Reward Prediction Error

Shrink Definition

Reward prediction error is the difference between the reward the brain expected and the reward it actually received. Wolfram Schultz's research showed that dopamine neurons signal this error rather than the reward itself. Positive prediction errors (better than expected) drive learning that this behavior is worth repeating. Negative prediction errors (worse than expected) drive updating in the other direction. Over time, once the reward becomes predictable, the dopamine signal moves back to the cue, not the reward.

Plain language

Reward prediction error is the gap between what your brain expected and what it got.

Shrink Insight

You don't chase the reward. You chase the anticipation of it.

Why it matters

Reward prediction error explains: habit formation, addiction, motivation, anhedonia, social media compulsion, and why a familiar reward eventually stops feeling like one.

Common misunderstanding

Dopamine isn't the pleasure chemical. It's the learning signal that tells your brain to update its expectations.

Shrink Perspective

The pull you feel is usually the anticipation, not the thing itself.

Shrink Reflection

What in your life do you enjoy less than the anticipation of it?

Shrink Journal

Pick a habit that used to feel rewarding and now feels flat. What did the surprise used to be?

Shrink Step

Today, name one moment where the anticipation was bigger than the actual experience. Notice without judgment.

Shrink Minute

Dopamine signals surprise, not pleasure.

Shrink Takeaway

Learning is driven by the gap, not the reward.

Medical boundary

Evidence summary

Reward prediction error has strong support from single neuron recordings, human neuroimaging, and computational modeling. It's one of the most robust findings in neuroscience.