Monoamine neurotransmitters such as dopamine, serotonin, and noradrenaline are released by small populations of neurons that send information to much of the brain. These neurons function like major news channels; both keeping many brain regions up to date on current affairs as well as making predictions of what the future will hold. But what’s on the news? What types of “stories” do these neurons report on?
For a long time, it’s been known that dopamine neurons broadcast information related to value (when an animal receives an unexpected reward or predicts that it will receive a reward in the future) to many regions of the cortex and striatum. Additionally, many addictive drugs are known to increase dopamine release by targeting dopamine neurons or the neurons that directly regulate them. Based on these observations, the prevailing theory has long been that dopamine is a “news channel” for the brain that focuses chiefly on reward and reinforces reward-seeking behavior. However, recent studies showed that some dopamine neurons fire in response to both rewarding and aversive stimuli. These surprising findings called to question the idea of “value” in dopamine signals. It was proposed that some dopamine neurons may signal “value”, whereas other dopamine neurons may signal “absolute value”. However, the idea has been controversial. More experiments were needed to understand what information dopamine actually broadcasts.
We hypothesized that, instead of acting as a “national” news channel reporting on value, distinct subpopulations of dopamine neurons could relay distinct information to sub-domains of the brain, acting as several “local” news channels reporting on a wider range of stories. To look for this sort of organization, we first used a meso-scale anatomical approach to label the monosynaptic inputs of dopamine neurons with different projection targets and found that dopamine neurons projecting to the posterior “tail” of the striatum had unique inputs compared to other populations (Menegas et al., 2015). The next question was what information each population is broadcasting? To find out, in this work, we infected dopamine neurons with GCaMP (a genetically encoded calcium indicator that can be used to read out neural activity) and placed optic fibers along the anterior-posterior axis of the striatum to record activity from dopamine neurons projecting to those areas while the animals interacted with their environment. We found a clear segregation between dopamine signals in the anterior striatum and those in the posterior striatum. Dopamine axons in the anterior striatum encoded value, whereas those in the posterior striatum encoded general stimulus salience, regardless of value. But even more interestingly, we found something that we hadn’t expected: opposite initialization to novel cues during learning in the anterior and posterior striatum.
Posterior striatum (salience coding) dopamine responded strongly to novel cues, whereas anterior striatum (value coding) dopamine didn’t respond to cues unless they reliably predicted reward. This fits beautifully with our intuition that novel stimuli are not necessarily rewarding (i.e. valuable), but that we still do pay more attention to them than to familiar ones (i.e. they are salient). These new results revised the notion that dopamine neurons broadcast a single message (“value” or “absolute value”) to the entire brain. These results suggest that dopamine could be involved in a more diverse set of cognitive functions than previously assumed.. While anterior striatum dopamine likely does guide value-based reward seeking, posterior striatum dopamine could guide a totally separate cognitive process, such as attentional allocation.
We demonstrated that dopamine broadcasts different types of information to different regions of the brain. Our anatomy and recording data will be a firm foundation for future studies to characterize the possible functional roles for dopamine in the brain, and their relation to learning, addiction, and disease states.