"When You're Smiling": How Posed Facial Expressions Affect Visual Recognition of Emotions

Abstract

Facial imitation occurs automatically during the perception of an emotional facial expression, and preventing it may interfere with the accuracy of emotion recognition. In the present fMRI study, we evaluated the effect of posing a facial expression on the recognition of ambiguous facial expressions. Since facial activity is affected by various factors, such as empathic aptitudes, the Interpersonal Reactivity Index (IRI) questionnaire was administered and scores were correlated with brain activity. Twenty-six healthy female subjects took part in the experiment. The volunteers were asked to pose a facial expression (happy, disgusted, neutral), then to watch an ambiguous emotional face, finally to indicate whether the emotion perceived was happiness or disgust. As stimuli, blends of happy and disgusted faces were used. Behavioral results showed that posing an emotional face increased the percentage of congruence with the perceived emotion. When participants posed a facial expression and perceived a non-congruent emotion, a neural network comprising bilateral anterior insula was activated. Brain activity was also correlated with empathic traits, particularly with empathic concern, fantasy and personal distress. Our findings support the idea that facial mimicry plays a crucial role in identifying emotions, and that empathic emotional abilities can modulate the brain circuits involved in this process.

Keywords: emotion recognition; emotions; empathy; fMRI; facial expressions.

Figure 1

Experimental protocol for the fMRI session.

Figure 2

Behavioral results. Left: % of disgust responses for participants posing disgust/neutral/happiness expression. Right: % of happiness responses for participants in happiness/neutral/disgust expression. * = p < 0.005, bars = standard errors.

Figure 3

Incongruent pose vs. neutral pose one-sample t-test; p < 0.05, FWE corrected. Functional results are shown on the SPM12 template; color bars represent T-values.

Figure 4

EC correlation for the contrast “posing happiness and perceiving happiness” vs. “posing happiness and perceiving disgust”; cluster size threshold k > 30, corrected at α < 0.05. Same overlay procedure as in Figure 3.

Figure 5

FS correlation with the contrast “posing disgust and perceiving happiness” vs. “posing and perceiving happiness”; cluster size threshold k > 31, corrected at α < 0.05. Same overlay procedure as in Figure 3.

Figure 6

PD correlation with the contrast “perceived happiness” vs. “perceived disgust”; cluster size threshold k > 31, corrected at α < 0.05. Same overlay procedure as in Figure 3.

Figure 7

PD correlation with the contrast “posing neutral and perceiving happiness” vs. “posing neutral and perceiving disgust”; cluster size threshold k > 27, corrected at α < 0.05. Same overlay procedure as in Figure 3.