Summary: Machine learning algorithms help researchers identify language patterns in children on the autism spectrum that are consistent across languages.
Source: Northwest University
A new study led by researchers at Northwestern University used machine learning – a branch of artificial intelligence – to identify language patterns in children with autism that were consistent between English and Cantonese, suggesting that language features are a useful diagnostic tool of the disease could be.
The study, conducted with collaborators in Hong Kong, provided insights that could help scientists distinguish between genetic and environmental factors that affect people with autism’s communication skills and potentially help them learn more about the origins of the condition experience and to develop new therapies.
Children with autism often speak more slowly than normally developing children and have different differences in pitch, intonation, and rhythm. But these differences (dubbed “prosodic differences” by researchers) were surprisingly difficult to characterize in a consistent, objective way, and their origins remained unclear for decades.
However, a team of researchers led by Northwestern scientists Molly Losh and Joseph CY Lau and Hong Kong-based collaborator Patrick Wong and his team successfully used supervised machine learning to identify language differences associated with autism.
The data used to train the algorithm were recordings of English- and Cantonese-speaking youth with and without autism telling their own version of the story set in a wordless children’s picture book called “Frog, Where Are You?”
The results were published in the journal Plus one on June 8, 2022.
“When you have languages that are so structurally different, any similarities in language patterns seen in autism in both languages are likely traits that are strongly influenced by genetic propensity for autism,” said Losh, who writes the Jo Ann G. and is the Peter F. Dolle Professor of Learning Disabilities at Northwestern.
“But equally interesting is the variability we observed, which may indicate features of language that are more malleable and potentially good targets for intervention.”
Lau added that using machine learning to identify the key elements of language that predict autism represents a significant advance for researchers who have been limited by the English language bias in autism research and people’s subjectivity in classifying language differences between people with autism and those without.
“Using this method, we were able to identify language features that can predict the diagnosis of autism,” said Lau, a postdoctoral researcher working with Losh in Northwestern’s Roxelyn and Richard Pepper Department of Communication Sciences and Disorders.
“The most outstanding of these characteristics is the rhythm. We hope this study can be the basis for future work on autism that uses machine learning.”
The researchers believe their work has the potential to contribute to a better understanding of autism. Artificial intelligence has the potential to make autism diagnosis easier by helping reduce the burden on healthcare professionals and making autism diagnosis accessible to more people, Lau said. Because of the computer’s ability to quantitatively analyze words and sounds independent of language, it could also be a tool that could one day transcend cultures.
Because the features of language identified by machine learning include those specific to English and Cantonese as well as to a language, machine learning could be useful for developing tools that not only identify aspects of language that are amenable to therapeutic interventions, but also measure the impact of these interventions by assessing a speaker’s progress over time.
Finally, the study’s findings could aid efforts to identify and understand the role of specific genes and brain processing mechanisms involved in genetic susceptibility to autism, the authors said. Ultimately, her goal is to create a more comprehensive picture of the factors that shape the language differences of people with autism.
“One brain network involved is the auditory pathway at the subcortical level, which is really strongly associated with differences in the processing of speech sounds in the brains of individuals with autism compared to those that typically evolve across cultures,” Lau said.
“A next step will be to identify whether these processing differences in the brain result in the behavioral patterns observed here and the underlying neural genetics.” We are excited to see what lies ahead.”
About this AI and ASD research news
Author: Max Witnski
Source: Northwest University
Contact: Max Witynski—Northwestern University
Picture: The image is in the public domain
Original research: Open access.
“Cross-language patterns of language-prosodic differences in autism: A study in machine learning” by Joseph C.Y. Lau et al. PLUS ONE
Cross-language patterns of language-prosodic differences in autism: A study in machine learning
Differences in speech prosody are a widely observed feature of autism spectrum disorder (ASD). However, it is unclear how prosodic differences in ASD manifest in different languages, showing cross-language variability in prosody.
Using a supervised analytical approach with machine learning, we examined acoustic features relevant to rhythmic and intonatory aspects of prosody derived from narrative samples collected in English and Cantonese, two typologically and prosodically distinct languages.
Our models demonstrated successful classification of ASD diagnosis using rhythm-related features within and between both languages. The classification with features relevant to intonation was significant for English but not for Cantonese.
The results highlight differences in rhythm as an important prosodic trait influenced in ASD, and also show important variability in other prosodic traits that appear to be modulated by language-specific differences such as intonation.