Purchase this article with an account.
Madeleine Sharp, Jayalakshmi Viswanathan, Linda Lanyon, Jason Barton; Reward probability and magnitude in saccadic decisions under risk: measuring bias and sensitivity to expected value. Journal of Vision 2011;11(11):210. doi: 10.1167/11.11.210.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Objective: Our goal was to evaluate how healthy subjects integrate information of reward magnitude and the likelihood of reward, to determine their threshold sensitivity to differences in value and their choice biases. Background: There are few clinical tests of decisions under risk. The most well-known, the Iowa Gambling task, assesses the ability to learn cumulative probabilities of gain or loss and to forego large rewards for smaller ones. However, tests that characterize sensitivity to expected value and biases between reward magnitude versus probability may provide insights in conditions with anomalous reward-related behaviour. Design/Methods: Twenty subjects were required in 170 trials to choose between two explicitly described prospects, one having higher probability of reward but lower magnitude of reward than the other. The sizes of reward and the degree of probability were varied so that the difference in expected value between the two prospects varied from 3% to 23%. We first plotted choice as a function of expected value. Second, we used Prospect Theory to evaluate choice as a function of perceived value, using an exponential function for perceived reward utility and a single-parameter Prelec function for perceived probability. Results: Subjects showed a threshold sensitivity of 9.0% difference in expected value. Regarding choice bias, we found a ‘risk premium’ of 9.4%, indicating a slight tendency to choose higher probability over higher reward. Prospect Theory analysis showed that this risk premium is the predicted outcome of non-linearities in the subjective perception of reward value and probability. Conclusions: This simple test provides a robust measure of discriminative value thresholds and biases in decisions under risk. Prospect Theory makes predictions about choice patterns when perception of reward or probability is anomalous, as may occur in populations with dopaminergic or striatal dysfunction, such as Parkinson's disease and schizophrenia.
This PDF is available to Subscribers Only