"New Retinas for Old," Feature Article, October 2003

FEATURE FOCUS: BioMEMS
new retinas for old

After years of research, scientists are one step closer to restoring vision to people suffering from certain types of blindness.

by Gayle Ehrenman,
Associate Editor

For roughly 1.3 million Americans who have lost their sight to age-related macular degeneration or to retinitis pigmentosa, there is a research project under way that may help them see again. According to some members of the project, these people could have some of their sight back before the decade is out.

The researchers are developing a retinal prosthesis, a "seeing eye" chip with as many as 1,000 tiny electrodes, to be implanted in the eye. It has the potential to help people who have lost their sight regain enough vision to function independently in the sighted world. Developers say a version of the device—not with a full 1,000 electrodes, but still potentially marketable—may enter clinical trials in about 18 months.

The Artificial Retina Project is a collaboration of five U.S. national laboratories, three universities, and the private sector. Project funding is being provided by the U.S. Department of Energy's Office of Science, which has anted up $9 million over three years to kick start the project.

The five DOE labs working on the project—Argonne, Oak Ridge, Lawrence Livermore, Sandia, and Los Alamos—are partnering with the University of Southern California; the University of California, Santa Cruz; and North Carolina State University to design a microelectromechanical device that can be implanted in the eye on the surface of the retina. In this artificial retina, a microelectrode array will perform the function of normal photoreceptor cells, to restore vision for people whose photoreceptors have been damaged. A private company, Second Sight LLC of Sylmar, Calif., is also involved in the project.

"The aim is to bring a blind person to the point where he or she can read, move around objects in the house, and do basic household chores," said Sandia's project leader, Kurt Wessendorf. "They won't be able to drive cars, at least in the near future, because instead of millions of pixels, they'll see approximately a thousand."

LIGHT INTO UNDERSTANDING

Although the retinal prosthesis itself will be a tiny device, creating it is no small task. The eye is a very complex organ that does a fair amount of processing to turn simple light into signals that the brain can understand. At its most basic, vision is produced when light enters the eye and gets turned into electrical signals, which the optic nerve carries to the brain's primary visual cortex. If the signals are disrupted on their way to the visual cortex, or the light is not properly converted, blindness results.

More specifically, light entering the eye passes through the cornea, the aqueous humor, the lens, and the vitreous humor. Finally, it reaches the retina, the light-sensing part of the eye. The retina contains two types of cells, rods and cones. Rods manage low-light vision, while cones manage color vision and detail. Light contacting these cells creates a series of complex chemical reactions. These reactions produce a chemical called activated rhodopsin, which creates electrical impulses in the optic nerve. The nerve fibers from the retina eventually reach the back of the brain, or the primary visual cortex, where vision is interpreted.

In retinal diseases, such as age-related macular degeneration and retinitis pigmentosa, problems with the retina prevent the electrical impulses from properly forming and transmitting to the primary visual cortex. Age-related macular degeneration is the leading cause of blindness in people over the age of 60; retinitis pigmentosa is the leading cause of blindness in those under 50. Together, the two causes account for roughly 75 to 80 percent of the 1.7 million Americans who are classified as legally blind, according to the National Eye Institute, a division of the National Institutes of Health.

The interface module and the antenna for future versions of the retinal prosthesis will all be implanted in the eye, instead of outside the eye.

In the case of age-related macular degeneration, photoreceptor cells in the macula (the center portion of the retina responsible for fine detail in the center of the visual field) deteriorate over time, and the patient ultimately loses all central vision.

Retinitis pigmentosa is a group of inherited retinal diseases that affect about 100,000 Americans and 1.5 million people worldwide. It causes the progressive deterioration of the photoreceptor cells, the specialized, light-absorbing cells in the retina.

As these cells slowly degenerate, people with retinitis pigmentosa develop night blindness and a gradual loss of peripheral vision. By about age 40, most have tunnel vision, although many may retain good central vision. Between the ages of 50 and 80, however, they typically lose their remaining sight. The disease generally is first diagnosed in childhood or adolescence.

Currently, there is no cure for macular degeneration or retinitis pigmentosa. That's where the artificial retina comes into the picture. It is intended to help patients who still have the neural wiring from the eye to the brain, but who lack photoreceptor activity, according to Mark Humayun, a professor of ophthalmology at the Doheny Eye Institute in the Keck School of Medicine at the University of Southern California, in Los Angeles. Humayun, who is also a bioengineer, is the lead researcher on the Artificial Retina Project.

"The artificial retina works by throwing a switch on," Humayun said. "If you've had vision, and then lost it, the moment you access the visual pathways, the brain becomes very interested. It remembers what vision is."

A VERY BASIC VISION

The current prototype of the artificial retina has pieces both inside and outside the eye. The patient wears a pair of tinted glasses that have a tiny video camera mounted on them. The camera captures images and sends the data to a visual processing unit, a microprocessor roughly the size of a cell phone, which is worn in a belt pack. The microprocessor converts the data to an electronic signal and transmits it via wire to a coil taped behind the patient's ear. An antenna in the lens of the glasses transmits the signal to a receiving antenna implanted in the eye. This coil transmits the signal using amplitude modulation (standard AM radio waves) that goes to a computer chip implanted behind the ear. This computer chip then sends the image data along a tiny wire to the retinal implant itself. The signal causes the implant to stimulate the remaining retinal cells. These cells send the image along the optic nerve to the brain. The result is a crude form of vision.

The retinal implant is an electrode array that is implanted in the back of the eye. In the first-generation prototype, the array has 16 electrodes arranged on a 4-by-4 grid. This array, which is made of silicon and platinum, measures 4 millimeters by 5 millimeters. This first-generation device was manufactured by Second Sight LLC.

Three patients have had the electrode implanted in one eye each, in the first round of clinical testing. In these tests, patients have been able to perceive light on each of the 16 electrodes. Patients have reported being able to detect when a light is turned on or off, to describe the motion of an object, and even to count discrete objects, according to Humayun.
The first tests of the prosthesis in all three patients involved computer-generated points of light sent directly to the implant, Humayun said. Once the patients were trained with the device, they received images from the video camera mounted on the glasses. All three patients have tolerated the device well, Humayun said. One patient has had the implant in place for 18 months, with no adverse reaction. Currently, Second Sight is petitioning the U.S. Food and Drug Administration to allow another five patients to be implanted with the retinal prosthesis.

Lawrence Livermore National Laboratory in Livermore, Calif., is focused on electrical array development. The ultimate goal, said Courtney Davidson, principal investigator for the lab's Retinal Prosthesis Effort, and group leader for advanced microfabrication for Lawrence Livermore's Center for Micro and Nanotechnology, is to create an array of 1,000 electrodes. Currently, prototype electrodes are 4,000 micrometers in diameter; the goal is an array that has 50 µm diameter electrodes. The 1,000-electrode array, according to Humayun, will deliver enough optical resolution for patients to read and recognize fine shapes.

Patients who have age-related macular degeneration lose their central vision. Tunnel vision is the end result of retinitis pigmentosa.

One of the challenges in developing this device, according to Davidson, is creating a microelectrode array that conforms to the curved shape of the retina, without damaging the delicate retinal tissue. To do this, the lab is building metal electrodes on a form of silicone rubber called poly(dimethysiloxane), or thin PDMS, which has the look and feel of plastic food wrap, Davidson said. It is fairly robust, highly impervious to water, somewhat gas permeable, and poses very little risk of yielding dangerous byproducts, all of which make it a good choice for a chronic (that is, permanent) implant, according to Davidson.

"The flexibility of PDMS allows us to create a highly flexible array and series of interconnects that will conform to the retina, without damaging it," Davidson said. "We've come up with an approach that allows us to lay down thin-film 'wires' on the PDMS that can take seven percent strain without breaking," he said. "Our electrical connectors can bend and still maintain electrical connectivity."

The prototype of the retinal prosthesis electrode array sandwiches the eight electrodes between layers of PDMS. One side of the sandwich has holes in it that allow contact with the electrodes. The other side is solid.

In thin-film metalizations, such as those in use with the retinal prosthesis, the wires are 100 micrometers wide, according to Davidson. There is still plenty of room to reduce the size of the wires and interconnects, and this will have to happen to reach the Holy Grail of a 1,000-electrode array, he said.

Sandia National Laboratories in Albuquerque, N.M., is also developing advanced electrodes. Its approach uses LIGA (a German acronym for lithography, electroplating, and molding—a method of MEMS fabrication) and surface micromachined silicon parts, according to Wessendorf.

BIOCHEMICAL CHALLENGES

Oak Ridge National Laboratory in Oak Ridge, Tenn., is the lead lab on the Artificial Retina Project. They're the folks responsible for fabricating and testing the electrodes, and making sure they're up to the challenge of being implanted long-term in a human body.

"The eye is a particularly hostile environment," said Elias Greenbaum, a corporate fellow for Oak Ridge Lab and its principal investigator for the Artificial Retina Project. "We don't know how happy the retinal tissue will be long-term. There's no real chance of infection from the implant, but we still need to determine the long-term stability of the device."

One of the biggest concerns for Greenbaum is toxic byproduct formation as a result of a biphasic pulse passing through the electrodes of the array. In use, first the array is hit with a negative pulse to stimulate the neural cells; this pulse is then quickly reversed to positive. The problem is that electrical stimulation creates an electrochemical reaction that dissolves the material of the interconnects and wires. And, Greenbaum said, "With a platinum electrode that's the size of a hair, you don't have much material to spare."

In particular, Greenbaum is concerned with the electrolysis of water, which results in the formation of hydrogen. Chlorine gas may be formed when you hit the electrodes with a positive pulse. During the negative pulse phase, molecular oxygen can react with the electrode, leading to hydrogen peroxide, he said.

A camera on the patient's glasses is able to capture images, which are transmitted via radio waves to a receiver that is implanted behind the patient's ear.

The trick to preventing the creation of these harmful substances is to use biphasic pulses intelligently. "There's a finite window of time when toxin formation can occur," Greenbaum said. "If we can reverse the polarity at the proper time, we can frustrate the formation of toxins by being faster than their formation."

Solving this problem will help the Artificial Retina Project move one step closer to the 1,000-electrode chip.
The other national labs working on the project are each tackling other challenges. The team at Argonne National Laboratory in Argonne, Ill., is working with its patented ultrananocrystalline diamond technology for the packaging of the implantable electronics and as electrode material. These ultrananocrystalline diamond films are said to be vital to overcoming the current size constraints of the implant. The diamond films have a low friction coefficient and surface adhesion, high electron emission, chemical inertness, and high conductivity, according to Argonne researchers. Argonne also is collaborating with Second Sight on soak testing of the device.

Los Alamos National Lab in Los Alamos, N.M., is developing advanced optical imaging techniques. The lab is modeling and simulating neural paths from the retina to the brain, to better understand how the retinal prosthesis will behave.

WHAT'S NEXT?

The earliest prototype of the retinal prosthesis is already in clinical tests. But this version, dubbed Model 1 by Humayun, who pioneered the concept of an electrode-based artificial retina, is only the beginning. Model 2, which has 60 to 100 electrodes, is being used in preclinical trials on blind dogs. In a change from the configuration of the device currently being tested, Model 2 would move the receiving coil, currently implanted behind the ear, to somewhere around the eye. That device is roughly 18 months away from human testing, according to Humayun. He expects the surgery to implant this device to take only 90 minutes, as opposed to the six hours required to implant Model 1. Model 2 also is the first retinal prosthesis that he considers "marketable." It should provide patients with the ability to distinguish shapes and light from dark.

The "chip" for Model 3, which has 1,000 electrodes, is just back from dunk testing, and is "functioning well," according to Humayun. Dunk testing consists of dunking the chip in a heated saline solution, which functions as an accelerated test environment. The next phase is mechanical testing, followed by preclinical trials, and then clinical trials, according to Humayun. This device moves the receiving coil into the eye.

The optimistic timeline for this model, which will give patients the ability to read and recognize fine detail, calls for the device to be available in five years. "One thousand electrodes is doable with the current technology," Humayun said.

Even the 1,000-electrode array won't restore full vision to patients. "We're not talking high-definition television," said Greenbaum at Oak Ridge. "We're talking about restoring enough vision to allow people to function in society."

Color vision is one area that probably won't be restored, but anything is possible, according to Humayun. "One patient reported seeing blues, oranges, deep reds, and greens with the implantation of the 16-electrode array. But there's no way for us to confirm that. Color vision is really of secondary importance."

There's also a need to explore what Livermore's Davidson calls the "psychophysics" of the artificial retina, in order to discover whether the brain will have to be taught how to see again, and how much it will be able to interpret and learn.

Humayun is optimistic that the combination of the 1,000-electrode array, new and improved software algorithms, and a more powerful camera with increased zoom range will provide better-than-expected vision for millions of people who haven't seen for years.

SIDEBAR: OTHER WAYS TO SEE