Value learning studies typically involve one of two forms of feedback. Learning could be guided by absolute feedback (specific reward values) or relative feedback (binary trial outcomes). This thesis compares both feedback types through two comparable four-day value learning tasks designed in a previous study. The question was whether distinct neural correlates of reward prediction errors, most commonly reported in the hippocampus and basal ganglia for the absolute and relative feedback task respectively, could be found in the neural data. Choice behavior was first modelled through reinforcement learning approaches, and prediction errors from the model were used to inform neural correlates. The nucleus accumbens (ventral striatum) was found to be correlated with signed prediction errors in both tasks on the first day of the learning task and on the second day for the relative feedback group's neural data only. There were no robust effects of signed prediction errors in the hippocampus for either feedback group. However, the hippocampus was correlated with unsigned prediction errors more significantly in the relative than absolute feedback task on the first day.