Dopamine Function Across Species and Behaviours: A Comparative Neuroscience Review

Reinforcement learning is the process by which animals change their behaviour from experience. When an action leads to a better-than-expected outcome, the animal becomes more likely to repeat that action. Here, I review studies that use self-stimulation, electrophysiological recordings, and fibre photometry to show that dopamine contributes to reward processing and reinforcement learning across rats, monkeys, mice, songbirds, and humans. First, stimulation of dopamine-related brain regions in rats reinforces actions in a way similar to natural rewards. Second, electrophysiological recordings from single dopamine neurons in thirsty monkeys show that dopamine spiking activity increases when juice is delivered, but only when that juice reward is unexpected, providing a learning signal that can reinforce the action associated with the reward. In mice, a fibre photometry study shows that dopamine signalling is also activated when a mother retrieves her crying pup. In songbirds, recordings show that dopamine responds to performance errors during birdsong learning and changes during courtship depending on social context. In humans, dopamine neuron recordings show that firing changes in response to unexpected financial gains and losses during decision-making. Across the recording studies, dopamine signals were shaped by whether an outcome was better or worse than expected. Together, these studies show a conserved core pattern that dopamine activity increases when an outcome is better than expected and decreases when an outcome is worse than expected, indicating that dopamine acts as a flexible learning signal across species and behaviours, although reward prediction error does not explain all of dopamine’s functions.