50. Perceptual Smile Asymmetry is Impacted by Five Smile Parameters

Abstract

PURPOSE: Facial expressions are integral to nonverbal communication. Dynamic smile reconstruction is a goal of care, and the aim of surgery is to create symmetry. This project evaluates how much asymmetry is tolerated in smile perception and quantifies that asymmetry. METHODS: Fifty-six facial photographs were categorized based on increasing smile asymmetry. Emotrics software was used to compute asymmetry in five mouth metrics. A Qualtrics survey consisting of 56 cropped photographs was disseminated using Amazon’s Mechanical Turk (MTurk). Three questions accompanied each photograph: 1) Is the person smiling? 2) What emotion is being expressed? 3) How symmetric is the facial expression? ANOVA, chi-squared, and multivariable linear modeling were performed. RESULTS: A total of 618 survey responses were analyzed. Smile recognition was 100%, 78%, and 33% in the symmetric, mild-moderately asymmetric, and asymmetric categories (p<0.001). Happy was identified in 89% of symmetric smiles, compared to 15% of asymmetric smiles (p<0.001). Disgust was conveyed in 26% of asymmetric smiles (p<0.001). If a smile was recognized, then 65% were happy; if not recognized, then disgust was chosen most frequently (p<0.001). Multivariable linear models using asymmetry in five mouth metrics predicted: smile recognition, happy emotion identification, and symmetry score (p<0.001). A 1% increase in asymmetry of lower lip height and commissure position decreased symmetry score by 32- and 18-points, respectively. CONCLUSIONS: Survey respondents are unable to recognize severely asymmetric smiles as smiles. Severely asymmetric smiles convey disgust rather than positivity. Mild differences in lower lip and commissure positions contribute to asymmetry perception, and models incorporating five smile parameters—nasolabial fold angle, interlabial distance, commissure height, commissure position, and lower lip height—were able to predict smile asymmetry.