Abstract
How does the visual system use speed to guide visual search? According to the absolute speed hypothesis, proposed by Ivry and Cohen (1992), the fastest moving object in the environment is the most salient and readily captures attention. This hypothesis is supported by Ivry and Cohen's visual search studies showing that faster-moving targets among slower-moving distractors are more readily detected than vice versa. An alternative is the relative speed hypothesis, proposed by Rosenholtz (1999, 2001), according to which an object's salience depends on the absolute difference between its speed and the average speed in the environment. Rosenholtz argued that the bidirectional motion of the stimuli used by Ivry and Cohen may have enhanced the salience of their faster-moving stimuli, thereby confounding the experimental design. Here, we present a series of new experiments in which we replicate Ivry and Cohen's studies with stimuli that all move in one direction only. These experiments eliminate the confound and allow us to pit the predictions of the two hypotheses against one another. We found that a target moving 3.2°/s slower than the distractors was detected as quickly and accurately as a target moving 3.2°/s faster. Furthermore, a target moving 0.8°/s faster than the distractors was more difficult to detect than the slower target. These results reject the absolute speed hypothesis and support the relative speed hypothesis (and, by extension, Rosenholtz' general salience model). The visual system is not biased towards faster-moving objects at the expense of slower-moving ones.
Supported by NIH MH65576 to TSH