Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Characterizing the effects of delayed auditory feedback and associations with related measures: A multiple case study
Poster Session E, Sunday, September 14, 11:00 am - 12:30 pm, Field House
Alexander Ocampo1, Jason W. Bohland1,2; 1University of Pittsburgh, 2Center for the Neural Basis of Cognition, Pittsburgh, PA, United States
Delayed auditory feedback (DAF) disrupts speech. However, the effects of DAF remain relatively poorly understood and vary across speakers. Here, we describe a multiple case study addressing DAF along with other related measures to characterize the sensorimotor control of speech in individual subjects. We present a detailed analysis of five native English speakers who completed several speech tasks. Speakers completed 248 trials of syllable sequence and 200 trials of sentence production tasks with 0-250 ms delays. Two phonetically balanced reading passages were used to extract habitual speaking rate (HSR). Speakers also completed an F1 perceptual acuity task, the Spontaneous Speech Synchronization (SSS) task (Assaneo et al., 2019), a reflexive formant perturbation task, and a demographic survey, providing detailed profiles of each participant. The speakers presented here were selected from a 56-participant dataset based on preliminary analyses. The selected participants exhibited HSRs within the lower, median, and upper quartiles of the full dataset and exhibited contrasting responses to DAF. The difference between the average speaking rates of 0 and 250 ms delay trials was used as a proxy for the extent of DAF’s effect on speech. Syllable sequence production was analyzed based on alignment to stimulus transcripts using Montreal Forced Aligner, with manual listening used to confirm automated measures. Log likelihood (LL) alignment scores were compared to a null model, with improvement in LL serving as the measure of interest. There is a strong positive correlation between HSR and improvement of LL at 250 ms delay (r = 0.956). This suggests slower speakers may depend more on auditory feedback during speech, therefore experiencing more induced errors under DAF. Of the five speakers, two strongly synchronized the rhythm of their whispered speech to an external syllable stream during the SSS task. These participants showed minimal change in rate across DAF conditions compared to the three non-synchronizers. There was a strong negative correlation between the participants’ SSS task phase-locking values and their reductions of speech rate with 250 ms delay. This correlation was found in both syllable sequences (r = -0.901) and sentences (r = -0.965). This relationship suggests that the non-synchronizers (who may have reduced overall sensorimotor integration), increase their dependence on feedback control in response to their auditory errors under DAF, resulting in an inverse relationship between rate and delay. However, the synchronizers (naturally stronger sensorimotor integration) may engage their feedback control system under normal speaking conditions and thus do not exhibit the same slowing. Only the two synchronizers of the five cases analyzed reported “moderate” to “extreme” proficiency in any musical instrument on the demographic survey. They also performed the best of the five participants in the F1 perceptual acuity task, both ranking within the 87th percentile of acuity scores across the larger dataset. The overall results suggest individual performance on speech production tasks is modulated by learned factors of auditory processing and sensory-motor integration. Having multiple related measures from each participant allows for a detailed analysis of individual speech patterns and how they are impacted under DAF.
Topic Areas: Speech Motor Control,