Discrepancy between visualization of freesurfer label and its vertices

✉️ Mailing List Archive (read-only)
1

Dear MNE users,

I observed a discrepancy between

(1) the visualization of Freesurfer generated anatomical labels (*.label
files) which can be loaded in mne_analyze with Labels --> Load Label

AND

(2) the actual vertices that the label contains. I visualize these
vertices by reading in the label file with mne_read_label_file() in MATLAB
and writing out an *.stc file with all vertices common between the label
file and the vertices in the source space set to '1' and all the other
source space vertices set to '0'.

To be specific, I'm having trouble with the 'S_central-lh.label' file
generated by Freesurfer 4.1.0. The visualizations (1) and (2) should look
the same, but I notice that (2) also contains a couple of vertices on the
medial wall.

Has anyone else observed this before? How am I mistaken? Thanks for your
comments.

Kind regards,
Pavan

2

Dear Pavan and others,

I have observed the same or similar problem. The the faces information sorted in the surf directory doesn't correspond to the either the .label or .stc files. This results in the wrong topographical stats and adjacency information been computed.

I am not sure if it is just a visualisation error or a problem with the adjacency/faces information from MNE/freesurfer.

Regards,

Li Su

3

Hi Pavan,
  I suspects this is as a result of the freesurfer indexing being zero
based and MNE's being 1 based. If that were indeed the case, if you add
1 to the vertex numbers read by mne_read_label_file() before finding the
common vertices that would resolve it..

Regards,
Hari

4

Thanks Hari! Your hunch is spot on and potentially saved me weeks.

5

Hi Pavan,