A computer scanner has learned to recognize visual images that a person thinks about

"I imagine it, but I don't know how to describe it in words." Common situation? Scientists at the University of Toronto believe they have found a way to solve this problem, and along with many others. They have developed a system that very accurately recognizes brain activity associated with mental processing of specific objects.

It is based on a classic brain activity scanner based on electroencephalography, because the speed of its response to changes in the brain is measured in milliseconds. This is exactly the level at which the formation of a mental imprint takes place, the recording of an image of a real object in memory or its call from there. That is, it allows you to isolate a specific trace among the stream of thoughts and assign a conditional code to it.

In the experiment, volunteers were asked to memorize several different human faces, and then mentally imagine them when the assistant calls the names of these people. The system read mental prints at the stage of memorizing faces, and then analyzed the work of the volunteer's brain, looked for coincidences and signaled about it. Technically, it is not a problem to expand the range of mental images to any number of different objects, this is a routine task.

If a disabled person, unable to speak or press buttons, learns the images of words, phrases, terms and receives a personal scanner-recognizer, the machine will literally be able to voice his thoughts. Instead of compiling a composite sketch from the words of an eyewitness, the computer will divide the bandit's face in the respondent's memory into thousands of known fragments, recognize them and transfer them to a graphic editor, immediately drawing a portrait. And this is just the first thing that comes to mind - real mind reading technologies are getting closer!