How We Read a Room

Perceiving the emotion on a face is a crucial skill for navigating our social world. But the facial cues to emotion are subtle — a slight bend of the lip, a wrinkling of the brow, a widening of the eyes. Being able to quickly pick up these subtle facial dynamics and integrate them into a holistic understanding of a person's emotional and mental state is a remarkable feat. Research has identified dedicated brain regions, such as the superior temporal sulcus (STS), that specialize in processing facial dynamics.

But what happens when you encounter not a single face, but a whole crowd of faces?

Navigating a work party

Source: Pixabay

Imagine you've just arrived at a work party. When you open the door, you see a small group of people chatting on a couch on the left. On the right there's a larger group standing in a circle, and there are also a few people in the hallway. Which group do you approach? How do you determine the "mood" of each group, and how your interaction is likely to play out?

It would not be practical to consider the situation "one face at a time" — it would take too long, and probably feel too awkward, to look at each face, one by one, determine its expression, and then make an overall assessment of the group, before moving on to assess the next group.

Perceiving the ensemble

Instead, research suggests that we have mechanisms for ensemble perception that allow us to quickly extract an average impression from a crowd. For example, in studies of gaze processing by Timothy Sweeny and David Whitney (2014), participants were briefly shown an array of four face images for one second. Each of the four faces was looking in a particular direction, and participants were instructed to estimate the average gaze direction of the ensemble before the faces disappeared. Participants were remarkably accurate at reporting the average gaze of the group, much more so than if they had only paid attention to one or two faces. The results suggest that we can quickly, at a glance, integrate facial information from at least four different faces simultaneously.

A more recent study by Elric Elias, Michael Dyer, and Timothy Sweeny (2017) pushed this idea even further. They presented observers with arrays of 12 dynamic faces on a screen, shown for only a second or less. Each of the 12 faces briefly displayed an emotional expression, before all the faces disappeared. Here, participants' task was to report the average expression of the group of faces. Again, people were remarkably accurate at estimating the group's average expression, even though a second is not nearly enough time to look at 12 faces.

This notion of "ensemble perception" is not entirely new. Vision researchers have known for decades that we can perceive averages in sets — the average orientation of lines, the average size of cookies on a sheet, the average color of leaves on a tree. What's new is determining that we can do this type of ensemble perception with something as complex and subtle as faces.

But of course, not all of us are equally adept at perceiving faces. For example, individuals diagnosed with Autism Spectrum Disorder typically show impairments in judging emotional expressions from faces. This impairment is usually assessed in the context of processing a single face, but the problem is likely exacerbated when an individual encounters a crowd.

Thus, our ability to "read the room" — something many of us take for granted — depends critically on a highly developed face perception system that can act upon a crowd.