Organization and plasticity of response properties in the primary visual cortex.

摘要

Understanding the organization of the primary visual cortex is critical for determining how visual scenes are represented in the brain (Issa et al., 2008). Furthermore, understanding how this organization can be disrupted during development has great clinical relevance for treating the resulting problems with vision (Lewis and Maurer, 2005). The work presented here examines, (1) the distributed organization of a visual response property, spatial frequency preference, in the primary visual cortex, (2) an improved model for how these maps of spatial frequency preference and other response properties can be used to predict cortical responses to complex stimuli, and (3) a novel gene that may be involved in the organizational changes caused by the disruption of vision during development. Specifically, in the first set of experiments, a new optical imaging technique, flavoprotein autofluorescence (FA), is applied to resolve a dispute about the organization of maps of spatial frequency (SF) preference in the primary visual cortex. SF maps constructed using FA confirm that spatial frequency is organized in a distributed fashion across the surface of the cortex. In the second set of experiments, the spatial frequency preference other response maps are applied as distributed linear filters in a spatiotemporal energy (STE) model that has been successfully used to predict cortical responses to complex moving images (Zhang et al., 2007). However, the critical assumption that the response maps used as linear filters are separable is shown to be incorrect, because spatial frequency preference depends on stimulus orientation. To address this issue, orientation and spatial frequency responses are reparameterized in a 2D Gabor RF model that provides separable parameters that account for the subtle dependence of spatial frequency responses on orientation. The result is an improved STE model that satisfies the critical assumptions underlying it and more accurately represents the relationship between its spatial response filters. Finally, moving beyond the normal organization of the primary visual cortex, cortical responses are measured in a mouse model to investigate the potential effects of a candidate gene, reelin, on ocular dominance plasticity due to monocular deprivation, an important clinical model for visual developmental disorders such as amblyopia.