Neuronal Responses to Visual Texture Features Across Rat Visual Cortex

Abstract

A fundamental goal of sensory neuroscience is to understand how stimulus features are encoded by neuronal populations as information arrives from the periphery and travels through successive stages of cortical processing. A guiding principle in the study of this relationship in visual perception is that the neuronal code is optimized to represent features commonly occurring in natural scenes. This idea has yet to be tested beyond striate visual cortex in rodents, as has been done in primate visual cortex. Here, I generate texture stimuli with different levels of naturalistic structure and use these stimuli to probe how visually-evoked responses chan ge across the visual cortex of rats – an organism shown to have a wide repertoire of complex visual behaviors. I report results from a systematic study of over 5,000 single neurons within visual areas V1, LM, and LI of the rat visual cortex. The results of this survey revealed three main ways in which evoked responses differed across the sampled areas of rat visual cortex. First, the level of naturalistic structure within visual stimuli modulates the magnitude and reliability of evoked responses within single neurons in extrastriate visual areas, but not neurons in V1. Second, the population response patterns within extrastriate visual areas contain relatively more information about visual texture stimuli with higher levels of naturalistic structure. Finally, multiple findings throughout this work point to a systematic change in the representation of visual texture stimuli across the ventral visual pathway of the rat visual cortex. Together, these findings provide support for the framework of a visual cortical hierarchy in the rat in which naturalistic image structure leads to more faithful encoding of visual textures within more lateral visual areas.