Abstract
Adaptation to a sinusoidally modulated contour produces a shift in the apparent shape frequency of a subsequently presented test contour, in a direction away from that of the adaptation stimulus. The phenomenon has been termed the ‘shape-frequency after-effect’ or SFAE (Kingdom & Prins, 2005, JOV, 5, 464). We describe an even larger after-effect of contour-shape amplitude, which we term the ‘shape-amplitude after-effect’ or SAAE. What underlies these after-effects? They occur even when the shape-phase of the contour is randomly changed every half second during the adaptation period, which would tend to lead one to reject the idea that local tilt (or orientation) after-effects (TAEs) are the underlying cause. However we show that even with adaptation contour phase-randomization, the geometrical relationships between adaptor and test are such that the TAE is difficult to rule out. We provide evidence against the TAE: sizeable after-effects are obtained for adaptors that are sine-wave-shaped and tests that are square-wave-shaped. In addition we test, and reject, three other candidates besides local orientation: global average curvature, local signed curvature and global spatial frequency. We suggest that contour shape after-effects result from adaptation to the sizes of local, partly-bounded regions defined by the contour's shape. This in turn implies that contour shape is processed by mechanisms that encode the sizes of partly-bounded regions of the stimulus.
Funded by NSERC grant OGP01217130 given to F.K.