Reverse-engineering a winning smile

Researchers from the University of Minnesota have figured out the features of a winning smile.

• It's a balance of the mouth's angle, smile width, how many teeth are visible and symmetry. Best bet: showing just a little bit of teeth.
• Smiles that form slightly asymmetrically were seen as more "genuine" and "pleasant" — to a point. If it happens relatively slowly (longer than 125 milliseconds), the crooked smile just starts to get "creepy."

Why the findings matter:

• They guide surgeries to reanimate someone's face after a stroke, says author Sofia Lyford-Pike, a facial plastic and reconstructive surgeon at the University of Minnesota. (People who are unable to smile effectively are at increased risk for depression.)
• It could inform computer graphics.
• More broadly, the work may help to understand how individual differences in facial expressions can lead to differences in how people perceive threats, trustworthiness and other socially relevant cues, something researchers have long studied.

What they did: The researchers created 3D computer animated faces with slightly different smiles constructed from combinations of different angles of the mouth, varying extents of the smile, how much teeth were showing and how symmetrically the smile formed on the face. 802 people were then asked to rate 250 millisecond-long video clips of the smiles forming for their effectiveness, how genuine they seemed, whether they were "pleasant" or "creepy," and what emotion they thought the smile conveyed. The best-rated smiles have lips angled between 13 and 17 degrees and extend between 55% and 62% of the distance between the eye's pupils.

Limitation: The study only modeled the mouth. Previous work has found the eyes influence people's perceptions but the study shows the mouth can transmit a lot of information on its own, says Lyford-Pike.

What's next: Determining whether the results translate to spontaneous smiles of different, real people, is a key question, says Aleix Martinez, a cognitive scientist and professor of electrical and computer engineering at Ohio State University. Martinez asks: "Can you predict a person's success (personal, profession) on the "niceness" of their smiles?"