Judgments of subtle facial expressions of emotion.

Abstract

Most studies on judgments of facial expressions of emotion have primarily utilized prototypical, high-intensity expressions. This paper examines judgments of subtle facial expressions of emotion, including not only low-intensity versions of full-face prototypes but also variants of those prototypes. A dynamic paradigm was used in which observers were shown a neutral expression followed by the target expression to judge, and then the neutral expression again, allowing for a simulation of the emergence of the expression from and then return to a baseline. We also examined how signal and intensity clarities of the expressions (explained more fully in the Introduction) were associated with judgment agreement levels. Low-intensity, full-face prototypical expressions of emotion were judged as the intended emotion at rates significantly greater than chance. A number of the proposed variants were also judged as the intended emotions. Both signal and intensity clarities were individually associated with agreement rates; when their interrelationships were taken into account, signal clarity independently predicted agreement rates but intensity clarity did not. The presence or absence of specific muscles appeared to be more important to agreement rates than their intensity levels, with the exception of the intensity of zygomatic major, which was positively correlated with agreement rates for judgments of joy.

Validation of the Amsterdam Dynamic Facial Expression Set – Bath Intensity Variations (ADFES-BIV): A Set of Videos Expressing Low, Intermediate, and High Intensity Emotions

“…However, the forced simultaneous changes of facial features that come with morphing pose a limitation for application in facial emotion recognition experiments. The naturalness of computer-generated expressions is questionable, as it is unclear whether the created movements are anatomically feasible [ 36 , 43 ]. This concerns the onsets of single facial action units, which can vary [ 44 ], and the speed of those action units in reaching apex, which varies between emotions (Hara and Kobayashi as cited by [ 45 ]).…”

Laughter and Smiling in 16 Positive Emotions

“…Moreover, [96] found an effect of age on smiling dynamics. Evidence is mixed on whether the AU6 can serve as a "reliable" muscle which is an inevitable sign of enjoyment as compared to posed smiles (AU6 also occurs in posed DDs [11] [17]), but it was shown that the action of the AU6 plays an important role in the perception of spontaneity, authenticity, genuineness of smiles and judgment accuracy [30] [31] [32] [33]. 2 [15] [19] [25] [34] defined the basis of joyful laughter (Duchenne laughter) to consist of a DD plus an audible, laughter-related vocalization and open mouth.…”

Naming emotions in motion: Alexithymic traits impact the perception of implied motion in facial displays of affect.

“…In our everyday interactions with others the facial displays of emotion that we encounter are dynamic and most people readily recognise a wide range of affective states, even from very slight and nuanced muscular contractions of the face (Matsumoto & Hwang, 2014). Despite this, much of the research into facial emotion recognition has relied on static photographs of posed expressions.…”