Sentinels in the Visual System

The human visual system, like that of other mammals, comprises a multiplicity of parallel visual pathways of old evolutionary origin that bypass the primary visual cortex (V1; Milner and Goodale, 2006; Tamietto and de Gelder, 2010). We currently know very little about these V1-independent pathways, but it makes sense from an evolutionary standpoint that these pathways are not simply vestigial, but may continue to be active and perform some visual functions even when V1 is damaged or denervated. To quote W. James " the main function [of such V1-independent systems] is that of sentinels which when beams of light move over then, cry, 'Who goes there' and calls the fovea to the spot " (James, 1893). A recent study by Schmid et al. (2010) published in Nature highlights the contribution of the lateral geniculate nucleus (LGN) in V1-independent vision in monkeys. They show that macaque monkeys with permanent V1 lesions are able to locate correctly high-contrast stimuli presented in the portion of the visual field affected by the lesion (scotoma). Critically, however, reversible inactivation of the LGN in the lesioned hemisphere abolishes behavioral detection and the associated fMRI responses in extras-triate visual cortex. What new light do these important animal findings throw on the understanding of homologous systems in humans and, more broadly, on the neural underpinnings of visual awareness? A first issue concerns the putative critical role of LGN in sustaining all forms V1-independent residual vision. In recent years, a broad spectrum of visual abilities persisting in humans after V1 lesions has been documented (blindsight; Weiskrantz, 2010). These include target detection, visually guided behavior, and the capacity to judge stimulus properties such as shape, color, motion direction, facial expression, and body postures. Given the existence of multiple V1-independent pathways, some of which do not involve LGN (Lyon et al., 2010), it makes sense to expect that these different abilities rely on different V1-independent pathways that are partially segregated starting already from early sub-cortical sites (Danckert and Rossetti, 2005). In fact, a recent study on a patient with V1 lesion and blindsight followed the same logic of the study by Schmid and colleagues, but selectively knocked-out the superior col-liculus (SC) instead of the LGN (Tamietto et al., 2010). Results showed that behavioral (i.e., visually guided manual responses) as well as fMRI evidence of V1-independent visual processing disappear when stimuli that are invisible to the SC, but normally visible to the …