Spatial filtering to detect brain sources from EEG measurements

In clinical analysis, neural activity in the brain due to groups of neurons located in the head is recorded by means of EEG electrodes positioned on the scalp of the patient. The identification of the sources responsible for this brain activity is of great importance, especially if neurophysiological disorders are detected. The problem is that there is usually an unknown number of signals simultaneously recorded from the electrodes, each one from unknown direction and with unknown amplitudes, and the measurements are always corrupted by noise too. In this work a method of spatial filtering which combines DOA estimation and Beamforming techniques is presented: a DOA estimation method allows to estimate the directions of activation of the sources, while spatial filters of type Beamforming are designed to pass the electrical activity from a given direction and stop that originated from other directions. In this way the brain is explored to detect sources that have determined the recorded activity from EEG.