The Predictive Brain: Broadcasting Expectations, Receiving Errors

Your brain does not wait for sensory data. It predicts what it expects to see, hear, and feel — then uses sensory input only to correct its mistakes. You are not experiencing reality. You are experiencing your brain's best guess about reality, constantly revised by error signals.

The Astonishing Fact About Perception

Here is a truth so counterintuitive that most people resist it: you do not see with your eyes. You see with your brain.

Your eyes do not transmit images to your brain like a video camera sending footage to a monitor. They transmit electrochemical signals — noisy, incomplete, ambiguous signals. The light hitting your retina is two-dimensional. Your blind spot has no photoreceptors at all. Your visual field is sharp only at the very center — the periphery is blurry and almost colorless.

Yet you experience a rich, stable, high-resolution, three-dimensional world, all in focus, all at once.

How? Construction. Your brain builds this experience from the bottom up using prediction — a process so fundamental to its operation that it may be the single most important thing to understand about how the brain works.

The Broadcast Model, Inverted

This series has been building the case for the brain as a broadcasting and receiving station. But so far, we've focused on receiving — the brain as a radio picking up consciousness. Predictive processing turns this inside out.

The brain is also a transmitter.

It constantly broadcasts predictions — cascading expectations from high-level conceptual models down to low-level sensory areas. And it receives only one thing back: prediction errors — the mismatch between what it expected and what the senses actually report.

In the broadcast metaphor:

| Component | Standard Model | Predictive Processing | |-----------|---------------|---------------------| | Transmitter | External world | Brain's internal models | | Broadcast signal | Sensory data | Top-down predictions | | Receiver | Sensory organs | Prediction error units | | Received signal | Neural representation of world | Error signals (mismatches) | | Tuning | Attention to relevant input | Precision weighting of errors |

The brain is both a broadcasting station and a receiving station, operating simultaneously in a continuous feedback loop.

The Free Energy Principle

Karl Friston, a neuroscientist at University College London, formalized this framework into the Free Energy Principle — which may be the most ambitious theory in all of neuroscience.

The core idea: living systems must resist the natural tendency toward disorder (entropy). To do this, they must minimize a quantity called free energy — which, in this context, is the same as surprise or prediction error.

A system that minimizes free energy is one that constantly improves its internal model of the world until predictions match sensory input as closely as possible.

This is achieved through two complementary processes:

Perception (Model Revision)

When sensory input doesn't match predictions, the brain updates its internal model to reduce the discrepancy. This is learning — the slow process of making your inner model better match reality.

Action (World Revision)

Alternatively, the brain can change the world to match its predictions. This is action — moving your body, manipulating objects, speaking — to make sensory input conform to expectations.

Active Inference is the name for this unified process: perception and action working together to minimize prediction error. Perception changes the brain. Action changes the world. Both minimize free energy.

Friston's radical claim: everything the brain does — perceiving, learning, deciding, moving, feeling — can be understood as free energy minimization.

The Hierarchical Broadcast Architecture

The brain is organized hierarchically. Higher cortical areas represent abstract, stable features of the world (objects, concepts, causal relationships). Lower areas represent concrete, rapidly changing sensory details (edges, colors, sounds).

Predictions flow downward through this hierarchy. Each level predicts the activity of the level below it. When predictions are accurate, the lower levels are "explained away" — they don't need to send signals upward.

Prediction errors flow upward. When lower levels detect a mismatch between predictions and actual input, they send error signals up the hierarchy, triggering revisions at higher levels.

This is astonishingly efficient. The brain doesn't need to represent the entire sensory world in detail. It only processes what it couldn't predict.

What This Means for Conscious Experience

The implications are profound:

Perception is controlled hallucination. What you experience is not the world as it is, but your brain's best prediction of the world, constrained by sensory feedback. This is why optical illusions work — they exploit your brain's predictive machinery, feeding it input that matches a specific (incorrect) prediction. This is why placebos produce real effects. This is why your brain can "fill in" your blind spot seamlessly.

Reality testing is error detection. The feeling of being in contact with reality is the experience of your predictions being confirmed, or gracefully updated, by sensory error signals. When predictions and sensory input diverge too far — as in psychosis, dreaming, or psychedelic states — the sense of reality breaks down.

Attention is precision weighting. Not all prediction errors are equally informative. Some are noise. Attention is the mechanism by which your brain assigns higher precision (reliability) to some error signals and lower precision to others. This is why you can be in a noisy room but attend to one conversation — you're upweighting the precision of speech signals and downweighting the background.

The Broadcast Metaphor, Extended

The radio metaphor from Part 1 now takes on new depth:

| Layer | Metaphor | Predictive Processing Equivalent | |-------|----------|-------------------------------| | Transmitter | The station broadcasting | Higher cortical areas broadcasting predictions | | Broadcast signal | The music being sent | Top-down predictions (the brain's hypotheses) | | Medium | The EM field carrying the signal | Hierarchical neural pathways | | Receiver | The radio picking up the signal | Lower sensory areas receiving predictions | | Signal detected | The music you hear | Prediction errors sent back upward | | Static/noise | Interference in the signal | Imprecise or noisy sensory data | | Tuning dial | Selecting a frequency | Precision weighting (attention) |

The brain is constantly broadcasting predictions outward (down the hierarchy) and receiving error signals back — an ongoing conversation between expectation and evidence. This is the broadcast model in action, but the broadcast is happening within the brain, between its own hierarchical levels.

What This Explains

Why Consciousness Is Unified but Selective

Predictive processing elegantly explains why conscious experience feels unified yet limited. The hierarchy of predictions creates a single "best guess" model of the world at any moment. But only unpredicted information needs to be consciously processed — the rest runs on autopilot.

This is why you can drive a familiar route and have no memory of the journey. Your predictions were accurate, so no prediction errors were generated, and nothing needed to be consciously processed.

Why Meditation Changes Experience

Meditation trains the brain to hold predictions more lightly. Advanced meditators show reduced prediction error responses — they are less captured by the automatic predictive machinery. This manifests as a sense of "seeing things as they are" — raw sensory experience with less conceptual overlay.

Why Psychedelics Expand Consciousness

Psychedelics disrupt the precision-weighting system. The brain can no longer reliably distinguish between high-precision (important) and low-precision (noise) signals. More prediction errors flood the hierarchy, and the brain generates more and more creative revisions to its models. The result: experience expands beyond normal bounds.

Why We Dream

Sleep, particularly REM sleep, may be a state where prediction errors from the day are being processed without the constraint of real-world sensory input. The brain generates predictions (dream content) and attempts to resolve them, but without external sensory data to anchor them, the process runs free — producing the bizarre, narrative quality of dreams.

The Sense of Self

The self, in Friston's framework, is an inference — a prediction that a persistent, bounded entity exists, generated by the brain to minimize free energy over long timescales. The feeling of being a "self" is not a brute fact of experience. It's a model, constantly updated and refined.

This is why the self can be altered by meditation, psychedelics, and brain damage. It's a construct, not a fundamental reality.

Criticisms and Limits

Predictive processing is not without critics:

It explains too much. Some argue that the framework is so flexible it can account for any finding post hoc. If a result fits the theory, it's evidence for predictive processing. If it doesn't, the precision-weighting parameters can be adjusted to explain the discrepancy. Falsification is difficult.

It's computationally demanding. The brain would need to implement Bayesian inference at scale — computationally expensive. Whether biology actually does this or merely approximates it is unclear.

It doesn't solve the Hard Problem. Predictive processing explains how the brain processes information efficiently. It doesn't explain why any of this processing feels like anything. As Anil Seth puts it, predictive processing is a theory of content — what experience is about — not a theory of consciousness itself.

This last point is crucial for our series. Predictive processing is the broadcasting and receiving mechanism. It's the architecture. But it doesn't tell us what the medium is, or what the ultimate source of awareness is.

Bringing It Back to the Series

We now have two layers of the broadcast model:

1. The EM field (Part 2) — the physical medium that integrates information and may be the substrate of conscious experience.
2. Predictive processing (Part 3) — the functional architecture that broadcasts predictions and receives errors.

These are complementary. The EM field may be what consciousness is. Predictive processing may be how it functions in the brain.

In Part 4, we'll see what happens when one brain's broadcast system meets another's — the phenomenon of neural coupling, where two brains literally synchronize their activity during communication. This is where the broadcast metaphor leaps from individual to social cognition, and where the empirical evidence becomes most striking.

Next in series: When Brains Sync — Neural Coupling and the Shared Mind

Previous: CEMI Field Theory

Back to Consciousness