ZOOMING IN ON DYSLEXIA

By J. MADELEINE NASH/NEWARK

AT AGE 5, KEILLAN LECKY DREADED kindergarten. So many of the words her playmates gleefully shouted or conspiratorially whispered seemed to hover just out of reach, as elusive as a vanishing rainbow. Her difficulty understanding them was starting to affect her schoolwork. Then, last summer, Keillan, along with 21 other language-impaired children, was enrolled in an experimental program at Rutgers University in Newark, New Jersey, in which the kids improved their auditory skills by playing computer games. The change in Keillan and the others was so remarkable, says Paula Tallal, co-director of the Center for Molecular and Behavioral Neuroscience at Rutgers, that even the scientists were stunned. After just four weeks of therapy, Tallal and her colleagues report in a recent issue of the journal Science, youngsters who were performing well below age level had jumped as much as two years.

For parents and teachers who have watched children like Keillan struggle so long with so little gain, the announcement was encouraging news. If larger studies bear out the results, then at least some cases of language impairment--those that stem from an inability to decode spoken words--may find effective treatment.

That in itself would be an important breakthrough. This form of language impairment afflicts up to 8% of otherwise normal children, most of whom go on to develop intractable problems with reading and writing. But Tallal and her colleagues take their findings one step further, and in doing so have aroused intense scientific controversy. They believe the same language-processing "glitch" may be the root of the more common problem of dyslexia, a reading disability that affects perhaps 15% of the population. If so, games like those that Keillan played could help at least some dyslexics whose impairment makes it hard for them to fully share in all the vital knowledge and pleasure that come with the printed word.

Can such a debilitating learning disorder really be remedied by playing games? Other experts, while praising Tallal's work as provocative and challenging, remain skeptical. Tallal, they point out, has not yet demonstrated that her therapy is effective for the broader population of dyslexics. But as Yale University's Dr. Sally Shaywitz, a behavioral pediatrician, acknowledges, "We don't need to speculate. We can carry out studies that answer this question."

The children enrolled in the Rutgers study had one thing in common: although they are of at least average intelligence, tests showed that they frequently had difficulty distinguishing among phonemes (the basic building blocks of language), especially those that begin with hard consonants like b, d and p. "There is nothing wrong with their ears," says Tallal. "They can hear these sounds, but the auditory centers of the brain can't process them."

The problem, Tallal believes, is all in the timing. Vowel-rich sounds resonate for 100 milliseconds, sometimes longer, and are thus easier to make out than hard consonants, which fly by in normal conversation at speeds of 40 milliseconds or less. Language-impaired children, Tallal has demonstrated, can more reliably identify fast consonants when the sounds are slowed to half their normal speed.

At the same time, neuroscientist Michael Merzenich, at the University of California, San Francisco, has shown that intensive training exercises can dramatically improve the ability of monkeys to identify minute differences in rapid-fire bursts of sound and that these improvements are accompanied by striking changes in the firing patterns of neurons in the monkeys' brains.

In 1994, with a $2.3 million grant from the Charles A. Dana Foundation, Tallal and Merzenich pooled their talents to design new therapies for language-impaired children. The researchers, led by William Jenkins of ucsf and Steve Miller of Rutgers, created computer programs that made the hard consonants easier to hear by elongating them, spacing them farther apart and making them louder. Then the researchers devised a series of computer and classroom games that enticed children to listen to the strange, synthetic sounds and gradually differentiate among them.

After a month of training, three hours a day, five days a week, the kids had made striking progress. The researchers believe the exercises--"aerobics for the brain," Tallal calls them--strengthened the connections between the neurons responsible for distinguishing fast-moving sounds. "It's possible," says Merzenich, "that we may actually be able to eliminate language impairment in a great majority of children."

This optimism seems extraordinary in view of the mounting evidence that both dyslexia and oral-language impairment are inherited disorders. In fact, it seems quite probable that a single gene on chromosome 6 may underlie at least some cases of dyslexia and perhaps other language-based learning problems as well. But a genetic susceptibility to dyslexia does not mean that the condition is inevitable or, after it occurs, that it is irreparable.

Merzenich believes something as simple as an inherited susceptibility to middle-ear infections in the first six months of life could explain at least some of the language problems young children experience. Other researchers think the problem is more fundamental. Autopsy work by Harvard neurologist Dr. Albert Galaburda shows that the brains of dyslexic people are dappled with tiny lesions and out-of-place cells, which suggests that the core problem may lie in the machinery that controls prenatal development.

Certainly difficulties in processing fast sounds show up very early. Developmental psychologist April Benasich, one of Tallal's colleagues at Rutgers, has conditioned six-month-old babies to turn their heads whenever they detect a change in a sequence of tones. As long as the tones are spaced well apart, all the babies do well. (Their reward is a toy that lights up or moves.) But when the interval between tones grows shorter, big differences emerge. Some babies cannot detect the changes unless the sounds are presented 300 milliseconds apart. Others do well when they are separated by a mere 40 milliseconds. The babies who do poorly at this test also have difficulty distinguishing between speech sounds like ba and da.

This difference in perception might just be critical, says University of Washington neuroscientist Patricia Kuhl. For it is during the first year of life that children form what Kuhl terms "mental magnets," which sweep up similar-sounding speech sounds and file them away in phonic bins. If language-impaired children never perceive ba and da as different, then they may form mental magnets that file these sounds into the same broad category, seriously undermining their ability to group sounds into words and sentences later on. Indeed, believes Benasich, the ability to make fine acoustic distinctions is one of the pilings on which language is built. "If the pilings are rickety," she says, "then language is not going to develop as well."

If this analysis is correct, then the possibilities for intervening early in a child's life multiply. "Wouldn't it be wonderful," asks Merzenich, in a burst of enthusiasm, "if we could treat dyslexia before a child started trying to read?" Or better still, before a child started trying to talk. Tallal and Merzenich go so far as to suggest that some forms of language impairment could turn out to be more correctable than poor hearing or poor eyesight. They point out that the earphones that transmit the exaggerated speech sounds to children's ears in the lab are only temporary aids. "When you take off eyeglasses, you can't see," observes Kuhl. "But when you take off these funny-looking earphones, then you might just proceed to understanding normal speech."

Perhaps. But, until Tallal and Merzenich know for certain, they may be giving more hope than is justified. Ever since the Science articles appeared in early January, thousands of desperate parents have flooded the Rutgers and University of California switchboards with calls, asking when the new therapy will be offered by local schools. To handle the overload, the researchers have set up an 800 number.

All the two neuroscientists can offer at the moment, however, is promising results from a very small research project. Whether the same approach will prove valuable, or even marginally useful, for the estimated 10 million dyslexic children in the U.S. remains an open question, and parents would be unwise to harbor unrealistic hopes. About one thing, however, there is no doubt. Tallal and Merzenich have made a difference in the lives of at least a few children. Keillan, the girl who hated kindergarten, is now 6 years old. She adores first grade. She runs to school smiling. And, with just a little difficulty, she is learning to read.