THE NEURAL-NETWORK OF THE LIMULUS RETINA - FROM COMPUTER TO BEHAVIOR

The visual system of the horseshoe crab, Limulus polyphemus, provides an excellent opportunity for studying the neural basis of behavior. Quantitative analysis of the animal's visual behavior is now possible as is theoretical analysis of information processing in its retina. We combine these theoretical and behavioral approaches to investigate the nature of the signals the eye transmits to the brain for the animal to see. Over the years theoretical studies of the Limulus eye were restricted by the limited capabilities of single processor digital computers. However, a breakthrough in technology with the advent of parallel computers greatly enhances the analysis of large neural networks such as that of the retina. We have developed a time-dependent model of the Limulus retina on the Connection Machine (Model CM-2), which is a massively parallel computer containing 32,768 processors. The model represents a matrix of 64 x 128 receptors and simulates interactions among receptors with digital filters and transduction and adaptation within a receptor by a multistage cascade. Neural response patterns computed with the Connection Machine model replicate to a first approximation the patterns of neural activity recorded in the laboratory. Behavioral studies of Limulus vision carried out in the field can be simulated on the Connection Machine. Neural responses recorded from behaving animals serve to test the accuracy of the model. Thus far we have developed just one model of the retina, but it eventually will have two forms, ''daytime'' and ''nighttime,'' to account for the known circadian rhythms in retinal function. With a combination of field, physiological, and theoretical studies, we hope to gain a better understanding of the neural mechanisms that underlie the animal's visually-guided behavior.