An ability to predict the consequences of one’s actions is the hallmark of adaptive intelligence. Learning involves the updating of predictions based on past experience, a process thought to be driven by dopaminergic neurons. Dopaminergic systems can also impart motivational control over learned behaviour. Drug-induced dopamine release can subvert these adaptive systems and link desire for addictive drugs with particular behaviours, environmental cues and motivational states.A fly learning to associate an odour with impending punishment requires synaptic input from the PPL1 cluster of dopaminergic neurons to mushroom body neurons, where olfactory memories are stored. A subset of PPL1 neurons is also responsible for satiety-dependent appetitive memory retrieval. The PPL1 cluster of dopaminergic neurons therefore represents a potential site where learning and motivational circuits intersect.I will use optogenetic and genetic methods to address three specific aims directed at understanding the learning algorithm and its subversion by addictive drugs. 1) I will establish transgenic lines providing genetic access to subsets of PPL1 neurons. 2) I will use an optogenetic reporter to monitor the activity of the PPL1 neurons in vivo, in order to test the hypothesis that PPL1 neurons encode the prediction errors used to update valuations during learning. The effect of drugs of abuse on this process will be determined. 3) I will use a single-fly behavioural assay to establish which cells are necessary, and which sufficient, to drive learning and to test the effects drugs of abuse on learning. A new learning based behavioural assay of addiction will also be developed for gaining access to the motivational component of addiction in Drosophila.