Researchers Build A Computer That Can 'Read People's Minds'

Researchers are one step closer to reading people's minds in real time. In a study published Thursday in the journal Plos Computational Biology, the team used computational software that managed to decode brain signals and accurately predict what they were seeing in real time.


Seven epilepsy patients had electrodes implanted in multiple locations in their temporal lobes. After this, they were shown 200 images of various things including human faces and houses in 400 millisecond flashes, being told to watch out for a picture of an upside-down house. When this was spotted, the software digitizing brain signals 1,000 times per second spotted it too, correlating brain activity.

This algorithm then gets to work learning and anticipating the changes in brain signals and managed to successfully predict their responses to another 100 images. Future uses of this could be helping to identify any familiar faces in the subconscious when a line-up just doesn't help.