Dear Victor,
I don't know why you think 10.9 mm^3 is reasonable and 1300 mm^3 is not. I think the opposite. I'm also not sure what you mean by "the computed form of the volume".
For instance, with a 1x1x1mm voxel size (fairly typical for structural imaging), a 10.9 mm^3 volume would be less than 11 voxels in total, which is very, very small. However, a 1300 mm^3 volume would be roughly the size of structure that was 26mm * 7 mm * 7 mm, which would be a pretty typical size for many structures (e.g. the hippocampus is quite similar to these dimensions). Note that with this voxel size the number of voxels and the mm^3 values should be the same.
Maybe we are just talking at crossed purposes, but if you are getting values like 1300 (mm^3) then these are the values you should use and not 10.9.
All the best,
Mark
On 22 Jan 2013, at 23:07, Victor DelBene <[log in to unmask]>
wrote:
> Hi FSL experts,
>
> I had a very quick question about the FSL - FIRST volume output when using fslstats. We are interested in obtaining the volumes of several regions that have been segmented in FIRST.
>
> To get these values, I used the command fslstats output_name_all_fast_firstseg -l 16.5 -u 17.5 -v, with voxels being the first number and volume being the second number. The literature I found on the FSL site awhile back, states that the output is in mm^3 for volume, but when all the volume output are around 1300 (this is a rough estimate of the mean for one of the regions), it is hard to believe the units of measurement is still mm^3, rather than the computed form of the volume.
>
> I took the cubic root of this mean, and got 10.9 mm^3, which is reasonable. Before continuing on with converting all the volumes to mm^3 and running statistical analyses on these values, I wanted to confirm that my line of reasoning is correct.
> Thank you for your help!
>
> Victor
|