God Soul Mind Brain

the human brain. What you see is the cerebral cortex, the wrinkly shell of cells around the outside of the brain. The primary visual cortex, or V1 as it is often called, is at the back of the brain (the left side of the drawing). This view of the brain is difficult to study because so much of the surface of the cortex is hidden inside the folds and wrinkles.

    This computer rendition of a monkey brain comes from the work of Van Essen and his colleagues. The front of the brain is at the right. The primary visual cortex is shown in black. Area STP, a possible site for the basics of social perception, is also shown in black.

    Diagram 6-2

    Diagram 6-3 shows the same cortex, but this time, using a computer graphics method, the wrinkled cortex has been stretched and flattened so that you can see everything including the cortex inside the folds.

    In addition to the primary visual cortex at the back of the brain, fifteen visual areas are shown, their borders outlined. Neuroscientists are not certain exactly how many visual areas a monkey has—whether some should be lumped together or further split apart—but this drawing gets across the basic organization. Each one of these visual areas processes the entire visual scene in some distinct manner. Together, they accomplish the job of seeing. This map of the visual cortex cost about fifty years, hundreds of laboratories, millions of scientist-hours. It is one of the great achievements of neuroscience. A similar map is now being constructed for the human brain.

    The computer graphics method of flattening the wrinkled cortex also comes from the work of Van Essen and his colleagues.This diagram shows fifteen possible cortical areas in the visual system of monkeys.

    Diagram 6-3

    At first sight, the mosaic of visual areas looks confusing, but some order can be made of it by applying two basic principles. These two principles were worked out by the scientists Ungerleider and Mishkin in 1982, and have formed the backbone of visual neuroscience ever since.

Hierarchy

    The first principle is hierarchy. Primary visual cortex, or V1, is the first cortical area to receive visual signals. V1 processes those signals to some degree, and sends its output to V2. V2 processes the signals further, and sends its output to V3. And so on. Information about the visual scene cascades from the back of the brain forward, through ranks of visual areas, step by step becoming more highly processed.

    For example, neurons in V1, at the beginning of the process, have simple properties. These neurons become active in reaction to spots and short line segments, as if they were monitoring pixels and fragments of the visual world. An active neuron in V1, firing off signals at a high rate, might be shouting out, so to speak, “There’s a vertical line segment projected onto a particular part of the retina!” or something else of that specificity.

    Neurons in TE, at the very end of the process, have much more complex properties. They fire off signals in reaction to whole images like hands and faces, as if they were recognizing actual objects in the world. An active neuron in TE might be effectively shouting out to the rest of the brain, “There’s a human face in view!”

    Somewhere in the stages between V1 and TE, in the step-by-step circuitry, the complexity is built up. Once again, nobody knows how the circuitry computes these visual properties, but there are some increasingly sophisticated computer models, artificial circuit diagrams that may be drawing close to an answer.

    The hierarchy is not absolute. Nothing in the brain is simple. The highest stages of processing, such as TE, send signals back to previous stages, and the backflow is at least as strong as the forward flow of information. The exact reason for the backflow of information is not understood, although there are many reasonable guesses.

Processing streams

    The second principle of organization is a division of the visual cortex into what are called processing streams. Originally Ungerleider and Mishkin proposed two processing streams, but now it looks as though there may be more. Recently, on the basis of a computer model of the monkey visual system, my students and I proposed that there are at least four streams, and therefore I will proceed in my description as though four is the correct number. Not everyone agrees on this particular number, but the general idea is the same in any case. The idea of processing streams is that