How are the MNI coordinates for the Desikan-Killiany atlas derived?

1,025 views
Skip to first unread message

Brains McGee

unread,
Aug 8, 2019, 5:22:09 PM8/8/19
to brainGraph-help
Does the x-y-z coordinate of each ROI represent the center of the ROI? I'm trying to come up with an ordered list of the ROIs which follow a rostral to caudal gradient and am ordering it by the y-coordinate.

Chris Watson

unread,
Aug 9, 2019, 5:01:55 AM8/9/19
to brainGr...@googlegroups.com
I actually don't quite remember, as I created the tables several years
ago and was not very diligent about documenting everything,
unfortunately. My best guess was outlined in this thread:
https://groups.google.com/forum/#!topic/braingraph-help/CJbbSbtx5Co

I just checked again, with "fsaverage" labels. If you read in all of
the ".label" files, the average of the coordinates for each region are
reasonably close to the coordinates in the data in brainGraph.
There are larger differences, though.

You can also run "mri_surfcluster" on the label files to get the
centroids (see e.g.
https://mail.nmr.mgh.harvard.edu/pipermail//freesurfer/2016-August/047063.html
).
These values are *extremely* close to the means of the coordinates in
all label files.

I'm fairly certain that I "eyeballed" the coordinates for the
dk{,.scgm} and dkt{,.scgm} atlases. That is, I got the coordinates
from somewhere (I don't remember unfortunately), and then I checked
all of them manually on a MNI152 volume, and manually changed them to
what looked "best". This is of course not at all rigorous, but using
the exact locations has never been a priority for me, except to be
"close enough" for visualization. It depends on how important the
exact value is for your purposes.

The "destrieux{,.scgm}" and all other atlases (except "dosenbach160")
appear to be the "correct" coordinates. For the latter, I think I got
the coordinates from a brainGraph user a few years ago.

Chris
> --
> You received this message because you are subscribed to the Google Groups
> "brainGraph-help" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to brainGraph-he...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/brainGraph-help/82f03b7f-c333-47f9-9a70-a5682c2a699f%40googlegroups.com.
>
Reply all
Reply to author
Forward
0 new messages