Researchers from UCLA have developed an innovative self-powered patch that translates silent speech — movements from the throat muscles — into audible speech through sophisticated machine-learning algorithms. This groundbreaking wearable technology has the potential to change the lives of individuals with speaking impairments due to various medical conditions, such as vocal fold injuries or disorders.
The human voice is generated as air passes through the throat, modulated by the complex interplay of laryngeal muscles. When the vocal cords, which are housed within the larynx, move, they can create a diverse array of sounds. Conditions like laryngitis or strain from overuse can render vocal cords less functional, inhibiting the ability to speak. However, other surrounding laryngeal muscles may still engage during attempted speech, and it’s these movements that the team at UCLA has set out to capture and interpret.
The smart patch created by the team uses several layers of materials to detect and respond to throat muscle movements. The core components include outer layers made from polydimethylsiloxane (PDMS), a type of silicone, surrounding copper coil layers that facilitate magnetic induction (MI). This assembly is complemented by an internal magneto-mechanical coupling (MC) layer with strategic incisions to enhance flexibility and movement sensitivity.
Weighing a mere 7.2 grams, this patch adheres to the neck and converts the muscular movements into electrical signals when a person tries to speak. These signals are then fed into a machine-learning system that has been trained with a selection of sentences. Within an impressive 40 milliseconds, the system interprets the silent speech and generates the spoken words with a reported accuracy rate of over 94 percent.
The current iteration of this technology has been trained with a limited range of words and sentences and, as such, will require further development before it can be used to vocalize the full spectrum of human speech. Individuals currently relying on alternative communication methods, such as Morse code or sign language, may find this an exciting area to watch for future developments.
Testing with a small group of participants has shown promising results, demonstrating the system’s ability to accurately discern and vocalize speech that was attempted without any sound. This level of accuracy in testing signifies a milestone in wearable assistive speech technologies.
This highly accurate, non-invasive, and user-friendly device could represent a transformative shift in how people with speaking disabilities communicate, offering new hope for more natural and accessible speech production. As the researchers continue to refine the technology, it’s likely that we’ll see broader applications and improvements, leading to greater inclusivity and quality of life improvements for those experiencing speech challenges.
