Search

The ability to navigate a spatial environment strongly depends on how well individuals learn, represent and make use of their knowledge about the environment. In the past, research investigated these aspects separately and often in a virtual environment. The current work studied these three aspects of navigation in a real-real world setting to understand how humans navigate naturally in a complex, urban environment like London, UK. Of particular interest was to determine if there was evidence of hierarchical representations during route planning as found in previous behavioural, neuroscientific or computational studies. Most past studies have explored knowledge for simplistic environments or fragmented knowledge of real-world environments. By contrast, licensed London taxi drivers acquire a unique, almost perfect mental representation of the street network, the location of places and the traffic rules that apply to it. Here, the rare knowledge of these navigation experts was explored in three studies with novel approaches. First, to gain an understanding of the training process of unqualified taxi drivers, information from an interview with a teacher, training lessons and study material was collected, summarised and reported. A range of learning strategies was identified that was linked to theoretical, map-based learning and practical, in-situ experiences of London and pointed towards a segmented planning of routes through subgoal selection. Second, a potential mental segregation of London was studied with qualified taxi drivers through boundary drawings of specific London districts with a paper map to understand a potential hierarchical representation. Higher agreement was found for geographical structures and topically distinct districts surrounded by a linear, almost rectangular street network, whereas agreement was lowest for irregularly shaped districts with similarities to neighbouring areas. Finally, taxi drivers were asked to plan and then verbally recall each street they would take along routes between selected origin destination pairs. Audio recordings of these routes made it possible to relate the response times between individual streets to specific street network properties. The analysis using a linear mixed model indicated slower responses at upcoming turns and entering main roads, whereas boundary streets were recalled faster, as were finial streets when compared to initial street. No effects of Euclidean distance or detours were found. Observations from the training process indicate that a potential segregation of the environment, which might impact on later route planning, might be formed already through specific learning strategies. Faster response times for boundary streets support models in which planning is hierarchical. These findings extend past work on route planning in lab-based networks to real-world city street networks and highlight avenues for future research to explore and make use of real-world data.

Social contexts often change how people engage with and evaluate available rewards, leading to behaviours that defy simple rules of reward maximization. The current thesis aims to characterize some of the principles that underlie reactions to rewards obtained in a social context and formalize them in computational models. In study 1, I explore how social reward distributions change the hedonic and motivational value of rewards. The study shows that people are often demotivated and distressed by the unfairness of the distribution, and are less willing to work for their offered rewards even if they are the ones benefiting from the unfair situation. I introduce a model that characterizes the responses to reward distributions as a linear combination of statistical dispersion and rank ordering of the rewards and show that its predictions fit more closely to observed behaviour than many other alternative models suggested in behavioural economics and psychology. In study 2, I test how people form subjective judgments about reward distributions. The study demonstrates that subjective judgments are biased by personal position in the distribution, and violate several normative axioms used in economics. In study 3, I demonstrate the effect of the international distribution of rewards on life-evaluations: the study shows that life evaluations are not only sensitive to comparisons with citizens in one’s own country, but also to comparisons with people in other countries. The model characterizing the response to reward distributions as a linear combination of statistical dispersion and rank ordering again is shown to fit well-being data better than any other alternative. Study 4 focuses on the influence of the distribution of beliefs about oneself on preferences for feedback. It shows that people sometimes might prefer negative feedback, and describes heuristics and learning mechanisms that lead to this behaviour. The four studies presented in this thesis expand our knowledge of how external and internal social contexts change our experience with rewards. They introduce computational models that aim to formalize such contextual influences, contributing to a more mechanistic understanding of these effects.

When making decisions under uncertainty, one common view is that people rely on simple heuristics that deliberately ignore information. One of the greatest puzzles in cognitive science concerns why heuristics can sometimes outperform full-information models, such as linear regression, which make full use of the available information. In this thesis, I will contribute the novel idea that heuristics can be thought of as embodying extreme Bayesian priors. Thereby, an explanation for less-is-more is that the heuristics’ relative simplicity and inflexibility amounts to a strong inductive bias, that is suitable for some learning and decision problems. I will formalize this idea by introducing Bayesian models within which heuristics are an extreme case along a continuum of model flexibility defined by the strength and nature of the prior. Crucially, the Bayesian models include heuristics at one of the Bayesian prior strength and classic full-information models at the other end of the Bayesian prior. This allows for a comparative test between the intermediate models along the continuum and the extremes of heuristics and full regression model. Indeed, I will show that intermediate models perform best across simulations, suggesting that down-weighting information is preferable to entirely ignoring it. These results refute an absolute version of less-is-more, demonstrating that heuristics will usually be outperformed by a model that takes into account the full information but weighs it appropriately. Thereby, the thesis provides a novel explanation for less-is-more: Heuristics work well because they embody a Bayesian prior that approximates the optimal prior. While the main contribution is formal, the final Chapter will explore whether less is more at the psychological level, and finds that people do not use heuristics, but rely on the full information instead. A consistent perspective will emerge throughout the whole thesis, which is that less is not more.

This thesis examines the factors (e.g., task load, distractor features, or individual differences) that determine the extent to which individuals are distracted and the manner in which these factors interact. In particular, established theories of attention are applied to two understudied forms of distraction common to daily life: distraction by entirely task-irrelevant external stimuli and internal distraction from unintentional task-unrelated thoughts (i.e. mind-wandering). The experiments reported in this thesis establish a new measure of distraction by entirely task-irrelevant stimuli, drawing on the attentional capture literature to demonstrate that this form of distraction can occur in the absence of any top down attentional settings relating to the distractor or task features. It is also demonstrated that both of this form of distraction and also mindwandering can be modulated by the level of perceptual task-load – an established determinant of other forms of distractor processing. In this manner the thesis integrates two previously separate bodies of literature on selective attention and mind-wandering. In addition, individual differences in both internal and external forms of distraction are shown to be correlated, suggesting a common underlying trait influencing susceptibility to distraction both from internal sources (in the form of mind-wandering) and from external task-irrelevant distractor stimuli.