Developing stimulation strategies for the bionic eye

Bionic Vision Australia is developing the Bionic Eye to target two degenerative retinal diseases: Retinitis Pigmentosa (RP) and Age-related Macular Degeneration (AMD). Both diseases result in loss of photoreceptors, the light-sensitive cells in the eye. Electrical stimulation of the remaining cells in the retina, usually retinal ganglion cells, can provide a sense of vision, where the perceived “light” is called “phosphenes”.

Stimulation strategies for bionic eyes typically encode image brightness by stimulating different electrodes with different amounts of electric current.  The resulting phosphenes are ideally round in shape, though the shapes can vary from electrode-to-electrode and from patient-to-patient.  It may be possible to control phosphene shapes by careful choice of electrodes to stimulate at any given time. In particular, elliptical phosphenes may be possible, giving the ability to directly present edge information to patients. This information about the edges of objects is known to be important to perceive the outline of objects.

Based on this, Isabell Kiral-Kornek undertook research in one of the psychophysics booths in the Centre for Neural Engineering on a novel stimulation strategy for the bionic eye that extracted edge information from images and encoded them using oriented elliptical phosphenes. She tested whether encoding edge orientation using elliptical phosphenes could lead to better alphabetic letter recognition than standard brightness-based phosphene encoding.

Isabell conducted a study with simulated phosphene vision with 12 normal-sighted volunteers. The picture shows an example of a volunteer reading some text off an iPad with virtual reality goggles; the computer screen shows what the volunteer is seeing at the instant the photo was taken.

The two different stimulation strategies were compared as Isabell varied several parameters including letter sizes and distances between phosphenes. She found that the average letter recognition accuracy was significantly better with the proposed elliptical phosphene strategy (65%) that standard brightness encoding (47%).

This research holds promise for assisting bionic eye users in the future to improve their vision performance.