Fighting Blindness with Microelectronics

See allHide authors and affiliations

Science Translational Medicine  06 Nov 2013:
Vol. 5, Issue 210, pp. 210ps16
DOI: 10.1126/scitranslmed.3007399


There is no approved cure for blindness caused by degeneration of the photoreceptor cells of the retina. However, there has been encouraging progress with attempts to restore vision using microelectronic retinal implant devices. Yet many questions remain to be addressed. Where is the best location to implant multielectrode arrays? How can spatial and temporal resolution be improved? What are the best ways to ensure the safety and longevity of these devices? Will color vision be possible? This Perspective discusses the current state of the art of retinal implants and attempts to address some of the outstanding questions.


More than two million people are affected by hereditary retinal diseases such as retinitis pigmentosa that result in loss of the rod and cone photoreceptor cells of the outer retina and the retinal pigmented epithelium. The photoreceptor cells transduce light into electrical signals that are processed in the retina and then carried by axons of the retinal ganglion cells along the optic nerve to the visual cortex (Fig. 1A) (1). Although molecular and genetic diagnostic tools can help to identify the exact genetic defect and mechanisms of photoreceptor cell death in the various forms of retinal degeneration, there is no established cure for these blinding diseases. Despite the loss of photoreceptor cells, the inner retina, comprising bipolar cells, amacrine cells, and the retinal ganglion cells, remains intact (2). This has raised hopes that it might be possible to connect light sensor devices by means of electrode arrays to the inner retinal layers, thereby restoring some visual sensation (Fig. 1).

Fig. 1 The visual pathway and locations of different retinal implants.

Shown is the human ascending visual pathway (A) and sites in the retina where epiretinal (B), subretinal (C), and suprachoroidal (D) implants are inserted. (B) With the epiretinal implant, processed signals from an external camera are transmitted to a non–light sensitive microelectrode array that directly touches the fibers of the retinal ganglion cells of the inner retina whose axons form the optic nerve. (C) In contrast, the subretinal implant has the image receiver within the retina in the region where photoreceptor cells have been lost. The light-induced signals from the photodiodes are amplified point by point and sent to the electrodes, which stimulate the bipolar cells of the inner retina directly, resulting in neuronal processing of the signals by the retinal bipolar, amacrine, and ganglion cells. (D) In contrast to the epiretinal and subretinal implants, the suprachoroidal implant is inserted from the back of the eye through a tunnel in the sclera. The electrodes do not touch the retinal nerve cells directly but are positioned in between the sclera and a blood-perfused layer called the choroid. Clinical studies with suprachoroidal implants have been done only with direct stimulation of electrodes in a laboratory.


Early attempts to directly connect electrode arrays to neurons of the visual cortex failed (3). More recent preclinical approaches placing needle-type multielectrodes into input layer 4C of the visual cortex look more promising (4). Cuff electrodes surrounding the optic nerve have been tested in two patients but enabled only very low spatial resolution (5). In the 1990s, several groups connected microelectronic devices to nerve cells in the retina (68), but as the electrodes stimulating these nerve cells were much larger than the fine mosaic of retinal synapses, the spatial resolution was poor (7,8). Cochlear implants containing 16 electrodes are used routinely to treat deafness, but it is technically much more difficult to generate a two-dimensional image transmitted simultaneously via hundreds or thousands of pixels, each one requiring a different signal to its respective retinal neuron. Successfully connecting microelectronic devices to the retina requires that there is both an intact optic nerve and an intact ascending visual pathway. In the future, electrodes implanted directly into the visual cortex may be possible in patients where the visual pathway from the retina to the visual cortex is damaged. However, as long as the visual pathway is intact, retinal implants are preferable and several different designs have been tested in clinical trials with some success.


There are three general ways in which electrode arrays can be connected to the retina (Fig. 1, B through D) (911). In the epiretinal implant, the electrode array is placed at the border between the vitreous humor and the neuroretina, with the electrodes placed on top of the nerve fibers of the retinal ganglion cells (Fig. 1B). In the subretinal implant, the electrode array is placed in the input region of the retina where photoreceptors have been lost (Fig. 1C). In the suprachoroidal implant, the electrode array is placed between the sclera and the choroid (Fig. 1D). Each approach has a principal problem, specifically the need to provide power and control signals to the electrode arrays. For example, transdermal inductive coils, serving as transmitting and receiving antennae, deliver energy and signals through intraocular cables to the electrode arrays of the epiretinal and subretinal implants. This challenge limits the number of electrodes that can be connected serially or in parallel, which then limits the spatial resolution of artificial vision. Attempts to use light falling onto the retina for powering the chip have been made (12, 13), but so far they have not been applied successfully in clinical trials because the light is not strong enough to stimulate cells directly without amplification. Therefore, complex surgical procedures are necessary to implant cables and inductive coils that provide energy and signals to the electrode arrays. Moreover, waterproof, biocompatible encapsulation of intra- or paraocular electronic circuits with very thin material is necessary, and heat production by electronic circuits has to be dealt with. Currently, more than 20 groups are developing various kinds of epiretinal, subretinal, or suprachoroidal implants (14). To date, only two retinal implants are available to patients and have received approval as commercial products: the ARGUS II epiretinal implant and the Alpha IMS subretinal implant.


The electrode array of the ARGUS II epiretinal implant (15) from Second Sight Medical Products Inc. (Sylmar, CA) is implanted on the surface of the retina (Fig. 2, #7) after removal of the vitreous body in an approach used in retinal detachment surgery. The electrode array is attached to the retina with a plastic tack that penetrates the retina and is anchored in the sclera by means of tiny barbed hooks. A foil with wires leads from the electrode array to the receiver electronics located in a capsule on the eye bulb (Fig. 2, #8). A pocket computer with batteries (Fig. 2, #3) translates the video image taken by an external camera (Fig. 2, #1) mounted on a spectacle frame into electrical impulses that are sent electromagnetically by a transmitter antenna (Fig. 2, #4) on the spectacle frame through the periorbital tissue to a receiver coil on the eye bulb positioned next to the episcleral capsule (Fig. 2, #6). From there, a foil with 60 wires leads directly to the 60-electrode array, which is attached to the retinal ganglion cells whose nerve fibers form the optic nerve (Fig. 2, #8, 9). Initially, a total of 30 patients either blind or with bare light perception due to hereditary retinal degeneration received the ARGUS II device in a clinical trial (9, 15). Postoperatively, all patients were able to perceive light. Subjects performed statistically better with the device switched on versus off in the following tasks (15): object localization (96% of subjects), motion discrimination (57%), and discrimination of oriented gratings (23%). The best recorded visual acuity was 20/1260. The mean performance of subjects on orientation and mobility tasks was significantly better when the device was switched on versus off; 70% of patients did not have any serious adverse events. The outdoor mobility of patients increased, as indicated by their ability to walk along pedestrian courses and to identify doors and posts. Second Sight gained European approval (CE Mark) in 2011 for marketing the Argus II as a medical device for treating patients with end-stage retinitis pigmentosa. This year, the Argus II also received approval by the U.S. Food and Drug Administration (FDA) for treating patients in the United States with end-stage retinitis pigmentosa ( The ARGUS II has meanwhile been implanted in a further 21 patients with retinitis pigmentosa in Europe.

Fig. 2 The epiretinal implant.

In the epiretinal approach exemplified by the ARGUS II device (Second Sight, Sylmar, CA), a camera is mounted on spectacle frames (#1). The camera images (pink arrow) are forwarded to a processor (#3) that the patient carries. Signals of the processed image are forwarded to a transmitter antenna (# 4) and are wirelessly transmitted through the periorbital tissue to a capsule (#8). The capsule contains electronic circuits to decode the image signals received by the receiver antenna (# 6) and provides electrical currents to each of the 60 electrodes of the implant (#7), corresponding to the brightness of each image pixel. The epiretinal electrode array receives the signals via an epiretinal foil with printed wires that originates from the capsule. The electrode array is fixed with a tack to the retinal surface and touches the superficial nerve fibers of the retinal ganglion cells that form the optic nerve (#9). The cells of the inner retina (bipolar cells, amacrine cells, horizontal cells, and inner photoreceptor segments) are not used in this approach.


The Alpha IMS subretinal implant (16, 17) is made by Retina Implant AG (Reutlingen, Germany) and was originally developed by our SUBRET consortium at the Universities of Tübingen and Stuttgart. The stimulation electrodes of the subretinal implant are placed in the layer where retinal photoreceptor cells have been lost and connect with the bipolar cells of the inner retina (Fig. 1C). The implant contains a chip with 1500 photodiodes, amplifiers, and electrodes in each pixel, spaced 70 μm apart (Fig. 3, #1). Power and control signals are generated in a small box that the patient carries (Fig. 3, #6). The transmitter antenna behind the ear (Fig. 3, #5) is kept in place by a magnet in the subdermal receiver box (Fig. 3, #3). Power and control signals are sent wirelessly to the subdermal receiver box from which a subdermal cable (Fig. 3, #7) leads to the eyeball, ending in a thin subretinal foil (Fig. 3, #2). The current injected into the retina is picked up by the return electrode (Fig. 3, #4). The chip is positioned on a foil that carries gold wires and is surgically implanted through a flap cut into the sclera. The implant´s tip is advanced transchoroidally beneath the fovea of the retina. In an initial pilot study, 11 patients received a wire-bound version of the device (16); the next 9 patients received a wireless device (17). The subretinal implant restored light perception in 8 out of 9 patients (8/9), light localization (7/9), motion detection (5/9; angular speed up to 35 degrees/second), grating acuity (6/9; up to 3.3 cycles per degree), and visual acuity measured with Landolt C-rings (2/9) up to a Snellen visual acuity of 20/546. Identification, localization, and discrimination of objects improved (P = 0.05 for each subtest) in repeated tests over a 9-month period. Three subjects were able to identify several large letters and combine them into a word (16, 17).

Fig. 3 The subretinal implant.

The subretinal implant exemplified by the Alpha-IMS device (Retina Implant AG, Reutlingen, Germany) is housed within the eye. It consists of a subretinal chip (#1) containing 1500 light-sensitive photodiodes, amplifiers, and electrodes that is placed in the region of the retina where photoreceptors have been lost. The brightness of the image projected by the lens of the eye through the retina onto the chip is analyzed point by point and, after amplification, is forwarded to the bipolar cells of the inner retina. From there, the neuronal network of retinal ganglion cells processes the “electrical image” normally and forwards it to the visual cortex. The electronic circuits in the chip are powered and controlled by a subdermal receiver behind the ear (#3) that receives power and signals wirelessly through the skin via a transmitter antenna coil (#5), which is kept in place behind the ear with a subdermal magnet. As seen in the X-ray image, an implanted cable (#7) leads from the subdermal receiver box to the intraorbital space, where it connects to a subretinal foil (#2) with printed gold wires. The battery box (#6) contains electronics that allow the patient to adjust brightness and contrast to the ambient light of the environment.


The Alpha-IMS implant recently received European approval (CE Mark) to be marketed as a medical device and is under review as an investigational device by the FDA. Meanwhile, a further 19 patients with retinitis pigmentosa have received the Alpha IMS device in an ongoing multicenter trial (Oxford, Hong Kong, London, Tübingen, and other sites).


Vision mediated by electronic retinal implants is different from normal vision. The new window opened by such devices to the world for blind persons is small, not larger than a laptop screen at reading distance. The perceived images reported by most patients consist of slightly flickering shades of grey with blurred borders of objects. Still, for a blind person, this is a great improvement and the overwhelming majority of patients have been happy to regain even this much vision. But there are plenty of opportunities for improvement and also many limitations that still must be overcome.

Spatial resolution. Current retinal implants cannot provide more than approximately 4% of normal vision due to the characteristics of electrical current distribution in multielectrode arrays. Stett et al. (18) have shown with multielectrode arrays in the rodent retina that the space constant, which reflects the spread of excitation picked up by the retinal output neurons, is ~50 to 70 μm. A 280-μm distance on the human retina reflects a one-degree visual angle. Electrode arrays with their current dimensions therefore have a theoretical maximum resolution of ~12 arc minutes (one-fifth of a degree), whereas normal vision readily resolves one arc minute. The excellent visual acuity of normal vision is provided by the high concentration of cones in the fovea. Electrodes cannot replace individual cones, which have a diameter of a few μm and make multiple synapses with the cells of the inner retina. Therefore, the limited spatial resolution of retinal implants is a biological problem rather than the technical problem of building greater density electrode arrays. There are also technical limitations to using high-density electrode arrays, including the heat generated by the wireless transmission of signals to the eye and the electronic circuits attached to the eye. Retinal implant designs where an image receiver, e.g., a camera, is positioned outside (or inside) the eye have to transmit a processed image point by point, that is, electrode by electrode, into the eye. The number of possible electrodes is limited by the number of wireless channels that can be received and decoded within the orbit and transmitted via cables into the eye. For designs with the image receiver outside of the eye, 256 individually accessible electrodes is a big challenge. Higher pixel density is easier to achieve if each pixel is autonomous, that is, if it consists of a light sensor, a processing unit, and an electrode within the eye, as is the case with the Alpha IMS retinal implant, because separate transmission of channels for each electrode is not necessary. On the other hand, due to the narrow subretinal space, encapsulation layers that protect subretinal electronics from body fluids need to be very thin (in the μm range), which is a challenge for material scientists. Longevity of the retinal implant devices is also an issue, as insertion requires surgery and replacing them every few years is not practical. Longevity is greater for the epiretinal implants, as the majority of the device is outside of the eye.

Contrast vision. The ability to discern two independent points depends on the distance of their luminance peaks and the amplitude of the luminance minimum between them. At low luminance, as much as 8 to 10% contrast may be necessary for detecting brightness differences with normal vision. At high luminance, a contrast of less than 1% is sufficient for discrimination in normal vision. This ability is due, in part, to the retina’s on/off system formed of two different pathways that consist of depolarizing and hyperpolarizing bipolar cells of the inner retina that actively signal on or off for light and send their signals separately to the brain’s higher visual centers. Because the specialized synapses at each cone pedicle that create the on/off pathways are lost in retinitis pigmentosa, vision mediated by implanted electrodes has only a limited ability to discern contrasts (currently 15%) (16).

Role of the fovea. In the center of the human retina is the fovea, which has the highest density of cones. The fovea enables us to fixate on an object and to resolve very fine details. The normal retina is built like an inhomogeneous receiver with highly compressed pixels in the center and a much lower density of pixels in the surrounding region. Moreover, although the fovea is only a tiny part of the retina, it feeds visual information into a large part of the primary visual cortex. Therefore, restoring vision in the foveal region is more effective than in the periphery of the retina (17). Future implants may utilize this natural inhomogeneity by having a high pixel density in the foveal region and a low pixel density in more peripheral parts of the retina. This approach would also increase the visual field without reducing spatial resolution in the visual field center.

Color vision. There are three spectrally different types of cones for differentiating chromatic information. Patients who have received subretinal implants typically report that images consist of different shades of grey, possibly a little bit yellowish, but they do not report color vision. Restoration of color vision by electronic implants would not be expected, as the minimal size of planar electrodes required for safe stimulation is >50 μm, whereas the diameter of cone outer segments is only several μm. Thus, all three types of cones are stimulated by each electrode. Even if it became possible to build subretinal implants with three types of spectrally different photodiodes or processors that would analyze the spectral content of a scene and separate the content of the three chromatic primaries individually for each pixel, it would be impossible to properly connect the spectrally specific implant output to the correct retinal input.

In epiretinal implants, the situation is different, as they stimulate the axons or cell bodies of numerous types of color-specific retinal ganglion cells where the center and surround is coded in a chromatically opposing manner, e.g., red on/green off cells, blue on/yellow off cells, green on/red off cells. As the brain has learned to link signals from these cells with respective chromatic information, patients with epiretinal implants often report color sensations. However, as each electrode stimulates a number of these cells simultaneously, the colors are related more or less randomly to various color coding retinal ganglion cells contacted by an electrode and do not reflect the chromatic characteristics of viewed objects.

Binocular vision and spatial resolution. All retinal devices so far have been implanted in only one eye. In principle, it should be possible to put retinal implants in both eyes, but obtaining binocular vision will be challenging. First, the implants need to be positioned in exactly the same location in both eyes, and each electrode of each implant needs to address the same corresponding points on the retina. If the electrode array is implanted in a different region of the retina in each eye, the visual field may be increased, but instead of depth perception double images may occur. Moreover, depth perception depends also on spatial resolution, which is limited with current retinal implants. Therefore, placing a retinal implant in one eye rather than both eyes is currently preferred.

Relearning hand-eye coordination. Most of the patients receiving retinal implants have been blind for many years. Patients have learned to use their auditory and tactile senses to locate objects, but often they have lost the ability to perceive the exact location of a certain object in the inner “imagined” visual space when they touch this object because feedback from the visual system has been absent for years. For a few days after initial activation of retinal implants, many patients, when asked to touch an object that they recognize through their newly acquired “electronic vision,” may try to grasp the object at a fixed angular distance from its actual location. However, we have observed that patients can learn to relate the object’s location in their visual cortex to the motor coordinates for correct grasping of that object within a few weeks.

Visual fields. Current retinal implants cover only a small region of the central retina: 19° by 18° for the epiretinal ARGUS II implant and 11° by 11° for the subretinal Alpha-IMS implant. Clinical experience shows that patients with retinal degeneration who have lost their peripheral retina can still locate objects quite well with a remaining field of 10° by using eye movements and scanning. However, reading requires a visual field extension of at least 5° horizontal and 3° vertical. With the current retinal implants, one to four individual large letters can be identified, but reading text fluently is not yet possible (16, 17, 19).

Temporal resolution. The rod cells of the human eye resolve a repetitive frequency of ~20 Hz, whereas the cone cells resolve repetitive frequencies of 50 Hz or higher depending on the object’s brightness. In the case of electronic implants, temporal resolution is much lower, as the stimulation of the remaining cells of the inner retina is not mediated by neurotransmitter release but rather by directly altering their membrane polarization. After external electrical stimulation, these cells may not return to their resting state immediately, rendering them nonresponsive for a period. Therefore, when the stimulation current is high, temporal resolution is limited and may be as low as 3 Hz. However, in most cases, retinal implants allow an image refreshment frequency of 5 to 20 Hz, although there may be “fading” of images after repetitive stimulation. With epiretinal implants, fading of images can be counteracted by head nodding (20), whereas with subretinal implants the natural microsaccades of the eye prevent fading to some extent (16).

Technical improvements. Supplying power to retinal implants has always been a tricky problem. Initial attempts using passive “solar cells” that directly stimulated inner retinal cells failed to provide vision in clinical studies (12). Newer approaches using passive subretinal elements comprising serially connected (stacked) light-sensitive systems to increase direct voltage output per pixel (13) are more promising. However, this approach needs light amplification in technically complicated laser-driven imaging goggles, which may limit possible application.

Theoretically, spatial resolution might be improved using 3D electrodes. Bionic Vision (BVA, Australia) is testing 3D needle electrodes implanted into retinal bipolar cells and is developing interactive electrode arrangements per pixel to enhance spatial resolution (21). However, implants with penetrating electrodes may damage the delicate retina and will probably face difficulties during explantation due to the danger of retinal detachment. BVA has reported their first clinical trial in three patients with cable-bound suprachoroidal planar electrode arrays. They observed loss of spatial resolution due to the large distance from the electrodes to the retinal nerve cells.

The longevity of retinal devices in the “hostile” environment of the body depends on the materials used to protect and encapsulate the device. The ARGUS II epiretinal implant has been monitored for more than 5 years in patients, and the Alpha-IMS subretinal implant has been monitored for 18 months so far. Newer materials such as liquid crystal polymers (22) look promising in preclinical studies.

Safety. No serious adverse events that have caused life-threatening injury or serious damage to the eye have been reported so far with either epiretinal or subretinal implants. Conjunctival erosion is the most common adverse event, which in most cases can be easily repaired. Increased intraocular pressure or loss of pressure and infection have been reported but have been treated successfully. Rare cases of local retinal detachment have been reported (14). The safety profile of electronic retinal devices so far can be considered good. On the other hand, the stimulation currents used may come close to safety thresholds (23).


Tongue stimulators

For the past 10 years, researchers at the University of Wisconsin have been developing a device that delivers spatially structured input to the tongue via a matrix of electrodes worn inside the mouth (24). Using a camera, a computer, and an input device, blind individuals are able to use this relatively simple and noninvasive device to recognize basic patterns (BrainPort Technology, WIBCAB Inc., Middleton, WI). A 1-year clinical trial with 75 patients is currently under way. In much the same way as blind people can use their finger tips to read Braille, subjects can recognize simple spatial patterns (squares, circles, triangles) using electrical stimulation of the tongue. Due to the electrotactile stimulation, the electrical stimuli on the tongue feel like a tickle or vibration but are not painful. However, the tongue stimulators may hamper speech, which limits their usefulness.

Neuroprotective therapies

Encapsulated cell technology. Various growth factors such as CNTF (ciliary neurotrophic factor) and BDNF (brain-derived neurotrophic factor) can slow down photoreceptor cell loss and boost photoreceptor cell function (25). Neurotech Inc. has developed a capsule filled with CNTF-producing cells, which is inserted into the vitreous of the eye. The capsule releases CNTF over months through small pores, and no immunological response to the cells inside the capsule has been reported. The encapsulated CNTF-producing cells are currently in phase II/III studies in patients with retinitis pigmentosa and age-related macular degeneration (

Electrostimulation. Preclinical studies have shown (26) that ocular electrostimulation using corneal electrodes can slow hereditary retinal degeneration. Electrostimulation releases endogenous growth factors (CNTF, BDNF, and others) that prolong survival of neurons. Corneal electrostimulation with 20-Hz biphasic pulses in patients with retinitis pigmentosa once a week showed statistically significant increases in the visual field area of up to 17% after treatment for 6 weeks (27). Electrostimulation devices have received the CE mark of approval for use for 30 min weekly, but larger studies are needed to explore the persistence of such effects.

Visual cycle modulators. Orally administered synthetic retinoids have been used to replace 11-cis-retinal, a key biochemical component of the visual cycle that cannot be made in patients lacking retinal pigment epithelium-specific protein (RPE65) or lecithin retinol acyltransferase (LRAT), resulting in early-onset retinal degeneration. It has been shown in dogs (28) that the biochemical defect in the retinoid cycle (RPE65 or LRAT deficiency) can be compensated for using synthetic retinoids. Clinical studies with synthetic retinoids have shown the first promising results, with an increase in rod sensitivity (29).

Gene replacement therapy

There are a number of animal and human gene therapy studies using viral or nonviral vectors to deliver genes encoding enzymes such as RPE65 that are mutated in diseases of retinal degeneration, including Leber’s congenital amaurosis (LCA), Stargardt’s disease, and Usher syndrome (30). The goal is to slow or prevent retinal degeneration by restoring the normal protein. The eye is particularly well suited for gene therapy because of its small size and self-contained structure. Adeno-associated viruses are the vectors most often used, and no major adverse effects have been reported in clinical studies (30). The availability of advanced tools for noninvasive clinical monitoring has improved the assessment of the therapeutic effects of gene therapy. After successful preclinical studies in a dog model of LCA (31), safety and efficacy of gene therapy with RPE65 have been established in clinical trials in LCA patients (32). However, it appears that retinal degeneration, albeit slowed, may still continue even as vision improves (33).

Stem cell therapy

Diverse cell populations—such as retinal stem cells, mesenchymal stem cells, or cells derived from Müller glial cells—have been used for retinal transplantation in preclinical animal models. Transplanted cells were shown to differentiate and survive for long time periods in the host retina. Rodent embryonic stem cells (ESCs) differentiated and integrated into retinal ganglion cell layers, plexiform layers, and outer nuclear layers after transplantation. ESCs could also be differentiated into photoreceptor cells in the laboratory, which could then be used for transplantation into the retina (34). In an early phase I clinical trial in two patients with macular degeneration, Schwartz and colleagues (35) performed subretinal transplantation of retinal pigmented epithelial cells derived from human embryonic stem cells in vitro. The transplanted cells integrated into the retinal pigmented epithelium and survived, and small increases in visual acuity were reported. Induced pluripotent stem cells (iPSCs), artificially derived from adult somatic cells by treatment with a cocktail of transcription factors, may be a useful source of cells for transplantation in the future. Although clinical applications are still in their infancy, stem cell therapy approaches show real potential (36).

Optogenetic approaches

Optogenetics is a new technique in which light-sensitive ion channels delivered to neurons by gene therapy are activated by an external light source, resulting in the neuron being switched on. Convincing proof of principle of the application of optogenetics to treating retinal degeneration has been established in animal experiments (37, 38). For example, Lagali et al. (37) showed that engineering “on” bipolar cells in the retina of mice lacking photoreceptor cells to express a light-sensitive channelrhodopsin resulted in light-driven responses in the mice. However, the light sensitivity provided by optogenetics is still several orders of magnitude lower than that of photoreceptor cells. Currently, restoring light sensitivity with high temporal resolution is not possible without light amplification. This could be accomplished using a camera and goggles that would project very bright images onto the retina. But toxicity and long-term stability have not been analyzed sufficiently to justify clinical studies at this time.


Electronic devices such as cardiac pacemakers, cochlear implants, and deep brain stimulators are routinely used in clinical practice. Considerable progress has been made using electronic epiretinal or subretinal implants to restore vision in blind patients. These devices have recently received market approval for treating patients with hereditary retinal degenerative diseases where retinal photoreceptor cells are lost but the inner retina and optic nerve are intact. Certainly, different forms of blindness will require different types of devices and biological therapies. Future technical developments and experience in application of these devices will lead to functional improvements and increased device safety, efficacy, and longevity. Useful vision with electronic devices is now possible, and the excitement created by this success will accelerate the further development of electronic retinal implants for treating inherited diseases of retinal degeneration as well as other forms of blindness.


  1. Acknowledgments: The author thanks the many colleagues involved in the Subretinal Implant Project. E.Z. is supported by the German Federal Ministry of Education and Research, the Bernstein Center for Computational Neuroscience, the Center for Integrative Neuroscience Center of Excellence (University of Tübingen), the Kerstan Foundation, ProRetina Deutschland e.V., and the German Research Foundation. Competing Interests: E.Z. is a co-inventor on patents related to subretinal implants, which were later acquired from the University of Tübingen by Retina Implant AG. E.Z. is an advisor to and holds stock in Retina Implant AG. E.Z. and the University of Tübingen have received grants from Retina Implant AG to perform preclinical experiments and clinical trials.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article