Decision making and learning under uncertainty in aging

Yale University, New Haven CT

Investigators

Linked publications, trials & patents

Abstract

While the proportion of elderly individuals in the population is constantly rising, considerable evidence suggests that older adults make decisions that may be detrimental to their financial wealth, health and general well-being. All of these decisions are made in a highly uncertain and ever-changing environment. Laboratory experiments demonstrate two age-related behavioral changes, which may contribute to the observed suboptimal behavior. One line of research suggests that attitudes towards uncertainty change with age, in both the financial and the medical domains. A separate line of research suggests that the ability to flexibly learn from feedback diminishes with age. What we do not know is whether and how age-related variations in learning ability are linked to age-related changes in uncertainty attitudes. In fact, we know very little, even in young adults, about the relationship between learning ability and uncertainty attitudes. This is surprising, because the primary goal of learning is to reduce uncertainty. The individual's ability to lower uncertainty by learning may therefore affect how they view un-certainty. At the same time, the individual's attitude towards uncertainty may influence their ability to learn. It is important to understand the relationship between learning ability and uncertainty attitudes, especially in aging, because this understanding will inform the development of successful interventions and decision aids. In this application we propose to study age-related changes in both uncertainty attitudes and learning ability, using behavioral tasks that we have developed, in conjunction with functional and structural MRI. We plan to examine monetary and medical decisions under uncertainty in the absence of learning, and simple appetitive reversal learning in the absence of choice, in adults between the ages of 18 and 90. This will enable us to examine gradual changes in uncertainty attitudes and learning ability, and to identify links between the two. In Aim 1 we will characterize age-related changes in uncertainty attitudes and identify structural and functional changes that reflect those behavioral changes. In Aim 2 we will characterize age-related changes in reversal learning and link those to neural changes. Finally, in Aim 3 we will tie the two together by having participants complete both tasks. The expected outcomes of the proposed work are twofold. At the behavioral level, we expect to identify, for the first time, links between individual differences in learning ability and individual differences in uncertainty preferences. At the neural level, we expect to identify structural features and functional mechanisms that underlie these individual and age-related differences, thus providing computational constraints on these processes. We expect these findings to guide the future development of decision aids and behavioral interventions that will enable older adults to make better financial and healthcare decisions in an uncertain, volatile world.