Search

Animals approach stimuli that predict a pleasant outcome. After the paired presentation of an odour and a reward, Drosophila melanogaster can develop a conditioned approach towards that odour. Despite recent advances in understanding the neural circuits for associative memory and appetitive motivation, the cellular mechanisms for reward processing in the fly brain are unknown. Here we show that a group of dopamine neurons in the protocerebral anterior medial (PAM) cluster signals sugar reward by transient activation and inactivation of target neurons in intact behaving flies. These dopamine neurons are selectively required for the reinforcing property of, but not a reflexive response to, the sugar stimulus. In vivo calcium imaging revealed that these neurons are activated by sugar ingestion and the activation is increased on starvation. The output sites of the PAM neurons are mainly localized to the medial lobes of the mushroom bodies (MBs), where appetitive olfactory associative memory is formed. We therefore propose that the PAM cluster neurons endow a positive predictive value to the odour in the MBs. Dopamine in insects is known to mediate aversive reinforcement signals. Our results highlight the cellular specificity underlying the various roles of dopamine and the importance of spatially segregated local circuits within the MBs.

Communication by means of pheromones plays predominant roles in colony integration by social insects. However, almost nothing is known about pheromone processing in the brains of social insects. In this study, we successfully applied intracellular recording and staining techniques to anatomically and physiologically characterize brain neurons of the ant Camponotus obscuripes. We identified 42 protocerebral neurons that responded to undecane and/or formic acid, components of alarm pheromones that evoke attraction or evasive behavior, respectively. Notably, 30 (71%) of these neurons were efferent (output) or feedback neurons of the mushroom body, and many of these exhibited different responses to formic acid and undecane. Eight of the remaining 12 neurons had arborizations in the lateral and/or medial protocerebrum, which receive terminations of efferent neurons of the mushroom body and from which premotor descending neurons originate. The remaining four neurons were bilateral neurons that connect lateral accessory lobes or dorsal protocerebrums of both hemispheres. We suggest that the mushroom body of the ant participates in the processing of alarm pheromones. Seventeen (40%) of 42 neurons exhibited responses to nonpheromonal odors, indicating that the pheromonal and nonpheromonal signals are not fully segregated when they are processed in the protocerebrum. This may be related to modulatory functions of alarm pheromones, i.e., they change alertness of the ant and change responses to a variety of sensory stimuli.