This came to my attention today: a house that looks like Hitler. And, weirdly enough, it does look strikingly like him – even though it bears none of the characteristics of any human face, let alone the subtle idiosyncrasies that make an individual’s face distinguishable. It’s got a slanted roof and a prominent lintel above the door.
As far as seeing faces in things goes, this is one of the most startling incidences of face pareidolia I can think of, by far. There have only been the fairly dubious images of the Virgin Mary burned onto toast, or the face on Mars (according to Wikipedia, taken by some to be evidence of a long-lost Martian civilisation. Hmm). Pareidolia is the phenomenon of seeing patterns or meaning in random objects or sounds, but – from my experience, anyway – it happens much more easily with seeing faces in things.
How does this happen? Visual images of objects that look like faces are, as you’d expect, processed in the same area of the brain that processes images of real faces – the fusiform face area. A study from 2009 looked at how pareidolia is produced by the brain.
The brain images on the left show where the activation is in response to seeing a face-like object (the top brain), a real face (the middle brain), and a non-face-like object (the bottom brain). The area circled is the fusiform face area (FFA), and you can see quite clearly that it shows roughly the same pattern for real faces and face-like objects – compared with no activity when the subject is looking at a non-face-like object. There’s the evidence.
So how are we able to tell that only one of them actually is a face, if both images are processed by the FFA? Well, they are processed differently. The graph on the right shows the level of activity in that area over time (a period of 0.8 seconds). The x=0 axis is the exact time that the image was shown, and the various lines show the current, which is indicative of the FFA’s response to those images. The pattern here is different – there is more activity when it’s a real face, and the shape of the peak is different. This means that slightly different neural circuits are activated in the FFA, and that’s what underlies the weird perception. More interestingly, the fact that the peak of activity happens so early (165ms is not a long time) means it is an immediate, low-level perceptual process. That’s why even when you know it’s not a face, you can’t help seeing Hitler.
Hadjikhani, N., Kveraga, K., Naik, P., & Ahlfors, S. P. (2009). Early (M170) activation of face-specific cortex by face-like objects. Neuroreport, 20 (4), 403-7.