Surprise and prediction error — how does the brain revise its expectations?

VZ Lens

From a VZ lens, this piece is not for passive trend tracking - it is a strategic decision input. The brain is constantly making predictions and learns through surprises (prediction errors). Discover how this mechanism reshapes our expectations and our perception of reality. Its advantage appears only when converted into concrete operating choices.

The brain does not passively receive stimuli—it constantly makes predictions and learns from the differences.

---

TL;DR

The human brain is not a camera that passively records reality. It constantly generates predictions—expectations of what will happen—and when reality deviates from these, a prediction error occurs. This error is not just an unpleasant sensation: it is also a learning signal, an attentional reflex, and a belief-updating mechanism. Surprise is the moment when the prediction error reaches the threshold of consciousness. The synthetic persona must also handle this mechanism if we want lifelike decision-making simulations.

---

The Silence of the Berlin Lab

The monitors are already turning off, the experimental chair is empty. The only sound in the room is the hum of the refrigerator, and the pages of an abandoned notebook flutter on the floor in the draft. I sit in my chair and watch the frozen brain activity curves on the screen. The lines suddenly spike at one point, then drop back down—a momentary outburst when reality didn’t match the expected pattern. The silence is particularly pronounced now. The ticking of the clock on the wall is exactly the same as it was an hour ago, but I am no longer the same. Because I saw how a brain reacts when the world doesn’t deliver what it expected. The curves show: surprise isn’t just a feeling. It’s a physical trace.

1. The Basis of Predictive Coding

One of the leading theories in neuroscience over the past twenty years is predictive coding. Karl Friston, a neuroscientist at UCL, developed its most sophisticated version.

The essence of the theory is simple: the brain does not passively receive incoming stimuli—it actively predicts what is coming. The difference between the information received from the senses and the prediction is the prediction error.

When the incoming stimulus matches the expectation: there is no error, no learning, and low energy consumption.

When the incoming stimulus deviates from the expectation: a prediction error occurs—and the system updates.

This update is not passive. The brain makes a decision: should it update its expectations (learn from the error), or classify the incoming signal as noisy, irrelevant, or erroneous (do not learn)?

---

2. Surprise as a threshold of consciousness

Prediction errors occur constantly—mostly unconsciously. While driving, our muscles are constantly correcting small prediction errors—but we don’t notice this.

A surprise is the moment when a prediction error reaches the threshold at which it generates conscious attention.

A familiar brand suddenly disappears: surprise. A well-known store sells a different product: surprise. A salesperson unexpectedly quotes a very low price: surprise.

Surprise grabs attention. According to evolutionary logic, an unexpected event is potentially important—whether dangerous or an opportunity.

---

3. Five Types of Surprise

Not all surprises are the same. In the synthetic persona system, it is useful to distinguish five types of surprise:

Type	Description	Consumer example
Informative surprise	A new fact that overrides a previous belief	“This product turned out to be more effective than I thought”
Social surprise	Unexpected behavior by another person	“My friend bought this too—I didn’t expect that”
Identity surprise	Affects the boundaries of self-image	“I wouldn’t have thought I’d like this style either”
Existential surprise	An event that challenges basic assumptions	“The brand I trusted has gone under”
Positive surprise	An actual experience that favorably exceeds expectations	“It was much better than I expected”

---

4. What happens after a surprise?

Post-surprise behavior is one of the least modeled—and one of the most important—consumer dynamics.

What typically follows:

1. Attentional freeze: A brief but measurable moment in which cognitive capacity focuses on the surprising stimulus. Other stimuli are filtered out.

2. Schema-mapping attempt: The brain attempts to fit the surprising event into an existing system of expectations (schema). If successful, the surprise causes a minor correction and subsides.

3. Schema revision or schema rejection: If the surprising event cannot be fitted into the existing schema, two things can happen:

• The schema overrides the event: “This must be an exception; it’s not typical.”
• The event overrides the schema: previously formed expectations change

4. Emotional processing: The emotional charge of the surprise is decisive. A positive surprise reinforces brand associations. A negative surprise—especially if it causes a schema revision—has a very strong, lasting negative effect.

---

5. Schema rigidity: the rigidity of expectations

Not everyone is the same in terms of how much they update their expectations in the face of a surprise.

Schema rigidity measures how resistant a person is to updating their existing system of expectations.

In cases of high schema rigidity:

• The surprise tends to cause schema rejection rather than schema revision
• The person defends their existing belief system against contradictory evidence
• In the case of a negative surprise, strong rationalization kicks in: “this must be an exception”

In the case of low schema rigidity:

• The surprise leads to a faster update
• The person adjusts their expectations to the experience more flexibly
• Both positive and negative surprises have a stronger, more lasting effect

Schema rigidity is related to the Big Five Openness dimension (low O → higher schema rigidity), but is not identical to it.

---

6. Belief update speed

Another important variable within the framework of predictive coding is belief update speed: how quickly does a person update their expectations in response to a prediction error?

This variable explains why two people react completely differently to the same surprise:

The first says, “OK, I’ve noted that; next time I’ll think differently”—a quick, flexible update.

The second says, “That’s not possible; it can’t be true”—and even weeks later, they stick to their original expectations—a slow, conservative update.

The speed of belief updating isn’t good or bad—it’s functionally different. Fast update: better in a rapidly changing environment. Slow update: better when the system is disturbed by noise (it maintains stability amid many conflicting signals).

---

7. Surprise and Decision

The direct effects of surprise on decision-making:

Positive surprise:

• Increases perceived satisfaction (the customer feels they’ve “gained” something)
• Creates a strong, lasting positive association
• Increases the likelihood of word-of-mouth (it’s worth sharing)

Negative surprise:

• Has a stronger impact than a negative experience that meets negative expectations
• Threatens the schema: if a person trusted the brand, the surprise affects that trust
• Triggers strong emotional reactivity that goes beyond the specific problem

[!WARNING] The asymmetry of negative surprises According to research, a negative surprise has, on average, 2–3 times the impact of a negative event of the same intensity that a person anticipated. This is because schema revision requires far more cognitive and emotional resources than processing a simple negative experience.

---

8. The Surprise Profile in the Synthetic Persona

In a dynamic synthetic persona system, surprise handling can be profiled:

Variable	Range	Example
Surprise sensitivity	0.0–1.0	Low O-value, high IoU → higher sensitivity
Schema rigidity	Low / Medium / High	How strongly it defends its expectations
Belief update speed	Fast / Medium / Slow	How quickly beliefs are updated after a surprise
Threat appraisal bias	Negative / Neutral / Positive	To what extent an ambivalent surprise is assessed as a threat
Emotional reaction amplitude	Low / Medium / High	How strong an emotional response the surprise generates

---

9. Application: How can this be used in market research?

Specific applications of surprise modeling:

Product experience testing: What surprised the tester (positively or negatively)? Where did the actual experience deviate from expectations? This difference can also be modeled in a simulation.

Communication test: Which message elicited surprise? In the best case, this generates attention; in the worst case, it leads to schema confrontation and rejection.

Price anchoring effect: The direction of the price surprise (it was more expensive than I thought / it was cheaper) is decisive in the purchasing decision.

Simulation scenario: “What happens if our brand’s market position changes?” — this can be simulated using the surprise profile: how does the target group persona process the shift in expectations?

---

10. Summary

Prediction error and surprise are the drivers of human learning and decision-making. When reality deviates from expectations, the brain updates—but not everyone does so at the same speed or in the same direction.

The synthetic persona system must also handle the surprise profile:

• surprise sensitivity
• schema rigidity (resistance to updating expectations)
• belief update speed
• amplitude of emotional reactivity

This layer allows us to model not only “normal operation” in the simulation, but also reactions to unexpected situations.

---

This article is the ninth part of the Synthetic Personas series. Next part: CAPS — if-then behavioral signatures and personality as a context-dependent system.

---

Zoltán Varga | vargazoltan.ai — Market research, artificial intelligence, synthetic thinking

Strategic Synthesis

• Convert the main claim into one concrete 30-day execution commitment.
• Set a lightweight review loop to detect drift early.
• Close the loop with one retrospective and one execution adjustment.