The strongest of these outputs terminate in the lateromedial (LM) and anterolateral (AL) extrastriate visual areas on the lateral side of V1. Outputs from V1 are distributed across many cortical regions, among these to at least nine extrastriate visual areas ( Olavarria and Montero, 1989 Wang and Burkhalter, 2007). Many of these areas receive inputs from the thalamus, but the most direct and strongest visual input to V1 derives from parallel retino-geniculate pathways ( Frost and Caviness, 1980 Simmons et al., 1982) specialized for high spatial/low temporal and low spatial/high temporal sensitivity ( Gao et al., 2010). Mouse visual cortex contains ≥10 areas ( Wang and Burkhalter, 2007), whose topographic organization was recently confirmed by calcium imaging of visual responses ( Marshel et al., 2011). Support for this organization comes from studies in rat that show that inputs to the amygdala, a nucleus involved in the acquisition and expression of fear, derive from ventral but not from dorsal extrastriate visual cortex ( McDonald and Mascagni, 1996). Indeed, it seems plausible that evolution has selected separate systems for recognizing predators and for navigating routes of escape ( Livingstone and Hubel, 1988). Based on these findings, it was proposed that rodents have functionally specialized cortical streams.
This explanation was later revised by experiments in rats, which showed that lesions in the temporal cortex interfere with object recognition, whereas lesions in the parietal cortex impair spatial orientation ( Kolb et al., 1994 Tees, 1999 Ho et al., 2011). In rodents, the notion of “two visual systems ” originated from studies in hamster, which showed that the cortex plays a role in recognizing “what ” an object is, whereas taking action “where ” to move was thought to be determined by the optic tectum ( Schneider, 1969). Studies in primates have shown that the ventral stream uses visual information for object recognition, whereas the dorsal stream is specialized for spatial perception and visually guided actions ( Kravitz et al., 2011). Although the scheme received support from behavioral observations in rat ( Kolb, 1990), cortical streams have not been demonstrated anatomically. The discovery that outputs from primate primary visual cortex (V1) flow through distinct dorsal and ventral streams ( Ungerleider and Mishkin, 1982 Goodale and Milner, 1992) has raised the question whether similar pathways exist in the visual system of afoveal rodents ( Livingstone and Hubel, 1988 McNaughton et al., 1989 Salinas-Navarro et al., 2009). These two subnetworks resemble dorsal and ventral cortical streams in primates, demonstrating that the basic layout of cortical networks is conserved across species. Importantly, the community structure was strongly dependent on the location of the source area: outputs from medial/anterior extrastriate areas were more strongly linked to parietal, motor, and limbic cortices, whereas lateral extrastriate areas were preferentially connected to temporal and parahippocampal cortices. Each source area showed a unique distribution of projection weights across its targets (i.e., connectivity profile) that was well fit by a lognormal function. We found a high density of the cortical graph that exceeded that shown previously in monkey.
Here, we traced the outputs of 10 visual areas and used quantitative graph analytic tools of modern network science to determine, from the projection strengths in 39 cortical targets, the community structure of the network. In primate visual cortex, information is processed in a dorsal stream specialized for visuospatial processing and guided action and a ventral stream for object recognition. To understand how this works in the normal brain and to determine the impact of disease, mice are promising models. Much of the information used for visual perception and visually guided actions is processed in complex networks of connections within the cortex.
0 kommentar(er)
