Lower-extremity movement restoration

through muscle closed-loop FES control using natural sensor feedback

Proposer: Christine Azevedo, Laboratory of Computer Science, Robotics, and Microelectronics ( Laboratoire d'Informatique, de Robotique et de Microélectroniquede Montpellier) (liRMM), Université Montpellier II, France

Accepted: 2005. Final report

Abstract

The ultimate objective of the project is to explore the possibility of using natural muscle sensors as feedback for the control functional electrostimulation (FES) in spinal cord injured (SCI) persons. This prospective research project aims at bringing together robotics and neurophysiology in order to improve neuroprostheses control.

The approach used by DEMAR (Montpellier, France) consists of an implanted stimulator connected via electrodes to the nerves and/or muscles situated below the spinal cord lesion (thus not controllable by the central nervous system). One paraplegic patient has already been equipped and is able to stand and execute steps using a walker through an open-loop controller. DEMAR is now working on the development of a closed-loop controller to improve the efficiency and robustness of the system, compensating for muscle fatigue and perturbations. That implies the use of sensors which choice and number are highly constrained by psychological and cosmesis considerations. An alternative to artificial sensors is the use of natural sensors already present, intact and active in the muscles below the spinal cord lesion. Current studies carried out by the partner SMI (Aalborg, Denmark) are aimed at developing monitoring of biological signals for the information feedback of electrical stimulation. We would like to develop a collaboration between our two laboratories in order to extend and improve this work by using more advanced automatic control and robotics tools.

We will focus our efforts to investigate the monitoring and interpretation of nerve activity in terms of muscle length variation. Recordings will be realized during animal experiments: intrafascicular electrodes will be implanted in acute rabbit nerves and records of the electrophysiological activity in different experiments involving passive and active contraction of muscle will be made. Muscle force and length variation will be measured in parallel with artificial sensors. The data collected will then be subjected to advanced signal processing techniques in order to identify the correlation between muscle/joint variables and neural activity. This model will complete the muscle model developed in parallel by DEMAR and will be validated in simulation. The next step will be to realize a series of experiments in rabbits which muscle will be stimulated on the basis of a closed-loop control system using the estimation of muscle position/force in relation with the electroneurograph (ENG) to adjust the actual position relatively to the desired one. The industrial partner MXM will ensure the technical support and expertise in terms of electrode and implantable electronics prototypes.

Consortium

• Christine Azevedo, Laboratory of Computer Science, Robotics, and Microelectronics (Laboratoire d'Informatique, de Robotique et de Microélectroniquede Montpellier) (liRMM), Université Montpellier II, France (Project co-ordinator),
• Ken Yoshida, Center for Sensory-Motor Interaction (SMI), Department of Health Sciences and Technology, Aalborg University, Denmark.