Pathological Voice Source Analysis System Using a Flow Waveform-Matched Biomechanical Model

Area of Science:

• Biomechanics of voice production
• Nonlinear dynamics in biological systems
• Pathological voice analysis

Background:

• Voice disorders stem from vocal cord lesions affecting vibration and airflow.
• Accurate modeling of vocal cord dynamics is crucial for understanding voice pathologies.
• Existing models may not fully capture the nonlinear complexities of pathological voice.

Purpose of the Study:

• To design and validate a pathological voice source analysis system.
• To investigate nonlinear characteristics of vocal cord vibration in pathological voices.
• To analyze acoustic parameter changes due to subglottal pressure and vocal cord paralysis.

Main Methods:

• Integration of nonlinear dynamics with an optimized asymmetric two-mass model.
• Utilizing a fitting procedure combining genetic particle swarm optimization and a quasi-Newton method.
• Analysis of sustained vowel /a/ samples from the MEEI database for normal and pathological voices.

Main Results:

• The proposed biomechanical model accurately reproduces vocal cord vibration.
• Paralyzed vocal cords were shown to increase model coupling stiffness.
• Distinct subglottal pressures and varying degrees of paralysis alter acoustic parameters like fundamental frequency.

Conclusions:

• The developed system effectively models pathological voice production.
• The study provides insights into the biomechanical effects of vocal cord paralysis.
• The optimized asymmetric two-mass model offers high accuracy in simulating vocal cord vibration.