Wednesday, July 1, 2015

It’s not what you say, but how you say it




Parents and therapists have long known that even when individuals with autism spectrum disorders (ASD) meet normal language milestones, there is still something odd or different about the way they talk. It is sometimes hard “to put a finger on” what exactly differs in the language of a person with ASD, even though people can hear it. Researchers have studied language differences using language tests that examine the grammar, syntax, gestures, the ability to communicate a story in a clear and logical order, and the prosody of speech. Prosody includes the emotional emphasis a person places on a word in order to highlight how important it is to what he or she is saying—it’s the tone, pitch, or emphasis on specific syllables.

In 2010, Dr. Van Santen and colleagues at the Oregon Health & Science University began to create new tools that could measure differences in language with computer algorithms. In this first study, children ages 4-8 years of age completed different language tasks. In some of them children had to imitate words or sentences spoken by a computer, and in one task children had to correct a computer when it made a mistake in describing a picture of an animal. The computer algorithms were better than human raters at picking up very subtle prosodic differences in the tone, pitch or emphasis of how children with ASD spoke. This was particularly true when children were coming up with their own words and not directly imitating the computer.

Here at the Center for Autism Research our scientists are starting to apply this research in more natural settings. Dr. Julia Parish-Morris recently received a grant from the Autism Science Foundation to record children with ASD having a conversation with our staff. She will then use cutting-edge computer algorithms to understand what aspects of language not only identify children with ASD from typically developing children, but can also serve as a measuring stick for treatment. This study will focus on how skilled children are at taking turns in conversation, and the use of contractions.

1 comment:

  1. Excellent work! I'm looking forward to seeing your results. I will be interesting to see how children with ASD compare to children without ASD in the normal variability in developing syntax, grammar and prosody.
    - Dr. J. Stone

    ReplyDelete