Hi Paul, Just as an aside, once you have a talairach_with_skull.lta file (whether from v3 or v4), you can calculate eTIV as scale_factor/determinant(lta), where the scale_factor employed is 2150.
Incidentally, using the -eTIV option of mri_label_volume, you can also specify other transforms (e.g., talairach.xfm, talairach.lta) to be used used instead of the default talairach_with_skull.lta in computing "ICV". I've looked at the effect of using different transforms in a handful of brains, and "ICV" can vary a bit depending on transform you choose to use.
I'm curious: Has anyone out there compared FS's ICV estimate to ICV estimated by another program?
Best, Mike H.
On Fri, 2008-06-20 at 14:35 -0700, Paul Greenberg wrote:
Hello Freesurfers,
I am interested in comparing eTIV calculated with version 3.0.5 to that calculated by version 4.0.5 (64 bit Cent OS).
In order to compare eTIV for a single brain estimated in version 3.0.5 to version 4.0.5, I copied a subject's previous freesurfer output (v3.0.5) into the freesurfer version 4.0.5 folder, deleted all the files within the "transforms" folder and carried out the following procedures as outlined in the recon-all dev table:
- Run the talairach procedure. Command = talairach_avi --i
mysubject --xfm talairach.auto.xfm 2) cp talairach.auto.xfm talairach.xfm 3) Run automatic failure detection. Command= talairach_afd -T 0.005 -xfm talairach.xfm 4) Create talairach_with_skull.lta. Command = mri_em_register -skull nu.mgz $freesurferhome/average/RB_all_withskull_2007-08-08.gca talairach_with_skull.lta 5) Check the new eTIV. Command = mri_label_volume -eTIV talairach_with_skull.lta aseg.mgz 17 53 As the aseg.mgz file exists from previous processing (versio 3.0.5), and the talairach_with_skull.lta was re-created with the new processing (version 4.0.5), the output of this command should list the eTIV based on the new talairach transform?
Are these procedures correct for generating new values of eTIV if that is all i want?
Thanks!
--
Paul Greenberg, Ph.D. Radiology Department, University of California, San Diego VA San Diego Healthcare System (116A-13) 3350 La Jolla Village Drive San Diego, CA 92161 Phone: (858) 552-8585 ext. 2903 Fax: (858) 642-3836 pgreenberg@ucsd.edu _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Hi, for those interested: we have correlated FS's ICV estimates and estimates from a procedure developed by Anders Dale (a deformable template procedure, similar to the Shrink Wrapping procedure previously described by Dale and colleagues (Dale et al, 1999; Dale and Sereno, 1993)).
Mean ICV from FS: 1567690 (SD = 161247) Mean ICV from Dale: 1584008 (SD = 136979)
The difference was not significant (p = .12, t = -1.596)
The correlation was .84. A scatterplot is attached (FS on the y-axis, Dale's method on the x).
Not too bad, but not perfect either.
The sample was a life-span sample (20-90 years, n = 74).
Hope this info was usefull.
Best, Anders MF
Hi Paul, Just as an aside, once you have a talairach_with_skull.lta file (whether from v3 or v4), you can calculate eTIV as scale_factor/determinant(lta), where the scale_factor employed is 2150.
Incidentally, using the -eTIV option of mri_label_volume, you can also specify other transforms (e.g., talairach.xfm, talairach.lta) to be used used instead of the default talairach_with_skull.lta in computing "ICV". I've looked at the effect of using different transforms in a handful of brains, and "ICV" can vary a bit depending on transform you choose to use.
I'm curious: Has anyone out there compared FS's ICV estimate to ICV estimated by another program?
Best, Mike H.
On Fri, 2008-06-20 at 14:35 -0700, Paul Greenberg wrote:
Hello Freesurfers,
I am interested in comparing eTIV calculated with version 3.0.5 to that calculated by version 4.0.5 (64 bit Cent OS).
In order to compare eTIV for a single brain estimated in version 3.0.5 to version 4.0.5, I copied a subject's previous freesurfer output (v3.0.5) into the freesurfer version 4.0.5 folder, deleted all the files within the "transforms" folder and carried out the following procedures as outlined in the recon-all dev table:
- Run the talairach procedure. Command = talairach_avi --i
mysubject --xfm talairach.auto.xfm 2) cp talairach.auto.xfm talairach.xfm 3) Run automatic failure detection. Command= talairach_afd -T 0.005 -xfm talairach.xfm 4) Create talairach_with_skull.lta. Command = mri_em_register -skull nu.mgz $freesurferhome/average/RB_all_withskull_2007-08-08.gca talairach_with_skull.lta 5) Check the new eTIV. Command = mri_label_volume -eTIV talairach_with_skull.lta aseg.mgz 17 53 As the aseg.mgz file exists from previous processing (versio 3.0.5), and the talairach_with_skull.lta was re-created with the new processing (version 4.0.5), the output of this command should list the eTIV based on the new talairach transform?
Are these procedures correct for generating new values of eTIV if that is all i want?
Thanks!
--
Paul Greenberg, Ph.D. Radiology Department, University of California, San Diego VA San Diego Healthcare System (116A-13) 3350 La Jolla Village Drive San Diego, CA 92161 Phone: (858) 552-8585 ext. 2903 Fax: (858) 642-3836 pgreenberg@ucsd.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
yes, the FS ICVs mostly seem to work :-)
I still think one should watch out for outliers. We had a sample of 28 kids aged 8-13 yrs, and the ICVs were quite a bit off and useless for two of those, so I would check all carefully especially if you have a group that may have smaller than normal adult ICV.
Best,
K
Hi, for those interested: we have correlated FS's ICV estimates and estimates from a procedure developed by Anders Dale (a deformable template procedure, similar to the Shrink Wrapping procedure previously described by Dale and colleagues (Dale et al, 1999; Dale and Sereno, 1993)).
Mean ICV from FS: 1567690 (SD = 161247) Mean ICV from Dale: 1584008 (SD = 136979)
The difference was not significant (p = .12, t = -1.596)
The correlation was .84. A scatterplot is attached (FS on the y-axis, Dale's method on the x).
Not too bad, but not perfect either.
The sample was a life-span sample (20-90 years, n = 74).
Hope this info was usefull.
Best, Anders MF
Hi Paul, Just as an aside, once you have a talairach_with_skull.lta file (whether from v3 or v4), you can calculate eTIV as scale_factor/determinant(lta), where the scale_factor employed is 2150.
Incidentally, using the -eTIV option of mri_label_volume, you can also specify other transforms (e.g., talairach.xfm, talairach.lta) to be used used instead of the default talairach_with_skull.lta in computing "ICV". I've looked at the effect of using different transforms in a handful of brains, and "ICV" can vary a bit depending on transform you choose to use.
I'm curious: Has anyone out there compared FS's ICV estimate to ICV estimated by another program?
Best, Mike H.
On Fri, 2008-06-20 at 14:35 -0700, Paul Greenberg wrote:
Hello Freesurfers,
I am interested in comparing eTIV calculated with version 3.0.5 to that calculated by version 4.0.5 (64 bit Cent OS).
In order to compare eTIV for a single brain estimated in version 3.0.5 to version 4.0.5, I copied a subject's previous freesurfer output (v3.0.5) into the freesurfer version 4.0.5 folder, deleted all the files within the "transforms" folder and carried out the following procedures as outlined in the recon-all dev table:
- Run the talairach procedure. Command = talairach_avi --i
mysubject --xfm talairach.auto.xfm 2) cp talairach.auto.xfm talairach.xfm 3) Run automatic failure detection. Command= talairach_afd -T 0.005 -xfm talairach.xfm 4) Create talairach_with_skull.lta. Command = mri_em_register -skull nu.mgz $freesurferhome/average/RB_all_withskull_2007-08-08.gca talairach_with_skull.lta 5) Check the new eTIV. Command = mri_label_volume -eTIV talairach_with_skull.lta aseg.mgz 17 53 As the aseg.mgz file exists from previous processing (versio 3.0.5), and the talairach_with_skull.lta was re-created with the new processing (version 4.0.5), the output of this command should list the eTIV based on the new talairach transform?
Are these procedures correct for generating new values of eTIV if that is all i want?
Thanks!
--
Paul Greenberg, Ph.D. Radiology Department, University of California, San Diego VA San Diego Healthcare System (116A-13) 3350 La Jolla Village Drive San Diego, CA 92161 Phone: (858) 552-8585 ext. 2903 Fax: (858) 642-3836 pgreenberg@ucsd.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
freesurfer@nmr.mgh.harvard.edu