| Summary: | Object perception is necessary to our understanding of the visual world, yet its neural mechanism remains poorly understood. The goal of this dissertation is to shed light on this mechanism. Current computational models of object perception suggest that regions on opposite sides of a shared border compete, with the winner perceived as the shaped object and the loser as its locally shapeless background (or ground). Recent behavioral work indicates that the result of this competition is suppression of the ground at the level of object shape--a finding not predicted by models. Here, I present three studies that extend this previous research on ground suppression as a mechanism by which object perception is accomplished. I first show that the amount of suppression applied to the ground depends on the amount of competition for object status (Salvagio, Cacciamani, & Peterson, 2012). I then provide the first neural evidence of ground suppression from shape-level competition at both high and low levels of the visual hierarchy, with the latter arising from top-down feedback (Cacciamani, Scalf, & Peterson, submitted). Finally, I show that semantic information pertaining to the ground is accessed prior to the assignment of object status, but unlike shape information, is not suppressed (Cacciamani, Mojica, Sanguinetti, & Peterson, 2014). Together, the three studies that comprise this dissertation demonstrate that ground suppression arising from shape-level competition underlies object perception. This research contradicts traditional theories stating that objects are processed unidirectionally through the visual system in a single feedforward pass; instead, it supports theories of object perception entailing dynamical feedforward and feedback processes.
|
|---|
