Psychologists at the University of Toronto have discovered a link between our aesthetic preferences and the amount of energy used by our brains to process what we see.
Using deep neural network models and functional magnetic resonance imaging (fMRI), the team of researchers found that we tend to prefer images that are “easy on the eyes”—or require less energy for the visual system to represent.
“Usually, people think we like specific images because of their cultural meanings, or specific signals we receive from them,” says Yikai Tang, a PhD student in the Department of Psychology and the first author of a new paper presenting a study conducted alongside Faculty of Arts & Science professors William Cunningham and Dirk Bernhardt-Walther.
“In this research, we look at the more objective measure of energy consumption to really quantify the processing cost of what we see and find pleasurable,” says Tang.
Their findings were recently published by PNAS Nexus in, “Less is more: Aesthetic liking is inversely related to metabolic expense by the visual system.”
Energy efficiency is key to the survival of all organisms, including humans. This involves balancing the costs of both physical and cognitive activities with their potential benefits. Scientists have previously proposed that instead of painstakingly calculating these ratios, organisms may use pleasure-based signals to guide their preferences toward optimal actions.
Translating the light captured by our eyes into a meaningful understanding of the environment is no small task. Our brains use 20% of the body’s energy, and the visual system accounts for nearly half of that. This study suggests that the evolutionary pull towards energy efficiency may also influence our aesthetic preferences.
“Our sense of visual pleasure is not just a matter of taste or culture, but also a built‑in energy-saving strategy of the brain that nudges us toward environments and designs it can process with minimal effort."
“Our work shows that part of what we experience as “beauty” in everyday images may simply be our visual system rewarding itself for doing its job efficiently, using less metabolic energy to process certain pictures,” says Bernhardt-Walther.
“Our sense of visual pleasure is not just a matter of taste or culture, but also a built‑in energy-saving strategy of the brain that nudges us toward environments and designs it can process with minimal effort.”
Determining the costs of seeing
The study involved a dataset of nearly 5000 real-world images and aesthetic pleasure ratings collected from over 1000 participants. Each image was rated on a scale from one to five by approximately 50 people from the total pool who were asked, “How much do you enjoy looking at this image?”
The researchers ran these images through two artificial neural networks that approximate the human visual processing system to estimate the “metabolic costs” required to process them, defined as the number of neurons active in representing an image. They then analyzed the metabolic costs needed by the human visual system to consume the same images using fMRI.
In both cases, they identified an inverse correlation between the metabolic cost associated with each image and its aesthetic pleasure rating. In other words, the less energy an image required to consume, the more it was liked on average.
Images that are recognizable, clear, symmetrical, and/or organized—like a human face or a natural vista—are easiest for the visual system to represent and tend to be rated highest in aesthetic pleasure. Complex ones, like a busy crowd or unintelligible scribbles, place a heavier burden on the visual system to process and tend to be rated lower.
More specifically, they found that the three brain regions responsible for the recognition of scenes and faces (fusiform face area, parahippocampal place area, and the occipital place area) displayed highly significant negative correlations between metabolic costs and aesthetic pleasure. This suggests that metabolic-related aesthetic pleasure may be driven primarily by high-level visual processing at the final stage of the visual pathway.
The researchers note that the correlation they have identified applies to the perceptual system, not deliberative cognitive processing or contemplation that may occur after a first glance at an image.
In this study, participants were asked to simply rate their level of pleasure when viewing each image, without any encouragement to contemplate them further. Previous research has shown that images initially seen as complex may become more aesthetically pleasing with more contemplation. Indeed, some of the most famous and studied sights in the world are incredibly complex, like the ceiling of the Sistine Chapel.
This research does not contradict those previous studies but complements them by underlining how aesthetic pleasure is linked to energy expenditure at the earliest stages of visual processing. It is particularly useful in understanding how aesthetic pleasure is defined while passively viewing an image or scene.
More Information
To learn more about this study or to speak to its authors, please contact:
- Michael Pereira
Communications Officer, Department of Psychology, University of Toronto
psy.communications.officer@utoronto.ca