Two groundbreaking studies demonstrate significant progress in decoding internal speech and restoring communication for individuals with paralysis, offering hope for those with conditions like ALS and spinal cord injuries.
Brain-computer interfaces (BCIs) have long been heralded as a potential solution for restoring communication in individuals who have lost the ability to speak or move due to paralysis or neurological conditions. Recent advancements in this field have brought us closer to turning this possibility into reality, with two separate studies showcasing remarkable progress in decoding internal speech and translating brain signals into comprehensible words.
Unprecedented accuracy in speech decoding
Researchers at UC Davis Health have developed a BCI system capable of translating brain signals into speech with up to 97% accuracy, marking it as the most precise system of its kind to date. The study, published in the New England Journal of Medicine [1], involved implanting sensors in the brain of a 45-year-old man with severely impaired speech due to amyotrophic lateral sclerosis (ALS).
The participant, Casey Harrell, had four microelectrode arrays implanted into his left precentral gyrus, a brain region crucial for coordinating speech. These arrays, designed to record brain activity from 256 cortical electrodes, allowed the BCI to interpret Harrell’s attempted speech and convert it into text that could be ‘spoken’ aloud by a computer.
Dr David Brandman, co-principal investigator and assistant professor in the UC Davis Department of Neurological Surgery, explained: “Our BCI technology helped a man with paralysis to communicate with friends, families and caregivers. Our paper demonstrates the most accurate speech neuroprosthesis (device) ever reported.”
Rapid training and expanded vocabulary
One of the most impressive aspects of this new BCI system is its ability to achieve high accuracy with minimal training time.
In the first speech data training session, the system achieved 99.6% word accuracy with a 50-word vocabulary in just 30 minutes. By the second session, the potential vocabulary expanded to 125,000 words, maintaining 90.2% accuracy with only an additional 1.4 hours of training.
Dr Sergey Stavisky, co-director of the UC Davis Neuroprosthetics Lab, noted the emotional impact of the system’s success: “The first time we tried the system, he cried with joy as the words he was trying to say correctly appeared on-screen. We all did.”
Decoding internal speech
While the UC Davis study focused on interpreting attempted speech, researchers at the
California Institute of Technology and the Feinstein Institutes for Medical Research have made strides in decoding internal speech – words that are thought but not spoken aloud. Their study, published in Nature Human Behaviour [2], demonstrated for the first time that a computerised brain implant can decode internal speech with minimal training.
The research team, led by Richard Andersen and Sarah Wandelt, implanted a device in the supramarginal gyrus of two individuals with tetraplegia. This brain area is known to be important for representing spoken words. The participants were asked to think about ‘saying’ specific words without speaking or moving, while the BCI measured their brain activity to predict the words being ‘spoken’ internally.
Challenges and variations
The study revealed varying levels of success between the two participants, with one achieving an average accuracy of 79% and the other 23%. Researchers attributed this discrepancy to differences in the unique patterns of brain activity associated with different words in each participant.
Despite these challenges, the findings offer proof-of-concept for a high-performance internal speech BCI. The device’s ability to decode nonsense words suggests that words are represented phonetically in this part of the brain, rather than based on their meanings.
Future implications
Both studies represent significant advancements in the field of BCIs and offer hope for individuals who have lost the ability to communicate due to paralysis or neurological conditions. As Dr Brandman noted: “This technology is transformative because it provides hope for people who want to speak but can’t. I hope that technology like this speech BCI will help future patients speak with their family and friends.”
While there is still much to learn about decoding internal speech reliably, these findings pave the way for future developments in restoring communication for those affected by conditions such as ALS, brain injuries, and spinal cord injuries.
As research in this field continues to progress, supported by initiatives like the NIH Brain Research Through Advancing Innovative Neurotechnologies (BRAIN) Initiative, we may be moving closer to a future where the ability to communicate using only one’s thoughts becomes a reality for those who need it most.
Watch video:
New brain-computer interface allows man with ALS to ‘speak’ again https://bit.ly/3yR4lA3
References:
- Card, N. S., Wairagkar, M., Iacobacci, C., et. al. (2024). An Accurate and Rapidly Calibrating Speech Neuroprosthesis.
New England Journal of Medicine, 391, 609-618. https://doi.org/10.1056/NEJMoa2314132 - Wandelt, S. K., et al. (2024). Representation of internal speech by single neurons in human supramarginal gyrus.
Nature Human Behaviour. https://doi.org/10.1038/s41562-024-01867-y
