A Caltech Library Service

Two themes in perceptual ecology : visual attention and awareness


Kolb, F. Christopher (1996) Two themes in perceptual ecology : visual attention and awareness. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/tv8e-g179.


The nature of consciousness poses the central problem for both neuroscience and philosophy of mind: how is the brain constructed to enable the mind, a collection of perceptual, emotional, and intentional functions that evolve around a fleeting stream of subjective (sensory) experience? This thesis is about how to approach mental processes, such as mindfulness (attention) and minding (awareness), for the primary human sense, vision, with reductionist tools. Neurobiological studies of consciousness have proven elusive in the past for the neural structures, by which visual attention and awareness operate, are difficult to observe. Although cellular studies of visual attention have come a long way to explain a specific instance of conscious awareness, where certain aspects of the world stand out in consciousness at the expense of others, there is no indication that cognitive neuroscience may soon stride the "anatomical path" to self-understanding. The difficulty lies in anatomically isolating one component of consciousness from the other in sensory interactions that are the product of complex neural mechanisms. It is here we make our prevalent contribution by experimentally unfolding the ecological range of sensory interactions that (potentially) reveal the cortical specificity of visual attention and awareness. Each chapter of this thesis delivers possible neural mechanisms and/or sites that pertain to component operations of visual awareness: "seeing", without looking (Chapter 2, perception outside the focus of attention), "looking" without seeing (Chapter 3, unconscious visual perception), "reading" without seeing the letters (Chapter 4, visual interpretation devoid cognition), "sensing" which eye sees (Chapter 5, unmasking the visual origin of eye), "learning" what the other eye sees (Chapter 6, dichoptic collaboration in "filling-in"). This thesis provides a conceptual framework that draws together results from visual neuroscience and psychophysics. It describes new approaches to determine the neural correlates and events responsible for imparting sense-data to the level of abstract thought and planned action. The central theme. however, is the discovery of "blindsight" in normals (Chapters 3 and 4). Visual behavior in the absence of visual awareness (blindsight) has been established unequivocally in patients with lesions in area V1, but remains controversial for normal observers ("subliminal perception"). We describe two visual displays that induce blindsight in normal observers, using an objective measure to distinguish conscious and unconscious performance. Both displays are designed to stimulate area V1 strongly and extrastriate areas poorly. In each display, one quadrant differs from the other three, and this "odd" quadrant is consciously perceived in one form of the display (C-form) but not another (NC-form) (100-200ms presentation time). The first display consists of paired dots moving on orthogonal trajectories (different directions in odd and other quadrants), either within 0.2° of each pair (NC-form) or separated by several deg (C-form), so that the two forms differentially stimulate units in area MT but not area V1 (Qian and Andersen 1994). The second display dichoptically presents arrays of oriented Gabor-elements (different orientations in odd and other quadrants), producing either binocular rivalry (NC-form) or fusion (C-form), so the two forms differentially stimulate binocular but not monocular units. To objectively demonstrate unconscious performance, we randomly interleave two forms of each display in a block of trials. Observers perform a forced-choice task (locating the odd quadrant), and rate their confidence in the choice of quadrant (scale of 1 to 10). Performance is comparable (and far above chance) for C- and NC-forms, but performance and confidence are correlated only for C-forms and essentially uncorrelated for NC-forms, confirming that performance in NC-forms is not accompanied by visual awareness. We surmise that performance may be based on stimulation of area V1 while conscious experience may result from stimulation of extrastriate areas.

Item Type:Thesis (Dissertation (Ph.D.))
Subject Keywords:perceptual ecology
Degree Grantor:California Institute of Technology
Division:Engineering and Applied Science
Major Option:Computation and Neural Systems
Thesis Availability:Public (worldwide access)
Research Advisor(s):
  • Perona, Pietro
Thesis Committee:
  • Perona, Pietro (chair)
  • Andersen, Richard A.
Defense Date:4 June 1996
Record Number:CaltechETD:etd-12202007-110933
Persistent URL:
Default Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:5088
Deposited By: Imported from ETD-db
Deposited On:23 Jan 2008
Last Modified:08 Nov 2023 00:44

Thesis Files

PDF (Kolb_fc_1996.pdf) - Final Version
See Usage Policy.


Repository Staff Only: item control page