2 - (SYM 7) Computational Modeling to Connect Decision Making Under Uncertainty with Reward Processing and Anxiety

Decreased reward processing can impair reward-based decision making, but it is unclear whether treatments targeting reward processing can improve decision making. In the present study, we use a novel experimental design based on published work (Zbozinek, et al., 2021) in which we computationally model facets of decision making under uncertainty. Socially anxious and anhedonic participants (N=33) complete the decision-making task before and after receiving therapy to enhance reward processing or relaxation (control). Preliminary data from an ongoing randomized controlled trial shows model accuracy (mean=79.4%) similar to published work (70%-80%; Zbozinek, et al., 2021). Pre-treatment (N = 33) analyses show individuals with decreased reward processing (Positive Valence Systems Scale) assigned lower value to ambiguous gain magnitudes (r = .382, p = .028), higher value to ambiguous loss magnitude (r = -.357, p = .042), and lower perceived likelihood of receiving a gain (r = .303, p = .086) or loss (r = .370, p = .034) that had ambiguous likelihood of occurring. Using the Depression, Anxiety, and Stress Scale, anxious individuals had a significant risk aversion (i.e., favoring certain vs probabilistic outcomes) for gains (r = -.476, p = .005) and marginal risk preference for losses (r = -.31, p = .079), assigned lower value to ambiguous gain magnitudes (r = -0.528, p = 0.001), and lower perceived likelihood of receiving a gain that had ambiguous likelihood of occurring (r = -0.411, p = 0.018). Comparing post- vs pre-treatment (n =16), participants with larger increases in reward processing assigned lower value to ambiguous loss magnitudes (r = -.577, p = .019) and had marginally lower unambiguous loss aversion (r = -.460, p = .073) and risk preference for unambiguous losses (r = -.466, p = .069). Participants with larger decreases in anxiety showed increases in unambiguous loss aversion (r = -.514, p = .042) and decreased perceived likelihood of receiving a gain that had ambiguous probability of occurring (r = 0.648 p = 0.007). In sum, reward processing is associated with optimism with gains and losses both at pre-treatment and resulting from treatment. Anxiety is associated with pessimism with gains and losses at pre-treatment, though increases in anxiety after treatment show some optimism. Results from this study would improve our understanding of decision-making tendencies, emotionality, and whether therapy that affects emotionality also changes decision making. This could improve targeted treatment.