This is an excerpt from the GDI study "Robotics and Disabilities". Download the complete study here.

Approximately three million people worldwide, in Germany alone about 30 000 – 40 000, suffer from the degenerative retinal disease Retinitis pigmentosa. The disease manifests itself in a slow darkening of vision, beginning with difficulties seeing well at night, through to complete blindness.

One possible procedure is to position a chip surgically on the retina. The patient wears glasses with a camera that transmits the images to the chip (interview with a person affected). The chip electrically stimulates the ganglion cells under the photoreceptors. These then pass the camera image on to the brain through the optic nerve. Such interventions are possible in the first years after becoming blind. The longer one waits, the more the nerves within the retina break down and the less receptive they are for the electrical stimulation through the chip.

But we must not imagine this as natural vision: The implant “Argus 2”, for example, has a resolution of 6x10 pixels. This means that the image you see consists of only 60 points (if all 60 electrodes are also properly received by the retina).
 
With a retina implant, it is therefore possible to see outlines at most. You can identify door frames, light sources, perhaps individual letters. How well an individual can actually implement
this relatively modest visual performance may be very different from person to person and is difficult to estimate. This is in contrast to, for example, a cochlear implant, a hearing prosthesis that functions as an artificial inner ear, which works about the same way for most users.

The fact that success cannot be predicted is problematic. On the one hand, eye surgery is not without risk. On the other hand, such an implant, including insertion, individual adaptation and the very important thorough training of the patient costs about 100 000 Euro. In Switzerland, in contrast to Germany and France, this is not paid for by health insurers. After insertion, each of the 60 electrodes must be calibrated separately. This means that the usable bandwidth of the signal strength must be elicited individually for each electrode.

Dr. Jörg Sommerhalder, who conducts research at the University Hospital of Geneva, is becoming increasingly sceptical of retina implants. “From 2000 to 2010 significant progress was certainly made, but since then it has been slow,” says the physicist, who is himself involved in this research. “The manufacturers appear to have temporarily come up against an obstacle.” At the moment, efforts are being made to extract improvements from the software, for example, to increase the contrast in image processing, so that outlines are clearer.

Despite the slow progress, the companies are now compelled to go to the market so that they can get back the many investments in research. This leads to a few tear jerking marketing videos of people who appear absolutely thrilled, alleging on camera that they “can see their own wife again for the first time.” Whether they could distinguish their own wife from another person is questionable.

Whether and when retina implants will overcome the technological and biological barriers that are impeding them at the moment is unclear. It is also conceivable that a different approach will win through, such as making the visual environment audible by an artificial intelligence. Or science comes up with a cure for degenerative retinal diseases. This example makes it generally clear that not every technology automatically gets better with increasing speed. Further developments may not build on a refinement of the existing one, but require a groundbreaking new idea.

Download the complete study here.