Three-dimensional Multi-dipole Model for Estimating Neuronal Generators Using Scalp EEG

One of the most important clinical uses of electroencephalography is detection of seizures in patients with epilepsy. In addition, some source localization methods have been proposed in order to estimate location of neuronal generators contributing to EEG spikes. These models are dependent on the geometry of the scalp (brain) as well as electromagnetic properties of the tissue. There has been quite extensive research including both realistic and approximate geometry inverse models for adults. In the neonatal patients due to a different structure of neurological pathways as well as presence of fontanel these models are not applicable. Several commercially available software solutions propose simple solutions by scaling the models. These solutions quite often do not work for inverse modelling as they may result in source localization that positions dipoles outside of the brain. To this purpose in this paper we develop a fully three-dimensional (3D) model using magnetic resonance image of a brain (scull).