UCLA engineers have pioneered a soft, flexible device that promises new hope for individuals with voice disorders, including those affected by vocal cord pathologies or recovering from laryngeal cancer surgeries.
Detailed in Nature Communications, this innovation, led by Jun Chen, assistant professor of bioengineering at UCLA Samueli School of Engineering, offers a method to convert larynx muscle movements into speech using machine learning, achieving nearly 95% accuracy.
The advancement follows Chen’s previous development of a glove that translates American Sign Language into spoken English, furthering his mission to assist those with communication disabilities.
Comprising a sensing component to capture muscle signals and an actuation component to produce voice, the device utilizes a soft magnetoelastic sensor to detect muscle movements. Encased in a biocompatible silicone compound equipped with magnetic induction coils, it translates physical movements into electrical signals and speech.
Compact and lightweight, this patch-like device easily attaches to the throat near the vocal cords, offering a non-invasive solution for voice restoration. This represents a significant advancement over current treatments, which can be invasive, uncomfortable, or require lengthy recovery periods.
The device accurately converted muscle movements into specific sentences in trials with healthy adults. This breakthrough could significantly benefit individuals with speech disorders, providing a wearable tool for communication during recovery.
Funded by several prestigious organizations, including the National Institutes of Health and the U.S. Office of Naval Research, this research paves the way for further development and testing in individuals with speech disorders, aiming to expand the device’s vocabulary through machine learning.
