Each source is in which layer of the mesh?

Discussions
,
#1

Hi brainstorm team,
I used a five-layer mesh for generating a head model, then I computed sources for EEG signal. Now I want to know that each source is in which layer of the head model? (for example sources in CSF layer)

#2

All the sources should be in the gray matter...

@tmedani Any suggestion for the easiest way to check which sources are outside of the GM?

1 Like
#3

I thought that when we use a five-layer mesh for generating the head model, some sources will be in gray/white matter and some of them will be in CSF. Am I wrong?
another question that I have is that when we use depth weighting with minimum norm imaging as a method and current density map as a measure, are we calculating weighted minimum norm estimate(wMNE)? If so, then what is the bst_wmne function? Does it do the same?

#4

I thought that when we use a five-layer mesh for generating the head model, some sources will be in gray/white matter and some of them will be in CSF. Am I wrong?

You are talking about a FEM mesh to be used with Duneuro, right?
No, I think the sources should not be located outside of the gray matter.

when we use depth weighting with minimum norm imaging as a method and current density map as a measure, are we calculating weighted minimum norm estimate(wMNE)?

Yes.

what is the bst_wmne function?

The function bst_wmne.m disappeared years ago.

#5

Thank you, dear Francois.:pray::pray:

#6

Hello all

For better accuracy and stable results of the FEM all the dipoles should be within the grey matter, this is guaranteed by checking the checkBox "Force inside the GM".

The realistic location of the brain source is within the GM and should not be in the CSF.

#7

a simple way is to display the "cortex" used as source space with the FEM mesh, and keep only the WM/GM/ with the cortex, then you can play with the colors and the visibility of the surfaces.

But when you compute the head model, and if you check the box "force inside", all the outside points are moved inside the GM during the FEM computation.

#8

Hi again,
if all the dipoles should be just in the gray matter, So what are the dipoles which are in the white matter? I mean there are some dipoles in the brain volume that seems to be in white matter.

sourceGrid
sourceGrid1730×919 286 KB

1 Like
#9

@tmedani Have you tested DUNEuro with volume head models?

1 Like
#10

Yes, technically it's working but we need to be careful with the results, here are some tips:

  • If the head model is simple just with 3 layers (brain, skull, and scalp), in this case, it's similar to the BEM process and then the volume source can be used and the results are sable.
  • If there is a distinction between the WM and GM, then some instabilities may appear in the results for the dipoles located near the interfaces GM / WM.

In order to reduce this error, I recommend using a finer mesh resolution (FEM mesh ==> extract surfaces, then re-mesh with iso2mesh with a low value ~0.01 or less, ideally the size of the mesh elements should be smaller than the distance between the dipoles grid).
Also, you can test both methods "Venant" and "Partial Integration".

The other and simple solution is just to assign the same conductivity value for the WM and the GM.

Also, for the volume source, you should not to check the box "Force Inside the GM" and the "Restrict" Venant's option, otherwise, the dipoles will be moved to the nearest GM location.

All this is for the modeling side, but we need to keep in mind that putting dipoles within the white matter is not realistic, the brain sources are mainly within the GM.