This dispute led to the design of critical experiments, for instance, examining the nature of learning that occurs in the absence of the driving force of reinforcement. SCR7 cell line The classic example here was the observation that an animal left to explore a maze environment, without ever experiencing a reinforcing reward contingency, can nevertheless be shown to be

engaging in what is known as latent learning (Blodgett, 1929 and Thistlethwaite, 1951). Latent learning is “unmasked” when the animal is subsequently tasked to navigate toward a rewarded goal state in this same environment. Critically, pre-exposed animals show facilitation in learning relative to naive animals, suggesting

that the preceding nonrewarded exposure epochs foster the formation of a cognitive map that aids subsequent attainment of the rewarded goal location (Tolman and Honzik, 1930). Latent learning about outcomes is also observed in procedures such as the irrelevant incentive effect (Krieckhaus and Wolf, 1968). Consider two groups of animals trained when thirsty, but not salt deprived, to press a lever to get either water or a sodium solution. Members of the latter group are found to press this website more avidly than those of the former when subsequently salt deprived, even if lever pressing is in extinction (when no solution of either sort is provided). This shows that latent learning occurred in relation to the salt characteristics of the solution, even when it was

irrelevant in the context of the then prevailing motivational state. At the time of the early studies, it was not easy to quantify how complicated the latent learning tasks were that the animals were being asked to perform. These experiments substantially predated the invention of dynamic programming (Bellman, 1957), which helped formalize the whole domain. The resulting theory, and particularly a computational variant called reinforcement learning (Sutton and Barto, 1998), has underpinned much of the impact of computational modeling in the later generations of studies that found has resulted in a considerable sharpening of experimental design and analysis. In terms of behavioral control, a cognitive map can be seen as a representational template that enables an animal, through mental search, to find the best possible action at a particular state. Some indirect evidence about search came from what is termed “vicarious trial and error” (VTE), a class of behavior evident at choice points that is manifest as motor hesitations and repetitive looking back and forth (Muenzinger, 1938). VTE behaviors are not merely incidental, since animals that express more VTE behaviors turn out to be better learners (Muenzinger, 1938 and Tolman, 1938).