A CBET Nugget Notable Accomplishments from CBET Awards

An Intraocular Camera for Retinal Prostheses Mark Humayun, Doheny Eye Institute - Affiliated with the University of Southern California, Keck School of Medicine

Background:  Loss of vision, and of the human experience and capabilities associated with vision, affects upwards of a hundred million people worldwide.  In two widespread causes of blindness, Age-Related Macular Degeneration (AMD) and Retinitis Pigmentosa (RP), the photoreceptor cells (rods and cones) within the retina of the eye progressively degenerate, leaving the retina insensitive to light.  Surprisingly, the inner retinal cells (including the retinal ganglion cells that provide output from the eye to the visual cortex within the brain), retain their viability, and can be electrically stimulated to provide the sensations of vision (visual percepts). At the University of Southern California (USC), researchers are developing an intraocular retinal prosthesis containing an array of tiny electrodes (a microstimulator array) that can be surgically attached to the inner surface of the retina.  When provided with electrical inputs derived from a camera recording an external scene and passed through the microstimulator array, the inner retinal cells can transmit the recorded image to the brain, thus providing prosthetic vision.  Currently, the retinal prosthesis is configured with an external camera mounted on a pair of eyeglasses.  Although functional, this configuration results in rapid head motion to look in different directions, as one would crossing a room or a road, which in turn can lead to disorientation and dizziness. Recently, a USC team of investigators has invented a miniature intraocular camera that is small enough to fit within the crystalline lens sac behind the cornea, and that can be electrically connected to the microstimulator array to provide a fully-implantable intraocular retinal prosthesis.  With such a device implanted, blind patients would not only have partial restoration of their vision, but also would be able to move their eyes naturally to scan the local environment, potentially greatly enhancing their mobility and capabilities. Results:  Extensive human psychophysical studies by the investigators and their students at USC have shown that the number of pixels, or picture elements, needed for most common tasks is about 625 (for example arranged in a 25 x 25 array), far lower than might be imagined.  Furthermore, and equally surprising, the investigators discovered that significant blurring of the resulting square-pixellated image improves rather than degrades visual perception, including the recognition of faces, locations, and common objects.  Taken together, these two psychophysical results significantly relax the design constraints on the envisioned intraocular camera, making an extremely compact design with a very short focal length (less than 2 mm) feasible.  By comparison, the human eye has a focal length of approximately 17 mm. The USC team also discovered that the miniature intraocular camera has tremendous depth of field, alleviating the need for an accommodative mechanism (such as the human eye employs) for focusing on objects both near and far away.  In fact, objects can be held very close to the eye without significant defocusing, allowing for enhanced magnification of small objects and features (like the numbers on a house key).  To date, several prototype intraocular cameras have been successfully fabricated.  The current prototype will employ an aspherical lens design, as well as lightweight optical materials and a miniature image sensor array.  The entire intraocular camera is designed to fit within the crystalline lens sac, and to be implanted using surgical techniques similar to those employed for the implantation of intraocular lenses following cataract removal.

A.  Image of a kitchen scene, 1024 x 1024 pixels. Permission Granted Credit: Armand R. Tanguay, Jr. and Noelle R. B. Stiles; University of Southern California

B.  Square pixellated image of a kitchen scene, 25 x 25 pixels. Permission Granted Credit: Armand R. Tanguay, Jr. and Noelle R. B. Stiles; University of Southern California

C.  Square pixellated image of a kitchen scene, 25 x 25 pixels with 34% Gaussian blur, showing significantly enhanced object recognition. Permission Granted Credit: Armand R. Tanguay, Jr. and Noelle R. B. Stiles; University of Southern California

Scientific Uniqueness:  The intraocular camera for retinal prostheses is unusual in that it is a fully biomimetic device, designed to use both the human (biological) corneal lens in conjunction with an implanted aspherical (inorganic) lens to provide the imaging function.  Furthermore, human visual psychophysics techniques have been used to discover the somewhat counterintuitive value of image blurring (at the point of stimulation of the retinal tissue) in enhancing object recognition, which in turn yields greatly relaxed design constraints for the intraocular camera itself.  In addition, the optical system design yields an expanded depth of field that eliminates the need for an accommodative mechanism to focus on near and distant objects. Impact on Industry and/or Society:  The successful development of a fully implantable intraocular retinal prosthesis would enable partial restoration of vision to those affected by the widespread blindness-inducing conditions of Age-Related Macular Degeneration and Retinitis Pigmentosa.  The incorporation of an intraocular camera is key to developing a fully implantable device that does not rely on connection to an external camera, and would in addition provide full use of natural eye movements in conjunction with greatly reduced head motions to acquire and track objects of interest within the natural environment. Work is notable because the development of an intraocular camera for use in conjunction with a microstimulator array to provide a visual prosthetic device represents a potentially significant advance in providing vision to the blind.  The development of such a miniaturized camera that can be implanted within the human eye would enable the entire intraocular retinal prosthesis to be fully implanted, yielding a prosthetic approach that would appear natural and not rely on coupling to potentially cumbersome or unsightly external devices.  Although several types of prostheses are available for various joints and body parts, an intraocular retinal prosthesis would represent only one of very few prostheses that interface to the central nervous system, as does the highly successful cochlear implant for those with loss of hearing. This work is multidisciplinary.  The intraocular retinal prosthesis research project in general, and the intraocular camera research project in particular, have required close cooperation among research scientists and medical doctors spanning a wide range of disciplines, including visual psychophysics, ophthalmology, optics, photonic technology, image processing, physics, electrical engineering, mechanical engineering, biomedical engineering, surgery, and neuroscience.  Graduate students, undergraduate students, and high school students have provided key breakthroughs at every stage of the project, and have been exposed to a broad range of theoretical, numerical, and experimental techniques across this range of disciplines. This nugget addresses the strategic outcome goals as described in the NSF Strategic Plan 2006-2011 as follows:  (1)  Discovery.  This research is at the cutting edge of a potentially viable biomimetic approach to at least partially restore vision loss by developing a fully implantable intraocular retinal prosthesis.  Prostheses for the central nervous system (vision, hearing, brain function) are at the very frontier of biomedical engineering. (2)  Learning.  Graduate students, undergraduate students, and high school students have greatly benefited by their participation in this highly multidisciplinary research project.  Students have been exposed to a broad range of theoretical, numerical, and experimental techniques across a wide range of disciplines (including visual psychophysics, ophthalmology, optics, photonic technology, image processing, physics, electrical engineering, mechanical engineering, biomedical engineering, surgery, and neuroscience).  Intensive learning is enhanced by weekly all-hands meetings and intensive seminars, the availability of common laboratories, involvement in collaborative experiments, and attendance at key surgical procedures and grand rounds. (3)  Research Infrastructure.  Research on the intraocular camera for retinal prostheses is in large part carried out in newly configured common laboratories that span the entire intraocular retinal prosthesis effort at USC.  Located on the Health Sciences Campus of USC, these common facilities encourage multidisciplinary collaboration, deep understanding of the entire research effort, and novel experimental approaches based on many disciplines. This Nugget represents transformative research.  Successful development of a fully-implantable intraocular retinal prosthesis could provide partial vision restoration to a very large population of those blinded by widespread conditions such as Age-Related Macular Degeneration and Retinitis Pigmentosa.  Key to full implantability is the development of a highly miniaturized intraocular camera.  No viable alternative therapeutic technique is currently available to restore lost vision.

Program Officer:		Leon Esterowitz
NSF Award Number:		0201927
Award Title:		Biocompatible Technology for a Light Sensitive Retinal Prosthesis
PI Name:		Mark Humayun
Institution Name:		Doheny Eye Institute - Affiliated with the University of Southern California, Keck School of Medicine
Program Element:		7236

Top of Page

This Nugget was Approved by ENG on 22 February 2007.