I had a question about gray matter and white matter volumes in FreeSurfer version 4.5. Sorry, we have too many subjects to rerun easily through 5.0. There is no gray matter listed in aseg.stats in 4.5. The two options I see are using mris_volume or summing the aparc gray matter structures, are there pros/cons to this? More specifically, should the sum of the aparc gm regions equal the surface-based gm volume from mris_volume or will they be different?
The aseg.stats also produces surface-based wm volumes, which I presume need the subcortical structures subtracted away, is there a command to do this with mri_segstats or do I just need to run a loop for every subcortical structure to get these corrected wm volumes?
One last question. The mri_segstats says BrainSeg is computed WITHOUT partial volume correction. My understanding was in the aseg pipeline every voxel got assigned to a specific structure in aseg.mgz. So, I don't undestand how partial voluming comes into play? In BrainSeg are some voxels counted twice as being part of two different structures, and then this is fixed when the volume of each structure is reported in the aseg table? Also, it would be nice to be able to directly compare BrainSeg (no partial volume correction) with BrainSegNotVent (partial volume correction), is this possible? Thanks!
Chris Bell University of Minnesota
One other thing. I did the addition of all the structures in aseg and am a bit confused. Here are the numbers for one subject.
BrainSeg= 1035602 BrainSegNotVent= 1034954 These two are almost identical but that is most likely due to the lack of pv correction in BrainSeg, so I think that makes sense. When I sum all the structures in aseg I get 1027260 (which includes the five CC regions). This number is smaller than BrainSegNotVent, which doesn't make sense to me since I would expect it to be larger than BrainSegNotVent. Sorry for all the questions, I think FreeSurfer is great!
Chris
On Nov 3 2010, bell0368@umn.edu wrote:
I had a question about gray matter and white matter volumes in FreeSurfer version 4.5. Sorry, we have too many subjects to rerun easily through 5.0. There is no gray matter listed in aseg.stats in 4.5. The two options I see are using mris_volume or summing the aparc gray matter structures, are there pros/cons to this? More specifically, should the sum of the aparc gm regions equal the surface-based gm volume from mris_volume or will they be different?
The aseg.stats also produces surface-based wm volumes, which I presume need
the subcortical structures subtracted away, is there a command to do this with mri_segstats or do I just need to run a loop for every subcortical structure to get these corrected wm volumes?
One last question. The mri_segstats says BrainSeg is computed WITHOUT partial volume correction. My understanding was in the aseg pipeline every voxel got assigned to a specific structure in aseg.mgz. So, I don't
undestand how partial voluming comes into play? In BrainSeg are some voxels
counted twice as being part of two different structures, and then this is fixed when the volume of each structure is reported in the aseg table? Also, it would be nice to be able to directly compare BrainSeg (no partial volume correction) with BrainSegNotVent (partial volume correction), is this possible? Thanks!
Chris Bell University of Minnesota
Freesurfers,
So you can ignore this last question. I double-checked my addition and got 1046270 for the sum of all structures which is indeed more than BrainSegNotVent. Oops.
Chris
On Nov 3 2010, bell0368@umn.edu wrote:
One other thing. I did the addition of all the structures in aseg and am a bit confused. Here are the numbers for one subject.
BrainSeg= 1035602 BrainSegNotVent= 1034954 These two are almost identical but that is most likely due to the lack of pv correction in BrainSeg, so I think that makes sense. When I sum all the structures in aseg I get 1027260 (which includes the five CC regions). This number is smaller than BrainSegNotVent, which doesn't make sense to me since I would expect it to be larger than BrainSegNotVent. Sorry for all the questions, I think FreeSurfer is great!
Chris
On Nov 3 2010, bell0368@umn.edu wrote:
I had a question about gray matter and white matter volumes in FreeSurfer
version 4.5. Sorry, we have too many subjects to rerun easily through 5.0. There is no gray matter listed in aseg.stats in 4.5. The two options I see
are using mris_volume or summing the aparc gray matter structures, are
there pros/cons to this? More specifically, should the sum of the aparc gm regions equal the surface-based gm volume from mris_volume or will they be
different?
The aseg.stats also produces surface-based wm volumes, which I presume need
the subcortical structures subtracted away, is there a command to do this with mri_segstats or do I just need to run a loop for every subcortical structure to get these corrected wm volumes?
One last question. The mri_segstats says BrainSeg is computed WITHOUT
partial volume correction. My understanding was in the aseg pipeline every
voxel got assigned to a specific structure in aseg.mgz. So, I don't
undestand how partial voluming comes into play? In BrainSeg are some voxels
counted twice as being part of two different structures, and then this is fixed when the volume of each structure is reported in the aseg table?
Also, it would be nice to be able to directly compare BrainSeg (no partial
volume correction) with BrainSegNotVent (partial volume correction), is this possible? Thanks!
Chris Bell University of Minnesota
freesurfer@nmr.mgh.harvard.edu
-
bell0368＠umn.edu