Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Neural Encoding of Sensory Prediction and Motor Performance Errors in Speech: Evidence from M100 and One-Shot Adaptation
Poster Session E, Friday, October 2, 11:00 am - 1:00 pm, Wangari Maathai
This poster is part of the Sandbox Series.
Caroline Niziolek1, Yubin Zhang1, Ding-Lan Tang2, Benjamin Parrell1; 1University of Wisconsin–Madison, 2University of Hong Kong
In motor control, two types of errors are often distinguished: sensory prediction error (SPE), corresponding to the mismatch between predicted and received sensory consequences, and motor performance error (MPE, or task error), which reflects deviation from a goal-directed target outcome. In speech, however, how these error signals are encoded is not well understood. Auditory cortical responses during speech production, specifically the speaking-induced suppression (SIS) of the response to speech onset compared with a passive listening baseline, have been used as a marker of auditory prediction error, but it is unknown to what extent this signal reflects sensory vs. task errors. In this study, we utilize a real-time auditory perturbation paradigm that shifts auditory feedback of vowel formants either toward the center (median) of a vowel’s distribution (inward perturbation) or away from it (outward perturbation). Both perturbations elicit SPE of the same magnitude, but the inward perturbation reduces MPE while the outward perturbation increases it. By comparing MEG and behavioral responses under these conditions, we test how different error types are encoded neurally at an early auditory level and drive behavioral adaptation. Fifteen American English speakers participated in an auditory perturbation experiment in the MEG scanner. In each of three separate sessions, a baseline phase in which participants spoke with normal feedback was followed by an exposure phase consisting of: (1) a control condition without perturbation, (2) an inward perturbation shifting formants towards the center of the vowel’s distribution in formant space, and (3) an outward perturbation shifting formants away from it. Interleaved with these speaking blocks were listening blocks in which participants heard playback of their speech. In the present analysis, we focused on the neural and adaptive behavioral response during the early exposure stage. For the neural response, we localized M100 source activity in the left auditory cortex and compared peak M100 amplitudes across perturbation conditions. For behavioral analysis, we focused on “one-shot” adaptation, i.e., the change from one trial to the next in the direction that corrected for the error on the previous trial. We computed the scalar projection of the change between the two trials at vowel onset onto the perturbation vector defined on the previous trial (i.e., the vector from the previous trial’s midpoint toward the center of the vowel distribution). Preliminary MEG analyses revealed a numerical increase in M100 amplitude from baseline to early exposure speaking for both inward and outward perturbation conditions compared with control. Behaviorally, preliminary results show that the outward perturbation condition elicited a larger one-shot adaptation response than the inward perturbation and control conditions, as measured by the vector projection magnitude. This result seems to suggest that while both perturbation conditions show neural evidence of auditory error, the outward perturbation leads to a larger adaptive change to produced speech than the inward perturbation. We plan to extend our analysis to additional datasets and examine whether these effects can be replicated across different experiments.
Topic Areas: Speech Motor Control, Multisensory or Sensorimotor Integration