The melding of a brain-computer interface with a prosthetic device will let paralyzed people walk again if IEEE Member Miguel Nicolelis has his way.
Codirector of the Center for Neuroengineering at Duke University Medical Center, in Durham, N.C., Nicolelis and his team are developing a real-time interface together with a full-body exoskeleton to be controlled by signals from a paraplegic’s brain.
Nicolelis has assembled an international team of neurophysiologists, computer scientists, engineers, roboticists, neurologists, and neurosurgeons at laboratories around the world for his project. Their goal is to enable a paralyzed person to walk again by the end of 2012. The Walk Again Project, the first worldwide nonprofit brain-research initiative of its kind, includes partners in Europe, Latin America, and the United States.
“We’ve created a consortium that will manage all the scientific and clinical aspects of this work with the goal of making someone walk again,” Nicolelis says.
Since he briefed reporters about his work at a media event held in March to celebrate IEEE’s 125 anniversary, the team has developed a simulation of an entire exoskeleton that can be controlled by a brain-machine interface (BMI).
Nicolelis’s project began to show results four years ago, when he and his group implanted electrodes in a rhesus monkey’s brain that detected when it intended to move an arm. Later the animal was able control with only its thoughts the reaching and grasping movements performed by a robotic arm.
That was accomplished by first decoding brain signals from the monkey’s cerebral cortex while it held a joystick to move a shape in a video game. The signals were sent to the robot’s arm, which then mimicked the monkey's movements, leading it to control the game. The monkey eventually realized that just thinking about moving the shape was enough to move it, so it stopped holding the joystick and instead controlled the game purely through thought.
Three years later, another rhesus monkey was implanted with a new BMI in her brain’s motor and sensory cortex to control a computer-screen image of a human-like figure walking on a treadmill—part of the Computational Brain Project of the Japan Science and Technology Agency in Kyoto. Watching the robotic figure walking on the treadmill on the screen at the Duke laboratory, the monkey was rewarded for walking in sync with the robot. The treadmill was turned off after an hour, but the monkey was able to direct the robot to walk for another few minutes. The experiment indicated that part of the monkey’s brain had become dedicated to controlling the robot, as if the robot were an extension of the monkey.
That showed it was possible to establish functional links between the brain and a robotic, or computational, device—which could lead to the ability to control the movement of limbs.
Nicolelis says his newest BMI has merged efforts to control arms and legs into a single interface for the entire body that could be applied to controlling the exoskeleton. So far, however, the merge has been accomplished only on a computer. During the next few months, Nicolelis plans to test a virtual whole-body interface, merging what “we have done for arms and legs into a single device and see if we can get these monkeys to control a whole body avatar,” he says. He is collaborating with Gordon Cheng, a professor of cognition for technical systems at Technical University, Munich.
“We want to be sure the signals coming from the brain will be able to control this simulation,” Nicolelis says. If successful, he adds, the next step will be to build a real exoskeleton.
His exoskeleton will be fundamentally different from others being developed, he says. Some are focused on the robotics aspect, while others work more with the brain. His version will employ a direct interface between the brain and the exoskeleton, he says: “A lot of people are building exoskeletons, but they don’t rely on brain signals like we do.”
If the animal testing proves successful, he plans to determine whether paralyzed people can learn to reproduce the types of signals needed to control the exoskeleton. He says he expects to have the results of clinical trials on human beings in 12 to 18 months.
He believes his work will someday allow communication from human brain to human brain, he says, adding, “Our neuroprosthetic device may restore motor functions initially, but in the future it will be used to restore other functions.”