The cerebral neocortex is a part of the brain that plays a major role in all cognitive activity. However, the structure of the cortex is remarkably uniform: the neural circuitry is roughly the same in all cortical areas. This thesis investigates the proposal that the cortex is a uniform cognitive architecture, providing a system of general-purpose computational mechanisms that are applied in diverse psychological domains. Evidence for cortical uniformity is reviewed, and assumptions and implications of the proposal are discussed. Objections to the proposal are considered, including modular organization, diversity of neuronal physiology, and the relation between computation and neural circuitry. A general theory of the cortex is then introduced. Among the general-purpose mechanisms are: a basic feedforward flow of information within layers 2-4 and between cortical areas, supplemented by feedback and lateral interactions; production of motor output in layer 5; representation of information by population coding; a form of gain control and enhancement of pattern-selectivity performed by localized excitatory and inhibitory circuitry; a form of synaptic plasticity causing neurons to learn patterns that occur frequently in the input; and a selective attention mechanism. This theory is then used to qualitatively explain a broad range of empirical data in visual perception, including: early processing of luminance, motion, spatial frequency, binocular disparity, texture, lightness and colour; general aspects of shape perception and object recognition; and special topics in shape perception, including the Gestalt grouping effects, illusory contours, multistability and hysteresis, stereopsis, the figure-ground phenomenon, and characteristics of selective attention as studied by visual search A simulation demonstrates how neurons can acquire simple and complex receptive fields, and organize themselves into a smooth map of orientation selectivity, from exposure to natural visual input. The thesis concludes with some speculations on the circuitry underlying conscious thought.
This paper is not available online