The Retinal Prosthesis for the Blind

October, 2012

Shawn K. Kelly, PhD

VA Center for Innovative Visual Rehabilitation

Carnegie Mellon University

The Retinal Prosthesis

An electronic implantable device to restore functional vision to patients with certain

forms of blindness, primarily retinal degenerative diseases.

The device works by stimulating nerves in the visual system based on an image from

an external camera.

Anatomy of the Eye

Anatomy of the Retina

Direction of Light

Inside/Front

Outside/Back

Sending Images to the Brain

Degenerative Diseasesof the Outer Retina

• Age-related Macular Degeneration (AMD)– Loss of photoreceptors in macula (center),

working outward

– Strikes at age 60-80 with increasing incidence– May take decades to go blind– 2 (?) million cases in US, tens of millions

worldwide

Degenerative Diseasesof the Outer Retina

• Retinitis Pigmentosa (RP)– Loss of photoreceptors in periphery, working

inward

– Genetic, strikes at age 15-60– Typically a decade or less to go blind– 100,000 cases in US, 1.7M worldwide

Living with RP

The Boston Retinal Implant

• Electronic prosthesis to restore functional vision to patients with retinitis pigmentosa and age-related macular degeneration

• 20-year collaboration: MIT, MEEI, VA, Cornell, Carnegie Mellon, and others

Subretinal Implant PlacementD

irection of Light

Inside/Front

Outside/Back

Retinal Prosthesis Function• Electrically stimulates ganglion nerves based on an

external camera image

ImageProcessing

PowerAnd DataTransmission

Other Visual Prostheses

Epiretinal Cortical

Visual ProsthesisResearch Worldwide

Optic Nerve

Cortical

Retina Epi

Retina Sub

Challenges

• Surgical– Place electrode array safely and securely– Place electronics where they won’t cause damage

• Microfabrication– Electrode materials that can safely deliver needed charge

but not physically damage tissue– Water-tight packaging of electronics

• Electrical Engineering– Deliver balanced electrical current to electrodes– Wirelessly deliver power and image data to implant

• Image Processing– Intelligently downsample camera images to 256 pixels– Convey saliency, maximum information about environment

Short-term HumanProof-of-concept Trials

• 1998 – 2000, Surgical trials on six volunteers

• Epiretinal stimulation for a few hours

• Reported spots, lines, not complex shapes

Rizzo, et al. IOVS, 2003 Rizzo et al. IOVS, 2003

Portable, 100-channelNeural Stimulator

Short-term HumanProof-of-concept Trials

Video - surgery

Short-term HumanProof-of-concept Trials

Video - clouds

Short-term HumanProof-of-concept Trials

Video - spots

What We Learned

• This concept works – patients can see spots and lines, distinguish from control

• We need to develop a chronic implant to allow patients to learn

• A number of new technologies needed to be developed for chronic

implantation

Wireless Power and Data

Power and data are delivered by inductively coupled coils via magnetic fields

Miniaturize:Integrated Circuit Design

Custom chip design is labor-intensive, but necessary given our size and power

restrictions

Kelly, et al. IEEE ISSCC, 2004 Theogarajan et al. IEEE ISSCC, 2006

Kelly, et al. IEEE TBioCAS, 2011 Theogarajan IEEE JSSC, 2008

First-Generation Implant• Implanted in 3 minipigs for up to 10 months in

2008• Wireless power and data telemetry• Coated in silicone – viable for many months, not

decades

Shire, et al. IEEE TBME, 2009

Improvements to First Generation Implant

• Hermetic barrier to protect circuits

• Larger telemetry coils

• Easier surgical access for electrode insertion

Second Generation ImplantAb externo approach

• Electrode array enters the space under the retina through the scleral wall of the eye.

Hermetic Titanium Case

• Ceramic 19-pin hermetic feedthrough

• Curved titanium case

• Laser-welded top and bottom lids

Sputtered Iridium OxideFilm Electrodes

• Higher charge capacity• 1 to 6 mC/cm2

• Pt is limited to 100 µC/cm2

• But the Shannon Limit is still in effect!

Prototype Implant

In vivo Results

• Recorded artifact of current pulses from cornea – shows implant was working

• Waveform is measured at implant and telemetered out

• Noisy, but you can see the step-ramp components, and variation of voltage with current

Kelly, et al. IEEE ISABEL, 2009 Kelly, et al. IEEE TBME, 2011

Kelly, et al. IEEE EMBC, 2009 Kelly, et al. BSPC, 2011

Third Generation Prosthesis• Hundreds of channels (>256)

• Smaller hermetic case

• Stimulator chip and electrode array under development

Image Processing

Future Research

• Finish 256+-channel retinal prosthesis, prepare for FDA clinical trials

• Design external camera system, portable telemetry system, etc.

• Image processing to convey maximum information with 256 pixels.

The Boston RetinalImplant Project

Engineering• John L. Wyatt, PhD (MIT)• Douglas B. Shire, PhD (VA, Cornell)• Shawn K. Kelly, PhD (VA, CMU)• Ashwati Krishnan, MS (CMU)• Marcus Gingerich, PhD (VA, Cornell)• Attila Priplata, PhD (VA, MIT)• William Drohan, MS (VA, MIT)• Bruce McKay, BS (MEEI, Cornell)• Oscar Mendoza (MIT)• Carmen Scholz, PhD (Alabama)

• Stuart Cogan, PhD (EIC Biomedical)• William Ellersick, PhD (Analog Circuit Works)• Sonny Behan (Sonny Behan Consulting)

Medicine and Biology• Joseph F. Rizzo, MD (VA, MEEI)• Jinghua Chen, MD, PhD (MEEI)• Hank Kaplan, MD (Louisville)• Vasiliki Poulaki, MD (MEEI)• Shelley Fried, PhD (VA, MGH)• Ralph Jensen, PhD (VA)• Ofer Ziv, PhD (VA, MIT)• Lotfi Merabet, OD, PhD (VA)

Thanks to:Department of Veterans Affairs, MOSIS, NIH, NSF, DoD, Catalyst Foundation

Discussion