With support from a 2022 CITRIS Seed Award, UC Davis neurobiologist Lee Miller is working to restore the voices of those who have lost theirs due to head and neck cancer, stroke or other causes.
While previous devices that help people learn to speak again have used artificial sounds to recreate speech, Miller draws on technology used for interpreting gestures and controlling robotic limbs, which is based on electromyographic (EMG) signals. These signals, recorded on a person’s skin and generated by muscle contractions from basic movements, can be decoded into digital instructions and used to control devices such as a robotic arm.
Using simplified EMG signals from mouth and face muscles in combination with recorded speech from patients, Miller and his team are training a computer to match pattern to speech, resulting in a tailored technology that allows an individual to speak in their unique, original voice.
Accessibility is key for Miller and his collaborators, including Daniel Cates, an assistant professor in the department of otolaryngology at UC Davis, and Ahmed Arif, an associate professor of computer science and engineering, who foresee the system working outside of the laboratory in several years.
“Ultimately, we want this to work easily for anybody,” said Miller.