Optimal selection of interesting intracerebral epileptic signals by means of a multi-agents system.

The paper presents a distributed approach for the classification and selection of the interesting epileptic signals based on a dynamical vectorial analysis method. The objective is to limit the instantaneous workload by avoiding redundant computations and ensuring a better distribution of the load. Our approach deals with the information recorded during the intracerebral exploration and it exploits a dynamic selection of the interesting information to optimize processes without curtailing the information. We associated signal processing algorithms (spectrum analysis, causality measure between signals) approved in the analysis of the epileptic signal in a multi-agent system.