Waiting for Vision

The prospect of developing and marketing a retinal prosthesis that restores at least some vision to blind individuals is one of the most exciting opportunities ever to confront the neurotechnology industry. Make no mistake—building a functional device will not be easy and the schedule is bound to slip as technical and organizational complications arise.

But for perhaps the first time, there is a widspread belief among the neural prosthesis community that it is at least feasible to construct a retinal implant. What no one knows for sure is how good the quality of vision will be once the first generation of devices beomes available. No amount of testing, modeling, simulation, or projection can adequately adderess the essentially subjective experience that accompanies human vision. Only when a sufficient number of users have been implanted with a device and report back on the quality of their visual experience will we really know if our goal is months or decades away.

As frustating as it may be to have to wait for user reports, the situation offers us some opportunities. First, we should begin to ponder ways of maximizing the perceived resolution achievable from a particular fixed resolution of stimulating electrodes. In the past 10 years, the digital imaging industry has accomplished just that for hard copy devices, image sensors, displays, and software. It may well be that a 25 by 25 electrode array could produce a perceived resolution of 100 by 100, if imaging techniques such as edge enhancement, dithering, and anti-aliasing are used.

Second, knowing that vision is a subjective experience should make us mindful that no device can by itself produce an effect that is consistent across subjects and at different times. The user’s memories, expectations, biases, and motivations will color—literally—the mental picture that emerges after the ganglion cells receive their input. What this means is that we should take pains to build in user input, feedback, and preference settings into whatever interface we provide to the device. Allowing the user to experiment with as many parameters as is practicable seems like a good idea here.

Third, creating a cellular environment where electrode and neural tissue can coexist harmoniously will only benefit performance and stability characteristics. This may mean looking at new coatings, electrode geometries, and surface characteristics of our stimulating electronics.

Finally, devising performance goals and standards into our first generation of devices will help us gauge our progress and guide our subsequent generations of devices. It will also help us communicate more realistic goals and timetables to the public, potential user base, and investment community. A user who’s expecting to see just a small subset of what’s stored in visual memory is likely to be more satisfied with any device than someone who’s expecting a return to normal.

In the end, a successful retinal prosthesis will do more than just restore some visual capabilities to blind people. It will tell us much about how the human sensory system is driven by external stimuli and it may well produce spinoff technology in machine vision. pattern recognition, and biomimetic neural networks. And that’s a sight that’s well worth the wait.

James Cavuoto
Editor and Publisher