A rat is trained over many sessions to press a lever after a tone to receive food. Initially, VTA dopamine neurons fire when the rat eats the food. After extensive training, what pattern would you expect?
ADopamine neurons continue to fire strongly when the rat eats — because eating is always pleasurable regardless of experience
BDopamine neurons now fire at the tone, and the response to eating the food diminishes toward baseline
CDopamine neurons fire throughout the entire tone-lever-food sequence with equal magnitude
DDopamine firing ceases entirely — the rat has learned the sequence and no longer needs a teaching signal
This temporal shift is the defining hallmark of reward prediction error signaling. Early in training, the food itself is unexpected, generating a large positive prediction error — dopamine fires at delivery. As learning proceeds, the tone reliably predicts the food, so the dopamine response shifts backward to the earliest predictor (the tone). The food itself, now fully expected, no longer generates a prediction error — dopamine firing at food delivery returns to baseline. This demonstrates that dopamine does not encode pleasure; it encodes the *surprise* of reward relative to expectation.
Question 2 Multiple Choice
A well-trained animal expects food after a tone. On a probe trial, the tone plays but no food is delivered. What happens to dopamine neuron activity at the time when the food would have arrived?
ADopamine firing increases — the animal is alerted and motivated to search for the missing reward
BDopamine firing stays at baseline — no reward, no signal, so the system is silent
CDopamine firing drops below baseline — a negative prediction error signaling the outcome was worse than expected
DDopamine firing first increases then rapidly decreases to encode the disappointment sequence
The reward prediction error model predicts three states: positive error (reward better than expected → firing above baseline), no error (reward matches expectation → firing at baseline), and negative error (reward worse than expected, or expected reward omitted → firing below baseline). The omission of an expected reward is a negative prediction error — the outcome was worse than what was predicted. This dip below baseline is the 'anti-reward' signal that teaches the system to reduce the value assigned to the cue. It is the mechanism by which extinction learning occurs.
Question 3 True / False
An addict who reports that heroin no longer produces euphoria may still compulsively seek the drug because dopamine signaling in the reward circuit has sensitized to drug-associated cues.
TTrue
FFalse
Answer: True
This dissociation between 'wanting' and 'liking' is a key feature of addiction. Tolerance reduces the hedonic impact of the drug itself (opioid receptor downregulation reduces 'liking'), but repeated drug use sensitizes the mesolimbic dopamine circuit to drug-associated cues — needles, locations, people, smells. These cues trigger strong dopamine responses (large positive prediction errors) that drive compulsive seeking behavior ('wanting') even as the drug itself delivers diminishing pleasure. Dopamine encodes incentive salience, not pleasure, which explains why craving can intensify even as enjoyment fades.
Question 4 True / False
Dopamine neurons in the VTA fire most strongly in response to the subjective pleasantness of a reward experience.
TTrue
FFalse
Answer: False
This is the central misconception about dopamine. VTA dopamine neurons encode reward prediction error — the difference between expected and received reward — not the absolute hedonic value of an experience. A reward that is fully predicted produces no increase in dopamine firing, regardless of how pleasant it is. The clearest evidence is the temporal shift in conditioning: after learning, dopamine fires to the predictive cue (which produces no pleasure itself), not to the reward. Pleasure (hedonic value) is thought to be more closely tied to opioid signaling in the nucleus accumbens shell, not dopamine.
Question 5 Short Answer
A patient with early Parkinson's disease has lost dopaminergic neurons primarily in the substantia nigra, but their neurologist notes they also show reduced motivation and difficulty learning which new behaviors lead to positive outcomes. How does the prediction error function of dopamine explain these non-motor symptoms?
Think about your answer, then reveal below.
Model answer: Dopamine prediction error signals are necessary for reinforcement learning — they update the value estimates stored in striatal and prefrontal circuits that guide motivated behavior. Without a reliable prediction error signal, the brain cannot update its model of which actions lead to reward. The patient would have difficulty learning that a new behavior produces a good outcome, because the dopamine 'teaching signal' that would reinforce the behavior is absent or degraded. Reduced motivation follows because the mesolimbic pathway (VTA → nucleus accumbens) encodes incentive salience — the drive to pursue predicted rewards. Depleted dopamine signaling reduces this drive, producing the amotivation and anhedonia characteristic of both Parkinson's disease and depression.
The non-motor symptoms of Parkinson's disease are among the strongest clinical evidence for the prediction error theory of dopamine. They also illustrate that different dopamine pathways serve different functions: the nigrostriatal pathway (substantia nigra → dorsal striatum) controls movement, the mesolimbic pathway (VTA → nucleus accumbens) controls reward motivation, and the mesocortical pathway (VTA → PFC) controls value-based decision making. Early Parkinson's affects the nigrostriatal pathway most severely, but as the disease progresses, VTA neurons also degenerate, producing the motivational and learning deficits described.