Real-world visual scenes display consistent first- and second-order statist
ical regularities to which visual neural representations may be perceptuall
y matched, but these lower-order regularities stem from constraints on imag
e power spectra, which appear to carry much less perceptual information tha
n image phase spectra. Natural scenes are shown also to display consistent
higher-order statistical regularities, and an analysis of these regularitie
s in terms of fourth-order spectra shows that they are strongly dependent o
n spatial frequency. These findings have important consequences for the des
ign of a visual system that aims to maximize sparseness in neural represent
ations. (C) 1999 Optical Society of America [S0740-3232(99)01207-7].