Justus Kerner, a German physician and poet, is known for dripping ink onto folded paper and interpreting the faces that emerged from the smudges. He published drawings in a book called Klecksographien, becoming one of the first people to systematically document something every human brain does without permission: hunt for faces in places where no faces exist. The pareidolic monkeys peering out of his inkblots were precursors to Rorschach’s test, but the underlying machinery — a knuckle of cortex behind your right ear called the fusiform face area — wouldn’t be named until much later.
That region was identified by Nancy Kanwisher at MIT, who used fMRI scanning to observe a specific patch of the fusiform gyrus activating when volunteers viewed faces on screen. Houses didn’t trigger it. Hands didn’t trigger it. Faces did. She called it the fusiform face area, or FFA, and the name stuck.
The strange part is what happens when there is no face at all.
The wall socket that watches you back
Show someone a photograph of a three-hole electrical outlet — two prongs above, a grounding slot below — and the FFA fires almost as if a real human face were on the screen. Same with the front of a Volkswagen Beetle. Same with the burn pattern on a piece of toast. Same with the dark crater shadows on the Moon. The brain isn’t being fooled, exactly. It’s being efficient.
Pareidolia is described as the cost of a detector tuned for false positives. A predator hiding in the leaves is expensive to miss and cheap to over-call, so evolution biased the face system toward generous interpretation. Better to flinch at a shadow that turned out to be nothing than to walk past a leopard.
A pattern wired in before birth
Newborns, only minutes old, will turn their heads toward a schematic of two dots above a curved line more readily than toward the same shapes scrambled. The preference is present before the infant has ever seen a face at all, which suggests the template is hardwired rather than learned. By adulthood, the FFA processes a familiar face in a fraction of a second — faster than conscious thought.
That speed is part of why pareidolia feels involuntary. By the time you’ve registered that the cabinet handles and drawer pull make a startled expression, the face has already been built. You can’t un-see it. You can only notice that you’ve seen it.
The Kanwisher experiment, and the argument that followed
Kanwisher’s work triggered a long debate that still hasn’t fully settled. Some neuroscientists argued the FFA wasn’t really a face module at all — it was an expertise module, lighting up for any category a person has trained on intensely. Bird experts showed FFA activation for birds. Car enthusiasts showed it for cars. Chess masters, for board positions.
More recent work using sharper imaging suggests both camps were partly right. The FFA does specialize in faces, but it sits inside a wider network — the occipital face area, the superior temporal sulcus, the amygdala — that handles different slices of the problem: geometry, gaze direction, emotional read. Pareidolia appears to activate that whole circuit, not just the FFA. A 2025 study summarised in Science Daily found that the same regions that respond to real faces also engage when people perceive faces in inanimate objects, with measurable overlap in timing and intensity.
The brain, in other words, doesn’t have a separate filing cabinet for real faces and face-shaped things. It has one face system that occasionally gets enthusiastic.
Why some people see more faces than others
Pareidolia is universal but not uniform. Some people report seeing faces in objects several times a day, while others can stare at the same socket and see only a socket. Researchers studying face perception have linked individual differences to a mix of factors — neural connectivity, attention style, even mood. Anxious states tend to push the detector toward more frequent face-spotting, particularly faces that look threatening or angry.
Women, on average, report more face pareidolia than men in lab studies, and they’re also faster to detect emotional expressions in the illusory faces they see. The reasons aren’t fully understood, but they map onto broader patterns in personality and perception research showing women score higher on average for sensitivity to social cues.
People with Parkinson’s disease, especially those who experience visual hallucinations, often score dramatically higher on pareidolia tests. So do people with certain forms of schizophrenia. The phenomenon has even been proposed as an early screening tool, since the over-active face-detection appears years before more obvious symptoms.
The Moon, the Shroud, the Martian mountain
Pareidolia explains a long list of cultural artifacts that resist debunking. The Man in the Moon, visible to anyone with clear skies, is a pattern of dark basaltic plains called maria that a face-hungry brain assembles into eyes and a mouth. The Face on Mars, photographed by Viking 1 in 1976, was a mile-wide mesa in the Cydonia region whose shadows happened to fall in a humanoid configuration. When the Mars Global Surveyor re-photographed it in 2001 at higher resolution, the face was gone — replaced by an eroded hill. The face had only existed in the angle of the light and the pattern-completion machinery of the people looking at it.
Religious pareidolia is its own subgenre. A grilled cheese sandwich said to bear the image of the Virgin Mary sold on eBay in 2004 for $28,000. The Shroud of Turin, the burn marks on Tokyo windows, the smoke patterns of the 9/11 collapse — all of them are, at the neural level, the FFA and its neighbors doing exactly the job they evolved to do, just with insufficient signal.
Animals do it too
Rhesus macaques, when shown photographs of objects with face-like configurations, fixate on the illusory eyes the same way they fixate on real eyes. Dogs orient toward face-like patterns. Even chickens, in experiments, pecked more readily at food arrangements that resembled a face than at scrambled controls.
The trait’s deep evolutionary roots suggest pareidolia isn’t a quirk of human culture or a side effect of staring at smartphones. It’s a feature of how vertebrate brains have organized themselves for hundreds of millions of years — a bet that mistaking a leaf-shadow for a predator costs less than the alternative.
What it feels like from the inside
Part of what makes pareidolia interesting is how it occupies a strange middle ground in conscious experience. You know the coat hanging on the back of the door isn’t a person. You can see, intellectually, that it’s fabric. But the face-shape your visual system has assembled out of the collar and shoulders has already done its work — it has produced a small startle, a brief social register, a flicker of recognition that someone is there.
This split between perceptual experience and rational belief is what some neuroscientists call cognitive impenetrability. The face module doesn’t take advice from the rest of the brain. It reports what it sees and the rest of the system has to sort out what to do with the report.
Recent work on how memory and perception share circuits, including a 2026 study on overlapping brain regions in different types of remembering, suggests the same principle may apply broadly: the brain runs many of its perceptual systems on shared real estate, which is efficient but means errors in one domain can color experience in another.
The pop-up faces of childhood
Children seem especially attuned to face pareidolia, which is part of why so much of toddler culture — toys, picture books, animated kitchenware — is built around anthropomorphized objects. A teapot with eyes. A car with a grille that grins. The brain is doing the same work it does in adults, just with less interference from the rational override that comes later. Building on research on individual differences in perception, the openness of childhood imagination and the looseness of the developing face detector are part of the same system.
That openness has consequences in how children play, how they form attachments to objects, and how easily they animate the inanimate. The stuffed bear with a face is a friend in a way the same bear without a face never could be.
The cloud you’re about to look at
Step outside in the next hour, find a cumulus drifting east, and your fusiform face area will be working before you’ve finished the thought. Sometimes it produces a dog. Sometimes a dragon. Often, statistically, a face — because the face detector is the loudest voice in the room and it has been there, working, since you were a few minutes old.
Kerner died in 1862 thinking his inkblots revealed something about the soul. The neuroscience that followed suggests he was looking in roughly the right place. The faces in the blots, like the faces in the toast and the wall sockets and the lunar maria, are not in the paper. They are in a small patch of cortex doing the job it was built for, slightly over-eagerly, on something that isn’t a face.
