Eye Tracking with Senior Data Scientist Irati R. Saez de Urabain

It turns out the old saying, ‘eyes are the window to the soul,’ is more true than we thought, especially as our understanding of the link between vision and cognitive behaviour grows. The importance of eye movements have been studied and observed from as early as the 1800s, where eye movements and motions were painstakingly tracked, manually. By the 1970s, psychologist Alfred L. Yarbus’ seminal book, Eye Movements and Vision, had revolutionized the study of eye tracking which led to a major expansion of research. Today, eye tracking is a method in which a subject’s eye movements are monitored as they respond to stimuli, using a device called an eye tracker.

So how does eye tracking actually work?

As a part of her PhD, Dalia’s Senior Data Scientist, Irati R. Saez de Urabain, investigated eye-movement development in babies and developed tools to analyze eye tracking data collected from babies (*). Because babies can’t communicate with words, eye tracking is one of the most essential tools for conducting cognitive developmental research among infants. Eye tracking has contributed to fascinating areas of research; from understanding at what age infants develop long term memory to exploring the communication skills of infants raised by blind parents.

When doing research with babies, a typical testing session starts with some light playtime that helps the infant (and the caregiver) feel comfortable with the experimenter and the lab. Once everyone is happy, the infant is set up in the caregiver’s lap to face the stimuli and the eye tracker. The eye tracker then uses some complex technology in order to catch the subject’s gaze as she/he responds to the stimuli being presented. “More specifically,” Irati explains, “most remote eye trackers use a technique called corneal reflection, which is based on recording the reflections on the cornea of a near infra-red light. As a result of the spherical properties of the eyeball, the illumination of the infrared light creates a bright glint on the back of the cornea that remains relatively stationary while the eye moves. It is this glint and its distance relative to the centre of the pupil that is used to estimate the gaze on the screen.”

eye diagram

The glint (the white dot) remains stationary while the pupil moves. The distance between the glint and the pupil (the red line) is used to track the eye’s movements.baby eye tracking

Irati’s photogenic young son was also one of the participants in her experiments!


But it’s not that simple. As Irati mentions, babies’ eyes are less developed and more watery than adults’ which leads to much lower data quality. Besides that, babies are notoriously… uncompliant. Infants do not respond to experimental instructions and are also easily distracted away from the experiment’s stimuli, making the process more complicated. One of Irati’s main topics was developing tools to analyze low quality eye tracking data to counter these setbacks.

Irati’s work on eye tracking with babies was featured in Wired, and has also been published in a chapter of the Cambridge Encyclopedia of Child Development entitled ‘Eye Tracking’ (**). She completed her PhD at the Center for Brain and Cognitive Development in London.

What else is eye tracking technology used for?

Today, eye tracking tools are incredibly efficient and can easily automate and track a subject’s gaze wherever it goes. With these major technological improvements, eye tracking has gone far beyond developmental research. Irati points out that eye tracking is also a fundamental tool for people with disabilities or impairments who can only communicate via their eyes, with a few low cost devices already in the market (costing about $100). Eye tracking is also a major component of human-computer interactive interfaces (ex. virtual reality, gaming, personal tech devices) and can be used to measure consumer behaviour for visual marketing or product design.


(*) Saez de Urabain, I.R. and Johnson, Mark H. and Smith, Tim J. (2014) GraFIX: a semiautomatic approach for parsing low- and high-quality eye-tracking data. Behavior Research Methods 47 (1), pp. 53-72. ISSN 1554-3528.

(**)Smith, Tim J. and Saez de Urabain, I.R. (2017) Eye tracking. In: Hopkins, B. and Geangu, E. and Linkenauger, S. (eds.) Cambridge Encyclopedia of Child Development. Cambridge, UK: Cambridge University Press, pp. 97-101. ISBN 9781107103412.

Photos from Wired