Neural mechanisms for reducing uncertainty in 3D depth perception

[摘要] In order to navigate and interact within their environment, animals must process and interpret sensory information to generate a representation or ‘percept’ of that environment. However, sensory information is invariably noisy, ambiguous, or incomplete due to the constraints of sensory apparatus, and this leads to uncertainty in perceptual interpretation. To overcome these problems, sensory systems have evolved multiple strategies for reducing perceptual uncertainty in the face of uncertain visual input, thus optimizing goal-oriented behaviours. Two available strategies have been observed even in the simplest of neural systems, and are represented in Bayesian formulations of perceptual inference: sensory integration and prior experience. In this thesis, I present a series of studies that examine these processes and the neural mechanisms underlying them in the primate visual system, by studying depth perception in human observers. Chapters 2 & 3 used functional brain imaging to localize cortical areas involved in integrating multiple visual depth cues, which enhance observers’ ability to judge depth. Specifically, we tested which of two possible computational methods the brain uses to combine depth cues. Based on the results we applied disruption techniques to examine whether these select brain regions are critical for depth cue integration. Chapters 4 & 5 addressed the question of how memory systems operating over different time scales interact to resolve perceptual ambiguity when the retinal signal is compatible with more than one 3D interpretation of the world. Finally, we examined the role of higher cortical regions (parietal cortex) in depth perception and the resolution of ambiguous visual input by testing patients with brain lesions.

[发布机构] University:University of Birmingham;Department:School of Psychology

[关键词] B Philosophy. Psychology. Religion;BF Psychology