Researchers to create tool for diagnosing children with apraxia of speech

Apr 27 2004

A new National Institutes of Health grant will enable Oregon Health & Science University researchers to create a tool for diagnosing children with apraxia of speech -- a neurologically based speech motor disorder. Children with apraxia of speech have great difficulty planning and producing the speech movements they need to make with their tongue, lips, jaw and velum, or soft palate, to form words. Children with apraxia of speech tend to speak less rhythmically than other children, including more even stress on syllables.

It is estimated that about one-tenth of 1 percent of children in the United States may have apraxia of speech. Apraxia of speech can occur as a result of a brain injury, but in children born with apraxia, the cause is unclear.

"Apraxia of speech is not very well understood or easily diagnosed," said John-Paul Hosom, Ph.D., the principal investigator of the three-year $492,000 grant from NIH's National Institute on Deafness and Other Communications Disorders, and an OGI School of Science & Engineering assistant professor of biomedical engineering. "It is especially difficult to diagnose apraxia of speech in young children because of the different rates at which young children learn language. It is suspected that many children with apraxia of speech go undiagnosed due to the lack of a diagnostic standard."

Hosom, a computer scientist in OGI's Center for Spoken Language Understanding (http://www.cslu.ogi.edu), will collaborate with one of the nation's leading experts on childhood speech disorders, Lawrence Shriberg, Ph.D. Shriberg is a professor of communicative disorders at the University of Madison-Wisconsin, director of the NIH-supported Phonology Project, and co-director of The Phonology Clinic at the Waisman Center (www.waisman.wisc.edu). Shriberg, who has been working with children with apraxia of speech for more than 20 years, is interested in the nature and origin of childhood speech disorders of currently unknown origin and ways to identify diagnostic markers in hopes of developing new treatments.

For the new NIH study, Hosom and Shriberg will examine the speech patterns of 3 to 8-year-olds whose speech has previously been recorded as part of Shriberg's long-term studies.

"Using a computer, we're going to automatically measure certain characteristics of a child's speech and determine whether those characteristics are consistent with apraxia of speech," said Hosom. "So, for example, when a child with apraxia of speech says the word 'ladder,' the duration of the 'ae' and 'er' sounds may be nearly equal with both having strong energy, rather than just the first syllable.

"We hope to develop a diagnostic tool for childhood apraxia of speech by automating existing measurements of the speech signal, developing new measurements and combining these different sources of information into a single diagnostic tool," said Hosom.

No reliable guidelines currently exist to help people diagnose childhood apraxia of speech. "Improving the reliability of the factors involved in diagnosing the disease is a very important by-product of this study," said Hosom. "If we can develop an automated diagnostic tool, we can reduce the human variability that makes this such a difficult disorder to identify, and get these children into treatment as early as possible."

Said Jan van Santen, Ph.D., director of OGI's Center for Spoken Language Understanding, "Paul's project is a good example of the new direction for both the Center and OGI as a whole--not just to apply existing technologies to health problems, but to invent entirely new technologies that we would have never thought of if we were not looking at health issues."