Purchase this article with an account.
Elizabeth Salvagio, Andrew J. Mojica, Mary A. Peterson; Context effects in figure-ground perception: The role of biased competition, suppression and long-range connections. Journal of Vision 2008;8(6):1007. doi: https://doi.org/10.1167/8.6.1007.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Does context affect figure-ground perception? And do the conditions under which context effects occur elucidate the underlying mechanisms? Salvagio, Peterson & Kim (VSS 2007) reported that context modulated the strength of the figural cue of convexity: Convex regions were increasingly likely to be seen as figure as the number of alternating convex and concave regions increased from 2 to 8 (57%–89%). Context effects were found when concave regions were homogeneous but not heterogeneous in color; homogeneity of the convex regions was irrelevant. On a biased competition model, convex and concave regions compete for figural status; concave regions are overpowered and suppressed. Our previous results suggested that context effects require long-range connections between neurons responding to homogeneous features of suppressed regions. We investigate whether the same pattern holds for more global figural cues, e.g., small area and symmetry. We found small area-context effects: As the number of alternating small and large regions increased, smaller regions were increasingly likely to be seen as figures. Like convexity-context effects, small area-context effects were observed when larger regions were homogeneous, p ps [[gt]] 0.05; homogeneity of the smaller regions was irrelevant. Again, biased competition (a bias toward small regions) and homogeneity of the suppressed regions (the larger regions) were necessary for context effects. In contrast, we did not obtain symmetry-context effects, p [[gt]] 0.10, perhaps because display width increased with region number. Symmetry resolution is impaired in the periphery relative to the fovea. Consequently, the symmetry bias is ineffective in the periphery and asymmetric regions are not suppressed. Without suppression, long-range connections linking homogeneous regions do not mediate context effects. Thus our tests of both small area and symmetry implicate long-range connections between neurons responding to homogeneous features of suppressed regions as the mechanism for context effects.
This PDF is available to Subscribers Only