What is Prediction error?
A prediction error is the difference between what the brain expected to happen and what actually happened. It is the fundamental learning signal in the brain, driving the updating of beliefs, expectations, and behaviors.
How it works
The brain is a prediction machine that constantly generates expectations about upcoming events. When reality matches predictions, little learning occurs. When reality differs (positive prediction error: better than expected; negative prediction error: worse than expected), the brain updates its model. Dopaminergic neurons in the midbrain encode reward prediction errors, firing above baseline for positive surprises, below baseline for disappointments, and at baseline for expected outcomes. This signal drives both reward learning and decision-making.
Applied example
A person who tries a restaurant expecting mediocre food and discovers it is excellent experiences a positive prediction error. The dopamine surge signals ‘this was better than expected’ and updates the brain’s model, making the person more likely to return. If the food had been exactly as expected, no learning signal would fire.
Why it matters
Prediction error is the brain’s universal learning currency, explaining how humans and animals learn from experience by identifying and correcting the gaps between expectation and reality.
