Youngsters experience huge swings in intelligence during their teenage years, a finding with widespread implications for education, scientists reported on Wednesday.
The results will not surprise parents who find their children are a sandwich short of a picnic at one stage of their adolescence but a few years later display shafts of brilliance -- or sometimes the reverse.
But the study say the shifts are linked to changes in the structure of the brain, not to hormonal surges, a finding that challenges an enshrined belief.
The theory that intellectual ability is stable throughout one's life underpins intelligence tests applied in primary schools that sometimes shape an individual's life by determining options in education and careers.
British researchers tested 33 adolescents in 2004, when the volunteers were aged between 12 and 16.
They used a standard intelligence quotient test which measures "verbal IQ" -- language skills, arithmetic, general knowledge and memory -- and "non-verbal IQ," such as identifying missing parts of a picture or solving a visual puzzle.
They put the youngsters through a second test in 2007-8.
On both occasions, the individuals were placed in a magnetic resonance imaging (MRI) scanner, which takes a 3-D snapshot of the brain.
In some cases, the teenagers' IQ over this period fell -- or rose -- by 20 points compared to a counterpart of a similar age.
Stunned by this discovery, the team turned to the MRI images and found the teens' brains had undergone significant changes during the intervening years.
For example, an increase in verbal IQ was associated with an increase in the density of grey matter in the left motor cortex, a part of the brain that helps process speech.
A rise in non-verbal IQ was linked to a boost in grey matter in the anterior cerebellum, which is associated with hand movements.
But an increase in one kind of IQ was not necessarily accompanied by an increase in another.
"We found a considerable amount of change in how our subjects performed on the IQ tests in 2008 compared to four years earlier," said Sue Ramsden of the Wellcome Trust Centre for Neuroimaging at University College London (UCL).
"Some subjects performed markedly better but some performed considerably worse. We found a clear correlation between this change in performance and changes in the structure of their brains, and so can say with some certainty that these changes in IQ are real."
Previous research has highlighted the plasticity of the brain, an organ that can change in structure throughout adult life. For instance, London taxi drivers who take the road-map test known as "The Knowledge" have a bigger hippocampus, a nodal point for navigation and memory.
"The question is, if our brain structure can change throughout our adult lives, can our IQ also change?" said lead researcher Cathy Price, a UCL professor.
"My guess is, yes. There is plenty of evidence to suggest that our brains can adapt and their structure changes, even in adulthood."
Still unclear, though, is whether the IQ changes in adolescence are due to early or late development, or whether education played a role.
If the findings, reported in the British journal Nature on Wednesday, are replicated elsewhere, the implications could be far-reaching.
"We have a tendency to assess children and determine their course of education relatively early in life, but here we have shown that their intelligence is likely to be still developing," said Price.
"We have to be careful not to write off poorer performers at an early stage when in fact their IQ may improve significantly given a few more years."