Credit: Jasiek Krzysztofiak/nature

Two monkeys sit at computer screens, eyeing one another as they wait for a promised reward: apple juice. Each has a choice — it can either select a symbol that results in juice being shared equally, or pick one that delivers most of the juice to itself. But being selfish is risky. If its partner also chooses not to share, neither gets much juice.

'Atomtronic' superfluid remembers how it has been stirred Genome of ‘Clovis boy’ raises questions on handling of Native American remains UK reveals decontamination procedures for terror attacks

This game, the ‘prisoner’s dilemma’, is a classic test of strategy that involves the simultaneous evaluation of an opponent’s thinking. Researchers have now discovered — and manipulated — specific brain circuits in rhesus macaques (Macaca mulatta) that seem to be involved in the animals’ choices, and in their assessments of their partners’ choices. Investigating the connections could shed light on how social context affects decision-making in humans, and how disorders that affect social skills, such as autism spectrum disorder, disrupt brain circuitry.

“Once we have identified that there are particular neural signals necessary to drive the processes, we can begin to tinker,” says Michael Platt, a neurobiologist at Duke University in Durham, North Carolina.

Neurobiologists Keren Haroush and Ziv Williams of Harvard Medical School in Boston, Massachusetts, zoomed in on neural circuits in rhesus macaques by implanting electrode arrays into a brain area called the dorsal anterior cingulate cortex (dACC), which is associated with rewards and decision-making. The arrays recorded the activity of hundreds of individual neurons. When the monkeys played the prisoner’s dilemma (see ‘A juicy experiment’) against a computer program, they rarely chose to cooperate. But when they played with another monkey that they could see, they were several times more likely to choose to share the juice.

As the monkey made its decision, a specific set of dACC neurons tended to fire if the animal was choosing to cooperate, allowing the researchers to predict its decision two-thirds of the time. A different set often fired if the monkey thought its partner would cooperate, and the predictions were 80% accurate.

Researchers have long known about mirror neurons, which are involved in copying other individual’s actions, but this is the first discovery of neurons that predict another individual’s unknown actions. The researchers also found that when they interfered with the cooperation circuit by delivering an electric shock to the area, the monkeys became less likely to co­operate — suggesting that the circuits for social inter­action normally override an inherent desire to self-indulge. Haroush and Williams will present their findings this week at the Computational and Systems Neuro­science meeting in Salt Lake City, Utah.

The ability to track individual circuits while monkeys assess one another is an important advance, says Robert Seyfarth, a psychologist at the University of Pennsylvania in Philadelphia, although he cautions that real social interactions are much more complex. Furthermore, he says, a monkey’s willingness to cooperate might depend on the other monkey — whether it is a dominant male, for instance, or a relative. “We really need to ask monkeys what they know about other monkeys,” says Seyfarth.

A closer look at the circuits could reveal how they respond to hormones or drugs, says Steve Chang, a psychologist at Yale University in New Haven, Connecticut. The neuro­hormone oxytocin, in particular, has been touted as a potential treatment for autism, because it seems to aid the formation of social bonds. For example, men playing the prisoner’s dilemma are more likely to co­operate if they have breathed in oxytocin1.

Chang and Platt have monitored the brains of monkeys playing a different game, in which they can choose to reward other monkeys without any sacrifice. Receiving a reward causes one set of neurons to fire, they found2; watching another monkey receive a reward triggers a different set. As in the prisoner’s dilemma, the monkeys wanted to reward another monkey only if it was physically present, and were unlikely to reward a computer. Chang is now studying how oxytocin, which has been shown to increase monkeys’ willingness to reward3, affects neural circuitry.

Besides drugs and neuro­hormones, electrical stimulation can also alter the brain. Platt’s lab is trying to map neural circuits, and show how they respond to electromagnetic stimulation. Previous work has shown4 that stimulating certain parts of the brain can increase people’s ability to perform empathetic tasks, such as assessing what someone else knows or likes.

But Platt says that scientists are just beginning to understand how such methods work. He adds that watching the response of neural circuits in monkeys is a good way to work out how much stimulation or hormone should be applied, and where. “If you were a parent and considering oxytocin or brain stimulation for your child, you would want to know the answers to all those questions,” he says.