Reduced Representational Momentum for Subtle Dynamic Facial Expressions In Autism Spectrum Disorders

Background:  Individuals with ASD are impaired in social communication via emotional facial expressions. Previous research have investigated facial expression processing in ASD, but reported inconsistent findings. Almost all of these studies used static facial expressions as stimuli. Emotional communication in daily life is mainly based on dynamic facial cues. A growing body of literature has shown that dynamic facial expressions enhance various psychological activities. The findings suggest that individuals with ASD may evidently have difficulty processing of dynamic facial expressions. A recent study (Yoshikawa & Sato, 2008) identified the representational momentum (RM) for dynamic facial expression. The study found that participants perceived the last image of dynamic facial expressions to be more emotionally exaggerated than the static facial expression.

Objectives: We examined the RM for dynamic facial expressions, which indicates more exaggerated subjective perception of last images of dynamic than static facial expressions. Based on the evidence that individuals with ASD are impaired in social communication via emotional facial expressions, we predicted that individuals with ASD show reduced RM for dynamic facial expressions

Methods: 12 individuals with ASD and 12 age- and gender-matched typically developing controls were participated in this study. We presented a total of 14, 20, and 26 image frames in succession in subtle (52%), medium (80%), and intense (108%) condition, respectively. In dynamic condition, each flame was presented for 10 ms. In static conditions, the last frame of dynamic facial expressions was presented. The total presentation time is the same as those of dynamic facial expression conditions corresponding with the intensity. Participants were asked to match a changeable emotional face display with the last image of presented dynamic and static facial expressions.

Results:  The ratios between the intensity of responded and presented images were calculated and subjected to the group × presentation × intensity ANOVA. Most important, the results revealed a significant three-way interaction (F(2,44) = 5.19, p < 0.05). Follow-up analysis revealed that the simple interactions between group and presentation condition were significant only in 52% condition (F(1,66) = 12.03, p < 0.05). Follow-up simple-simple main effect analysis of group in 52% condition revealed that typically developing controls perceived more exaggerated images than individuals with ASD did in the dynamic condition (F(1,132) = 6.27, p < 0.05), but not in the static condition (p > 0.10).

Conclusions:  Our results revealed reduced RM for subtle dynamic facial expressions in ASD. This group difference is in line with previous studies suggesting that dynamic information is more important for the processing of subtle, than intense, emotional expressions. Facial expressions are often displayed with subtle intensity in daily life communication. The RM for dynamic facial expressions might be useful to detect emotion in another’s face, and to predict the behavior what person will take and the sudden change in the environment. The reduced RM for subtle dynamic facial expressions might prevent individuals with ASD to interact other people smoothly in real life communication.