electrodes2bem.md

title	tags	redirect_from
Align EEG electrode positions to BEM headmodel	example eeg mri source headmodel	/example/align_eeg_electrode_positions_to_bem_headmodel/

Align EEG electrode positions to BEM headmodel

Description

To use accurate individual volume conduction models of the head (or "head models") in inverse source reconstruction of EEG data, we construct a head model from the individual MRI. Furthermore, it is necessary to express the electrode positions in the same coordinate system and units as the MRI and head model.

This script demonstrates how to transform electrode positions to the individuals MRI space using the 3 fiducial points 'nasion', 'preauricular left' and 'preauricular right'. In the example below the position of these three anatomical marker locations is specified in the electrode structure as if they were normal electrodes.

Furthermore, it is necessary to provide the voxel coordinates of the same fiducial points in the MRI. Typically the MRI is aligned to MNI space, thus the resulting electrode positions are as well aligned to MNI space.

MATLAB script

% fit electrode coordinates to an individual MRI according to the same
% fiducials (nasion, left & right preauricular points) in both systems

% ensure that the electrode coordinates are in mm
elec = ft_convert_units(elec,'mm'); % should be the same unit as MRI

% these are expressed in the coordinate system of the electrode position capture device
Nas = elec.chanpos(strcmp(elec.label, 'nasion'),:);
Lpa = elec.chanpos(strcmp(elec.label, 'left'),:);
Rpa = elec.chanpos(strcmp(elec.label, 'right'),:);

% determine the same marker locations in voxel coordinates, e.g., [57,127,15])
% find fiducials e.g., by using ft_sourceplot(cfg, mri) which plots a figure in which
% you can interactively select slices of the mri
% You can also use ft_volumerealign with cfg.interactive='yes' and obtain the fiducials from the output.cfg.fiducial

vox_Nas = mri.nasion;  % fiducials saved in mri structure
vox_Lpa = mri.left;
vox_Rpa = mri.right;

vox2head = mri.transform; % transformation matrix of individual MRI

% transform voxel indices to MRI head coordinates
head_Nas          = warp_apply(vox2head, vox_Nas, 'homogenous'); % nasion
head_Lpa          = warp_apply(vox2head, vox_Lpa, 'homogenous'); % Left preauricular
head_Rpa          = warp_apply(vox2head, vox_Rpa, 'homogenous'); % Right preauricular

elec_mri.chanpos = [
  head_Nas
  head_Lpa
  head_Rpa
  ];
elec_mri.label = {'nasion', 'left', 'right'};
elec_mri.unit  = 'mm';

% coregister the electrodes to the MRI using fiducials
cfg = [];
cfg.method   = 'fiducial';
cfg.template = elec_mri;
cfg.elec     = elec;
cfg.fiducial = {'nasion', 'left', 'right'};
elec_new = ft_electroderealign(cfg);

% interactively coregister the electrodes to the BEM head model
% this is a visual check and refinement step
cfg = [];
cfg.method    = 'interactive';
cfg.elec      = elec;
cfg.headshape = vol.bnd(1);
elec_new = ft_electroderealign(cfg);