During Pavlovian conditioning, dopamine (DA) neurons learn to predict the expected time of reward occurrence relative to a conditioned stimulus and encode error in this prediction via bidirectional changes in firing behavior, a signal that is believed to be analogous to the reward prediction error term in the temporal difference model of machine learning. The mechanism by which DA neurons acquire and store temporal information is incompletely understood but likely reflects long-term synaptic plasticity. With repeated exposure to a stimulus-reward contingency, the burst response at the time of primary reward delivery decays as a new burst time-locked to the reward-predictive cue develops. Omission of an expected reward causes a transient decrease in DA neuron firing at the predicted time of reward presentation, corresponding to a negative prediction error. The mechanism by which DA neurons cease to respond to the primary reward in Pavlovian conditioning is unresolved, as is the mechanism of negative prediction error. I use whole-cell patch clamp techniques to investigate burst-timing dependent plasticity of GABAergic inputs to DA neurons as a potential temporal marker of expected reward.