Adaptive foraging in humans: behavioural flexibility under constraints, volatility, and structural demands

Foraging is a universal adaptive behaviour that is fundamental to the survival of most species (Mobbs et al., 2018; Hayden, 2018). At its core, foraging involves a continual trade-off between exploration, seeking new opportunities or, in so-called patch-leaving problems, moving to new areas, and exploitation, the efficient use of currently available resources. This exploration–exploitation dilemma lies at the heart of Optimal Foraging Theory (OFT) (Charnov, 1976), where it is formalised as a problem of maximizing long-term reward under ecological constraints. Because resolving this trade-off requires sensitivity to environmental structure, learning from experience, and flexible adjustment of behaviour, it is widely regarded as a hallmark of cognitive flexibility. While extensively studied across the animal kingdom (Stephens and Krebs, 1986) and despite its growing interest, foraging frameworks are increasingly being applied to human decision-making and im- portant questions remain about how foraging behaviour unfolds in environments that are complex, uncertain, volatile, or structured. In particular, it is still unclear which cognitive mechanisms enable humans to flexibly adapt their behaviour when environmental constraints, changing contingencies, and structural demands jointly shape behaviour. This thesis brings together three studies that, although differing in their specific experimental manipulations, are unified by a single overarching research question: how do humans flexibly adapt their foraging strategies across ecological contexts characterized by resource constraints, environmental volatility, and structural orga- nization, and which cognitive mechanisms account for systematic deviations from optimal behavior in these contexts? Studying foraging provides a powerful lens through which to investigate flex- ible, goal-directed behaviour in real-world contexts. Just as a lion must decide whether to pursue a gazelle, humans face comparable foraging-like dilemmas in daily life: selecting a job candidate, deciding whether to invest in a relationship, choosing a restaurant among many alternatives, browsing e-commerce platforms for the best deal, or even deciding which movie to watch on a streaming service. In his spirit, the studies presented in this thesis are framed through the everyday-life metaphor of a supermarket trip (Schlender et al., 2024), one of the most familiar human foraging situations, where aisles can be conceived as analogous to resource patches. Imagine a shopper entering a supermarket with limited time and uncertain knowledge of which aisles contain the most profitable items. In this kind of scenario, the shopper must decide whether to continue collecting items in the current aisle or explore another, gradually learning which areas are most rewarding and how to best allocate their time. Analogously, in our first study participants navigated a four-area environment to collect coins from treasure boxes, adapting to resource distributions and time constraints. Their strategies improved with experience, approaching, though not fully reaching, the performance of an optimal reward-maximizing agent. Now, consider a more dynamic supermarket, with the location of the most profitable items shifting according to a hidden alternating rule (e.g., fruit and veg- etables on even weeks, meat and fish on odd weeks). This scenario introduces environmental volatility, where contingencies change over time and require contin- ual monitoring and adaptation. Similarly, in our second experiment, the location of rich areas varied according to a hidden rule. Most participants detected this volatility and flexibly adjusted their strategies, although individual differences emerged: participants with higher autistic traits, measured by the Autism Spectrum Quotient (AQ) (Baron-Cohen et al., 2001), exhibited slower adjustments following rule changes than transient rule violations. Finally, imagine two supermarkets with contrasting layouts: one structured, with items organized by category (e.g., fruit and vegetables in one aisle, meat and fish in another), and another unstructured, with items scattered randomly. In the structured store, shoppers may adopt hierarchical plans (e.g., collect vegetables first, then meat), whereas in the unstructured store, they may simply follow the shortest path to nearby items. Likewise, in our third study participants collected coloured spheres in a grid-like virtual environment, adapting their strategies to the environmental structure: hierarchical approaches predominated in structured layouts, while shortest-path strategies were favoured in unstructured ones. Together, these studies show that human foraging is highly adaptive but system- atically shaped by ecological demands and cognitive constraints. People flexibly adjust their behaviour to resource distribution, time limits, volatility, and envi- ronmental structure, yet do not always conform to normative optimal models. These deviations highlight the role of learning, cognitive limitations, and individual differences in shaping adaptive decision-making. By integrating evidence across constraints, volatility, and structural demands, this thesis advances our understand- ing of how humans navigate complex, uncertain, and dynamic environments.