Scientists Demonstrate Animal Mind-Melds – The New York Times


A single neuron can’t do much on its own, but link billions of them together into a network and you’ve got a brain.

But why stop there?

In recent years, scientists have wondered what brains could do if they were linked together into even bigger networks. Miguel A. Nicolelis, director of the Center for Neuroengineering at Duke University, and his colleagues have now made the idea a bit more tangible by linking together animal brains with electrodes.

In a pair of studies published on Thursday in the journal Scientific Reports, the researchers report that rats and monkeys can coordinate their brains to carry out such tasks as moving a simulated arm or recognizing simple patterns. In many of the trials, the networked animals performed better than individuals.

“At least some times, more brains are better than one,” said Karen S. Rommelfanger, director of the Neuroethics Program at the Center for Ethics at Emory University, who was not involved in the study.

Brain-networking research might someday allow people to join together in useful ways, Dr. Rommelfanger noted. Police officers might be able to make collective decisions on search-and-rescue missions. Surgeons might collectively operate on a single patient.

But she also warned that brain networks could create a host of exotic ethical quandaries involving privacy and legal responsibility. If a brain network were to commit a crime, for example, who exactly would be guilty?

“It’s really important to address these issues before they come up, because when you try to play catch-up, it can take a decade before something’s in place,” she said.

For the past 25 years, Dr. Nicolelis and his colleagues have been designing devices that decipher signals recorded by electrodes implanted in brains. With these instruments, monkeys can learn to control robot arms and even entire robotic exoskeletons.

Monkeys get better at these tasks as more brain neurons join in the effort to produce commands. The scientists wondered if combining brains into a network might bring an even greater supply of neurons to bear on the tasks.

Dr. Nicolelis and his colleagues began by implanting two sets of electrodes in the brains of four rats. One set delivered a signal into one part of each brain, while the other eavesdropped on a different brain patch.

The four rats received the same signal, and then a computer monitored how their brains responded. If all four rats produced synchronized signals in their brains, they were rewarded with a sip of water.

Through trial and error, the rats learned how to consistently synchronize their brains, making it possible for the rats to act like a simple computer. In one experiment, the animals learned how to produce different brain responses to two different signals: a single burst of electric pulses, or four bursts.

The rats learned how to produce synchronized brain activity in response to one of the signals, and unsynchronized activity in the other. Their collective response was correct as often as 87 percent of the time — substantially better than an individual rat learning on its own.

The scientists also found that the brains of three rats could be linked into an information-processing chain. First, they trained one rat to produce the correct kind of brain activity to two different electrical bursts in the brain. Then they linked the first rat’s brain signals to the brain of a second rat.

The second rat learned to produce the same response as the first rat, the scientists found, and a third rat could reliably interpret the second rat’s brain responses. And when they delivered the third rat’s brain signals back to the first rat, it also responded correctly much of the time.

Dr. Nicolelis and his colleagues then turned from rats to monkeys, with a new twist on earlier experiments in which individual monkeys learned to control a robot arm. This time the scientists implanted electrodes into two monkeys instead of one.

Each monkey looked at a computer screen on which there were images of an arm and a ball. The computer combined the brain signals from both monkeys to move the arm. The two monkeys learned to work together to reliably move the arm to the ball, which produced a reward.

In another trial, one monkey learned to control the horizontal movement of the arm while the other monkey controlled its vertical movement by means of electrical brain impulses. In an even more ambitious test, the scientists programmed a virtual arm in three-dimensional space, allowing three monkeys to share control of different aspects of its movement.

Once again, the monkeys learned to move the arm to the ball. Even when one of the monkeys did a bad job of controlling the arm, the other two compensated to keep it on track.

The idea of brain nets is not new. In a 2013 study, for example, French scientists created a video game that two people could play with their brains.

Each player donned an EEG recorder that detected brain waves from the scalp. With practice, the players could learn how to combine their EEG signals to move a ball into a goal on a computer screen as well as or better than a single player.

Rajesh P.N. Rao, a professor of computer science and engineering at the University of Washington, said that it was the complexity of Dr. Nicolelis’s new studies that made them important.

“What’s different here is that he’s able to demonstrate that more than a pair of brains can be yoked together,” he said.

Dr. Nicolelis speculated that our brains can naturally join together when we share the same experiences. “When people are watching television — millions of people watching the same images — we may be synchronizing millions of brains,” he said.

By understanding this capacity of the brain, it may someday be possible to combine the power of many human brains. “One can imagine that these experiments are paving the way for people to solve problems by literally putting our heads together,” Dr. Rao said.

Dr. Rommelfanger considers it unlikely that people would be willing to have brain surgery to join a network. “But I think this is a starting point for being able to move to less invasive technologies,” she said.

Already, scientists have developed powerful tattoo-like EEG devices that can stick to the skin and pick up brain activity. It’s also possible to deliver magnetic pulses to different regions of the brain through the scalp with a technology known as transcranial magnetic stimulation.

The video game industry might drive the development of better devices that could allow players to join large-scale networks, Dr. Rommelfanger said. Such games would create ethical problems similar to those we now face on the Internet.

If a company records people’s brain signals as part of a network, what guarantees will there be that someone can’t steal the data?

“I think that neural privacy is something we should worry about,” she said.