ABSTRACT
The lamina of the insect visual system is the site of the first synaptic transmission of the dynamic visual image sampled by the photoreceptors of the retinal array. This chapter analyses the nonlinear aspects of the image/signal processing taking place in the lamina system. It shows what inferences can be drawn from the empirically derived kernels, by examining the transmission and generation of linear and nonlinear components and developing model systems consisting of networks of polynomial time-invariant operators. The chapter presents a study of the signal processing taking place within a neural system in terms of polynomial time-invariant operators. In the fly, the axons of six receptors sampling the same region of space terminate in one lamina cartridge. White-noise stimulation on a checkerboard domain produced estimated spatiotemporal first-order kernels for the system, showing precisely the distribution of sensitivity and antagonism on space and time.
