Advancing Voice Health: High-Fidelity Modeling of Human Voice Production from Neuromuscular Control to Acoustic Resonance*

The human voice produces a remarkably diverse array of sounds, making it the primary channel for human communication and a potential interface for human-computer interaction. Voice production starts with neural stimulation of the intrinsic laryngeal muscles, which control the position and biomechanics of the vocal folds. Air from the lungs interacts with the vocal folds, generating self-sustained vibrations and creating a pulsatile jet in the larynx. These glottal air pulses form the primary sound sources, which pass through the supraglottal vocal tract, reshaping the sound spectrum as an acoustic resonator. This intricate process requires precise neuromuscular control of complex interactions among glottal flow dynamics, vocal fold vibrations, and vocal tract acoustics. 

Developing a comprehensive model of voice production remains a significant challenge, limiting our ability to accurately analyze and simulate voice sounds for medical and technological applications. This talk presents a cutting-edge research endeavor focused on creating a high-fidelity computer model to simulate the entire voice production process, from laryngeal muscle contraction to flow-structure interactions, integrating realistic laryngeal anatomy and physiologically quantifiable inputs. The model has been used to investigate the cause-effect relationships between muscle contraction, glottal aerodynamics, vocal fold vibrations, and resulting voice acoustics. By capturing these intricate dynamics, the model provides critical insights into voice production mechanisms. Additionally, this work represents a significant step towards developing a patient-specific voice simulator, offering personalized simulation-guided voice management, training, and surgical planning.