University of Texas at Dallas: Researchers Aim To Detect, Address Autism Language Delays Earlier


Research has shown that early intervention for children with autism is linked to improved development, especially in language skills. With that in mind, researchers have pushed to discover methods for detecting the potential for autism in infants in the first two years of life, before symptoms appear.

A University of Texas at Dallas researcher is spearheading a nationwide effort to identify language and communication markers for autism in infants. The project team will also use brain imaging to identify and monitor potential biomarkers for autism and will develop speech intervention strategies for children’s caregivers.

Dr. Meghan Swanson, assistant professor of psychology in the School of Behavioral and Brain Sciences, has received a five-year, $3.7 million grant from the National Institute on Deafness and Other Communication Disorders — a component of the National Institutes of Health (NIH) — to inform future interventions with a three-pronged, large-scale study.

Dr. Meghan Swanson directs the Baby Brain Lab at UT Dallas, which seeks to better understand how the infant brain develops and how to support optimal development.
“More than a third of children with autism are either nonverbal or minimally verbal for their entire lives,” Swanson said. “Even though this is no longer part of the autism diagnostic criteria, a large part of the population has significant language struggles. That’s what this grant specifically targets — supporting skills in the domain of language.”

Dr. Hervé Abdi, professor of psychology, serves as the project’s expert in advanced multivariate analysis, building the statistical method for combining audio and imaging data to allow the researchers to draw valid conclusions.

The project will include researchers from five other member institutions of the Infant Brain Imaging Study (IBIS) Network: the University of North Carolina at Chapel Hill, University of Minnesota, University of Washington, Washington University in St. Louis and Children’s Hospital of Philadelphia. IBIS, supported by the NIH Autism Centers of Excellence Program, investigates brain and behavior development of infants who have an older sibling with autism. A 2019 NIH grant for $9.5 million has supported extensive imaging work to help identify the symptoms of autism at an earlier age.

“There’s foundational work on infant language neurobiology that will be performed under this grant,” Swanson said. “From a technical standpoint, the imaging data in this project is well beyond what we have collected previously.”

Dr. Hervé Abdi will build a statistical method for combining audio and imaging data in the study.
The study involves children, starting at age 6 months, who have an older sibling with autism. The study has three aims; the first involves searching for links between the infants’ vocal sounds and either later autism diagnosis or language and communication issues.

“This methodology is the bread and butter of my lab,” Swanson said. “Our lightweight recorders worn by children are the size of a credit card. They record 16 continuous hours, picking up every sound the child makes as well as many of the words addressed to the child.”

UT Dallas’ specific role in the project is to collect and process these home language recordings. They are first analyzed by an automated processing tool, then by the researchers.

“The automated tools give us the high-level view — how many words are being addressed or spoken to or near the child. Algorithms break child vocalizations into speech-like and non-speech-like,” Swanson said. “Then researchers do a more detailed annotation of a 10-minute play session at the beginning of the recording. That’s where we get into distinguishing babbling, distress, delight, laughing and so on.”

The study’s second goal is to examine the association between caregiver speech — both quantity and quality — and later language and communication skills in the child. Swanson said that 30 years of research on the topic has made that link “irrefutable” across the general population.

“There’s foundational work on infant language neurobiology that will be performed under this grant. From a technical standpoint, the imaging data in this project is well beyond what we have collected previously.”

Dr. Meghan Swanson, assistant professor of psychology in the School of Behavioral and Brain Sciences

“We know babies learn language by hearing people talk to them, and that if you talk more and use higher quality speech, that improves their language skills,” she said. “There’s also research showing that through training, caregivers can increase how much they’re talking to their children. It’s an actionable intervention target. We can teach these skills.”

The challenge is to refine the advice given to caregivers, who may have multiple responsibilities. Swanson emphasized that her study is hoping to clarify what works best.

“If I could tell parents to do one thing, what would it be?” she said. “We don’t know the most salient aspect of caregiver speech. Most existing models are based on neurotypical children. We hope to distill down to the best kind of support for language skills for this specific population.”

The final aim of the study is to use MRI scans of children’s brains at 6, 12 and 24 months to explore the association between caregiver speech and white matter development. Swanson said this component is at the heart of the study.

“Our preliminary small-sample data shows a significant association between caregiver speech and brain development in infancy,” she said. “That’s important because understanding this mechanism can tell us all sorts of useful things. If we understand the mechanisms, for instance, we may get important information on timing in terms of ideal intervention for brain development.”

Abdi explained his role in the study as revolving around designing methods for automatically determining which of the myriad variables “are truly telling you something” and which are redundant.

“The complex data from brain imaging in this study is of the highest quality,” he said. “The major challenge from a statistical standpoint is figuring out how to put very different kinds of data together, both qualitative and quantitative, then see if you can use one to predict the other.”

Swanson said the study is unique in several ways. Among them, using MRI to detect neurological changes in specific regions of the brain may be a powerful way to show that interventions are working.

“Studies showing the neurological connection between caregiver speech and child language skills just haven’t been done because any such study is long, expensive and hard to do,” she said. “There are also limitations to the conventional ways of measuring language skills in a 2-year-old. Showing change is extremely difficult, so showing effectiveness of an intervention is equally difficult.”

Among the study’s cohort of 250 infants at high familial risk, statistics indicate approximately 50 will later be diagnosed with autism.

“Autism is increasingly prevalent, and intervening before symptoms arise can improve outcomes, and potentially improve quality of life,” Swanson said. “But we need to establish the evidence for initiating intervention. Having a well-developed sample of home language recordings with longitudinal data will allow us to ask questions of that data in innovative ways.”

Leave A Reply

Your email address will not be published.