Source current directions

✉️ Mailing List Archive (read-only)
1

Hi all,

I am new to MEG and MNE. I have a question about the normal directions in a
surface source space. Are the normal directions at all source positions
aligned with the pial to white matter direction, i.e the pyramidal cell
axons in the cortex, or do they flip signs depending on the curvature?

If they all point to the white matter, I would expect that the recovered
source currents by MNE or dSPM to have the same signs of activation (all
positive or all negative). However, when trying the example

"Compute MNE-dSPM inverse solution on single epochs",
http://martinos.org/mne/stable/auto_examples/inverse/plot_compute_mne_inverse_epochs_in_label.html#compute-mne-dspm-inverse-solution-on-single-epochs

the currents in the auditory cortex split into a postive group and a
negative group.
Does it mean that some of the source currents are along the axons and some
are in reversed direction?

Thanks a lot.

Best
Ying
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20140409/e82168e8/attachment.html

2

Hi Ying,

The orientations follow the cortical surface. Thus, the activity from a dipole on one side of a sulcus or gyrus will show the opposite sign on the other side of the sulcus or gyrus (because the euclidean orientation of the dipole does not change, but its orientation relative to the cortical surface did).

Best wishes,
Avniel

3

Hi Avniel and other experts,

Thanks for the quick reply. So is there any method to transform the
forward matrix G in some way, so that all currents have the same sign? I
mean the cortical structure is known, we known whether two sources are on
different side of a gyrus or sulci.

Best,
Ying

Ying Yang

Graduate Student in Neural Computation,
Center for the Neural Basis of Cognition,
Carnegie Mellon University.
5000 Forbes Ave, Pittsburgh, PA, 15213

4

hi,

Thanks for the quick reply. So is there any method to transform the forward
matrix G in some way, so that all currents have the same sign?

not really. See however the sign flip trick that MNE uses to avoid cancelation
when summing currents in a region of interest.

http://martinos.org/mne/stable/auto_examples/inverse/plot_compute_mne_inverse_epochs_in_label.html
http://martinos.org/mne/stable/generated/mne.label_sign_flip.html#mne.label_sign_flip
http://martinos.org/mne/stable/generated/mne.extract_label_time_course.html

HTH
Alex