��������

How the brain feels about the world around it is the subject of a new paper published in PNAS (Proceedings of the National Academy of Sciences), co-authored by Eugene Surowitz Assistant Professor of Computational Cognitive Neuroscience Edward A. Vessel of the Colin Powell School for Civic and Global Leadership.

The study, “,” sought insight into the link between perceptual processes and affective (emotional) responses.

To predict human affective responses to a diverse set of natural images, the researchers employed machine learning models trained to categorize the content of images – “visual machines” that only see and cannot feel, as Vessel explained it. “These systems are tuned to understand the world by mimicking human perception,” he said. “They are good at perception because they are good models for how our brains take in information.”

“When you look at something, your brain compares it to the vast, structured knowledge about the world that each of us accumulate over a lifetime of seeing: Is it familiar? Is it unique? Do I understand it?” the authors wrote. “Since we can’t directly measure a person’s internal visual knowledge, we used machine learning models as stand-ins.”

Their results showed that this perceptual process is enough to predict much of our affective responses, such as whether one finds an object beautiful, positive or negative, or arousing.

“A lot of affect may be less ‘emotional’ than we think it is,” said Vessel, suggesting that perceptual processes, which are built on rich sensory experiences, may shape what we feel about the world around us far more than many psychological theories suggest.

Their findings have direct implications for artificial intelligence and machine learning, and could lead to new ideas for how such systems could learn.

Vessel’s research group, the , uses behavior, brain imaging and computation to study the psychological and neural basis of aesthetic experiences, creative insight and curiosity. Before coming to CCNY, Vessel was a senior research scientist at the Max Planck Institute for Empirical Aesthetics in Frankfurt, Germany, and was formerly co-director of the New York University Artlab. He is general secretary of the International Association of Empirical Aesthetics and an editorial board reviewer for the journal Psychology of Aesthetics, Creativity and the Arts.

His co-authors are: Daniel Graham, associate professor of psychology at Hobart and William Smith Colleges; Colin Conwell, research scientist at MIT and Harvard; and Chelsea Boccagno, postdoctoral research fellow in epidemiology at the Harvard T.H. Chan School of Public Health.