Dopamine is a neurotransmitter molecule that influences brain pathways that are involved in motivation, movement, cognition and reward-driven learning. How it contributes to such varied behaviours is the subject of ongoing investigation. Writing in Nature, Mohebi et al.1 shed light on how dopamine release is regulated in the rat brain to accomplish different functions.
Dopamine is produced by neurons located in the midbrain, in regions known as the ventral tegmental area (VTA) and the substantia nigra pars compacta. The long axons of these neurons extend to other parts of the brain, including the nucleus accumbens, the dorsal striatum and the prefrontal cortex. Within these target sites, the axons branch extensively2 to form a structure known as an arbor. The textbook description of dopamine signalling suggests that the activation of dopamine-producing neurons in the midbrain generates electrical signals that travel along their axons to their target regions, where they cause a dopamine release that is ‘broadcast’ throughout the territories covered by the axonal arbors. This concept is fundamental to current ideas about how reward-based learning occurs: an unexpected reward leads to an increase in the activity of dopamine neurons that is assumed to transmit a dopamine signal throughout the target regions to facilitate learning3,4.
Yet dopamine release in the target regions is more complicated than its textbook description. For example, it can be regulated locally by neurotransmitters and other molecules5. Moreover, studies in animals of the activity of dopamine neurons using an imaging approach to monitor the activity of dopamine neurons or a microelectrode method to assess dopamine release indicate that an unexpected reward can cause the predicted increase in the activity of the axonal arbor, and dopamine release in the nucleus accumbens6,7. However, these features are absent in the neighbouring dorsal striatum6,7, providing an argument against a role for dopamine as a universal broadcast signal.
In addition to reward-dependent increases in dopamine release, dopamine release in the nucleus accumbens ramps up slowly8–13 as an animal approaches a reward site, before it obtains a reward. The amount of dopamine that is released during this ramping-up phase provides information about the value of the anticipated reward and motivates the amount of effort that the animal makes to attain it.
Most previous studies that tracked dopamine release during motivated behaviours used freely moving animals. By contrast, experiments that recorded the activity of dopamine neurons typically studied animals whose heads were held in a fixed position owing to technical constraints. Mohebi et al. took the key step of recording both neuronal activity and dopamine release in freely moving rats, although both properties were not recorded in the same animals.
The authors studied animals that were engaged in a behavioural task called a bandit task. The experimental apparatus used has ports into which a rat can poke its nose (Fig. 1). The central port is illuminated at the start of the task, which provides a cue for the animal to make a nose poke into that port, and then to stay in position until a sound cue prompts it to make a nose poke into one of the adjacent ports. Nose pokes into a neighbouring port might result in a food reward from a separate food port. The animal could therefore learn, for example, that a particular port was associated with a high probability of a food reward.
Mohebi et al. observed that, when the rate of reward increased, there was also an increase in the animals’ response rate, with a decrease in the time taken by the animal to start the task by making a nose poke into the central port after receiving the light cue. An increased response rate indicates that the animals are more motivated to take part in the task. Using a microdialysis technique to assess extracellular concentrations of dopamine in the brain revealed a pattern of dopamine elevation in distinct areas that correlated with reward availability during the task.
Dopamine increases were observed in the nucleus accumbens core, but not in the adjacent nucleus accumbens shell or in the dorsal striatum. Dopamine elevation was recorded in a part of the prefrontal cortex called the ventral prelimbic cortex, but a rise in dopamine levels was not observed in other areas of the prefrontal cortex. These data should put to rest the prevailing view that increased activity of dopamine neurons in response to rewards, or to cues that predict reward delivery, causes the broadcast of a dopamine signal to all of the brain regions that these neurons target.
The regions of the rat brain that showed an increase in dopamine release correspond to the human nucleus accumbens and ventromedial prefrontal cortex. In human brain-imaging studies, changes in activity in these regions are observed that correlate with the subjective value of rewards in decision-making tasks14. This offers a hint that Mohebi and colleagues’ findings might have relevance across species.
The authors found that neuronal activity in the VTA of rats was unaffected by an increase in the reward rate, however, suggesting that the motivation-related dopamine release is dissociated from the activation of dopamine neurons. To test this hypothesis, the authors examined dopamine release in the rat nucleus accumbens using a method for the rapid imaging of dopamine release15 that enables monitoring on a subsecond timescale, rather than the timescale of minutes that is possible with microdialysis.
As expected, the authors found that cues for reward availability, as well as the reward itself, were linked to an increase in the activity of dopamine neurons in the VTA (Fig. 1). These cues were also associated with an increase in dopamine release in the nucleus accumbens (the only brain region that was examined for dopamine release in this experiment) that serves as a learning signal to influence future behaviour.
As the reward rate increased, extracellular dopamine levels ramped up progressively as the rats approached either the central port or the port where food was dispensed, which is consistent with dopamine driving the motivation, as has been proposed previously8–13. However, there was no ramping up of the activity of dopamine neurons in the VTA (Fig. 1). This evidence for dopamine release in the absence of an increase in activity in dopamine neurons provides further support for the authors’ model that neuronal activity and dopamine release can be dissociated. Importantly, these data point to the possibility of a local regulation of dopamine release that is independent of the activity of dopamine neurons. The power of such local regulation5 is familiar to those who study the cells and circuits of the striatum.
Unanswered by Mohebi and colleagues’ study is which local factors in the nucleus accumbens might generate the ramped-up increases in dopamine release — or, in other words, what it is that motivates motivation. One possibility might be the release of the neurotransmitter acetylcholine from cells called cholinergic interneurons16,17. Of course, if the authors had reported evidence to support this, then the key question would have become which parts of the brain are conveying information about motivation to those neurons. Mohebi et al. report that there is a ramping up of activity of some non-dopamine neurons in the VTA before the animal carries out a nose poke. Perhaps those neurons have a role in facilitating dopamine release, which could be a topic for future research.
As well as providing evidence for the textbook view that a spike in the activity of dopamine neurons is accompanied by dopamine release, albeit not in all target regions, the unexpected observation of dopamine release in the absence of activity of dopamine neurons provides a new depth of understanding of dopamine signalling in the brain. Like the ramping up of dopamine, this is bound to provide the motivation for more work.