Neuronal Mechanisms of Decision-Making

Human decision-making has been the subject of research in many disciplines including psychology, economics, and neuroscience. Over the last 50 years, different mathematical models of decision-making were proposed such as the race model, DDM, and LCA. Twenty years ago, XiaoJing Wang developed a physiologically realistic model based on a recurrent attractor network with leaky integrate-and-fire neurons. I have developed an alternative implementation of the recurrent attractor network model based on binary neurons. We shall discuss the construction of this model and its features. I will show that the model predictions are consistent with the results of dot motion discrimination experiments. We shall also discuss two applications of the model for studying neuronal mechanisms of decision-making: 1) the computational analysis of speed-accuracy tradeoff (my paper on this subject has just been accepted for publication in Scientific Reports), and 2) modeling the probability weighting function postulated by the Prospect Theory (Kahneman and Tversky).