The idea that there are parallel learning and decision-making systems in the brain is ubiquitous in neuroscience, where a division is often made between systems that support habitual behaviour and those that allow for more deliberative strategies. In one recent interpretation (Daw et al, 2005) this has been framed as a coexistence of “model-free” and “model-based” reinforcement learning strategies. With this project we try to elucidate the neural substrate of these two modes of behaviour by investigating how unilateral temporal lobectomy (surgical removal of the anterior temporal lobe) in human epilepsy patients influences learning.