Volume Currents in Forward and Inverse Magnetoencephalographic Simulations Using Realistic Head Models

Abstract

Volume currents are important for the accurate calculation of magnetoencephalographic (MEG) forward or inverse simulations in realistic head models. We verify the accuracy of our finite element method implementation for MEG simulations by comparing its results for spheres containing dipoles to those obtained from the analytic solution. We then use this finite element method to show that, in an inhomogeneous, nonspherical realistic head model, the magnetic field normal to the MEG detector due to volume currents often has a magnitude on the same order or greater than the magnitude of the normal component of the primary magnetic field from the dipole. We also demonstrate the disparity in forward solutions between a model that employs spheres, one that uses the realistic head and primary currents alone, and a realistic head model that incorporates both primary and volume currents. In forward and inverse MEG simulations using the inhomogeneous realistic model, the results obtained from calculations containing volume currents are more accurate than those derived without considering volume currents.

Solution of the Forward Problem in Magnetic-Field Tomography (MFT) Based on Magnetoencephalography (MEG)

“…The volume current 0018-9464/$25.00 © 2009 IEEE is attributed to the effect of the macroscopic electric field on charge carriers [4]. It has been shown that for a realistic brain model with inhomogeneous conductivity distribution, the magnetic field from this volume current can be comparable with that from the primary current source (e.g., dipole) [5]. Thus, the total current density becomes (5) where is the electric scalar potential.…”

Investigation of optimal parameters for finite element solution of the forward problem in magnetic field tomography based on magnetoencephalography

“…In case of Section 4 below, conductivity of the brain material plays a major role and, therefore has to be carefully considered [13]. Due to high complexity and anisotropy level, each element has to have its own material properties which leads to a complicated structure of matrix for material properties.…”

Influence of brain conductivity on magnetoencephalographic simulations in realistic head models

“…Indeed, dipoles that were "deep" in the head were generally more influenced by changes in conductivity values, whether they were changes in the gray matter or white matter, than were more superficially located dipoles. The smaller localization errors found with superficial dipoles would be expected based on equation (1), which implies that superficially located dipoles should produce magnetic fields with relatively low contributions from volume currents or influence from conductivity.…”

“…The finite element realistic head model was created from 256 volume magnetic resonance image (MRI) slices and consisted of 72,745 nodes, 406,493 tetrahedral elements, and 61 magnetic field detectors placed over the head [1]. The model consisted of five conductivity values: scalp (σ=1.0 S/m), skull (σ=0.05 S/m), cerebrospinal fluid (CSF) (σ=4.62 S/m), gray matter (σ=1.0 S/m), and white matter (σ=0.43 S/m) [2]; these conductivity values were considered the baseline conductivities.…”

“…With more realistic, inhomogeneous, nonspherical head models, however, volume currents become of critical importance in determining magnetic fields [1], and tissue conductivity values must be considered in calculations of volume currents. In the literature, gray and white matter conductivity values have been reported to be between 0.33 S/m -1.0 S/m and 0.31 S/m -0.48 S/m, respectively [2][3][4][5][6].…”