University of Rochester

"Star Wars" Technology Finds A Surprise in Our Eyes
Clearest Images of Retina Reveal Remarkable Randomness

February 10, 1999

The best images ever obtained of the living human retina show an unexpected randomness in the way that individual cells called cones are arranged. The results, obtained by a University of Rochester team using a technology called adaptive optics that was originally developed by the military, are in the Feb. 11 issue of Nature.

"We've known for 200 years that there are three types of cones in the human eye, but we've never been able to see how they're arranged before," says first author Austin Roorda, who did the work as a post-doctoral researcher at Rochester. "We're hopeful that someday the same technology that allowed us to do this work will help physicians more effectively diagnose and treat diseases that cause blindness."

The findings were possible thanks to a laser-based system developed by David Williams and colleagues at the University over the last decade that maps out the topography of the inner eye in exquisite detail. The team built on technology known as adaptive optics, initially proposed by astronomer H.W. Babcock in 1953, then developed by the U.S. military to clear up images from spy satellites. The idea is to correct for aberrations in the atmosphere so that rays of light travel in parallel lines and converge at a single point, delivering a sharp image. Astronomers use the technique in telescopes to grab ever-better photos of the heavens. Williams, director of the University's Center for Visual Science, leads an effort to apply the same technology to human vision.

Roorda and Williams took unprecedented photographs of the retinas of three people, two with normal color vision and one with color vision problems. The retina is basically a screen inside the eyeball that is packed with millions of cells known as cones and rods that detect light signals and convert them into electrical signals that the brain puts together to form images. There are three kinds of cones: S, or short-wavelength, to see blue light; M, or medium-wavelength, to see green light; and L, or long-wavelength, to see red light. Most cones detect either green or red light, and their arrangement has been completely unknown until now.

The team's biggest surprise was the tremendous difference between the two people with normal vision: One had about the same number of cones to detect green and red, while the other had four times as many cones to detect red light. Not only that: The cones seemed randomly arranged, and in both people large areas of the retina lacked one type or the other. Nevertheless, both see colors the same, and neither lacks any color vision.

"This random arrangement is not at all what we expected," says Williams. "If you were to design a digital camera, you'd never choose a random geometry like this. Yet it turns out that that's what the most sophisticated imaging device in the world the human eye uses."

The idea that a relatively large area of our retina is blind to one color is not completely new: The section of our retina with the sharpest vision can't see blue. Yet somehow the brain fills in that information, enabling us to see blue wherever we look. That's one possible explanation why most of us detect no gap in the way we perceive colors even though there appear to be clumps of certain types of cones.

"The brain does an awful lot behind the scenes," says Roorda, who is now an assistant professor of optics at the University of Houston College of Optometry. "It takes in a tremendous amount of raw visual information and processes it before providing us with our view of the world. It's no wonder that about one-third of our brain is devoted to vision."

In the experiment scientists briefly shine a laser beam into a person's eye, then break the light reflected back into more than 200 closely packed laser beams. Then they use a sophisticated sensor to analyze deviations in each beam's path as they travel into an electronic camera. The analysis reveals imperfections or aberrations that exist in the person's cornea and lens.

The system then counters this aberration with the help of a sensitive "deformable mirror," whose 37 tiny computer-controlled pistons move minute regions of the two-inch-wide device ever so slightly based on the customized information about the person's optical system. The result is a system that can send light unimpeded into and out of the eyeball. The system gives scientists a clear view of the retina and also allows them to shine different colors of light into a subject's eye and watch how individual cones react.

Such information should be invaluable to ophthalmologists, who scan the retina for hints of diseases like glaucoma, retinitis pigmentosa, and age-related macular degeneration, which is the leading cause of blindness in the elderly. The technique also offers a clearer view than ever before of the capillaries in the back of the eye, says Roorda, making it much easier to catch the early signs of diabetic retinopathy, which can lead to harmful bleeding in the eyes of diabetics if left untreated.

The approach could also lead to new tools to give the best vision possible to pilots, athletes, and others who need outstanding vision. "Adaptive optics makes the world look crisper," says Williams, who is the William G. Allyn Professor of Medical Optics and a faculty member in the departments of brain and cognitive science, optics, and ophthalmology. "It may allow for a level of vision correction that's just not available today."

The team has presented the research at several seminars, where the reception by ophthalmologists often evolves from skepticism to surprise and wonder at the clarity of the images. The University has patented the technology.

The team's work is funded by the National Institutes of Health, Prevent Blindness America, and Research to Prevent Blindness.