The Numerical Methods Of Electromagnetic Inverse Problems In Biomedicine

Bioelectromagnetical computation, the combination of the computing electromagnetic theory and the biomedical science, has shown its great promising perspective in clinical applications and biomedical research these years. In this paper, the numerical methods for bioelectromagnetism are investigated. This paper consists of: (1) The development and application of bioelectromagnetic problems are reviewed. (2) The technique of biomagnetic measurement system is introduced, which including the development, the principle, the structure and the applications of the biomagnetic measure system. (3) The mechanism of bioelectromagnetic field, the source model of neural activity, and the theory of forward computation are analyzed, with concentration on the EEG forward computations. Studies include the establishment of the multiple spherical head model, the forward computation of a single point source and the lead field based on the linear superposition. (4) The numerical methods and the regularization methods for inverse problem of bioelectromagnetic field are studied. The basic algorithms for the inverse problem of bioelectromagnetic field are analyzed and the LORETA method based on voxel imaging is studied. The regularization methods are used for the localization of sources either in brain volume or on cortex, and the cortical-LORETA method is extended from plane model to arbitrary shaped surface. The promising localization results based on multiple spherical head model are obtained.