Complexity analysis of eye-tracking trajectories: Permutation entropy may unravel cognitive styles

Abstract

We propose a novel adaptation of permutation entropy analysis applied to eye-tracking data. Eye movements arising during cognitive tasks are characterized as sequences of trajectories within a space of ordinal trajectory patterns, thus taking advantage of recent advancements in the study of complex processes in terms of statistical complexity. Results show correlations between the permutation entropies of the eye-tracking trajectories and the type of cognitive task being performed by the subjects. Moreover, the behavior of subjects along all the experiments cluster together into two groups within a projection of the ordinal pattern space in the three principal components. This strongly suggests the existence of two different underlying problem solving styles among the subjects, which are expressed in how the movement sequences are organized.