Poster #37 - Exploring Mechanisms of Phonetic Category Learning Through Perceptual Attunement

• In Event: Poster Session 20

Abstract

Infants begin life able to discriminate both native and non-native speech sounds. By 10-months, infants improve at discriminating similar sounding native speech sound differences (e.g. ‘ba’ vs ‘da’) and decline in discriminating speech sound differences that are not used in the native language. “Acquired Distinctiveness” (AD), in which two similar speech sounds are consistently paired with two different objects, is a perceptual learning mechanism that boosts discrimination (Werker & Tees, 1984). “Acquired Equivalence” (AE), in which two similar speech sounds are inconsistently paired (each paired with both objects), is a learning mechanism that diminishes discrimination (Honey & Hall, 1989). AD (Yeung & Werker, 2009) has been shown to be effective in changing speech sound discrimination in infants aged 6-8 months. In a recent study using the same EEG discrimination paradigm employed here, we found that a passive statistical learning mechanism, distributional learning, was effective in changing speech sound discrimination at the younger ages, but no longer effective after 10-months of age (Reh et al., 2021), consistent with the possibility of a critical period that begins closing by 10-months (Werker & Hensch, 2015). The current study was designed to test whether the efficacy of AE and AD similarly changes across the first year of life or, because they involve linking sound to meaning, they remain effective at an older age. We are testing this by comparing English-learning 6-, 9-, and 12-month-old monolingual infants. As in Reh, et al. (2021), we are using a native English, but acoustically difficult, phonetic contrast (English ‘ra’ vs ‘’la’). We are first presenting the infants with either consistent or inconsistent speech sound/object pairings, and then testing phonetic discrimination. Again, as in Reh, et al. (2021), we are measuring the discriminative response by measuring the ERP (event-related potential) response to change trials in an oddball task. To date, we have tested 60 of the proposed 80 infants, with more in the younger age (33 six-month-olds; 16 nine-month-olds; 11 twelve-month-olds). If AE/AD are effective learning mechanisms across all ages, then AD will boost, and AE will diminish, discrimination of ‘ra’ vs ‘la’ in all age groups. If the efficacy of these learning mechanisms is delimited to a 6–10-month sensitive window in perceptual development (Werker & Hensch, 2015), then these mechanisms will only be effective in the younger age period (as was shown for distributional learning in Reh, et al., 2021). The results of this study will advance our knowledge of how infants become adept “native language listeners” and will provide insight as to whether the kind of learning opportunity infants encounter (passive listening versus word learning) impacts the timing of plasticity.

Authors

• Sarvenaz Oloomi, University of British Columbia (UBC)
  Presenting Author

• Janet F Werker, University of British Columbia (UBC)