AI converts brain signals to text
A recent study has been published about a non-invasive technique that uses artificial intelligence to convert brain activity into written text, without the need to implant sensors inside the brain. The technology relies on EEG and MEG recordings and utilizes a deep learning algorithm called "Brain2Qwerty," which predicts letters based on brain activity. The system achieved an error rate of 29% with MEG and 65% with EEG, with notable improvements for early participants, reducing the error rate to 18%. The results demonstrate potential for developing communication tools for individuals with motor disabilities and designing AI systems capable of translating non-invasive brain activity into commands or text, while enhancing safety and making communication more accessible for those who are unable to communicate.
Notice: This Is an AI-Generated Summary
Comments (0)