Abstract
The Tilt After-Effect (TAE) is maximal when the adaptor and test are presented in the same part of the visual field (Gibson, 1937) and contain the same spatial frequencies (Ware and Mitchell, 1974). These results implicate low-level, spatial frequency (SF) selective processes in the TAE, but they do not exclude the possibility that higher-level, global processes may also contribute to the TAE. To quantify that contribution, we presented adaptor and tests in different parts of the visual field. TAEs were measured using a two-alternative forced choice method that eliminates many forms of non-perceptual bias. Observers adapted to stimuli tilted ±15° of upright (for houses) or ±15° of vertical (for gratings). Two tests were presented on each trial and observers selected the test with the smallest apparent tilt. In Experiment 1 we tested adaptation to tilted images of houses (taken in the fronto-parallel plane at eye level) in conditions where the adaptor and tests were: (a) the same house, (b) different houses, or (c) differently filtered versions of the same house (separated by more than 2 octaves for spatial frequency). In Experiment 2 we examined TAEs for gratings when the adaptor and tests had either (d) identical or (e) different spatial frequencies (separated by 2 octaves). In all conditions the TAE was significant (p < 0.05) for most observers (a: 4/5, b: 4/5, c: 5/5, d: 8/8, e: 5/6). In all conditions the average TAE was approximately 1° (a: 1.01°, b: 1.18°, c: 1.49°, d: 1.27°, e: 1.19°). We conclude that a generalized mechanism can contribute to the TAE. It is relatively unselective for spatial frequency and does not require the adaptor and test to be presented in the same part of the visual field.
Meeting abstract presented at VSS 2016