A method for rapid production of subject specific finite element meshes for electrical impedance tomography of the human head.

ABSTRACT

Finite element (FE) methods are widely used in electrical impedance tomography (EIT) to enable rapid image reconstruction of different tissues based on their electrical conductivity. For EIT of brain function, anatomically-accurate (head-shaped) FE meshes have been shown to improve the quality of the reconstructed images. Unfortunately, given the lack of a computational protocol to generate patient-specific meshes suitable for EIT, production of such meshes is currently ad hoc and therefore very time consuming. Here we describe a robust protocol for rapid generation of patient-specific FE meshes from MRI or CT scan data. Most of the mesh generation process is automated and uses freely available user-friendly software. Other necessary custom scripts are provided as supplementary online data and are fully documented. The patient scan data is segmented into four surfaces brain, cerebrospinal fluid, skull and scalp. The segmented surfaces are then triangulated and used to generate a global mesh of tetrahedral elements. The resulting meshes exhibit high quality when tested with different criteria and were validated in computational simulations. The proposed protocol provides a rapid and practicable method for generation of patient-specific FE meshes of the human head that are suitable for EIT. This method could eventually be extended to other body regions and might confer benefits with other imaging techniques such as optical tomography or EEG inverse source imaging.