Year: 2018
Type: dissertation
Abstract: A fundamental question of cognition and perception has asked how the human brain represents the local visual environment. Investigations seeking to answer this question have often explored the contributions of spatial information, such as geometry and layout, in forming a holistic percept of the visual world, yet a growing body of work suggests multiple properties may be crucial to the underlying neural mechanisms supporting scene recognition. This dissertation aims to examine the role of common and ubiquitous visual features, described here as geometry and surface properties, in shaping neural representations of the visual environment across space and time. Investigations in Chapter 2 sought to elucidate the weighting of geometry and surface properties in object and scene perception, and demonstrated equal sensitivity in the scene-selective parahippocampal place area for these features within a scene, and greater sensitivity to the surface information of an object, over its geometry. In Chapter 3, the interactions of these properties with goal-directed behaviour and higher-order scene attributes were examined, with results indicating that scene-selective cortex constructs a flexible representation of the visual environment by integrating diagnostic visual information with task context. Following this investigation, Chapter 4 examined these features within the temporal domain, wherein dissociable neural activity patterns emerged demonstrating how these features may diverge across time to influence the time course of scene perception. Together, these results demonstrate that multiple visual properties shape neural representations responsible for successful scene recognition, and these properties interact with goal-directed behaviour, and unfold across time, to influence our perception of the visual world.
Source: TSpace (University of Toronto)
Author: Matthew X. Lowe
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W2911406961