'Mind reading' neurones help predict behaviour of others

© memerym/iStockphoto

A new study on rhesus monkeys sheds light on co-operative behaviour.

Scientists have discovered a group of neurones that enable one monkey to predict what another monkey is about to do - the first-known instance of neurones calculating another animal's behaviour.

The discovery may be fundamental for understanding social behaviour and could lead to better treatments for conditions like autism spectrum disorder.

US neuroscientists got pairs of monkeys to play a game based on classic game theory known as 'the prisoner's dilemma.'

Their findings are published today in the journal .

Decisions... decisions

In the game, the monkeys sit side by side facing a computer screens. They can choose either to cooperate (signified by pressing a hexagon on their screen) or to be selfish (by pressing a triangle).

Although they are well aware of each other's presence, neither monkey can see the other's facial expressions, nor can they see the choice the other monkey makes as they make it, explains study co-author neuroscientist Dr Keren Haroush of Harvard Medical School and Massachusetts General Hospital, Boston, USA.

Their reward depends on their combined choices. If one monkey chooses to be selfish and the other to cooperate, the selfish monkey wins hands down, getting six drops of juice as a reward while the other (cooperative) monkey gets only one drop.

But if they both choose the selfish option they get just two drops each. Both deciding to cooperate, however, wins them each four drops of juice.

"The only follow-up was at the end of the trial: once they had both made their selections, they got to see what the other one chose." says Haroush.

Not only could they see the choice the other monkey made, they could also hear the drops of juice that it got as a reward.

Listening in

While the monkeys were making up their minds, the researchers were eavesdropping on the neurones in a brain region known as the cingulate cortex.

"We had small microchips inside the brains and we were able to record many neurones at the same time. We could basically 'listen in' on their activity as the monkeys were performing this task," says co-author Associate Professor Ziv Williams, also of Harvard Medical School and Massachusetts General Hospital, Boston, USA.

"We found some of these neurones fired differently based on what the other monkey's yet unknown decision was predicted to be. That was pretty remarkable and surprising...we could actually tell what the first monkey thought the other would do way before that decision was revealed," he says.

The researchers recorded from 363 neurones during the games. Of these 32.4 per cent appeared to be involved in predicting what the other monkey would do - attempting to 'mind read' - while a (largely different) 24.3 per cent of neurones seemed to be encoding the monkey's own decision.

"Basically what's happening is these neurones are building up a predictive model of what the other monkey is likely to do in a situation, given their past interactions," says Williams.

Furthermore, the 'mind-reading' neurones appear to be activated more in social situations. When the monkeys played the game in separate rooms (and did not know about each other) or played against a computer, considerably fewer neurones activated by the task were 'mind-reading' ones.

"Our best guess is that these types of neurones are likely to be found in other social animals including humans. However, without testing this hypothesis directly, we currently do not know for certain," remarks Williams.

Social disruption

Haroush and Williams also tried disrupting the cingulate cortex by sending electrical pulses to that part of the brain while the monkeys were doing the task.

They found the monkeys were much less likely to cooperate with each other, even if they had both cooperated on the immediately preceding trial.

The pulses "selectively inhibited their ability to cooperate and to reciprocate past cooperation," says Williams. The pair thinks this demonstrates the key role of the cingulate cortex in social behaviour.

"Ultimately we want to see if we can use some of these insights to develop treatment for social behavioural disorders such as autism and antisocial personality disorders," concludes Williams.