not follow the volume correctly, presumably the thing to do is white surface correction + re-running recon. But what should one do if, due to anerroneous
averaging between timepoints, you see blurs/ghosts in your base template? Many thanks! Tudor On 9 May 2014 21:33, Martin Reuter <mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu>> wrote: Hi Tudor, the longitudinal pipeline in FS is actually one of the beston
the planet as far as I know :-). If there is anycontradictory
information on the wiki, can you point me to that so I cansee
what causes the misconception. Really: compared toindependent
processing, it significantly increases sensitivity. Also we have designed it to be unbiased with respect to a singletime
point or directionality. It is quite mature by now. You should definitely use the longitudinal pipeline for the analysis of your data. Now to your questions 1. QDEC: I am not too familiar with qdec. You candefinitely
try the 2-stage approach described on the wiki. There you first compute a measure of change (e.g. hippocampal volume change during your week) and then compare that measureacross
groups similar to a cross sectional volume/thicknessanalysis.
You can also use our tools to run a linear mixed effectsmodel
if you want to do that (it is more involved and requiresyou
to use matlab tools). In your case, you probably have 2time
points for all subjects and the time distance is probablythe
same for all subjects, so the 2-stage approach should befine.
2. The image processing is done via the longitudinalpipeline
(three steps: cross, base, long), to prepare the data lookat
the description of the 2-stage model http://freesurfer.net/fswiki/LongitudinalTwoStageModel and also the longitudinal tutorialhttp://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial
3. At the recon all level in FS you get (after the 3 steps) measurement for all time points. So you would compare those results across time in the stats. Hope that helps, Martin On 05/08/2014 08:14 AM, Tudor Popescu wrote:Sorry for the repeat, wasn't sure whether this wasreceived
the first time. Tudor On 6 May 2014 19:55, Tudor Popescu <tudor3@gmail.com <mailto:tudor3@gmail.com>> wrote: Dear FS list, I have structural data from a learning study (pre&post-training scans, with 3 groups). Although the training was only one week, I'm guessing from ananalysis
point of view, this still qualifies as longitudinal. I want to check for * the main within-subjects effect of time point(pre&post) * the main between-subjects effect of group
(treatment
A, treatment B, control), * the time x group interaction I intend to look at thickness, surface area, volume,and
lGI.
I read on the wikihttp://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing
that FS is currently not optimal for longitudinal analyses. I intend my FreeSurfer analysis tosupplement a
VBM analysis done in FSL. In case it is in fact a good idea to do this, my questions (not covered in the 'longitudinal' wiki page) are: 1) Can QDEC be used for such an analysis, and if so,what
would be different as compared to a cross-sectional(no
temporal/within factor) study? 2) Also, is the pre-processing stage any different? 3) In FSL, for longitudinal designs you do stats on images obtained as the difference between consecutive time points. Does this have to be done in FreeSurferas
well, and if so, is it done at the recon-all level or only at the stats (QDEC) level? Thanks! Tudor _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Martin Reuter, Ph.D. Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129 Phone:+1-617-724-5652 <tel:%2B1-617-724-5652> Email: mreuter@nmr.mgh.harvard.edumailto:mreuter@nmr.mgh.harvard.edu reuter@mit.edu mailto:reuter@mit.edu Web :http://reuter.mit.edu
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu>https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for theperson
to whom it is addressed. If you believe this e-mail was sent to you inerror
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 contactthe
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
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 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.
<02-06-2014, 22.17.45.png><02-06-2014, 22.17.55.png><only LH posterior.png><norm.png>_______________________________________________
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Dr. Martin Reuter Assistant in Neuroscience - Massachusetts General Hospital Instructor in Neurology - Harvard Medical School MGH / HMS / MIT
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 Tudor,
1. not sure, this is a more general question for people more familiar with surfaces and edits
2. nothing is necessary, however, if you have regions where you find difficulties in some subjects, I'd look at that region in all subjects (e.g. you mention the posterior end). Usually if surfaces on the base are good, the long should be high quality also.
3. looks good (I'd do space separated, also some columns merged in your email and fsid-base for subject3 is wrong).
4. using the two-stage model you could do a pairwise analysis between your groups in the second stage, so linear mixed effects model is not necessary. In the two stage model, just think of it as a cross sectional analysis of e.g. thickness change, instead of thickness.
5. yes, is sutiable. You'd have to run it several times, once for each of the measures you want to analyze. You can drop some of:
--do-avg --do-rate --do-pc1 --do-spc --do-stack
and only keep e.g. the rate or the pc1 (depending on what you want to analyze). You can also use --stack-rate to generate a stack of the rate maps.
Best, Martin
On 06/04/2014 01:19 PM, Tudor Popescu wrote:
Hi Martin,
I reran everything from scratch and this time the recons ended without error. Few more questions:
- Is the brainmask.mgz vs wm.mgz problem that I described in my
previous email (surfaces only starting to show rather late at the posterior end of some subjects) something that I need to correct (e.g., by editing wm.mgz), or is it fine as it is? 2) Given that I visually inspected all bases, Is it necessary to inspect /all /slices for the longitudinal overlaps (vols+surfs) of all subjects, or is it enough to just check the central coronal slice for each subject? I ask because it's not just /opening /freeview with all those layers that is very slow, but also navigating between slices 3) The tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorialcovers the creation of long.qdec.table.dat for a single-group scenario; I have 3 groups, with scans on the first and final (fifth) day of training, and was wondering whether the following table looks correct
*fsid* *fsid-base* *group* *day* *ageDemeaned
- *gender* *score*
subj1_day1 subj1 treatmentA 1 1.34 f 109 subj1_day5 subj1 treatmentA 5 1.34 f 109 subj2_day1 subj2 treatmentB 1 -5.85 m 102 subj2_day5 subj2 treatmentB 5 -5.85 m 102 subj3_day1 subj2 control 1 3.06 f 98 subj3_day5 subj2 control 5 3.06 f 98
- Is this mixed design (3 groups and 2 time points, with both effect
of group and of time being relevant) suitably analysed with only a simple 2 stage model, or is a linear mixed effects model necessary? Also, how are the stats different given there are 3 groups and not the usual 2 than most FS tutorials are based on? 5) Is the default long_mris_slopes command (which I understand does preliminary analyses, before going to QDEC) mentioned in the tutorial, i.e. long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas thickness --hemi lh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time years --qcache fsaverage --sd $SUBJECTS_DIR suitable for my particular case? I will also look at surface area and lGI, so I suspect it's at least "--meas area" and "--meas lGI" that I need to add to the above command?
Thanks so much! Tudor
On 3 June 2014 02:02, Martin Reuter <mreuter@nmr.mgh.harvard.edu mailto:mreuter@nmr.mgh.harvard.edu> wrote:
HI Tudor, you can run the -long processes in parallel. Something else is wrong. Check if you have enough disk space. Re-run the long from scratch (to make sure everything is generated new). Best, Martin On Jun 2, 2014, at 5:34 PM, Tudor Popescu <tudor3@gmail.com <mailto:tudor3@gmail.com>> wrote:Hi Bruce I suspected this might be the case, since assessing 3D surfaces from a succession of 2D slices can be tricky, which is why I was wondering whether there's a better way to do these visual inspections. I checked those apparent artefacts in the other 2 planes (axial, saggital), but did not find it helpful in deciding whether the surfs was as they should or not. Something that I think is more likely to be a real problem (i.e., more than just a viewing plane disorientation), is that for many subjects, posterior slices only start to be followed by the surfaces quite late, e.g. from coronal slice 80 onward. This seems to be because those slices only appear in the brainmask.mgz, but not also in the wm.mgz (see screenshots attached). Also, after having run the cross and base steps, and ended up with folders of the type subj1_scanA, subj1_scanB etc, for the "long" step I launched simultaneously (for all subjects&scans) commands of the type recon-all -long subj1_scanA subj1 -all recon-all -long subj1_scanB subj1 -all however these resulted in two types of error messages: /mghRead(/vols/Scratch/tpopescu/learning/FS/subj1_scanA.long.subj1/mri/aseg.mgz, -1): read error/ (those particular aseg.mgz files had size 0kB), and / (standard_in) 2: Error: comparison in expression/ I initially had disk space issues, but after making space I relaunched the same commands with the -no-isrunning flag, but with the same results. Is it the case that the long commands for the different scans of a subject have to be launched sequentially rather than in parallel, i.e. the command for scanB should be run after the command for scanA has finished? Otherwise, what could the problem be? Thanks again for your help. Tudor On 1 June 2014 15:17, Bruce Fischl <fischl@nmr.mgh.harvard.edu <mailto:fischl@nmr.mgh.harvard.edu>> wrote: Hi Tudor are you assessing the thickness visually? If so, you are probably being misled by the orientation of the surface w.r.t. the viewing plane. Same fo the apparent bubbles. You need to look in a different orientation cheers Bruce On Sun, 1 Jun 2014, Tudor Popescu wrote: Many thanks Martin and Nick! The rendering option was already checked, but the freeview inspection commands were still very very slow, so I used commands of the following form, which as far as I understand should be equivalent: tkmedit subjID norm.mgz -surfs For several subjects (maybe 20 out of the total of 72), I noticed that the surfaces don\t seem to quite follow the GM/WM demarcation line as expected - for instance, by having portions of white surface (yellow line, see attached screenshots) surrounded by pial surface (red line), or by having very large cortical thickness in some parts of the brain. Is it possible that these apparent artefacts make sense "in context" (by looking at the adjacent slides), or is it the case that I need to go back and do white surface correction, and then redo all 3 steps of the longitudinal process? Thanks again for your help! Tudor On 29 May 2014 17:39, Nick Schmansky <nicks@nmr.mgh.harvard.edu <mailto:nicks@nmr.mgh.harvard.edu>> wrote: Tudor, in your virtual machine, make sure the 3D rendering option, or 'use hardware rendering' (or something like that) is enabled. this would be on the Windows VM config side of things. otherwise it will do software rendering which is painfully slow. Nick > Hi Tudor, > > I don't think there is a way to speed things up. > Let me know if you find a case where the template is blurry or has > ghosts. It should not happen, but if it does it indicates a bad > registratration, you'd have to run the mri_robust_template command with > different parameters manually then. > > Best, Martin > > On 05/27/2014 06:13 PM, Tudor Popescu wrote: >> >> Hi Martin, >> Wasn't sure whether you'd seen my reply below, look forward to hear >> back your thoughts, thanks! >> Tudor >> >> On 25 May 2014 21:40, "Tudor Popescu" <tudor3@gmail.com <mailto:tudor3@gmail.com> >> <mailto:tudor3@gmail.com <mailto:tudor3@gmail.com>>> wrote: >> >> Thanks very much Martin and Bruce. I guess I'd misread the Wiki >> (my own fault, not the text's), and am glad to hear that the >> longitudinal pipeline is in fact perfectly suitable for my needs >> here. >> >> Having run the first 2 steps (cross and base), I'm a bit unclear >> how the output so far has to be manually inspected. It says in the >> tutorial >> <http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial> >> that you should load each subject's base volume + surfs in >> freeview and then "move back and forth a few slices". However, >> even just loading each base in this manner takes ~1 min on my PC >> (CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving >> with PgUp/PgDn between all coronal slices (starting from the >> default slice=128, going all the way posterior and then all the >> way anterior) is excruciatingly slow. All of this would have to be >> repeated for all my 72 subjects - is there any way to optimise >> this manual inspection? >> >> Also, if the surfs turn out to not follow the volume correctly, >> presumably the thing to do is white surface correction + >> re-running recon. But what should one do if, due to an erroneous >> averaging between timepoints, you see blurs/ghosts in your base >> template? >> >> Many thanks! >> Tudor >> >> >> On 9 May 2014 21:33, Martin Reuter <mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> >> <mailto:mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu>>> wrote: >> >> Hi Tudor, >> >> the longitudinal pipeline in FS is actually one of the best on >> the planet as far as I know :-). If there is any contradictory >> information on the wiki, can you point me to that so I can see >> what causes the misconception. Really: compared to independent >> processing, it significantly increases sensitivity. Also we >> have designed it to be unbiased with respect to a single time >> point or directionality. It is quite mature by now. >> >> You should definitely use the longitudinal pipeline for the >> analysis of your data. Now to your questions >> >> 1. QDEC: I am not too familiar with qdec. You can definitely >> try the 2-stage approach described on the wiki. There you >> first compute a measure of change (e.g. hippocampal volume >> change during your week) and then compare that measure across >> groups similar to a cross sectional volume/thickness analysis. >> You can also use our tools to run a linear mixed effects model >> if you want to do that (it is more involved and requires you >> to use matlab tools). In your case, you probably have 2 time >> points for all subjects and the time distance is probably the >> same for all subjects, so the 2-stage approach should be fine. >> >> 2. The image processing is done via the longitudinal pipeline >> (three steps: cross, base, long), to prepare the data look at >> the description of the 2-stage model >> http://freesurfer.net/fswiki/LongitudinalTwoStageModel >> and also the longitudinal tutorial >> http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial >> >> 3. At the recon all level in FS you get (after the 3 steps) >> measurement for all time points. So you would compare those >> results across time in the stats. >> >> Hope that helps, Martin >> >> >> On 05/08/2014 08:14 AM, Tudor Popescu wrote: >>> Sorry for the repeat, wasn't sure whether this was received >>> the first time. >>> Tudor >>> >>> >>> On 6 May 2014 19:55, Tudor Popescu <tudor3@gmail.com <mailto:tudor3@gmail.com> >>> <mailto:tudor3@gmail.com <mailto:tudor3@gmail.com>>> wrote: >>> >>> Dear FS list, >>> >>> I have structural data from a learning study >>> (pre&post-training scans, with 3 groups). Although the >>> training was only one week, I'm guessing from an analysis >>> point of view, this still qualifies as longitudinal. I >>> want to check for >>> >>> * the main within-subjects effect of time point >>> (pre&post) >>> * the main between-subjects effect of group (treatment >>> A, treatment B, control), >>> * the time x group interaction >>> >>> I intend to look at thickness, surface area, volume, and >>> lGI. >>> >>> I read on the wiki >>> <http://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing> >>> that FS is currently not optimal for longitudinal >>> analyses. I intend my FreeSurfer analysis to supplement a >>> VBM analysis done in FSL. In case it is in fact a good >>> idea to do this, my questions (not covered in the >>> 'longitudinal' wiki page) are: >>> >>> 1) Can QDEC be used for such an analysis, and if so, what >>> would be different as compared to a cross-sectional (no >>> temporal/within factor) study? >>> >>> 2) Also, is the pre-processing stage any different? >>> >>> 3) In FSL, for longitudinal designs you do stats on >>> images obtained as the difference between consecutive >>> time points. Does this have to be done in FreeSurfer as >>> well, and if so, is it done at the recon-all level or >>> only at the stats (QDEC) level? >>> >>> Thanks! >>> >>> Tudor >>> >>> >>> >>> >>> >>> _______________________________________________ >>> Freesurfer mailing list >>> Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> >>> <mailto:Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu>> >>> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer >> >> -- >> Martin Reuter, Ph.D. >> >> Instructor in Neurology >> Harvard Medical School >> Assistant in Neuroscience >> Dept. of Radiology, Massachusetts General Hospital >> Dept. of Neurology, Massachusetts General Hospital >> Research Affiliate >> Computer Science and Artificial Intelligence Lab, >> Dept. of Electrical Engineering and Computer Science, >> Massachusetts Institute of Technology >> >> A.A.Martinos Center for Biomedical Imaging >> 149 Thirteenth Street, Suite 2301 >> Charlestown, MA 02129 >> >> Phone:+1-617-724-5652 <tel:%2B1-617-724-5652> <tel:%2B1-617-724-5652> >> Email: >> mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> >> <mailto:mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu>> >> reuter@mit.edu <mailto:reuter@mit.edu> <mailto:reuter@mit.edu <mailto:reuter@mit.edu>> >> Web :http://reuter.mit.edu <http://reuter.mit.edu/> >> >> >> _______________________________________________ >> Freesurfer mailing list >> Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> >> <mailto:Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> >> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer > > -- > Martin Reuter, Ph.D. > > Instructor in Neurology > Harvard Medical School > Assistant in Neuroscience > Dept. of Radiology, Massachusetts General Hospital > Dept. of Neurology, Massachusetts General Hospital > Research Affiliate > Computer Science and Artificial Intelligence Lab, > Dept. of Electrical Engineering and Computer Science, > Massachusetts Institute of Technology > > A.A.Martinos Center for Biomedical Imaging > 149 Thirteenth Street, Suite 2301 > Charlestown, MA 02129 > > Phone: +1-617-724-5652 <tel:%2B1-617-724-5652> > Email: > mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> > reuter@mit.edu <mailto:reuter@mit.edu> > Web : http://reuter.mit.edu <http://reuter.mit.edu/> > > _______________________________________________ > Freesurfer mailing list > Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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. <02-06-2014, 22.17.45.png><02-06-2014, 22.17.55.png><only LH posterior.png><norm.png>_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer--------------------------------- Dr. Martin Reuter Assistant in Neuroscience - Massachusetts General Hospital Instructor in Neurology - Harvard Medical School MGH / HMS / MIT A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129 Phone: +1-617-724-5652 <tel:%2B1-617-724-5652> Email: mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> reuter@mit.edu <mailto:reuter@mit.edu> Web : http://reuter.mit.edu <http://reuter.mit.edu/> _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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
Hi Martin, thanks a lot for your answers. The slopes command is giving me errors though, see below. I also attach the long.qdec.table.dat that I'm using. I should point out that those ?h.long.cortex.label files mentioned in the error messages are in fact at their expected locations, and appear to be fine (i.e. are not size 0)
In case anyone else can advice about question 1, I'd be really grateful!
*Command*
*Error*
long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas area --hemi lh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time day --qcache fsaverage --sd $SUBJECTS_DIR
No such file or directory
mris_label_calc: could not open label file /vols/Scratch/tpopescu/learning/FS/F02/label/lh.long.cortex.label
Invalid argument
mris_label_calc: mris_label_calc.cpp:127: int main(int, char**): Assertion `l2 != __null' failed.
long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas area --hemi rh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time day --qcache fsaverage --sd $SUBJECTS_DIR
mris_calc: could not open label file /vols/Scratch/tpopescu/learning/FS/F23/label/rh.long.cortex.label
Invalid argument
mris_calc: could not read label file /vols/Scratch/tpopescu/learning/FS/F23/label/rh.long.cortex.label
Invalid argument
long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas lGI --hemi lh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time day --qcache fsaverage --sd $SUBJECTS_DIR
mris_label_calc: mris_label_calc.cpp:127: int main(int, char**): Assertion `l2 != __null' failed.
SUBJECT F02 Intersecting Within-Subject Cortex Label
cp /vols/Scratch/tpopescu/learning/FS/F02_d1.long.F02/label/lh.cortex.label /vols/Scratch/tpopescu/learning/FS/F02/label/lh.long.cortex.label
mris_label_calc intersect /vols/Scratch/tpopescu/learning/FS/F02_d5.long.F02/label/lh.cortex.label /vols/Scratch/tpopescu/learning/FS/F02/label/lh.long.cortex.label /vols/Scratch/tpopescu/learning/FS/F02/label/lh.long.cortex.label
ERROR -6 : mris_label_calc intersect did not work?
long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas lGI --hemi rh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time day --qcache fsaverage --sd $SUBJECTS_DIR
No such file or directory
mris_label_calc: could not open label file /vols/Scratch/tpopescu/learning/FS/F02/label/rh.long.cortex.label
Invalid argument
mris_label_calc: mris_label_calc.cpp:127: int main(int, char**): Assertion `l2 != __null' failed.
On 5 June 2014 16:00, Martin Reuter mreuter@nmr.mgh.harvard.edu wrote:
Hi Tudor,
- not sure, this is a more general question for people more familiar with
surfaces and edits
- nothing is necessary, however, if you have regions where you find
difficulties in some subjects, I'd look at that region in all subjects (e.g. you mention the posterior end). Usually if surfaces on the base are good, the long should be high quality also.
- looks good (I'd do space separated, also some columns merged in your
email and fsid-base for subject3 is wrong).
- using the two-stage model you could do a pairwise analysis between your
groups in the second stage, so linear mixed effects model is not necessary. In the two stage model, just think of it as a cross sectional analysis of e.g. thickness change, instead of thickness.
- yes, is sutiable. You'd have to run it several times, once for each of
the measures you want to analyze. You can drop some of:
--do-avg --do-rate --do-pc1 --do-spc --do-stack
and only keep e.g. the rate or the pc1 (depending on what you want to analyze). You can also use --stack-rate to generate a stack of the rate maps.
Best, Martin
On 06/04/2014 01:19 PM, Tudor Popescu wrote:
Hi Martin,I reran everything from scratch and this time the recons ended without error. Few more questions:
- Is the brainmask.mgz vs wm.mgz problem that I described in my
previous email (surfaces only starting to show rather late at the posterior end of some subjects) something that I need to correct (e.g., by editing wm.mgz), or is it fine as it is? 2) Given that I visually inspected all bases, Is it necessary to inspect *all *slices for the longitudinal overlaps (vols+surfs) of all subjects, or is it enough to just check the central coronal slice for each subject? I ask because it's not just *opening *freeview with all those layers that is very slow, but also navigating between slices 3) The tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorialcovers the creation of long.qdec.table.dat for a single-group scenario; I have 3 groups, with scans on the first and final (fifth) day of training, and was wondering whether the following table looks correct
*fsid* *fsid-base* *group* *day**ageDemeaned * *gender* *score* subj1_day1 subj1 treatmentA 1 1.34 f 109 subj1_day5 subj1 treatmentA 5 1.34 f 109 subj2_day1 subj2 treatmentB 1 -5.85 m 102 subj2_day5 subj2 treatmentB 5 -5.85 m 102 subj3_day1 subj2 control 1 3.06 f 98 subj3_day5 subj2 control 5 3.06 f 98 4) Is this mixed design (3 groups and 2 time points, with both effect of group and of time being relevant) suitably analysed with only a simple 2 stage model, or is a linear mixed effects model necessary? Also, how are the stats different given there are 3 groups and not the usual 2 than most FS tutorials are based on? 5) Is the default long_mris_slopes command (which I understand does preliminary analyses, before going to QDEC) mentioned in the tutorial, i.e.
long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas thickness --hemi lh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time years --qcache fsaverage --sd $SUBJECTS_DIR
suitable for my particular case? I will also look at surface area and lGI, so I suspect it's at least "--meas area" and "--meas lGI" that I need to add to the above command?
Thanks so much! Tudor
On 3 June 2014 02:02, Martin Reuter mreuter@nmr.mgh.harvard.edu wrote:
HI Tudor,
you can run the -long processes in parallel. Something else is wrong. Check if you have enough disk space. Re-run the long from scratch (to make sure everything is generated new).
Best, Martin
On Jun 2, 2014, at 5:34 PM, Tudor Popescu tudor3@gmail.com wrote:
Hi BruceI suspected this might be the case, since assessing 3D surfaces from a succession of 2D slices can be tricky, which is why I was wondering whether there's a better way to do these visual inspections. I checked those apparent artefacts in the other 2 planes (axial, saggital), but did not find it helpful in deciding whether the surfs was as they should or not.
Something that I think is more likely to be a real problem (i.e., more than just a viewing plane disorientation), is that for many subjects, posterior slices only start to be followed by the surfaces quite late, e.g. from coronal slice 80 onward. This seems to be because those slices only appear in the brainmask.mgz, but not also in the wm.mgz (see screenshots attached).
Also, after having run the cross and base steps, and ended up with folders of the type subj1_scanA, subj1_scanB etc, for the "long" step I launched simultaneously (for all subjects&scans) commands of the type recon-all -long subj1_scanA subj1 -all recon-all -long subj1_scanB subj1 -all however these resulted in two types of error messages:
mghRead(/vols/Scratch/tpopescu/learning/FS/subj1_scanA.long.subj1/mri/aseg.mgz, -1): read error* (those particular aseg.mgz files had size 0kB), and
- (standard_in) 2: Error: comparison in expression*
I initially had disk space issues, but after making space I relaunched the same commands with the -no-isrunning flag, but with the same results. Is it the case that the long commands for the different scans of a subject have to be launched sequentially rather than in parallel, i.e. the command for scanB should be run after the command for scanA has finished? Otherwise, what could the problem be?
Thanks again for your help.
Tudor
On 1 June 2014 15:17, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Tudor
are you assessing the thickness visually? If so, you are probably being misled by the orientation of the surface w.r.t. the viewing plane. Same fo the apparent bubbles. You need to look in a different orientation
cheers Bruce
On Sun, 1 Jun 2014, Tudor Popescu wrote:
Many thanks Martin and Nick! The rendering option was already checked,
but the freeview inspection commands were still very very slow, so I used commands of the following form, which as far as I understand should be equivalent: tkmedit subjID norm.mgz -surfs For several subjects (maybe 20 out of the total of 72), I noticed that the surfaces don\t seem to quite follow the GM/WM demarcation line as expected - for instance, by having portions of white surface (yellow line, see attached screenshots) surrounded by pial surface (red line), or by having very large cortical thickness in some parts of the brain. Is it possible that these apparent artefacts make sense "in context" (by looking at the adjacent slides), or is it the case that I need to go back and do white surface correction, and then redo all 3 steps of the longitudinal process?
Thanks again for your help! Tudor
On 29 May 2014 17:39, Nick Schmansky nicks@nmr.mgh.harvard.edu wrote: Tudor,
in your virtual machine, make sure the 3D rendering option, or 'use hardware rendering' (or something like that) is enabled. this would be on the Windows VM config side of things. otherwise it will do software rendering which is painfully slow. Nick > Hi Tudor, > > I don't think there is a way to speed things up. > Let me know if you find a case where the template is blurry or has > ghosts. It should not happen, but if it does it indicates a bad > registratration, you'd have to run the mri_robust_template command with > different parameters manually then. > > Best, Martin > > On 05/27/2014 06:13 PM, Tudor Popescu wrote: >> >> Hi Martin, >> Wasn't sure whether you'd seen my reply below, look forward to hear >> back your thoughts, thanks! >> Tudor >> >> On 25 May 2014 21:40, "Tudor Popescu" <tudor3@gmail.commailto:tudor3@gmail.com> wrote:
Thanks very much Martin and Bruce. I guess I'd misread the Wiki (my own fault, not the text's), and am glad to hear that the longitudinal pipeline is in fact perfectly suitable for myneeds
here. Having run the first 2 steps (cross and base), I'm a bitunclear
how the output so far has to be manually inspected. It says inthe
tutorial <http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial> that you should load each subject's base volume + surfs in freeview and then "move back and forth a few slices". However, even just loading each base in this manner takes ~1 min on myPC
(CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving with PgUp/PgDn between all coronal slices (starting from the default slice=128, going all the way posterior and then all the way anterior) is excruciatingly slow. All of this would have tobe
repeated for all my 72 subjects - is there any way to optimise this manual inspection? Also, if the surfs turn out to not follow the volume correctly, presumably the thing to do is white surface correction + re-running recon. But what should one do if, due to anerroneous
averaging between timepoints, you see blurs/ghosts in your base template? Many thanks! Tudor On 9 May 2014 21:33, Martin Reuter <mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu>> wrote: Hi Tudor, the longitudinal pipeline in FS is actually one of the beston
the planet as far as I know :-). If there is anycontradictory
information on the wiki, can you point me to that so I cansee
what causes the misconception. Really: compared toindependent
processing, it significantly increases sensitivity. Also we have designed it to be unbiased with respect to a singletime
point or directionality. It is quite mature by now. You should definitely use the longitudinal pipeline for the analysis of your data. Now to your questions 1. QDEC: I am not too familiar with qdec. You candefinitely
try the 2-stage approach described on the wiki. There you first compute a measure of change (e.g. hippocampal volume change during your week) and then compare that measureacross
groups similar to a cross sectional volume/thicknessanalysis.
You can also use our tools to run a linear mixed effectsmodel
if you want to do that (it is more involved and requiresyou
to use matlab tools). In your case, you probably have 2time
points for all subjects and the time distance is probablythe
same for all subjects, so the 2-stage approach should befine.
2. The image processing is done via the longitudinalpipeline
(three steps: cross, base, long), to prepare the data lookat
the description of the 2-stage model http://freesurfer.net/fswiki/LongitudinalTwoStageModel and also the longitudinal tutorialhttp://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial
3. At the recon all level in FS you get (after the 3 steps) measurement for all time points. So you would compare those results across time in the stats. Hope that helps, Martin On 05/08/2014 08:14 AM, Tudor Popescu wrote:> Sorry for the repeat, wasn't sure whether this was
received
> the first time. > Tudor > > > On 6 May 2014 19:55, Tudor Popescu <tudor3@gmail.com > mailto:tudor3@gmail.com> wrote: > > Dear FS list, > > I have structural data from a learning study > (pre&post-training scans, with 3 groups). Although the > training was only one week, I'm guessing from an
analysis
> point of view, this still qualifies as longitudinal. I > want to check for > > * the main within-subjects effect of time point > (pre&post) > * the main between-subjects effect of group
(treatment
> A, treatment B, control), > * the time x group interaction > > I intend to look at thickness, surface area, volume,
and
> lGI. > > I read on the wiki >
http://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing
> that FS is currently not optimal for longitudinal > analyses. I intend my FreeSurfer analysis to
supplement a
> VBM analysis done in FSL. In case it is in fact a good > idea to do this, my questions (not covered in the > 'longitudinal' wiki page) are: > > 1) Can QDEC be used for such an analysis, and if so,
what
> would be different as compared to a cross-sectional
(no
> temporal/within factor) study? > > 2) Also, is the pre-processing stage any different? > > 3) In FSL, for longitudinal designs you do stats on > images obtained as the difference between consecutive > time points. Does this have to be done in FreeSurfer
as
> well, and if so, is it done at the recon-all level or > only at the stats (QDEC) level? > > Thanks! > > Tudor > > > > > > _______________________________________________ > Freesurfer mailing list > Freesurfer@nmr.mgh.harvard.edu > mailto:Freesurfer@nmr.mgh.harvard.edu >
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Martin Reuter, Ph.D. Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129 Phone:+1-617-724-5652 <tel:%2B1-617-724-5652> Email: mreuter@nmr.mgh.harvard.edumailto:mreuter@nmr.mgh.harvard.edu reuter@mit.edu mailto:reuter@mit.edu Web :http://reuter.mit.edu
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu>https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for theperson
to whom it is addressed. If you believe this e-mail was sent to you inerror
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 contactthe
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
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 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.
<02-06-2014, 22.17.45.png><02-06-2014, 22.17.55.png><only LH posterior.png><norm.png>_______________________________________________
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Dr. Martin Reuter Assistant in Neuroscience - Massachusetts General Hospital Instructor in Neurology - Harvard Medical School MGH / HMS / MIT
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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.
I should add that the commands for thickness, i.e. long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas thickness --hemi ?h --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time day --qcache fsaverage --sd $SUBJECTS_DIR finished without error, and for some reason, so did the surface area commands, when I ran them a second time. Also, I realised that for lGI, the flag for --meas was pial_lgi and not just lgi. However, the command long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas pial_lgi --hemi lh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time day --qcache fsaverage --sd $SUBJECTS_DIR resulted in this error, similar for both hemispheres: SUBJECT F02 Stacking Within-Subject Maps mri_concat /vols/Scratch/tpopescu/learning/FS/F02_d1.long.F02/surf/rh.pial_lgi /vols/Scratch/tpopescu/learning/FS/F02_d5.long.F02/surf/rh.pial_lgi --o /vols/Scratch/tpopescu/learning/FS/F02/surf/rh.long.pial_lgi-stack.mgh ERROR 1 : mri_concat stacking did not work?
Thanks again for your help!
On 6 June 2014 17:45, Tudor Popescu tudor3@gmail.com wrote:
Hi Martin, thanks a lot for your answers. The slopes command is giving me errors though, see below. I also attach the long.qdec.table.dat that I'm using. I should point out that those ?h.long.cortex.label files mentioned in the error messages are in fact at their expected locations, and appear to be fine (i.e. are not size 0)
In case anyone else can advice about question 1, I'd be really grateful!
*Command*
*Error*
long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas area --hemi lh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time day --qcache fsaverage --sd $SUBJECTS_DIR
No such file or directory
mris_label_calc: could not open label file /vols/Scratch/tpopescu/learning/FS/F02/label/lh.long.cortex.label
Invalid argument
mris_label_calc: mris_label_calc.cpp:127: int main(int, char**): Assertion `l2 != __null' failed.
long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas area --hemi rh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time day --qcache fsaverage --sd $SUBJECTS_DIR
mris_calc: could not open label file /vols/Scratch/tpopescu/learning/FS/F23/label/rh.long.cortex.label
Invalid argument
mris_calc: could not read label file /vols/Scratch/tpopescu/learning/FS/F23/label/rh.long.cortex.label
Invalid argument
long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas lGI --hemi lh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time day --qcache fsaverage --sd $SUBJECTS_DIR
mris_label_calc: mris_label_calc.cpp:127: int main(int, char**): Assertion `l2 != __null' failed.
SUBJECT F02 Intersecting Within-Subject Cortex Label
cp /vols/Scratch/tpopescu/learning/FS/F02_d1.long.F02/label/lh.cortex.label /vols/Scratch/tpopescu/learning/FS/F02/label/lh.long.cortex.label
mris_label_calc intersect /vols/Scratch/tpopescu/learning/FS/F02_d5.long.F02/label/lh.cortex.label /vols/Scratch/tpopescu/learning/FS/F02/label/lh.long.cortex.label /vols/Scratch/tpopescu/learning/FS/F02/label/lh.long.cortex.label
ERROR -6 : mris_label_calc intersect did not work?
long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas lGI --hemi rh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time day --qcache fsaverage --sd $SUBJECTS_DIR
No such file or directory
mris_label_calc: could not open label file /vols/Scratch/tpopescu/learning/FS/F02/label/rh.long.cortex.label
Invalid argument
mris_label_calc: mris_label_calc.cpp:127: int main(int, char**): Assertion `l2 != __null' failed.
On 5 June 2014 16:00, Martin Reuter mreuter@nmr.mgh.harvard.edu wrote:
Hi Tudor,
- not sure, this is a more general question for people more familiar
with surfaces and edits
- nothing is necessary, however, if you have regions where you find
difficulties in some subjects, I'd look at that region in all subjects (e.g. you mention the posterior end). Usually if surfaces on the base are good, the long should be high quality also.
- looks good (I'd do space separated, also some columns merged in your
email and fsid-base for subject3 is wrong).
- using the two-stage model you could do a pairwise analysis between
your groups in the second stage, so linear mixed effects model is not necessary. In the two stage model, just think of it as a cross sectional analysis of e.g. thickness change, instead of thickness.
- yes, is sutiable. You'd have to run it several times, once for each of
the measures you want to analyze. You can drop some of:
--do-avg --do-rate --do-pc1 --do-spc --do-stack
and only keep e.g. the rate or the pc1 (depending on what you want to analyze). You can also use --stack-rate to generate a stack of the rate maps.
Best, Martin
On 06/04/2014 01:19 PM, Tudor Popescu wrote:
Hi Martin,I reran everything from scratch and this time the recons ended without error. Few more questions:
- Is the brainmask.mgz vs wm.mgz problem that I described in my
previous email (surfaces only starting to show rather late at the posterior end of some subjects) something that I need to correct (e.g., by editing wm.mgz), or is it fine as it is? 2) Given that I visually inspected all bases, Is it necessary to inspect *all *slices for the longitudinal overlaps (vols+surfs) of all subjects, or is it enough to just check the central coronal slice for each subject? I ask because it's not just *opening *freeview with all those layers that is very slow, but also navigating between slices 3) The tutorial http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorialcovers the creation of long.qdec.table.dat for a single-group scenario; I have 3 groups, with scans on the first and final (fifth) day of training, and was wondering whether the following table looks correct
*fsid* *fsid-base* *group* *day**ageDemeaned * *gender* *score* subj1_day1 subj1 treatmentA 1 1.34 f 109 subj1_day5 subj1 treatmentA 5 1.34 f 109 subj2_day1 subj2 treatmentB 1 -5.85 m 102 subj2_day5 subj2 treatmentB 5 -5.85 m 102 subj3_day1 subj2 control 1 3.06 f 98 subj3_day5 subj2 control 5 3.06 f 98 4) Is this mixed design (3 groups and 2 time points, with both effect of group and of time being relevant) suitably analysed with only a simple 2 stage model, or is a linear mixed effects model necessary? Also, how are the stats different given there are 3 groups and not the usual 2 than most FS tutorials are based on? 5) Is the default long_mris_slopes command (which I understand does preliminary analyses, before going to QDEC) mentioned in the tutorial, i.e.
long_mris_slopes --qdec ./qdec/long.qdec.table.dat --meas thickness --hemi lh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time years --qcache fsaverage --sd $SUBJECTS_DIR
suitable for my particular case? I will also look at surface area and lGI, so I suspect it's at least "--meas area" and "--meas lGI" that I need to add to the above command?
Thanks so much! Tudor
On 3 June 2014 02:02, Martin Reuter mreuter@nmr.mgh.harvard.edu wrote:
HI Tudor,
you can run the -long processes in parallel. Something else is wrong. Check if you have enough disk space. Re-run the long from scratch (to make sure everything is generated new).
Best, Martin
On Jun 2, 2014, at 5:34 PM, Tudor Popescu tudor3@gmail.com wrote:
Hi BruceI suspected this might be the case, since assessing 3D surfaces from a succession of 2D slices can be tricky, which is why I was wondering whether there's a better way to do these visual inspections. I checked those apparent artefacts in the other 2 planes (axial, saggital), but did not find it helpful in deciding whether the surfs was as they should or not.
Something that I think is more likely to be a real problem (i.e., more than just a viewing plane disorientation), is that for many subjects, posterior slices only start to be followed by the surfaces quite late, e.g. from coronal slice 80 onward. This seems to be because those slices only appear in the brainmask.mgz, but not also in the wm.mgz (see screenshots attached).
Also, after having run the cross and base steps, and ended up with folders of the type subj1_scanA, subj1_scanB etc, for the "long" step I launched simultaneously (for all subjects&scans) commands of the type recon-all -long subj1_scanA subj1 -all recon-all -long subj1_scanB subj1 -all however these resulted in two types of error messages:
mghRead(/vols/Scratch/tpopescu/learning/FS/subj1_scanA.long.subj1/mri/aseg.mgz, -1): read error* (those particular aseg.mgz files had size 0kB), and
- (standard_in) 2: Error: comparison in expression*
I initially had disk space issues, but after making space I relaunched the same commands with the -no-isrunning flag, but with the same results. Is it the case that the long commands for the different scans of a subject have to be launched sequentially rather than in parallel, i.e. the command for scanB should be run after the command for scanA has finished? Otherwise, what could the problem be?
Thanks again for your help.
Tudor
On 1 June 2014 15:17, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote:
Hi Tudor
are you assessing the thickness visually? If so, you are probably being misled by the orientation of the surface w.r.t. the viewing plane. Same fo the apparent bubbles. You need to look in a different orientation
cheers Bruce
On Sun, 1 Jun 2014, Tudor Popescu wrote:
Many thanks Martin and Nick! The rendering option was already checked,
but the freeview inspection commands were still very very slow, so I used commands of the following form, which as far as I understand should be equivalent: tkmedit subjID norm.mgz -surfs For several subjects (maybe 20 out of the total of 72), I noticed that the surfaces don\t seem to quite follow the GM/WM demarcation line as expected - for instance, by having portions of white surface (yellow line, see attached screenshots) surrounded by pial surface (red line), or by having very large cortical thickness in some parts of the brain. Is it possible that these apparent artefacts make sense "in context" (by looking at the adjacent slides), or is it the case that I need to go back and do white surface correction, and then redo all 3 steps of the longitudinal process?
Thanks again for your help! Tudor
On 29 May 2014 17:39, Nick Schmansky nicks@nmr.mgh.harvard.edu wrote: Tudor,
in your virtual machine, make sure the 3D rendering option, or 'use hardware rendering' (or something like that) is enabled. this would be on the Windows VM config side of things. otherwise it will do software rendering which is painfully slow. Nick > Hi Tudor, > > I don't think there is a way to speed things up. > Let me know if you find a case where the template is blurry or has > ghosts. It should not happen, but if it does it indicates a bad > registratration, you'd have to run the mri_robust_template command with > different parameters manually then. > > Best, Martin > > On 05/27/2014 06:13 PM, Tudor Popescu wrote: >> >> Hi Martin, >> Wasn't sure whether you'd seen my reply below, look forward to hear >> back your thoughts, thanks! >> Tudor >> >> On 25 May 2014 21:40, "Tudor Popescu" <tudor3@gmail.com> mailto:tudor3@gmail.com> wrote: > > Thanks very much Martin and Bruce. I guess I'd misread the Wiki > (my own fault, not the text's), and am glad to hear that the > longitudinal pipeline is in fact perfectly suitable for my
needs
> here. > > Having run the first 2 steps (cross and base), I'm a bit
unclear
> how the output so far has to be manually inspected. It says in
the
> tutorial > http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial > that you should load each subject's base volume + surfs in > freeview and then "move back and forth a few slices". However, > even just loading each base in this manner takes ~1 min on my
PC
> (CoreDuo, 4GB, Ubuntu Virtualbox in Windows 7), and then moving > with PgUp/PgDn between all coronal slices (starting from the > default slice=128, going all the way posterior and then all the > way anterior) is excruciatingly slow. All of this would have to
be
> repeated for all my 72 subjects - is there any way to optimise > this manual inspection? > > Also, if the surfs turn out to not follow the volume correctly, > presumably the thing to do is white surface correction + > re-running recon. But what should one do if, due to an
erroneous
> averaging between timepoints, you see blurs/ghosts in your base > template? > > Many thanks! > Tudor > > > On 9 May 2014 21:33, Martin Reuter <mreuter@nmr.mgh.harvard.edu > mailto:mreuter@nmr.mgh.harvard.edu> wrote: > > Hi Tudor, > > the longitudinal pipeline in FS is actually one of the best
on
> the planet as far as I know :-). If there is any
contradictory
> information on the wiki, can you point me to that so I can
see
> what causes the misconception. Really: compared to
independent
> processing, it significantly increases sensitivity. Also we > have designed it to be unbiased with respect to a single
time
> point or directionality. It is quite mature by now. > > You should definitely use the longitudinal pipeline for the > analysis of your data. Now to your questions > > 1. QDEC: I am not too familiar with qdec. You can
definitely
> try the 2-stage approach described on the wiki. There you > first compute a measure of change (e.g. hippocampal volume > change during your week) and then compare that measure
across
> groups similar to a cross sectional volume/thickness
analysis.
> You can also use our tools to run a linear mixed effects
model
> if you want to do that (it is more involved and requires
you
> to use matlab tools). In your case, you probably have 2
time
> points for all subjects and the time distance is probably
the
> same for all subjects, so the 2-stage approach should be
fine.
> > 2. The image processing is done via the longitudinal
pipeline
> (three steps: cross, base, long), to prepare the data look
at
> the description of the 2-stage model > http://freesurfer.net/fswiki/LongitudinalTwoStageModel > and also the longitudinal tutorial >
http://freesurfer.net/fswiki/FsTutorial/LongitudinalTutorial
> > 3. At the recon all level in FS you get (after the 3 steps) > measurement for all time points. So you would compare those > results across time in the stats. > > Hope that helps, Martin > > > On 05/08/2014 08:14 AM, Tudor Popescu wrote: >> Sorry for the repeat, wasn't sure whether this was
received
>> the first time. >> Tudor >> >> >> On 6 May 2014 19:55, Tudor Popescu <tudor3@gmail.com >> mailto:tudor3@gmail.com> wrote: >> >> Dear FS list, >> >> I have structural data from a learning study >> (pre&post-training scans, with 3 groups). Although the >> training was only one week, I'm guessing from an
analysis
>> point of view, this still qualifies as longitudinal. I >> want to check for >> >> * the main within-subjects effect of time point >> (pre&post) >> * the main between-subjects effect of group
(treatment
>> A, treatment B, control), >> * the time x group interaction >> >> I intend to look at thickness, surface area, volume,
and
>> lGI. >> >> I read on the wiki >>
http://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing
>> that FS is currently not optimal for longitudinal >> analyses. I intend my FreeSurfer analysis to
supplement a
>> VBM analysis done in FSL. In case it is in fact a good >> idea to do this, my questions (not covered in the >> 'longitudinal' wiki page) are: >> >> 1) Can QDEC be used for such an analysis, and if so,
what
>> would be different as compared to a cross-sectional
(no
>> temporal/within factor) study? >> >> 2) Also, is the pre-processing stage any different? >> >> 3) In FSL, for longitudinal designs you do stats on >> images obtained as the difference between consecutive >> time points. Does this have to be done in FreeSurfer
as
>> well, and if so, is it done at the recon-all level or >> only at the stats (QDEC) level? >> >> Thanks! >> >> Tudor >> >> >> >> >> >> _______________________________________________ >> Freesurfer mailing list >> Freesurfer@nmr.mgh.harvard.edu >> mailto:Freesurfer@nmr.mgh.harvard.edu >>
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
> > -- > Martin Reuter, Ph.D. > > Instructor in Neurology > Harvard Medical School > Assistant in Neuroscience > Dept. of Radiology, Massachusetts General Hospital > Dept. of Neurology, Massachusetts General Hospital > Research Affiliate > Computer Science and Artificial Intelligence Lab, > Dept. of Electrical Engineering and Computer Science, > Massachusetts Institute of Technology > > A.A.Martinos Center for Biomedical Imaging > 149 Thirteenth Street, Suite 2301 > Charlestown, MA 02129 > > Phone:+1-617-724-5652 tel:%2B1-617-724-5652 > Email: > mreuter@nmr.mgh.harvard.edu > mailto:mreuter@nmr.mgh.harvard.edu > reuter@mit.edu mailto:reuter@mit.edu > Web :http://reuter.mit.edu > > > _______________________________________________ > Freesurfer mailing list > Freesurfer@nmr.mgh.harvard.edu > mailto: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
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 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.
<02-06-2014, 22.17.45.png><02-06-2014, 22.17.55.png><only LH posterior.png><norm.png>_______________________________________________
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Dr. Martin Reuter Assistant in Neuroscience - Massachusetts General Hospital Instructor in Neurology - Harvard Medical School MGH / HMS / MIT
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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 listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Martin Reuter, Ph.D.
Instructor in Neurology Harvard Medical School Assistant in Neuroscience Dept. of Radiology, Massachusetts General Hospital Dept. of Neurology, Massachusetts General Hospital Research Affiliate Computer Science and Artificial Intelligence Lab, Dept. of Electrical Engineering and Computer Science, Massachusetts Institute of Technology
A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129
Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu
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@nmr.mgh.harvard.edu
-
Bruce Fischl -
Martin Reuter -
Nick Schmansky -
Tudor Popescu