A Simple 3-Parameter Model for Examining Adaptation in Speech and Voice Production

Area of Science:

• Motor control research
• Speech production
• Computational neuroscience

Background:

• Sensorimotor adaptation experiments reveal motor learning and control mechanisms.
• Speech adaptation involves auditory feedback perturbations, relying on complex feedback and feedforward control.
• Disentangling these control mechanisms in speech is challenging.

Purpose of the Study:

• To describe and test a simple 3-parameter mathematical model, SimpleDIVA, for quantifying feedback and feedforward control in sensorimotor adaptation.
• To assess the model's utility in interpreting various adaptation experiment results and predicting responses.
• To provide a mechanistic explanation for behavioral responses in speech adaptation.

Main Methods:

• Developed SimpleDIVA, a simplified adaptive neural network model based on the DIVA model.
• Utilized computer simulations to fit the model to six existing sensorimotor adaptation datasets.
• Analyzed model parameters related to auditory feedback, somatosensory feedback, and feedforward control.

Main Results:

• SimpleDIVA simulations produced excellent fits to real data across diverse perturbations and paradigms (mean correlation = 0.95 ± 0.02).
• The model successfully interpreted adaptation experiments involving formant frequencies and fundamental frequency.
• Demonstrated utility in assessing noise effects, fitting multi-dimensional perturbations, examining different timepoints, and predicting responses across experiments.

Conclusions:

• SimpleDIVA effectively quantifies the relative contributions of feedback and feedforward control in speech sensorimotor adaptation.
• Model parameters offer mechanistic insights into behavioral responses not evident from data alone.
• The publicly available SimpleDIVA software can advance speech motor control research and inform speech rehabilitation strategies.