The Becker-DeGroot-Marschak method (BDM) is an auction-like mechanism widely used in behavioural economics, marketing research, and, more recently, in neuroimaging studies of human decision making. The BDM has never been used with animal subjects before, yet its application in monkeys would allow for comparison of studies across species while providing a direct measure of what a reward is worth to a monkey in a single experimental trial.

In the BDM, a subject is given a budget with which they can place a bid for some reward, and a computer then randomly selects a competing bid. If the subject’s bid is higher than the computer’s bid then the subject pays an amount equal to the computer’s bid, receives the reward object, and gets to keep the remaining budget. If the subject’s bid is lower than the computer’s bid, the subject does not gain the reward object but retains the entire budget.

To adapt the task for monkeys, two rhesus macaques were taught to use water as a budget, and to use a joystick to place a bid in terms of this budget for different volumes of fruit-juice reward.

The BDM ensures that the subject’s optimal action is to place a bid equal to their value for the reward-object. This property of truthful value revelation is the BDM’s most important feature in the context of value-based decision making.

Currently, the only method of eliciting a monkey’s value for one reward in terms of another depends upon inference of the magnitudes at which the two rewards are chosen with equal probability. Using this ‘binary-choice’ method, many trials are needed to infer a single value: pairwise comparisons of many different magnitudes must be made and choices of each pair must be repeated so that the probability of choosing a reward can be estimated.

In contrast, the BDM provides a direct measure of the monkey’s value for the reward as they explicitly state this value on each trial by selecting an equivalent bid. Therefore, the BDM more efficiently utilises the limited time in which a monkey’s behaviour can be assessed in each experimental session, as animals lose the motivation to participate when they become sated.

The thesis summarised here describes the training and performance of two rhesus macaques on a novel version of the BDM, specifically designed for a subject that cannot be instructed on the optimal strategy. The technical steps and intermediate tasks that are needed to train a monkey to flexibly place bids by operating a joystick are also detailed, as well as the development of different versions of the task over three years of testing. The results of the final version of the BDM are then presented for both monkeys, showing rational bidding behaviour consistent with an understanding of the method’s contingencies. Theoretical concerns and limitations of the BDM in such a setting are also discussed and the thesis outlines how future experiments can make use of and adapt this version of the BDM for neuronal recording experiments.