The goal of this research proposal is to develop an innovative voice simulator which can accurately and rapidly simulate highly complex spatiotemporal dynamics of glottal flow, vocal fold (VF) vibrations and acoustics in realistic geometries of the VFs and tract over a broad range of phonatory conditions. We envision that the successful development of this voice simulator can benefit both fundamental research and clinical applications in the following aspects:(1) predict/estimate the outcomes of different surgical options that may affect vibration patterns, i.e. VF injection or voice therapy, and suggesting the more efficacious therapy without relying on trial and error; (2) assist in designing new phono surgical-procedures by estimating their impacts prior to animal and human trials; (3) examine clinically relevant quantities and features that cannot be assessed by traditional diagnostic tools, such as mechanical stress; (4) establish the links between VF physiology, VF vibrations, acoustics and perception of the produced voice. Continuum VF models can represent realistic VF anatomies and vibratory dynamics, thus having a great potential for assisting in clinical intervention. However, its current use in modeling voice disorders is very limited due to the lack of an accurate and rapid computation of flow pressures inside highly irregular glottal shapes. Most past voice production models assumed symmetric, inviscid 1D flow and a single-channel glottis to justify the use of the simplified Bernoulli equation. However, these assumptions would be erroneous in highly irregular glottal shapes, such as the hourglass shape or severely asymmetric shapes, in which the intraglottal flow can be severely curved and the glottis often presents multiple flow channels. The Navier-Stokes (N-S) equation on the other hand considers the full complexity in glottal flows and can predict accurate pressures in any shapes. Unfortunately, it suffers from a very high computational cost that only allows explorations of limited behaviors.
This research aims to address this problem by developing a simplified flow model that can provide an accurate and near real-time solution of flow pressures inside varied glottal shapes. We propose a simplified flow model which assumes curved streamlines based on the variation of the cross-sectional shapes. The effect of curvatures will be modeled through centripetal pressures. The effect of multiple flow channels will be modeled using a serial-parallel-network approach. Furthermore, the effects of dynamic flow separations and viscous losses, which are generally missing in the Bernoulli equation, will be included through a series of empirical functions. The empirical functions will be developed by using a large number of high-fidelity, N-S equation based flow simulations in varied glottal configurations. The new flow model will be coupled to a fiber-gel VF tissue model and acoustic wave propagation model to simulate flow-structure-acoustics interactions (FSAI). The model will be quantitatively validated against scaled-up physical models and excised canine models, and verified against an N-S equation based FSAI model in both static and dynamic configurations.
A computer model which can represent realistic vocal fold (VF) physiology/pathology and accurately simulate the spatiotemporal dynamics of glottal flow, VF vibrations and acoustics over a broad range of phonatory conditions can be useful for studying the relationships between VF physiology/pathology, VF vibratory dynamics, acoustics and speech perception. Such knowledge is essential for understanding how VF physiology changes finally affect the produced voice. It also can help better predict the voice outcomes of different clinical management of voice disorders, thus improving both diagnosis and treatment.
