Thanks again Mark.
The pixel dimension of a voxel (pixdim 1,2,3 ) is 1 mm in all data sets. So it appears that the voxel count outputof your script will not need to be multiplied by the volume of a single voxel. Please let me know if this is not correct.
Best,
Lance
On Wed, May 18, 2016 at 5:00 AM, Mark Jenkinson <[log in to unmask]> wrote:
Hi,
The *_origsegs* file contains segmentations that are not boundary corrected, so you do not want to use these. If you want to calculate volumes then you should use the boundary corrected output, which are in the *_firstseg* file. The individual structure files such as L_Hipp_corr.nii.gz also contain boundary corrected results and can be used.
If you want a quick calculation of all the volumes then the following script will do it:
im=YOUR_IMAGE_NAME_firstseg.nii.gzmaxn=`fslstats $im -P 100`fslstats $im -h $maxn | grep [1-9] | sed -n '2,$p'
The easiest way to do this is in your terminal. Just type out the first line but replace "YOUR_IMAGE_NAME_firstseg.nii.gz" with the name/path of your firstseg file. Once you've done that you can just cut and paste the next two lines into your terminal. Note that some email clients do bad things with quotes (backward quotes on the second line, and forward ones in the third line) so you may need to type/fix instead of a simple paste.
The volumes that you get out are voxel counts, so you'll need to multiply them by the volume of a single voxel to convert them to mm^3 (or unit of your choice). The order of the values is the same as the order of the labels on the wiki:
All the best,Mark
From: FSL - FMRIB's Software Library <[log in to unmask]> on behalf of Lance Stevens <[log in to unmask]>
Reply-To: FSL - FMRIB's Software Library <[log in to unmask]>
Date: Thursday, 12 May 2016 21:59
To: "[log in to unmask]" <[log in to unmask]>
Subject: [FSL] subcortical volume
The FIRST FSL course material online indicates that volume would be calculated from a mask/ROI of the total output_name_all_fast_origsegs.nii.gz or on individual structures (e.g L_Hipp_corr.nii.gz).
To assess a group of subcortical structures (e.g. L_Accu L_Amyg L_Caud L_Hipp) between groups it looks like the "for loop" could be used, as indicated in the FIRST user guide, and masks made of each individual's all_fast_origsegs.nii.gz output file; the mask volumes could then be calculated and compared between or among groups in Matlab or some other program. Also, while the structure labels and locations distinguish the individual structures, the all_fast_origsegs.nii.gz output file is a collective of all the structures; masks of individual structures can not be readily made.
Further, to compare individual subcortical structures between (or among) groups, masks of individual structures can be made from running first on individual subjects and individual structures; again all volume could be grouped and calculated in Matlab or some other program.
Are these assumptions correct?
Lance