Identification Of Brain Areas Involved In Epileptic Activity By Means Of Multi-Agents System

This paper describe an experimental approach of vectorial analysis of the epileptics signals. based on it distributed method. The objective aim to identify and to localize the various intericial and ictal epileptiform events (pathological and/or normal) contained in the SEEG signals on a whole of 128 recording SEEG channels. This approach associates a multi agent system and signal processing algorithms. This association involves a double classification of the cerebral activity, one according to the segmentation and the other according to the activity. This classification is followed by a coloured representation of the seizure. BY taking into account the complexity by which SEEG signals are generated, we defined 3 great groups of activity: background activity (slow rhythms), pre ictal activity and critical activity fast rhythms). All original association Of these different elements (classification and representation) thus brings lighting on the settings concerned the seizure propagation and the association of the results provided by the synchronized (agents) give raise topology seizure analyzed