Parts of the brain responsible for decision-making and impulse control are still developing during a person's teens

The age of criminal responsibility in England, Wales and Northern Ireland could be "unreasonably low" given the emerging understanding of how slowly the brains of children mature, according to a report by the Royal Society. Widespread differences between individuals also mean that the cut-off age at which children are deemed fit to stand trial, at 10 years old, might not be justifiable in all cases.

The comments are part of an assessment carried out by a panel of scientists, lawyers and ethicists of how developments in neuroscience and brain imaging should inform the future practice of law. Neuroscience and the Law, published on Tuesday, examines how scientific understanding of the brain has advanced in recent decades and the light this has shed on behaviour. The report also assesses the reliability of lie detector tests.

In England, Wales and Northern Ireland, a child is deemed fit to stand trial at the age of 10, but in recent years it has been shown that important changes in the brain's neural circuits go on well into a person's teens. In Scotland children cannot be convicted until they are 12.

"A number of psychologists have already shown that adolescents are not wholly responsible individuals and are inclined to take risks and behave in irresponsible ways," said Nicholas Mackintosh, an emeritus professor in the department of experimental psychology at the University of Cambridge and chair of the Royal Society panel. "What neuroscience has shown in the last 10 years is that this is at least associated with the fact that the brain continues to develop throughout adolescence."

In particular, the prefrontal cortex, which is responsible for decision-making, impulse control and cognitive control, is among the slowest parts of the brain to mature and is not fully developed until around the age of 20. "Neuroscience adds to the evidence that a 10 or 12 or 15-year-old does not have a fully adult brain in many important respects," said Mackintosh.

Research has also shown that there is huge variation between individuals and that the development of the slowest-developing parts of the brain is associated with comparable changes in mental functions such as IQ, suggestibility, impulsivity, memory and decision-making.

In contrast, the amygdala, an area of the brain responsible for reward and emotional processing, develops during early adolescence.

"It is clear that at the age of 10 the brain is developmentally immature, and continues to undergo important changes linked to regulating one's own behaviour," said the report. "There is concern among some professionals in this field that the age of criminal responsibility in the UK is unreasonably low, and the evidence of individual differences suggests that an arbitrary cut-off age may not be justifiable."

Roger Brownsword, a former professor of law at King's College London, said that the question of criminal responsibility showed how neuroscience should form part of general policy debates around criminal justice. "It wouldn't be so much neuroscience driving that but it might be that neuroscience became very relevant in the background as we review that particular question," he said.

Another emerging uses of brain imaging is in lie detection. Several companies in the US offer services that claim to reveal whether a defendant is telling the truth. "We take a similarly sceptical view of that," said Mackintosh. "Most brain imaging data might tell you the difference between someone who is telling the truth and someone who is lying, though even that is uncertain. But it is quite certainly the case that some guilty offenders who protest their innocence actually believe they are innocent. They are not lying as far as they're concerned. It is also possible that, as soon as you're told what brain imaging is picking up in your brain, it's quite easy to fool it."

He added that the large majority of behavioural scientists and neuroscientists agree that brain imaging is not yet ? and perhaps never will be ? a reliable lie detector.

This has implications for criminal trials, said Brownsword, where the perceptions of jurors might be dangerously incorrect unless they have been tutored in exactly what this science can and cannot do.

"Lawyers and neuroscientists have to be educated in the nature of behavioural and neuroscientific evidence that all it can do is, by and large, indicate changes in probability," added Mackintosh. "The law is not only concerned with passing sentence in court, it's also concerned with prediction and risk assessment. Is this person likely to re-offend if now released from prison? It's impossible to make this prediction with any degree of certainty or accuracy. Risk assessment is a risky business and is notoriously inaccurate."

One of the main recommendations of the Royal Society report is that there should be an international meeting of neuroscientists and lawyers every three years to discuss the latest advances in "areas at the intersection of neuroscience and the law to identify practical applications that need to be addressed".