A review of gaze entropy as a measure of visual scanning efficiency. (January 2019)

• Record Type: Journal Article
• Authors: Shiferaw, Brook; Downey, Luke; Crewther, David
• Is Part Of: Neuroscience and biobehavioral reviews. Volume 96(2019)
• Page Start: 353
• Page End: 366
• Publication Date: 2019-01
• Subjects: Gaze entropy -- Gaze -- Fixations -- Visual scanning; Psychophysiology -- Periodicals; Human behavior -- Periodicals; Animal behavior -- Periodicals; Neurology -- Periodicals; Behavior -- Periodicals; Ethology -- Periodicals; Neurology -- Periodicals; Psychophysiologie -- Périodiques; Comportement humain -- Périodiques; Animaux -- Mœurs et comportement -- Périodiques; Neurologie -- Périodiques; Animal behavior; Human behavior; Neurology; Psychophysiology; Periodicals; Electronic journals; 573.8
• DOI: 10.1016/j.neubiorev.2018.12.007
• Languages: English
• ISSNs: 0149-7634

Abstract

Highlights: There are two types of gaze entropy: GTE and SGE. There may be an optimal range of GTE for a given task. GTE indicates level of efficiency in visual scanning. SGE reflects overall spatial dispersion of gaze. Normalising observed entropy improves interpretability. Abstract: While the concept of entropy has been applied to gaze analysis, it is unclear what aspects of visual scanning it measures. In this review, we first outline gaze control as a complex system of spatial prediction. Second, we provide a brief introduction to the concept of entropy within the context of information theory as the foundation for gaze entropy measures; with a specific focus on equations for Shannon's entropy and conditional entropy. The application of these equations to gaze data is described as stationary gaze entropy (SGE) and gaze transition entropy (GTE) respectively. Third, we present an updated model of gaze orientation and propose an adaptable definition of GTE as a measure of visual scanning efficiency that underlies overall gaze dispersion measured by SGE. Finally, we review studies that have utilised GTE and SGE to assess visual scanning and discuss their results in relation to our proposed definitions and associated hypotheses. Methodological limitations in gaze entropy measures are discussed and suggestions provided to improve interpretability and generalisability of future studies.