Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Poster Slams
Speech variability is mediated by auditory error sensitivity
Poster B56 in Poster Session B and Reception, Thursday, October 6, 6:30 - 8:30 pm EDT, Millennium Hall
Ding-Lan Tang1, Benjamin Parrell1, Caroline A Niziolek1; 1University of Wisconsin-Madison
Introduction: Previous work has shown that speakers adjust their vowel formant variability when exposed to real-time auditory perturbations that increase (outward-pushing) or decrease (inward-pushing) their perceived variability, suggesting that speech variability is actively regulated. Simulations using a state-space model of motor behaviour identified two distinct mechanisms that could lead to the observed increase in variability: a direct increase in controlled variability or an increase in the sensitivity to auditory errors, which would amplify the corrective changes from one trial to the next. Here, we aim to distinguish these mechanisms by examining how these perturbations affect neural responsiveness to errors as measured by magnetoencephalography (MEG). Based on the model simulation, we hypothesise that exposure to only the outward-pushing perturbation will result in an increased sensitivity to auditory error. Methods: Acoustic, MEG, and structural MRI data were collected from 8 participants (data collection ongoing; target sample size is 15). In each of 3 separate sessions (control, inward-pushing and outward-pushing perturbation, order counterbalanced), participants produced and listened to playback of the stimulus words ease, add and odd during 3 phases (baseline, exposure and test; perturbation applied only during exposure phase). As a neural measure of auditory error sensitivity, we use the magnitude of speaking-induced suppression (SIS). SIS is calculated by subtracting the peak auditory evoked response (M100) to subjects’ own speech from the peak response to playback of the same acoustic stimuli, and is thought to arise from the cancellation of incoming auditory feedback during speaking by a prediction of that feedback. Such suppression is diminished when the incoming feedback has been altered to create a mismatch (i.e. prediction error). Thus, the magnitude of SIS has been argued to reflect neural sensitivity to auditory prediction error: speakers who are more sensitive to the auditory errors would generate less suppression. Results: Consistent with previous behavioral results, both inward and outward perturbations caused participants to increase their produced variability during the exposure phase. At the neural level, SIS in left auditory cortex was attenuated only after exposure to outward perturbations (decrease of 20.5 ± 14.7% s.e.), suggesting speakers became more sensitive to auditory errors in this condition. Conversely, SIS increased by 10.6% (± 15.7 s.e.) after exposure to inward-pushing perturbations, and by 31.4% (± 33.3 s.e.) in a control session with no perturbations. We also plan to examine whether the change in neural responsiveness to errors (i.e. SIS) is correlated with the change in behavioural variability. Conclusion: Auditory cortical responses to speech productions were less suppressed after perturbations that increase perceived variability, supporting our hypothesis that the increase in produced variability induced by this perturbation is mediated by a change in auditory error sensitivity. The current results not only enhance our basic understanding of the speech motor system, but also establish the validity of modulating the nervous system’s sensitivity to errors by alterations of perceived variability. Modifying error sensitivity has the potential to enhance speech motor learning and rehabilitation.
Topic Areas: Speech Motor Control, Speech Perception