How Your Brain Uses Predictions To Construct Your Perceptions and Emotions

There are some astounding discoveries coming out of neuroscience that I project will have far-reaching implications in the world of psychology, but also just every-day human interaction. One of the most interesting is the way our brain produces our experience of reality is actually the opposite of the way we experience reality. Put another way, we feel like “things happen to us” and we “react.” But if that is not how our brains work, what is actually happening?

It turns out, the neurocircuitry of the brain is too slow to successfully execute a stimulus response model. It would not work. Instead, the brain runs a kind of continuous “prediction loop” where it is constantly trying to anticipate what is about to happen and then make continuous, micro adjustments as sensory information becomes available. These “adjustments” are really about responding to prediction error.

Think of it this way. When I was a coach, I taught our players to be “moving on the pitch.” As the pitcher would rear back to throw the ball, we told the players to try and anticipate where the batter might hit it. We wanted them to already be in motion when the batter swung. It is much easier to change direction once you are moving, than to start moving from a dead stop in the direction you need to go. We called this, “getting a jump on it.” The brain does something very similar. It is constantly trying to anticipate what is next, preparing to make the right move, and like a ball player, if the information comes in that tells the brain to do something different, it is more energetically efficient (i.e. easier) to change course, than to start from a dead stop.

What does this mean in practical terms? In practice, this means the brain is continuously constructing our experience of reality from two massive streams of chaotic, confusing, and incomplete data. One set is our senses. Our eyes, ears, skin, smell, and so on. But it also includes the interoceptive network, the giant bundle of signals coming from inside our body. In practical terms, your brain is continuously trying to make sense of what is going on. It spends a lot of its computational bandwidth trying to anticipate what will come next.

Now, this feels completely counterintuitive. The world definitely feels like something happens to us, and we respond. I know that is how it feels to me. However, it turns out there is a gap between what happens to us and what we perceive. This is different from Stephen Covey’s Gap between stimulus and response. This is a gap between sensory data and what I will call, “sense making.” This gap happens at a level outside conscious awareness. I don’t want to call it subconscious because “the subconscious” is a suitcase word. It can hold lots of different meanings, and different people pack different meanings into that word. So let’s use the word preconscious. Your brain is trying to raise the information to the level of conscious awareness as quickly as possible, but it takes time for it to run through all the layers it has to before you can “become aware” of what is happening.

Putting it this way, your brain has information to work on: what just happened. It makes an “educated guess” as to what is most likely to happen next, then it compares its predication against input from the external senses and the interoceptive network. Basically, it compares what is actually happening to what it predicted would happen. The deviations are called “prediction errors,” and the brain uses prediction errors to adjust its next cycle of predictions.

Going back to our baseball analogy, the outfielder starts moving on the pitch, anticipating a line drive in front of him when the batter hits the ball deep to his left. The player anticipated the ball moving in one direction, but the sensory information from his eyes, transmitted through his visual cortex, tells him that the ball is going to be over his head to his right, so he immediately adjusts to the new information. He predicted one thing, but the facts indicated something else (prediction error), so he adjusted.

Here, however, is the catch. The exact same mechanism is used to determine how we feel. The reason that feelings seem to “trigger” us is that we only become consciously aware of them very late in the processing game. Basically, feelings, far from being this built-in, predetermined biological response to stimulus, are the result of how our brain tries to interpret the interoceptive signals coming to it from inside our body. In practice, the model of " we think than we feel" is very likely wrong. It feels this way because we become aware of the story much faster than the emotion, but the process is the same. What this means, and it is really mind-bending, is that our brain constructs our experiences from the dynamic interaction of predicting what will happen and adjusting to what really happens. Both perception and emotion are processed at the same time, but typically, we become aware of our thoughts before we become aware of the feelings giving rise to our belief that our thinking or what happens to us causes our feelings.

We experience our emotions as something that happens to us, but the science says our feelings are the brain’s way of making sense from the interoceptive signals from inside the body.

Okay, I know this is deep, but let’s go back to the ball player. He moves in anticipation of what he believes is mostly likely to happen next, then while he is in motion, his senses give him new information which is different from what he expected, but he adjusts. But how does he feel? Well, in order to execute, his body has to respond to his predictions. Muscles need to move, which means breathing has to provide enough oxygen, the heart has to pump that oxyinated blood, and other things as well need to prepare the body to perform. It was his anticipation of what would happen, and his intention to execute that kicked all of this off, but the channels of communication between the body and the brain are not one way, they are bidirectional, so the body can report back to the brain, “hey this is how we are executing.” This is what it feels like, and the brain then tries to interpret those sensory inputs within the context of its predictions. It is asking what does this mean? Those “feelings” or “emotions” get integrated into, or put another way, they color our perceptions. Our thoughts and intentions are imbued with the signal information from inside the body, which we experience as emotion.

The surprising result of the ball going over the head signals a large prediction error, so the brain kicks more energy so the player can turn faster, and run harder to close the gap and catch the ball. All of that physical action inside the body generates signals back to the brain - the interoceptive networks in the brain interpret these signals as emotion. But what emotion?

And that is the surprising key to this whole thing. Remember when I said the brain makes “educated guesses” about what will happen next? What education? The education of experience. Our mental models, our predictive engine, are shaped by our life experience. Not only cognitively, but emotionally. In essence, we learn what those signals in our body mean, and we assign concept words to those feelings. Those concept words are emotion words.

In short, we all, everyone one of us, are constructing our experience of reality, continuously, dynamically, all the time. And how we interpret the signals that make it to us is a combination of what has happened to us before, as well as what options we believe are afforded to us in the present moment. (These are called affordances.)

So what you might ask?

What good is knowing this? Well, once you begin to understand the way the brain, experience, and emotion really work, you can begin to use that information to work with your brain, (and others) to get better outcomes. This is a new model for understanding the human condition, and one reason I believe it is so powerful is that it has consistently demonstrated not only an explanatory capacity (it explains observed phenomena), it is the first time that I am aware of we have a predictive model for human behavior.

In a rather obscure experiment conducted by Pascal Wallisch, a neuroscientist at New York University, he demonstrated that our “priors” (prior experience) absolutely affect our perceptions below the level of conscious awareness.¹ He was trying to understand the phenomenon behind the controversial internet meme “The Dress.” (Google it.) His experiment was groundbreaking because he used the model of constructed perception (as opposed to stimulus-response) to construct an experiment to test whether or not people’s past experience shaped how they literally saw “reality.” The answer: that is exactly what happens. Our brain uses what we think can happen, informed by what has happened to us in the past, to construct predictions that we use to perceive the world. Another term for perception in this context could be “disambiguation.” The world is noisy and chaotic, yet our brain finds ways to make sense of it, in real time, and move forward with tremendous confidence. Daniel Kahneman wrote about this in his book Thinking, Fast and Slow. We have to sift through an extraordinary number of sensory data points and make sense of it all. Our brain does this not by merely reacting, but by anticipating (projecting) and adjusting (predictive error correction). The really astounding discovery is that we process emotions the same way.

Your brain doesn’t generate emotions, it constructs them, at the exact same time, in pretty much the same way, that is, makes sense of what is happening in the world around us, it tries to make sense of the world inside us. Only we become aware of one before the other. Most of the work of constructing emotions happens preconsciously, which gives rise to the sensation that emotions “just hit us.” We rarely, if ever, see them as they form. However, there are times we can sense the ambiguity. There are times when the emotions are not clear. And that realization opens the door to a new way to think about emotion regulation, and how perhaps we might be able to manage ourselves, and our relationships with others a little more effectively.

What if Everything You Knew About Emotion was Wrong?

Over the past week, I have been reading the book, How Emotions are Made by Lisa Feldman Barrett. And when I say, “read,” I mean “listen to.” However, the reality is that I do a bit of both. I listen, then pause and go to the Kindle version of the book and highlight or mark key concepts.

While listening to Dr. Barrett’s book, I was struck by the amazing similarity between her work and that of Stephen Grossberg. You probably have never heard of Stephen Grossberg because he works tirelessly in a very obscure niche between psychology, neuroscience, and physics.

Grossberg’s goal was to understand how the observed behavior of neurons (neuroscience) can explain the observed behaviors of humans (the discipline of psychology). Very few psychologists are trained in neuroscience, and very few neuroscientists are trained in psychology. Neither of those groups is trained in physics.

What does physics have to do with the brain? One of Isaac Newton’s great contributions (similar to Einstein’s) was not only a theory of how physical reality is structured but the mathematics to model it. We have amazing equations like Force = Mass * Velocity^2. I used to tell my baseball players that a lighter bat moving faster can hit with more force than a heavy bat moving slowly.

This, it turns out, is counterintuitive. Early baseball players played with very large, heavy bats. They believed a big thing (the big bat) hitting a little thing (the ball) would produce the best results. However, controversial home run king Barry Bonds swung the lightest bat he was allowed to swing. Modern professional players focus on swing speed and exit velocity. Newton’s laws of physics shape how they play. Early players’ intuitions were wrong. And this, sadly, turns out to be a case for many of our “intuitive” senses about reality.

A great model can unlock better understanding of the way the world works. Of course, there are limits. Newton’s classic mechanics gave way to quantum mechanics, which challenges our understanding of a great many things.

Einstein’s great discovery of relativity is known by its equation, e = mc^2 . But you typically do not see equations like that in medicine, neuroscience, or psychology. The most common type of mathematics used in the “people” sciences is Bayesian statistics, a formal set of disciplines designed to help scientists make sense of data gathered from experiments. This means most psychologists and neuroscientists tend to form a hypothesis and test it for validity. Physicists build models which have valid predictive capabilities.

Grossberg is perhaps the only person working in the “soft science” fields I know of building such models. He wanted to understand how a group of neurons, wired together, can produce many of the psychological effects researchers gather from their studies.

In essence, using cross-disciplinary skills, he is trying to connect the dots. How do bundles of neurons collected into a three-pound (1.5-kilogram) ball make a human mind?

Grossberg is not well known to most people; however, he was the first to educate other scientists and engineers on his discoveries about how neural networks work. Because he also focused on creating mathematical models to explain how these networks work, engineers could build software simulations of his networks. Thus, Grossberg seeded a generation of work which has led, more or less in a straight line, to our modern artificial intelligence systems, the “large language models” or LLMs behind ChatGPT, Gemini, Claude, and others.

But Grossberg was not trying to build artificial intelligence; he was trying to understand human intelligence.

I have been reading his textbook for the last year a little at a time, and it is dense. What’s more, Grossberg suffers from a lot of self-reference. He mostly refers to his own papers, and he is very rarely cited in other people’s papers. This has the unfortunate side effect of making him look like a bit of a crank, or at the very least an oddball.

Sadly, even science seems to be subject to the laws of human prestige. If we can’t know for certain the truth (because every claim is hard to verify), we resort to good old-fashioned popularity contests. If enough people not only build on your work but say they build on your work and give you credit for doing so, then you must be right.

From my point of view, since Grossberg toils away in the gaps between disciplines, effectively belonging to multiple camps, he belongs to no camps. However, recently, the work of Dr. Lisa Feldman Barrett on understanding human emotion seems to me at least to lend tremendous credence to Grossberg’s work. While she never directly cites him or mentions him, her work is consistent with Grossberg’s assertions and other recent neuroscience discoveries that the brain does not work the way we believe it does.

Just as the early baseball players managed to play the game despite the mistaken belief that a heavier bat was more effective than a lighter one, many scientists and clinicians have been able to work with people using incorrect understandings of the human brain’s wiring.

The first major myth of the brain, one that is near and dear to any determinist’s heart, is the idea that the brain is a stimulus and response engine. B.F. Skinner built his entire career around this idea, that the brain is stimulated and responds. End of story.

Only… the evidence - for quite a while has been mounting that this is not actually how the brain works. It is not fast enough to be only a stimulus and response engine. So it does something even more incredible.

The human brain is a proactive prediction engine. Now, that won’t mean very much to you at this point in time, but the difference is profoundly important. And it is this understanding that is of particular interest to me, because I think it directly relates to why kids are having so much trouble with video games, especially free games designed to “hook” the user.