Science & Technology: Artificial Vision
Now, Electronic "Eyes" for the Blind
Brain-implant systems may soon hit the market
Blind people using a white cane to move through the underground labyrinth known as the New York City subway system are fairly common. But there's no one like Jerry. He doesn't have a cane or guide dog. Instead, he uses a prototype artificial eye. Later this year, an advanced version of his artificial-vision system should hit the market. The price will be steep, initially--close to $50,000--but could drop as production increases in volume.
Jerry, who asks that his last name not be used, lost all sight in an accident in 1974, when he was 36. A couple of years later, he began working with William H. Dobelle, then director of artificial-organ research at New York's Columbia-Presbyterian Medical Center and now CEO of Dobelle Institute Inc. Dobelle has waged a three-decade-long crusade, investing $25 million in technology to restore sight to the blind. Since 1978, Jerry has been a star guinea pig. "I'm in this with Bill for the long haul," he says.
The key to the artificial-vision system is a printed-circuit board, about the size of a jumbo postage stamp, implanted in Jerry's brain. The card has a grid of electrodes that stimulate the visual cortex, creating coarse black-and-white images. Individual electrodes are turned on and off on the basis of signals from a tiny TV camera mounted in the right lens of Jerry's sunglasses.
First, though, the video images must be digitized and processed by a notebook computer on Jerry's belt or in a backpack. Last summer, the system got a major overhaul. Thanks to new software and faster hardware, Jerry can now read 2-inch-high letters from 5 feet away--roughly equivalent to 20/400 vision. And in recent weeks, he has been learning to read the video signals from a computer so he can plug into the Internet. When Jerry says "plug into" the Internet, he means it literally: He yanks out the wire to his TV-camera eye and inserts the end of the video cable from a PC. His backpack computer must still act as interpreter, though. Normal Internet screens involve way too much detail. So Dobelle's programmers are experimenting to see how much "noise" can be expunged without destroying a screen's basic functions.
The technology has improved dramatically. Twenty years ago, the computer was as big as a refrigerator. And it would often crunch data for minutes on end to figure out an image's most salient features--the first step in determining which electrodes to stimulate in Jerry's head. As a result, the experience of artificial vision had to be savored within a single room. Moreover, Jerry then could only discern letters bigger than 6 inches at 5 feet.PIXEL POWER. Jerry can barely wait to get the commercial version of Dobelle's electronic eye. Its brain card will feature 512 electrodes, eight times the number now inside Jerry's skull. That should provide a major increase in image quality. Of course, an image with 512 pixels is a far cry from the 200,000 pixels on today's television sets, let alone the 20 million in a 35mm photographic negative. But if a grainy, low-resolution image can restore some mobility and help blind people hold down jobs that involve using a PC, then even 512 pixels would be worth a billion words.
Dr. William J. Heetderks, head of a National Institutes of Health program focused on developing electronic implants, is optimistic that high-tech implants "will be able to provide significant function to blind people." There's also an alternative to brain implants. Retinal stimulation uses a small imaging chip implanted in or behind the eye. One NIH-backed team, headed by Drs. Eugene DeJuan and Mark Humayun at Johns Hopkins Medical School, has placed temporary 25-pixel prototypes in more than a dozen patients. Similar technology is being developed by a group at Massachusetts Eye & Ear Infirmary and Massachusetts Institute of Technology--and by a team in Germany that includes the University of Bonn and startup Intelligent Implants.OLD MODEL. Retinal implants, however, require the eye and optic nerve to be healthy. They can't help Jerry or others who lost their eyes, or those with diseased optic nerves. That's why the NIH is also funding work on brain implants at Illinois Institute of Technology, although this research is still at the exploratory stage, says Heetderks.
Dobelle began his artificial-vision research in 1968 at the University of Utah. For the first dozen years, his main help came from a few farsighted executives. Harold Geneen, former chairman of ITT Corp., had his factories produce custom parts for Dobelle. Digital Equipment Corp. kept him supplied with the latest computers. And Polaroid Corp. founder Edwin H. Land gave Dobelle some stock, which Dobelle sold to finance his early work.
These benefactors suggested that Dobelle become self-supporting by adopting the Wright brothers' model: Wilbur and Orville Wright paid for their airplane with profits from their bicycle factory. So Dobelle created the Dobelle Institute and developed a much less risky line of medical products. The main one is a pacemaker-like gadget that regulates breathing for patients with severe respiratory problems. Since 1981, sales to some 100,000 patients worldwide have provided $18 million for artificial-vision R&D. The Dobelle Institute is privately held, and Dobelle says only that annual revenues are "in the millions, but not tens of millions." That's peanuts compared to what may be coming. Dobelle expects to sell 10,000 artificial-vision systems a year in fairly short order.
Americans, though, will have to travel overseas for the operation. The Food & Drug Administration will likely be wary of allowing a brain implant with a connection through the skull--a potential pathway for infection. Still, the outlook for blind people has never been brighter.By Otis Port in New YorkReturn to top