Content area
Full Text
The primate visual system contains dozens of distinct areas in the cerebral cortex and several major subcortical structures. These subdivisions are extensively interconnected in a distributed hierarchical network that contains several intertwined processing streams. A number of strategies are used for efficient information processing within this hierarchy. These include linear and nonlinear filtering, passage through information bottlenecks, and coordinated use of multiple types of information. In addition, dynamic regulation of information flow within and between visual areas may provide the computational flexibility needed for the visual system to perform a broad spectrum of tasks accurately and at high resolution.
Visual images provide an immensely rich source of information about the external world. We use this information so effortlessly and efficiently that it is easy to underestimate the computational complexity of ordinary visual routines. For instance, driving a car on a busy street requires coordination of a number of ongoing visual tasks: reading traffic signs, recognizing familiar landmarks, localizing and tracking vehicles, and being alert to pedestrians and other sources of sudden danger. The results of these ongoing analyses must be appropriately routed to brain regions involved in perception, motor control, and planning. Our ability to perform such tasks accurately, rapidly, and reliably in the natural environment requires an extremely sophisticated and well-engineered visual system.
The challenge of understanding vision has prompted widespread interest in interdisciplinary approaches that attack the problem from complementary viewpoints. In this article we discuss a combined neurobiological and systems engineering approach to studying the primate visual system. The neurobiological approach provides a detailed anatomical and physiological description of the visual system and suggests a number of key principles, including modular design, hierarchical organization, and the presence of distinct but intertwined processing streams. The systems engineering perspective provides a framework for analyzing and interpreting these and other aspects of visual system organization. It emphasizes the need for computationally sound models that are grounded in basic principles of signal processing and respect both the power and the limitations of the underlying neural circuitry.
ANATOMICAL OVERVIEW
The anatomy of the primate visual system has been intensively studied in the macaque monkey, whose visual system is similar in many ways to that of humans. Figure 1 provides an overview of the major components of the...