Taken together, functional brain scans and tests of reading skills strongly predict which children will have ongoing reading problems. What's more, the two methods work better together than either one alone, according to new research in the June issue of Behavioral Neuroscience, which is published by the American Psychological Association (APA).
Neuroscientists at Stanford and Carnegie Mellon universities think this double-barreled diagnostic can help identify at-risk readers as early as possible. That way, schools can step in before those children fail to learn to read or develop poor reading habits that might interfere with remediation, such as relying on memory for words rather than sounding out new ones. Early identification and systematic intervention can very often turn likely non-readers into readers, according to the study authors.
This study of 73 Pittsburgh-area children of ages 8 to 12, all identified as struggling readers, ran for a school year. At the start of the year, the researchers administered standard tests of early literacy skills, including word identification, fluency, comprehension, vocabulary, efficiency, and phonological processing -- this last a critical measure of how well children process the sounds of letters and letter combinations.
The researchers also used functional MRIs (fMRIs) to depict how the children's brains' worked when they had to read two words and say whether they rhymed, a test of phonological awareness. To make the fMRI results more sensitive to differences among children, the authors further analyzed the images using a method called "voxel-based morphometry" that uses the density of the brain's white and grey matter to zero in on activation patterns in specific parts of key brain regions.
At the end of the school year, the team, led by Fumiko Hoeft, MD, PhD, of the Stanford University School of Medicine, tested the children's ability to decode text using the Word Attack subtest of the Woodcock Reading Mastery Test, a standardized measure of decoding. Hoeft's team then determined which test method (either or both) predicted reading skill more strongly. The model combining the behavioral and neuroimaging measures predicted future decoding significantly better than either of those methods alone.
The behavioral predictors alone accounted for 65 percent of the variance in end-of-year performance, which means they could tell future good from poor readers nearly two out of three times. Brain imaging (a composite of the fMRI and voxel analysis scores) accounted for 57 percent of the later variance, thus accurately predicting more than half the time whether they would still have problems reading after a year of regular instruction. Both figures are respectable. However, together they explained an impressive 81 percent of the variance. In other words, the combined tests were able to predict the children's future decoding skill more than four out of five times.
Although MRIs might not be suitable as widespread screening instruments, they might be considered for use in children showing early reading problems, especially to differentiate children who have a true language disorder from those who simply need time to mature. Hoeft points out that the cost of a brain scan might compare favorably to the expense of hiring trained personnel to run full batteries of neuropsychological testing, the more common mode of problem identification.
Conceivably, she adds, if researchers can run a similar but very large study assigning children to different remedial reading programs, sometime in the future they may be able to determine which programs will work best for which children through understanding both their reading behavior and their specific patterns of brain activation.
Article: "Prediction of Children's Reading Skills Using Behavioral, Functional, and Structural Neuroimaging Measures;" Fumiko Hoeft, MD, PhD, Stanford University School of Medicine and Stanford University; Takefumi Ueno, MD, PhD, and Allan L. Reiss, MD, Stanford University School of Medicine; Ann Meyler, PhD, Carnegie Mellon University; Susan Whitfield-Gabrieli, BSc, MIT; Gary H. Glover, PhD, Stanford University School of Medicine; Timothy A. Keller, PhD, Carnegie Mellon University; Nobuhisa Kobayashi, MD, PhD, Paul Mazaika, and Booil Jo, PhD, Stanford University School of Medicine; Marcel Adam Just, PhD, Carnegie Mellon University; and John D. E. Gabrieli, PhD, MIT; Behavioral Neuroscience.
Cite This Page: