The late neuroscientist Paul Bach-Y-Rita hypothesized in the 1960s that “we see with our brains not our eyes.” Now a noninvasive device trades on that thinking and aims to partially restore the experience of seeing for the visually impaired by relying on the nerves on the tongue’s surface to send light signals to the brain.

First demonstrated in 2003 by neuroscientists at Middleton, Wis.–based Wicab (a company co-founded by Bach-y-Rita), the device could finally be ready for sale at the end of the year. Called BrainPort, it tries to substitute for the two million optic nerves that transmit visual signals from the retina to the brain’s primary visual cortex. Visual data are collected through a small digital video camera mounted on the center of sunglasses worn by the user. Bypassing the eyes, the data go to a handheld base unit, which houses such features as zoom control and light settings as well as a central processing unit (CPU), which converts digital signals into electrical pulses.

From the CPU, the signals are sent to the tongue via a “lollipop,” an electrode array about nine square centimeters that sits directly on the tongue, which seems to be an ideal organ for sensing electric current. (Saliva is also a good conductor.) Moreover, the tongue’s nerve fibers are densely packed and are closer to the surface relative to other touch organs. The surfaces of fingers, for example, are covered with a layer of dead cells called the stratum corneum.

Each electrode on the lollipop corresponds to a set of pixels. White pixels yield a strong electrical pulse, whereas black pixels translate into no signal. The nerves at the tongue surface receive the incoming electrical signals, which feel a little like Pop Rocks or champagne bubbles to the user.

Typically within 15 minutes of using the device, blind people can begin interpreting spatial information via BrainPort, says William Seiple, research director at the nonprofit vision health care and research organization Lighthouse International. The electrodes spatially correlate with the pixels so that if the camera detects light fixtures in the middle of a dark hallway, electrical stimulations will occur along the center of the tongue. “It becomes a task of learning, no different than learning to ride a bike,” says Wicab neuroscientist Aimee Arnoldussen, adding that the “process is similar to how a baby learns to see. Things may be strange at first, but over time they become familiar.”

Seiple works with four patients who are training with BrainPort once a week. He notes that his patients have learned how to quickly find doorways and elevator buttons, read letters and numbers, and pick out cups and forks at the dinner table without having to fumble around. “At first, I was amazed at what the device could do,” he says. “One guy started to cry when he saw his first letter.” The researchers have yet to figure out if the electrical information is transferred to the brain’s visual cortex, where sight information is normally sent, or to its somatosensory cortex, where touch data from the tongue are interpreted.

To develop criteria for monitoring the progress of artificial sight, optometrist Amy Nau of the University of Pittsburgh Medical Center’s Eye Center will further test BrainPort, along with other devices such as retinal and cortical implant chips. “We can’t just throw up an eye chart. We have to take a step back and describe the rudimentary precepts that these people are getting,” she says. Nau is particularly interested in BrainPort because it is noninvasive, unlike implants.

“Many people who have acquired blindness are desperate to get their vision back,” she points out. According to the National Institutes of Health, at least one million Americans older than 40 are legally blind, with vision that is 20/200 or worse or that has a field of view of less than 20 degrees. Adult vision loss costs the country about $51.4 billion a year.

Although sensory substitution techniques cannot fully restore sight, they do provide the information necessary for spatial orientation. Wicab had planned to submit BrainPort to the U.S. Food and Drug Administration for approval at the end of August, says Robert Beckman, president and chief executive officer of the company. He notes that the device could be approved for market by the end of 2009 for about $10,000 a machine.