Editors Pick: Top Three Presentations at Workshop

by Warren Grill, senior technical editor

Direct Brain Interface for Control of a Virtual Robot
Dawn Taylor and Andy Schwartz, Arizona State University
The motor cortex contains neurons that are activated before and during arm movements, and, as a population, these neurons code the intended direction of arm movement. Over the past 10 years, several groups have recorded signals directly from the motor cortex, using arrays of penetrating microelectrodes, and using these signals to control a prosthesis, robot, or computer. Taylor, of ASU, reported on remarkable progress toward this goal. She demonstrated high-fidelity control of a computer-based virtual environment, using signals recorded from monkey motor cortex. Importantly, this control was achieved without the monkey moving its arm (i.e., the intent to move was used as the command signal). This finding is critical to the ability to use such signals in persons who are paralyzed, and thus unable to move their limbs. Further, Taylor demonstrated that the control improved over time and that the neurons actually change their tuning properties to better execute the virtual task.

Implantation of Electrode Arrays in the Visual Cortex of a Non-Human Primate
Phil Troyk, Illinois Institute of Technology
The NIH Neural Prosthesis Program was inaugurated to focus on the development of a visual prosthesis using electrical stimulation of the visual cortex. Thirty-two years later, the field of neural prostheses has undergone much change, but this still remains a mission. Continuing work that was undertaken by F. Terry Hambrecht and colleagues at the NIH, Phil Troyk and colleagues are taking important steps toward realizing this long-standing goal. In addition to development of the necessary implantable technology, Dr. Troyk has assembled an impressive team of top investigators from around the country, as well as a pre-clinical animal model to evaluate chronic implantation of arrays of penetrating microelectrodes in primate visual cortex. The team successfully implanted and maintained over 100 penetrating microelectrodes, and has demonstrated that stimulating through the electrodes produces visual sensations. This is an important step toward chronic implantation in a person with blindness.

Use of Noise to Improve the Function of Cochlear Implants
Jay Rubinstein, University of Iowa
In traditional engineering systems, noise degrades performance, and designers attempt to maximize the signal-to-noise ratio. However, recent work in biological systems has shown that sensory detection is actually improved in the presence of certain amounts of noise. This phenomenon, termed stochastic resonance, has found application in improving the performance of cochlear implants, which stimulate the auditory nerve to restore the hearing sense to deaf people. Rubinstein and colleagues, from the University of Iowa, demonstrated that adding noise to the stimulation pulse train applied to the auditory nerve lowered perceptual thresholds and increased the dynamic range between threshold stimuli and stimuli that are perceived as being the limit of comfortable loudness. Computer-based modeling suggests that this effect occurs because the noise de-synchronizes the activity in different fibers of the auditory nerve, making the pattern seen by the brain more like what occurs under acoustic stimulation. This finding holds promise to improve the function of cochlear implants, and demonstrates that engineers always have something to learn from nature.