Inducing Speech: The Challenges of First Words

To watch a video interview with Dr. Rhea Paul, click here.

Autism spectrum disorders (ASD) affect one in 110 children in the United States. Prominent symptoms include impaired communication and delayed acquisition of spoken language. Twenty to forty percent of preschool children with ASD do not develop speech, and rectifying this is the primary concern for many parents of these children. Children who are able to communicate primarily through speech by age 5 or 6 have more opportunities for mainstream integration, vocational opportunities, and advanced schooling than those who cannot use spoken language at the same ages.

At the forefront of this quest for communication is speech-language pathologist Dr. Rhea Paul, a professor at the Yale Child Study Center. She is interested in helping children with autism develop spoken language by studying methods of language intervention. In a recent study, she contrasted two methods of inducing speech in nonspeaking preschoolers.

Overview of Autism

Autism is a developmental disorder that was first identified in 1943 in socially isolated children with stereotypical behavior and motor patterns. As clinicians learned more about these children however, they found that autistic children show a broad range of disabilities and levels of social engagement, giving rise to the term “autism spectrum disorder.” But at the core of the disorder were social disabilities, and this aspect unifies this broad spectrum of neuropsychiatric disorders. Paul emphasizes that these children are not all alike, quoting a saying used in the field: “When you’ve met one child with autism, then you’ve met one child with autism.”

The Four Approaches

Treatments for children with autism derive from different schools of thought. The four main approaches to work with young children with ASD are the behaviorist approach, naturalistic approach, developmental-or child-centered approach, and parent training approach.

The behaviorist approach utilizes a stimulus-response-reinforcement paradigm, which attempts to change behaviors in a systematic and explicit way. While a structured and intensive behavioral program can effectively teach children, children can become dependent on the structured environment, failing to carry over the learned skills to everyday life.

By contrast, naturalistic approaches use the advantages of behavioral methods – structure, explicitness, and careful sequencing of goals – but then build these approaches into a more natural setting.

Instead of using just a stimulus-response-reinforcement paradigm, they arrange the environment with objects and activities the child finds interesting and attractive such that the child cannot access them without help from an adult. Thus, the impetus to perform the desired behavior does not come from the adult but comes instead from the child, who initiates the sequence of training.

The third developmental approach takes the opposite point of view, with goals of teaching children with autism by following their lead and identifying their interests. While effective, the research behind this is weaker. The fourth approach involves training parents. The philosophy behind this approach is that the parent is with the child most often and has the greatest opportunity to influence the child’s behavior. As a result, if parents can be trained to use these everyday natural situations, then the children will have more opportunities to learn.

The Experiment

In her experiment, Paul examined two different approaches to help children with autism learn language. She did not try to show that one was better than the other; rather she tried to enhance each of these approaches to make them more functional and successful for the autistic children and their families.

The first approach was Rapid Motor Imitation Antecedent therapy (RMIA), which is a form of classic behavioral intervention. Since it is difficult for children with ASD to produce and put together sounds, RMIA uses motor imitation as a catalyst for vocal behavior. The child is taught to imitate a series of rapid, simple motor actions before anything else and then is taught to say a simple word at the end of the motor sequence in order to request a highly preferred object. However, even this procedure is very hard for some children with ASD. Around forty percent of the children recruited for the experiment were disqualified because they were unable to master simple motor imitation.

The second approach was a standard, well-established naturalistic treatment called Milieu Communication Training (MCT), in which the clinician engineers the environment so that items of interest are in sight but out of reach, and the child has to initiate a request to solicit the help of an adult to reach an object. Children from ages three to six were randomized into one of the two treatments, with a third group as a control.

The wrinkle in this study was that, in addition to the treatments, all parents, regardless of the group, received training in naturalistic techniques that could be used at home with the child. Since it is hard for behavioral techniques to carry over to real life, this ele¬ment was built into the intervention from the beginning to see if the children would take what they had been learning from the inter¬ventions and use it outside of the therapy room in real world situations.

The children each received 36 hour-long intervention sessions – three times a week over twelve weeks. Before the study began, the children underwent an intensive assessment including standardized tests and diagnostic assessments and had a sample of their communication taken in a natural situation that revealed spontaneous word usage. This was repeated at the end of the study, comparing the number of words used at the beginning and at the end. The parents also recorded the number of words the child used at home in natural settings without prompting, both before and after treatment.

The Results

Paul first made the point that not all the children originally recruited were able to participate in the study because many could not complete the RMIA training. “Skills that lay a basis for more complex learning may be very difficult for these children to acquire, and if they can’t acquire the foundation skills, then we can’t even attempt to teach them the more complex ones.”

The study required that participating children demonstrate a low number of words, as reported by their parents. On average, both groups started with around five words used. However, after the intervention sessions, the RMIA group produced an average of 100 words, the MCT group produced an average of 75 words, and the control group revealed no significant change in words produced. Nevertheless, there was a great amount of variance in the results, reflecting the wide degree of effectiveness of the RMIA treatment.

Another variable Paul tracked was the number of words the child used in a 20-minute interview with a clinician. The clinicians were blinded to which treatment, if any, the children received, in order to minimize experimenter bias. The children started with two words, and while the control group had no change over time, the MCT and RMIA groups went to six and seven words, respectively, but again there was much variance.

Yale Child Study Center at night. Photo courtesy of the Yale Child Study Center.

Significance of the Results and Looking to the Future

One important conclusion of the study was that there is no real difference between the two treatments – both work equally well. However, it is important to remember that both treatments were supplemented with parent training. Paul reasons, “I think that the parent training is what made the children that got the RMIA – a very structured behavioral treatment – look so good on these two measures, which are measures of natural language.”

These results show that a variety of means can work and is good news for the treatment of ASD. Yet, three conditions are essential for successful treatment. First, the means must be intensive; children in the study were seen frequently for long periods. Second, the clinicians must be exceptionally well-trained, and in this study, they showed very high fidelity in the way they delivered the treatment. Clinicians received consistent feedback from experts and had many opportunities to compare ideas and receive critique and support. Third, carryover at home is important; the parents participated in generalizing what the children were learning and gave their children opportunities and encouragement to use their new skills outside of the therapy session. “It is not so much picking one treatment or another,” explains Paul, “as developing the appropriate mix of approaches.” Still, she is quick to caution that this treatment only works in some of the children who were involved in the study, and additional methods need to be developed for those unable to participate in this approach.

In the future, Paul would like to analyze the data to see what factors predict how children will do in treatment. She is trying to determine what allows one child to show strong gains and what holds other children back. In another study, which involves older individuals with ASD who can talk, Paul is evaluating a treatment that looks not only at behavioral outcomes but also at brain responses to the types of stimuli that the children worked on in the intervention in order to see if brain function changes in addition to behavioral change. Finally, Paul is looking for ways to take these interventions and apply them to younger children.

When looking at children with ASD as a whole, Paul wants people to understand that even though they may act differently, they still have the same feelings, desires, and needs as other people, even if they do not show them in the same way. They have a different way of looking at the world, and they can be better at certain things than the rest of us. Society needs to see what strengths individuals with ASD have and help them use those strengths in a functional way. As we take steps towards this vision, we can only await what Paul and the clinicians at the Yale Child Study Center discover next.

Yale speech-language pathologist Rhea Paul is interested in helping children with autism by studying methods of language intervention. Photo courtesy of Rhea Paul.

About the Author
Sudhakar Nuti is a Classical Civilizations major in Trumbull College. He is a researcher in Dr. Gil Mor’s lab studying the cellular mechanisms of ovarian cancer and loves everything about science.

The author would like to thank Professor Paul for her time, patience, and openness throughout the writing process.

Further Reading
“Autism Speaks, Be Informed, What Is Autism, FAQs.” Autism Speaks, Home Page. <>.
Chawarska, Katarzyna, Ami Klin, Rhea Paul, Suzanne Macari, and Fred Volkmar. “A Prospective Study of Toddlers with ASD: Short-term Diagnostic and Cognitive Outcomes.” Journal of Child Psychology and Psychiatry 50.10 (2009): 1235-245.
“NIMH • NIMH Statistics.” NIMH • Home. <>.
Paul, Rhea. “Interventions to Improve Communication in Autism.” Child and Adolescent Psychiatric Clinics of North America 17.4 (2008): 835-56.
Paul, Rhea. “Talk to Me: Issues in Acquiring Spoken Language for Young Children with Autism Spectrum Disorders.” American Speech-Language-Hearing Association | ASHA. <>.