Abstract
After prolonged viewing of an expanding pattern, a stationary test pattern shows a negative after-effect of apparent contraction. This after-effect could be the result of mechanisms responding to the contours moving outwards across the retina — a conventional MAE — or, alternatively, it could be due to mechanisms sensitive to the changing size (or spatial frequency) of local texture elements. To test for the latter possibility, we measured the after-effects produced by adapting the visual system to an expanding, bandpass filtered, texture pattern in which successive frames were completely uncorrelated, like expanding twinkling noise, so that there was no consistent pattern of outward motion. We also created movie sequences in which the proportion of coherent transitions (same pattern in subsequent frame) to uncorrelated transitions (different pattern in subsequent frame) was systematically varied from 100:0 to 0:100. The after-effects were measured using a sensitive nulling procedure. Results were clear. After adaptation for 60s to an expansion sequence with 100% coherent transitions (as occurs in the normal world), there was a substantial after-effect such that observers judged a test sequence, which physically expanded at 15–20% of the adapting rate of expansion, to be unchanging in size. The strength of the after-effect decreased linearly with an increase in the proportion of uncorrelated transitions and none of our eight observers saw any significant after-effect when 100% of the transitions were uncorrelated. Size change alone was insufficient to create an after-effect. In sum, we found no evidence for a separate mechanism that responds to the change-of-size (as proposed by Schrater et al. 2001) in the absence of outwardly moving contours. This is consistent with Regan and Beverley's model of a size change mechanism based on detecting opposite directions of motion rather than a size change per se.