Whole or skim milk? Take your normal route home or an unfamiliar potential shortcut? Flip a hypothetical trolley switch that would kill one person, saving five others.
All of these decisions have two things in common: uncertainty and anxiety.
Whether you consciously feel anxious or not in deciding between two similar options at the grocery store, I mean anxious in the technical sense. Anxiety is your brain’s response to uncertainty, big or small.
The Fantastic Organ
Neuroscientist Karl Friston has called the brain a “fantastic organ.” The brain is not only an organ that is fantastically complex, but one which is constantly generating fantasies. Your brain is a prediction machine—this is the basis for predictive processing theories of consciousness.
Most of what your brain and body does is unconscious: There is high predictability, and little uncertainty, and attending to these details would be exhausting. We can choose to take conscious control of our breathing, but can you imagine if we had to do this all the time? How tiring, and perhaps even deadly, would it be to exist in a world where our brain couldn’t take over and regulate our breathing unconsciously?
But when there is uncertainty, prediction becomes much harder. There is no unconscious smooth sailing. The whole reason we have conscious thought is to deliberate between actions when multiple decisions are possible. Our fantastic organ must generate the right fantasy map to navigate us safely to our goals.
Entropy Equals Uncertainty
From a computational perspective, uncertainty increases entropy. You may recognize this term from physics. It is a law of nature that entropy always increases. Entropy is chaos and disorder. Heat dissipates, the universe expands, and order does not stay ordered. Sandcastles collapse every day, but there is no universe in which sand spontaneously arranges itself into castles.
The caveat, of course, is that it is possible to build sandcastles and create order in a disordered universe. We cannot create order from anything; it comes at the cost of energy. Expending energy into work still increases entropy, globally, but we can trade energy from the outside world to maintain order in a narrow context.
This anti-entropic process of consuming energy to maintain order is the whole basis of life. Homeostasis is an organism’s process of expending energy to maintain itself, whether that energy is gathered from photosynthesis, or from eating other organisms.
And at a much higher level, your brain, the prediction machine, is fundamentally in pursuit of minimizing entropy. What does a prediction machine want more than anything else? To be right. What does it need to be right? Certainty. Why does it want certainty? There are two answers to this. From a low-level computational perspective, the less uncertainty there is in a problem, the more straightforward the answer. Your brain saves energy. From a high-level evolutionary perspective, the more uncertain your environment is, the less likely you are to survive. We need certainty.
Of course, if all we wanted was a certainty and to minimize brain power, we would be in a constant state of hibernation. This is not a long-term adaptive solution. The best types of prediction machines are those which can handle high amounts of uncertainty, and still come out on top.
Three Types of Brain
The struggle to navigate uncertainty is the best way to understand the evolution of our large brains and to understand the neuropsychology of anxiety.
The brain can broadly be divided into three layers:
- First, there is the brainstem, which governs basic survival processes. This includes largely unconscious processes like breathing, heartbeat, digestion, and reflexive movement. This most ancient brain system is common to all vertebrates and is sometimes referred to as the “reptilian” or “lizard” brain.
- Next, there is the limbic system. The limbic system controls our emotions, ranging from pain and pleasure, love and fear, and hunger and sex drives. Compared to the basic reptilian brain systems shared across all vertebrates, the limbic system is more developed in social species. Whether predator or prey, parent or child, if your survival depends on receiving care from or avoiding harm from others, it pays to have emotions.
- Finally, there is the cortex or the “rational brain.” (Cortex has the same Latin root as corona, meaning crown; it sits on top of the rest of the brain.) The more intelligent an animal is, the larger its brain (relative to its body), and the more cortex it has. We see the largest cortical brain areas (again, relative to body size) in highly social birds and primates, with humans at the forefront.
Looking at this in the context of brains as prediction machines, this makes perfect sense. Think about the primitive lizard brain, or better yet, even more, primitive fish vertebrates. Yes, they need to survive and reproduce like the rest of us, but these simple organisms may lay dozens, hundreds, or even thousands of eggs at a time. Their prediction machines are straight and narrow; they live largely by reflex. When prediction fails, they die. But what they lack in cognitive flexibility, they make up for in numbers.
More complex social mammals, like rats, have limbic brain systems (and some cortex) much more similar to us humans. They truly experience anxiety in the face of threat and uncertainty. Their tiny but fantastic organs do have a vested interest in generating fantasies of safety, security, nourishment, and social bonding. We know from decades of animal research that rodents experience anxiety in uncertain environments and that their anxiety response is very similar to ours. We release the same stress hormone, cortisol, and the same anti-anxiety medications that humans are treated with work on rats. (That is, after all, where they are first tested to make sure they are safe and effective on mammals.)
Anxiety Is Prediction Error
The fact that all ties it together, posed by Friston’s model of predictive processing in the brain, is that anxiety is the felt experience of entropy. Your brain is constantly making predictions about the world, and judging those predictions against what actually happens. The greater the prediction error, the greater the entropy. Prediction error can be as simple as a single neuron firing or misfiring at the wrong time, or as dramatic as an entire belief system falling apart.
This is where humans’ uniquely large prefrontal cortex comes into play for our unique relationship with anxiety. Even for an animal with a sophisticated limbic system, such as a rat, uncertainty is marked by the direct presence or absence of a threat. Even for animals with highly developed frontal lobes, such as chimpanzees, uncertainty is marked by uncertainty about the attitudes of others. Where am I in the dominance hierarchy? Is this chimp friend or foe? If I share my food, will the favor be repaid? All of the uncertainty an animal has to keep track of is magnified when living in a social environment. There is uncertainty not only about its own thoughts, feelings, and security but about everyone else’s as well.
The most socially complex species have had to evolve the most advanced prediction machines in order to keep up with uncertainty. This is the case for humans as well as primates. But where humans excel—and this is arguably our greatest strength and greatest weakness, when thinking about anxiety—is metacognition, or the ability to think about our own thoughts, and to think in terms of abstract symbols.
What other species can have an anxiety attack brought on by existential dread—just thinking about the meaning of life, or what happens after death? Our ancient brain regions respond physiologically to uncertainty with anxiety, just as they evolved to. But we are no longer only dealing with uncertainty about an immediate threat. Uncertainty can mean losing your job, or hearing bad news about stock market futures, or flunking a test. It doesn’t matter; our fantastic brain, the prediction machine, still produces the physiological response of anxiety that is meant to protect us from immediate physical threat: Cortisol levels rise. Your heart rate increases. Your pupils dilate. You begin to sweat. All of these are genuinely adaptive responses when faced with uncertainty, in an evolutionary landscape in which uncertainty means to fight or flight.
The problem is that this system is too good at its job. The theory of psychological entropy states that uncertainty is always felt as anxiety, no matter the cause. It does not matter if you are anxious because of a prediction error caused by an immediate threat when you were expecting safety, or because you have begun questioning a worldview that used to give you a sense of security in the world. It does not matter if you are uncertain about something trivial, like whether to buy whole or skim milk. Your brain evolved to deal with uncertainty as a threat, and anxiety is the natural response to that.
Where does this leave us? Are we doomed to become anxious over everything we can’t predict with absolute certainty—which is, literally, everything—like the neurotic philosopher Chidi in The Good Place? Not quite.
Living Out the Fantasy
The optimistic part about the theory of psychological entropy is that it tells us where all emotions are rooted, positive or negative. Fundamentally, it all has to do with prediction error. We inherently experience greater prediction error, greater uncertainty, and greater entropy, as negative. But on the flip side, we inherently experience reduced prediction error as positive. This is why (as mediated by dopamine) it feels so good when you achieve a goal.
This is also why practicing mindfulness and routine improves well-being: It minimizes prediction error, directly or indirectly. Gratitude and humility help rein in your fantasies, leaving less room for disappointment. And journaling and self-reflection help you more clearly map out your thoughts and worldview, leaving less room for error. And healthy habits leave less room for uncertainty. For the fantastic organ, reducing psychological entropy is the key to well-being.
Friston, K. (2010). The free-energy principle: a unified brain theory? Nature Reviews Neuroscience, 11(2), 127-138.
Gray, J. A., & McNaughton, N. (2000). The Neuropsychology of Anxiety: An Enquiry into the Functions of the Septo-Hippocampal System, 2nd Edn Oxford: Oxford University Press.
Hirsh, J. B., Mar, R. A., & Peterson, J. B. (2012). Psychological entropy: a framework for understanding uncertainty-related anxiety. Psychological review, 119(2), 304–320. https://doi.org/10.1037/a0026767
Omary, A. (Host). (2023, April 24). Dr. Karl Friston – Active Inference & Free Energy (No. 99) [Audio podcast episode]. In The Nature & Nurture Podcast. https://www.youtube.com/watch?v=aWAvZWGolcg&authuser=0
Parr, T., Pezzulo, G., & Friston, K. J. (2022). Active Inference: The Free Energy Principle in Mind. Brain, and Behavior. Cambridge, MA: MIT Press.
Sapolsky, R. M. (2017). Behave: The biology of humans at our best and worst. Penguin.