100 • CHAPTER 4 Attention
BOTTOM-UP DETERMINANTS
OF EYE MOVEMENTS
Attention can be infl uenced by stimulus salience—the
physical properties of the stimulus, such as color, contrast,
or movement. Capturing attention by stimulus salience is
a bottom-up process because it depends solely on the pat-
tern of light and dark, color and contrast in a stimulus. For
example, the task of fi nding the people wearing yellow hats in
Figure 4.22 would involve bottom-up processing because it
involves responding to the physical property of color, without
considering the meaning of the image (Parkhurst et al., 2002).
But where we look is not determined only by the bottom-
up processes triggered by stimulus salience. We can see that
this is true by checking the eye movements caused by the foun-
tain in Figure 4.23. Notice that the person never looks at the
fence in the foreground, even though it is very salient because
of its high contrast and its position near the front of the scene.
Instead, the person focuses on aspects of the fountain that are
more interesting, such as the horses. In this example, it is the
meaning of the horses that attracts attention.TOP-DOWN DETERMINANTS OF EYE MOVEMENTS
Top-down processing is also associated with scene schemas—an observ-
er’s knowledge about what is contained in typical scenes (remember
“environmental regularities” from Chapter 3, page 63). Thus, when
Melissa Vo and John Henderson showed observers pictures like the ones
in ● Figure 4.24, observers looked longer at the printer in Figure 4.24b
than the pan in Figure 4.24a because a printer is less likely to be found
in a kitchen.
The fact that people look longer at things that seem out of place
in a scene means that attention is being affected by their knowledge of
what is usually found in the scene. Consider, for example where the per-
son looked when presented with the baseball scene in ● Figure 4.25. The
person’s attention appears to have been initially captured by the bright
yellow band (perhaps an example of stimulus salience at work), but the
person’s gaze immediately shifts to the fi eld, fi xating on the various play-
ers. It seems probable that the person’s knowledge of the layout of the
bases and positions of players played a role in determining where he or
she looked.
The way the person scanned the baseball scene also suggests that
attention is infl uenced by a particular person’s knowledge and interests.
This person appears to be interested in baseball and have some knowl-
edge about it. Where do you think a person with little interest in baseball
but a great deal of interest in architecture might have looked? Most likely
the person would be more interested in, and therefore pay more attention
to, the buildings in the upper part of the scene.
Attention occurs not only as we view static scenes, but as we carry
out actions. The development of portable eye trackers like the one in
● Figure 4.26 makes it possible to track people’s eye movements as they
perform tasks. This research shows that when a person is carrying out a
task, the demands of the task override factors such as stimulus saliency.
● Figure 4.27 shows the fi xations and eye movements that occurred as
a person was making a peanut butter sandwich. The process of mak-
ing the sandwich begins with the movement of a slice of bread from the
bag to the plate. Notice that this operation is accompanied by an eye● FIGURE 4.24 Stimuli used by Vo and Henderson
(2009). Observers spent more time looking at the
printer (in B) than at the pot (in A), shown inside
the yellow rectangles (which were not visible to the
observers). (Source: M. L. Vo & J. M. Henderson, “Does Gravity
Matter? Effects of Semantic and Syntactic Inconsistencies on
the Allocation of Attention During Scene Perception,” Journal of
Vision, 9, 3, Article 24, Figure 1A & B, 1–15, 2009.)● FIGURE 4.23 Scan path of a person viewing a fountain in
Bordeaux, France.First fixationCourtesy of John M. Henderson, University of EdinburghCopyright 2011 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.