What is it about?
This study explored how the way people use their eyes at close distances, for example, while reading or studying, can reflect the kind of educational system they came from. We compared male college students who had attended ultra-Orthodox (intensive) schools versus non-ultra-Orthodox (standard) schools in Israel. Using wearable sensors, we objectively recorded how long and how far students looked at near or far objects during study sessions. A machine learning model was then trained to predict which educational background each participant had, based solely on these visual behavior patterns. The model achieved 80% accuracy (AUC), showing that shorter viewing distances and fewer long breaks between near-work sessions were strong indicators of an intensive schooling background
Featured Image
Photo by 3D Render on Unsplash
Why is it important?
The number of children and young adults with myopia (nearsightedness) has risen dramatically worldwide in recent decades, reaching epidemic levels in parts of East Asia and increasing steadily in Western countries as well. Myopia is not just a matter of wearing glasses; when it progresses to higher levels, it becomes a major risk factor for serious eye diseases, including retinal detachment, glaucoma, and macular degeneration, which can cause irreversible vision loss. Understanding what behavioral and environmental factors drive myopia is therefore essential for prevention. This study highlights how early educational intensity and near-viewing habits, how closely and how long children focus on near tasks, can leave long-lasting imprints on visual behavior. By using objective measurements and machine learning, the research identifies potential early-life behavioral markers that could help predict and prevent future myopia risk. In the long term, insights like these could guide public health interventions and education policy, balancing academic demands with visual health in children.
Perspectives
We were intrigued by how habits shaped during childhood, such as reading distance and time spent focusing on near tasks, can persist into adulthood and be objectively measured using modern wearable technology. Seeing that these visual behaviors alone could distinguish between educational systems was both surprising and thought-provoking. Given the rapid worldwide increase in childhood myopia, often linked to prolonged near work and limited time outdoors, understanding these behavioral patterns has become more critical than ever. Our findings suggest that machine learning and behavioral sensing can provide new ways to detect and even prevent risk factors for myopia before significant vision changes occur.
Ayelet Goldstein
Jerusalem Multidisciplinary College
Read the Original
This page is a summary of: Near viewing behaviors predict educational system in a machine learning model, Scientific Reports, August 2025, Springer Science + Business Media,
DOI: 10.1038/s41598-025-10108-9.
You can read the full text:
Resources
Contributors
The following have contributed to this page
