AI model predicts brain activity during human conversations

livescience.com

Researchers have developed an AI model called Whisper that predicts human brain activity during conversations. Unlike traditional models, Whisper learns from audio and text without pre-coded language features, revealing how language structures emerge naturally. The study involved four epilepsy patients, capturing over 100 hours of real-life conversations. This approach allowed scientists to observe brain activity during speech production and comprehension, supporting the idea that different brain regions work together rather than separately. Whisper's predictions of brain activity surpassed those of traditional models. The findings suggest a connection between AI language processing and human cognition, although further research is needed to explore these similarities.


With a significance score of 5.5, this news ranks in the top 1% of today's 15705 analyzed articles.

Get summaries of news with significance over 5.5 (usually ~10 stories per week). Read by 9000 minimalists.


loading...