LAUSR

Individuals with anxiety disorders tend to focus on threatrelated information and are more likely to interpret ambiguous information as negative than as positive (1). Therefore, it is natural to assume that when anxious individuals make economic decisions, they preferentially attend to potential negative outcomes rather than positive outcomes. Consistent with this notion, a few studies have documented reduced economic risk taking in anxiety (2). However, enhanced perception of potential losses—or loss aversion, as economists call it—is only one of several basic cognitive processes that may suppress risk taking. Prior studies have employed experimental paradigms that did not allow independent evaluation of each of these cognitive processes. In a study reported in this issue of Biological Psychiatry, Charpentier et al. (3) take a behavioral economic approach to decision making under risk in generalized anxiety disorder (GAD). Their primary goal is to distinguish between the effects of loss aversion and risk aversion (4) on choice behavior. To understand these effects, let us consider the choice of whether to accept a mixed lottery —a lottery that offers a potential gain but also a potential loss. Figure 1A presents such a mixed lottery with 50% chance of winning $8 and 50% chance of losing $4. The subjective value, or utility, of the lottery depends not only on these amounts and probabilities but also on the individual’s attitudes toward these amounts and probabilities. Figure 1B–D presents the utility curves for gains and losses of three different individuals and marks the utilities of an $8 gain (green) and a $4 loss (red). Utility functions are typically concave in the gain domain and convex in the loss domain, indicating diminished sensitivity for increased value. In the gain domain, this diminished sensitivity is translated into risk aversion. Participant 1 (Figure 1B) exhibits slight risk aversion in the gain domain, obtaining just over 5 utility units from a gain of $8. This participant places the same weight on gains and on losses and thus expects the utility of a $4 loss to equal half of that of the $8 gain, with a negative sign. The lottery’s expected utility for this participant therefore is high, and she is likely to accept the lottery. Participant 2 (Figure 1C) exhibits a similar degree of risk aversion. For this participant, however, losses loom larger than gains, such that his negative utility from a $4 loss is quite high, leading to an overall negative expected utility for the lottery. Thus, increased loss aversion may drive reduced risk taking in this participant compared with Participant 1. Participant 3 (Figure 1D) is also less likely to accept the lottery, but for a different reason. Like Participant 1, Participant 3 weighs gains and losses equally. This participant, however, exhibits increased risk aversion (reflected in a more curved utility function in the gain domain), which decreases the utility of the gain and reduces the overall desirability of the lottery. Charpentier et al. (3) had both unmedicated anxious individuals and healthy individuals make a series of choices with