Twenty years after Dean Lloyd's world slipped into total darkness, his brain is learning to see what his eyes cannot.
A camera in Lloyd's glasses is focused on UCSF researcher Jaque Duncan. As she moves, her image is being sent through a video processor. From there, an experimental technology will deliver it to the back of Lloyd's right eye, not as a complete picture, but as flashes of light.
"We try to connect the stars to get some sort of an image," said Lloyd.
"The device allows him to see very high contrast, large targets relatively well," said Duncan.
Located in the back of his eye is a surgically implanted strip containing dozens of electrodes. Dubbed the Argus-2, it relays electronic impulses from the camera on Lloyd's glasses directly to the neurons, essentially bypassing the eye.
"So the beauty of it is that it uses the patient's own wiring and connective neurons to get the message to the brain," said Duncan.
The device functions something like an artificial retina, but the images it sends are complicated to interpret. In fact, patients need to literally retrain their brains.
Working with a computer image in a darkened room, Lloyd will exercise his brain's ability to follow the shifting light patterns, ultimately interpreting them in his mind as shapes and movement.
"For example, if there is a contrast of dark into light area, then I know I've got a border," said Lloyd. "Usually I put my hand to the doorway to see what kind of a thing I'm dealing with."
"He can see very bright white lines on a black surface or he can see a big black door in a white wall," said Duncan.
Currently, Lloyd is one of just 10 people in the U.S. fitted with the implant, and after more than a year of training, he is now able to distinguish some colors and shades of gray and white.
It is a beginning he believes could someday allow others to effectively regain their site.
"I rather enjoy being a part of this project as much as the benefit I receive for myself," said Lloyd.
The device that Lloyd had implanted contains about 60 electrodes. The company says future versions could contain thousands, potentially producing far more detailed images.