Seemingly random perceptual switching in bistable events triggered by transient unconscious cues
Here are some demonstrations of the various conditions discussed in the following paper:
Ward, E. J., & Scholl, B. J. (2015). Stochastic or systematic?: Seemingly random perceptual switching in bistable events triggered by transient unconscious cues. Journal of Experimental Psychology: Human Perception & Performance, 41(4), 929-939.
These demonstrations are provided as Quicktime movies, which can be downloaded or viewed directly in most web-browsers. These movies are a bit large and choppy, but they should be sufficient to illustrate the basic conditions. As compressed versions of the original stimuli, these movies may not preserve the precise spatial and temporal characteristics of the originals.  
What we see is a function not only of incoming stimulation, but of unconscious inferences in visual processing. Among the most powerful demonstrations of this are bistable events, but what causes the percepts of such events to switch? Beyond voluntary effort and stochastic processing, we explore the ways in which ongoing dynamic percepts may switch as a function of the content of brief, unconscious, independent cues. We introduced transient disambiguating occlusion cues into the Spinning Dancer silhouette animation. The dancer is bistable in terms of depth and rotation direction, but many observers see extended rotation in the same direction, interrupted only rarely by involuntary switches. Observers failed to notice these occasional disambiguating cues, but their impact was strong and systematic: cues typically led to seemingly stochastic perceptual switches shortly thereafter, especially when conflicting with the current percept. These results show how the content of incoming information determines and constrains online conscious perception -- even when neither the content nor the brute existence of that information ever reaches awareness. Thus, just as phenomenological ease does not imply a corresponding lack of underlying effortful computation, phenomenological randomness should not be taken to imply a corresponding lack of underlying systematicity.

Baseline Temporal Dynamics (1.3 MB)
In all experiments, observers continuously monitored the Spinning Dancer animation and were instructed to press and hold keys to indicate at every moment which direction of rotation they perceived. To obtain measeurements of baseline perceptual switching, observers viewed the Spinning Dancer animation without any cues. Given the enduring interest in the initial Spinning Dancer illusion, we used a version of the original animation (which can be obtained from Wikipedia or from Nobuyuki Kayahara's page). To experience switching, you'll want to download this movie locally and then set it to loop.  
Transient Disambiguating Cues (60 ms) (1.3 MB)
Transient Disambiguating Cues (300 ms) (1.3 MB)
In Experiments 2-4, we explored the influence of explicit occlusion cues on perceptual switching. These cues were taken directly from the Wikipedia entry for the illusion, where they were used (with no attribution) as a simple way of disambiguating the image. We presented these cues in a subtle and transient fashion so that observers were unaware of their existence (throughout the entire experiment), and we varied the nature of the cues to be either consistent or conflicting with the observer's current percept. Conflicting cues led to perceptual switching immediately afterwards -- even though these switches still seemed to be stochastic to the observers. Both of these movies contain two cues -- first a line segment that suggests counterclockwise rotation, and then (after a few rotations) a line segment that suggests clockwise rotation. The "60 ms" version presents these cues for roughly the duration for which they appeared in Experiments 2 and 4; the "300 ms" version presents them for 5 times as long, to emphasize the effect.