Brain-Computer Interface Technology Makes Breakthrough: Aussie Scientists Develop AI System That Decodes Thoughts

Scientists at the University of Technology in Sydney have made a groundbreaking discovery in brain-computer interface technology, developing an artificial intelligence system that can decode brainwaves into text, effectively reading thoughts.

The team, led by PhD students Charles Zhou and his supervisors Chin-Teng Lin and Dr Leong, has created an AI model that uses deep learning to translate brain signals from electroencephalography (EEG) into specific words. In a remarkable demonstration of the technology's capabilities, the AI model successfully decoded the phrase "I am jumping happily, it's just me" when Dr Leong wore a 128-electrode EEG cap and didn't utter a single word.

Currently, the AI model has been trained on a limited set of words and sentences to simplify the recognition process. However, researchers argue that this is a crucial step to filter out noise and clarify brain signals, which are often overlapped by different brain sources on the skull's surface.

While Elon Musk's Neuralink is also working on developing similar technology, the Sydney team's innovation lies in its non-invasive nature, avoiding the need for implantable devices. According to Mr Lin, "We can't get very precise because with non-invasive, you can't actually put it into that part of the brain that decodes words."

The implications of this breakthrough are immense, with potential applications in stroke rehabilitation, speech therapy for autism patients, and restoring communication for paralysis patients.

This is not an isolated discovery, as scientists around the globe have been exploring the combination of EEG and AI to achieve impressive results. For instance, researchers at Mass General Brigham recently developed an AI tool capable of predicting brain decline in patients years in advance, using subtle changes in brain activity during sleep recorded through EEG.

The AI tool accurately flagged 85 per cent of individuals who eventually experienced cognitive decline, with an overall accuracy of 77 per cent. As this technology continues to evolve, it offers a promising avenue for improving our understanding and treatment of various neurological conditions.