Motion dictates that the spatial structure of natural image sequences is highly structured over time. In principle, knowledge of this spatiotemporal regularity, combined with local estimates of image structure and velocity, could be exploited by the brain to construct a forward model describing the likely future pattern of visual input. We have recently proposed that predictive signals generated from a drifting sinusoidal grating stimulus interfere with visual input in adjacent regions of space, resulting in a pattern-specific modulation of contrast sensitivity (Roach et al 2011, Curr Biol 21 740–745). Detection thresholds are reduced for targets with a spatial form consistent with continuation of the moving pattern (constructive interference) but increased for targets that are inconsistent (destructive interference). To gain a more detailed understanding of the underlying mechanisms, here we extend our examination of these effects using contrast discrimination, equivalent noise masking and contrast matching paradigms. In keeping with our previous work, predictive interference was inferred from a phase-dependent modulation of performance at the leading edge (but not the trailing edge) of a drifting sinusoidal inducing stimulus. Results suggest that predictive interference (i) is evident at all points within the facilitatory range of the threshold versus pedestal contrast (TvC) function (ie, the ‘dip’), but not within the masking range (the ‘handle’); (ii) is robust to the addition of small amounts of external contrast noise, but disappear beyond the elbow of the threshold versus noise (TvN) function; and (iii) exhibits no measurable effect on supra-threshold perceived contrast.
