Many everyday decisions have to be made in the face of uncertainty about the eventual outcome of the chosen action. Such decisions are strongly influenced by an individual's risk attitude. Risk attitude is flexible and depends on contextual factors, such as whether the gamble outcome represents a potential gain or a loss, and the momentary wealth level. Impairments in the ability to properly assess risk can lead to severe behavioral disorders including various addictions and pathological gambling as well as an increased tendency toward criminal behavior. Such self-defeating behavior creates an enormous medical and economic toll on the individual as well as on society. The long-term goal of this research project is to understand the neural mechanisms controlling risk attitude. Human imaging experiments and lesion studies suggested that two functionally different, but connected networks guide decisions under risk. One `risk-attitude' network monitors the contextual factors that influence risk attitude. A central node within this risk-attitude network is anterior insular cortex. The risk-attitude network signals the momentary value of seeking or avoiding risk to a second `risk-decision' network, centered on the lateral and medial frontal cortex, which represents the option and action values of risky and sure (risk-free) options and selects a particular action. Our central hypothesis is that: (1) Anterior insular cortex (AIC) monitors the behaviorally salient factors that modulate risk-attitude. The neuronal signals serve as input variables into the risky decision process. They are likely represented in a non- spatial reference frame, not linked to specific actions. These signals guide activity in lateral prefrontal cortex (LPFC). (2) LPFC uses the risk-attitude-relevant signals to estimate and compare the value of the risky and sure option. These transformed signals guide activity in supplementary eye field (SEF), the oculomotor subsection of the medial frontal cortex. (3) SEF uses the option value input from LPFC to generate action value signals that reflect the momentary contextual risk-attitude and guides the final saccade action selection process, which indicates the choice between seeking and avoiding a risky option. We have developed the token- based gambling task, an animal model of context-dependent risky decision making. In this task, the monkey has to acquire a number of tokens over multiple trials to obtain reward by making decisions under risk. The trial outcomes can either be a gain or loss of tokens. Behavioral data show a clear effect of both gain/loss context and currently owned token number on the monkey's risk attitude. Using the token-based gambling task, we can use a combination of recording (Aim 1) and reversible inactivation (Aim 2) to test whether and how neural activity in AIC, LPFC and SEF is causally involved in risk-related behavior. These experiments provide a novel approach to understanding the competition between risk-seeking and risk-avoidance behavior at the neural level. Understanding the neural basis of variations in risk-related behavior will provide a road map for precise therapeutic interventions and early diagnosis of pathological risk-seeking behaviors.
Impairments in the ability to properly assess risk can lead to severe behavioral disorders including various addictions and pathological gambling as well as an increased tendency toward criminal behavior. Our experiments provide a novel approach to understanding the competition between risk-seeking and risk- avoidance behavior at the neural level. Understanding the neural basis of variations in risk-related behavior will provide a road map for precise therapeutic interventions and early diagnosis of pathological risk-seeking behaviors.
