(Credit: Yale University)
Researchers at Yale have developed a method to accurately reconstruct faces locked in the memories of other people.
Based on the idea that the human brain repeats activity when repeating thoughts, scientists have already found a method to reconstruct a person's dreams — and even stream in real-time what a person sees when they watch film trailers.
Both of these experiments, however, could only reconstruct generalities about what a person was visualising; for example, the brain activity might show that a person was thinking about a bird, but the technology has not been sophisticated enough to show that the bird is a parrot.
However, researchers at Yale have just breached this barrier, at least as pertains to human faces. Using — as previous experiments have — functional magnetic resonance imaging (fMRI), a team of cognitive psychologists have managed to accurately reconstruct human faces as viewed by other people.
It was no easy task. The brain uses large areas to process information about human faces, and each brain operates differently. In order to figure out what the brain activity of a particular individual means, fairly extensive research is required on that individual. Each of the six subjects in the study was given 300 faces to look at while their brains were being scanned using fMRI.
This information was then compiled cross-referenced to create a statistical library of what certain brain activity probably means. For example, activity in a certain region of the brain might indicate a moustache, a skin tone or a bulbous nose. The subjects were then shown a second set of faces while being scanned. Based purely on the fMRI data from the second set of faces, the researchers used the statistical library to reconstruct the faces the test subjects were seeing.
There are several potential uses for this technology. The Yale team believes it could be used to determine how autistic children react to faces, while on a broader level, seeing into the human mind may be able to help reveal what happens in the minds of coma patients, or those with neurodegenerative conditions such as Alzheimer's.
But the technology is a long way from being at that point; and, before anyone starts getting concerned about governmental brain surveillance, it's not quite so simple: each individual brain works to individual patterns, so a subject would have to consent to pretty comprehensive testing before their thoughts were readable.