Human eye movements follow a pattern similar to foraging by hunter-gatherers

Human cognitive processes are often approximated as Brownian motion. But in certain cases, such as visual searches, this approximation is insufficient, as it ignores underlying statistical properties. Using data from two independent eye-tracking experiments, Piotr Błażejczyk and Marcin Magdziarz studied visual search dynamics and developed a model for gaze direction and velocity.

When instructed to find a specific picture or word on a screen, individuals tend to rest their gaze in one of five distinct areas for a random length of time, before shifting their eyes to a different area at a constant speed. Błażejczyk and Magdziarz describe the dynamics as a Lévy walk – a type of random motion with constant velocity – with the pause periods determined by a random Markov chain distribution. Lévy walks have been used to describe other human and animal behaviors before, such as for modeling movement during foraging activities.

“When we first heard of experiments describing visual search dynamics, our first thought was to check if this type of dynamics can be described by Lévy walk,” said Magdziarz. “The idea comes from the well-known fact that there is a linear dependence between the peak velocity and the magnitude of human saccadic eye movements.”

Results from their combined Lévy walk and Markov chain model agreed remarkably well with experimental data, confirming their methodology accurately describes an individual’s eye movements.

“Our results show that the dynamics of human eye movement has the same optimal properties as macroscale foraging patterns performed by a variety of animals,” Magdziarz said.

The pair hopes to extend their eye-tracking research to monitor where the eye fixates while browsing the web to help optimize webpage design.

Source: “Stochastic modeling of Lévy-like human eye movements,” by Piotr Błażejczyk and Marcin Magdziarz, Chaos (2021). The article can be accessed at https://doi.org/10.1063/5.0036491 .