Abstract
Schultz, Dayan, and Montague show that dopamine neurons fire for a better-than-predicted reward and pause when an expected reward is omitted. The signal teaches the cue. It does not score the whole life.
Schultz, W., Dayan, P., & Montague, P. R. (1997). Science, 275(5306), 1593–1599. DOI
Key findings
- Prediction, not pleasure — The response tracks the difference between expected and received reward.
- Omission teaches — A missing expected reward produces a dip. That dip trains the cue to stop predicting pay.
Limits
Primate electrophysiology plus a computational model (Science, 1997, 275, 1593–1599). Crossref DOI 10.1126/science.275.5306.1593. It does not say which personal project to keep, and it does not prescribe a spare habit.
Application
When a miss tells you to leave the tree, separate the cue the error is training from the project you still mean to keep (Reward prediction error).