Seeing Things in Pictures

By Neuroskeptic | June 1, 2012 7:22 pm

A team of Japanese neurologists propose a new method to detect visual hallucinations – the Pareidolia Test.

Pareidolia means perceiving things that aren’t there, in random or unrelated stimuli. Uchiyama et al created a set of 25 photos, each of which contains things that kind of look like faces, animals, or other objects… but not really. As you can see, the flowers and the birds look like faces. I can’t work out what the leopard and the trees are meant to be, though…

The authors showed the pictures to some patients with dementia. They had one minute to describe what they saw in each image. Compared to healthy controls, patients with Alzheimer’s disease did not experience any more pareidolia than controls.

But people with Lewy Body dementia – a disorder in which visual hallucinations and misclassifications are more common than in Alzheimer’s – reported seeing numerous faces, people and creatures that weren’t there.

They didn’t just say that the images “looked like” these things: they actually thought they were pictures of the illusionary objects. This is an interesting test which might help doctors to diagnose visual hallucinations, which are often under-reported by patients. Some degree of pareidolia, especially for faces, is entirely normal, however, as the popularity of Jesus’s in snacks shows. When specifically told to expect it, people can even “see” faces in random black and white patterns.

ResearchBlogging.orgUchiyama, M., Nishio, Y., Yokoi, K., Hirayama, K., Imamura, T., Shimomura, T., & Mori, E. (2012). Pareidolias: complex visual illusions in dementia with Lewy bodies Brain DOI: 10.1093/brain/aws126

CATEGORIZED UNDER: mental health, papers
  • Anonymous

    I think the leopard was picked for what might look like two smileys near the indicating arrows. The tree image may have been picked for what might look like large faces behind the trees on the left. Don't know what folks might imagine seeing at the rightmost arrow.

  • Anonymous

    Looked again at the leopard and trees at 1.25 and 1.5 magnification. Don't see anything that could be imagined to be there now.

  • Nick Byrd

    I wish they would have control pictures and not just control subjects. The bottom two pictures, when compared to the top two, are not the same type. The bottom pair ccontain something different: a face-like pattern that is noticeable without any serious study of the picture. How these patients perceive these two types of images differently is also worth teasing out.

  • Anonymous

    Why didn't they point a great big arrow at the very obvious ape face in the tree picture? It's just below the rightmost arrow.

  • petrossa

    Till you pointed it out i didn't see an ape there. Still i have trouble seeing anything there actually. Short visit to a shrink maybe? :)

  • Michael Browning

    This is a really interesting approach to pareidolia:

    the guy who runs the blog gets people to send in tons of photos which have induced pareidolia (and which are really interesting in themselves), then writes a face recognition program and shows the same pictures to the program. In the linked blog he gives some examples of pictures where the software sees the same faces as the humans and others where the software finds alternative faces.

    The author of the blog is particularly interested in the machine perception side of things– although I think it may also be relevant in regards to human perception, particularly the pictures where the program sees other faces. Someone (can't remember who) said that robotics is the engineering of psychology– because if you truly understand a process you can build a robot to emulate it. This project suggests to me that the face identification program may roughly emulate some of the algorithms involved in human face perception but is quite often either employing a different algorithm or a more limited set of algorithms– and it is the faces where different faces are perceived that are likely to tell us what those differences are.

    Although, as an interested reader with absolutely no expertise in this area I have no idea what it is telling us…

  • Anonymous

    I guess they haven't heard of the Rorschach test. . .

  • katiedid

    If you tilt your head to the right you might see a face at the arrow on the right in the tree picture. To me, it looks like a giant smile similar to a freaky clown smile.

    I see one face in the leopard (the bottom, also by tilting right) but not the top or the other trees.

  • katiedid

    Oh! And I see the ape face too!

  • Neuroskeptic

    Michael Browning: Thanks for the link, I'd seen that before and it is indeed fascinating! Everyone should follow that link for machine pareidolia fun.

    katiedid: Oh yeah… I see the ape. Still can't see anything in the trees, though.

  • omg

    I see triangles.

  • Anonymous

    Here, have a look:

    If you stare at either of the last two pictures for too long, you end up with the yellow-marked areas (plus many more) until all you see is faces instead of the intended focus of the picture.

    I blame this on caffeine and anyone telling me otherwise can go jump off a cliff. :)

  • Anonymous

    From someone who is used to seeing things that aren't there…… (I'm a mum, it goes with the territory)

    Leopard: a leopard face, facing the same direction, almost a tiny copy of the leopard head.

    Trees: Shadowy humanoid figures next to each tree.

    Facinating blog – thank you



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar