Using neuroscientific methods to study human decision-making is a new and exciting research area. Today it is dominated by observational approaches driven by the fMRI and other brain-imaging technologies, but it already links empirical data with concepts and findings from established paradigms such as utility theory and prospect theory. Developed in parallel over many decades, mathematical neuroscience can now offer a next generation of theories that help better understand the cognitive

more »

... processes in general and those involved in economic choices in particular. However, the methods and techniques needed to implement sophisticated neuroscientific models for studying observational or experimental data are still in their infancy. This paper presents one effort in that direction and shows how the mathematical theory of reflex conditioning by Grossberg and Schmajuk [ 1 ] can account for important aspects of people's decisions in a laboratory experiment. A stochastically calibrated neurocomputational model emulated the cognitive mechanism for choosing one economic option among four, and was successful for 70% of the subjects in a sample of 80 people. Against a baseline probability of 0.25, its prediction rates for 9% of the people exceeded 0.75, and ranged between 0.37 and 0.70 for approximately 60% of the people. These results were very robust both for the model's impressive achievements as well as for its lackluster performances. We hypothesize that it predicted well only those persons who took decisions intuitively (as understood by Kahneman [ 2, 3 ]), and failed with those who made efforts to develop a decision strategy. We anticipate our study to be followed by further steps in discovering the exact cognitive mechanisms of economic choice.