Deuxième conférence plénière française de Neurosciences Computationnelles, "Neurocomp08"

8-11 octobre 2008

Marseille, France - http://2008.neurocomp.fr

Neural decoding for Brain-Machine Interfaces

Saturday, October 11th, 9h00 - 13h00 , Faculté de Médecine (Timone)

Paralysis of the upper limb is a frequent consequence of lesions in the nervous system. Recent advances in neurophysiology and multi-electrode recording techniques suggest that Brain-Machine Interfaces (BMIs) are a promising therapeutic option for the restoration of motor functions. The basic principle of BMIs is to decode the real-time modulation of brain activity for the control of an external actuator. Such technique can possibly be used to bypass lesioned motor structures and to allow paralyzed patients to act within their environment. Non-invasive BMI are based on electroencephalographic (EEG) recordings and have been used in patients for the 2D control of a cursor on a computer screen or for the control of other simple devices. The non-invasive techniques are limited by the low spatial as well as temporal resolution of the recorded signals due to overlapping activity in multiple areas, and high susceptibility to eye movement artifacts. Alternatively, invasive BMIs are based on intra-cortical recordings of multiple single neurons, of multi-unit activity or of local field potentials (LFP). Most of the invasive BMIs rely on the on-line extraction of control signals from motor cortical areas for the control of a robotic device by a non-human primate. In order to decode the cortical signals into an efficient artificial motor command, different models and mappings between input and output signals have been and can be considered.

The goal of this half-day workshop is to explore the most recent techniques available for deriving movement-control signals from invasive or non-invasive cortical recordings. During the first part of the workshop, three invited speakers (Mehring C, Scherberger H and Schwartz AB) will present their most recent research on BMI, focusing their presentation on algorithms used for feature extraction from the electrophysiological data. In the second part of the workshop, all participants will be invited to join an open round-table discussion to explore the contribution of computational neuroscience to progress in BMI development, to address critical issues and to outline future perspectives.