Improved Identification of Complex Temporal Systems with Dynamic Recurrent Neural Networks. Application to the Identification of Electromyography and Human Arm Trajectory Relationship Jean-Philippe Draye* t , Guy Cheron*, Marc Bourgeois", Davor Pavisic* and Gaötan Libert* *' Parallel Information Processing" Laboratory, FacuM Polytechnique de Möns (Belgium) # Laboratory of Biomechanics, University of Brussels (Belgium) *Senior Research Assistant of the Belgian National Fund for Scientific Research Corresponding author: Jean-Philippe Draye Facultä Polytechnique de Möns "Parallel Information Processing" Laboratory Rue de Houdain, 9 B-700 Möns (Belgium) Phone: +32-65-37.40.56 Fax: +32-65-37.45.00 E-mail: jpd@pip.fpms. ac. be ABSTRACT We propose a new approach based on dynamic recurrent neural networks (DRNN) to identify, in humans, the relationship between the muscle electromyographic (EMG) activities and the arm kinematics during the drawing of the figure eight using an extended arm. After learning, the DRNN simulations showed the efficiency of the model. We demonstrated its generalization ability to draw unlearned movements. We developed a test of its physiological plausibility by computing the error velocity vectors when small artificial lesions in the EMG signals were created. These lesion experiments demonstrated that the DRNN has identified the preferential 83 Brought to you by | University of Arizona Authenticated Download Date | 5/28/15 6:50 AM