Download

Download Poster (445 KB)

Student Class

2029

Student Affiliation

Other

First Advisor

Alireza Soltani

First Advisor Department

Department of Psychological and Brain Sciences

Abstract

This research project aims to discover how a person’s brain continuously learns about its changing environment and how those memories dynamically alter their subsequent willingness to take risks. By testing 73 participants on a two-stage computer module, the study evaluates whether humans evaluate risk absolutely or relatively to their historical surroundings. In Stage 1, participants use reinforcement learning to track a hidden, shifting "lucky room" context. In contrast, in Stage 2, they face explicit two-option gamble cards in which they must weigh different point amounts and win percentages. Computational maximum likelihood estimation proved that a Subjective Utility (SU) model, which captures how the human brain subjectively shrinks the emotional value of giant numbers while non-linearly warping percentages, significantly outperforms basic, robot-like Expected Value (EV) models. Ultimately, by locking these stages together in an integrated framework, the project tests the hypothesis that the environmental values learned in Stage 1 act as a moving internal baseline or dynamic reference point that continuously reframes downstream risk calculations, offering mathematical insights into why decision-making can become rigid in psychiatric conditions like anxiety, OCD, and addiction.

Publication Date

2026

Keywords

decision-making, computational neuroscience, neuroeconomics, prospect theory, subjective utility, reinforcement learning

Disciplines

Applied Behavior Analysis | Cognitive Psychology | Experimental Analysis of Behavior

Risky Decision-Making Under Uncertainty

Share

COinS