Skip to content
Search

Latest Stories

Add Good to your Google News feed.

'Intelligent' neck device detects words and emotions, giving stroke survivors a literal new voice

The wearable device achieved 96 percent accuracy.

stroke, speech therapy, medical breakthrough, wearable device, accessibility

A new neck wearable can help stroke victims speak better with less fatigue.

Photo credit: Canva/Cambridge University via X

The effects of a stroke can vary widely from person to person, causing lifelong challenges with vision, mobility, and other functions depending on the severity and location of the stroke in the brain. For those who lose the ability to speak clearly, however, there is new hope in the form of an AI-assisted neck device.

Researchers from Beihang University in China and the University of Cambridge in the UK have developed a neck-worn device designed to help stroke survivors with dysarthria. Dysarthria is a speech disorder caused by weakened strength and control of the muscles used for speaking, often resulting in speech that sounds slow, slurred, or quiet. This can be deeply frustrating for stroke survivors, who know exactly what they want to say and how they want to say it, but find their bodies unable to cooperate.


So far, tests of the device have been promising, allowing five stroke patients to speak more clearly and with the tones, words, and inflections they intended.

- YouTube youtube.com

The "Intelligent Throat" (IT) wearable device, dubbed "Revoice," was tested using a 47-word vocabulary of common phrases. Results showed that the IT system correctly decoded 95.8% of individual words and 97.1% of complete sentences users attempted to vocalize. The device was also able to interpret emotional context, allowing speech to sound more natural and less robotic. Participants reported a 55% higher satisfaction rate compared to word-by-word speech output systems.

The Revoice device works by sensing vibrations in the throat muscles along with pulses from the carotid artery. Sensors on the elastic band capture the tiny movements users make as they mouth words without producing sound. That data is then transmitted wirelessly to an AI system, which interprets not only the intended speech but also its emotional intent.

@poststrokeorg

A common misconception about stroke survivors is that if they have a speech problem, then their cognition and intelligence as been impacted, too. On the contrary, this is simply not true. They may understand everything you say. They might just need a little more time to get their words out, or they might need to use a communication board. The same person is still in there even though they might not communicate in the same way they did before their stroke. And this doesn’t mean they’re dumb. Exercise patience and compassion. Don’t speak to them like a child or finish their sentences for them. You’ll often be wrong and just make them frustrated! #stroke #strokerecovery #strokerehab #strokerehabilitation #strokerecovery #strokeawareness #recovery #speech #communication #communicate #aphasia #braininjury #neuro #brain #poststroke #occupationaltherapy #occupationaltherapist #nonprofit #nonprofitorganization

The AI can be trained to translate speech word for word or customized to expand short phrases into full sentences based on factors such as the user's word choice, pulse patterns, and even time of day. For example, a user might mouth "Go hospital now," which the system could transform into a fuller statement like, "I still don't feel well or comfortable. Can we go to the hospital now?" This allows users to communicate more clearly with less strain on their vocal muscles, using a device that's comfortable enough to wear all day and powered by a long-lasting battery.

@cmalugen

#strokesurvivor #strokeawareness #strokerecovery #massivestroke #stroke #iamasurvivor

The device improves on earlier technologies that forced users to pause between words, resulting in speech that was clear but slow and robotic. By contrast, the IT device enables more fluid conversation, allowing users to be heard not only through their words but also through emotional nuance, making communication more enjoyable overall. Researchers hope to expand this technology by increasing the device's vocabulary and testing it with a more diverse group of dysarthria patients.

"We are actively expanding our study cohort to include a broader range of dysarthria patients with varying neuromuscular conditions, ensuring that the system is robust across different symptom severities," the researchers wrote in their study. "We also plan to recruit participants from diverse linguistic and ethnic backgrounds to capture a wider range of speech patterns and cultural communication norms, enabling multilingual evaluation of the decoding and synthesis pipelines."

Over time and with investment, this technology could provide a clearer, faster, and more conversational voice for the voiceless.