MIT Study Explores Flexible Thinking in Everyday Decision-Making

A recent study conducted by researchers at the Massachusetts Institute of Technology (MIT) has shed light on the cognitive strategies employed by humans when navigating complex decision-making tasks. Published in the journal *Nature Human Behavior* on June 12, 2025, the study emphasizes the human brain's remarkable ability to break down intricate problems into manageable subtasks, ultimately aiding in effective problem-solving.

The research, led by Mehrdad Jazayeri, a professor of brain and cognitive sciences at MIT and a member of the McGovern Institute for Brain Research, investigates the decision-making processes involved when individuals predict the trajectory of a ball moving through a maze. The study's findings indicate that humans utilize two primary reasoning strategies: hierarchical reasoning and counterfactual reasoning, which allow them to approach complex tasks efficiently.

According to Jazayeri, "When faced with simple tasks, such as categorizing objects, individuals typically perform exceptionally well. However, as tasks grow more complex, the absence of a singular optimal solution necessitates the use of problem-solving shortcuts, or heuristics."

The experimental design involved approximately 150 volunteers tasked with predicting the path of a ball that was hidden from view within a maze. The participants were required to make their predictions based on auditory cues that indicated the ball's position at various junctions. This task was intentionally devised to be complex enough to require the application of hierarchical and counterfactual reasoning, yet simple enough for the outcomes to be accurately measured.

The study revealed that when participants engaged with the task, they first employed hierarchical reasoning to make an initial prediction about the ball's trajectory. Based on the auditory feedback they received, they would then sometimes switch to counterfactual reasoning to revise their predictions. This flexibility in reasoning showcases the adaptive strategies that humans utilize when faced with uncertainty and cognitive limitations.

Co-authors of the study include Mahdi Ramadan, a PhD candidate at MIT, and Cheng Tang, a graduate student, with Nicholas Watters also contributing as a co-author. Their collaborative work highlights the brain's capacity to make rational decisions under cognitive constraints, a finding that has significant implications for understanding human behavior in various contexts, from daily decision-making to complex problem-solving scenarios.

The researchers also created a machine-learning model designed to mimic human decision-making. By imposing similar cognitive constraints on the model, they found that it began to adopt strategies akin to those employed by humans, further supporting the idea that human cognitive processes can be rationalized through the lens of computational constraints.

This research not only advances the understanding of cognitive strategies but also opens avenues for future exploration into how these decision-making processes operate within the brain. As Jazayeri notes, "The question remains: how do we problem-solve in a suboptimal way, chaining together clever heuristics to approach solutions?"

Moving forward, the research team intends to investigate the neural mechanisms underlying these heuristic strategies, aiming to delineate how humans shift between different reasoning approaches in response to varying cognitive demands. The implications of this study extend beyond neuroscience; they touch on fields such as artificial intelligence, cognitive psychology, and behavioral economics, providing a foundation for interdisciplinary exploration of decision-making processes.

This study was supported by various funding sources, including the Lisa K. Yang ICoN Fellowship and the Howard Hughes Medical Institute. As research in this area continues to evolve, further understanding of human cognitive flexibility could lead to enhanced methodologies in education, technology development, and psychological health.