Source analysis

Abstract

The scalp electroencephalogram (EEG) records electrical activity generated by the ensemble of a great number of pyramidal neurons within the brain. Sampling of electromagnetic brain signals in milliseconds has already been achieved. Unfortunately, the spatial resolution of EEG is very poor, which is limited by the relatively small number of spatial measurements (only a few hundred in EEG) and the inherent ambiguity of the underlying static electromagnetic inverse problem. In fact, localizing these potentials from the scalp EEG within the brain is an ill-posed inverse problem. This chapter firstly provides a brief introduction to the topic with an overview of the inverse problem from a signal processing perspective. In the next two sections, we describe the source models and head models of EEG source analysis, followed by various approaches to the inverse problem with which the properties of the neural current generators are estimated from the data. Finally, we discuss about the recent developments and the emerging signal processing issues of EEG data analysis.