The outcome-representation learning model: impairments in decision-making in adolescents with excess weight

Abstract

Impairments in decision-making have been suggested as a predisposing factor to obesity development. Individuals with excess weight display riskier decisions than normal weight people. Furthermore, adolescence is a period of life in which risky behavior may increase. We aimed to investigate decision making applying the Outcome-Representation-Learning (ORL) model to the Iowa Gambling Task (IGT) in adolescents with excess weight. Twenty-nine excess weight and twenty-eight normal weight adolescents, classified according to their age-adjusted body mass index (BMI) percentile, participated in the study. Decision-making was measured using the IGT. A Bayesian computational ORL model was applied to assess reward learning, punishment learning, forgetfulness, win perseverance and deck perseverance. The IGT net score was lower in excess weight than normal weight adolescents (β = 2.85; p < .027). Reward learning (95% HDI [0.011, 0.232]) was higher, while forgetfulness (95% HDI [− 0.711, − 0.181]) and deck perseverance (95% HDI [− 3.349, − 0.203]) were lower, in excess weight than normal weight adolescents. Excess weight adolescents seemed better at learning the most rewarding choices and showed a random strategy based on reward and novelty seeking. Consequently, excess weight adolescents made more disadvantageous selections, and performed worse in the IGT.