11/7/10

DIVA

Speech science offers clues to managing the problems of stuttering and cluttering.

Speech scientists study how we transform ideas into spoken words. They construct hypothetical models to explain these processes. A recent research development will change my speech therapy. It has to do with a computer model that can actually produce both fluent and stuttered speech. This model suggests that persons who stutter may not have accurate motor commands – feedforward commands – telling the oral motor system what to do. And, another related article proposes that inadequate feedback affects speech. (1) To my mind, this means speech therapy needs to spend a whole lot more time on whatever tasks elicit the most fluency instead of pressing onward to more difficult speaking tasks.

The speech model I am referring to is called the DIVA model. (2) Computer simulations of this model produced stuttered speech when manipulated in very specific ways. I cannot explain the mathematical calculations that caused the speech changes. However, the basic idea is easy to understand…

The DIVA model proposes that our brains issue motor commands to speak. The motor command seems to be for an entire sentence, so as to plan for “co-articulation.” Co-articulation refers to the slight changes we make when pronouncing an /i/, for example, based on what sounds come before and after it. The way we articulate /i/ in the word “mine” is slightly different that the way we articulate /i/ in “fight.” The DIVA model speculates that when a person mispronounces a sound, an internal “speech monitor” tries to correct the error by turning off the voice, “repositioning the articulators,” and starting over again. (3) This causes sound and syllable repetitions.

The ‘speech monitor’ depends on feedback. We rely on feedback for a variety of tasks. When I turn on the water to wash dishes, feedback from my hands tells me if the water is too hot or too cold. I use visual feedback while driving when I look at the speedometer to check how fast I am driving. And when speaking, I rely heavily on auditory feedback to be sure my speech is clear and my words express my message.

The Fall 2010 issue of The Stuttering Foundation newsletter refers to the DIVA model in a front page article, “Feedforward Strategy in Children.” (4) The author explains that he is going to study co-articulation skills in children who stutter to learn more about their speech motor planning skills. If children who stutter show poorer co-articulation skills compared to fluent children, then this would support the theory that inadequate feedforward commands could be to blame. In other words – if this theory is correct - children who stutter find themselves in mid-sentence having to change their articulation of sounds because of faulty motor planning at the very beginning of the sentence.

The April 2010 International Cluttering Awareness Day online conference includes a paper about the DIVA model. It even includes a diagram of the model. This paper suggests that a speaker’s “…feedforward mechanism may be relatively intact, so that he is aware of the appropriate sounds to produce…however, if the feedback system that includes both auditory [hearing] and somatosensory [touch] has not been providing the appropriate feedback, then the feedforward mechanism may be ‘faulty ‘ in that it may not be properly tuned.’” The article recommends a variety of ways to help the speaker increase awareness and improve feedback monitoring.

I’ll continue to experiment with new ways to provide feedback: digital video and audio recordings, delayed auditory feedback, and Audacity® in addition to traditional token reinforcement and verbal praise. But, now I will think in terms of practicing the correct output much more frequently.
also, I will be doing alot more modeling in order to provide the fluent model. Maybe this will help fine-tune the feedforward system. As usual, results will be different for each student. It will be interesting to see what session data reveals.

(1) Leahy, M. (2010) “Monitoring feedback as you speak: how DIVA contributes to explaining a part of the problem of cluttering, and to developing a therapy plan. http://www.mnsu.edu/comdis/ical/papers/leahyc.html

(2) Boston University Speech lab, http://speechlab.bu.edu/diva.hph (This link was broken today.)

(3) Civier, O. , et. al, (2010) “Overreliance on auditory feedback may lead to sound/syllable repetitions: Simulations of stuttering and fluency-inducing conditions with a neural model of speech production” Journal of Fluency Disorders, 35, p. 266

(4) broken link
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.