Low-resolution retinal implants aimed at helping the blind gain mobility are getting ready to appear on the market.
click the image to enlarge
"If it is determined that an ASR microchip implanted into the subretinal space of the eye in an RP patient confers a neuroprotective effect, then RP patients around the world may have a treatment that may not only improve but prevent further deterioration of their vision," says Alan Chow at Optobionics.
Numerous researchers are dedicating their time and efforts to develop microelectromechanical systems (MEMS)-based visual prostheses that will improve the quality of life of visually impaired individuals. Three approaches are generally being explored, with the prosthesis being implanted in the brain's visual cortex, the optic nerve, or in the retina. A retinal prosthesis can be fixated either on the retinal surface (epiretinal) or below the retina (subretinal).
With retinal implants, most scientists are focusing on two degenerative conditions, retinitis pigmentosa (RP) and age-related macular degeneration (MD). These ailments affect about 30 million people around the world and have no effective treatment. A collective term for different diseases that mainly affect the retina's photoreceptors, or photosensitive cells, RP usually runs in families and occurs earlier in life. It affects peripheral and night vision— and in some cases, central vision as well— potentially leading to blindness. MD afflicts mainly the elderly and impacts the central part of the eye's retina, the macula, resulting in diminished visual acuity, and possibly the loss of central vision. In both RP and MD, the retina's photoreceptors decline, preventing the brain from receiving the necessary information for sight.
Most retinal prostheses are designed to replace the function of the receding photoreceptors by stimulating the remaining viable retinal cells to send signals to the brain. But there are those who do not subscribe to the retinal implant approach, like Richard Normann, professor of bioengineering at the Univ. of Utah, Salt Lake City, who is involved in cortical implant research. "Recent work has shown that there's extensive degeneration of the retina associated with these pathologies. So the retina starts to fall apart as soon as the photoreceptors start to fall apart," says Normann. As a result, Normann sees a number of difficulties with the concept of retinal prosthesis.
Although "there are several labs working on a retinal prosthesis, it will be several years before any device will be approved by the Food and Drug Administration (FDA) and useful to patients," says Peter Dudley, an expert in retinal diseases at the National Eye Institute, Bethesda, Md. Further more, everyone in this field agrees that it is too early "to talk about the percentage of sight restored since the prosthesis is rudimentary." The consensus is that higher resolution devices, which will further allow patients to recognize faces and possibly read, could be available in five years. Consequently, research endeavors, some of which are highlighted here, are concentrating their efforts on enhancing the resolution and biocompatibility of these implants.
Epiretinal approach
At the Doheny Retina Institute, which was established at the Doheny Eye Institute, Keck School of Medicine, Univ. of Southern Calif., Los Angeles, a lower resolution retinal prosthesis has been implanted in three blind patients. Devised by Mark Humayun, associate director of research at Doheny, and Eugene de Juan, the organization's CEO, the device was developed by Second Sight, Sylmar, Calif.
Spotted with 16 metal electrodes in a 4 x 4 array embedded in silicon, the 4- x 5-mm retinal prosthesis is placed on the surface of an individual's damaged retina. A receiver is also implanted behind the person's ear, and a stimulator is attached to the implant via a threadlike cable. The implant's electrodes are stimulated by images received from a small camera housed in a pair of shaded glasses. These electrodes, in turn, electrically stimulate the retina's remaining functioning cells, which forward the visual information to the brain through the optic nerve.
Preliminary tests in the three patients implanted in February and July of 2002, and March 2003, indicated that these individuals were able to see a point of light when an electrical current was passed through each of the 16 electrodes in real-time. As a result, they were able to distinguish dark from light and detect the presence and motion of large objects.
The team at Doheny is now currently working on numerous challenges, among them the digital signal processing of images with higher resolution retinal prostheses, which will allow patients to gain more than mobility. Humayun is trying to figure out the best way for such information to be processed since once these higher resolution devices are implanted, digital signal processing will need to occur in real-time. "If a person is about to cross a street, and there's a bus coming, we can't take a minute to process this information," says Humayun. Finding a solution is not simple, however, because "it's not straightforward digital signal processing like the one on your computer or camera— we have to interface it in real-time and see if the brain can make sense out of it."
Another problem to tackle is the design of a processor that consumes low power and is small enough, so a patient can easily wear it. But the biggest challenge facing these scientists is the manufacture of electrodes with higher density because all the electrodes are handmade. The solution to this problem may lie in photolithography-type processes. However, the use of such methods comes with its own set of problems. "The stability of very thin electrode arrays that are microfabricated using photolithography-type techniques is great. But when you put them in water, especially the warm saline water in your eye, then a lot of these techniques fail. The metal, for instance, will flake off or delaminate," says Humayun. This brings up the issue of packaging that the researchers are working on, trying to come up with biological skins that will seamlessly interface with the retinal tissue.
MEMS' evolution in optical networking
click the image to enlarge
The market for MEMS-based components in optical networks is expected to grow gradually in the future. This steady slow rise will be driven by the recovery of the telecom industry, which will increase customer activity, and be tempered by the timeframe needed for product evaluation.
Despite all these hurdles, "a 32 x 32 array, or more than a thousand electrodes, is our goal," says Humayun. The low-resolution devices that have been implanted so far will help people gain mobility. "But in terms of going beyond that, being able to read or recognize faces, you need that high, 32 x 32, electrode count." However, the 32 x 32 array is not an immediate goal for the center and will be offered maybe in the next five years. "But the immediate types of devices that we're implanting for mobility will probably be available in the next year or so."
Helping the center meet all these challenges is a group of national labs, including Sandia, Albuquerque, N.M., who is providing parts to Doheny, and Second Sight. These laboratories are "supplying piece parts, not functional in the clinical sense, but useful to test biocompatibility," says Kurt Wessendorf, a staff researcher at Sandia.
A similar project, the Boston Retinal Implant, is being conducted by a partnership between researchers at the Harvard Medical School, Boston, and MIT, Cambridge, Mass. The team of scientists is led by Joseph Rizzo at the Massachusetts Eye and Ear Infirmary, an affiliate of the Harvard Medical School, and John Wyatt of MIT's Research Laboratory of Electronics. The Boston Retinal Implant project researchers implanted their retinal prosthetic device into five people afflicted with RP, and one, whose vision loss is due to cancer. They placed a tiny thin wire array in contact with still-functioning retinal nerve cells to send electric signals to produce the "sight" of lines and spots.
Fully awake during the test, patients were able to depict what they saw. "A little more than half the time, our patients saw patterns of light that matched the geometric pattern of electric stimulation," says Rizzo. "However, in some cases, where we expected a single electrode to produce a single spot, the patient saw a cluster of two or three spots."
Univ. of Utah's Normann sees this as a major issue in retinal implants. "Does patterned electrical stimulation of the visual pathways evoke discriminatable patterned percepts?" says Normann. "No one has shown this yet, and until someone does, all this work is not a reality."
Bypassing the damage
Another project aiming to develop implants that electrically stimulate retinal ganglion cells, the retina's output cells, is the Epiretinal Implant (EPIRET) project, which is funded by the German Ministry of Education and Research. The work, being conducted by an interdisciplinary research team from numerous universities and coordinated by Mokwa Wilfried at the Aachen Univ. of Technology, Germany, is currently focused on people afflicted with RP. "Later, the system will be adapted to patients suffering from MD," says Mokwa.
Similar to the other epiretinal prostheses, the EPIRET system uses MEMS to evoke action potentials in retinal ganglion cells, creating a visual sensation. It is comprised of an intraocular implant and an extraocular part (in a pair of glasses), which consists of a complementary metal-oxide semiconductor image sensor that records visual images. Once recorded, these images are then turned into corresponding signals by an artificial neural net (retina encoder), which imitates the functions of the retina's ganglion layers. Using telemetery, these signals, along with the energy, are then transmitted to the intraocular implant's receiver unit, the retina stimulator. "Radio frequency links are now used, but optical links are under investigation for future use," says Mokwa.
The retina stimulator's receiver circuitry decodes the signals before transferring the data through a micro-cable to a stimulation unit. The decoded information then chooses a programmed electrode, applying a stimulation current with a programmed pulse width and height to that electrode.
So far, complete systems have been implanted into cats and mini pigs, with cortical recordings being measured 25 msec after telemetric stimulation of the mini pig's retinal ganglion cells (similar results were observed in cats). But what have the results shown so far? They have indicated "that an epiretinal prosthesis based on a telemetric system has been successfully developed," says Mokwa.
He predicts the system to be commercially available in the next three to four years, although there are still unanswered questions regarding a total retinae implant system. "Questions concerning a long-term stable fixation, a long-term stable stimulation, and the minimization of power consumption of the implant," says Mokwa.
click the image to enlarge
Following a recent FDA approval, USC Doheny Eye Institute's implanted patients have been able to take home the wearable camera part and digital processor and activate the entire system.
Subretinal approach
At Optobionics Corp., Naperville, Ill., the Chow brothers, Alan and Vincent, are taking a subretinal approach to help people afflicted with RP and age-related MD. They invented a microchip, Artificial Silicon Retina (ASR), which can be implanted under the retina to stimulate damaged retinal cells. Powered solely by incident light, the 25-µm-thick, 2-mm-dia ASR chip contains 5,000 microphotodiodes, or microscopic solar cells. Each microphotodiode has its own stimulating electrode and converts the light energy from images into electrical/chemical impulses. These artificial photoelectric signals then generate biological visual signals in retinal cells not affected by RP and MD.
Not yet commercially available, the experimental ASR device was implanted in 2000 in 10 RP test patients following permission by the FDA. Enhanced perceptions of movement, shape, and brightness were measured with none of these individuals showing problematic signs like inflammation or rejection. Currently, the Chow team is "evaluating the long-term visual function results of the first 10 patients implanted with the ASR microchip, several of whom are approaching the four-year mark after surgery. Our challenge is to determine if the beneficial results observed so far in the first 10 patients is a consistent trend," says Alan Chow.
This is where Univ. of Utah's Normann recommends caution, suggesting further studies to rule out a placebo effect. "Perhaps when patients undergo this procedure, there is this tremendous desire to regain sight, and maybe they delude themselves," says Normann.
To ensure that this is not happening, Normann believes a number of experiments could be done. For instance, a simple piece of silicon that lacks the photodiodes could be implanted, or the photodiode array could be implanted backwards to see if positive results are obtained. Another possible explanation for the positive results that needs to be eliminated is that of a wound healing response. "Whenever you do any kind of intervention in the eye, you produce a short-term rescue, a physiological response, and that could produce positive results," says Normann.
Building better chips
Helping the pursuit of higher resolution retinal implants is a team of researchers and clinicians at Wayne State Univ., Detroit, Mich., who has already created a 2-mm chip package, which provides a 1,024 array. The package, the Stimulating Array Integrator for Neurological Tissue (SAINT), is the result of a partnership between the College of Engineering and the School of Medicine.
Part of this chip set is a microarray that spatially stimulates the retina (or other neural tissue) and senses neural signals. The stimulation is achieved through a microbump electrode system, which interfaces with neural tissue.
The array is linked to a telemetry chip, which uses magnetic induction for wireless power with a digital overlay for communication. Modifications in the telemetry coil's induced current are used to send information to the reading coil. "Since the reading and telemetry coil are magnetically coupled, the current change can be sensed by the reading system. A reverse communication is coupled to the sending unit," says Greg Auner, director of Wayne State Univ.'s Smart Sensors and Integrated Devices. Thus, the data to create a patterned stimulation in the retina is transmitted as a digital magnetic field perturbation. "This, in turn, controls the SAINT chip to form a patterned stimulation in the neural tissue for a visual perception."
A third generation, 32 x 32, prototype has been built and is currently being tested both in vivo and in vitro .
