Abstract
Visual perception is continuously shaped by both bottom-up and top-down information. A classic example for top-down modulation is the influence of context on the perception of an ambiguous stimulus: the latter can appear as an utterly different object when it is embedded in different contexts. Can such contextual effects on perception occur even when the context is not consciously perceived? While some low-level contextual effects (brightness and tilt illusions) were found even without awareness of the inducing context, high-level visual illusions (namely, perceiving Kanizsa shapes) were not. Here, we investigated this question using non-illusory, higher-level stimuli, while also manipulating the physical properties of the critical ambiguous stimulus. Observers were presented with six versions of the well-known B/13 ambiguous stimulus. These versions varied in distance between the two parts of the stimulus, hereby making it more similar to B or to 13. This ambiguous stimulus was presented for 33 ms together with context inducers which were either letters ("A" and "C") or numbers ("12" and "14"). While the ambiguous stimulus was always visible, the context inducers were sandwich masked. In Experiment 1 (conscious; N=12), 100 ms blanks were introduced between the context inducers and the masks, so the inducers were visible. In Experiment 2 (unconscious; N =24), the order of masks and blanks was switched, rendering the inducers invisible. Subjects' awareness of the inducers was assessed using the Perceptual Awareness Scale and an additional objective visibility test in which observers judged whether the inducers were letters or numbers. Experiment 1 revealed both context and distance effects (all ps< 0.01). Crucially, this pattern was also observed in Experiment 2 (all ps< 0.01), demonstrating that subjects' interpretation of an ambiguous stimulus can be biased by the context in which it appears, even when this context is not consciously perceived.
Meeting abstract presented at VSS 2017