Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Prototypicality effects of allophonic variants on Mismatch Negativity (MMN)
Poster Session A, Friday, September 12, 11:00 am - 12:30 pm, Field House
This poster is part of the Sandbox Series.
Yasuaki Shinohara1, Ebony Goldman2, Kevin Guzzo2, Valerie L. Shafer2; 1Waseda University, 2The Graduate Center, City University of New York
[Introduction] Mismatch Negativity (MMN) has been found to reflect the discriminability of speech sounds (Dehaene-Lambertz, 1997; Näätänen et al., 1997; Winkler et al., 1999). However, MMN amplitude does not always correlate with discriminability—for example when allophonic variants are presented (Shinohara et al., 2022). We propose that MMN also reflects the prototypicality (i.e., familiarity) of the standard and deviant stimuli. When a more familiar allophonic variant is presented as the standard stimulus, the brain can more efficiently form a short-term memory trace, leading to a larger prediction error in response to the deviant stimulus—and thus a greater MMN amplitude—compared to the reverse order (i.e., an unfamiliar standard followed by a familiar deviant). Additionally, we hypothesize that allophonic variants are stored in long-term memory. Therefore, even under an adapted oddball paradigm in which MMN is not typically observed (Phillips et al., 2000), we expect MMN to be elicited. When various examples of a familiar allophone are used as standard stimuli, the brain can still categorize them as the same allophone and generate a prediction to hear another example of the same allophone. If the deviant stimulus is an unfamiliar token from the same phonological category but not perceived as the same allophone, a prediction error may still occur, eliciting MMN. We predict a larger MMN when an unfamiliar deviant is presented in a series of familiar, varying standards, compared to the reverse. [Method] [Participants] Twenty General American English and 20 Southern British English speakers will participate in both behavioral perception and EEG recording sessions. [Stimuli] A ‘fleeing’–‘fleeting’ stimulus continuum was created in Praat using a recording from a Southern British English speaker. The /t/ duration was manipulated from 0 ms to 145 ms in 5 ms increments. Shorter durations resembled the General American English flap, while longer durations resembled the Southern British English plosive. [Procedure] Participants will behaviorally judge whether each stimulus matches their native accent. Based on these responses, 11 stimuli will be selected: one ‘fleeing’, five American-accented ‘fleeting’, and five British-accented ‘fleeting’ stimuli. EEG recordings will be conducted under four experimental conditions. [Expected Results] We predict a larger MMN in American English speakers under Condition A, in which five varying flap-like /t/ sounds are used as standards and one plosive /t/ as the deviant in the word ‘fleeting’, compared to Condition B—the reverse of A, where five varying plosive /t/ sounds serve as standards and one flap-like /t/ as the deviant. For British English speakers, a larger MMN is expected in Condition B than in Condition A. In Condition C, where a single flap-like /t/ is used as the standard and a single plosive /t/ as the deviant, American English speakers are expected to show a larger MMN than in Condition D—the reverse of C (i.e., a single plosive /t/ as the standard and a flap-like /t/ as the deviant). Conversely, British English speakers are expected to show a larger MMN in Condition D than in Condition C. We make similar predictions for the cross-boundary conditions (i.e., ‘fleeting’ vs. ‘fleeing’).
Topic Areas: Speech Perception,