The neurons in the next layer might respond most intensely to a simple pattern, or density of lines. Through the cortex, the complexity increases, but the mechanisms remain the same: neurons are tuned to certain phenomena, whether that's a line of a certain orientation or a face.1
Much of this has been known for decades.2 But here's where the science has gotten really interesting in the last five or ten years. When neuroscientists look at the connections between the cells, they don't just see information passing up the complexity chain. There is information running down from the neocortex's higher levels to the lower ones.
"For every axon coming from the retina into our thalamus before entering our 'consciousness' in the primary visual cortex, the primary visual cortex sends at least twice as many axons back onto the thalamus to modulate the raw signal," explained UC San Diego neuroscientist Bradley Voytek.3
Why is that significant? "Our cortex is already changing the raw visual information before that information gets into our consciousness," Voytek concluded.
You're not only seeing what is actually before you; you're seeing what your brain is telling you is there.
Specifically, the cortex is sending a cascade of predictions about what should be seen at all the different layers of complexity. So what travels back up from the eyes is not raw visuals of the environment, but how the world deviates from what the brain is expecting.
According to University of Edinburgh philosopher Andy Clark's masterful 2013 summary of the state of cognitive science, this emerging idea about the brain is called the "bidirectional hierarchical network model." It holds that every level of the brain is engaged in making predictions, so the expectation of seeing a house feeds down through the cortex to the eyes, which are then more likely to perceive a sloping roof instead of something else. But if something is amiss with the prediction, that information gets transmitted and the brain tries to find a better organizational paradigm for the visual input. Knowledge feeds perception and back again. There are loops everywhere strengthening and weakening according to how well they seem to reflect exterior reality.
Now perhaps you see why the facepalm trick works: if you know to look for a hand, there it is.
University of Sheffield cognitive scientist Tom Stafford likes to show people this picture in talks. He "warns the audience that [he's] about to 'rewire their brains.'"
Then, he says, "It's a frog."
At that moment, the frog jumps out at most people, and they'll find it easy to see the frog from then onward. They cannot unsee it. And though he's joking about the brain wiring stuff—playing off the media trope/tripe that this or that thing rewires the brain—he's not wrong. Somewhere in your neocortex, the predictive model that knows what a frog looks like is influencing the cascade of neuronal activity, turning circles into eyes and an arc into a mouth. If you still can't see him, here's the original grayscale:
So, when Holly Brockwell tells us that she can't unsee the facepalm in the World Cup logo, she is, quite literally, rewiring your brain.
As is Richard Shearwood here:
And this one, oh god, this one is a great example of someone directing your conscious attention to a previously unnoticed alternative image interpretation.
CANNOT UNSEE OMG pic.twitter.com/aQMfW2tqc3— bye (@bootytbh) March 16, 2014
These silly images tell us something significant about the way we are. Shakespeare was onto something with his email-signature-famous line, delivered by Hamlet, "for there is nothing either good or bad, but thinking makes it so."
Cognitive scientists have other ways of putting it. Here are a couple: "Sensory stimulation might be the minor task of the cortex, whereas its major task is to ... predict upcoming stimulation as precisely as possible," write Lars Muckli and colleagues at the University of Glasgow.
Or Karsten Rauss at the University of Tubingen in Germany and collaborators: "Neural signals are related less to a stimulus per se than to its congruence with internal goals and predictions, calculated on the basis of previous input to the system."
To paraphrase all of them: It is not that the real world doesn't exist, but more that we experience it as a hybrid reality: our top-down categories and imagination of the world and our bottom-up sensory experience of the world blend seamlessly into the experience of walking outside into the sunshine or seeing a bird on a wire or eating an oyster or seeing Jesus in a tortilla.
"All this makes the line between perception and cognition fuzzy, perhaps even vanishing," is the conclusion Clark the philosopher draws.
So, the next time someone tweets an image they can't unsee, know that your brain will never be the same.
I really can't unsee pic.twitter.com/1TLPctWbAj— shai (@shaileebd) January 3, 2014
2. Thanks to neuroscientist and machine learning expert Beau Cronin for his guidance through the literature.
3. Voytek published a fascinating paper on the "hypothetical neuroscience" of the residents of China Mieville's novel The City and The City, in which two towns co-exist ignoring each other's existence.