• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site

Article

Eye movement strategies in facial expression recognition are not related to the strength of inversion and thatcherization effects

The Russian Journal of Cognitive Science. 2019. Vol. 6. No. 3. P. 6-13.
Лунякова Е. Г., Гани-Заде Дж. С.

The present research focuses on the mechanisms of facial expression recognition. We explored the relationship between eye movement strategies in face perception processes and the intensity of holistic perception effects — namely, the inversion effect. It was assumed that if holistic and feature-based mechanisms rely on certain specific image viewing strategies, the  intensity of the inversion effect would be associated with certain eye movement characteristics (the number of examined facial features and the number of gaze transitions between them). The strength of the inversion effect indicated the dominance of the mechanisms of holistic perception. This was measured as a decrease in the accuracy of expression recognition of inverted images. In a facial expression recognition experiment, we analyzed responses and eye tracking data of 92 participants. Photographs of four characters from the WSEFEP database (Olszanowski et al., 2015) were used as stimuli. Each model displayed seven basic expressions. Stimuli were presented in three conditions: upright, inverted and thatcherized. A within-subjects design was used. The results showed a significant correlation between the effects of inversion and thatcherization, which argues in favor of the universality of the mechanism used by a particular person in face expression recognition. We found a high correlation between the eye movement characteristics under the three conditions of presentation, which indicates an individual-specific type of oculomotor activity. However, no correlation was found between the strength of holistic processing and certain eye movement characteristics. Most likely, oculomotor strategies for collecting information do not reflect the analytic or holistic mechanisms of its processing in facial expression recognition.