Towards a Clinically Applicable Computational Larynx Model

Abstract

The enormous computational power and time required for simulating the complex phonation process preclude the effective clinical use of computational larynx models. The aim of this study was to evaluate the potential of a numerical larynx model, considering the computational time and resources required. Using Large Eddy Simulations (LES) in a 3D numerical larynx model with prescribed motion of vocal folds, the complicated fluid-structure interaction problem in phonation was reduced to a pure flow simulation with moving boundaries. The simulated laryngeal flow field is in good agreement with the experimental results obtained from authors’ synthetic larynx model. By systematically decreasing the spatial and temporal resolutions of the numerical model and optimizing the computational resources of the simulations, the elapsed simulation time was reduced by 90% to less than 70 h for 10 oscillation cycles of the vocal folds. The proposed computational larynx model with reduced mesh resolution is still able to capture the essential laryngeal flow characteristics and produce results with sufficiently good accuracy in a significant shorter time-to-solution. The reduction in computational time achieved is a promising step towards the clinical application of these computational larynx models in the near future.

3D-FV-FE Aeroacoustic Larynx Model for Investigation of Functional Based Voice Disorders

“…As described by Sadeghi et al (2018), there must be a small area between both vocal folds of 0.5 mm 2 at GC1 to reach a numerically stable simulation. Nevertheless, this small gap still interrupts the flow through the glottis during the closed phase, as shown by Sadeghi et al (2019b). For GC2, GC3, and GC4, the initial glottal gaps possess a triangular and for GC5 a rectangular shape, see Figure 3.…”

Geometry of the Vocal Tract and Properties of Phonation near Threshold: Calculations and Measurements

“…The effects of supraglottal structures as ventricular folds or characteristic geometrical conditions associated with specific vowels (pharyngeal constrictions and partially obstructed channel exits) were neglected. Previous studies have shown that the presence of the ventricular folds significantly reduces the phonation threshold pressure as long as their positions and the gap in between is optimally selected [35,44,45]. Furthermore, the acoustical driving effect of vocal tract resonances was minimized.…”

High-fidelity continuum modeling predicts avian voiced sound production

“…During postnatal vocal development, gradual body changes, such as lung growth or VF stiffness in marmosets (46,47) or increased muscle speed in songbirds (48), can drive changes in vocal 180 behavior. Embodied human VF models have direct clinical relevance to pathological physical behaviors with abnormal vocal output (1) and patient specific model-assisted phonosurgery (49,50), because most laryngeal human voice disorders are caused by changes in VF geometry, structural integrity, or kinematics (4). The embodied approach to voice production as presented here thus improves our understanding how neural mechanisms and biomechanics interact to 185 drive vocal behavior in vertebrates.…”

“…Causally linking descending motor control to voiced sound production requires 50 computational biophysical models to explore the multidimensional control space (13), which the brain also learns to navigate. Particularly high-fidelity continuum models that include the full fluid-structure-acoustics interaction (FSAI) complexity of voiced sound production in anatomically realistic geometries of vocal fold and tract (4,(14)(15)(16)(17)(18)(19) are essential to develop when realistic representations of voice physiology and biomechanics are essential, such as in the 55 clinical management of voice disorders (1), or understanding motor control of voice (20,21) or (bird)song (22).…”