Abstract

This thesis tested a model of how brain reward circuitry operates. The proportion of a subject's time that it is willing to dedicate to performance of an operant task (time allocation) depends on the strength and cost of the reward. Plotting time allocation as a function of pulse frequency and price yields a three-dimensional structure called the "mountain." According to the model tested by the experiment, the mountain reflects processing of information about reward strength and price in a multi-stage network. An initial stage translates impulse flow in the directly activated neurons into a neural signal representing reward intensity. A later stage combines this intensity signal with the subjective mapping of reward price to yield a net payoff, and the final stage translates payoffs into behaviour. To understand how manipulations that alter operant performance produce their effects, it is necessary to determine at which stage(s) of processing they act. The primary goal of this experiment was to test the multistage model by establishing whether a manipulation that acts on the initial stage shifts the mountain along only one of its axes, as predicted. The results support the principal prediction of the mountain model. The demonstration that the mountain can shift along one axis without shifting along the other supports the notion that the integrative model on which the mountain is based can distinguish manipulations that affect the reward system at different stages.