Untangling the Language Challenges of Autism Delayed talking alone is not considered a risk factor for autism. So what accounts for high rates of language impairment in the disorder? Features
Features  |   April 01, 2017
Untangling the Language Challenges of Autism
Author Notes
  • Nancy Volkers is a freelance medical writer based in Vermont. nvolkers@nasw.org
    Nancy Volkers is a freelance medical writer based in Vermont. nvolkers@nasw.org×
Article Information
Special Populations / Autism Spectrum / Language Disorders / Features
Features   |   April 01, 2017
Untangling the Language Challenges of Autism
The ASHA Leader, April 2017, Vol. 22, 54-59. doi:10.1044/leader.FTR2.22042017.54
The ASHA Leader, April 2017, Vol. 22, 54-59. doi:10.1044/leader.FTR2.22042017.54
“My toddler isn’t speaking yet. Should I be worried?”
If you’re a speech-language pathologist working with young children, this likely sounds familiar. When a toddler hasn’t uttered a word by 24 months, many parents get understandably concerned and start raising the possibility of autism spectrum disorder (ASD).
But unless the toddler has other behavioral signs, being a “late talker” is not a risk factor for ASD.
Late talking was historically linked with autism; in fact, language delays used to be a necessary part of diagnosis. The fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM) listed three domains of autism symptoms: language impairment, social impairment and repetitive behaviors.
When the DSM-5 came out in 2013, however, it eliminated the language impairment domain. Now, the areas of impairment required for diagnosis are “social communication and social interaction across multiple contexts” and “restricted, repetitive patterns of behavior, interests or activities.”
So what do we know about language delays and impairment in autism, based on the research to date? We know that while language delays don’t predict autism, they are common among diagnosed children, with about 75 percent showing some type of language delay. We know that these delays can be dramatic; on average, children on the spectrum produce their first word one to two years later than typically developing children. And we know that language ability among children on the spectrum varies greatly—some may be fluent conversationalists, while others can barely speak.
Yet despite an explosion of research on developmental differences among people with ASD, little is known about why their language ability varies so much. Are the language issues primary, or are they effects of the social and behavioral issues? Two decades of studies are just beginning to shed some light on these questions, suggesting that both factors are at work.
Apart from language delays, children on the spectrum also have deficits in social pragmatics and discourse processes. It’s logical to think that these deficits would affect language development.
“The social interaction perspective is important—kids learn language in interactions with other people,” says Connie Kasari, professor of psychological studies in education and psychiatry at the University of California, Los Angeles. “So if socializing is impaired, language can be impaired.”

Language ability among children on the spectrum varies greatly—some may be fluent conversationalists, while others can barely speak.

Behavior observation
Among those seeking to learn more about the language issues in autism is Sudha Arunachalam, assistant professor of speech, language and hearing sciences at Boston University. In collaboration with Rhiannon Luyster at Emerson College, she’s assessing vocabulary knowledge in children with ASD by studying their eye tracking on computer screens.
“Getting kids with ASD to do standard assessments, like picking up objects or pointing, requires a lot of cooperation and motivation,” says Arunachalam. “If they’re looking at a screen and they hear something about one of the objects on the screen, it’s natural for their eyes to go there.”
One side of the screen might show a video of a girl dancing; on the other side, a girl is bouncing a ball. “I’ll ask, ‘Where is she dancing?’ and measure where they look,” Arunachalam explains.
It’s known that typically developing children can learn vocabulary by watching and listening to other people talk, but not participating. To see if children with ASD also can learn this way, Arunachalam is comparing two conditions:
  • A researcher sits across from a child with a pile of toys. The researcher picks up and talks to the child about each toy. One toy is labeled with an unusual word: “I like this garbit. It’s yellow.” The child is then given all the toys and is asked to find the garbit.

  • The researcher goes through the same process while talking to another adult. The child is in the same room, and not required to watch or listen. Once the researcher is done, the child is given the toys and asked to find the unusually labeled one.

If the assumption is that children with ASD are not interested in other people, says Arunachalam, the second scenario would result in less learning. But so far, results suggest that most children tune in and learn the new vocabulary under both sets of conditions. So, it’s possible that challenges learning vocabulary in ASD don’t necessarily stem from social detachment.
But there may be sensory processing differences, indicates research by Stephen Camarata, a professor of hearing and speech sciences and psychiatry and behavioral sciences at Vanderbilt University, and his colleagues. In one study they conducted, they presented two groups of children (with and without ASD) with various audiovisual stimuli and asked them if the “audio” and the “visual” events happened at the same time or not. The children with autism had problems processing simultaneous stimuli.
“It’s like they’re watching a movie that’s been dubbed, and the sound lags behind the picture,” Camarata explains. This lack of a “time lock” between a spoken word and a visual explanation could hamper language ability and learning.
Brain investigations
Another line of research seeks to shed light on language differences in autism by investigating the structure and function of the brain. For example, a 2014 study of 80 men with autism found that those with a history of language delay had reduced brain volume in the temporal lobe, insula and ventral basal ganglia, as well as larger brainstem structures, compared with those who had no language delays. The study’s senior author was Simon Baron-Cohen, director of the University of Cambridge Autism Research Centre.
A number of findings suggest that parts of the cerebral cortex may develop differently in people with autism. A 2016 study of boys with autism found that certain areas of the cerebral cortex acquire more folds in childhood. Boys with more severe autism symptoms also tended to have more folding. These areas are associated with social skills and language.
Other research has investigated the function of the brain during certain tasks, usually with electroencephalogram (EEG) or functional magnetic resonance imaging (fMRI). Researchers at the University of California San Diego Autism Center of Excellence, led by its co-director Eric Courchesne, used fMRI to measure the brain’s response to speech in toddlers visiting an autism clinic. Among the toddlers later diagnosed with autism, those with good language ability had strong brain responses to speech on their earlier fMRI scans. In kids with minimal language skills, the same brain regions did not activate.
Researchers from Carnegie Mellon University (CMU) used fMRI and machine learning to measure brain activation in 135 areas while participants thought about words representing social interactions, such as “hug,” “adore” and “humiliate.” They found that they could use these brain-activation patterns to diagnose autism with 97 percent accuracy.
The study included 34 young adults, 17 with autism and 17 without. When the young adults with ASD were given one of the words to think about, an area associated with the representation of self did not activate. Study author Marcel Just, director of the university’s Center for Cognitive Brain Imaging, likened their thoughts to “considering a dictionary definition or watching a play.”
The CMU study and others suggest that in people with autism, the concepts of “self” and “others” aren’t developed in the same way as in other people, speculates Kasari. This could affect the use of language, even if it doesn’t affect the size of a person’s vocabulary. Consider that there’s more to having a conversation than knowing words: We also must know how to put them together, when to speak and when to listen, and what’s expected of us in a dialogue or group setting.
While some children with autism know a lot of words, they are less skilled in how to put them to use, Kasari explains. “They will echo sentences, or pick up ‘scripts’ from TV that aren’t functional,” she says. “They may talk a lot, but [the words] may not fit the context.”

Some studies suggest that in people with autism, the concepts of “self” and “others” aren’t developed in the same way as in other people. This could affect the use of language.

Early brain changes
To find early clues about how people with ASD develop language, and how their brains develop, some researchers—like Helen Tager-Flusberg, director of Boston University’s Center for Autism Research Excellence—are focusing on the younger siblings of children with autism, starting practically from birth.
Tager-Flusberg is principal investigator with the Baby Siblings Research Consortium, which is using multiple methods, including EEG, to study this group, which has a high risk of ASD: Overall, about one in five younger siblings will be diagnosed. So far, the consortium’s results indicate differences in two areas related to language: lateralization and brain connectivity.
While most people have a language center in the left hemisphere of the brain, people with autism are more likely to have a right-hemisphere or mixed-hemisphere language localization. Tager-Flusberg’s research confirmed this finding. The implications aren’t clear, though, and there may be some confounding with handedness (left-handers tend to have right- or mixed-hemisphere language centers, too, and children with autism are more likely to be left-handed or ambidextrous).
Tager-Flusberg’s group also found that by 12 months of age, younger siblings later diagnosed with autism had the lowest measures of brain connectivity between regions associated with language, as measured by EEG. This means their brains may not be processing language as quickly or as well as the brains of people who do not have autism.
However, at least in this study, the differences didn’t translate to severe language impairments. Tager-Flusberg notes that every child in the study who showed signs of language issues received high-quality early intervention. When the study ended, 23 children had been diagnosed, but none had serious language problems.
The group is now studying 12-month-old infants at risk for ASD but without a family history, to see if they find the same types of brain differences. The infants are flagged for the study using a language and social communication screening tool.

The social communication issues inherent in autism spectrum disorder can make it tough to get reliable information. Children may struggle with motivation or attention during a study.

“There are multiple ways that language can be compromised, and ultimately we’re going to see patterns and links between different kinds of brain differences and how they relate to different outcomes,” says Tager-Flusberg.
Though research has shown that many components of language are affected in people on the spectrum, making generalizations is difficult. Most studies are relatively small, and children on the spectrum compose a heterogenous group, with some unable to communicate and others high-functioning. Plus, the social communication issues inherent in ASD can make it tough to get reliable information. Children may struggle with motivation or attention during a study. Some may focus on irrelevant aspects of the testing situation or ignore the experimenter. In a study with a small sample size, these issues are magnified.
Additionally, most studies compare children on the spectrum with a “control group” of typically developing children. The definition of “typically developing” can vary by study. Some match children by age, some by IQ, and some by other measures.
Like many conditions, autism is a complex mix of genetic and environmental factors. Social impairments may affect a child’s language because language is social; without others to communicate with, we would have little need for it. But some language issues also appear to be neurologically based, able to be nudged with intervention but not completely corrected. And of course, notes Kasari, “corrected” is in the eye of the beholder.
2017 Research Symposium Focuses on Autism

Sudha Arunachalam will present her research findings at the 2017 Research Symposium, “Advances in Autism Research: From Learning Mechanisms to Novel Interventions,” at the ASHA Convention in Los Angeles. Helen Tager-Flusberg is coordinating the symposium. Other presenters include Amy Wetherby, Jana Iverson and Susan Ellis Weismer.

Submit a Comment
Submit A Comment
Comment Title

This feature is available to Subscribers Only
Sign In or Create an Account ×
April 2017
Volume 22, Issue 4