Dopamine is the first character in our cast, and the most misunderstood. You have probably read, in the kind of magazine article that flattens neuroscience for a general audience, that dopamine is the "pleasure chemical." That is wrong, or at least so incomplete that it produces bad predictions at the bedside.
Pleasure, in the brain, is largely an opioid story. We will get to opioids in a few minutes. Dopamine does something different, and once you understand what dopamine actually does, every other thing you learn about it — addiction, Parkinson's disease, schizophrenia, ADHD — will fit into place.
Three words tell you what dopamine does: want, predict, learn.
- In three words
- Want · Predict · Learn
- Core signal
- Reward prediction error — fires when reward beats prediction, silent when it matches, dips below baseline when it falls short
- Pathways
- Mesolimbic (reward), mesocortical (cognition), nigrostriatal (movement), tuberoinfundibular (prolactin)
- Receptors
- D1-like (D1, D5) excitatory; D2-like (D2, D3, D4) inhibitory
- Synthesis
- Tyrosine → L-DOPA → dopamine
- Breaks in
- Parkinson's (movement), addiction (reward), schizophrenia (salience), ADHD (motivation)
The prediction-error framework (Schultz, 1990s) is the lens for reward, motivation, learning, and much of psychiatric symptomatology.
Dopamine signals the gap between what you expected and what you got. When the reward you receive is bigger than you predicted, dopamine fires. When the reward is exactly as predicted, dopamine is silent. When the reward is smaller than predicted, or absent entirely, dopamine dips below baseline.
This pattern — the asymmetric response to predicted versus unpredicted outcomes — is called the reward prediction error signal. It was first described in the work of Wolfram Schultz, who in the 1990s recorded from single dopamine neurons in monkeys. His experiments are some of the most influential in modern motivational neuroscience, and the framework he established is the lens through which we now understand reward, motivation, learning, addiction, and a surprising amount of psychiatric symptomatology.
Here is why this matters clinically.
The gambler at the slot machine is not pulling the lever because winning feels good. The gambler is pulling the lever because the unpredictable timing of wins keeps the dopamine system firing prediction errors. The addict craving at the corner is not seeking pleasure. They are responding to a learned cue that predicts reward, and the prediction itself drives the want. The Parkinson's patient on dopamine replacement therapy who develops a sudden gambling problem is not displaying a hidden character flaw — they are showing the predictable result of pushing extra dopamine into a system that codes for reward prediction. The patient with schizophrenia who hears voices that feel deeply meaningful is, in one influential model, experiencing aberrant dopamine signals that attach false significance to random sensory data.
The same molecule, the same signal, different territories. Hold onto the three words: want, predict, learn.