Definition
A prediction error is the mismatch between what the brain expected and what actually happened. Prediction errors are the currency of learning: they signal that the model of the world needs updating.
Why it matters
When you understand that surprise — not repetition — is what teaches the brain, therapy, education, and personal change all look different.
The Science
Dopaminergic signaling encodes reward prediction errors; broader cortical networks encode sensory and cognitive prediction errors. The magnitude and precision of the error, and the precision of the prior, together determine how much the model updates.
The NSI Perspective
NSI holds prediction error as the moment of possible change. Care organized around producing meaningful, tolerable prediction errors is aligned with the biology.
Clinical Implications
Exposure therapies, corrective emotional experiences, and behavioral experiments all trade in prediction errors.
Practical Application
If you want change, you need experiences that gently, safely violate your predictions. Repetition of the expected does not update the model.
References
- 1.Schultz W. Dopamine reward prediction-error signaling: a two-component response. Nat Rev Neurosci. 2016;17(3):183–195.