People with high levels of social anxiety tend to perceive individuals and groups as walking straight towards them, even when the actual movement is diagonally away. The study, published in Cognition and Emotion, reveals that these perceptual biases occur at both the individual and crowd level. The results suggest that social anxiety sensitizes the brain to potential social threats by changing how basic spatial logic is applied to moving objects.
People with social anxiety have a strong fear of being judged by others. Psychological models suggest that these people have perceptual biases related to threat. They naturally prioritize information that could suggest interpersonal danger in social situations. While past research has looked closely at how anxious people read facial expressions, people often rely on broader body language to infer intentions from a distance.
Researchers use a concept called biological motion to study this phenomenon. To separate movement from physical appearance, scientists use simulated human figures made entirely of moving dots placed at major joints such as knees, hips, and shoulders. Although visually sparse, this animation style allows the human brain to instantly recognize walking patterns. Observers can infer the intentions and emotional state of a simulated person simply by observing the rhythmic changes in the graphic dots.
Previous research on this topic relied on two-dimensional dot animation. Those particular images lacked visual depth, so the figures often appeared vague to the viewer. An observer can interpret the exact same walking animation as either walking towards you or walking directly towards you. In these older studies, socially anxious viewers exhibited conflicting responses that alternated between intense sensitivity to approaching figures and a pattern of psychological avoidance in which they only looked at the person walking away.
Error management theory suggests that human cognition has evolved to make the least costly mistakes when faced with visual uncertainty. In the context of human ancestral survival, preparing for the approach of a hostile figure is a low-cost mental action. However, failure to detect incoming physical threats can be extremely dangerous and potentially deadly. To determine whether socially anxious people were overapplying this ancient safety mechanism, scientists needed to fundamentally remove the visual ambiguities of the flat image displays used in previous behavioral literature.
A team of psychologists at the Catholic University of Korea designed a new visual experiment using three-dimensional representations. The research team, including psychologists Jae-Won Yang, Jisu Choi, and Jiyeon Park, added shading and spherical perspective cues to the walking dot. This allows the animated figures to always appear to be moving forward on a realistic surface. Viewers had to judge the exact angle of approach, rather than guessing at the absolute direction.
The researchers recruited 147 college students and assessed them using a widely accepted psychological questionnaire. They divided students into high and low social anxiety groups based on survey scores. The researchers also measured baseline depressive symptoms to ensure that the observed differences were strictly related to social anxiety, as anxiety and depression often overlap in the general public.
In the first part of the experiment, participants watched a short animation of a single dot-based shape. Characters either walk straight towards the viewer or walk slightly away in various increments up to 24 degrees. The animation lasted only a quarter of a second. Participants then pressed a key to indicate whether they felt the pedestrian was coming towards them or whether they felt it was safe to pass.
This test allowed researchers to set individual detection limits for everyone. Participants in the high social anxiety group were found to have significantly lower spatial thresholds than those in the low anxiety group. Highly anxious individuals classified pedestrians as approaching even when they were at an angle away from the direct collision path. Their visual processing supported the hypothesis that a lone human was walking straight towards them.
The experiment then moved on to testing a phenomenon known as ensemble perception. The human visual system uses cognitive shortcuts to quickly extract summary statistics from groups. This allows a person to quickly determine the average direction in which a flock of birds is flying, or assess the overall mood of a crowd of humans. The researchers wanted to know whether the proximity bias seen in isolated individuals would extend to larger crowds of moving strangers.
Participants simultaneously viewed a digital crowd of 10 dot-based shapes on a screen for 0.5 seconds. The researchers customized this crowd for each person based on their original spatial detection thresholds. A certain percentage of pedestrians in the crowd moved in a straight line, while others moved at an angle slightly wider than the participant’s testing limit. Participants had to judge whether the entire crowd was moving closer to them on average.
The high social anxiety group again showed a measurable bias compared to the low anxiety group. Socially anxious participants needed fewer direct walkers in the group to judge the entire crowd as approaching. Their brains naturally distorted the average movement of the broader group toward the looming threat. Because the researchers calibrated the crowd test using each person’s specific baseline, the group-level bias exhibited another distortion in how the anxious brain averages out visual information.
Survey data provided further details about this visual bias. The researchers assessed two different types of anxiety: fear of being observed by others and fear of actively interacting with others. Perceptual approach bias was only correlated with fear of being observed. The researchers believe that because the point-based figures lacked facial features and clear social context, they may have caused evaluation fear rather than conversational anxiety.
Certain physical behaviors seen in people with social anxiety may be explained by a perceptual bias toward approaching strangers. Overestimating the physical distance of an approaching person may increase the subjective sense of danger. This visual distortion can lead to avoidance of physical actions and involuntary postural sway, which are common in clinical observations. Taking these automatic visual judgments into account may ultimately aid in the development of new physical therapy and exposure strategies.
This study has some methodological limitations that require future scientific investigation. Point-light pedestrians are incredibly useful abstractions, but they inevitably fall short of naturalistic, everyday visual perception. The intentionally short exposure times also meant that the animation only depicted a fraction of a normal human walking stride. In the real world, the human brain relies on continuous visual feedback to adjust spatial judgment over long periods of time. Future experimental studies may test these spatial biases in highly realistic interactive settings using fully rendered virtual reality environments and real actors.
Participants consisted only of college students, not individuals with clinical psychiatric diagnoses. Replicating experiments in clinically defined patients will help validate the practical application of these findings. The effect sizes recorded during the task also indicate that additional psychological or biological variables influence how people perceive biological motion. Still, the data support the idea that social anxiety actively reshapes basic spatial judgments in ambiguous situations.
The study, “People Are Approaching Me: Biased Ensemble Perception of Biological Motion in Social Anxiety,” was authored by Jisu Choi, Jiyeon Park, and Jae-Won Yang.
