Event Types in Perception: Containment vs. Occlusion
 
 
Here are some demonstrations of the various phenomena and manipulations discussed in the following paper:
Strickland, B., & Scholl, B. J. (2015). Visual perception involves 'event type' representations: The case of containment vs. occlusion. Journal of Experimental Psychology: General, 144(3), 570-580.
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.  
 
Recent infant cognition research suggests that our core knowledge involves "event type" representations: during perception, the mind automatically categorizes events into types (e.g. occlusion and containment), which then guide attention to different properties (e.g. with width processed at a younger age than height in containment events, but not occlusion events). We tested whether this aspect of infant cognition also structures visual processing in adults. In 6 experiments, adults had to detect occasional changes in ongoing dynamic displays that depicted repeating occlusion or containment events.  
 
Experiment #1: Vertical Events (3 MB)
Experiment #1: Vertical Events (Cheat Version) (2.9 MB)
 
In containment events, a rectangular object continuously moved from the top edge of the display into a depicted container near the center of the display, and back. Occlusion events were identical, except that the object repeatedly appeared to move behind the container instead of into it. Occasionally, a moving object changed either its height or width slightly while invisible, and observers had to press a key whenever they detected such a change. (The first movie depicts a sample trial excerpt. The second movie is simliar, except that multiple changes are made to the leftmost object, starting with its second disappearance. Subjects would not see this, but the demo may help readers orient to hard-to-detect changes in order to see what they look like.) Mirroring the developmental progression, change detection was better for width vs. height changes in containment events, but no such difference was found for otherwise equivalent occlusion events. This pattern held even for observers who did not notice the difference between occlusion and containment. These results suggest that event-type representations are a part of the underlying currency of adult visual cognition.  
 
Experiment #2: Horizontal Events (3.5 MB)
Experiment #2: Horizontal Events (Cheat Version) (3.3 MB)
 
We hypothesized that width was prioritized over height during containment events in Experiment 1 because only width was relevant to whether the object could fit inside the container. (In contrast, both variables were equally relevant to occlusion.) This makes the strong prediction that the relative prioritization of width and height should flip for containment events that involve horizontal containers, as tested here. (The first movie depicts a sample trial excerpt. The second movie is simliar, except that multiple changes are made to the top object, starting with its fourth disappearance. Subjects would not see this, but the demo may help readers orient to hard-to-detect changes in order to see what they look like.) This experiment functionally replicated Experiment 1, with height now prioritized over width during containment events -- i.e. so that the prioritized variable during containment was again the one that was relevant to whether the rectangle could 'fit' inside the container.  
 
Experiment #3: Shading Control (3.1 MB)  
In order to depict containers in our simple animations, we had to depict the inner and outer surfaces of the container differently, and we did so via distinct shading of the same hue, as might be produced by simple lighting differences. This meant, though, that the rectangles were disappearing behind surfaces with different shades in containment vs. occlusion. Could those shading differences alone have driven our results? We tested this possibility by directly contrasting change detection with uniform surfaces of each of the two possible shades. This had no influence on change detection, ruling out the possibility that the containment vs. occlusion contrast could be explained in terms of simple shading.  
 
Experiment #4: Timing Control (1.1 MB)  
The basic physics and optics of occlusion vs. containment in Experiments 1 and 2 required another confounded difference: in containment events, the moving rectangle stayed hidden for a slightly shorter amount of time than in the occlusion events. We directly tested whether this factor influenced change detection, by matching the temporal dynamics of Experiments 1 and 2, but testing only occlusion behind uniform surfaces of various sizes. The subtle timing differences had no influence on change detection, ruling out the possibility that the containment vs. occlusion contrast from Experiments 1 and 2 reflects simple timing differences.  
 
Experiment #6: Sensitivity vs. Bias (3.2 MB)  
The differences in attentional prioritization observed with the demos above could be realized in principle by enhanced sensitivity or by an increased bias to see such information (even when not present), where both possibilities are consistent with differences in hit rates. In Experiments 1 and 2, false alarms could not be isolated to any particular category, but the present experiment replicated Experiment 2 used a blocked design in order to untangle sensitivity vs. bias. In any given trial, all of the depicted events were of one event type (either containment or occlusion), and all of the changes were of one spatial dimension (either height or width). Here the demo depicts height changes in containment events. Robust differences in sensitivity were observed, consistent with the possibility that these differences reflect relatively automatic perceptual processing. Furthermore, isolating sensitivity in this way makes the nature of this effect especially clear: all of the 'action' in the differences is driven only by increased prioritization of the dimension relevant to 'fit' in the containment events (which was height in this study, since it employed horizontal trajectories).