With the right interface, our brains can adopt all sorts of prosthetics as our own – letting us escape our physical constraints
ANSWER quickly: what links the internet, the stock market, democratic elections, a perfect soccer play, the big bang theory, the frescoes of the Sistine Chapel and the iPad? Most people guess that the only possible link is they are all created by humans. While this is technically correct, it doesn’t credit the true creator of such macro structures and exquisite tools: the human brain.
As well as the almost infinite catalogue of artificial tools and beliefs that rule most of our lives, our cherished social, political, and economic systems also blossom as by-products of the incessant electrochemical storms brewed by the brain circuits formed by billions of interconnected cellular elements. These neurons make up an organic structure so majestic and mysterious that its only true rival in complexity and power is the cosmos that hosts us all.
Advertisement
For the past 200 years or so, neuroscientists have been obsessed with understanding how the roots of all our glory and disgrace, as individuals and as a species, emerge from waves of neuronal electrical activity that propagate through a neural ocean. Just how do they morph into what is conventionally known as thinking, the main currency of our primate brains?
In the early 19th century, Franz Joseph Gall in Germany and Thomas Young in Britain pioneered the modern age of neuroscience with opposing theories of how the brain worked. Gall’s phrenology proposed that brain functions were localised in particular spatial territories of the human cortex, the most superficial part of the nervous system, just beneath the skull. Gall and his disciples made a living by claiming to ascertain the key personality traits of his patients by palpating the bumps on their heads.
In 1802, on the other side of the English Channel, the polymath Young unveiled his trichromatic theory of colour vision, just a year after carrying out what physicist Richard Feynman considered the inaugural event of quantum mechanics, the double-slit experiment. Young’s new theory proposed that information about any type of light wavelength, that is, different colours, could be represented by the combined action of only three colour receptors in the human eye.
Young reasoned that each of these receptors would not only respond to a particular colour but could also indicate the presence of a broad range of other colours, through responses of different strength. Hence, the whole visible spectrum could be represented by the combined response of the three receptors.
More than a century later, the trichromatic theory, derived without even examining the anatomy or physiology of a human eye, was proved right by the discovery of the three classes of retinal cones. So Young can be seen as the founder of the “distributed” view of brain function, whereby the widespread populations of interconnected neurons, spread across many locations, are responsible for any function or behaviour.
To be a disciple of Young was and is to reject Gall’s phrenology and all those who have taken advantage of sophisticated imaging technologies over the past century to defend the “localisationist” theory. In this view, brain functions correlate strongly with specific spatial locations of the cortex. Young’s ideas also challenge another central dogma of 20th-century neuroscience: that an individual brain cell is the functional unit of the central nervous system, the one capable of generating CNS functions and behaviours.
About 12 years ago, my colleagues and I at Duke University, North Carolina, devised what was then considered an unorthodox experimental paradigm to test whether Young’s assumption also applied to motor behaviours. We called this the brain-machine interface (BMI) paradigm. It involves sampling the simultaneous electrical activity produced by hundreds of individual neurons located in different locations of the brain, while animals perform a well-rehearsed motor task. By feeding the neuronal activity into computer models running in real time, we can extract basic motor commands conveyed by these neuronal electrical storms.
Once extracted and transformed into digital signals, these commands can be transmitted to artificial robotic, electronic or virtual tools capable of mimicking, in real time, the voluntary motor intentions generated in our subjects’ brains. Using BMI, we have shown that just by thinking, monkeys can learn to move robotic arms and legs using their brain activity directly, without any movement of their bodies.
By sending visual, tactile or direct electrical feedback signals from the artificial actuators back to the subject’s brain, we have also shown that the signals can be assimilated as true extensions of the subject’s sense of self. In other words, the tools that are controlled directly by brain activity in these experiments seamlessly become extensions of the internal brain model.
One of our most beloved subjects, an owl monkey named Belle, was the first primate to use her brain to control an artificial device. By controlling the movements of a robotic arm placed a few metres from her own body, just by thinking, Belle showed that the brain can be liberated from the physical constraints of the body and its limits on our daily actions with the world that surrounds us.
A few years later, we took this concept a step further with Belle’s rhesus monkey “cousin”, Idoya. We used raw cortical electrical activity, recorded while she walked on a treadmill in our lab in North Carolina, to control the walking behaviour of CB-1 in real time. CB-1 is one of the world’s most sophisticated bipedal robots, based at the labs of one of the leaders in this field, ATR Neural Information Analysis, in Kyoto, Japan. This BMI turned out to work a little faster than the original biological control circuit in Idoya’s brain that she uses to move her own legs – even though her cortical neurons were controlling a much bigger and stronger body than her own, thousands of miles across the planet. As in Belle’s case, Idoya continued to control CB-1’s legs even when her own legs stopped moving at Duke. Just by imagining the leg movements, Idoya kept CB-1 walking.
“Even by imagining the leg movements, monkey Idoya kept robot CB-1 walking”
Recently, we have moved further on still, demonstrating that the monkeys can use brain activity to control a virtual avatar arm or hand that explores the texture of what appear to be visually identical objects in a virtual world. As the avatar hand explores those surfaces, an electrical signal describing the texture is generated and transmitted through arrays of microfilaments. These are already used to record electrical brain signals to the brain region that processes tactile information generated by the monkey’s “real” body.
In closing this loop between a brain and a virtual body using BMI technology, we now know the primate brain can operate beyond the boundaries and physical constraints of its body and interact with any world presented to it. No single neuron could make BMI work properly, and only populations of neurons, distributed across multiple cortical areas, could achieve the best results in terms of the motor control of artificial devices. Hence, we think we’ve generated yet another experimental validation for Young’s insights.
Besides providing further ammunition for the distributionist camp, BMIs have opened up many other areas of scientific inquiry. For starters, there is renewed hope of restoring mobility to people who are paralysed. Indeed, a non-profit research consortium, the Walk Again Project, has been formed by universities and research institutes in the US, Europe, and Brazil. It has a three to five year goal of helping a person who is quadriplegic due to a spinal cord injury regain full body mobility with a BMI. The plan is to use the individual’s cortical motor activity to control the movements of a whole-body robotic vest.
This “robotic vest”, an exoskeleton, will become the patient’s “body”, carrying them through the world controlled by thought alone. Moreover, sensors spread all over the vest’s surface will provide the kind of sensory feedback to the brain that the person’s biological body used to do. As with Belle and Idoya, we fully expect the patient’s brain to seamlessly incorporate the exoskeleton into its internal neural model of self.
Deeper into the future, BMIs may become the conduit through which our brains control all our tools, to extend our reach, presence and communication with the universe. I do hope that my descendants have the opportunity to feel what it is to be climbing hills covered with fine red Martian sand while actually sitting on their favourite beach in northern Brazil. By then, they all may wonder, in astonishment and mild disbelief, how we ever lived without the ability to do things just by thinking.
Profile
Miguel Nicolelis is Anne W. Deane Professor of Neuroscience, and founder of the Center for Neuroengineering, Duke University, North Carolina. Pope Benedict XVI appointed him an ordinary member of the Pontifical Academy of Sciences. This essay is based on his latest book, Beyond Boundaries (Times Books)
