Contributions of Motility to Large-scale Search and Navigation

Abstract

Navigation and search are fundamental daily activities, however, their ubiquity belies their complexity. These behaviours are reliant upon multiple underlying cognitive processes including our ability to reliably employ viewpoint invariant allocentric and viewpoint dependent egocentric cues. The ability to use these sources of information have been studied extensively in the human spatial cognition literature, largely on desktop PC platforms. Whilst this provides researchers with an immense amount of experimental control, participants remain static, differing fundamentally to the demands of navigation and search in real-world contexts. This thesis presents a series of experiments that interrogate place learning and probabilistic cueing on both desktop PC and immersive virtual reality (VR) platforms. Chapter 2 reports the development of a novel place learning task based on the Blue Velvet Arena task (BVA; Kalová, 2005) for both immersive VR and desktop PC platforms. The task included experimental conditions that interrogated the ability to accurately use allocentric or egocentric information in isolation, or both types of information in conjunction. Across a series of experiments, participants appeared to be able to more accurately identify a target location in the immersive VR task, relative to performance on the desktop PC equivalent. The same immersive VR system was employed in Chapter 3 for a replication of Smith et al.'s (2010) probabilistic cueing task, in which participants learned the statistical contingency underpinning a target's spatial distribution. Chapter 3 replicated results reported by Smith et al. (2010) and provide further evidence that allocentric probabilistic cueing may be contingent on supplementary visual information. The experiments reported in this thesis demonstrate the utility of immersive VR as a tool for interrogating human spatial cognition in large-scale environments. Importantly, the cross-platform inconsistencies revealed by the place learning task demonstrate the need to comprehensively test theories of human spatial behaviour across a range of conditions and platforms.