Fruitful Visual Search: Inhibition of Return in a Virtual Foraging Task

Article excerpt

Inhibition of return (IOR) has long been viewed as a foraging facilitator in visual search. We investigated the contribution of IOR in a task that approximates natural foraging more closely than typical visual search tasks. Participants in a fully immersive virtual reality environment manually searched an array of leaves for a hidden piece of fruit, using a wand to select and examine each leaf location. Search was slower than in typical IOR paradigms, taking seconds instead of a few hundred milliseconds. Participants also made a speeded response when they detected a flashing leaf that either was or was not in a previously searched location. Responses were slower when the flashing leaf was in a previously searched location than when it was in an unvisited location. These results generalize IOR to an approximation of a naturalistic visual search setting and support the hypothesis that IOR can facilitate foraging. The experiment also constitutes the first use of a fully immersive virtual reality display in the study of IOR.

When attention is focused on a location, stimuli at that location are detected more readily. For example, when their attention is summoned to a location by a flashed cue, people are faster to detect stimuli presented near this cue. However, when the delay between the flashed cue and the subsequent appearance of the stimulus is long enough, people are actually slower to detect the stimulus at the cued location than at other locations in the display. Presumably, after a delay, attention is no longer focused on the cued location, and that location thus receives no processing benefit. Moreover, processing at that location might even be inhibited, making people less likely to refocus attention on that location (Posner & Cohen, 1984). Such an inhibitory mechanism, now known as inhibition of return (IOR; Posner, Rafal, Choate, & Vaughan, 1985), could bias processing in favor of sampling of new information in the visual field.

IOR is most practically relevant in the context of visual search performance. Search is more efficient if participants attend to new items rather than repeatedly searching previously examined ones, and IOR encourages orienting toward new items and away from searched items (Klein, 1988; Klein & MacInnes, 1999; MacInnes & Klein, 2003; Müller & von Mühlenen, 2000; Takeda & Yagi, 2000). When participants must detect a probe during a difficult visual search task, they detect probes in empty locations faster than those in locations occupied by visual search items. However, this effect occurs only if the probe appears while the search array is still present, suggesting that this visual search IOR is object based (see Klein & Taylor, 1994, and Wolfe & Pokorny, 1990, for failures to replicate Klein, 1988, that are resolved with this explanation). The presence of IOR in simple, computer-based visual search tasks is imbued with practical significance; it might serve to facilitate foraging (see, e.g., Klein, 2000; Klein & Machines, 1999; Tipper, Weaver, Jerreat, & Burak, 1994).

In one test of the foraging facilitator hypothesis (Klein & MacInnes, 1999), participants searched for a character in the cluttered drawings of a Where's Waldo? book while experimenters monitored their eye movements. In addition to searching for their target, participants had to make a saccade to a flashed probe whenever it appeared. When the probe appeared in a previously fixated location, saccadic latencies were longer than when the probe appeared at other locations. Furthermore, saccades made prior to probe onset were typically biased away from previous saccadic directions, providing further support for the notion that IOR directs attention to new locations during visual search. IOR and the saccadic direction bias occur even when the probe appears after participants have stopped their visual search and are inspecting a target of their choice (Maclnnes & Klein, 2003). …