SAN FRANCISCO -- Any geek worthy of the moniker has dreamed of connecting his or her brain directly to a computer for blissful freedom from keyboard and mouse. For quadriplegics, that ability would give life a whole new dimension.
If people with physical handicaps could control a computer by just thinking, they could also operate light switches, television, even a robotic arm -- something the 160,000 people in the United States who can't move their arms and legs would surely welcome.
Work in that brain-computer interface, or BCI, technology has ramped up considerably in the past five years. More than half of the scientific papers on the topic were published in just the past two years. Also, by connecting their patients' brains directly to a computer, researchers have seen improvement in patients' ability to control a cursor.
Cyberkinetics is leading research on BCIs in the private sector. Last year the company enrolled its first patient, Matthew Nagle, in a clinical trial to test its BrainGate system. From his wheelchair, Nagle can now open e-mail, change TV channels, turn on lights, play video games like Tetris and even move a robotic hand, just by thinking.
"Not bad, man, not bad at all," Nagle says in a video as he uses BrainGate to control a hand for the first time since he was stabbed in the neck during a fight at Wessagussett Beach in Weymouth, Massachusetts. The stab wound severed his spine and left him paralyzed and on a respirator.
The device, which is implanted underneath the skull in the motor cortex, consists of a computer chip that is essentially a 2-mm-by-2-mm array that consists of 100 electrodes. Surgeons attached the electrode array like Velcro to neurons in Nagle's motor cortex, which is located in the brain just above the right ear. The array is attached by a wire to a plug that protrudes from the top of Nagle's head.
The electrodes transmit information from 50 to 150 neurons through a fiber-optic cable to a device about the size of a VHS tape that digitizes the signals. Another cable runs from the digitizer to a computer that translates the signal.
The Matrix -like device protruding from Nagle's head seems little price to pay for the new abilities he's gained thanks to BrainGate.
But other researchers are working on simpler, noninvasive BCIs. Jonathan Wolpaw , a professor at the Wadsworth Center in New York, published a paper in December 2004 in the Proceedings of the National Academy of Sciences showing that his noninvasive electroencephalogram, or EEG, cap could pick up brain signals at least as well as Cyberkinetics' invasive technology.
Both patients and their doctors would prefer not to open the skull to implant a BCI, but it's not yet clear whether a BCI sitting outside the head will be as good at picking up brain waves as an implanted device. Experts generally thought the answer was no until Wolpaw published his results.
"It's clear that noninvasive methods can be a lot better than most people gave them credit for," he said. "How much better they can get and how much better invasive methods can get is all up in the air."
This is an important question for patients balancing the potential to greatly expand their physical abilities with the possibility of infection or even brain damage.
"Noninvasive would be important to me," wrote Steven Edwards, who lost use of his arms and legs after a car accident in 1998, in an e-mail. "I would not want something implanted into my brain unless it significantly enhanced the experience (think virtual reality) or allowed me to be more efficient (think a 3-D card allowing me to do large amounts of trig in my head nearly instantaneously)."
External BCIs might also have their own advantages, because they can retrieve signals from many points in the brain rather than just a specific site.
"Implanted electrodes are very specific, so they can record activity relative to the intended muscle or motor movements, and that has its uses," said Charles Anderson , a researcher at Colorado State University. "We're hoping to identify a higher level of cognitive activity like different mental tasks. That would take too many implants."
While Cyberkinetics may not be able to satisfy Edwards, at least in the near term, the company is planning other next-generation technologies, said CEO Tim Surgenor in an interview during the JPMorgan Healthcare Conference in San Francisco. Nagle can already control a robot hand, and Surgenor says he can imagine the possibilities if technicians could calibrate the motion with more precision: from pouring a cup of coffee to swinging a tennis racket or writing a letter. And eventually, researchers hope to implant electrical probes directly into muscles, so patients can use the system to control their own limbs.
Such technologies are decades off, but in the meantime, Surgenor said, the company is developing a brain-controlled wireless handheld. Such devices would be a good fit with its technology, since handhelds often rely on left, right, up and down commands.
Cyberkinetics needs four more patients to complete enrollment in its Food and Drug Administration-approved BrainGate clinical trial. The company is also seeking FDA approval for a study testing BrainGate on patients with amyotrophic lateral sclerosis (often called Lou Gehrig's disease, or ALS). They hope to enroll one patient by the end of this year.
Another private company, Neural Signals , has developed a BCI that employs a small screw inserted 2 mm beneath the skull. The $50,000 device (including about $30,000 for surgery) is FDA-approved. The best candidates for the device are patients who are " locked-in " and have no movement at all, such as people with ALS. The device allows patients to move a cursor and turn a switch on and off.
The market for a convenient, reliable BCI is estimated at about $2 billion, Surgenor said. But before companies can pursue that market, researchers must first make the bulky equipment smaller, more accurate and more automated, so patients can turn on and calibrate the systems themselves. Cyberkinetics has already outlined plans for a prototype that would be implanted behind the ear, much like a cochlear implant , and attach to external equipment via a magnet, so patients would not have a device protruding through their skin. Cyberkinetics wouldn't estimate how much its BCI will cost, but Wolpaw said his noninvasive system would likely cost around $10,000.