Scientists at the University of California, San Francisco have developed a groundbreaking Bilingual Brain Implant. This innovative program utilizes artificial intelligence to assist stroke patients in communicating in both Spanish and English for the first time.
A team of nearly a dozen scientists at the university’s Center for Neuroengineering and Prosthetics has spent several years creating a decoding system that can translate the man’s brain activity into text, both spoken and visual. This technology was recently demonstrated on screen.
An article published in Nature Biomedical Engineering on May 20th reveals that the man, named Pancho, suffered a stroke in the early 2000s at the age of 20, which left him severely paralyzed. While Pancho can make sounds, he struggles to speak clearly. As a native Spanish speaker who learned English later in life, this new technology has been life-changing for him.
Led by Dr. Edward Chang, a neurosurgeon and co-director of the Center for Neuroengineering and Prosthetics, the team was able to begin tracking Pancho’s brain activity with neural implants in February 2019.
Using an AI technique known as neural networks, researchers trained Pancho’s implant to interpret words based on his brain activity when attempting to speak them. This AI training allows the brain implant, also known as a Brain-Computer Interface Device, to process information similarly to the human brain.
By 2021, the technology had significantly improved Pancho’s ability to communicate, but only in English, as reported by NEJM.
The team realized that speech decoding was mostly for monolingual speakers, and they aimed to allow bilingual individuals like Pancho to communicate in both languages, as each language shapes a person’s identity. This led to the development of a bilingual decoding system for Pancho’s brain implant.
This advancement enables language switching based on preference
After identifying bilingual brain activity in Pancho’s brain, researchers discovered they could train bilingual brain implants without the need for separate language-specific decoding systems. This transfer learning between languages proved to be successful, as data from one language significantly facilitated training in the second language.
In 2022, scientists utilized artificial neural networks to further train Pancho’s brain implant on unique bilingual neural activity generated during speech. The results indicated that Pancho could participate in conversations and select his preferred language with ease.
The study, outlined in the May 20 article, introduces the concept of a “bilingual speech neuroprosthesis” or bilingual brain implant, illustrating how this technology has the potential to restore natural communication among bilingual individuals with paralysis.
For more information from NBC Latino, subscribe to our weekly newsletter.
Source: www.nbcnews.com
