Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Individual Differences in Neural Speech Encoding Are Present at Birth
Poster Session E, Friday, October 2, 11:00 am - 1:00 pm, Wangari Maathai
Shaoqi PAN1,2, Leyan Xie1,3, Fei Yuan1,4, Tak Fei Mau1, Hugh Simon Lam5, Tak Yeung Leung6, Nikolay Novitskiy1,4, Ting Fan Leung5,7, Patrick Wong1,4; 1Brain and Mind Institute, The Chinese University of Hong Kong, Hong Kong SAR, China, 2Department of Psychology, The Chinese University of Hong Kong, Hong Kong SAR, China, 3Department of Education, The Chinese University of Hong Kong, Hong Kong SAR, China, 4Department of Linguistics and Modern Languages, The Chinese University of Hong Kong, Hong Kong SAR, China, 5Department of Paediatrics, Prince of Wales Hospital, The Chinese University of Hong Kong, Hong Kong SAR, China, 6Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong, Hong Kong SAR, China, 7Hong Kong Hub of Paediatric Excellence, The Chinese University of Hong Kong, Hong Kong SAR, China
Introduction: Language acquisition theories emphasize the role of postnatal input in shaping native-language learning. Consistent with this view, developmental research has shown that infants are broadly sensitive to speech contrasts early in life, before perception becomes increasingly tuned to native-language categories over the first year. However, speech perception is not shaped by experience alone: some contrasts are more perceptually distinct and may therefore be easier to discriminate. Newborns also show sensitivity to speech structure and language-relevant regularities, indicating that speech-processing mechanisms are active from the earliest stages of life. This raises an important question: do newborns begin life with relatively uniform speech-processing capacities, or do they already differ in the precision with which they encode and discriminate speech sounds? Here, we used frequency-following responses (FFRs) to test whether neonates show individual differences in neural encoding and discrimination of speech-tone categories, and whether this early variation prospectively predicts later language development. Methods: We recorded FFRs from 518 neonates (mean:1.64 (0–8) days) while they listened to three speech-tone stimuli: Cantonese Tone 2, Tone 4, and Mandarin Tone 3. Neural encoding strength was quantified using temporal response function (TRF) prediction correlations. Tone-pair discrimination was quantified using logistic-regression classification AUC for three contrasts: ga2–ga3, ga2–ga4, and ga3–ga4. Encoding and discrimination significance were assessed against permutation-based null distributions. Later language outcomes were available for a subset of infants using Bayley-III language scores (mean assessment age:17.33 (9.40–31.77) months). Results: Most newborns showed significant neural encoding: 498/518 infants encoded at least one tone, and 327/518 encoded all three tones. Nevertheless, encoding strength varied substantially across individuals. Encoding prevalence differed modestly across tones (Cochran’s Q = 6.88, p = .032), with ga3 encoded more often than ga2 (p = .043). Encoding strength alone marginally predicted later language outcomes in cross-validation (mean r = .146, permutation p = .071). Tone discrimination showed stronger evidence of individual heterogeneity. Newborns were distributed across one-pair, two-pair, and three-pair discrimination profiles rather than showing a uniform success-or-failure pattern. The three contrasts were not equally discriminable, with ga2–ga3 emerging as the least discriminable contrast despite acoustic similarity patterns suggesting that ga3 and ga4 should be more difficult. Encoding appeared necessary but not sufficient for discrimination: all infants who discriminated at least one contrast also showed significant encoding, P(encoding | discrimination) = 100%, whereas no infant without significant encoding showed discrimination. Pair-level encoding strength predicted discrimination AUC (β = .589, p < .001). Importantly, individual discrimination measures predicted later Bayley language scores (r = .199, p = .031), although the encoding–discrimination–language mediation pathway was not significant (p = .108). Conclusion: Neural speech processing is not uniform at birth. Although most newborns encode speech tones, they differ in encoding strength and, especially, in which contrasts they neurally discriminate. These early discrimination differences prospectively predict later language outcomes, suggesting that meaningful variability in language neurodevelopment is already present in the newborn period.
Topic Areas: Language Development/Acquisition, Speech Perception