search

thomshaw92 / LASHiS

Longitudinal Automatic Segmentation of Hippocampal Subfields (LASHiS) using multi-contrast MRI.
GNU General Public License v2.0
7 stars 4 forks source link

Qeustions about the subfield label stats #7

Closed ash077 closed 1 year ago

ash077 commented 2 years ago

Hi, I'm confused with the output of subfield label stats. Why do different subfields contain similar/same numbers of voxels but their volumes are largerly different? Here is one subject's output:

sub-004 left CA1 22 1604.080 sub-004 left CA2 10 13.046 sub-004 left DG 18 989.380 sub-004 left CA3 9 57.983 sub-004 left misc 17 263.138 sub-004 left SUB 21 556.259 sub-004 left ERC 13 611.343 sub-004 left BA35 14 655.289 sub-004 left BA36 15 2907.791 sub-004 left PHC 12 1468.201 sub-004 left sulcus 23 758.514

And the readme file says: Columns: {side} {subfield name} {number of voxels containing the label} {volume in mm^2} I think here should be volume in mm^3?

Thanks!

thomshaw92 commented 1 year ago

Thanks for the feedback @ash077 I'll have a look through and get back to you soon.

Would you mind please sending me the output from your run?

I've updated the README - thanks for spotting the error!

thomshaw92 commented 1 year ago

Hi @ash077 I've dug into the original code for stats export: https://github.com/pyushkevich/ashs/blob/4595fdaefcbe49ed6d3c8e788f22dbb429c26d40/bin/ashs_extractstats_qsub.sh

Seems like we are actually exporting the 'extent along the Z-axis.'

I've updated the README to reflect this and apologise for the confusion! You can use ANTs or similar to get the voxel count of each label.