The center coordinate point of ROI of tables_CHARM of NMT_v2.

AFNI Message Board
1

Dear expert
I'm studying the homology relationship between human and macaque. At present, I need to find the coordinate points of tables_CHARM of NMTv2 in various ROI of CHARM3 and CHARM6. But I looked for the relevant literature and documents, but I couldn't find them. NMTv2 only gave CHARM the name. I would like to ask if the relevant coordinates are available.
Thank you very much.

2

Howdy-

OK. The AFNI extension part of the NIFTI header contains the ROI information in an "atlas points" format, with "center of gravity" (COG) listed for each ROI, in what I believe is RAI coordinate notation (right, anterior and inferior are all negative). For example, this is the top of the file:

3dinfo -atlas_points CHARM_3_in_NMT_v2.0_asym_fh.nii.gz
<atlas_point_list
 ni_form="ni_group">

<ATLAS_POINT
 data_type="atlas_point"
 STRUCT="ACC"
 VAL="3"
 LONGNAME="anterior_cingulate_cortex"
 OKEY="3"
 GYoAR="0"
 COG="2.64479 -31.4568 25.2082"
/>

<ATLAS_POINT
 data_type="atlas_point"
 STRUCT="MCC"
 VAL="11"
 LONGNAME="midcingulate_cortex"
 OKEY="11"
 GYoAR="0"
 COG="2.92232 -16.2771 28.6221"
/>

<ATLAS_POINT
 data_type="atlas_point"
 STRUCT="med_OFC"
 VAL="17"
 LONGNAME="medial_orbital_frontal_cortex"
 OKEY="17"
 GYoAR="0"
 COG="1.73304 -39.1361 19.7625"
/>

We can get those out programmatically using scripts and Linux commands, but I see that those regions are not left/right separated; that is, region 17, which has the string label "med_OFC" is not separated into "left-med_OFC" and "right-med_OFC". However, I think they might all be the coordinates of the left hemisphere half of the regions, because otherwise the midpoint of each symmetric or nearly symmetric region would be closer to zero for the x-coord.

So, my question is: do you want the region center-of-masses separately for left and right hemisphere?

--pt

3

These should be the "Internal Center" - Icent described in 3dCM's help:

      the location of the center of a voxel closest to the center of mass
      It will be the same or similar to a center of mass
      if the CM lies within the volume. It will lie necessarily
      on an edge voxel if the CMass lies outside the volume

The coordinates should be in RAI coordinates. If the regions are not separated by left and right, the centers are usually computed for one hemisphere to avoid having every center at x=0. For symmetric atlases, like the CHARM atlas for NMT symmetric atlases.

For the asymmetric versions of the template and atlas, you would have to separate the atlas by the left/right masks or by x=0 with 3dcalc and recompute the centers with 3dCM or with one of the methods described here:

4

Yes, thank you very much for your help. Now I have successfully calculated the center point of ROI I need. I have a new question to mask you. When I used NMTv2 as a template to register my fmri data, and after preprocessing, I resampled the fmri data of macaques to 1.3751.3751.375, but I encountered a problem. The CHARM_6_in_NMT_v2.0_sym_05mm of NMT is 0.50.50.5. I tried some methods. But it seems that the effect is not good, so, what is a feasible method to resample the mask of CHARM_6 to 1.375mm, and the result will not be greatly missing? Thank you very much for your help.

5

There is no perfect solution for this difference in resolutions. 3dresample is the method to resample from one grid to another, but there is an inherent loss of information from both the coarseness, the difference in grid locations and in the process of resampling. Using the larger ROIs from the coarser resolutions of the CHARM atlas is one of the main purposes of that atlas, so consider using something other than the CHARM_6, the finest resolution. See the information from one of our class presentations (slide 17) for more.

https://afni.nimh.nih.gov/pub/dist/edu/latest/afni11_roi/afni11_roi.pdf