Summary: As people grow older, they increasingly rely on whether a face is centered on a screen when deciding where to look while watching videos.
Source: UCR
How do infants, children, and adults differ in where they look when watching faces on television and video?
Researchers at the University of California, Riverside examined how viewers of different ages attend to faces in dynamic media. Their research shows that a viewer’s reliance on face centering — a face appearing near the middle of the screen — grows with age. In contrast, sensitivity to a face’s visual saliency, or how much a face stands out from its surroundings, remains consistently important from infancy through young adulthood.
In the study, lead author John M. Franchak and graduate student Kellan Kadooka analyzed two visual cues that might guide gaze during video viewing: face saliency and face centering. Face saliency captures contrast, motion, color, and other image features that make a face visually prominent. Face centering captures the spatial position of a face within the frame — whether the face is near the center or at the periphery.
Previous psychological research had suggested that attention to salient parts of a scene decreases with age. Franchak and Kadooka’s findings challenge that assumption: across participants aged six months to young adults, visual saliency consistently predicted where observers looked. What changed with age was how much observers used centering as a cue to look at faces.
“Our results indicate that what develops is not sensitivity to saliency, but the tendency to treat centered faces as especially important,” said Franchak, an associate professor of psychology. “Adults reliably look more at faces placed in the center of the screen, likely because years of watching edited media teach them that directors position key characters centrally to signal importance.”
Infants and very young children were just as likely to look at faces at the edge of the screen as at the center. With increasing age, however, centering became a stronger predictor of face-looking. The researchers note that additional work is needed to determine exactly how and when this centering bias develops and which experiences foster it.
The study recruited 79 children and 20 adults. Each participant watched a set of clips that included two Sesame Street segments, two music videos, and a children’s science demonstration, with many faces appearing throughout. Eye-tracking recorded where viewers looked. For every video frame and every face, the team computed a visual saliency score and measured the face’s distance from the image center. The analysis tested how saliency and centering influenced gaze across ages.
Practical applications
Understanding which visual cues drive attention to faces in videos has clear implications for educational media and for clinical assessment. Because adults have extensive experience with edited television, they learn conventions such as centering important characters in the frame. Young children may take years to acquire these conventions, which matters for how they learn from television.
“Many studies show infants and young children can struggle to learn from television,” Franchak said. “If infants rely more on saliency than on centering, children’s educational media might increase learning by making primary characters visually salient rather than relying on centering alone.”
The findings also bear on autism research. Clinicians sometimes measure attention to faces on screens as a behavioral marker for autism spectrum disorder (ASD). Prior work shows people with ASD often look less at faces in video than neurotypical peers. By clarifying how face centering and saliency shape gaze across development, the study can inform the design of better diagnostic tasks that control for these visual features.
Franchak emphasized that distinguishing between centering and saliency is important for both research and practice. “If clinicians and researchers want to use videos to assess social attention, they need to know which face features influence looking at different ages so stimuli can be appropriately controlled,” he said.
“Viewing experience gives us automatic shortcuts for where to look,” added Kadooka. “Even when a scene contains many competing elements, past exposure to edited media biases viewers toward particular locations and characters.”
The research team plans to extend this work by testing how centering and saliency affect learning from characters in videos. They predict that older children will learn more from centered characters when those characters convey key information, and they will test that hypothesis in follow-up studies.
About this neuroscience research news
Author: Iqbal Pittalwala
Source: UCR
Contact: Iqbal Pittalwala – UCR
Image: The image is in the public domain
Original Research: Closed access.
“Age differences in orienting to faces in dynamic scenes depend on face centering, not visual saliency” by John M. Franchak et al., published in Infancy
Abstract
Age differences in orienting to faces in dynamic scenes depend on face centering, not visual saliency
This study examined how infants (6–24 months), children (2–12 years), and adults differ in the influence of two visual cues—visual saliency and spatial centering—on attention to faces in video. Using a secondary analysis of eye-tracking data, the researchers quantified each face’s saliency and measured its distance from the frame center on every video frame.
Results showed that visual saliency increased the probability that viewers of all ages would look at a face: salient faces attracted gaze consistently from infancy through adulthood. In contrast, centering affected face-looking only in older children and adults; infants did not show a greater tendency to look at centered faces. A control analysis confirmed that a general bias to look toward image center was present at all ages and did not account for the age-related increase in centering for faces.
The findings highlight that developmental changes in attention to faces in dynamic scenes are driven by growing sensitivity to face centering, rather than a decline in sensitivity to saliency. The study discusses implications for using videos in educational contexts and for refining diagnostic approaches that rely on gaze to faces.