Poster E14, Saturday, August 18, 3:00 – 4:45 pm, Room 2000AB
Neurophysiological correlates of directional asymmetries in adult vowel perception: An auditory brainstem study
Matthew Masapollo1, T. Christina Zhao2, Linda Polka3,4, Lucie Ménard4,5;1Boston University, 2Institute for Learning & Brain Sciences, University of Washington, 3McGill University, 4Center for Research on Brain, Language & Music, 5University of Quebec at Montreal
Cross-language studies of vowel perception with adults and infants have consistently shown that for many vowel contrasts, discrimination is easier when the same pair of vowels is presented in one direction compared to the reverse direction. According to one account, these directional asymmetries reflect a universal bias favoring “focal” vowels (vowels with adjacent formants close in frequency, Polka & Bohn, 2011). An alternative, but not mutually exclusive, account is that such effects reflect an experience-dependent bias favoring prototypical exemplars of native-language vowel categories (Kuhl, 1991; Miller & Eimas, 1996). In a recent study designed to distinguish these two accounts, Masapollo, Polka, Molnar and Ménard (2017) identified a region in phonetic space where vowels were consistently categorized as /u/ by both Canadian-English and Canadian-French listeners, while French /u/ prototypes were more focal compared to the English ones. Results showed that English and French listeners performed better at discriminating changes from less-focal to more-focal /u/ compared to the reverse, regardless of prototypicality, supporting the focalization account. The present investigation examined the effects of focalization and prototypicality on subcortical neural encoding and discrimination of vowels. Using a passive oddball/reversed oddball paradigm, we examined the frequency-following response (FFR) in the auditory brainstem to a less-focal/English prototypic /u/ and a more-focal/French prototypic /u/. Sixteen healthy monolingual English speakers were tested. The stimuli were synthesized with the Variable Linear Articulatory Model (Ménard et al., 2004) using the same parameter values from Masapollo et al. (2017). Each vowel was 100-ms long with a 10-ms onset/offset ramp and had a mean fundamental frequency (f0) of 130 Hz. The mean F1/F2 values for the English /u/ prototype were 300 and 979 Hz, and 275 and 746 Hz for the French prototype. All other acoustic properties were identical. We utilized a standard complex ABR recording method (Ground-Forehead, Ref-Right ear lobe, CZ). We conducted two analyses to examine the FFR as a function of stimulus type (English vs. French prototype) and condition (Standard vs. Deviant). First, we conducted a cross-correlation analysis to quantify the overall similarity between the stimuli and the FFR. Overall, higher correlational values were observed for the English prototype than the French prototype, suggesting more robust neural encoding of the native prototype. Yet, at the same time, deviants overall exhibited higher correlational values than standards. This effect was largely driven by the French prototype, suggesting an influence of focalization on discrimination. Second, we examined the magnitude of the individual harmonics in the frequency domain by comparing the spectral peaks at the f0 and the first harmonic (h1) (the peak within the F1 region). For the f0 peak, only an effect of native-language was observed (larger f0 peak for English prototype), whereas for the h1 peak, only a condition effect was observed (larger h1 for deviant than standard). The latter effect was largely driven by the French prototype. Altogether, these findings suggest that the subcortical encoding and discrimination of vowels in adulthood is shaped by a complex interplay between universal biases and experiential influences.
Topic Area: Perception: Speech Perception and Audiovisual Integration