I don’t think he needed a machine to tell him what I was feeling right then. While the research is far from creating a pocket-sized device for spouses, researchers in CMU’s Dietrich College of Humanities and Social Sciences successfully developed a computer model that can read and identify a person’s emotions more accurately than when those feelings are self-reported. It uses a combination of functional magnetic resonance imaging (fMRI) and machine learning to measure brain signals triggered by emotional stimuli. Karim Kassam, PhD, assistant professor of social and decision sciences and lead author of the study, says there are many possible applications for this type of research, particularly in marketing where product input often relies upon self-reported emotions. However, he notes that the technology offers diagnostic possibilities for mood and emotional disorders as well. He says he’d like to see someone who works with clinical populations take this idea further, and see how the research applies in the mental health field. “We consider the marketing implications first because our work establishes that as a more immediate possibility,” says Dr. Kassam. “However, this research has a lot of implications for psychopsychology. One intriguing facet of the research were our findings of how emotion is organized in the brain. That understanding of what emotions are can lead to a better understanding of psychopsychology and may lead to different, more accurate methods of diagnosing disorders that can affect your mood and emotions.” The computer model’s rank accuracy levels were well above the 50% accuracy that would result from random guessing, but Dr. Kassam said there are no statistics available as to the accuracy of self-reported emotions. However, he says individuals tend to be less accurate in reporting their emotions when they are under stress, or when the emotions may not be socially acceptable, such as in cases of prejudice, or with certain emotions that may have a negative connotation, such as lust. “There are times and situations where self-reported emotions are very accurate; other times its more limited,” says Dr. Kassam. “Sometimes people are not completely in touch with how they’re feeling and they may not even be aware of it. This can offer another data set that can give the researchers a more accurate, scientific measure of what an individual is really feeling.” More from Prevention: Could This Be The Depression Wonder Drug?
