Photovoltaic Retinal Prosthesis for Restoration of Sight in Retinal Degeneration

Retinal degenerative diseases lead to blindness due to loss of the “image capturing” photoreceptors, while neurons in the “image-processing” inner retinal layers are relatively well preserved. Information can be reintroduced into the visual system using electrical stimulation of the surviving inner retinal neurons. We developed a photovoltaic subretinal prosthesis which converts light into pulsed electric current, stimulating the nearby inner retinal neurons. Results of the clinical trial with our implants (PRIMA, Pixium Vision) having 100um pixels, as well as preclinical measurements in rodents with 75 and 55um pixels, confirm that spatial resolution of prosthetic vision can reach the sampling density limit.

For a broad acceptance of this technology by patients who lost central vision due to age-related macular degeneration, visual acuity should exceed 20/100, which requires pixels smaller than 25um. Radial expansion of electric field in flat arrays precludes scaling the pixels to such small dimensions. We are working on 3-dimensional electro-neural interfaces which should enable such a high resolution, and may even reach single-cell selectivity.

Recent 12-minutes talk summarizing the project

System Design

In our photovoltaic system, data stream from a video camera is processed by a pocket PC, and the resulting images are displayed on the augmented-reality glasses. Images are then projected from the microdisplay onto the retina using pulsed (1-10 ms) near-infrared (~880 nm) light. These light pulses are photovoltaicly converted into bi-phasic pulses of electric current flowing between the active and return electrode in each pixel, which stimulate the nearby inner retinal neurons, and thereby introduce visual information into the retinal neural network.

Optical delivery of the information and power allows for simultaneous activation of thousands of pixels in the implant, and retains the natural link between the eye movements and visual perception. Since each photovoltaic pixel operates independently, they do not need to be physically connected to each other. Thus, small (1-2 mm) modules can be separately placed into the subretinal space to tile a large visual field, greatly simplifying surgery.

Animation illustrating the concept

Image on the right shows a 1-mm wide and 30 um thick array implanted subretinally in a rat eye. SEM demonstrates a high magnification of the array with 55 um pixels placed on retinal pigment epithelium.

Color insert on the right shows a single pixel in the hexagonal array. Each pixel includes 2 photodiodes connected in series between the central active electrode (brown disk) and the circumferential return electrode made of SIROF.

We study the mechanisms of neural stimulation and characteristics of prosthetic vision ex-vivio and in-vivo, and optimize the system to enable high resolution prosthetic vision. These studies include modeling of electric field in tissue, ion channel dynamics, electrode-electrolyte interface, circuit performance, fabrication of the implants, and electrophysiological assessement of the retinal, cortical and behavioral responses to visual stimuli. We also participate in the design and data analys of the clinical trials of our PRIMA system manufactured by Pixium Vision.

3-dimensional electro-neural interface

High-resolution retinal prostheses require small, densely packed pixels, but limited penetration depth of the electric field formed by a planar electrode array constrains such miniaturization.


We developed a novel 3-D honeycomb confguration of an electrode array with vertically separated active (red) and return (blue) electrodes. This geometry is designed to leverage migration of the inner retinal cells into voids in the subretinal space. Insulating walls surrounding each pixel align the electric field vertically, thereby decoupling the field penetration depth from the pixel width. Alignment of electric field along the bipolar cells reduces the stimulation threshold and enables scaling the pixel size down to cellular dimensions.

Confocal microscopy demonstrating 3-D view of the retina integrated with the noneycomb implant. Inner nuclear layer fills the wells, while the inner plexiform layer and ganglion cells remain above the implant.


Retinal cells from the inner nuclear layer migrated into the honeycomb wells within 6 weeks of implantation.