EVALUATION OF TWO-DIMENSIONAL MODEL OF THE FACIAL EXPRESSIONS USING AFFECT GRID
 
Shibui, S., Shigemasu, K., The University of Tokyo, Yamada, H., Nihon University, Japan
 
It is a fundamental question whether facial expressions are processed categorically or continuously in the multidimensional space. According to Young et al (1997 Cognition 63 271 - 313), two-dimensional continuous model predicts that: (i) transitions between expressions should be continuous, and (ii) at least some transitions between expressions should pass through a central neutral region or a region corresponding to a third emotion. We investigated this issue using schematic facial expressions which were used by Yamada (1993 Applied Cognitive Psychology 7 257 - 270). These stimuli were made to deform the shapes of facial elements from the standard ones by changing eight feature points. We measured the difference between the neutral and each of six facial stimuli with basic expressions and created opposite faces with reference to the neutral face [100% (original, e.g. happy), 50%, 0% (neutral), -50%, -100% (opposite, e.g. unhappy), for each expression]. Our previous study using semantic differential method (Shibui et al. 2001 Perception 30 53-54) showed the results in line with the two- dimensional continuous model of facial perception. In this study, subjects rated these stimuli using "affect grid". (Russel et al. 1989 Journal of Personality and Social Psychology 57, 493-502) Subjects put points directly onto the dimensions constituted by 'pleasantness' and 'activity'. The configuration of each emotional stimulus showed that each stimulus (100%) passed the neutral face and reached the opposite face (-100). This result supports the two-dimensional continuous model of facial perception.