Higher order visual processing in the visual cortex.

This chapter covers how the cortex combines visual signals from parallel retinocortical channels into perceptually useful neural representations. The retina sends signals to the visual cortex about the surfaces of objects, including shape, lightness and color, and location in space. One of the most important properties of early visual cortex is the visuotopic map of the contralateral visual field built from the neural images of the visual world in the left and right eyes that are brought into spatial register in the lateral geniculate nucleus (LGN) of the thalamus. The visual cortex computes the orientation of edges and the direction of motion of edges and objects, and the color of object surfaces. It does this by combining signals from ON and OFF feedforward LGN inputs, and by combining color signals from parvocellular, koniocellular, and magnocellular LGN cells. The cortex also computes the depth of objects through stereopsis, which depends on combination of signals from left and right eyes. The cortex is a highly recurrent network in which excitatory and inhibitory neurons make many synapses on each other. Excitatory and inhibitory synaptic currents are approximately balanced, a consequence of the cortex's recurrence. A balanced network like the cortex is very sensitive to feedforward inputs that tip the balance one way or the other.