Healthy speech requires proper function of the velopharyngeal mechanism, which consists of the hard palate, velum (soft palate), lateral and posterior pharyngeal walls, and the musculature that drives movement of these structures. The function of the VP mechanism is to close the VP port, an orifice between the velum and pharyngeal walls, and completely separate the nasal and oral cavities. When the VP mechanism cannot achieve this closure, VP dysfunction results. In speech, VPD manifests as hypernasality, nasal air emission, and fatigue, in addition to frequent unintelligibility. VP dysfunction is a common occurrence for children with repaired cleft palate; approximately 1 in 1000 children are born with cleft palate and 25% have VP dysfunction after primary palate repair. VPD is also associated with 22q11.2 deletion syndrome, and VPD in this population is not easily treated. The critical barrier for prevention and effective treatment of VPD has been the primarily observational nature of studies examining the VP mechanism. Insights into VP function and its relationship with anatomy are limited by the number and types of measurements that are feasible in vivo, and causal relationships cannot be examined due to the sheer number of clinical cases required to isolate the effects of pre-repair or surgically reconstructed anatomy.
Our understanding of the complex relationship between VP structure and in vivo function remains limited. In this dissertation, I developed methods to empower this investigation. The levator veli palatini is the primary muscle of VP closure, so knowledge of its in vivo function is essential to understanding VP mechanics during speech. Therefore, I developed a method to measure LVP lengths and velocities during speech production using dynamic MRI. Results obtained using this method revealed that LVP shortening and contraction velocity scale with VP port depth. In the 22q11.2 DS population, the relationship between anatomy and function is likely more complex than in healthy individuals. Simulations using a computational model optimized for anatomical parameter sensitivity revealed that LVP cross-sectional area is a disadvantageous feature in all 22q11.2 DS anatomies. However, no other anatomical measure was consistently disadvantageous for VP closure across all anatomies, supporting an anatomy-informed, rather than “one size fits all”, approach to treatment of VPD in children with 22q11.2 DS. Finally, we do not understand how each muscle of the VP mechanism affects VP closure, which limits our ability to prevent and treat VPD. I developed a novel MRI-finite element modeling framework to probe the roles of two VP muscles – the palatopharyngeus and palatoglossus – in VP closure.
Ultimately, the framework developed in this dissertation integrates the wealth of literature with MRI-based anatomy and validation to provide new insights into VP biomechanics. Coupling imaging and computational modeling empowers us to unravel the complexities of the VP structure-function relationship and improve the lives of children born with cleft palate and those living with VPD.