Bringing Visuospatial Research Into the Real World

Socially relevant relations • Navigating with cognitive maps

It has long been said that the eyes are the window to the soul, providing an unparalleled glimpse into the soul of a human being. Today, psychological scientists continue to plumb the depths of all that these sensory organs have to tell us about how we experience the world. 

Research on visual attention often involves the use of static images, but if researchers want to fully understand human perception, they need to use experimental stimuli that better reflect the complexity of the real world, wrote Nicolas Roth (Technische Universität Berlin) and colleagues in a forthcoming Psychological Science article. 

“Even if the scene before our eyes remains static for some time, we might explore it differently compared to static images, which are commonly used in studies on visual attention,” explained Roth and colleagues. “Our results reveal observers’ experience-based expectations, reinforcing that we can learn most about attention in a dynamic world by studying it in dynamic scenes.” 

To do so, the researchers monitored the eye movement of 20 participants while they viewed either images or brief videos that unfroze after 5 seconds. In both cases, participants were informed whether or not the set of images they were about to view would begin moving after 5 seconds. The researchers then compared participants’ gaze behavior during the first 5 seconds of each trial, when both the images and videos were still static. 

During the first second of each trial, Roth and colleagues found that participants in both conditions preferred to look at objects at the center of the screen before scanning the entire image for other objects that may be important to the scene and areas with a high potential for movement. In the remaining 4 seconds, participants then focused on rich visual features like text in the static images and continued to focus on the areas with a high potential for change (PfC) in the frozen videos. 

Participants’ experience-based expectations caused them to scan parts of the frozen videos where movement was most likely to occur, Roth and colleagues wrote. 

“This means that images, as commonly used in psychological studies, actually conceal experience-based expectations that affect gaze behavior in the potentially dynamic real world,” Roth said in an interview. 

These findings align with previous research suggesting that eye movement often anticipates rather than responds to movement in the real world, the researchers noted.  

Related content: Collected research on Vision

“Our results suggest that how we behave is constantly affected by what we expect to happen in the (near) future,” Roth said. 

The results also impact our understanding of the animate-monitoring hypothesis of vision, Roth added. This hypothesis suggests that our gaze is attracted to animate objects because it was evolutionarily advantageous to attend to entities with the ability to move themselves, such as other animals. When Roth and colleagues further analyzed participants’ gaze behavior, however, they found that participants focused on high PfC areas regardless of whether the object that was potentially about to move had animacy or not. 

This suggests that the preference for movement described by the animate-monitoring hypothesis may reflect humans’ tendency to attend to the potential for change, rather than a preference for animacy in particular (Roth et al., in press). 

Because of the robust strength of this effect, PfC-related gaze behavior could be used in future research to gauge how well participants understand an image, Roth said. The ability to measure participants’ expectations using gaze behavior could be especially valuable in research with groups that cannot understand complex task instructions, including children and nonhuman animals, he added. The group’s ongoing research on gaze behavior and Parkinson’s disease suggests that experience-guided gaze behavior could also be used to measure how well patients are responding to treatment for neurological conditions. 

Socially relevant relations 

The sight of a person facing us is another element that draws our attention to a scene, wrote Nicolas Goupil (Université Claude Bernard Lyon 1) and colleagues in a 2024 Psychological Science article. People also prefer to look at pairs of people who are facing toward, rather than away from, each other, the researchers noted. 

“Facing another is a powerful signal of social engagement,” Goupil and colleagues wrote. “Social interaction, even when concerning others, would retain high social relevance because the way in which others interact provides important information for regulating one’s own behavior, for social learning and knowledge.” 

With that in mind, we may have evolved a visual preference for facingness—when two people can simultaneously perceive one another—so that we don’t miss out on valuable social stimuli, Goupil and colleagues explained. This is supported by previous research that found that we are more likely to notice a pair of facing versus nonfacing bodies. We also visually process pairs of facing people more quickly because our brains interpret them as one visual unit rather than two separate bodies, the researchers added. 

Through a series of five experiments, Goupil and colleagues found that humans consistently demonstrate a preference for facingness by the time we are 5 years old, but that it may emerge in infants as young as 7 months. 

In their study of 24 adults, eye-tracking software confirmed that participants looked at images in which pairs of people were facing each other longer than when they were facing away from one another. Goupil and colleagues also replicated these findings through a study of 21 juvenile macaque monkeys, suggesting that preference for facingness may have emerged in our shared evolutionary past. 

And in a subsequent study, 138 participating adults rated images of people facing each other high in social semantic dimensions like meaningfulness and emotion, reinforcing that these images held higher social value. 

Finally, Goupil and colleagues shifted their focus to the gaze behavior of children. Through a study of 80 human infants, the researchers found that while infants under 10 months old tended to spend more time looking at nonfacing dyads, infants between 15 and 18 months showed no preference for facing or nonfacing dyads. When the researchers repeated this study with a group of twenty 3-year-olds and thirty 5-year-olds, they found that the 3-year-olds also demonstrated no preference for facingness. The 5-year-olds, however, demonstrated the same preference for facingness as adults in previous studies. 

This change could represent an important milestone in children’s social development, and assessing children for this behavioral marker of social cognition could provide further insight into typical and atypical developmental pathways, the researchers wrote. 

“In this spirit, the preference for facingness would mark a milestone in social cognition development, signaling that a child represents others as social agents who attend to, engage with, and act upon one another,” the researchers explained (Goupil et al., 2024). 

Navigating with cognitive maps 

Our visuospatial abilities are also an important aspect of how we navigate our physical environment. The existing research on path integration has established that people have an intuitive understanding of the layout of 2D flat environments, also referred to as Euclidean, Misun Kim (University College London and Max Planck Institute for Human Cognitive and Brain Sciences) said in an interview about her Psychological Science article. 

“This means that when people move around, they could update their location in their mental map and find the efficient route back to the start location even when they cannot see the prominent landmarks,” Kim said. 

But there has been considerably less research on how people build cognitive maps of the non-Euclidean, or spherical, environments that exist in the real world, Kim continued. To extend this research, Kim and Christian F. Doeller (Max Planck Institute for Human Cognitive and Brain Sciences) tasked a group of 20 participants with navigating two distinct environments using a virtual reality headset and an omnidirectional treadmill. Though both environments featured a starry sky and a set of 12 stationary animals facing toward a central point of the map, one was a Euclidean planar environment and the other was a small, non-Euclidean spherical planet. 

After being given 15 minutes to explore the current environment, participants completed memory training trials to help them learn the location of each animal by starting at various locations and using the other animals as landmarks. During the test phase, however, the animals were hidden once participants started moving, requiring them to navigate to where they believed a specific animal would be without any visual landmarks. 

Next, participants completed a path-integration task known as the triangle-completion task. During this task, participants were instructed to walk down an L-shaped hallway. When they reached the end of the hallway, the walls became invisible, and they were tasked with walking back to the point where they believed they had started. 

When Kim and Doeller analyzed the paths used by participants in each of these tasks, they found that participants’ behavior more closely aligned with that predicted by Euclidean, planar geometry regardless of whether they were in the Euclidean or non-Euclidean environment.  

“People have difficulty in finding the optimal route in this unfamiliar space,” Kim said. “They seem to very much rely on the path-integration system built for 2D flat surfaces, not adjusting to the curved spherical surface. It implies that people use multiple patches of 2D cognitive maps rather than building a perfect 3D map containing the sphere.” 

Participants were able to use these 2D cognitive maps to locate objects in the 3D spherical environment with a reasonable degree of accuracy, Kim and Doeller wrote. 

During the animal-location task in particular, participants’ use of planar maps to navigate spherical environments led them to make large errors when choosing which direction to move at the start of the task. These initial direction errors ultimately resulted in relatively small position errors at the end of the task, however, allowing participants to perform better than chance on both maps, the researchers explained. This was because of the converging nature of spherical geometry, Kim said, as all straight lines intersect on a sphere. 

“In the natural world that humans live in, surfaces often have modest curvature that can be closely approximated by Euclidean geometry,” Kim and Doeller wrote. “If animals can achieve reasonable navigation accuracy by utilizing an existing path integration system and multiple planar maps, there might be a limited demand for building more complex high-dimensional maps.” 

Maintaining multiple small cognitive maps of a single space may also allow humans and other animals to more easily update their cognitive maps as they continuously interact with an environment. 

“Building a precise map for high-dimensional space can be costly and inefficient when only a part of the space is relevant for behavior,” the researchers explained. “Thus, a dynamic and flexible arrangement of cognitive maps becomes more desirable.” 

Previous studies of people without formal math education, as well as rats raised in artificial spherical environments, have found evidence of the same behavior, Kim and Doeller noted, suggesting that this may arise from an intuitive understanding of Euclidean geometry. 

The neural mechanisms and behavioral strategies used to navigate flat and spherical environments could be used for solving other cognitive tasks as well, they said (Kim and Doeller, 2024).  

“Multiple complementary cognitive strategies and dynamic interaction with environment enables adaptive behavior in humans,” Kim said. “Approximation and dividing the high-dimensional task into multiple low-dimensional tasks could be a general principle of our mind.” 

Going forward, Kim would like to explore how senses other than vision may influence human path integration. For example, running a spherical navigation experiment in a zero-gravity environment like the International Space Station could provide valuable insight into how gravity and vestibular sensation may influence our sense of orientation, Kim said. She would also like to test if participants develop different types of cognitive maps after spending long periods in non-Euclidean spherical environments and how neurons in the brain that encode head direction respond to flat and spherical surfaces. 

After all, our eyes may be able to reveal a lot about how we perceive the world around us, but they are just one of the many windows through which we perceive our lives. 

Back to top

Feedback on this article? Email [email protected] or login to comment.

References 

Goupil, N., Rayson, H., Serraille, É., Massera, A., Ferrari, P. F., Hochmann, J. R., & Papeo, L. (2024). Visual preference for socially relevant spatial relations in humans and monkeys. Psychological Science, 35(6), 681–693. https://doi.org/10.1177/09567976241242995 

Kim, M., & Doeller, C. F. (2024) Multiple planar maps for a non-Euclidean environment: human path integration and spatial memory on a sphere. Psychological Science, 0(0). https://doi.org/10.1177/09567976241279291  

Roth, N., McLaughlin, J., Obermayer, K., & Rolfs, M. (2024). Gaze Behavior Reveals Expectations of Potential Scene Changes. Psychological Science, 0(0). https://doi.org/10.1177/09567976241279198