Right now, there is no "universal scientific method" to quantify a change in someone's voice quality. That is what a team of researchers at three universities, lead by MSU's College of Communication Arts & Sciences, is trying to solve so clinicians have more precise, consistent tools for accurately diagnosing and treating illness and disease.
"People have tried to do this for decades (quantify changes in voice quality) and have come as close as 50-percent accuracy. We are hoping to achieve a whole new level of precision and are shooting for 90-percent accuracy," said Rahul Shrivastav, Chair of the Communicative Sciences and Disorders department and lead researcher on the project.
"Once we achieve precision in measurement, there are all sorts of things we can do with that in terms of diagnosis, treatment and prevention. It will give us the tools that we have never had before."
The key difference between the approach taken by this research group and others lies in understanding how a person's voice is represented in a listener's ear and parts of the brain. Historically, scientists have captured a speaker's voice using a microphone. This recording is analyzed to identify characteristics that influence how it sounds. In contrast, the team led by Shrivastav uses computer programs to simulate how a speaker's voice is encoded in the neural pathways associated with hearing. This "internal representation" of the speaker's voice is used to determine the quality of sound.
The way the human ear captures sounds is very different from the way a microphone works. By replicating the biological processes in sound perception, the researchers are able to explain how voices sound with much greater accuracy.
The findings of the research also highlight a close collaboration between different experts, including those in speech production and speech-language pathology, hearing science and electrical/computer engineering. To be successful, the team needs a good understanding of how voice is produced as well as how the ear works. They need to be able to simulate complex biological processes on computers, which requires a high level of skill in programming and digital signal processing. They also need to understand how such tools can be utilized in clinical practice, requiring the team to work closely with front-line clinicians.
The research is funded by a five-year, $2.8 million grant from the National Institutes of Health. The NIH funding helps put together a team that is capable of solving such complex problems.
In the end, the researchers hope to give clinicians the tools they need to quickly and accurately capture and measure the changes in a person's voice quality. Once the team is ready, they want to test the results of their research with a small number of clinicians to get their input.
The universities sharing the grant with MSU include the University of South Florida and Medical University of South Carolina.
The other researchers involved in the project from the College of Communication Arts & Sciences include Mark Skowronski, Assistant Professor of Communicative Sciences and Disorders, and doctoral student Lisa Kopf. The team also includes David Eddins, Associate Professor of Communication Sciences and Disorders at the University of South Florida, and Heather Bonilha, Assistant Professor at Medical University of South Carolina.Share via these networks: