Particularly Exciting Experiments in Psychology
September 28, 2017
Facial Expression Recognition
Faces are processed differently than other stimuli, perhaps due to the challenges in discriminating between highly similar stimuli to recognize individuals, and the social necessity of doing so successfully. For instance, your husband Bob and his brother Bill may both have two eyes, a nose, and a mouth, but it would be quite awkward if you got them confused.
In addition to recognizing face identity, successful communicative interactions also depend on the ability to recognize and interpret changes in a social partner's facial expressions.
Calcutt and colleagues (2017, Journal of Comparative Psychology) tested facial expression discrimination ability in tufted capuchin monkeys, a species that is able to recognize individual faces and that, as a highly social species, relies on a range of facial expressions for communication.
On each trial, four conspecific faces were shown. Three faces depicted a neutral facial expression, and the fourth depicted either a scalp lift affiliative or open mouth threat facial expression.
Capuchins were highly accurate at selecting the emotional image as the odd image, although performance was better for threat expressions (97%) vs. affiliative expressions (81%). This difference may be due to greater salience of exposed teeth in the threat expression, or the fact that recognizing threat is more critical because incorrect identification can have greater consequences. When faces were inverted, performance discriminating emotional from neutral faces was still above chance, but was greatly reduced relative to upright faces (only 35% accuracy).
This suggests that, like in humans, expression discrimination in capuchin monkeys may rely on processing configural relations between individual face features.
Calcutt and colleagues found that tufted capuchins can discriminate facial expressions, and may rely on similar processes as humans to do so. Palermo and colleagues (2018, Journal of Experimental Psychology: Human Perception and Performance) examined perceptual underpinnings of individual differences in this ability in humans.
Specifically, they were interested in adaptive coding, which reflects calibration of perception to optimize sensitivity to novel input. On each trial participants saw an "antiexpression" adaptor face, and then had to judge the expression of a subsequent neutral face as happy, angry, sad, or scared.
The extent to which the antiexpression biased perception of the subsequent face (e.g., anti-happy face causing subsequent neutral face to appear happy) was correlated with three measures of facial expression recognition ability, and a task-independent index of facial expression recognition extracted from principal component analysis; stronger adaptive coding of expression was associated with better facial expression recognition.
The authors suggest that such adaptation may facilitate recognition of subtle differences in expressions. In addition, anxiety, but not depression or general affect, was negatively associated with facial expression recognition, independently of adaptive coding. These results suggest that facial expression recognition ability is influenced by both perceptual and affective factors.
- Calcutt, S. E., Rubin, T. L., Pokorny, J. J., & de Waal, F. B. M. (2017). Discrimination of emotional facial expressions by tufted capuchin monkeys (Sapajus apella). Journal of Comparative Psychology, 131(1), 40–49. http://dx.doi.org/10.1037/com0000055
- Palermo, R., Jeffery, L., Lewandowsky, J., Fiorentini, C., Irons, J. L., Dawel, A., Burton, N., McKone, E., & Rhodes, G. (2018). Adaptive face coding contributes to individual differences in facial expression recognition independently of affective factors. Journal of Experimental Psychology: Human Perception and Performance, 44(4), 503–517. http://dx.doi.org/10.1037/xhp0000463
Particularly Exciting Experiments in Psychology™ (PeePs) is a free summary of ongoing research trends common to six APA journals that focus on experimental psychology.