The target: a retina that has lost its sensors
In inherited diseases such as retinitis pigmentosa, the rods and cones that sense light gradually die, but many of the cells further along the retina survive for years. That observation is the basis of retinal implants. If the lost sensors can be replaced by electrodes, or by tiny light-sensitive chips, the surviving cells may carry signals to the brain.
Developers had to solve problems at every level: how to place a fragile array on a curved, thin tissue; how to deliver current without damaging cells; how to seal electronics in a salty, warm environment for years; and how to get power and images into the eye without a bundle of wires through its wall.
Epiretinal designs: electrodes on the inner surface
The epiretinal approach places an array on the surface of the retina facing the middle of the eye. It is often combined with an external camera on glasses, a small processor worn on the body, and a wireless link to a receiver that sits on the eye.
Work on this approach in the United States, led by Mark Humayun and colleagues in Los Angeles, with Eugene de Juan and others, began in the 1990s. A first-generation device with 16 electrodes was implanted in a small number of volunteers from around 2002. A second-generation design with 60 electrodes, known by its commercial name Argus II, was tested in a clinical trial that began in the late 2000s. It received a CE mark in Europe in 2011 and approval in the United States in 2013 under a pathway intended for devices serving small patient groups.
The design used a metal tack to hold the array against the retina and a cable running to electronics outside the eye wall. Trials reported that many participants could detect light, locate bright objects and sometimes follow lines, though the quality of vision remained very limited. Surgery was complex, and trial reports listed adverse events such as conjunctival erosion and low eye pressure.
Subretinal designs: chips beneath the retina
The subretinal approach places a light-sensitive chip where the lost photoreceptors used to be. Its appeal is that the chip receives light directly through the eye’s own optics, so the person can use natural eye movements rather than turning the head toward a camera.
Optobionics, a company founded by Alan and Vincent Chow, implanted an “artificial silicon retina” in patients from the year 2000. It was a small chip carrying thousands of microscopic photodiodes. Participants reported some improvements, but the chip produced very little current, and later discussion suggested that any benefit might have come from a general stimulating effect on the retina rather than from the chip acting as a camera. The company eventually stopped operating.
A team at the University of Tübingen led by Eberhart Zrenner developed a more powerful subretinal chip with around 1,500 light-sensing elements, each with its own amplifier, powered through a cable passing behind the ear. Under the name Alpha IMS and its successor, it received a CE mark in 2013. Some participants could identify large letters and objects in trials, and results varied widely. The company behind it later ceased operations around 2019.
A newer approach, called PRIMA, uses tiny photovoltaic implants beneath the retina that are powered by near-infrared light projected from glasses. It was designed mainly for atrophic age-related macular degeneration. Trial results were reported from about 2025, with some participants able to read letters through the system; its regulatory status and long-term performance should be checked against current sources.
Suprachoroidal designs
A third approach places electrodes in the space between the choroid and the sclera, outside the retina. Surgery in this location may be less risky because the array never enters the eye itself, but the electrodes are farther from the cells they are trying to excite, so more current is needed and the resulting perception is less precise. Groups in Australia, Japan and elsewhere ran early trials with this design.
Engineering problems that kept recurring
Several problems appeared repeatedly.
- Resolution. Each electrode stimulates a patch of cells, not a single point, and neighboring electrodes can overlap or distort the picture. Adding electrodes does not automatically add resolution.
- Axon stimulation. Current on the inner retina can activate nerve fibers passing underneath, producing streaks rather than dots.
- Longevity. Corrosion and moisture ingress in the eye’s environment can break sealed electronics over time.
- Fading perception. In some users, the percepts dim during continuous stimulation, which led to research on better pulse patterns.
- Fixation. The array must remain in the right place against a moving eye for years.
As in the development of other implantable devices, these issues are as much about materials and packaging as about electronics. Advances in silicon chips and microfabrication supported the push to smaller and denser arrays, while diagnostic tools similar to those used in routine retinal examination helped surgeons plan the placement and check the result.
What the trials showed and did not show
Clinical trials for these devices were small, because the eligible population is small. Outcomes were measured with tests such as locating a light square on a screen, following a line or identifying large letters. Many participants improved on these tests while the device was on, and some reported benefits in orientation or in finding doors and people. Others gained little.
Researchers have also noted that these tests do not necessarily match daily life, and that reports of quality-of-life benefits are harder to measure than test scores. The approved systems did not restore anything close to normal sight.
A field in transition
The collapse of some of the companies, and the difficulty of supporting implanted devices without a commercial maker, gave a cautionary lesson about how long-term technology support matters for people who receive an implant. Meanwhile, parts of the field turned toward alternatives such as gene editing and other gene therapies for inherited retinal disease, cortical implants and optogenetics. The retinal implant remains a milestone in neural engineering, but its place in future care is still uncertain.