Device Revoice
The Revoice device uses two AI agents to detect the user’s silent speech and emotional state, emitting speech through a synthetic voice module.
Speech disorder, known as dysarthriais a very common consequence of strokes (AVC)affecting nearly half of all survivors.
A stroke can cause weakness in the facial muscles and vocal cords, making it difficult to speak fluently, clearly, or in complete sentences. Although most people eventually recover, the process is often slow and frustrating, significantly affecting quality of life during rehabilitation.
Scientists at the University of Cambridge have been working to change this reality. In a study this week, they announced the development of a wearable device called Revoice – designed to help people with dysarthria communicate more naturally.
“Dysarthria after a stroke can be extremely frustrating because people they know exactly what they want to saybut they have physical difficulties in saying it, because the signals between the brain and the throat have been messed up by the stroke”, explained the corresponding author of the study to New Atlas, Luigi Occhipinti.
Reading and reconstructing these signals is precisely what Revoice – a non-invasive, wearable alternative – is intended to do.
The device looks like a soft, adjustable collar, with textile strain sensors and a wireless printed circuit board.
Device Revoice
As New Atlas details, its integrated AI system decodes speech signals using two AI agents:
- one reconstructs words from silently articulated speech and predicts sentences by reading vibrations in the throat muscles;
- another detects the patient’s emotional state by measuring the pulse in the neck.
This allows the device to not only reconstruct complete sentences, but also make them emotionally expressive and logical.
Totally innovative system
Previous silent speech systems were mostly tested on healthy participants and lacked real-world application. They required users to pause one to three seconds between words, creating strange and unnatural delays in speech.
Revoice overcomes these limitations by using a AI-driven throat sensor system and a large, lightweight linguistic model to instantly convert muted articulated words into complete sentences.
After initial tests with healthy participants, the device underwent some optimization and was subsequently tested on five stroke patients, showing Incredibly low error rates: 4.2% for words and 2.9% for sentences.
Participants reported a 55% increase in user satisfaction and confirmed that the device allowed them to communicate with the same fluency they had before their stroke.
Before the product can reach the market, it will have to undergo more extensive clinical trials.
