A UC Davis study shows ALS patient Casey Harrell used a speech neurointerface to communicate nearly 2 million words over 19 months, enhancing his quality of life.
A team of researchers from the University of California, Davis, has successfully implemented a speech neurointerface that has allowed an amyotrophic lateral sclerosis (ALS) patient, Casey Harrell, to communicate extensively from home. Over a period of 19 months, Harrell managed to send nearly 1.96 million words, comprising 183,060 sentences, with an average communication speed of 56 words per minute.
Published in Nature Medicine, the study highlights the practical application of the neurointerface in everyday life, emphasizing its independent use by the patient rather than just laboratory settings.
Communication Achievements
Throughout more than 3,800 hours of home use, Harrell utilized the system to interact with family, friends, and medical professionals, enabling him to send messages, engage in video calls, and browse the internet despite his paralysis. He reported that the system correctly interpreted 92% of his sentences as mostly accurate, while formal assessments indicated accuracy rates exceeding 99% for a vocabulary of 125,000 words.
Since implantation, Harrell has used the system for 444 days out of 653. The average time spent interacting with the device surged from 3.7 to 9.5 hours daily after his assistants were given the autonomy to manage the equipment without researcher oversight.
Technical Overview
The neurointerface consists of four microelectrode arrays implanted in Harrell's left precentral gyrus, which is crucial for speech coordination. It reads signals from 256 cortical electrodes when the patient tries to speak. An algorithm processes these neural signals every 80 milliseconds, converting them into phoneme probabilities, which a language model uses to generate text in real-time. Once a phrase is formed, the system can vocalize it using a synthesized version of Harrell's voice prior to his illness.
Additionally, Harrell controlled a computer cursor through this system, showcasing that both speech and movement can be managed through signals from the speech motor cortex, challenging previous assumptions about the brain's functional regions.
Shifting to Practical Applications
Earlier research had demonstrated high accuracy of the neurointerface in controlled environments, achieving 99.6% accuracy during initial training sessions. This latest study moves the focus towards its long-term use in home settings, marking a significant milestone for neurointerfaces aimed at helping those with severe motor impairments. Co-author Sergey Stavisky noted that the recorded brain activity represents the largest individual dataset with single-neuron resolution to date.
While the study provides valuable insights, it is important to note that it describes a single clinical case. The applicability of these results to other patients remains uncertain, as factors such as electrode types and neurological conditions may vary. The current system remains experimental, requiring daily setup and wired connections, making it suitable for home use only due to its size.
In related advancements, Neuralink CEO Elon Musk recently shared that their brain-computer interface could restore speech to those who have lost it. Reports indicated a patient used their device to play World of Warcraft using only their thoughts within 80 days of implantation. Similarly, Galena Bakwalter, another neurointerface user, learned to compose music through her thoughts. Furthermore, a startup founded by former Neuralink leadership is expected to commence human trials of a biohybrid chip soon.
