Predicting phonological outcomes in late talkers: the role of speech perception and production skills

The National Institute on Deafness and Other Communication Disorders awarded Purdue University $159,116 on September 30, 2025, under the Research Related to Deafness and Communication Disorders program (CFDA 93.173) to investigate patterns of speech perception and production abilities among late-talking children and whether these skills predict later phonological development and speech-sound disorder.

The project examines a large cohort of both late talkers and typical talkers at ages three and four years, measuring speech perception at age three using pupillometry mispronunciation tasks and measuring speech production at both ages using fine-grained acoustic analysis. Late talkers—children with expressive vocabulary delays in the absence of neurodevelopmental disorders—frequently show phonological delays including later onset of canonical babbling, reduced phoneme inventory, reduced syllable structure inventory, and atypical phonological patterns. Up to 50 percent of late talkers are later classified as having a speech-sound disorder, with associated negative developmental outcomes including delays in preliteracy skills, phonological awareness, and psychosocial difficulties. Current diagnostic measures cannot accurately forecast which late talkers will develop speech-sound disorder, making it difficult to allocate intervention resources efficiently. This research aims to identify whether late talkers' early phonological delays stem from underlying difficulties in speech perception, speech production, or both.

Performance of research occurs in Indiana. The award period runs from September 30, 2025, through September 29, 2026. This is a Project Grant, an assistance type supporting research conducted by eligible institutions.

Abstract

Late talkers (LTS) are a heterogeneous group of children who display delays in their expressive vocabulary development in the absence of neurodevelopmental disorders.1 in addition to their delays in vocabulary growth, LTS frequently show delays in multiple areas of phonological development. These delays include a later onset of canonical babbling,2 reduced phoneme inventory,3 reduced syllable structure inventory,4,5 and the presence of atypical phonological patterns.6 up to 50% of LTS will later be classified as having a speech-sound disorder (SSD),2,5,7 with associated negative developmental outcomes including delays in preliteracy skills,8 phonological awareness,9 and psychosocial difficulties.10 while effective early speech-sound interventions exist for young children,11 current diagnostic measures cannot accurately forecast which LTS will later develop SSD, making IT difficult to efficiently allocate intervention resources early to mitigate these negative outcomes.12 to become competent language users, children must become adult-like in both speech perception, (e.g., the ability to differentiate between speech sounds), as well as in speech production (i.e., the ability to accurately produce speech sounds). LTS' early phonological delays could be due to underlying difficulties in either speech perception, speech production, or both. The primary objective of the proposed project is to explore patterns among speech perception and production abilities amongst children with a history of LT and whether these skills predict later phonological development and the presence of SSD. A large group of both LTS and typical talkers will complete study visits at 3 and 4 years of age. At age 3, we will measure speech perception using a pupillometry mispronunciation task. At age 3 and 4, we will measure speech production using fine-grained acoustic analyses of children's speech productions that can detect subtle articulatory differences that may be missed on perceptual-based measures. Aim 1 will assess the concurrent relationship between speech perception and production on both early-acquired (/k/-/t/) and later-acquired (/s/-/ʃ/) sounds at age 3. Aim 2 will assess the degree to which age-3 speech perception and speech production skills predict age-4 speech production skills. Aim 3 will utilize latent profile analyses to identify unique profiles of speech perception and speech production skills at age 3 that predict the presence of SSD at age 4. This project will contribute to current theoretical models of speech production by empirically testing the association between perception and production both concurrently and longitudinally and will identify profiles of skills at age 3 that can be used to assess risk for later SSD.