One way to expand our research to more natural situations could be by changing the cost function to mimic an ecological survival game with perishable

outcomes. Such a paradigm would allow one to determine if subjects indeed follow a variance minimizing strategy and incorporate Paclitaxel cell line information about reward correlations. The recent financial crisis has amply demonstrated that even experts have difficulties regulating correlated risks in the financial domain and investors often deviate from rationality when making financial decisions (Daniel et al., 2002 and Kuhnen and Knutson, 2005). In contrast, we show here that individuals are adept at detecting and responding to correlations and appropriately selecting actions to minimize risk in an intricate learning task. Indeed, this exquisite sensitivity taps into an adaptive and evolutionary conserved ability of implicit neurobiological systems to learn environmental reward structure through trial-by-trial sampling; intrinsic behavior that might even supersede that of financial experts deciding about explicitly described statistics. Sixteen healthy subjects

(7 female; 18–35 years old) with no history of neurological or psychiatric illness participated in the study. Two additional pilot subjects from the lab were excluded from the final analysis, as they were already familiar with the hypotheses in the experiment. The study was approved by the Institute of Neurology (University College London) Research Ethics Committee. IWR-1 To investigate whether and how subjects learn the reward structure in the environment we designed a portfolio-mixing task in which knowledge of the correlation between two resource outcomes could improve Rolziracetam performance. Subjects’ task was to keep the combined output of two power

stations as stable as possible (i.e., minimize the variance of an energy portfolio) by mixing the fluctuating outcomes of these two individual resources. They accomplished this by adjusting weights that determined how the resources were linearly combined. A normative best performance is achievable in this task by finding a solution that directly depends on knowledge of the covariance structure of these resources, a task design that approximates a simple portfolio problem in finance (Markowitz, 1952). We presented the task to subjects as a resource management game that invoked a scenario whereby a power company generates fluctuating amounts of electricity from two renewable energy sources, a solar plant and a wind park. The resource outputs rsun and rwind were drawn as random numbers in every trial from distributions with a common mean M and variances σ2sun and σ2wind.