[FieldTrip] Source reconstruction on BEM mesh of EEG data

[Schoffelen, J.M. (Jan Mathijs)]

Dear Arno,

I think you are mixing up the sourcemodel and the volume conduction model. These are 2 separate things.

The BEM is a volume conduction model, and describes the geometries at which the potential distributions are computed that are the consequence of the volume currents, due to the primary currents in the brain.

The sourcemodel describes the set of positions at which an active dipole is assumed, and for which a ‘leadfield’ is computed (in the forward modelling step) / for which the source power is estimated (for the inverse modelling step).

The distributed sourcemodel can be defined as a 3D grid (suboptimal), or as a 2D surface. The latter can be obtained for instance using Freesurfer, as per the
'Construction of a source model based on a surface description of the cortical sheet'

section of

http://www.fieldtriptoolbox.org/tutorial/sourcemodel/

This piece of documentation describes how the ‘Subject01_sourcemodel_15684.mat’ has been generated.

I hope this helps.

Best wishes,
Jan-mathijs




On 15 May 2019, at 23:24, Arnaud Delorme <arno at ucsd.edu<mailto:arno at ucsd.edu>> wrote:

Dear all,

I am trying to perform source reconstruction on a BEM mesh of EEG data
The page below describes how to localize source on a BEM mesh for MEG data using minimal norm estimate on the BEM mesh itself. This is what I would like to do but for EEG data.

http://www.fieldtriptoolbox.org/tutorial/minimumnormestimate/

This page describes how to compute a leadfield matrix for EEG

http://www.fieldtriptoolbox.org/workshop/baci2017/forwardproblem/

However, the leadfield matrix created for EEG is for a 3-D grid, that can then be projected on a 3-D MRI volume - so not the projection on BEM mesh I would like. I have not been able so far to generate a leadfield matrix for a BEM surface. Now the tutorial page below in theory describes how to create one for MEG from a BEM mesh (section "Realistic single-shell model, using brain surface from segmented mri").

http://www.fieldtriptoolbox.org/example/make_leadfields_using_different_headmodels/

However, again I obtained a 3-D source model on a grid. I figured that the sourcemodel needs to look like this (the content below is for tutorial MEG file named Subject01_sourcemodel_15684.mat). However, I do not know how this file was generated and how it could be transposed for EEG (which needs 3 layers instead of 1 etc…).

sourcemodel =

  struct with fields:

                    pos: [15684×3 double]
                    tri: [31360×3 double]
                   sulc: [15684×1 double]
                   curv: [15684×1 double]
              thickness: [15684×1 double]
         brainstructure: [15684×1 double]
    brainstructurelabel: {2×1 cell}
                 inside: [15684×1 logical]

Any hint would be appreciated.
Cheers,

Arno
_______________________________________________
fieldtrip mailing list
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
https://doi.org/10.1371/journal.pcbi.1002202

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20190516/96e14364/attachment-0002.html>
-------------- next part --------------
_______________________________________________
fieldtrip mailing list
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
https://doi.org/10.1371/journal.pcbi.1002202