Kyoto, Japan -- Theft is almost universally considered an exploitative behavior. The conventional view underlying judicial systems is that the act of taking resources from others without offering compensation stems from moral and personal failings. As a result, these systems rely on the assumption that the threat or imposition of punishment will motivate criminal offenders to adjust their future behavior. Yet the outcome is not always success.
Despite the threat of severe penalties, the rate of theft recidivism, or repeat offenses, remains high at 70 percent within five years. These recidivists tend to fall into two distinct subgroups: kleptomania and instrumental theft. While individuals with kleptomania are overcome by uncontrollable urges to steal items of little to no monetary value, instrumental theft is inherently profit-seeking.
While it is clear that compulsive and irresistible urges drive recidivism for kleptomania, why instrumental thieves continue to reoffend despite facing severe punishment has remained a bit of a conundrum. This motivated a team of researchers at Kyoto University to investigate.
The scientists recruited three groups of participants: theft recidivists with a kleptomania diagnosis, theft recidivists without kleptomania, and a control group of individuals with no criminal history. The participants were asked to complete a psychological task known as the "4-arm bandit," in which they are repeatedly presented with four choices of fluctuating reward probabilities, where they had to select the highest payoff option each time. The task involves balancing decisions between exploration -- investigating unknown options that may yield high payoffs -- and exploitation -- settling for safe options with known payoffs -- in order to achieve the highest payoff.
During the task, the researchers used functional Near-Infrared Spectroscopy -- fNIRS -- to assess the participants' neural responses in the prefrontal cortex region of the brain. The team then used a reinforcement learning model to evaluate the data and compare the differences between the groups based on learning rate; in other words, they examined how efficiently the participants were able to adjust their actions based on reward feedback.
Surprisingly, the results revealed that theft recidivists without kleptomania are actually unable to adjust their behaviors based on feedback, indicating that impaired reinforcement learning may be preventing instrumental recidivists from changing their behavior to avoid the consequences.
"Given that instrumental theft is a profit-seeking behavior, reinforcement learning has been superficially believed to play a critical role, or even be abnormally enhanced in these individuals," says corresponding author Yukiori Goto. "Yet our findings have clearly demonstrated that recurrent instrumental theft involves neurobehavioral deficits."
These findings challenge the conventional views underlying the current judicial system and raise crucial ethical questions about how society should address crime. If certain criminal acts stem from neurobehavioral dysfunction, such impairments will need to be taken into consideration.
###
The paper "Lower Reward-Based Reinforcement Learning Underlying Explore-Exploit Decision Making in Non-Kleptomanic Theft Recidivists" appeared on 9 November 2026 in Frontiers in Psychology, with doi: 10.3389/fpsyg.2026.1997570
About Kyoto University
Kyoto University is one of Japan and Asia's premier research institutions, founded in 1897 and responsible for producing numerous Nobel laureates and winners of other prestigious international prizes. A broad curriculum across the arts and sciences at undergraduate and graduate levels complements several research centers, facilities, and offices around Japan and the world. For more information, please see: http://www.kyoto-u.ac.jp/en
Journal
Frontiers in Psychology
Article Title
Lower reward-based reinforcement learning underlying explore-exploit decision making in non-kleptomanic theft recidivists
Article Publication Date
9-Nov-2026
COI Statement
Author(s) CK was/were employed by Shogakukan Academy Co., Ltd. The remaining author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The author YG declared that they were an editorial board member of Frontiers at the time of submission. This had no impact on the peer review process and the final decision.