Summary:
UCLA researchers in the Department of Bioengineering have developed a novel, machine learning-assisted wearable loudspeaker to capture and translate muscle movement into audible speech.
Background:
Dysphonia, or the impairment of voice production, is commonly caused by disorders of the vocal folds or postoperative recovery of laryngeal surgery and causes significant communication difficulties. It is estimated that 30% of people will suffer from a form of dysphonia during their lives. Current treatments include vocal therapy and surgery, after which a period of absolute voice rest of up to three months is required. Existing devices for translating muscle movement into audible speech are often inconvenient, invasive, or uncomfortable. There is an obvious need for a comfortable, low-profile device capable of generating speech in patients suffering from and recovering from dysphonia.
Innovation:
UCLA researchers led by Professor Jun Chen have created a lightweight, waterproof system proficient in generating speech without the use of vocal folds through lip-synching or regular speech. The device measures just 1.2 inches across, weighs 7.2 grams, has mechanical properties that mimic skin, and has high stretchability to ensure adhesion to the skin. The sensing component effectively captures extrinsic muscle movement across all three dimensions and converts them into high-fidelity electrical signals. These signals are converted to speech with the assistance of machine learning with an accuracy of 95%. This innovative device facilitates the restoration of voice function and increases the quality of life in patients with vocal fold dysfunction.
Press Release:
UCLA Newsroom: Speaking without vocal cords, thanks to a new AI-assisted wearable device
Potential Applications:
• Postoperative laryngeal surgery recovery
• Vocal fold disorder recovery
• Vocal fold disorder pre-treatment communication
Advantages:
• Lightweight
• 95% accuracy in speech generation
• Stable against perspiration
• Stretchable and flexible to move with skin
Development-To-Date:
The device has been fabricated and described in a peer-reviewed publication.
Related Papers:
Che, Z., Wan, X., Xu, J. et al. Speaking without vocal folds using a machine-learning-assisted wearable sensing-actuation system. Nat Commun 15, 1873 (2024). https://doi.org/10.1038/s41467-024-45915-7.
Reference:
UCLA Case No. 2024-087
Lead Inventor: Jun Chen