Exploiting large language models for diagnosing autism associated language disorders and identifying distinct features

1

Large language models (LLMs) enhance the diagnosis of language disorders linked to autism by improving sensitivity and profiling linguistic features.
2

The study demonstrated over a 10% increase in sensitivity and positive predictive value using LLMs in a zero-shot learning configuration.
3

Ten key features of autism-associated language disorders, such as echolalia and pronoun reversal, were identified as critical for diagnosis and treatment.
4

LLMs can analyze ASD-related communication behaviors effectively, even in low-resource settings, without extensive labeled datasets.
5

The framework developed using LLMs supports personalized therapeutic strategies, addressing individual patient needs in autism treatment.

1