Neuroscience: Opening the Mind’s Closed Circuits
Thought-controlled prosthetics, implanted in the motor cortex, are a source of new hope for paralysis patients.
Sam Kaplan for Proto
Minutes after a blood clot blocked an artery in Cathy Hutchinson’s brain stem, the 44-year-old single mother of two could no longer move her limbs. Neither could she speak. She could, however, think—and by moving her eyes, she asked Massachusetts General Hospital physicians to do everything in their power to help.
Not long ago, that power amounted to almost nothing, at least in terms of altering the fundamental parameters of Hutchinson’s new life. She was trapped in the prison of her body, fully aware but unable to move more than her eyes, communicating only through the direction of her gaze. Radiological scans showed a gap where the brain stem should have connected her otherwise functioning brain to her otherwise functioning body. Only an optimistic prognosis would suggest she might one day control a wheelchair by blowing through a straw. She was “locked in,” a phrase coined in 1966, around the time that a few visionary scientists started searching for a key.
Among them was Don Humphrey, a National Institutes of Health neuroscientist who, with others, showed it was possible to predict the force, direction and velocity of movement of a monkey’s limbs from recordings of electrical activity in its brain. Moreover, it wasn’t necessary to monitor each of the brain’s tens of billions of neurons, a task of unapproachable complexity, but rather predictions could be derived by tracking just a few. That raised the possibility of conveying neural messages from the motor cortex—the brain’s control center for physical movement, which remains active in locked-in people even though its signals have nowhere to go—to a wheelchair, computer, prosthetic limb or anything else that could be rigged to receive signals. “The motor cortex is quite good at controlling external devices,” says MGH neurologist Leigh Hochberg. “Those devices just usually happen to be arms and legs.”
Humphrey, three decades into his research when Hochberg arrived at his laboratory as a Ph.D. student in the early 1990s, was using an array of microscopic electrodes inserted into the motor cortex of monkeys as each animal played video games using a joystick to detect differences in the pattern of electrical activity. Hochberg and others continued on to decode this cortical ballet of movement in humans, converting its codas into code and designing machines to follow the instructions.
By 1996, when Hutchinson had her stroke, John Donoghue, a Brown University neuroscientist who had served on Hochberg’s doctoral thesis committee, was discussing with Hochberg how to translate into the human realm what had been learned in the monkey experiments. Six years later, Donoghue and his colleagues were recording and decoding activity in dozens of neurons simultaneously, translating their signals into machine-readable commands. Then they changed the monkeys’ video game playing in a remarkable way—by disconnecting the joysticks. The monkeys continued to move them but controlled the cursors with thoughts alone.
Donoghue co-founded a company called Cyberkinetics to develop BrainGate, a thought-controlled prosthetic. Hochberg, with support from MGH, then helped design a pilot trial for BrainGate. When Hutchinson learned of the trial, she—spelling out her words by locking eyes with Hochberg through translucent plastic printed with the alphabet—told him she wanted to join. On Nov. 30, 2005, doctors at Rhode Island Hospital implanted the device in her brain.
The BrainGate sensor consists of 100 platinum-tipped silicon microelectrodes sheathed in a polymer coating, inserted into the arm-controlling area of Hutchinson’s motor cortex. A cable plugs into the sensor, carrying neural signals from Hutchinson’s brain to a computer that decodes the patterns, then translates them into computer-controlling commands. To make a cursor move in a particular direction, Hutchinson imagines it moving; to click on a button, she imagines closing her hand. When asked recently if she had a message for other people with paralysis, she spelled out: “There is hope.”
For now, Hutchinson uses BrainGate only during twice-weekly sessions with Hochberg’s team, which took over the BrainGate trials after Cyberkinetics wound down. The investigational system isn’t yet ready for full-time use, though Hutchinson, in a research setting, can use it to write e-mail, type conversations and control iTunes. She spends much of the time calibrating new software by running a cursor across targets on a computer screen and reporting how easy or difficult each task is to perform. “She’s helping us understand the motor cortex even better, to see how it can work in somebody who hasn’t moved her arms and legs in over a decade,” says Hochberg. “She also provides feedback. Just because we see a cursor moving beautifully on a screen doesn’t mean it’s easy for her to do.”
One major challenge for BrainGate’s developers and users is day-to-day consistency. Each microelectrode must rest within a tenth of a millimeter of a target neuron, and an infinitesimal drift can alter a signal. Brain patterns also vary naturally from one day to the next. “If we’re recording on Monday, we’ll build a neural decoder that morning and use it that day,” says Hochberg. “If we come back on Tuesday, we might build a new neural decoder, because the signals are slightly different. An ideal system, in the future, wouldn’t require a caregiver to boot the computer and run a new version of the software.”
Hutchinson and her BrainGate device reached 1,000 days together in 2008, a proof-of-concept benchmark celebrated in a Journal of Neural Engineering article. She continues to use the system, though Hochberg is reluctant to describe its latest refinements before he publishes the work. For the current BrainGate trial, he hopes to recruit 10 or more people, including those who have spinal cord injuries, brain stem stroke, amyotrophic lateral sclerosis and muscular dystrophy. The ultimate goal is to build a wireless, implanted system that can be used every day, directly controlling computer communication systems and robotic limbs.
In the meantime, the technology might help people who are not locked in but who struggle to use prosthetics. “There have been some great prosthetic limbs developed during the past five years, but the controllers are still far too rudimentary,” Hochberg says. The difficulty of connecting prosthetics to damaged nerves and muscles means that “somebody using one to replace her right arm might need to wiggle her left shoulder to make it bend. The ideal controller is the original controller: the brain.”
For people with paralysis, it might even be possible someday to abandon prosthetics altogether, sending neurological signals—after a brief technological detour—back into the limbs for which they were intended. For someone like Hutchinson, the hope, Hochberg says, “is that one day she will simply reach out for a cup of coffee, pick it up with her own hand and drink.”