UC Berkeley

Vision is a fundamental component of the human sensory experience. In order to construct visual percepts and mediate appropriate behaviors, the brain uses myriad neural circuits to enact diverse computations and process visual information. Motion is a particularly prevalent feature of the visual world, representations for which are seen through the visual systems of many mammals. By studying neural circuits for motion processing, we gain insights into an important visual information channel utilized by the brain, and thus establish a foothold from which to further understand mechanisms of perception and behavior. The retina in particular provides the opportunity for study of relatively tractable circuits by which to understand how neural signals are integrated in space and time. This work investigates the mechanisms of neural processing within the early visual system. Chapter 1 provides an introduction to motion processing in the retina. This chapter reviews the literature on the diversity of retinal motion detectors which have thus far been catalogued, and discusses what is presently known about their mechanisms of computation. Chapter 2 focuses on a comparative study of two retinal motion detectors in particular: the ON and ON-OFF direction selective ganglion cells. This chapter presents original research findings on the synaptic mechanisms by which these circuits jointly encode the direction and velocity of motion using electrophysiology, pharmacology, genetic knockout animals, and computational modeling. Chapter 3 concludes with a discussion of the computational role of the retina compared to central brain areas, and examines modern physiological findings alongside insights from evolutionary biology, comparative neuroanatomy and theory studies. In sum, the work within this thesis expands our understanding of the mechanisms by which retinal circuits encode motion, and in so doing contributes to our knowledge of how neural signals are integrated by the brain to mediate vision.