Like it or not, we often have positive or negative feelings about a total stranger based solely on the looks of his or her face. A person’s features, expressions, gender, and skin color can generate emotions that we can’t control in that moment. Because, let’s face it: We’re biased about faces, and that can affect everything from getting a job to getting a date.
Last week, researchers described, in the journal PLOS Biology, a brain training system that can alter emotions evoked by the sight of someone’s face. With just a few days of training, study volunteers felt more positively or negatively about a photo of a stranger. The change in feeling was slight but statistically significant, and could become more pronounced with further training, says Mitsuo Kawato, a computational neuroscientist at the Advanced Telecommunications Research Institute International in Kyoto, Japan. Kawato is one of the paper’s coauthors.
The study brought Kawato and his colleagues one step closer to their end goal: helping people with post-traumatic stress disorder (PTSD), anxiety, depression, phobia, and other mental disorders by manipulating their feelings about past events. A soldier who saw awful things during combat, or a person who was abused as a child could theoretically retrain their brain to feel more neutral about those experiences.
“If we can manipulate brain memory unconsciously by this technique, it could be very beneficial for fear extinction,” says Kawato.
The experiment involved functional magnetic resonance imaging (fMRI), a software decoding program, and tricking 24 volunteers. Kawato’s team showed the participants 400 photos of faces, and asked them to rate their sentiments about each one on a scale of 1 to 10—with 1 representing strong dislike, 5 being neutral, and 10 for strong positive feelings.
Meanwhile, an fMRI machine recorded the participants’ neuronal activity in a region of the brain called the cingulate cortex. (fMRI measures brain activity by detecting changes in blood flow, which correlates with neuronal activity.) The software, based on a machine learning algorithm, then decoded the brain activity of each participant, noting the patterns associated with positive and negative emotions.
The researchers later called the participants back to the lab, put them in the fMRI scanner and showed them a subset of faces that they had rated as neutral. After each face, researchers showed the volunteers an image of a green disc. And here’s where the trick—and the training—happens. The researchers told the participants that if they could somehow make the image of the green disc bigger with their minds, they would receive a monetary reward. Volunteers were given no further instructions and did not know the purpose of the study.
By this stage of the experiment, the volunteers had been divided into three groups: some in a “positive” group, some in a “negative” group, and others in a control group.
The only way the green disc would get larger is if the volunteer managed to generate the same brain activity patterns he or she used when feeling positive or negative emotions about a face. Volunteers in the positive group had to generate the patterns they generated for faces they rated as positive, and vice versa for the negative group.
The neurofeedback system worked. When the volunteers later rated the neutral faces again, the positive group rated the formerly neutral faces as slightly more positive, and the negative group rated them as slightly more negative. The control group showed no change in feelings.
The fact that the brain activity monitored over the course of the experiment was going on in the cingulate cortex helped land the group’s work a spot in PLOS Biology. Previous research suggested that brain activity associated with positive and negative reactions to faces comes from multiple brain regions. Kawato’s group says it is the first to show that focusing on just the cingulate cortex is sufficient for manipulating both positive and negative facial preferences.
That region of the brain is already known to play an important role in affective disorders such as depression, PTSD, and phobias, says Kawato. Being able to manipulate that part of the brain in order to affect facial preference (which is an emotional condition), bodes well for the group’s chances for targeting debilitating mental disorders.
Earlier this year, the team used the same fMRI and decoding software technique, which they call DecNef, to brain train another group of volunteers. That group of subjects was trained to associate a color with a pattern. The experiment worked so well that, eventually, the volunteers would see a color cued by a pattern, even though the color was not there. Kawato’s team first demonstrated the technology in 2011 in Science.
Kawato is aware that his technique sounds a lot like brainwashing. Indeed, one of the reasons his group published the paper was to notify the public and give people a chance to discuss the ethical implications of the technology, he says. In this particular experiment, because subjects had voluntarily lied in an MRI scanner for 1.5 hours at a time, “It’s difficult to imagine this technique being used for brainwashing against a subject’s will,” says Kawato.
As scientists get better at brain training, the public will have to decide what kind of emotion and memory manipulation is acceptable. Altering ourselves to like people we shouldn’t like can be dangerous. But of course, in a way, we already have that technology. It’s called alcohol. Just ask anyone wearing one of these.