Berkeley, California - Donald Trump, Ted Cruz, Bernie Sanders and Hillary Clinton may want to take note.

If a quick scan of the crowd tells you your audience looks riveted, bored or angry, trust your eyes. New research from UC Berkeley shows that we average out expressions in a sea of faces to get a reliable mood reading.

The study, published in the Journal of Vision, shows the human brain is wired to speedily pick up on the overall emotional tenor of a crowd, registering everything from alarm over impending danger to fascination, humor and anger as well as mixed emotions.

“Our results suggest that we can recognize how united or divided the crowd is, which is useful for guiding our actions,” said study senior author David Whitney, a psychology professor at UC Berkeley.

For example, he said, “Politicians should be able to gauge how polarizing their speech is, and use that to help modify their talking points.”

In a series of experiments, study participants looked fleetingly at pictures of crowds. In some images, all the faces in the crowd showed similar emotions, say, delight. In others, they showed a variety of emotions ranging from happiness to anger.

Study volunteers quickly differentiated between emotional uniformity and emotional diversity in the pictures. The experiments were controlled for such variations as the angle or brightness of faces.

The study was conducted at UC Berkeley. The lead author is Jason Haberman, an assistant professor of psychology at Rhodes College in Tennessee and a former student of Whitney. Pegan Lee, a histotechnologist at UCSF and former Whitney lab member, is a co-author.