Receptive and expressive language phenotyping in over 62,000 autistic Individuals

Unraveling the phenotypic complexity of autism is essential for advancing our understanding of its underlying biology, developmental trajectories, and diverse clinical presentations. Traditional classifications of language profiles in autism distinguish three groups: 1) verbal individuals without structural language impairment, 2) verbal individuals with structural language impairment, and 3) minimally verbal individuals. However, this tripartite framework hides substantial linguistic heterogeneity. Leveraging data-driven clustering on novel comprehensive syntactic language assessment from a large cohort of over 62,414 autistic individuals aged 4–21 years, we have identified more nuanced, clinically meaningful subtypes. For receptive language, three distinct phenotypes emerged: 1) Command, limited to understanding single words and simple commands; 2) Modifier, that extends to integrating nouns with adjectives but lacks full syntactic processing; and 3) Syntactic, supporting integration of nouns with spatial prepositions and complex syntactic structures. For expressive language, four phenotypes were delineated: 1) Nonverbal, 2) Single-Word, 3) Single-Sentence, and 4) Multi-Sentence. Adopting a two-dimensional framework that separately evaluates receptive and expressive abilities offers greater precision than the conventional tripartite system. This approach better captures individual variability and can guide more targeted language interventions addressing specific strengths and deficits in each domain. Such refinement holds promise for personalized therapies and improved outcomes.