This effort is part of a growing movement to create direct connections between the human brain and computers. The Texas-based company behind the implant is developing technology aimed at restoring communication abilities in individuals affected by spinal cord injuries, stroke, or neurodegenerative diseases such as ALS. The device translates neural signals into synthesized speech, typed text, or cursor movements.
After several years of testing on animal models, particularly sheep, this marked the first use of the implant in a human brain. The procedure was conducted during a scheduled brain surgery for epilepsy treatment, where the patient consented to a temporary insertion of the device into a brain region involved in auditory processing and memory. The company used a specialized tool resembling an automatic injector to implant the device. Researchers confirmed that the device successfully recorded brain signals.
BCIs do not read a person’s thoughts directly but instead interpret neural activity related to motor intentions. For example, someone with paralysis who tries to speak—even without physical movement—generates brain signals that can be decoded into speech.
Recent research efforts have demonstrated significant progress in decoding speech from brain activity. Some teams have achieved word rates of over 60 words per minute in individuals with paralysis—an encouraging result given that normal speech averages about 130 words per minute. The startup aims to match or exceed such benchmarks in future trials, which will involve long-term implantation in volunteers with paralysis.
The implant is smaller than a dime and contains hundreds of tiny electrodes capable of recording signals from individual neurons. Unlike older systems such as the Utah array, which required a visible pedestal on the skull and posed risks of long-term tissue damage, the newer design is more compact, less invasive, and capable of capturing more detailed data.
The ultimate goal is to improve signal quality by positioning the electrodes as close as possible to individual neurons, allowing for high-resolution decoding of intended speech. Competing approaches include implants placed on the brain’s surface or even inside blood vessels, which capture broader patterns of neural activity rather than signals from single neurons.
This initial test served as a dry run to validate the implant procedure, confirm that the device remained functional during use, and ensure it could be safely removed. Looking ahead, the company plans to assess the safety of long-term implantation and explore the possibility of using multiple implants to enhance performance.
