Pareidolia: Why We See Faces in Almost Everything
The world is full of faces. Faces in wall outlets. Faces in lamp switches. Faces in cheese graters. Sometimes these faces have religious significance, like the women who found an image of the Virgin Mary in her grilled cheese (and then sold it for $28,000), or the Cheeto that looks convincingly like Jesus.
The phenomenon of seeing faces where they’re not supposed to be — in clouds, on buildings, in tacos — is so common and widespread that it has a name: pareidolia. In Greek, pareidolia translates as “beyond form or image,” and it means finding meanings or patterns where there aren’t any, like hearing a heartbeat in white noise or believing that a seat cushion is mad at you.
It’s easy to dismiss pareidolia as a fun optical illusion, or worse, as a psychotic delusion. But some scientists now believe that our uncanny ability to find faces in everyday objects points to a new understanding of how our brains process the outside world. Instead of taking in visual cues and then making sense of them as an apple, a tree or a face, it might be the other way around. What if our brains are actually telling our eyes what to see?
We Are Programmed to See Faces
Kang Lee is a professor of applied psychology and human development at the University of Toronto. In addition to giving a popular TED Talk on how to tell if kids are lying, Lee has spent decades studying how infants, children and adults process faces.
“As soon as we’re born, we start to look for faces,” says Lee, explaining that it’s a product of millions of years of evolution. “One reason is that our ancestors needed to avoid predators or find prey, all of which have faces. And a second reason is that humans are very social animals. When we interact with each other, we need to know if the other person is a friend or foe.”
Evolution might explain pareidolia, too. Since the ability to quickly recognize and respond to different faces could be a matter of life and death, there’s a much higher cost for not seeing the lion’s face in the underbrush than for mistaking an orange-and-black flower for a lion’s face. The brain is better off making a “false positive” (pareidolia) if it means that you’re primed to recognize real danger, too.
Which Comes First, the Eyes or the Brain?
It’s clear that evolution has programmed our brains to prioritize faces, but how exactly does it all work under the hood? That’s what Lee wanted to find out.
The conventional wisdom is that the eyes take in visual stimuli from the outside world — light, colors, shapes, movement — and send that information to the visual cortex located in a region of the brain known as the occipital lobe. After the occipital lobe translates the raw data into images, those images are sent to the frontal lobe, which does the high-level processing. Is that a rock outcropping or is it a giant head?
That conventional model is what Lee calls “bottom-up” processing, in which the brain’s role is to passively take in information and make sense of it. If the brain sees faces everywhere, it’s because the brain is responding to facelike stimuli — basically any cluster of spots and spaces that roughly look like two eyes, a nose and a mouth.
But Kang and other researchers began to question the bottom-up processing model. They wondered if it wasn’t the other way around; a “top-down” process in which the brain is calling the shots.
“We wanted to know whether the frontal lobe actually plays a very important role in helping us to see faces,” says Lee. “Instead of the face imagery coming from the outside, the brain generates some kind of expectation from the frontal lobe, then goes back to the occipital lobe and finally to our eyes and then we see faces.”
‘Seeing Jesus in Toast’
That question is what made Lee think about pareidolia. He’d read those stories of people seeing images of Jesus, Elvis and angels in their toast and tortillas, and wondered if he could build an experiment around it.
So, Lee recruited a bunch of regular people, hooked them up to an fMRI scanner, and showed them a series of grainy images, some of which contained hidden faces and some of which were pure noise. The participants were told that exactly half of the images contained a face (not true) and were asked with each new image, “Do you see a face?” As a result of this prodding, participants reported seeing a face 34 percent of the time when there was nothing but static.
What was most interesting to Lee were the images coming back from the real-time fMRI scan. When participants reported seeing a face, the “face area” of their visual cortex lit up, even when there was no face in the image at all. That told Lee that another part of the brain must be telling the visual cortex to see a face.
In a paper provocatively titled, “Seeing Jesus in toast: Neural and behavioral correlates of face pareidolia,” Lee and his colleagues reported that when the brain was properly “primed” to see faces, then the expectation to see a face was coming from the frontal lobe, specifically an area called the inferior frontal gyrus.
“The inferior frontal gyrus is a very interesting area,” says Lee. “It’s related to generating some kind of idea and then instructing our visual cortex to see things. If the idea is a face, then it would see a face. If the idea is Jesus, I’m pretty sure the cortex is going to see Jesus. If the idea is Elvis, then it’s going to see Elvis.”
The “Jesus in toast” paper won Lee a 2014 Ig Nobel prize, a cheeky award handed out by the humorous science magazine Annals of Improbable Research, but Lee says the pareidolia experiment proved that top-down processing plays a critical role in how we experience the world around us.
“A lot of things we see in the world aren’t coming from our sight, but are coming from inside our minds,” says Lee.
Friend or Foe?
Lee has also run research on babies and racial bias. He found the very youngest babies were able to recognize differences between faces of all races but lost that ability as they grew older. By 9 months, they could only differentiate between faces that were their same race. The rest started to blur together. The reason is that they had only been exposed to same-race faces (in most cases, mom and dad) for the first nine months of their lives.
From his research, Lee now believes racial biases aren’t biological; we simply learn to trust people that look like the faces we saw when our brains were first developing. Unfortunately, this can develop later into different kinds of biases based on societal messaging and stereotypes.
“The reason there are racial biases is because of early experiences,” says Lee. “If we created a diverse visual and social experience for children, then they would be less likely to have biases.”
The good news is that parents and educators can combat racial bias by exposing infants and toddlers to faces of all races and identifying them as “Jill” or “Derek,” not as a “white person” or a “Black person.”