Virtuosity in Vision Research

At Einstein, research into vision extends from the latest in anatomy (a heretofore undiscovered eye structure) to discovering an impressive physiology feat (the brain’s ability to suppress redundant visual stimuli). This article describes both of those recent findings along with a notable Einstein contribution to corneal transplantation.

Chuck’s Anatomy

Roy S. Chuck, M.D., Ph.D.Leafing through Gray’s Anatomy, the classic and seemingly exhaustive atlas of the human body, you might reasonably conclude that everything of any anatomical consequence has been discovered. But last year, Einstein ophthalmology researchers identified new structures in the cornea, the transparent membrane that is front and center in the eye. The newly found structures form an impossibly thin layer of tissue, barely a tenth the thickness of a human hair. But they may play a key role in maintaining corneal health.

Team leader Roy S. Chuck, M.D., Ph.D., the Paul Henkind Chair in Ophthalmology in the department of ophthalmology and visual sciences and professor of genetics at Einstein, was as surprised as anyone to find something new in the eye. Dr. Chuck also chairs the ophthalmology department at Montefiore Health System.

“It’s not every day that you discover new anatomy,” he said. “The eye is a little organ, and its anatomy has been studied to death.”

Dr. Chuck wasn’t looking for new ophthalmic landmarks. He was interested in better clinical images of the cornea, one of his areas of expertise. Using surplus eye bank corneas as his test material, he started experimenting with second harmonic generation imaging—an advanced microscopy technique that captures extremely detailed three-dimensional images of living tissue.

Newly discovered structures within the limbus, the transition zone between the clear cornea and the opaque sclera (arrows). Photo: Chuck lab.“We kept seeing these odd structures in the limbus—the transition zone where the cornea meets the sclera, the white of the eye,” said Dr. Chuck, whose study was supported by unrestricted funding from the Research to Prevent Blindness foundation. “They weren’t in any of the textbooks.”

Dr. Chuck has yet to discover the function of this tissue layer. But its composition (collagen and elastin), location (below a layer of stem cells in the limbus), and architecture (think Swiss cheese with tiny blood vessels passing through the holes) offer intriguing clues.

“All of this suggests that this tissue supports and feeds the limbal stem cells, which are responsible for replenishing cells in the epithelium, the cornea’s outermost layer,” he said.

Dr. Chuck and his colleagues, including Choul Yong Park, M.D., Ph.D., a visiting scientist from Dongguk University in South Korea, dubbed the tissue the anterior limbal cribriform layer. (Anatomists use the term “cribriform” to describe tissues laced with holes.)

The next challenge is to adopt harmonic generation imaging for clinical studies. “There are all kinds of movements in the living human eye, which makes it hard to obtain really good images,” he said. “But once we start to identify the layer reliably and reproducibly, then we can start looking at what happens to the layer in response to disease or injury.”

For good measure, the Einstein team identified a second structure: tiny fibrils that presumably anchor the anterior limbal cribriform layer in place. The two discoveries were reported in Investigative Ophthalmology & Visual Science.

Are there still other structures in the eye yet to be discovered?  “My gut feeling says there are,” he said.

A Revolution in Corneal Transplantation

Graph showing the domestic distribution of corneal grafts from 2005 through 2014 according to the Eye Bank Association of America. After 2011, endothelial keratoplasty (EK) surpassed penetrating keratoplasty (PK). Anterior lamellar keratoplasty (ALK) remained constant. © OphthalmologyOphthalmologists have performed corneal transplants for more than a century with steadily improving results. So it’s no wonder that this eyesight-saving surgery triggers little media coverage. But in recent years, this obscure corner of the transplant field has undergone a critical improvement that every potential transplant recipient should know about, and it’s due in large part to Dr. Chuck’s efforts.

The cornea has two main functions: to protect the rest of the eye and to help focus light on its way to the retina. The cornea is quite durable, but disease, injury and aging can cloud the cornea or render it thin and misshapen. If the damage is extensive, the only remedy is a cadaver cornea transplant.

For most of the last century, ophthalmologists performed corneal transplants with a procedure called a penetrating (full-thickness) keratoplasty, in which all five layers of a donor cornea are transplanted. It requires a relatively large incision, which exposes the eye to infection and other risks. It’s also a bit of overkill, since corneal damage is usually confined to the vitally important endothelium—the innermost layer that keeps the rest of the cornea clear by pumping out excess fluid.

In the 1990s, scientists learned how to tease out just the endothelial layer from a donor cornea, allowing for single-layer, or lamellar, keratoplasties—reducing the risk of rejection, infection, and astigmatism and speeding recoveries. But there was a hitch: Eye surgeons had to perform the delicate task of extracting the extremely thin endothelial layer in the operating room, just prior to the transplant, and few of them were up to the job.

That’s where Dr. Chuck comes in. About 12 years ago, he and his colleagues developed a system allowing eye banks to harvest and preserve corneal endothelial layers so they can be safely shipped anywhere in the world. “When that happened, use of single-layer transplants skyrocketed,” he said.

A decade ago, 95 percent of corneal transplants were full thickness and just 5 percent were single layer. By 2014, those numbers were 42 percent and 58 percent, respectively, as Dr. Chuck reports in a recent review published in Ophthalmology. This trend is expected to continue, sparing thousands of patients from surgical complications.

What Color Is That Dress?

Adam Kohn, Ph.D.Early last year a woman in Scotland posted on Tumblr, the social networking site, a washed-out photo of a dress. She wanted to settle an argument among friends about the garment’s colors. The post quickly went viral, garnering tens of millions of hits. Half of the online commenters saw the dress as white with gold lace, while the other half saw it as blue with black lace—and almost everyone was confused. Which side was right? How could so many people not see what others were clearly seeing?

The befuddlement was not surprising. Neuroscientists themselves are just beginning to understand how we see. Common sense would suggest that vision occurs when light hits the retina. But the real business of seeing begins an instant later, when electrical signals from the retina’s rods and cones are relayed to the visual cortex, the part of the cerebral cortex that receives sensory nerve impulses from the eyes. A near-miraculous, quasi-computational process occurs there in which millions of nerve impulses are transformed into coherent views of the world, image after image, millisecond after millisecond.

One researcher trying to make sense of all this is Adam Kohn, Ph.D., associate professor in the Dominick P. Purpura Department of Neuroscience and of ophthalmology and visual sciences. He is particularly interested in a sensory phenomenon called surround suppression.

It was once thought that each neuron in the brain’s visual cortex responds to a localized area of the visual field—much like a pixel in a camera sensor. “But the eye is not a camera and the brain is not a computer,” said Dr. Kohn. Researchers have since shown that a neuron can be suppressed by a stimulus outside its receptive field, an area known as the surround. In other words, a neuron’s signal will be suppressed if its input is similar to that of surrounding neurons. “That way, the brain doesn’t need to encode every single point in the image,” he explains. “It can use a kind of compressed code.”

The Dress.At least, this is what happens when the brain processes simple images. In a recent study funded by the National Institutes of Health, with additional support from the Research to Prevent Blindness foundation, Dr. Kohn and his team investigated whether that’s also true for complex images.

To test this hypothesis, macaque monkeys were presented with photos of animals in the wild, landscapes, buildings, and other real-world scenes. Each image was quantified using a technique called natural scene statistics, which provides a measure of similarities and differences within the image. As they presented each photo to the macaques, the researchers made individual recordings from about one hundred neurons in the animals’ primary visual cortex, the first region of the brain that processes visual input.

The results, published in Nature Neuroscience, show that surround suppression was stronger for images that had statistically redundant signals. Neuroscientists now had a new model for predicting how the brain will respond to natural images at the neuronal level. The key take-home message is that how we interpret visual sensory information depends on the environmental context.

Which brings us back to the dress—which is actually blue and black.

“The reason people see the dress in different ways is because of assumptions about the lighting conditions in which the image was taken,” said Dr. Kohn. “If you think the photo was taken indoors, with a certain type of light, you might see the dress one way; and if you think it was taken outdoors, you might see it another way. The context in which we see things matters. That’s true for everything we see.”

Dr. Kohn’s next challenge: to measure how neurons interact when the visual system is presented with complex images. This will give neuroscientists a fuller picture of how we see and perhaps how other senses work as well.

“If we can understand computation in the brain,” he said, “we will be able to implement those rules in artificial systems for a host of applications, including better brain-machine interfaces for people who are paralyzed or have lost their eyesight.”