Finite Element Modeling of a Realistic Head Based on Medical Images

Accurate modeling of head volume conductor plays an important role in EEG/MEG forward and inverse problems. The finite element (FE) modeling of realistic head is a key issue for the finite element analysis of brain electromagnetic field. Various FE modeling methods have been investigated and developed recently to model the realistic head volume conductor, but these methods either use approximation for the tissues, which causes large geometry distortions, or need substantial manual interventions, which makes the modeling process time consuming. In the present study, we have developed a new method to generate subject specific FE head models based on their magnetic resonance (MR) and computer tomographic (CT) image data. The present approach consists of three parts: segmentation of MR and CT images, co-registration of MR and CT images, and mesh generation for all tissue volumes. Numerical experiments using MRI/CT data in two human subjects demonstrate the ability and feasibility of the present FE modeling method to automatically construct realistic head FE models including the scalp, skull, cerebrospinal fluid, gray matter and white matter tissues.