Multiple coordinates per cluster for cluster report

AFNI Message Board
1

Hello AFNI team,

Is there a way to output more than one coordinate per cluster in the 3dClusterize cluster report?

I’m using the whereami function to locate atlas regions on a group level cluster map. For each cluster, I use the center of mass from the cluster report as well as manually entering other coordinates into the -coord_file flag that seem to cover the extent of the cluster. Instead of manually clicking around the cluster to find additional coordinates, I was hoping there was a way to output this information per voxel.

2

Hi-

I would agree that a single coordinate is often not enough for really summarizing a cluster accurately. We make the case here:
https://www.biorxiv.org/content/10.1101/2022.10.26.513929v2
… that the whereami functionality of looking for overlaps with an ROI atlas is a much more informative way to go. See in particular the discussion starting on page 17 of that draft (“One additional aspect is worth noting…”), comparing the relative information content of Tables 1 and 2.

For example, to get the list of ROIs that overlap with each region in the Clust_mask+tlrc dataset in the Glasser MNI atlas, you could use:


whereami                                          \
            -omask Clust_mask+tlrc.HEAD                \
            -atlas MNI_Glasser_HCP_v1.0

You could also do that ROI-by-ROI, e.g., for the ROI with integer value 4 in your cluster mask:


whereami                                          \
            -omask Clust_mask+tlrc.HEAD"<4>"                \
            -atlas MNI_Glasser_HCP_v1.0

–pt

3

Thank you Paul!

I tried it with the Glasser atlas and was able to get a list of the overlapping anatomical regions for each ROI. Is there a way to do the same thing for TT_Daemon? When trying the command it didn’t list any anatomical regions.

whereami -omask ordered_cluster_map_p001_alpha05000_1+tlrc.HEAD"<1>" -atlas TT_Daemon
++ Input coordinates orientation set by default rules to RAI
++ In ordered mode …
++ Have 2 unique values of:
0 1
++ Skipping unique value of 0
++ ========================================================================
++ Processing unique value of 1
++ 4143 voxels in ROI

4

Hmm, OK. What is the output of each of these commands:


# A
afni -ver

# B
@FindAfniDsetPath TTatlas+tlrc

# C
afni -no_detach -q -VAFNI_GLOBAL_SESSION=

# D
afni -no_detach -q -VAFNI_ATLAS_PATH=

?

thanks,
pt

5

Hi PT,

I am working with Catherine and can provide you with the information you requested to expedite this query.

Here is the output of each command:

A

afni -ver
Precompiled binary linux_ubuntu_16_64: Mar 24 2021 (Version AFNI_21.0.21 ‘Titus’)

B

@FindAfniDsetPath TTatlas+tlrc
/usr/local/afni/bin

C

afni -no_detach -q -VAFNI_GLOBAL_SESSION=
this command did not actually do anything, it just returned the command line

D

afni -no_detach -q -VAFNI_ATLAS_PATH=
this command did not actually do anything, it just returned the command line

Regards,
Christine Smith

6

Hi, Christine-

I think if you put the following into ~/.afnirc, setting the environment variable:


AFNI_ATLAS_PATH =  /usr/local/afni/bin

… that should do it. I tend to have both that and AFNI_GLOBAL_SESSION set to the location of my main directory of reference atlases and templates; in this case, that would mean also setting:


AFNI_GLOBAL_SESSION =  /usr/local/afni/bin

there.

Can you please let me/us know how that goes?

–pt

7

Hello,

Here is the contents of our .afnirc. We are still getting the same output as the initial post where there are no anatomical regions listed for TT_Daemon but using MNI Glasser works.

***ENVIRONMENT
AFNI_ATLAS_PATH = /usr/local/afni/bin
AFNI_GLOBAL_SESSION = /usr/local/afni/bin

We also tried running these commands again.

C

afni -no_detach -q -VAFNI_GLOBAL_SESSION=
now we see /usr/local/afni/bin

D

afni -no_detach -q -VAFNI_ATLAS_PATH=
now we see /usr/local/afni/bin