For children with dyslexia, reading doesn't come naturally, and only about 20% of them grow into normal readers by adulthood. No one knows why this is, and standard reading tests can't predict which kids will outgrow their reading problems. But brain scans can, according to a new study. The researchers say their findings provide clues about the neurobiology of dyslexia and could one day help educators identify students who could benefit from more intensive help.
Neuroscientist Fumiko Hoeft of Stanford University in Palo Alto, California, and her colleagues used functional magnetic resonance imaging to investigate brain activity during reading in 45 children between the ages of 7 and 16, of whom 25 had dyslexia. After 2.5 years, the same children returned to the lab to take a battery of standardized reading tests. Then the researchers went back to the brain scans to see if there were differences in the kids with dyslexia who had made the biggest improvements.
And indeed there were. Inside the scanner, participants had seen pairs of words flashed on a screen and been asked to indicate whether they rhymed. Children with dyslexia take longer and make more mistakes, especially when different spellings produce a rhyme, as in "gate" and "bait." To analyze brain activity evoked by this task, Hoeft's group used a method called multivariate pattern analysis, a statistical technique for comparing patterns of activation across the entire brain (Science, 13 June 2008, p. 1412). The patterns predicted with 92% accuracy which dyslexic kids exhibited above or below average gains in reading over the next 2.5 years, the researchers report online this week in the Proceedings of the National Academy of Sciences. This particular type of analysis doesn't reveal which brain regions are responsible, however.
The differences in reading progress can't be explained by differences in tutoring and other types of remediation, Hoeft says. The researchers found no correlation between participation in such programs and reading gains. Moreover, 17 conventional tests of reading and language also failed to predict reading progress.
That makes the findings especially impressive, says Rajeev Raizada, a neuroscientist at Dartmouth College. Using brain scans to make more accurate diagnoses and predictions about behavior than can be made with more conventional tests should be a major priority for neuroscientists, Raizada says. But so far very few studies have successfully done this. "Fumiko's paper is probably the closest to neurally based diagnosis that our field has come so far."
In additional experiments, the Stanford researchers found clues about what's different about the brains of dyslexic children who show the biggest reading improvements. For example, those who had greater activation in the right inferior frontal gyrus (IFG) during the rhyming task improved more. Previous studies have found that the counterpart of this region on the left side of the brain is underactive in people with dyslexia, Hoeft says. She thinks the increased activation in the right IFG may reflect a compensation for a deficit on the left side of the brain, which typically plays a bigger role in language.
Another type of scan suggested that dyslexic children with more connectivity among brain regions on the right side of the brain are better able to overcome their deficit. All in all, the findings suggest that kids who are somehow able to shift more language function from the left to the right side of the brain are better able to overcome their reading deficit, Hoeft says.
She hopes that refinements to these methods will one day lead to tests that could be used to identify which students are most in need of help, and perhaps even determine what type of intervention is most likely to work for a given student. That's probably still a long way off, says Raizada. He notes that it will be important to replicate the findings in larger, more diverse groups of dyslexic children. But, he says, "this type of study lays the groundwork for showing that it is not a pie-in-the-sky idea."