Using computer games, CD-ROM books and audiotapes, a team of language researchers has developed a therapy that they say has the potential to correct language-learning disabilities that afflict as many as 7 million children in the United States.
Such children, who have trouble understanding and using the spoken word, typically go on to develop the reading problems characteristic of an impairment of reading ability called dyslexia.
Current therapies involve intensive, expensive, one-on-one interventions and may continue for years.
The new approach, which uses computers to train the brain to recognize hard-to-hear sounds, has been shown to accomplish in a month the same goals that conventional therapy achieves in two or more years.
The techniques, developed at Rutgers University in New Brunswick, N.J., and the University of California in San Francisco, are described in two papers in today's issue of the journal Science. The Charles A. Dana Foundation funded the studies.
The new therapy is based on more than two decades of research into the causes of language-learning disabilities by neuropsychologist Paula Tallal of Rutgers. She reported last year that the major problem in perhaps 85 percent of these children is that their brains simply do not operate fast enough to distinguish many different sounds.
When an individual hears a sound, "brain cells require a certain amount of time to respond, recover and be ready to respond again," Ms. Tallal said. In normal individuals, this response occurs very quickly, on the order of 30 to 40 milliseconds. For language-learning impaired children, the process takes 100 milliseconds or longer.
But many speech sounds, particularly the so-called stop-consonant syllables such as ba, da, ga, pa, ta and ka, have a very short transitional period in which the initial spoken consonant sound changes to the vowel. The initial "b" in ba, for example, lasts for only 40 milliseconds before switching to the "ah" sound.
Because the brains of the language-learning impaired children do not respond rapidly enough, they have great difficulty distinguishing these sounds, Ms. Tallal has found. This language-learning disorder translates into dyslexia, as children enter school and learn to associate speech sounds with letters.
Based on this insight, neuroscientist Michael Merzenich and his colleagues at UCSF used computers to alter speech, accenting the stop-consonant syllables so they stand out from the rest of the speech and lengthening them so that the children could recognize them.
Using the altered speech, researchers developed video games that reward the user when he or she recognizes the stop-consonant sounds and responds appropriately. As the user's proficiency increases, the syllables are progressively shortened until they are recognized in real time.