Vocal wow in an adapted reflex resonance model

TL;DR

This paper presents a simple reflex resonance model explaining vocal wow as a bifurcation phenomenon influenced by auditory feedback, with formulas for parameter recovery and experimental validation.

Contribution

It introduces a novel reflex resonance model for vocal wow, linking bifurcation theory with vocal modulation and providing methods to estimate feedback parameters from acoustic signals.

Findings

Vocal wow can arise from an Andronov-Hopf bifurcation in a reflex model.

Feedback gain and delay differently affect modulation depth and frequency.

Model predictions are validated with data from two subjects.

Abstract

Vocal wow is a slow 0.2 - 3 Hz modulation of the voice that may be distinguished from the 4 - 7 Hz modulation of tremor or vibrato. We use a simple model of laryngeal muscle activation, mediated by time-delayed auditory feedback, to show that wow may arise due to Andronov-Hopf bifurcation. The model demonstrates a differential effect of feedback gain and delay on modulation depth and frequency, respectively. Parametric formulas for recovering feedback parameters from the acoustic signal are presented. Interactions between reflex and auditory parameters are also assessed in a full model that includes a neuromuscular reflex loop. Model predictions are tested in two subjects