Brain Bases of Natural Scenes's Visual Perception of Natural Scenes

Using the available data from psychophysics, cellular electrophysiology and functionnal neuroanatomy of visual pathway, current models of visual recognition suppose that the perception of scenes start with a parallel extraction of differents elementary visual characteristics to different spatial frequencies according to a default processing principle named : 'coarse-to-fine'. According to this principle, the visual scene's analysis would be decomposed in two steps. Fisrt, the fast analysis of the global information borne by low frequency of the scene will provide an overview of the scene's structure and would enable a first perceptive categorisation which would be then refined, approved or denied by the latest analysis of the most local, detailed and precise information, carried by the very high spatial frequency of the scene. The research carried out since several years is preparing a biologically plausible model and to find brain bases by different imaging techniques among healthy subjects but also patients with a brain lesion and patients with a peripheral lesion. The main goal of this Magnetic Resonance Imaging study is to find brain bases of natural scenes's visual perception of the natural scenes. Three studies in Magnetic Resonance Imaging will be conducted, during which subjects will have to categorize pictures of natural scenes filtered in spatial frequencies. The outcome of this study will allow to refine models of visual recognition, most of them based on analysis of spatial frequencies.