Speech Screening May Help Diagnose Autism in Children

Posted on Aug 2, 2010 in Autism

July 20, 2010 by Mallory CrevelingDespite difficulties with diagnosing autism, researchers believe they have developed a new method to screen for the disorder based on how a child talks.

A team of researchers from the University of Memphis created a pocket-sized recorder that examines the words a child uses throughout the day, as well as a computer program that analyzes the child’s sounds, Reuters reports.

Currently, doctors diagnose autistic children based on an in-depth look at three problem areas, including socialization, behavior and language, according to Dr. Donald Gallo, a clinical child psychologist who specializes in autism and author of “Diagnosing Autism Spectrum Disorders: A Lifespan Perspective.” Some early symptoms of the disorder include avoiding eye contact, a disinterest in others, speech delays and repetitive movements like flapping hands or rocking back and forth. 

The authors of the study, published in Proceedings of the National Academy of Sciences, evaluated data collected from nearly 1,500 recordings of 232 children aged 10 months to 4 years who expressed more than 3 million syllables, according to Reuters.

Using the software program, the researchers correctly identified autism 86 percent of the time.

Kimbrough Oller, professor of audiology and speech-language pathology and the study’s lead author, identified how children form different syllables in their first four years of life. But based on her recent study — which looked at day-long recordings in order to analyze a child’s natural speech — autistic children do not follow the typical pronunciation patterns, according to Reuters.

Oller added that parents could record their child for a day and send the device back to the company to have it screened for language development and autism. And although the tests were conducted in English, it could apply to other languages as well.

While all children tend to mix up their syllables when learning to talk, the researchers found that autistic children did this for a longer amount of time than normal children — an attribute that allowed the computer program to easily distinguish which children were autistic, AOL News reports.

“[This software] is the first kind of system that’s totally objective,” Oller told AOL News. “I don’t know of any other system that doesn’t involve judgments being made by people.”

Both Gallo and Oller agree that this device should not be the only diagnostic tool for autism, but it could help to diagnose autistic children sooner, which would help them receive services earlier and therefore help to accelerate the child’s improvements.

Gallo told AOL Health that although other institutions need to test this software to make sure they get the same positive results, it is a promising advancement for autism.

“I’m 100 percent for all the research we can do to get better at diagnosing autism,” he said. “The more we can learn about autism, the better. And the more we can learn about how normal children develop, the more we can use that information to identify the delays of those on the autism spectrum.”

The nonprofit LENA Foundation funded this research and sells the device along with clothing to hold the recorder.