Processing EEG signals through Beamforming techniques for seizure diagnosis

Epilepsy is one of the most common neurological disorders. It is very difficult the identification of seizures because they are infrequent events to be detected and an electroencephalogram (EEG) does not always show abnormalities due to interferences and noise. In this work, methods based on principles of spatial filtering are developed to allow a more accurate diagnosis. For this aim techniques based on Beamforming are taken into account for their ability to constrain the passing of the electrical activity from a specified location attenuating the activity to others. The problem is set as a waveform estimation, so information about amplitude and frequency of the signal are extracted from measured EEG signal. The reconstruction of the desired signal, that is the possible signal originated from epileptogenic focus, will highlight abnormalities or not in the electrical activity of the brain, and will help to have accurate diagnosis. For this scope, in this work, EEG data by scalp are simulated, in a preliminary way, through the solution of a forward problem related to geometry of the head and to the definition of a model for the envisaged source. A simple geometry is considered and, simulations and results are presented, showing the performance of several algorithms. Following the methods have been applied on real data, In this case it is necessary to know the position of the source for to reconstructing of the signal of interest. As per alternative, to estimate the position of the sources, methods of a DOA estimation have been studied. In a preliminary way this problem is faced with consideration on the resolution issue of the sources when they are very close together.