Supervised Neuronal Approaches for EEG Signal Classification: Experimental Studies

A. Frédéric (France), K. Nizar (France, Tunisia), B.K. Khaled, B.M. Hedi (Tunisia), B. Laurent (France), and D. Mohamed (Tunisia)


Neural networks, selforganizing maps, multilayer perceptrons, electroencephalographic signal interpretation, medical application


Using artificial neural networks for Electroencephalogram (EEG) signal interpretation is a very challenging tasks for several reasons. The first class of reasons refers to the nature of data. Such signals are complex and difficult to process. The second class of reasons refers to the nature of underlying knowledge. Expertise is manifold and difficult to formalize and to be made compatible with a numerical processing. In previous studies we have deeply described that expertise and explained, from theoretical and bibliographical studies, why artificial neural networks could be interesting candidates to perform such a signal interpretation. In this paper, we report recent experiments that we have made on real EEG data in a classification framework. These results are interesting with regard to the state of the art. They also indicate that further work must be done on expertise integration in our neuronal platform.

Important Links:

Go Back