Dysphonia (i.e., chronic hoarseness) negatively affects the speaker's intelligibility, especially in noisy communication environments. While the intelligibility deficit is clinically well-recognized, there is no method for evaluating this aspect of the disorder. Consequently, clinicians have not been able to determine the baseline performance of their patients nor keep track of their progress in improving intelligibility. This project focuses on characterizing how various types and levels of background noise affect the intelligibility of individuals with dysphonia.
We are developing a database of speech recordings from individuals with dysphonia. Our database contains linguistically comprehensive samples, which will help the development of an automatic voice and speech analysis system for clinical use.
In collaboration with Speech Technology and Applied Research Corp (speechmrk.com/), our lab is developing automatic speech analysis software for the clinical evaluation of dysphonic speech. Our system is based on the landmark theory of speech production and perception. Unlike common speech recognition software, our goal is to develop a system that informs clinicians by describing clinically relevant features of disordered speech.
Telehealth is a rapidly growing area of healthcare that has improved patient access to care, clinical outcomes, and patient satisfaction. The recent medical office closures due to the COVID-19 pandemic underscored its vital role in service delivery. However, service delivery through telehealth is not without challenges. In collaboration with Professors Yvonne Redman and Sarah Wigley, and Dr. Bridget Sweet at the School of Music, we have developed a standardized patient training program for graduate students in speech-language pathology.
We appreciate the Provost's Office for the generous funding that made this project possible.