As part of a neuronal relay race, electrical signals travel from the brain to the muscles that control movement. Break any neuronal connection in the chain — through injury or disease — and the control may be lost.

Eberhard Fetz and his colleagues at the University of Washington, Seattle, attempted to use a small computer chip to replace these lost connections. But in experiments described on page 56, they found that their device could also serve to strengthen and rewire existing ones.

The collaboration between Fetz, Andrew Jackson and Jaideep Mavoori began in 2003, when Mavoori, an electrical engineer, was developing a computer chip that could be implanted in moths to record and stimulate neuronal activities as they flew. “We heard about the work and thought something that small should be easy to implant into a monkey's brain,” recalls Fetz. That insight turned out to be correct, although implementing it was not easy.

The team spent more than a year redesigning the chip so that it could be programmed to record signals from neurons firing in an area of the brain's motor cortex responsible for moving certain wrist muscles in one direction (site A). The chip, which was positioned on top of the brain, then delivered these signals a few milliseconds later to a nearby site responsible for controlling a different set of wrist muscles (site B). The chip made the recordings and stimulations through two implanted electrodes, thus providing a constant connection between the two sites in the brain. The device had to be small, but possess enough battery power to run for several days, and enough computational power to record signals and analyse the data in real time, according to Mavoori.

The chip also had to be sufficiently stable to function in living animals. “Monkeys do not sit very still. They don't seem to care if you are trying to do science,” laughs Fetz, who worried that such movements would create too much background noise. But once implanted, the chip worked. “I remember the first time we downloaded the data from the chip and got a glimpse of a day in the life of a brain cell,” recalls Jackson. “It was very exciting.”

They next asked whether the microchip was having any effect on brain functions. At the start of the experiment, they used an external stimulator to test different sites in the monkeys' motor cortex. As expected, stimulation of site A caused wrist muscles to twitch in one direction, and stimulation of site B caused twitching in another. Then the monkeys were left to go about their business with the chip on their brain, where it recorded and transmitted signals. After a couple of days, the researchers once again tested different areas of the brain with the external stimulator. They found that now stimulation of site A was causing the wrist to move in the same direction as stimulation of site B. Thus the artificial connection between A and B had changed A's function to be more like B's.

“I was rather surprised when we got the result,” says Jackson. “We saw a change of output from a whole region of the cortex and this caused a lasting change in the movement produced by stimulation.” This change in function, called plasticity, is thought to occur normally during learning and memory. The researchers speculate that in people who have suffered brain damage, the chip could be used to help reorganize information flow and boost the functioning of spared neurons.