THE PROCESSING OF VISUAL INFORMATION IN A SIMPLE RETINA

Our lighted surroundings give us an infinite variety of ever-changing patterns of light and shade and color. Imaged on the mosaic of photoreceptors in our retinas, these are translated into patterns of nervous activity, transformed and utilized in various ways by our nervous system to become ultimately the patterns in our brains that underlie visual responses and visual experience. Our own visual system, and those of all other vertebrates and of higher invertebrates, are extremely complex. Even the first steps of visual processing, in the retina, are only slowly being unraveled. It has been useful to search for simpler visual systems. The lateral eye of the common horseshoe crab, Lirnulus, provides such a system. We and our collcagues have been studying it over a nunibers of years.’,.’ The two lateral eyes of Lirriulus are compound, coarsely faceted, rigidly fixed in the shell, each with a long optic nerve running to the central optic lobes of the brain. Behind each facet is a cluster of about 12 receptor cells, and (usually) one eccentrically placed bipolar neuron. Each such “omniatidiuni” acts as a single receptor unit. From the large axon of the eccentric cell can be recorded electrically trains of nerve impulses elicited by illuminating the receptor unit. The individual receptor units are studied either by restricting illumination to each corresponding facet, by dissecting their individual nerve fibers from the optic nerve, or by impaling the eccentric cell in an onimatidiuni by a niicropipette electrode. Illuminated steadily, an isolated receptor unit responds with a steady discharge of impulses, which appear on record as spikelike “action potentials” of uniform amplitude. The rate of impulse “firing” is approximately linear with the logarithm of the light intensity. A range of 1 0 in intensity elicits frequencies of firing that range from a few impulses per second to 60 or 70. There is thus considerable “data compression” by the receptor. Abrupt changes in intensity are accented transiently: Light suddenly turned on elicits a strong initial outburst of neural discharge, which subsides in a second or so to a steady level. Such “sensory adaptation” to a greater or lesser degree is alniost universal for sense organs. In the visual receptor, such adaptation is in part a property of the transducing mechanism (light to receptor excitation). There is, however, a large component of “neural adaptation,” which in Lir?icrIus is a negative feedback-“self inhibition”-exerted on the eccentric cell by its own discharge. The receptor units in the eye of LirnuIus are interconnected by a network of nerve-fiber branches that constitute a true “retina,” rich in synaptic interconnections. This plexus lies just back of the layer of ommatidia. Functionally, the receptor units do not act entirely independently, but interact with one another. These interactions, exerted laterally over the fibers of the plexus, provide another step of data processing in this eye. The influence that the receptor units exert on one another mutually is inhibitory. This is a contrast-enhancing mechanism. The rate of impulse dis-