Title :
An improved boundary element method for realistic volume-conductor modeling
Author :
Fuchs, Manfred ; Drenckhahn, Ralf ; Wischmann, Hans-Aloys ; Wagner, Michael
Author_Institution :
Philips Res. Lab., Hamburg, Germany
Abstract :
An improved boundary element method (BEM) with a virtual triangle refinement using the vertex normals, an optimized auto solid angle approximation, and a weighted isolated problem approach is presented. The performance of this new approach is compared to analytically solvable spherical shell models and highly refined reference BEM models for tangentially and radially oriented dipoles at different eccentricities. The lead fields of several electroencephalography (EEG) and magnetoencephalography (MEG) setups are analyzed by singular-value decompositions for realistically shaped volume-conductor models. Dipole mislocalizations due to simplified volume-conductor models are investigated for EEG and MEG examinations for points on a three dimensional (3-D) grid with 10-mm spacing inside the conductor and all principal dipole orientations. The applicability of the BEM in view of the computational effort is tested with a standard workstation. Finally, an application of the new method to epileptic spike data is studied and the results are compared to the spherical-shells approximation.
Keywords :
boundary-elements methods; electroencephalography; magnetoencephalography; medical signal processing; physiological models; analytically solvable spherical shell models; dipole mislocalizations; electroencephalography setups; epileptic spike data; highly refined reference boundary element method models; lead fields; magnetoencephalography setups; optimized auto solid angle approximation; principal dipole orientations; radially oriented dipoles; realistically shaped volume-conductor models; simplified volume-conductor models; singular-value decompositions; spherical-shells approximation; standard workstation; tangentially oriented dipoles; three dimensional grid; vertex normals; virtual triangle refinement; weighted isolated problem approach; Boundary element methods; Brain modeling; Conductors; Electroencephalography; Magnetic analysis; Magnetoencephalography; Optimization methods; Performance analysis; Solids; Testing; Body Fluid Compartments; Computer Simulation; Electric Conductivity; Electroencephalography; Epilepsy; Head; Humans; Image Processing, Computer-Assisted; Magnetic Resonance Imaging; Magnetoencephalography; Models, Biological; Surface Properties;
Journal_Title :
Biomedical Engineering, IEEE Transactions on