Our ability to imitate facial expressions depends on learning and visual feedback, say psychologists. Marketing people knew that already. The 'chameleon effect' is commonly used in interpersonal negotiations because imitating another person's postures and expressions is an important social lubricant.

How do we learn to imitate with accuracy when we can't see our own facial expressions and we can't feel the facial expressions of others?

In an experiment, the psychologists videotaped participants as they recited jokes and then asked them to imitate four randomly selected facial expressions from their videos. When they achieved what they perceived to be the target expression, the participants recorded the attempt with the click of a computer mouse. A computer program evaluated the accuracy of participants' imitation attempts against a map of the target expression. In contrast to previous studies that relied on subjective assessments, this new technology allowed for automated and objective measurement of imitative accuracy.

They found that participants who were able to see their imitation attempts through visual feedback improved over successive attempts. But participants who had to rely solely on proprioception – sensing the relative position of their facial features – got progressively worse. These results are consistent with the associative sequence-learning model, which holds that our ability to imitate accurately depends on learned associations between what we see (in the mirror or through feedback from others) and what we feel.

They conclude that contingent visual feedback may be a useful component of rehabilitation and skill-training programs that are designed to improve individuals' ability to imitate facial gestures.

Published in Psychological Science.