MIMO Systems and Applications to Brain Computer Interface by Using EEG

Brain Computer Interface (BCI) is, as proposed by Vidal in 1973, a multitechnology discipline able to use signals generated by the brain to dialogue with intelligent devices in order to support the person in controlling external apparatus as e.g.: prosthetic devices. Also, it could permit to have a better knowledge of neurophysiological phenomena that govern the production and the control of observable neuroelectric events. A non invasive way to capture brain signals is to measure the electric voltage between one electrode placed on the scalp and a reference point, by the so called Electroencephalogram (EEG). The signals we are revealing are sustained by fluctuation of either electrical potential or current generated by neurons. Looking at EEG we note that when the person is in a relaxed state there is an almost periodic signal due to the spontaneous brain activities. If the person is subject to sensory messages, such as sounds, light, etc., this periodic signal almost disappear and is substituted by a short waveforms that are evoked from the sensory messages. In order to detect these signals, which have a duration of 0,5-2 seconds, we need to know other information like: the area of brain where they have been generated; the standard parametered signal shapes, if any, the typical signal bandwidth, etc. Many papers treated in the past these topics but they do not reach a fully complete result. For example many papers try to get the position inside the brain of the point where the evoked potential has been generated. In order to solve the, so called, inverse problem, first the electric field on the skull, generated by either a dipole or a current source in a known position within the brain, has been computed. In doing this computation the brain has been considered as a homogeneous material with known and linear electric characteristics. In the analysis two discontinuities made by two layers of homogeneous materials have been considered. The first one represents the cerebral cortex, which appears as a good conductor compared to the brain and, the second is the skull which appears as a good insulator compared to the brain. The result is that the electric field measured on the scalp is deterministically related to the position and the intensity of the source. To solve the inverse problem the knowledge of the electric field all over the scalp is required. Also, we remark that to the received signals generated by the brain overlap noise and interference. To solve the inverse problem is a challenging engineering task, but, is this the problem we want to solve? Specific points of the brain devoted to perform specific tasks depend on several factors; most of them are still unknown. What is known is that a given task is done in a volume of the brain, while the specific point where it is done depends on the organization of the brain of the person considered. Then, in many cases we are interested to the know area where the signal originated not to the specific point. In the following we limit our analysis to EEG signals and, moving from the study done we introduce new ideas in order to apply to EEG the results of MIMO and radar technologies for a further improvement in the analysis of EEG signals.