Stanford researchers demonstrate a neural implant enabling paralyzed individuals to communicate at 90 words per minute with 95% accuracy.
Researchers at Stanford University have achieved a landmark breakthrough in brain-computer interface technology, demonstrating a neural implant that allows completely paralyzed individuals to communicate at 90 words per minute with 95% accuracy. The system decodes intended speech directly from neural activity in the motor cortex, translating thoughts into text and synthesized voice in real time.
The study, published in Nature, involved five participants with amyotrophic lateral sclerosis (ALS) and brainstem stroke who had lost the ability to speak. Each received a small implant containing 256 electrodes placed on the surface of the brain's speech motor cortex, the region that normally controls the muscles of the mouth, tongue, and larynx during speech.
When participants attempt to speak, the implant records the neural patterns associated with each intended phoneme. A deep learning algorithm running on a pocket-sized external processor decodes these patterns in real time and generates both text and natural-sounding synthesized speech. The system was trained using just 30 minutes of attempted speech data from each participant.
Dr. Jaimie Henderson, the neurosurgeon who performed the implantations, described the emotional moment when the first participant spoke to his family for the first time in three years. The technology represents more than a communication tool; it restores a fundamental aspect of human identity and social connection.
The 90-words-per-minute rate represents a tenfold improvement over previous BCI communication systems and approaches the average natural speaking rate of 150 words per minute. The 95% accuracy makes the system practical for everyday conversation without constant error correction.
The research team is now working on a fully wireless version of the implant that would eliminate the need for a percutaneous connector passing through the skull. This wireless design, expected to enter clinical trials within two years, would significantly reduce infection risk and improve the cosmetic acceptability of the device.
The technology has implications beyond communication. The same neural decoding approach could potentially be used to control prosthetic limbs, operate computers, and interact with smart home devices, giving individuals with severe paralysis unprecedented independence.
Several companies, including Neuralink, Synchron, and Blackrock Neurotech, are developing competing BCI technologies. The competitive landscape is accelerating innovation and driving down costs, with the goal of making neural interfaces accessible to all patients who could benefit.