Abstract
Selective adaptation is a powerful psychophysical technique that has provided important insights into the organisation and properties of information processing in the visual system. One classic visual adaptation phenomenon is the tilt-aftereffect (TAE), whereby prolonged exposure to an oriented stimulus alters the perceived orientation of subsequently viewed stimuli. The properties of the TAE point towards a locus early in cortical visual processing, and extant models of the effect propose a central role for changes in excitatory and/or inhibitory interactions between, orientation-selective neurons in primary visual cortex. Here, using adapting stimuli containing complex global structure, we report sizeable TAEs that exhibit markedly different characteristics to those traditionally reported. Whereas the presence of a TAE following adaptation to a Cartesian grating requires spatial overlap between adaptor and test stimuli, aftereffects induced by adaptation to a polar (circular, radial) grating extend to remote, unadapted spatial regions. We further demonstrate that the direction and magnitude of these spatially remote TAEs are consistent with the local orientation implied by the global structure, rather than simple propagation of effects from nearby adapted regions. Manipulation of the relative spatial frequency of adapting and test stimuli results in a tight spatial frequency tuning profile for the traditional TAE. In contrast, remote TAEs resulting from global form adaptation show little or no spatial frequency tuning. These findings demonstrate that mechanisms which pool orientation information across space in order to extract global structure exert a marked influence on local orientation coding.