Williams' adaptive optics system may eventually replace the phoropter, the device used by eye doctors to prescribe glasses and contact lenses. It is also being used to provide ultra-high resolution pictures of the retina, which will help in diagnosing and treating eye disease.
But adaptive optics will be most useful when it can be used to correct vision permanently outside the lab. That's why eye surgeons are so interested in Williams' work. His technique is fairly easily translated to surgery (and may also work for contact lenses). A laser surgeon can follow the map of errors revealed by the wavefront sensor, making minuscule, precise corrections on the corneal surface. No longer will laser surgery be limited to the big aberrations that surgeons can now eliminate: It could erase every error in the eye.
Williams is collaborating with Dr. Scott MacRae, a celebrated University of Rochester eye surgeon, who recently finished a laser surgery trial on more than 300 patients using the wavefront sensor. MacRae and Williams are also working with Bausch & Lomb to develop contact lenses that would eliminate higher-order aberrations. At a January conference, MacRae reported that 91.5 percent of the eyes treated in his study attained 20/20 or better vision six months after the surgery. And in another surgical trial of 340 patients, Dr. Stephen Slade reported that more than 70 percent achieved 20/16 vision or better.
Even if the surgery proves to be a successful technique, it's unrealistic to expect every patient to achieve 20/10 vision with no higher-order aberrations: Surgery is always risky, patients respond differently to treatment, and many have other visual problems that could weaken their sight.
Bausch & Lomb and its competitors are puzzling out how to apply wavefront sensor technology to contact lenses, which float around on the eye. To eradicate higher-order aberrations, the lenses must lock in one place—which remains an engineering challenge. MacRae says that Bausch & Lomb "is making inroads" on the customized contact lenses.
Neither Williams nor McRae would predict when such surgery or contact lenses might be used widely, but it will probably happen a few years from now, not a few decades. In November 2002, the FDA recently approved its first wavefront-guided surgical system, LADAR-Vision, which is made by Alcon, Bausch & Lomb's competitor.
2) The Cyborg Eye
A dozen teams of American scientists are working on implants to restore sight to the blind. The scientists all begin with the same basic notion: Sight is essentially an electrochemical signal in the brain, and in the blind, something prevents the light that enters the eye from becoming a usable signal. (Usually, the source of this failure is the retina's rods and cones, the delicate cells that detect light and color.) These implant scientists want to circumvent the broken parts of the eye and send a signal to the cells that still work.
The implant researchers take two basic approaches. Some want to place miniature light-sensitive photo arrays in the retina itself. This "retinal implant" replaces broken rods and cones. When light strikes the photo array, its sensors transform the light into an electrical pulse that fires the optic nerve, which in turn stimulates the visual field in the brain. Due to the difficulty of transmitting signals from machine to cell, today's implants contain no more than 1,000 photoreceptors. The fovea, by contrast, has more than 100,000 rods and cones, meaning current implants might provide only a blurry, low-definition approximation of normal vision. (So far, only one outfit, a company called Optobionics, claims to have put a working implant in a human eye.)
Other groups are targeting the brain, not the eye. Rather than repair or implant eyes, they mount a camera on the head of the patient or build it into a pair of glasses. This camera transmits a picture to electrodes that have been implanted in the brain's visual center or to the optic nerve. William Dobelle, a New York scientist, has just installed the first successful brain implant. He put electrodes—about 100 of them, according to a Wired reporter who witnessed the experiment—in a man who'd been blind for 20 years, then hooked the electrodes up to a signal processor and a camera. The blind man was able to see—very crudely, to be sure, but enough to walk around unguided. Dobelle's implants suffer from the same shortcoming as retinal implants: With only 100-odd electrodes, they deliver a fuzzy facsimile of normal sight.
Brain imaging and miniaturization are improving rapidly. It's possible that Dobelle and others will figure out how to feed 1,000 electrodes, or even 100,000 electrodes, into the brain or build retinal arrays with 100,000 sensors.