Year: 2017
Type: dissertation
Abstract: The structural nature of the world often provides a clear guide to where sought objects are likely to appear, as well as the kind of objects they may repeatedly appear in the presence of. The relationship between the targets, distractors and the landscape provides context, which ensures efficient search. This thesis will explore the dynamics of how knowledge of the environment ahead will inform search on future presentation of those scenes, as well as explore how several factors between individuals (such as cognitive resources, or tendencies towards anxiety) may influence search and learning processes. This thesis reports three studies using a new eye movement experimental paradigm termed the repeated scenes search task (RSST). This task presented scenes taken on a route around a suburban neighbourhood as search arrays, while participants searched for target superimposed in naturalistic locations. The scenes were presented on 8 occasions in each experiment, and performance improved with number of repeats. In the experimental chapters the influence of scene order on search was examined with targets appearing in several contingencies with relation to scene identity and compared between the scenes appearing in a consistent or randomised order. Subtle benefits to search were found when scenes were presented in a consistent order. The influence of boosting WM and inducing a state of anxiety upon participant responses (via more efficient eye movements) were also examined. The impact of these findings upon the general literature and with regard to individuals searching in dangerous environments are discussed, with the key finding that attentional networks, working memory and a state of anxiety are important factors to consider in search through familiar environments.
Author: Oliver Tew
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2799805041