Hi Chris,
I am working on a project where we would like to incorporate
anatomical
information into our results. We essentially need to be able to
load the
ROI's produced with mne_add_patch_info and the Deskian ROIs in
the same coordinate system.We sourced MEG data to agnostically assigned ROIs (labels; using
mne_add_patch_info producing *-l/rh.label) and would like to find a
relatively easy and automated way to figure out to which anatomical
regions these ROIs/labels correspond. We thought we could find some
coordinates in the .label files that might correspond to the
anatomical
Desikan ROIs automatically produced by freesurfer. The
agnostic .label files
seem to have coordinates in them. However the anatomical Desikan
ROIs do
not seem to have any coordinates in them, yet we are able to view
them in
mne_analyze. Does anybody know where those coordinates come from
so that I can programmatically attribute the agnostic regions to the
anatomical Desikan regions they are located in? Has anybody any other
ideas how to easily and automatically figure out to which anatomical
region the .label files correspond? We have 162 labels per hemisphere
and a number of subjects so looking them up through mne_analyze
would be
quite time consuming.
You do not, actually, need coordinates to do what you describe. You
can work entirely with vertex numbers. However, you need to do a
little bit of Matlab coding to achieve the result. Here is how:
1. Convert the Desikan parcellation to labels using mne_annot2labels
2. The newest MNE version 2.6.1 contains mne_read_label_file which you
can use to read label files.
3. Load all the Desikan parcellation labels.
4. Using the 'intersect' Matlab function determine which Desikan
parcellation labels has the biggest number of vertices common to your
agnostic label of interest.
I hope this helps.
- Matti