Quantitatively characterizing reflexive responses to pitch perturbations.

Area of Science:

• Neuroscience
• Speech Science
• Computational Auditory Neuroscience

Background:

• Reflexive pitch perturbation experiments probe vocal motor control mechanisms.
• These experiments involve real-time auditory feedback of altered vocal pitch.
• Speakers naturally adjust vocal pitch to counteract unexpected shifts.

Purpose of the Study:

• Develop a quantitative model for vocal responses to pitch perturbations.
• Interpret model parameters in terms of physiological mechanisms.
• Capture both group-mean and individual subject response variations.

Main Methods:

• Established a model framework using Proportional-Integral-Derivative and State-Space/Directions Into Velocities of Articulators (DIVA) models.
• Compared the performance of 19 models against data from two published studies.
• Evaluated models on their ability to fit population-level data and individual differences.

Main Results:

• A three-parameter DIVA model best fit group-mean data and individual differences.
• This model corresponds to state-space and low-pass filter models.
• Parameters reflect auditory feedback gain/delay and somatosensory feedback gain.

Conclusions:

• Individual responses to pitch perturbations are stereotyped.
• The identified model stably characterizes population and individual vocal control.
• Future work will apply the model to communication disorders.