UT Austin researchers have devised an implant-less method to decode the brain process into a stream of text.

Researchers at the University of Texas at Austin have developed an artificial intelligence system called a semantic decoder that can transform an individual’s neural signals, either while they are listening to a narrative or mentally visualizing narrating a story, into a coherent and uninterrupted text.

The system, which uses a transformer model similar to those in ChatGPT and Google’s Bard, may help those who are mentally conscious but unable to physically speak, such as people debilitated by strokes, to communicate effectively again. Unlike other language decoding systems, the new system does not require surgical implants, making the process noninvasive. (The University of Texas at Austin)

“For a noninvasive method, this is a real leap forward compared to what’s been done before, which is typically single words or short sentences,” said Alex Huth, an assistant professor of neuroscience and computer science at UT Austin. “We’re getting the model to decode continuous language for extended periods of time with complicated ideas.”

Brain activity is measured using an fMRI scanner after the individual has listened to hours of podcasts in the scanner. Later, when the participant listens to a new story or imagines telling a story, the machine generates corresponding text from the brain activity alone. The system captures the gist of what is being said or thought, rather than producing a word-for-word transcript. Although the decoder can accurately reproduce some words and phrases and capture the essence of many more, its results are not perfect.

While the system is not practical for use outside the laboratory because of its reliance on fMRI machines, the researchers believe the work could be adapted for more portable brain-imaging systems.

The UT Austin researchers addressed potential concerns about the technology’s misuse, noting that it worked only with willing participants and that the results for individuals on whom the decoder had not been trained were unintelligible.

Related Story:  Elon Musk’s Neuralink Raises $205M For Its Brain-Machine Interface

Image by Peace,love,happiness from Pixabay