Our goal is to examine deliberative decision-making in rats from neurophysiological, cognitive, and mathematical perspectives. Deliberative decision-making entails behaviors in which agents explicitly consider multiple possibilities before acting. Early experiments identified behaviors in which a rat paused, looking back and forth, before acting (Tolman1948). This behavior occurred during learning and after changes in experimental contingencies, and was suggested to create expectancies of the consequences of the available choices, but the computational and experimental techniques needed to address this question were not available at the time. An interest in this behavior has been revived in the past few years based on new computational models of decision-making systems (Daw2005, Niv2006, Redish2007a, Redish2008), on the availability of new behavioral measures (e.g. Lauwereyns2006), and on recent discoveries of prospective encoding in the hippocampus (Wood2000; Johnson2007a) and other structures (Ramus2007, vanderMeer2009, vanDuuren2009). Our team brings together expertise in behavioral neuroscience, neurophysiology, computational neuroscience, cognitive psychology, and mathematics. Experiments will examine prediction, evaluation, and action-selection mechanisms in rats making deliberative decisions. Computational projects will examine mechanisms by which these systems can shift suddenly from consideration of one possibility to the other, based on models of chaotic attractors (Tsuda2001a).