A major leap forward in assistive technology: researchers have developed a brain–computer interface (BCI) that decodes not only what someone wants to say, but how they say it. The result? A paralysed participant can literally speak and sing with emotion again.
How it works:
A grid of 256 silicon electrodes was implanted in the speech region of a 45‑year‑old man with ALS.
AI deciphers neural signals every 10 milliseconds, translating them into synthetic voice almost instantly—just 10 ms after thinking the words.
Crucially, the system captures natural intonation, pitch, and emphasis, bringing back expressive communication.
“This is real, spontaneous, continuous speech,” says Christian Herff, a computational neuroscientist.
Why it matters:
Previous BCIs had noticeable lag or only produced robotic speech after entire phrases were thought. This new device speaks in real time, preserving the natural cadence of human language.
The synthetic voice is personalized, modeled on the participant’s recorded voice before his illness, making the result more authentic.
What’s next:
This milestone sets the stage for more advanced BCIs that could restore rich communication for people with severe speech impairments.
Key challenges remain, including long-term safety, expanding to more users, and refining emotional nuance across languages.
This breakthrough marks a massive step toward giving voice back to those who've lost it—not just any voice, but their own, fully expressive and in real time.