A team of researchers from UC San Francisco, UC Berkeley, and Speech Graphics has developed an innovative communication system that allows a paralyzed stroke patient to communicate by using a digital avatar controlled through a brain-computer interface (BCI). BCIs are devices that monitor brain signals and translate them into digital signals understood by computers. In this study, researchers implanted an electrode array into the patient’s speech center, which captured electrical signals responsible for jaw, lip, and tongue movements. These signals were then routed to a bank of processors, including an AI system that learned to recognize over 1,000 words based on the patient’s unique signals.

Using the AI interface, the patient can now generate written responses and “speak” through a synthesized voice trained on recordings of her pre-paralysis voice. The research team collaborated with Speech Graphics to create a digital avatar for the patient, which can produce realistic facial expressions based on audio input. The patient’s mental signals directly control the avatar, enabling her to convey emotions and communicate non-verbally.

BCIs have evolved since the 1970s, with recent advancements in processing and computing technology driving progress in the field. Startups like Synchron and Neuralink are competing to secure FDA regulatory approval for their BCIs. The research team’s breakthrough in enabling paralyzed individuals to communicate effectively demonstrates the potential for AI-driven avatars in restoring embodiment and control to patients who have lost their ability to speak.

Sources:
– UC Berkeley: https://news.berkeley.edu/2022/07/01/uc-berkeley-ucsf-team-unveils-brain-implant-that-lets-a-paralyzed-stroke-patient-speak
– Speech Graphics: https://www.speech-graphics.com/news/2022/ucsf-uc-berkeley-team-unveils-brain-implant-that-lets-a-paralyzed-stroke-patient-speak

The post Paralyzed Stroke Patient Uses AI-Powered Avatar to Communicate appeared first on TS2 SPACE.