We are getting closer and closer to becoming mind readers, especially after research carried out by scientists at Carnegie Mellon University (USA) in which they have identified some of the complex brain patterns that are responsible for the formation of prayers in our heads. Experts have managed to predict thoughts with 87% accuracy.
How did they do it?
We enter the field of artificial intelligence. The experts designed an algorithm with the ability to read a human mind thanks to the neural signals that our brain emits. Specifically, they had the AI collect information from a functional magnetic resonance imaging or fMRI machine. Thus, with the appropriate algorithm, he was able to detect activation patterns of the brain and the way in which these are activated when creating complex thoughts.
Thus, just by looking at an MRI, artificial intelligence predicted with 87% accuracy what the 7 participants of the experiment were thinking. The algorithm that works the ‘miracle’ works with an alphabet of 42 different concepts, including size, location, colors, actions and other elements that make up complex thoughts. As an example, one of the phrases that the AI elucidated was “the witness screamed during the trial”, involving both a subject, a setting and an action.
Although it is not yet possible for the AI to unravel what you are thinking about right now – at least not yet -, it is capable of deciphering the type of thinking that the user is having based on the brain scan.
“One of the great advances of the human brain was the ability to combine individual concepts into complex thoughts, that is, not just thinking about ‘bananas’, but something like ‘I like to eat bananas at night with my friends,'” he says. Marcel Just, leader of the work.
Identifying the building blocks of a thought in the brain is something that has not been done before, and the researchers argue that these blocks appear to be the same, regardless of the mother tongue or culture of the person involved. “This breakthrough makes it possible for the first time to decode thoughts that contain various concepts,” says Just.
As if this were not enough, the system could also work in reverse: receive a sentence and create a correct image of how the human brain had been activated to form it.
Reference: Predicting the brain activation pattern associated with the propositional content of a sentence: Modeling neural representations of events and States. June 2017. H uman Brain Mapping. doi / 10.1002 / hbm.23692
