Abstract
Previous studies have revealed that dynamic facial expressions are better recognized (e.g. Ambadar et al., 2005), and are processed in partially different brain areas (e.g. Schultz & Pilz, 2009), than static expressions. Still unknown is if the visual strategies underlying the recognition of dynamic and static expressions differ. Here, the ocular fixation pattern (Exp. 1) and spatial frequency (SF) utilization (Exp. 2) of 20 participants were measured with static and dynamic expressions. In both experiments, participants categorized pictures or videos (block design) of the six basic facial expressions and neutrality. In Exp. 1, the stimuli were presented unaltered and in Exp. 2, they were randomly filtered using SF Bubbles (Willenbockel et al., 2010). In both experiments, stimuli were presented for a duration of 500 ms. Fixation patterns were analyzed using iMap4 (Lao, et al., 2015). A repeated measures ANOVA revealed a main effect of condition (p< 0.05), indicating more fixations on the eye area with static than dynamic expressions, and more fixations on the nose area with dynamic than static expressions. SF tunings were obtained by conducting a multiple regression analysis on the random SF filters and accuracies across trials. Statistical thresholds were found with the Stat4Ci (Chauvin et al., 2005). A SF band peaking at 17.7 cycles per face (cpf) with a full-width-half-max (FWHM) of 30.3 cpf, and a SF band peaking at 16.0 cpf with a FWHM of 29.0 cpf, were found with static and dynamic facial expressions, respectively. SF between 3.7 and 5.7 cpf were more utilized with dynamic than with static expressions, and those between 18.7 and 27.3 cpf were more utilized with static than with dynamic expressions (p< 0.025). Together, these results suggest that the recognition of dynamic facial expressions can be performed using lower spatial frequencies, decreasing the need to directly fixate on facial features.
Meeting abstract presented at VSS 2016