The brain adaptively integrates present sensory input, past experience, and options for future action. The insect mushroom body exemplifies how a central brain structure brings about such integration. Here we use a combination of systematic single-cell labeling, connectomics, transgenic silencing, and activation experiments to study the mushroom body at single-cell resolution, focusing on the behavioral architecture of its input and output neurons (MBINs and MBONs), and of the mushroom body intrinsic APL neuron. Our results reveal the identity and morphology of almost all of these 44 neurons in stage 3 Drosophila larvae. Upon an initial screen, functional analyses focusing on the mushroom body medial lobe uncover sparse and specific functions of its dopaminergic MBINs, its MBONs, and of the GABAergic APL neuron across three behavioral tasks, namely odor preference, taste preference, and associative learning between odor and taste. Our results thus provide a cellular-resolution study case of how brains organize behavior.

Dopaminergic neurons serve multiple functions, including reinforcement processing during associative learning [1-12]. It is thus warranted to understand which dopaminergic neurons mediate which function. We study larval Drosophila, in which only approximately 120 of a total of 10,000 neurons are dopaminergic, as judged by the expression of tyrosine hydroxylase (TH), the rate-limiting enzyme of dopamine biosynthesis [5, 13]. Dopaminergic neurons mediating reinforcement in insect olfactory learning target the mushroom bodies, a higher-order "cortical" brain region [1-5, 11, 12, 14, 15]. We discover four previously undescribed paired neurons, the primary protocerebral anterior medial (pPAM) neurons. These neurons are TH positive and subdivide the medial lobe of the mushroom body into four distinct subunits. These pPAM neurons are acutely necessary for odor-sugar reward learning and require intact TH function in this process. However, they are dispensable for aversive learning and innate behavior toward the odors and sugars employed. Optogenetical activation of pPAM neurons is sufficient as a reward. Thus, the pPAM neurons convey a likely dopaminergic reward signal. In contrast, DL1 cluster neurons convey a corresponding punishment signal [5], suggesting a cellular division of labor to convey dopaminergic reward and punishment signals. On the level of individually identified neurons, this uncovers an organizational principle shared with adult Drosophila and mammals [1-4, 7, 9, 10] (but see [6]). The numerical simplicity and connectomic tractability of the larval nervous system [16-19] now offers a prospect for studying circuit principles of dopamine function at unprecedented resolution.

The Drosophila larva has turned into a particularly simple model system for studying the neuronal basis of innate behaviors and higher brain functions. Neuronal networks involved in olfaction, gustation, vision and learning and memory have been described during the last decade, often up to the single-cell level. Thus, most of these sensory networks are substantially defined, from the sensory level up to third-order neurons. This is especially true for the olfactory system of the larva. Given the wealth of genetic tools in Drosophila it is now possible to address the question how modulatory systems interfere with sensory systems and affect learning and memory. Here we focus on the serotonergic system that was shown to be involved in mammalian and insect sensory perception as well as learning and memory. Larval studies suggested that the serotonergic system is involved in the modulation of olfaction, feeding, vision and heart rate regulation. In a dual anatomical and behavioral approach we describe the basic anatomy of the larval serotonergic system, down to the single-cell level. In parallel, by expressing apoptosis-inducing genes during embryonic and larval development, we ablate most of the serotonergic neurons within the larval central nervous system. When testing these animals for naïve odor, sugar, salt and light perception, no profound phenotype was detectable; even appetitive and aversive learning was normal. Our results provide the first comprehensive description of the neuronal network of the larval serotonergic system. Moreover, they suggest that serotonin per se is not necessary for any of the behaviors tested. However, our data do not exclude that this system may modulate or fine-tune a wide set of behaviors, similar to its reported function in other insect species or in mammals. Based on our observations and the availability of a wide variety of genetic tools, this issue can now be addressed.