What is it about?
People with autism may have difficulties developing social language skills, particularly narrative and conversational abilities. The study investigated how people with autism tell stories. The researchers used natural language processing techniques to analyze the narratives of individuals with autism and those with typical development. Computational analysis confirmed differences between narratives in the two groups, including the use of word with emotional component and the level of language abstraction. The results suggest that sentiment and language abstraction analysis can be useful for quantitatively assessing narrative abilities in individuals with ASD, providing insight into their language and social communication difficulties.
Featured Image
Photo by Susan Holt Simpson on Unsplash
Why is it important?
The study sheds light on the specific narrative impairments in individuals with Autism Spectrum Disorder (ASD), which is an important aspect of social communication difficulties that many people with ASD experience. By using natural language processing techniques to quantify these impairments, the study provides an objective and precise way to assess narrative abilities in individuals with ASD. Second, the study highlights the potential of sentiment and language abstraction analyses as useful tools for studying and evaluating the language and social communication difficulties that people with ASD face. Overall, the study contributes to our understanding of the specific narrative impairments in individuals with ASD and provides a promising approach to assessing and addressing these difficulties.
Perspectives
These findings show that the sentiment and language abstraction analyses seem to be promising methods to characterize selected aspects of narratives in ASD. The use of fully automated natural language processing techniques to capture this in a quantitative way is novel and in our opinion worth studying. A computational approach allows for large-scale empirical research and an in-depth characterization of large cohorts that could not be carried out using hand-coding methods, where the coding is made by a human expert. The results reported in this paper form the basis for future implementations of automated tools supporting clinicians and researchers working with ASD. Our study confirmed the usability of selected linguistic information (sentiment and language abstraction), but one may easily envision linking it with relevant non-linguistic variables to combine as an input for machine learning models related to ASD. These tools could be used to improve screening, diagnosis and intervention planning. Further studies are needed to investigate this, especially collecting multimodal and historical data.
Izabela Chojnicka
Read the Original
This page is a summary of: Social language in autism spectrum disorder: A computational analysis of sentiment and linguistic abstraction, PLoS ONE, March 2020, PLOS,
DOI: 10.1371/journal.pone.0229985.
You can read the full text:
Resources
Contributors
The following have contributed to this page
