Brain stimulation device lets monkeys ‘see’ shapes without using eyes

By Michael Le Page

Shutterstock / Blueton

Two monkeys are able to “see” and recognise letter shapes generated by arrays of electrodes implanted in their visual cortex rather than relying on light hitting their retina. It is the highest resolution achieved with implants in the brain, rather than the retina.

“That’s really good news,” says Pieter Roelfsema at the Netherlands Institute for Neuroscience, whose team aims to restore some vision to people who have lost their sight.

Many research groups around the world are working on restoring some sight in people who are blind by sending signals from a head-mounted camera to arrays of electrodes that stimulate the appropriate nerve cells. There have been numerous trials in people already, and one 60-electrode device, called the Argus II, was approved for use in the US in 2013.

Advertisement

Most implants, including the Argus II, are designed to be placed in the retina of an eye, but this approach won’t work for people whose optic nerve has been damaged, for instance. So some groups like Roelfsema’s are focusing on the visual cortex instead.

The visual cortex is a bit like a cinema screen in our heads. Each area on its surface maps to the visual field, so activating an A-shaped pattern of electrodes in contact with the visual cortex will, in principle, make people “see” an A-shaped pattern of dots.

However, if electrodes are simply placed on the surface of the visual cortex, a relatively strong current is required to stimulate the nerves below, and it is hard to generate a perception of more than two dots.

Roelfsema and colleagues have instead used arrays of needle-like silicon electrodes that are 1.5 millimetres long. These electrodes are pushed into the cortex so that they make better contact with the nerve cells. The team implanted 16 arrays, each with 64 electrodes, across the visual cortex of two rhesus macaques, for a total of 1024 electrodes in each monkey.

These monkeys had already been trained to recognise 16 letter shapes made from dots on a computer screen and to move their eyes in specific ways in response to each one. They responded in the same way to letter shapes created by the electrodes.

Unfortunately, achieving the same in humans will be harder because the central part of the visual field is deeper within the human brain than in macaques. What’s more, the electrodes would have less and less effect over time as scar tissue builds up around them.

This work is a clear step forward in increasing the number of electrodes, says John Pezaris at Harvard Medical School, but it doesn’t solve the issue of how to reach the central visual field in human brains. “It is in a particularly challenging location to access surgically,” he says.

Roelfsema hopes both these issues could be solved by using very thin, flexible electrodes, inserted using rods that are then withdrawn. These could be inserted more deeply into the brain, and should last longer too.

“The idea is that they would have much better longevity, that they would stay in operation for many years,” says Roelfsema.

Even if this approach succeeds, a device with 1000 electrodes won’t come close to matching the resolution of human vision – our eyes have the equivalent of a million pixels. Nor it is yet possible to control colour or depth perception.

“High-fidelity artificial vision through cortical stimulation is a difficult goal,” says Pezaris. “One thing that we are learning as a field is that our idea that any level of artificial vision is better than being blind, and therefore crude devices are worth developing, is not shared by the blind community.”

Journal reference: Science, DOI: 10.1126/science.abd7435