Linking Auditory Processing and Lexical Representation Via Phonological Discrimination

Background: Auditory Brainstem Responses (ABRs) are auditory evoked potentials recorded on the scalp that reveal the earliest stages of auditory processing. Prolonged conduction time to click-evoked ABRs as well as differences to speech-evoked ABRs have been found in children with autism spectrum disorders (ASD). The Auditory Stability Hypothesis of Language Development, proposes that inconsistent neural processing of sound negatively impacts the ability to form a stable representation of the auditory world, thus disturbing higher-level functions, such as language, that depend on that sensory signal. Unstable neural responses to sound may be characteristic of reading-impaired populations, and recent evidence suggest that other language-impaired individuals, including those with ASD, may likewise have unstable neural responses to sound, suggesting that stability of auditory processing may support language development in both typical and atypical populations.

Objectives: The purpose of this project is to study the connections between auditory processing and language learning using a phonological discrimination task, which by hypothesis links early auditory processing and macro-level language competency. The overarching goal is to understand the role of consistency of sound processing in early word learning by looking at its effect on phonological discrimination.

Methods: Sixteen children participated in the study, eight typically developing (TD) (two girls) with a mean age of 11.25 (1.9) years and eight diagnosed with ASD (one girl) with a mean age of 12.5 (3.59) years. All participants were first screened for normal hearing thresholds. ABRs were recorded in response to a “da” stimulus (10.9/sec, 6000 trials) presented at 80 dB SPL. To assess phonological discrimination, pairs of bisyllabic CVC-CVC (consonant-vowel-consonant) novel words that differed by only one phonemic unit (biskar vs. bisdar) or were identical (selzim vs. selzim) were presented, and the child was asked if they were the same or different. Current language ability was measured using four subtests of the Clinical Evaluation of Language Fundamentals 5 (CELF-5); Word Classes, Formulated Sentences, Repeated Sentences, and Following Directions.

Results: The TD group performed significantly more accurately on phonological discrimination than the ASD group (t(14)= 2.84, p= 0.023). Neural response consistency was positively correlated with phonological discrimination ability in the ASD group (r=0.825, p= 0.014) but not in the TD group (r= 0.213, ns). Bivariate correlations including both groups found that children with better phonological discrimination also had higher standard scores on all CELF 5 subtests (rs> 0.747, ps < 0.001). The relationship between neural response consistency and language ability (Formulated Sentences and Word Classes) was mediated by phonological discrimination, as indicated by a Sobel test (z’s>2.28, ps <0.02).

Conclusions: Phonological discrimination appears to mediate the relationship between neural response consistency to speech sounds and current language ability. Neural consistency may engender firmer phonological representations that encourage word-learning as well as more advanced language development. As group differences were not observed between TD and ASD this may indicate that similar processes are at work in both typical and atypical populations. Greater heterogeneity within the ASD sample may account for more pronounced relationships between phonological discrimination and neural responses.