Brain-computer link lets paralyzed patients convert thoughts into actions

Brain-computer link lets paralyzed patients convert thoughts into actions

A multi-institutional team of researchers has found that people with long-standing, severe paralysis can generate signals in the area of the brain responsible for voluntary movement and these signals can be detected, recorded, routed out of the brain to a computer and converted into actions--enabling a paralyzed patient to perform basic tasks.

In the 13 July 2006 issue of Nature, the researchers present the first published results from the initial participants in a clinical trial of the BrainGate Neural Interface System, a "neuromotor prosthesis" developed by Cyberkinetics Neurotechnology Systems, Inc., of Foxborough, Mass.

The first patient, Matthew Nagle, a 25-year-old Massachusetts man with a severe spinal cord injury, has been paralyzed from the neck down since 2001. After having the BrainGate sensor implanted on the surface of his brain at Rhode Island Hospital in June 2004, he learned to control a computer cursor simply by thinking about moving it.

During 57 sessions, from July 2004 to April 2005, at New England Sinai Hospital and Rehabilitation Center, Nagle learned to open simulated e-mail, draw circular shapes using a paint program on the computer and play a simple video game, "neural Pong," using only his thoughts. He could change the channel and adjust the volume on a television, even while conversing. He was ultimately able to open and close the fingers of a prosthetic hand and use a robotic limb to grasp and move objects. Despite a decline in neural signals after 6.5 months, Nagle remained an active participant in the trial and continued to aid the clinical team in producing valuable feedback concerning the BrainGate technology.

The second patient, a 55-year-old man with a similar injury, had the sensor implanted by surgeons at the University of Chicago in April 2005 and was followed by researchers from the Rehabilitation Institute of Chicago and Cyberkinetics. Although his device initially had electrical problems, these were repaired and he was able to learn to control the cursor from months seven through 10 of the trial, until a technical issue caused signal loss at most electrodes after 11 months.

"The results," said senior author of the paper, John Donoghue, professor and director of the brain science program at Brown University and chief scientific officer of Cyberkinetics, "hold promise to one day be able to activate limb muscles with these brain signals, effectively restoring brain-to-muscle control via a physical nervous system."

"Our researchers initiated the clinical trial with the hope of being able to develop a non-obtrusive system that would one day provide more freedom to those with severe paralysis," said Timothy Surgenor, president and CEO for Cyberkinetics. "We are eager to expand on this initial proof-of-concept toward one day providing improved independence and overall quality of life."

The BrainGate System consists of a 4x4 millimeter sensor, about the size of a baby aspirin, with 100 tiny electrodes, each thinner than a human hair. The sensor is implanted on the surface of the area of the brain responsible for voluntary movement, the motor cortex. The electrodes penetrate about 1 mm into the surface of the brain where they pick up electrical signals--known as neural spiking, the language of the brain--from nearby neurons and transmit them through thin gold wires to a titanium pedestal that protrudes about an inch above the patient's scalp. An external cable connects the pedestal to computers, signal processors and monitors.

Converting digitized intentions into meaningful action, however, is not simple. Active neurons fire between 20 and 200 times a second and they work in teams.

Although scientists have long been able to eavesdrop on individual nerve cells, before 1996 no one had developed a reliable system for directly collecting precise data from large groups of brain cells. That year, Donohue and his post-doctoral student Nicholas Hatsopoulos modified an existing sensor and used it for the first time to record signals from multiple brain cells in monkeys.

At the time, Donohue assigned Hatsopoulos--now an assistant professor of organismal biology and anatomy at the University of Chicago--the task of creating algorithms to translate the chatter between neurons in the motor cortex into a language the computer could understand and use to control other devices. Hatsopoulos and other students in the Donoghue lab were slowly able to match neuronal signal patterns with specific arm movements. In 2002, he, Donohue and colleagues at Brown showed that monkeys could learn to control the cursor without moving a muscle.

Meanwhile, in order to move into human trials, Donohue, Hatsopoulos, Gerhard Friehs and Mijail Serruya, all then at Brown, formed Cyberkinetics, which was incorporated in 2001. In 2002, they merged with Bionics, makers of the sensor, raised $5 million, and applied to the FDA for approval to conduct a pilot clinical trial. The trial began in 2004. So far, four patients have enrolled.

For each trial patient, training sessions begin soon after the sensor is inserted. The volunteer is asked to imagine moving one hand as if he were controlling the computer mouse. The researchers study the data and build filters to convert patterns of neural spikes into two- dimensional commands.

"Training patients to move things with their minds is different with each patient," said Maryam Saleh, who worked with the first two patients as a Cyberkinetics technician and is now a doctoral student in Hatsopoulos's Chicago lab.

The current BrainGate System is still in its infancy and is far from perfect. It is bulky and cumbersome. The quality of the signal can vary from patient to patient and from day to day. A great deal of work remains to be done to extend the longevity and reliability of the sensor. Patient two never developed as much control as Nagle, and even Nagle's level of control, the authors note, "is considerably less than that of an able-bodied person using a manually controlled computer cursor."

Despite remarkable progress, "this isn't being done for the patient's benefit," said University of Chicago neurosurgeon Richard Penn, who implanted the sensor in the second patient. "It's being done for mankind's benefit."

"Most people involved in this study think of themselves as pioneers," said Saleh. "They see the prospects for future applications. That's why they do it."

Nevertheless, Penn added, "this is the strangest, most interesting surgery I've ever done. Not the technical stuff, but the data that we get from the neurons firing in different patterns when you're thinking in different ways. And seeing it is only the beginning."

The system is constantly being improved. Next steps include faster and more precise algorithms to help the computer keep pace with the neuronal inputs, and a more portable wireless system. The researchers are also looking at new applications, such as enabling the brain-computer combo to control a wheelchair or other gadgets that will restore some control and freedom to patients with severe paralysis.

At the American Spinal Cord Injury Association meeting in June, David Chen, medical director of the spinal cord injury rehabilitation program at the Rehabilitation Institute of Chicago, presented preliminary results from a third participant in the trial. This patient, who is unable to speak because of a brainstem stroke, can control a cursor with significantly greater stability than the first two. He can stop cursor movement at will, "click" on icons and type messages using assistive software.

"We believe these advances could ultimately enable a paralyzed person to control communication devices, medical devices, computer-controlled robotics, wheel chairs – and even their own limbs," said Cyberkinetics's Surgenor.

"As a physician," said Harvard's Leigh R. Hochberg, lead author of the Nature paper and a principal investigator in this pilot trial, "I do whatever I can to optimize the recovery of patients with paralyzing disorders such as stroke, spinal cord injury or neuromuscular disease. The available assistive technologies, however, provide neither sufficient independence nor mobility.

"Thanks to the generosity and pioneering spirit of our initial trial participants, who have volunteered without expecting to derive any personal benefit, important progress is being made in developing a real-time neuromotor prosthesis," Hochberg said. "Though much work remains to be done, hopefully one day, I'll be able to say: 'We have a technology that will allow you to move again.'"

Co-authors of the study include: Leigh R. Hochberg, MD, PhD, the Massachusetts General Hospital, Spaulding Rehabilitation Hospital and Instructor, Harvard Medical School; John P. Donoghue, PhD, Founder and Chief Scientific Officer of Cyberkinetics, Professor and Director of the Brain Science Program, Brown University; David Chen, MD, Medical Director, Spinal Cord Injury Program, Rehabilitation Institute of Chicago; Richard Penn, MD, Professor of Neurosurgery, University of Chicago Hospitals; Mijail D. Serruya, MD, PhD, Founder of Cyberkinetics and Department of Neuroscience, Brown University; Jon Mukand, MD, PhD, Sargent Rehabilitation Center; Gerhard Friehs, MD, Founder of Cyberkinetics, Associate Professor of Clinical Neuroscience, Brown Medical School and Director of Functional Neurosurgery at Rhode Island Hospital; Maryam Saleh, Cyberkinetics and the University of Chicago; Almut Branner, PhD, Cyberkinetics; and Abraham H. Caplan, Cyberkinetics.