Rules for controlling low-dimensional vocal fold models with muscle activation

Abstract

A low-dimensional, self-oscillation model of the vocal folds is used to capture three primary modes of vibration, a shear mode and two compressional modes. The shear mode is implemented with either two vertical masses or a rotating plate, and the compressional modes are implemented with an additional bar mass between the vertically stacked masses and the lateral boundary. The combination of these elements allows for the anatomically important body-cover differentiation of vocal fold tissues. It also allows for reconciliation of lumped-element mechanics with continuum mechanics, but in this reconciliation the oscillation region is restricted to a nearly rectangular glottis (as in all low-dimensional models) and a small effective thickness of vibration (<3 mm). The model is controlled by normalized activation levels of the cricothyroid (CT), thyroarytenoid (TA), lateral cricoarytenoid (LCA), and posterior cricoarytenoid (PCA) muscles, and lung pressure. An empirically derived set of rules converts these muscle activities into physical quantities such as vocal fold strain, adduction, glottal convergence, mass, thickness, depth, and stiffness. Results show that oscillation regions in muscle activation control spaces are similar to those measured by other investigations on human subjects.