Map subcortical dlabel into right/left

25 views
Skip to first unread message

amirhussein.a

unread,
Jul 10, 2024, 9:09:31 AM (6 days ago) Jul 10
to HCP-Users
Hi,

I'm using a subcortical dlabel file, that has not divided the structures into right/left (i.e., only one Putamen label comprising the bilateral structures). Is there a way to generate a new dlabel file based on the existing one, with new label keys for the structures in each hemisphere (e.g., L_putamen and R_putamen)?

Best,
Amir

Tim Coalson

unread,
Jul 10, 2024, 3:07:18 PM (5 days ago) Jul 10
to hcp-...@humanconnectome.org
You need to give the voxels in the left versus right structures different values (and also values that aren't used by any other structure).  If you have a "dividing line" between left and right, you could represent that as a volumetric ROI, use it to modify the voxel values, and re-import the modified values as a label volume or dlabel file.

The freesurfer structure parcellation already separates left and right structures, is there a reason you don't want to use that as a starting point?

Tim


--
You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.
To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/14b9dc8b-cf63-400c-a116-defad5df912dn%40humanconnectome.org.

amirhussein.a

unread,
Jul 15, 2024, 5:41:28 AM (20 hours ago) Jul 15
to HCP-Users, tim.c...@gmail.com
Hi Tim,

Thanks for your reply. I'm using CIT168 CIFTI parcellation dlabels provided here: https://osf.io/am6vz
Our goal is to apply those dlabels on HCP's young adult rsfMRI dtseries data, and check connectivity profiles of the subthalamic nucleus. However, to test our hypothesis we have to have a division between right and left which that dlabel file has not applied.

Are there any right vs. left masks in the atlas space in HCP? I couldn't find any in MNINonLinear folder that I can use as a mask. Or possibly a dscalar file indicating which CIFTI-indices indicate a hemisphere for the subcortical structure.

Best,
Amir

Tim Coalson

unread,
Jul 15, 2024, 4:59:53 PM (9 hours ago) Jul 15
to amirhussein.a, HCP-Users
Based on the provenance, they used our Atlas_ROIs.2.nii.gz file to designate the subcortical voxels, so you can reuse that to separate right from left.  Take that file from here:


Use wb_command -volume-label-export-table on it, manually delete the pairs of lines that aren't part of one hemisphere (note, brain stem is a structure that we don't split into hemispheric parts), and use -volume-label-import with the -discard-others option to get a single-hemisphere version of it, then use -volume-math 'x > 0' to turn that into a binary ROI.

Tim

Reply all
Reply to author
Forward
0 new messages