Humans are masters of seeing faces in any old thing - a handbag, TV static, toasted white bread. Scientists want to know why.
A few years ago, as the category 5 Hurricane Milton bore down on the Florida coast, the internet noticed something strange.
Doesn't the satellite image of the storm look a bit like an angry bald guy?
Or a skull?
Seeing faces in everyday objects is so common scientists have a word for it - "face pareidolia" - but why we do it is a mystery.
"People see all sorts of things," says researcher from the UNSW School of Psychology Dr Lindsay Peterson, author of a recent study examining the phenomenon published in Royal Society Open Science .
Dr Peterson's study drew on two experiments involving around 70 participants, who were asked to identify faces and assign traits such as age, gender and emotion across both object images and abstract "visual noise".
In one experiment, respondents were shown a picture of a handbag in whose zip, folds, and buckles they consistently saw a young and happy smiling face, and a picture of "visual noise".
In the noise, interpretations quickly diverged.
"Buddha, angels, demons, dragons. It's amazing you can have these quite rich responses to a stimulus that is essentially noise," she says.
"It is quite remarkable [what we see] given that in the noise stimulus, it is just noise. There really isn't anything there."
In experiment two, the researchers introduced vertical symmetry - a subtle structural cue that loosely mirrors the layout of a human face.
Even that small amount of order shifted perception, with participants more likely to see angry faces in these images.
While there was more diversity in the noise images, patterns did begin to emerge.
Even without any real features to guide them, participants repeatedly converged on similar interpretations, including a strong bias towards perceiving male faces and, in many cases, expressions of anger.
"For whatever reason it seems like we've got this template stored in our brains about what a basic human face is, and it resembles a male face," says Dr Peterson.
There are a number of theories as to why this might be, including social and cultural biases.
But Dr Peterson's isn't the first study to pick up on this.
"The male bias exists across generations and in children as young as four years old, which suggests that it's hard wired."
And the fact that more than a fifth of respondents saw the perceived face as angry is intriguing to Dr Peterson.
She suggests that there could be an evolutionary bias to assuming something is a threat before you've had a chance to evaluate it properly.
"Your lizard brain is telling you that the safest thing is to assume it's a threat and then deal with it."
The study gives us a window into how our brains parse ambiguity: face first, ask questions later.
Why we do it is a bit harder to determine, but this study and multiple others over years underscore the fact that the brain is pre-determined to recognise faces even when none are there, suggesting face recognition was an extremely important skill for our ancestors and remains so today.
The team now plans to explore how these biases shift under different conditions, including whether factors like spatial detail or motion change the kinds of faces people think they see.
