Developing Novel Brain Robot Interfaces (BRI) to improve the Control of Neuro-Prosthetics

Thomas Martineau


The Advent of Neuro-technologies

Ever since Hodgkin and Huxley first described the firing dynamics of neurons in 1952, neuroscientists have been able to gradually map the inner circuitry of our brain and understand the complex modalities under which information is communicated and processed within neuronal layers. The digital revolution offered then an opportunity for both engineers and scientists to collaborate in the creation of new and exciting neuro-technologies which exploit, in order to function, our constantly expanding knowledge of human neuro-dynamics. This is the mission of the Imperial College Center for Neurotechnology through which this project is funded. 

What is a BMI?

One of the greatest challenges in neuro-technology today is to create efficient and practical direct Brain Machine Interfaces or (BMIs). A BMI records electrical signals within the brain. Those patterns of neuronal activity are broken-down and translated into useful volitional commands which reflects the intents of the BMI’s user. Put simply, a BMI user is capable to control a digital device using his mind. What would have sound as shear science-fiction non-later than a decade ago, is now possible: after the implantation of miniature electrode arrays in targeted motor regions of the brain, patients suffering from extreme paralysis are now able to control a robotic arm with a simple train of thought. There is hope that one day, those suffering from the most severe forms of locomotive impairment, such as multi-limb amputees or or tetraplegics, could use BRIs to fully control robotic prosthesis as natural extensions of their own body.

Achieving fine Motor Control using a BRI

Unlike any other interfaces developed for biomechatronics, BMIs have permitted the direct control of robotic manipulators with large numbers of degrees of freedom. Although models have been successively formulated and later implemented to extract high-order kinematics directly from the motor regions of the brain, the same cannot be said for the decoding of kinetic variables using a BMI. Humans are able to achieve dexterous tasks not only by guiding the position of their limbs in space but also by controlling haptic interaction forces and modulating muscle tone (or stiffness). Our objective is to create a novel class of invasive BMIs which will decode those particular kinetic signals from the central nervous system. This would notably enable BMIs users to control their force grip when manipulating objects. In the past, we have studied how those ideas could be apply to the control of exoskeletons and are now extenting this work to the development of BMIs for the control of robotic prosthesis.