Scientists at the University of Texas have developed an AI-powered algorithm that allows people to decode what they're hearing and thinking through brain scans using fMRI.
The first non-invasive method can facilitate communication for the first time for people who cannot speak.
The technology used three brain regions associated with natural language to extract the data. The model recreates the sounds or thoughts a person hears or has using natural language.
As a result, the system can generate plain text of the person's thoughts.
For example, while participants were listening to narrative stories, scientists used fMRI to scan parts of the brain associated with natural language.
They then fed the scanned data into an AI-powered decoder to provide a summary of what the person was hearing.
Thoughts can be as simple as one word, such as dog, or as complex as "I have to walk the dog."
Until now, the only way to accomplish this technique was to implant electrodes into the brain.
Additionally, the brain has its own alphabet of 42 distinct elements. Each refers to a different concept, such as size, color, or location, and works together to create complex ideas.
0 Comments