Individual differences in infants’ face viewing reflect domain-general visual biases

Before a baby can reach out to grasp a toy, they learn about the world by shifting their gaze. This directed attention is how they begin to actively shape their experience, and because so much of an infant's waking time is spent near other people, human faces become a primary source of visual information. During these early months, the brain's visual system is rapidly tuning itself to make sense of what it sees. When adults watch a face, they tend to look at the eyes and mouth, regions that hold the most social clues. For infants, looking at the eyes is considered a hallmark of healthy social development, while the ability to shift attention to the mouth often depends on the situation, such as whether the person is speaking. Yet, even within these general trends, there are surprising differences between individual babies. Some infants stare intently at the eyes, while others seem more interested in the mouth or other parts of the face. For decades, scientists have assumed these differences reflect variations in how a baby processes social information or their motivation to engage with others.

A new series of studies challenges this long-held view by suggesting that these individual differences might not be about social thinking at all. Instead, the research indicates that the way a baby looks at a face is deeply connected to how they look at non-social objects, like a chair or a tree. The findings suggest that what we see as "social" looking is actually driven by broad, general visual habits that are unique to each person. These habits appear very early in life and seem to be rooted in the basic physical structure of how the eye and brain sample visual information, rather than in the specific social meaning of the object being viewed.

To test this idea, researchers at Uppsala University conducted three separate experiments involving hundreds of infants. In the first study, they observed 68 babies who were around seven months old. Using a specialized eye-tracking system that recorded where the babies looked on a screen, the researchers showed the infants a series of images. These images included both human faces and various non-social objects, such as tools and natural items. The babies watched these pictures for four seconds each. The researchers measured exactly where the babies' eyes landed, paying close attention to whether they looked higher or lower on the screen, and whether they preferred the left or right side.

The results were striking. The babies who looked higher up on the pictures of faces also tended to look higher up on the pictures of objects. Similarly, those who looked toward the left side of a face also looked toward the left side of an object. This connection was strong and consistent, suggesting that a baby's visual habits are not specific to faces but are a general way of seeing the world. Even when the researchers looked at the very first moment a baby looked at a picture, the same pattern held true. The study also examined a specific measure often used to gauge social interest: the ratio of time spent looking at the eyes versus the mouth. They found that a baby's preference for the eyes over the mouth was also linked to how high they looked on the non-social objects. This implies that the tendency to focus on the eyes might simply be a side effect of a general habit of looking higher up, rather than a specific drive to engage socially.

To ensure these habits were not just learned over time, the researchers repeated the experiment with a much younger group of infants, only three months old. At this age, babies have had very little visual experience, their vision is still developing, and they cannot yet sit up without support. Despite these limitations, the same strong connection appeared. The three-month-olds who looked higher on faces also looked higher on objects. This finding is crucial because it shows that these general visual tendencies emerge almost immediately after birth, long before complex social learning could take place. It suggests that the foundation for these individual differences is laid down very early, likely in the physical wiring of the visual system itself.

The final experiment pushed the investigation further to see if these patterns held up in older babies and with more realistic stimuli. The researchers studied 97 infants who were between eleven and twenty months old. Unlike the previous studies which used static, still images, this experiment used moving videos of people talking and making facial expressions. In real life, faces are dynamic and often accompanied by sound and emotion. When babies watch a person speaking, they typically shift their attention from the eyes to the mouth to catch the words. The researchers wanted to know if the general visual habits would still be visible even when the whole group of babies was looking lower on the face to watch the mouth.

The answer was yes. Even though the babies as a group looked more at the mouths of the speaking actors, the individual differences remained. A baby who naturally looked higher on the moving face also looked higher on the static objects shown in the same session. This held true for the first glance and for the overall pattern of looking. The study also checked whether these looking patterns were related to the babies' actual social skills or language development, as reported by their parents. Surprisingly, there was no link. The babies who looked more at the eyes or the mouth did not show better or worse social or language abilities than those who looked elsewhere. This lack of connection further supports the idea that these looking patterns are driven by general visual mechanics rather than social competence.

The researchers took great care to rule out other explanations for their findings. They considered whether the results could be caused by errors in the eye-tracking equipment, such as a slight miscalibration that might make all babies look in the same wrong direction. They ran multiple checks and found that even when accounting for these technical factors, the strong link between looking at faces and looking at objects remained. They also considered whether the babies were simply moving their eyes in a certain way due to physical limitations, but the consistency of the patterns across different ages and types of stimuli suggests a deeper, more fundamental cause.

These findings suggest that the way an infant scans a face is largely shaped by domain-general visual biases. In other words, the unique way a baby samples visual information applies to everything they see, not just people. This does not mean that social development is unimportant, but it does mean that individual differences in face-looking are not necessarily a direct window into a baby's social mind. Instead, these differences may reflect basic variations in how the eye and brain are built, such as differences in the density of light-sensitive cells in the retina or how the visual map is organized in the brain.

The implications of this research are significant for how we understand early development. If a baby's gaze is guided by these general visual habits, then the visual experience they receive is unique to them from the very beginning. Some babies might naturally sample the eyes more often, while others might sample the mouth or the cheeks, simply because of where their eyes tend to land. This early experience could then shape how their brains develop, potentially influencing how they learn to recognize faces or process language later on. However, the study also cautions against assuming that a baby's looking pattern is a sign of a social disorder or a specific social strength. The results call for a re-evaluation of how we interpret individual differences in infant gaze, reminding us that what looks like a social choice might actually be a fundamental quirk of human vision.