Phase-locked characterization of intraglottal flow separation vortices and medial surface motion during phonation

This study investigates vocal fold vibration by characterizing glottal geometry and aerodynamic properties at different coronal planes along the posterior to anterior axis in excised canine larynges. Multi-plane, phase locked two-dimensional (2D) particle image velocimetry was used to reconstruct the medial glottal wall geometry and intraglottal flow fields across equidistant coronal planes during phonation. The results reveal significant spatial variation in glottal shape and flow dynamics, with larger glottal openings and divergence angles observed in the mid-coronal and anterior planes compared to the posterior plane. Flow separation vortices (FSV) consistently formed along the medial glottal wall during the closing phase but varied in size and location depending on coronal position. In the mid-coronal and anterior regions, FSV formed closer to the inferior edge and produced stronger flow entrainment, higher circulation, and greater negative gauge pressures, leading to more pronounced flow skewing during closure. In contrast, FSV in the posterior region formed near the superior aspect, exhibited weaker entrainment and circulation, and resulted in less pronounced glottal skewing. These findings highlight fluid structure interactions within the glottis and underscore the importance of considering the posterior to anterior variations to improve our understanding of voice production mechanisms.

After completing the geometric, aerodynamic, and acoustic measurements, the larynges were bisected along the sagittal plane to create hemilaryngeal configurations. The elastic properties of the vocal folds in the excised canine larynges were measured following the procedure outlined by (Dion , 2019). Each hemilarynx was secured to a custom plaster mold using cyanoacrylate adhesive. To maintain tissue integrity, the samples were kept moist with a phosphate buffered saline (PBS) solution before and during testing. To account for buoyancy effects, the load cell was zeroed with the indentation probe submerged in PBS just above the vocal fold surface.

A Biomomentum Mach-1 (Biomomentum Inc., Laval, Quebec, Canada) mechanical tester with a 150 g single axis load cell and a 7.5 mg resolution performed automated indentations on the nonplanar hemilaryngeal surface. The folds were indented 2 mm normal to the surface using a spherical indenter with a radius of 2.5 mm at each specific data point [shown by the dots in Fig. 2(b)]. The force displacement data were collected along the length (z axis) and height (H) (y axis) of the vocal folds at the superior, middle, and inferior aspects’ normalized locations: y/H = 0.83, 0.50, and 0.17, respectively [Fig. 2(b)]. A normal indentation of 2 mm was chosen since the typical glottal width for excised canine vocal folds during vibration is between 3 and 4 mm (Oren , 2021), which agrees with the collected superior glottal width measurements captured with PIV in the present study.

The surface slope of each vocal fold was determined by probing the spherical indenter at four locations around each data collection point to identify the gradient of the tissue surface. The probe was then indented normal to this gradient at the specified measurement location at a speed of 0.2 mm/s, held at maximum displacement for 1 s, and then unloaded at the same speed. Load cell data were sampled at 1 kHz throughout the loading, pause, and unloading phases.