Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Evaluating a causal role for right temporal cortex in accessing talker-specific phonetic signatures
Poster Session E, Sunday, September 14, 11:00 am - 12:30 pm, Field House
Sahil Luthra1, Hannah Mechtenberg2, Hannah Olson2, Holly Zaharchuk2, Alexa Feder2, Emily B. Myers2; 1Carnegie Mellon University, 2University of Connecticut
Individual talkers may exhibit substantial variation in how they produce their speech sounds, and listeners are exquisitely sensitive to the idiosyncratic phonetic signatures of different talkers. For example, listeners can explicitly indicate whether the voice-onset time (VOT) of a voiceless stop consonant is typical or atypical of a given talker (e.g., whether a talker tends to produce sounds like /k/ with a relatively short or long VOT; Allen & Miller, 2004). Such sensitivity to talker-specific phonetic variation may be advantageous for conditioning speech perception on talker information. A growing body of neuroimaging data suggests that the integration of talker information during speech perception recruits the brain’s right hemisphere, especially the right temporal cortex. For example, one study found that during a phonetic categorization task, the right middle temporal gyrus (MTG) was differentially activated for VOT variants that were typical or atypical of a given talker (Myers & Theodore, 2017). However, an open question is whether recruitment of the right hemisphere is strictly necessary for integrating talker information and phonetic detail or simply incidental to this integration process. Recently, our team conducted a transcranial magnetic stimulation (TMS) study (Luthra et al., 2023) to test for a causal role of the right posterior temporal cortex in learning what phonetic variation is typical of a talker. During an initial exposure phase, participants heard two talkers, one of whom produced voiceless stop consonants with a relatively short VOT and one with a relatively long VOT. During exposure, TMS was applied to the right MTG, left MTG, or a control site (scalp vertex). In a subsequent test phase, participants judged whether VOT variants were typical or atypical of a learned talker. Listeners exhibited near-ceiling performance in this talker typicality task regardless of stimulation site; disruption to the right MTG did not have consequences for learning which phonetic variants were typical of a talker. This finding leaves the specific contributions of the right MTG to talker-specific phonetic processing unclear. In the current study (planned N=36, data collection ongoing), we follow up on our previous work to further characterize the role of the right temporal cortex in talker-specific processing. Specifically, we test the hypothesis that the right MTG may play a causal role not in learning what phonetic variation is typical of a talker but rather in accessing a listener’s beliefs about a talker’s phonetic signature. The design of our study closely resembles our previous work, but with the important difference that we now apply TMS during the test phase (talker typicality judgment) rather than during the exposure phase. Preliminary results (N=17) indicate that listeners may be best able to determine what phonetic variation is typical of a talker following right MTG stimulation and worst following left MTG stimulation. If this pattern holds, such a finding would suggest that the right and left MTG may play distinct causal roles in accessing talker-specific phonetic beliefs. Ultimately, results will provide insight into the differential contributions of the brain’s two hemispheres for conditioning speech perception on talker information.
Topic Areas: Speech Perception,