Representations of Reward and Movement in Drosophila Dopaminergic Neurons

The neuromodulator dopamine is known to influence both immediate and future behavior, motivating and invigorating an animal’s ongoing movement but also serving as a reinforcement signal to instruct learning. Yet it remains unclear whether this dual role of dopamine involves the same dopaminergic pathways. Although reward-responsive dopaminergic neurons display movement-related activity, debate continues as to what features of an individual’s experience these motor-correlates correspond and how they influence concurrent behavior. The mushroom body, a prominent neuropil in the brain of the fruit fly Drosophila melanogaster, is richly innervated by dopaminergic neurons that play an essential role in the formation of olfactory associations. While dopaminergic neurons respond to reward and punishment to drive associative learning, they have also been implicated in a number of adaptive behaviors and their activity correlates with the behavioral state of an animal and its coarse motor actions. Here, we take advantage of the concise circuit architecture of the Drosophila mushroom body to investigate the nature of motor-related signals in dopaminergic neurons that drive associative learning. In vivo functional imaging during naturalistic tethered locomotion reveals that the activity of different subsets of mushroom body dopaminergic neurons reflects distinct aspects of movement. To gain insight into what facets of an animal’s experience are represented by these movement-related signals, we employed a closed loop virtual reality paradigm to monitor neural activity as animals track an olfactory stimulus and are actively engaged in a goal-directed and sensory-motivated behavior. We discover that odor responses in dopaminergic neurons correlate with the extent to which an animal tracks upwind towards the fictive odor source. In different experimental contexts where distinct motor actions were required to track the odor, dopaminergic neurons become emergently linked to the behavioral metric most relevant for effective olfactory navigation. Subsets of dopaminergic neurons were correlated with the strength of upwind tracking regardless of the identity of the odor and remained so even after the satiety state of an animal was altered. We proceed to demonstrate that transient inhibition of dopaminergic neurons that are positively correlated with upwind tracking significantly diminishes the normal approach responses to an appetitive olfactory cue. Accordingly, activation of those same dopaminergic neurons enhances approach to an odor and even drives upwind tracking in clean air alone. Together, these results reveal that the same dopaminergic pathways that convey reinforcements to instruct learning also carry representations of an animal’s moment-by-moment movements and actively influence behavior. The complex activity patterns of mushroom body dopaminergic neurons therefore represent neither purely sensory nor motor variables but rather reflect the goal or motivation underlying an animal’s movements. Our data suggest a fundamental coupling between reinforcement signals and motivation-related locomotor representations within dopaminergic circuitry, drawing a striking parallel between the mushroom body dopaminergic neurons described here and the emerging understanding of mammalian dopaminergic pathways. The apparent conservation in dopaminergic neuromodulatory networks between mammals and insects suggests a shared logic for how neural circuits assign meaning to both sensory stimuli and motor actions to generate flexible and adaptive behavior