Study of an Extensive Set of Eye Movement Features: Extraction Methods and Statistical Analysis

  • Ioannis Rigas Texas State University
  • Lee Friedman Texas State University
  • Oleg Komogortsev Texas State University
Keywords: eye movements, feature extraction, saccades, fixations, post-saccadic oscillations, variability, test-retest reliability, factor analysis

Abstract

This work presents a study of an extensive set of 101 categories of eye movement features from three types of eye movement events: fixations, saccades, and post-saccadic oscillations. We present a unified framework of methods for the extraction of features that describe the temporal, positional and dynamic characteristics of eye movements. We perform statistical analysis of feature values by employing eye movement data from a normative population of 298 subjects, recorded during a text reading task. We present overall measures for the central tendency and variability of feature values, and we quantify the test-retest reliability of features using either the Intraclass Correlation Coefficient (for normally distributed and normalized features) or Kendall’s coefficient of concordance (for non-normally distributed features). Finally, for the case of normally distributed and normalized features we additionally perform factor analysis and provide interpretations of the resulting factors. The presented methods and analysis can provide a valuable tool for researchers in various fields that explore eye movements, such as in behavioral studies, attention and cognition research, medical research, biometric recognition, and human-computer interaction.
Published
20-03-2018
How to Cite
Rigas, I., Friedman, L., & Komogortsev, O. (2018). Study of an Extensive Set of Eye Movement Features: Extraction Methods and Statistical Analysis. Journal of Eye Movement Research, 11(1). https://doi.org/10.16910/jemr.11.1.3
Section
Articles