The coarse-to-fine hypothesis revisited: Evidence from neuro-computational modeling Martial Mermillod a, * , Nathalie Guyader b,c , Alan Chauvin b,c a Cognitive Science Department, Universite ´ de Lie ` ge, Belgium b Laboratoire de Psychologie et Neuro-Cognition, CNRS, UMR 5105, Grenoble, France c Laboratoire des Images et des Signaux, CNRS, ESA 5083, INPG, 27 Av Fe ´lix Viallet, Grenoble, France Accepted 12 August 2004 Available online 6 November 2004 Abstract The human perceptual system seems to be driven by a coarse-to-fine integration of visual information. Different results have shown a faster integration of low-spatial frequency compared with high-spatial frequency (HSF) information, starting at early retinal pro- cesses. The difference in spatial scale decomposition remains throughout the lateral geniculate nucleus (Hubel & Wiesel, 1977) and V1 (Tootell, Silverman, & De Valois, 1981). During the last decade, a debate has emerged concerning the origin of the coarse-to-fine integration. Is it a constant, perceptually driven integration (Parker, Lishman, & Hughes, 1992, 1996)? Instead, the flexible use hypothesis suggests that different spatial frequency channels could be enhanced depending on the requirement of the task for high-level cognitive processes like categorization (Oliva & Schyns, 1997; Schyns & Oliva, 1999). In two connectionist simulations, we have shown that global categorization performance could actually be better performed with HSF information when the amount of information is normalized across the different spatial frequency channels. Those results suggest that high-level requirement alone could not explain the coarse-to-fine bias toward LSF information. A hypothesis is proposed concerning the possible implication of the amount of data provided by different spatial frequency channel that might provide the perceptual bias toward LSF information. Ó 2004 Elsevier Inc. All rights reserved. 1. Introduction During recent years, several results have shown the importance a perception belonging to different scales to account for both neurophysiological and psychological data. Those studies allowed us to understand how visual and cognitive processes operate (Morrison & Schyns, 2001; Oliva & Schyns, 1997; Torralba & Oliva, 2003). Starting at the early stages of visual processing, different results have shown a spatial scales decomposition of retinal information which is very robust to illumination or contrast modification (Smirnakis, Berry, Warland, Bialek, & Meister, 1997). That spatial frequency (or scale) decomposition is then processed through the lat- eral geniculate nucleus (Hubel & Wiesel, 1977) and visual cortex (De Valois & De Valois, 1988; Lauritzen & Miller, 2003) and is provided as input for higher level cognition, such as categorization and recognition cortical areas. In other words, it seems that accounting for high-level pro- cesses requires a good understanding of early visual pro- cesses and beyond, studying scale usage could serve as a tool to investigate the interactions between perceptual and cognitive processes. 2. The coarse-to-fine hypothesis: Neurophysiological and psychological data 2.1. Neurophyiological and computational data A dominant hypothesis in cognitive sciences suggests a coarse-to-fine bias in scale processing, meaning an 0278-2626/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.bandc.2004.08.035 * Corresponding author. E-mail address: mmermillod@ulg.ac.be (M. Mermillod). www.elsevier.com/locate/b&c Brain and Cognition 57 (2005) 151–157