AI enables paralysed woman to regain speech and facial expressions

AI enables paralysed woman to regain speech and facial expressions

AI enables paralysed woman to regain speech and facial expressions

Scientists have made a remarkable breakthrough in the field of artificial intelligence. Through the power of AI, they have successfully restored the ability to communicate for a woman who had been paralysed.

This achievement, published in the journal Nature on August 23, 2023, marks the first time that both speech and facial expressions have been synthesised from brain signals, opening new doors for communication for individuals with severe paralysis.

At the heart of this innovation lies a revolutionary technology known as a brain-computer interface (BCI). Dr. Edward Chang, chair of neurological surgery at UCSF, and his team have dedicated over a decade to perfecting this technology. They achieved a major breakthrough by implanting a thin rectangle containing 253 electrodes onto the surface of a woman's brain. These electrodes intercepted brain signals that would normally control muscles responsible for speech and facial expressions.

Connected to a bank of computers via a cable, this BCI decoded these brain signals into the richness of speech and the nuanced facial movements that characterize conversation. The participant worked tirelessly with the research team, training artificial intelligence algorithms to recognize her unique brain signals for speech. This process involved repeating phrases from a 1,024-word conversational vocabulary until the computer could identify the brain activity patterns associated with specific sounds.

Deciphering phonemes

Unlike traditional approaches that teach AI to recognize whole words, the researchers adopted a strategy to decode words from phonemes—the building blocks of speech. For example, the word "hello" consists of four phonemes: "HH," "AH," "L," and "OW." This approach significantly boosted accuracy and speed, making communication nearly as fast as natural conversation.

Personalised voice and facial expressions

To recreate her voice before the injury, the team developed an algorithm for synthesizing speech, using a recording of the woman speaking at her wedding. The digital avatar, representing the woman, was brought to life with the help of software from Speech Graphics, a company specializing in AI-driven facial animation. Customized machine-learning processes enabled the software to mimic the movements of the woman's face, including jaw, lip, and tongue movements, as well as expressions like happiness, sadness, and surprise.

This remarkable achievement not only restores communication but also promises profound effects on independence and social interactions. Future goals for the team include creating a wireless version of the BCI, eliminating the need for physical connections and offering more freedom to users.

For hose who have lost their ability to speak due to conditions like brainstem strokes, this breakthrough brings hope and the prospect of reconnecting with the world through natural conversation. Dr. Chang's vision of "a full, embodied way of communicating" is coming closer to reality, and the impact on countless lives could be immeasurable.