The mushroom bodies of insect brains are essential for memory formation. The sensory input region of the mushroom bodies, the calyx, processes information received from stereotypic olfactory input channels, the Projection Neurons (PNs). Mushroom body neurons, Kenyon cells (KCs), integrate these inputs through a combinatorial coding mechanism that allows discrimination among them. The calyx is innervated by extrinsic neurons that regulate its activity; these include two octopaminergic (OA) SEZ Ventral Unpaired Median 1 (sVUM1) neurons that are presynaptic in the calyx and antennal lobes (AL), and postsynaptic in the suboesophageal zone (SEZ). OA, functionally homologous to noradrenaline in mammals, is also a mediator of behavioural state in insects. The aim of my work was to understand the role of the OA neurons in the calyx circuitry, by testing their behavioural roles, testing for their effects on calyx activity, and characterising their inputs. Since the calyx is the site where the mushroom body can discriminate among different odours, I tested whether OA input might affect olfactory discrimination during learning. Building on preliminary results in our lab, I found that optogenetic activation of five neurons in the SEZ, including the two sVUM1 neurons, sVUMmd1 and sVUMmx1, impairs discrimination among similar but not dissimilar odours. To develop optogenetic stimulation of larval brains as a tool for investigating functionality of anatomical circuits, I used ChR2-XXL to activate the calyx-innervating inhibitory neuron APL, and recorded Ca2+ responses in KCs. I then applied this strategy to activate OA neurons, and observe the responses in projection neuron (PN) terminals or KCs. I found a tendency towards a potentiation of PN input to the calyx, but no effect on KC activity. By analysing the trajectories of sVUM1 neuron 1st-order upstream neurons in the Drosophila first-instar (L1) connectome, I found three classes of neuron synapsed onto the sVUM1 neurons in their dendritic region within the SEZ: one class of neurons originated in the protocerebrum; a second class were local SEZ interneurons with cell bodies and arborizations in the SEZ, and a third class originated in the Ventral Nerve Cord (VNC). To understand the upstream connectivity of the neurons directly synapsing onto the sVUM1 neurons, I used a criterion of following the highest number of synapses between neurons, to identify strongly connected pathways upstream of the sVUM1 neuron dendritic trees in the SEZ. I identified one pathway connecting the MB output neuron (MBON-i1), of the mushroom body medial lobe, to the sVUM1 neurons. I also identified a sensory input pathway originating in the pharyngeal nerve with the sensory neuron MN-L-Sens-B1-ACpl-01, potentially transmitting processed sensory input to the sVUM1 neurons. I screened for GAL4 drivers that target neurons identified from the connectomics data. I found potential GAL4 lines for two of the first upstream neurons in these pathways, MB2IN-19 in the protocerebrum and MB2IN-104, a SEZ local neuron, in the third-instar (L3) larva. My behavioural analysis has shown that activation of a small subset of neurons including the sVUM1 neurons impair fine discrimination of odours during learning, and therefore might regulate KC activity in the calyx. However, imaging experiments did not succeed in ascribing this behavioural effect to specific connections of the sVUM1 neurons. My connectomic analysis suggests that sVUM1 neurons are not signalling direct sensory inputs, but that their dendritic regions receive and process a range of inputs: inputs from the protocerebrum; local circuits that can process the input to sVUM1s, and inputs from the VNC. One model that emerges from my studies is that sVUM1s are potentially regulated by learned MB output. I generated tools for functional calcium imaging, and these can facilitate the dissection of the circuits that activate sVUM1s, and link behavioural context to specific neurons.