Modeling Age Dependent Shifts in Visual Attention

15 concepts7 papers1 proposal

Computer vision and saliency models traditionally treat human vision as a static, universal process. They operate under the assumption that an image which draws the eye of a 25-year-old will capture the attention of a 5-year-old or an 80-year-old in the exact same way. However, a growing body of developmental and gerontological research reveals that visual attention undergoes profound, systematic shifts across the human lifespan.

Childhood: From Saliency to Exploration

In early development, visual attention is heavily constrained by both physiological maturity and immediate physical utility. Infants initially rely on low-level, bottom-up visual saliency to guide their gaze, but as Yisi Zhang et al. demonstrate in An automatic sustained attention prediction (ASAP) method for infants and toddlers using wearable device signals, this reliance decreases over time as older infants engage higher-level, top-down cognitive processes.

Young children also exhibit a powerful "center bias"—a tendency to focus predominantly on the middle of a visual frame, regardless of peripheral content. In Gaze distribution analysis and saliency prediction across age groups, Onkar Krishna et al. showed that standard, age-blind saliency models fail to predict children's gaze because they do not account for this intense central focus and lower overall explorativeness. By introducing a mathematically variable center-bias weight that decreases with age, the authors significantly improved saliency prediction accuracy.

Furthermore, children's gaze is driven by immediate physical interaction rather than purely social cues. In developing the ChildPlay: A New Benchmark for Understanding Children's Gaze Behaviour dataset, Samy Tafasca et al. found that children look at faces much less frequently than adults do; instead, their attention gravitates toward objects within their immediate physical reach.

The Aging Eye: Feature Shifts and Centralization

At the other end of the lifespan, visual strategies shift again, often narrowing or reallocating attention in response to cognitive and neural changes. A primary example occurs during face perception. While young adults heavily prioritize the eye region to decode emotions, older adults systematically shift their attention downward toward the mouth.

As Wenfeng Chen et al. note in Automatic and Controlled Processing of Facial Expressions in the Aging Brain, this mouth-centric strategy allows older adults to maintain proficiency in identifying mouth-heavy emotions like disgust, even as they struggle with eye-heavy emotions like fear. However, this is not merely a conscious change in strategy. Using a "moving window" technique to restrict the visible field of view in Exploring emotional expression recognition in aging adults using the Moving Window Technique, Elina Birmingham et al. demonstrated that the drop in emotional recognition accuracy among older adults stems from deeper differences in how visual information is processed, rather than where they look.

As aging progresses, visual exploration often becomes increasingly centralized and less feature-specific. In Gaze Fixation Patterns Correlate with Visual Attention and Memory: the Results of a Pilot Study in Healthy Subjects, Gleb Danilov et al. found that older adults tend to shift attention from informative peripheral details to the central areas of objects. This loss of edge-sensitivity directly impairs "pattern separation"—the ability to distinguish between highly similar memories because the diagnostic details located at the boundaries of objects are missed.

Gaze Dynamics as Digital Biomarkers

Understanding these lifespan-dependent shifts does more than just improve computer vision models; it provides a window into cognitive health. When the natural variance of aging transitions into pathological cognitive decline, gaze patterns alter in predictable ways.

In Eye-movement patterns in face recognition are associated with cognitive decline in older adults, Antoni B. Chan et al. demonstrated that a pronounced shift away from scanning specific, diagnostic facial features toward a holistic, center-of-the-face fixation pattern correlates strongly with lower cognitive status in older adults. Because this behavioral shift tracks with cognitive decline independently of chronological age, gaze dynamics can serve as an objective, non-invasive digital biomarker for tracking neurodegenerative changes.

Key Foundational Work

Gaze distribution analysis and saliency prediction across age groups (2018) — contributed 8 concept elements to this direction.
Exploring emotional expression recognition in aging adults using the Moving Window Technique (2018) — contributed 1 concept elements to this direction.
Eye-movement patterns in face recognition are associated with cognitive decline in older adults (2018) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_age_specific_spatial_scale_preference_d46ac2ef,
  title = {age-specific spatial scale preference},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/age-specific-spatial-scale-preference/rdr_d46ac2ef58024896a12887a9f294007d},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_age_specific_spatial_scale_preference_d46ac2ef_v1,
  title = {age-specific spatial scale preference},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/age-specific-spatial-scale-preference/rdr_d46ac2ef58024896a12887a9f294007d/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake