UC Irvine

Preattentive processing is an essential step of filtering information from the world into readibly parsable chunks that demand attentional resources. In a visual domain, this skill enables a person to easily devote resources to items that are captured by one or more attention filters. This phenomenon can be historically seen through differences in accuracy and reaction time amongst visual search tasks. However, identifying specifically what filters exist in human vision, how strong these filters are, how many filters there are, and what stimuli maximally activate each of these filters are questions that have largely been unanswered. This dissertation explores several experiments that have developed frameworks designed to begin answering these questions. Chapter 1 uses the centroid paradigm –- a psychophysical method aimed at discovering the visual influence of stimulus items through weighted center-of-mass calculations –- to explore what features drive discrimination amongst textures with equal energy spectra. We discover that large homogeneous regions, rather than high-order texture statistics, drive texture salience in this domain. Chapter 2 seeks to further understand the mental processes that underlie comparative salience computations for visual texture displays. We do so by proposing a new paradigm that compares two visual textures to a background uniform texture and to each other, which enables us to quantify the specific strength and tuning functions for mechanisms whose existence had been strongly suggested by existing literature. Chapter 3 again uses the centroid method to examine several atomic micropatterns consisting of various lines and line terminators and asks whether the human visual system contains mechanisms that are selective towards a subset of micropatterns. We find that certain micropatterns exhibit strong salience and “popout” effects, while others are difficult to attend to under similar conditions, suggesting that there exists a small number of mechanisms sensitive to these atomic patterns. Overall, these experiments serve as a “proof-of-concept” to study how preattentive processing affects a subset of textures and atomic patterns. The groundwork is then laid to explore many more texture types and feature dimensions (such as color) in the future.