Abstract
Center surround cells in the retina and LGN have been shown to exhibit speed tuning (Frishman, Schweitzer-Tong, Goldstein, J. Neurophysiol. 50: 1393-1414, 1983). Understanding the mechanism of this speed tuning is necessary to understand the visual system’s responses to moving stimuli. Recent experiments in the Van Hooser lab showed that speed tuning in the visual cortex could not be adjusted with visual experience (Ritter et al., 2017), and we would like to understand the neural mechanisms that underlie speed tuning in the cortex. Modeling the retina and LGN, the primary inputs to the visual cortex, is important because knowing what information the cortex receives is necessary to understand how the visual system processes speed and direction. In order to understand the properties of the center-surround cells that give rise to cortical responses, first a simulation approach was taken. However, kernel based simulations provided little intuition as to how velocity tuning arises. To address this, we developed a Fourier analysis approach that can be analyzed more easily. The function of the kernel was decomposed into multiple functions, each with only one or two variables, allowing for easier analysis. Fourier analysis allowed for quicker computation and better prediction.