You can teach an old dyslexic new tricks. Intensive coaching may train a brain region on the right-hand side of the brain to compensate for a related, malfunctioning one on the left side, a new imaging study has revealed.

Three-and-a-half hours of drill per day, five days a week for eight weeks, improved the reading ability of severely dyslexic adults "by between 10 and 20% -- bringing them into the normal range for reading", Georgetown University Medical Centre researchers found.

The team told the annual meeting of the American Association for the Advancement of Science in San Francisco today that a regime of 'phonological' training -- breaking down and manipulating the sound components of words -- seems to increase activity in the right temporoparietal cortex, which processes spatial tasks.

This is the sister region of the left temporoparietal cortex, which deals with visual motion and which is underactive in many dyslexics. These findings add to a growing body of evidence linking visual areas with the subtle portmanteau of motion-detection and motor-control impairments that make up dyslexia.

About 5% of the world's population suffers from dyslexia. But despite its educational and social effects it is still only poorly understood. Several genetic regions -- on chromosomes 1 and 6, for example -- have been linked to the condition.

"We now plan to follow up the adults to see if their improvements persist," says Guinevere Eden, one of the Georgetown team. "We hope that ultimately some modified version of this training will be used in schools."

To this end, a large-scale clinical trial of the same coaching and screening method will start later this year. It too will use behavioural tests combined with functional magnetic resonance imaging, this time to assess the reading skill and neural changes in several hundred dyslexic children over three to four years.

"Behavioural tests alone are sensitive but 'noisy'," explains Thomas Zeffiro, also of the Georgetown team. Carrying out brain imaging as well, he adds, is about 10 times more expensive but much more informative. Indeed, the team believes that a combination of the two could one day offer a powerful pre-symptomatic test for potentially dyslexic toddlers.