Comparison Of Multi-layer Brain Model Of Ellipsoidal And Spherical Brain Model Eeg, Meg Problem

EEG (Electroencephalography) and MEG(Magnetoenphalography) are two important medical technologies ininvestigation of the physiological activities of the brain. They are mainlyused in epilepsy, syncope, mental illness and substance in brain lesions(such as cancer) and so on. The location of neural source in the brain (the inverse problem) requires a lot calculation of EEG forward problem and MEG forward problem (that is, the prediction of the potentials and the magnetic fields outside the head from the given current source).Most previous work of the EEG and MEG forward problem used single or multilayer spherical head to study the scalp potential and the magnetic fields outside head versus dipole position moving along the coordinate axis. After all, the human head is not spherical and the single spherical model is too unrealistic for the head due to the large difference between the conductivities of brain and skull. Some studies suggest that using uniform spherical model or multilayer spherical model to approximate the head will bring relatively large localization error of the neural current sources. Thus, a more realistic model to simulate the human head is quite necessary, such as the ellipsoid head model, the realistic shape head model and so on. However, the realistic head models based on MRI image require excessive computing time and heavy computation burden. Therefore, these kinds of models are not convenient for mathematically analyzing the effects of head shape and dipole source parameters on EEG/MEG forward problems. In a addition, the past work usually apply one kind of head model to test its numerical method. The results of the effects of different head shapes on the potentials and magnetic fields are limited. In this paper, we use multiplayer ellipsoid and spherical head models to study the distribution of scalp potentials and the external magnetic fields versus different dipole parameters. As far as numerical methods are concerned, direct boundary element methods with constant or linear basis functions are commonly used in previous work. Here, we use boundary element method based on weighted residuals with trigonometric basis functions. Our method has the advantages of the simple procedures, less work and convenient to implement.This paper is divided into four chapters: The first chapter introduces the EEG/MEG the history of the development and the present situations of the research. The second chapter describes the field equations of the forward problems and gives two different head models, the ellipsoid model and the spherical head model. In the third chapter, the formulas of solving the EEG/MEG forward problems are developed using BEM based on the Weighted Residual. In the fourth chapter, the numerical simulations are implemented in three-layer and two-layer head models for ellipsoid models and sphere models. We investigate the effect of the head shape on scalp potentials and the external magnetic fields and the differences of potentials and magnetic fields between three-layer head model and two-layer head model. And the comparisons of the performance of potential and magnetic field versus the different dipole parameters are made. The results show that the head shapes, the number of layer of head models and the dipole parameters affect the solutions of the EEG/MEG forward problems. The peak values of potentials and the locations of the peak values are obviously different in different head models. The head models and the dipole parameters also significantly influence the external magnetic fields. The comparison of RDM (relative difference measure) of potentials and magnetic fields in different head models show that our method is reliable and efficiency.