Abstract
In this paper, the focus is put on a mechanism that performs spatio-temporal integration of local retinal image-motion to meditated high temporal acuity. The idea of retinal image motion is divided into two categories: the first-order motion (luminance defined) which drives the local parallel processes, and the second-order motion (non-luminance defined) which is computed on time-dependent basis, modulates system response sensitivity to enhance retinal stimulus-induced graded activation for signaling one pattern rather than another. This extends the scope of our hypothetical retinal model ARO in an attempt to explain how the early processes at the retina may impose organization on multi-sensor data to generate encoded maps on different perspectives of the surrounding visual world to send instructive rather than sensory cues to higher centers in the brain.