The Predictive Brain and the Self

You wake up one morning and the person you have been performing for the last decade just does not fit anymore. The career, the relationship, the careful self-presentation — all of it feels borrowed. You are not depressed, exactly. You are not sure what you are. What you do know is that the story you told about yourself has stopped working. Neuroscience has a very specific explanation for why that happens, and for why it feels so disorienting. It also has something to say about what comes next.

Your Brain Is Not a Camera. It Is a Forecasting Machine.

There is a version of how the brain works that most people carry around without ever questioning it. In that version, the world sends information inward through your senses, and the brain processes it. Reality arrives first. The brain responds second.

That version is mostly wrong.

What neuroscientists now believe — based on decades of converging evidence — is closer to the opposite. Your brain does not wait for the world to tell it what is happening. It runs a model. It makes a continuous prediction about what it expects to sense next, based on everything it has experienced before. Then it checks that prediction against the incoming signal. If there is a gap between the guess and reality, the brain registers what researchers call a prediction error. And it uses that error to update the model.

This framework is called predictive processing. The German physiologist Hermann von Helmholtz sketched the idea in the 1860s, describing perception as a form of unconscious inference. It was formalised much later by Karl Friston at University College London, whose work through the 2000s and 2010s built it into one of the most influential theories of brain function in modern neuroscience.

The key point is this: what you perceive, at any given moment, is mostly your brain's guess. The outside world only gets a vote when the guess is wrong.

That drive toward predictability is not a flaw. It is extraordinarily efficient. The brain cannot process every signal arriving every second at full resolution. So it shortcuts. It fills in. It uses prior experience to generate the most plausible version of reality and only devotes full attention to the parts that do not match.

Most of the time, this works beautifully. You do not have to consciously calculate that the floor will hold your weight each morning. You do not have to re-learn your name. The model handles all of that in the background.

But the model also handles you.

The Self Is a Model the Brain Maintains

Every model has structure. At the bottom level, the brain predicts simple things: the specific colour of that surface, the pitch of that sound. At higher levels, it predicts more complex things: whether that face is friendly, whether this situation is dangerous. And at the very top of the hierarchy, it holds the most general predictions of all — predictions about what kind of person you are, how you tend to respond, what you value, what you deserve.

That is what the self is, inside a predictive processing framework. It is the brain's highest-level, most generalised model of the organism doing the experiencing.

This idea has been developed carefully by philosopher and cognitive scientist Andy Clark, particularly in his 2016 book Surfing Uncertainty. Clark argues that the sense of being a continuous, stable self — the feeling that there is a "you" persisting through time — is itself a prediction. It is the brain's best running guess about the agent at the centre of all this experience.

That is a strange thing to sit with. You are not simply receiving a self. You are generating one, constantly, based on a model.

Now consider what happens when that model stops matching your life. When the person you have been predicting — the reliable one, the ambitious one, the one who holds it all together — keeps running into evidence that she is not quite real anymore.

The brain experiences that as a cascade of prediction errors. Not just surprises about the world. Surprises about the self.

This is one reason why identity upheaval feels so physically destabilising. It is not just emotionally difficult. The brain's core operating model is failing. That failure registers in the body — in sleep, in appetite, in the strange feeling that you are watching yourself from a slight distance, that nothing quite lands the way it used to.

If that last sentence sounds familiar, there is a dedicated piece on this site about depersonalisation and identity transformation that goes deeper into that specific experience.

Why the Old Self Holds On So Stubbornly

Here is what makes this genuinely hard. Prediction errors at the lower levels of the hierarchy update fast. You see an unexpected object on your desk. Your brain corrects. Done. But prediction errors at the very top of the hierarchy — at the level of identity — move slowly. The brain weights its highest-level beliefs heavily. It takes a lot of contradictory evidence to shift them.

This is not irrationality. It is a design feature. A brain that revised its deepest assumptions every time something unexpected happened would be a chaotic brain. Stability at the top level allows the lower levels to absorb daily surprises without throwing the whole system into disarray.

But that stability becomes a cage when the assumptions at the top are simply wrong.

Think about the person who has spent fifteen years defining themselves through their career. Their brain's model includes, at its deepest level, something like: I am someone whose work matters. My value is tied to my output. When the job ends — through burnout, redundancy, or a quiet realization that the work no longer means anything — the world keeps generating evidence that the old model is failing. But the model resists revision. It keeps predicting the old self. And the gap between that prediction and reality costs energy, every single day.

Psychiatrist and neuroscientist Mark Solms, in his 2021 book The Hidden Spring, connects this resistance to survival mechanisms deep in the brain. The brain treats its own predictions as anchors for safe navigation. Updating them feels, at a bodily level, like danger. That is why change — even wanted change — so often arrives accompanied by dread.

This is not weakness. It is the same mechanism that keeps you upright and competent on an ordinary Tuesday. The tragedy is that it also keeps you trapped in a story you have outgrown.

When the Model Breaks: Burnout, Grief, and the Necessity of Crisis

There comes a point, for many people, when the gap between the brain's self-model and lived reality becomes too large to paper over. The system cannot suppress the prediction errors anymore. Something breaks open.

Burnout often works this way. Not as a failure of willpower or resilience, but as a collapse of the predictive model. The brain has been running a version of the self — hardworking, indispensable, capable of absorbing everything — that no longer corresponds to what the body can actually sustain. When the model breaks down, it does not feel like relief. It feels like falling.

Grief does something similar. The death of someone you loved deeply, or the end of a marriage, or the loss of a future you had built in your imagination — these all strip out pieces of the self-model that were never visible until they were gone. Your brain had been predicting a world that included that person, that relationship, that future. Now those predictions keep firing and finding nothing. The ache of grief is partly the ache of a model being forcibly rewritten.

The neuroscientist Lisa Feldman Barrett, in her 2017 book How Emotions Are Made, argues that emotion itself is a prediction. Fear is not a response to danger — it is a prediction that danger is probable. Grief is not just sadness — it is a continuous error signal firing where something familiar used to be. Her work, rooted in the predictive processing framework, gives emotional suffering a strange kind of dignity. It is real, and it is structural. The pain is the sound of the model being rebuilt.

What Jung called the nigredo — the blackening, the first stage of alchemical transformation where the old form must dissolve before a new one can take shape — maps onto this moment with uncomfortable precision. Not as metaphor, but as description. The old model is failing. The new one has not yet formed. The space between them is real, and it is genuinely dark.

For more on how that Jungian stage connects to what happens in the body during collapse, the article on burnout and the nigredo lays it out in detail.

How the Model Actually Changes

The predictive processing framework is not pessimistic. A model that can be wrong can also be revised. The brain that locked you into an outdated self is the same brain that can build a new one. But that process has real structure. It does not happen through positive thinking or a decision made at midnight.

First, the brain needs what researchers call high-precision prediction errors — errors it cannot dismiss. Small contradictions can be absorbed and explained away. But sustained, undeniable evidence that the model is failing forces revision. This is one reason why crisis so often precedes genuine change. The crisis provides errors too large to ignore.

Second, the new model has to be practiced. The brain encodes patterns through repetition. The self-concept is not just a thought — it is a set of deeply rehearsed predictions embedded in neural pathways (the connections between brain cells that strengthen the more they are used). Shifting identity means building those pathways in a new direction. That is slow. It requires new behaviour, new context, new relationships that generate different feedback.

Michael Merzenich has argued across decades of work, summarised in a popular book, that the adult brain remains more changeable than scientists once believed. New skills, new environments, and new relational contexts all produce measurable structural change. Merzenich, M. (2013). Soft-Wired: How the New Science of Brain Plasticity Can Change Your Life. Parnassus. The changes are real. They are also proportional to the depth and consistency of the new experience.

Third — and this is where it gets uncomfortable — the brain also changes through what it refuses. The predictive processing account of repression is not Freudian in its mechanics, but it arrives at a similar place. When a prediction error is consistently too threatening to be incorporated, the brain finds ways to suppress it. It lowers the weight given to the incoming signal. It redirects attention. It builds habits that prevent the error from arising at all.

Jung recognised this in clinical terms, calling it the shadow — the parts of the self that the conscious mind will not acknowledge. He described individuation, the lifelong process of becoming more fully oneself, as requiring repeated encounters with exactly those suppressed parts. The predictive processing model gives that a neurological grounding. What you suppress is not gone. It is a persistent error your model has learned to mute. And muting costs energy, every day.

For a deeper account of what the default mode network — the brain's self-referential system — does with these suppressed patterns, see the article on the default mode network and identity.

What This Means When You Are Trying to Change

None of this is a roadmap. The brain does not update on schedule. But understanding the mechanism does change how you relate to the difficulty.

When you feel stuck — when you know, intellectually, that you want a different life and find yourself repeating the same patterns — that is not a moral failing. It is the predictive machinery doing exactly what it was built to do: protect the existing model. The model resists because revision carries risk. Every deeply held belief about yourself is also a load-bearing wall. Pull it out and the whole structure shakes.

When people emerge from serious identity crisis — the divorce, the burnout, the collapse of a faith — they often describe a period of profound disorientation followed, eventually, by something that feels like clarity. That is not a metaphor for psychological health. It is, in part, a description of a model that has been revised. The old predictions stop firing. New ones, built from new experience, start fitting the world more accurately. The sense of being a stranger to yourself lifts.

That process is not linear and it is rarely fast. But it is real. And it is the brain doing something it was built to do — not breaking, but updating.

The philosopher Thomas Metzinger, in his 2003 work Being No One, argued that the self is not a thing you have. It is a process the brain performs. The process can be interrupted, revised, and rebuilt. That is not terrifying. It is, once you can hold it, the ground of possibility.

Researchers studying identity development across adulthood — including the late-career work of psychologist Dan McAdams at Northwestern University — have consistently found that the people who navigate major life transitions most coherently are not those who changed the least, but those who were able to revise the narrative story they held about themselves. McAdams calls this a "narrative identity" — the story you tell yourself about who you are and how you got here. When that story can be rewritten to include the crisis, rather than treating it as a rupture, the brain has something to work with. McAdams, D. P. (2001). The psychology of life stories. Review of General Psychology, 5(2), 100–122.

This connects to what active imagination, as Jung described it, actually does in practice. It is not about producing meaningful dreams. It is about giving the suppressed prediction errors a form — a voice, an image, a conversation — so the model can finally process what it has been refusing. For a clear account of that practice, see the article on active imagination as a Jungian technique.

Frequently Asked Questions

What is predictive processing, in plain terms?

Your brain does not wait for the world to tell it what is happening. It runs a model — a continuous guess about what is out there, built from everything you have lived through before. What you see, hear, and feel is mostly that guess. Only the gaps between the guess and reality update the model. Predictive processing is the name for this loop: predict, sense the error, correct, repeat.

Why does identity feel so fixed if the brain is always updating?

The model updates constantly for small things, but identity-level beliefs sit very deep in the hierarchy. They carry huge weight. The brain needs enormous, repeated evidence before it will revise something that foundational. That is why you can know, rationally, that your life has changed and still feel, moment to moment, exactly like the person you used to be.

Is burnout connected to predictive processing?

Yes. Burnout often happens when the brain's model is running on a set of assumptions — about what you owe others, what success looks like, what makes you safe — that no longer match your actual life. The mismatch between that internal model and reality generates a constant drain. The brain keeps predicting a version of you that no longer exists and keeps getting it wrong.

Can the self actually change, or is that wishful thinking?

The self is not a fixed object. It is a model the brain maintains and updates. Research on neuroplasticity — the brain's physical capacity to rewire itself — confirms that sustained new experience genuinely changes brain structure. Identity change is real and measurable. But it is slow, and it tends to require the kind of discomfort that most stabilising habits are designed to avoid.

What does any of this have to do with Jung or alchemy?

Jung described individuation — the lifelong process of becoming more fully yourself — as a series of encounters with parts of the self that the ego refuses to recognise. Predictive processing gives that a neurological frame: the parts you refuse to recognise are the persistent prediction errors your model has learned to suppress. Crisis, in both frameworks, is the moment suppression fails. And that failure is often where real change begins.