Brain decoder
A study published today in journal Nature introduced a non-invasive semantic decoder based on functional magnetic resonance imaging (fMRI) and technology similar to the one powering ChatGPT. The decoder can convert a person's brain activity when they are listening to or silently imagine telling a story into a continuous stream of text. The method created by the team from the University of Texas at Austin may enable people to speak with their minds. Given novel brain recordings, this decoder generates intelligible word sequences that recover the meaning of perceived speech, imagined speech and even silent videos, demonstrating that a single decoder can be applied to a range of tasks. And it can fruitfully work with multiple brain regions. This demonstrates the viability of non-invasive language brain–computer interfaces. Computational task was performed by Generative Pre-trained Transformer (GPT), a 12-layer neural network using multi-head self-attention to combine represent...