The guessing game 20 Questions might be won with a quick scan of an opponent’s brain. Researchers have developed a way to measure how the human brain reacts to pictures—and to predict what images people are gazing at.
Though a long way off, the technology could help scientists reconstruct dreams or repressed memories, says Jack Gallant, a neuroscientist at the University of California, Berkeley, who led the study, published in the March 20 Nature.
As a first step, two members of Gallant’s team viewed hundreds of black-and-white photos from everyday experience, such as people, animals, and fruits, while a functional magnetic resonance imaging (fMRI) scanner recorded the activity of their brains. By analyzing the activity patterns that the images sparked, the researchers made a model that could predict how the brain might respond to any image.
Next, the same two subjects viewed 120 pictures they had never seen before, while their brains were scanned. Based on the model, Gallant’s team predicted how each person’s brain would respond to the 120 new images, then matched the brain scans to these predictions. The researchers guessed right 92 percent of the time for one person and 72 percent for the other. When they upped the ante to 1,620 pictures, accuracy dropped considerably. The team read one person’s mind half the time and the other subject’s mind a third of the time.