New Brain Implant Decodes ‘Inner Monologue’ of People With Paralysis
1 min read
Summary
Scientists from Stanford University have developed a system using AI to translate a person’s “inner monologues” into speech in real time, with up to 74% accuracy, according to a Cell study.
The aim of the technology, which includes a “neural password” that users can activate to prevent their thoughts being broadcast, is to help people with paralysis communicate with their loved ones and find it easier to communicate using current brain-to-speech systems.
Four people with multiple 64-channel microelectrode arrays implanted in their brains took part in the study, during which the participants were asked to think words, or attempt to say them aloud.
The team then trained an AI decoder to decipher the thoughts, with the system labelling attempted speech differently to inner speech as a safety mechanism.
