Purchase this article with an account.
Motoyasu Honma, Yoshihisa Osada; The effect of the dynamic property of a face on the recognition of facial expressions and eye movements. Journal of Vision 2004;4(8):910. doi: https://doi.org/10.1167/4.8.910.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
We focused on the effect of the dynamic property of moving faces on face processing, particularly on eye movements as the cognitive strategy of an observer. We investigated what cue observers use to judge moving facial expressions when presented band-pass-filtered faces. METHOD: A 2AFC task was employed. Face images were produced by morphing the stimulus ranging from the neutral to the prototype face of each emotion [morphing rates (ms); 10 (336), 30 (1008), 50 (1680), 100% (3360)]. Moving and static face images were presented at the same intervals of particular morphing rates. We used a 2 × 2 × 3 factorial design with presentation formats (dynamic and static), facial expressions (happy and sad), and spatial frequencies (low-pass filtered, high-pass filtered, and original). We also recorded eye movements to monitor saccades made by the observer, and analyzed the data by the VVF (Variance Value of Fixation-points) equation. The VVF changes with the extent of eye movements: the increase of the VVF means that observer fixates many different locations of a face. RESULT: [Original faces] The VVF decreased on moving faces in comparison with that obtained on static faces. The VVF on happy faces was smaller than that on sad faces. [High-pass filtered faces] The VVF increased for moving faces in comparison with that obtained for static faces. The VVF on happy faces were larger than that on sad faces. In addition, the VVF on happy faces did not increase with morphing rate, although the VVF on sad faces increased. DISCUSSION: The results suggest that humans may utilize different cues to recognize moving and static faces. This implies that the effect of the dynamic property of a face changes with the spatial frequency of facial expressions. In particular, in the case of a high-pass filtered happy face, the dynamic property may facilitate the integration of local components of the face, and make the configural information of the face.
This PDF is available to Subscribers Only