Spatial constraints on learning in visual search: modeling contextual cuing

Predictive visual context facilitates visual search, a benefit termed contextual cuing (M. M. Chun & Y. Jiang, 1998). In the original task, search arrays were repeated across blocks such that the spatial configuration (context) of all of the distractors in a display predicted an embedded target location. The authors modeled existing results using a connectionist architecture and then designed new behavioral experiments to test the model's assumptions. The modeling and behavioral results indicate that learning may be restricted to the local context even when the entire configuration is predictive of target location. Local learning constrains how much guidance is produced by contextual cuing. The modeling and new data also demonstrate that local learning requires that the local context maintain its location in the overall global context.     

Awareness in contextual cuing with extended and concurrent explicit tests

The term contextual cuing refers to improved visual search performance with repeated exposure to a configuration of objects. Participants use predictive cues-derived from learned associations between target locations and the spatial arrangement of the surrounding distractors in a configuration--to efficiently guide search behavior. Researchers have claimed that contextual cuing can occur implicitly. The present experiments examined two explicit measures--generation and recognition. In Experiment 1, we found that contextual cuing information was consciously retrievable when the number of trials used in a generation test was increased, and the results also suggested that the shorter tests that were used previously were not statistically powerful enough to detect a true awareness effect. In Experiment 2, concurrent implicit and explicit (generation and recognition) tests were employed. At a group level, learning did not precede awareness. Although contextual cuing was evident in participants who were selected post hoc as having no explicit awareness, and for specific configurations that did not support awareness, we argue that awareness may nevertheless be a necessary concomitant of contextual cuing. These results demonstrate that contextual cuing knowledge is accessible to awareness.     

Implicit,long-term spatial contextual memory

Learning and memory of novel spatial configurations aids behaviors such as visual search through an implicit process called contextual cuing (M. M. Chun & Y. Jiang, 1998). The present study provides rigorous tests of the implicit nature ofcontextual cuing. Experiment 1 used a recognition test that closely matched the learning task, confirming that memory traces of predictive spatial context were not accessible to conscious retrieval. Experiment 2 gave explicit instructions to encode visual context during learning, but learning was not improved and conscious memory remained undetectable. Experiment 3 illustrates that memory traces for spatial context may persist for at least 1 week, suggesting along-term component of contextual cuing. These experiments indicate that the learning and memory of spatial context in the contextual cuing task are indeed implicit. The results have implications for understanding the neural substrate of spatial contextual learning, which may depend on an intact medial temporal lobe system that includes the hippocampus (Mi. M. Chun & E. A. Phelps, 1999).     

Use of spatial context is restricted by relative position in implicit learning

For inefficient search, target detection is faster for repeated than for regenerated layouts. This effect, calledcontextual cuing, was assumed to arise from implicit learning of local spatial relationships between targets and distractors. However, a more global influence from distractors far from the target has not been tested. In this study, the search field was divided into upper and lower halves containing a repeated and a regenerated configuration set, respectively. The positions of the two sets were or were not exchanged, meaning that their relative as well as their absolute positions were the same or different (Experiment 1). In Experiment 2, the repeated set appeared alone in either the same or the other half of the screen (same or different absolute position). The contextual cuing effect remained when only absolute position was changed, but not when both absolute and relative positions were changed. These results suggest that contextual cuing depends on relative positional information.     

Does contextual cuing guide the deployment of attention?

Contextual cuing experiments show that when displays are repeated, reaction times to find a target decrease over time even when observers are not aware of the repetition. It has been thought that the context of the display guides attention to the target. The authors tested this hypothesis by comparing the effects of guidance in a standard search task with the effects of contextual cuing. First, in standard search, an improvement in guidance causes search slopes (derived from Reaction Time x Set Size functions) to decrease. In contrast, the authors found that search slopes in contextual cuing did not become more efficient over time (Experiment 1). Second, when guidance was optimal (e.g., in easy feature search), they still found a small but reliable contextual cuing effect (Experiments 2a and 2b), suggesting that other factors, such as response selection, contribute to the effect. Experiment 3 supported this hypothesis by showing that the contextual cuing effect disappeared when the authors added interference to the response selection process. Overall, the data suggest that the relationship between guidance and contextual cuing is weak and that response selection can account for part of the effect.     

When a reappearance is old news: visual marking survives occlusion

Performance in a visual search task becomes more efficient if half of the distractors are presented before the rest of the stimuli. This "preview benefit" may partly be due to inhibition of the old (previewed) items. The preview effect is abolished, however, if the old items offset briefly before reappearing (D. G. Watson & G. W. Humphreys, 1997). The authors examined whether this offset effect still occurred if the old items undergo occlusion. Results show that a preview benefit was found when the old items were occluded but not otherwise, consistent with the idea of top-down attentional inhibition being applied to the old items. The preview benefit is attenuated, however, by movement of the irrelevant stimuli in the displays.     

Categorical implicit learning in real-world scenes: evidence from contextual cueing

The present study examined the extent to which learning mechanisms are deployed on semantic-categorical regularities during a visual searching within real-world scenes. The contextual cueing paradigm was used with photographs of indoor scenes in which the semantic category did or did not predict the target position on the screen. No evidence of a facilitation effect was observed in the predictive condition compared to the nonpredictive condition when participants were merely instructed to search for a target T or L (Experiment 1). However, a rapid contextual cueing effect occurred when each display containing the search target was preceded by a preview of the scene on which participants had to make a decision regarding the scene's category (Experiment 2). A follow-up explicit memory task indicated that this benefit resulted from implicit learning. Similar implicit contextual cueing effects were also obtained when the scene to categorize was different from the subsequent search scene (Experiment 3) and when a mere preview of the search scene preceded the visual searching (Experiment 4). These results suggested that if enhancing the processing of the scene was required with the present material, such implicit semantic learning can nevertheless take place when the category is task irrelevant.     

Region segmentation and contextual cuing

Contextual information provides an important source for behavioral orienting. For instance, in the contextualcuing paradigm, repetitions of the spatial layout of elements in a search display can guide attention to the target location. The present study explored how this contextual-cuing effect is influenced by the grouping of search elements. In Experiment 1, four nontarget items could be arranged collinearly to form an imaginary square. The presence of such a square eliminated the contextual-cuing effect, despite the fact that the square’s location still had a predictive value for the target location. Three follow-up experiments demonstrated that other types of grouping abolished contextual cuing in a similar way and that the mere presence of a task-irrelevant singleton had only a diminishing effect (by half ) on contextual cuing. These findings suggest that a segmented, salient region can interfere with contextual cuing, reducing its predictive impact on search.     

Visual marking across eye blinks

Visual search for a conjunction target can be made efficient by presenting one initial set of distractors as a preview, prior to the onset of the other items in the search display Watson & Humphreys (Psychological Review 104:90–122, 1997). However, this “preview advantage” is lost if the initial items are offset for a brief period before onsetting again with the search display Kunar, Humphreys, & Smith (Psychological Science 14:181–185, 2003). Researchers have long disputed whether the preview advantage reflects a process of internally coding and suppressing the old items or of the onset of the new items capturing attention Donk & Theeuwes (Perception & Psychophysics 63:891–900, 2001). In this study, we assessed whether an internally driven blink (in which participants close their eyes) acts in the same manner as an external blink produced by offsetting and then onsetting the preview. In the novel blink conditions, participants searched feature, conjunction, and preview displays after being cued to blink their eyes. The search displays were presented during the eye blink, and so were immediately available once participants opened their eyes. Having participants make an eye blink generally slowed search but had no effect on the search slopes. In contrast, imposing an externally driven blink disrupted preview search. The data indicated that visual attention can compensate for internally driven blinks, and this does not lead to the loss of the representations of distractors across time. Moreover, efficient preview search occurred when the search items had no abrupt onsets, demonstrating that onsets of new search items are not critical for the preview benefit.     

Attentional guidance of the eyes by contextual information and abrupt onsets.

Contextual cuing is a memory-based phenomenon in which previously encountered global pattern information in a display can automatically guide attention to the location of a target (Chun & Jiang, 1998), leading to rapid and accurate responses. What is not clear is how contextual cuing works. By monitoring eye movements, we investigated the roles that recognition and guidance play in contextual cuing. Recognition does not appear to occur on every trial and sometimes does not have its effects until later in the search process. When recognition does occur, attention is guided straight to the target rather than in the general direction. In Experiment 2, we investigated the interaction between memory-driven search (contextual cuing) and stimulus-driven attentional capture by abrupt onsets. Contextual cuing was able to override capture by abrupt onsets. In contrast, onsets had almost no effect on the degree of contextual cuing. These data are discussed in terms of the role of top-down and bottom-up factors in the guidance of attention in visual search.     

The role of peripheral vision in implicit contextual cuing

Implicit contextual cuing refers to the ability to learn the association between contextual information of our environment and a specific target, which can be used to guide attention during visual search. It was recently suggested that the storage of a snapshot image of the local context of a target underlies implicit contextual cuing. To make such a snapshot, it is necessary to use peripheral vision. In order to test whether peripheral vision can underlie implicit contextual cuing, we used a covert visual search task, in which participants were required to indicate the orientation of a target stimulus while foveating a fixation cross. The response times were shorter when the configuration of the stimuli was repeated than when the configuration was new. Importantly, this effect was still found after 10 days, indicating that peripherally perceived spatial context information can be stored in memory for long periods of time. These results indicate that peripheral vision can be used to make a snapshot of the local context of a target.     

Awareness of the continuously visible: information acquisition during preview

What can we learn about a scene while we stare at it, but before we know what we will be looking for? Three experiments were performed to investigate whether previewing a search array prior to knowing the target allows search to operate more quickly (lower reaction time [RT]), more efficiently (reduced set size slope), and/or by consulting abstract mental representations. Experiment 1 compared RTs for previewed and nonpreviewed arrays, some of which were highly degraded with visual noise. Preview reduced RTs for the noisy displays but did not affect search efficiency. Limited interactions of visual quality and preview suggested that prior exposure allowed the extraction and maintenance of about three abstract identities. If the target was one of those items, the observer responded without searching; if not, the observer searched the remaining items as if there had been no preview. Experiment 2 replicated these findings with less extreme noise. In Experiment 3, subjects previewed 0-6 items of a 12-item display. RTs decreased linearly as the number of previewed items increased from 0 to 3 and then reached a plateau, confirming that the capacity of the representation was about 3 items. Implications for visual awareness are discussed.     

Auditory-visual contextual cuing effect

Under incidental learning conditions, a spatial layout can be acquired implicitly and facilitate visual searches (the contextual cuing effect). Whereas previous studies have shown a cuing effect in the visual domain, the present study examined whether a contextual cuing effect could develop from association between auditory events and visual target locations (Experiments 1 and 2). In the training phase, participants searched for a T among Ls, preceded by 2 sec of auditory stimulus. The target location could be predicted from the preceding auditory stimulus. In the test phase, the auditory-visual association pairings were disrupted. The results revealed that a contextual cuing effect occurs by auditory-visual association. Participants did not notice the auditory-visual association. Experiment 3 explored a boundary condition for the auditory-visual contextual cuing effect. These results suggest that visual attention can be guided implicitly by crossmodal association, and they extend the idea that the visual system is sensitive to all kinds of statistical consistency.     

Response time modeling reveals multiple contextual cuing mechanisms

Contextual cuing refers to a response time (RT) benefit that occurs when observers search through displays that have been repeated over the course of an experiment. Although it is generally agreed that contextual cuing arises via an associative learning mechanism, there is uncertainty about the type(s) of process(es) that allow learning to influence RT. We contrast two leading accounts of the contextual cuing effect that differ in terms of the general process that is credited with producing the effect. The first, the expedited search account, attributes the cuing effect to an increase in the speed with which the target is acquired. The second, the decision threshold account, attributes the cuing effect to a reduction in the response threshold used by observers when making a subsequent decision about the target (e.g., judging its orientation). We use the diffusion model to contrast the quantitative predictions of these two accounts at the level of individual observers. Our use of the diffusion model allows us to also explore a novel decision-level locus of the cuing effect based on perceptual learning. This novel account attributes the RT benefit to a perceptual learning process that increases the quality of information used to drive the decision process. Our results reveal both individual differences in the process(es) involved in contextual cuing but also identify several striking regularities across observers. We find strong support for both the decision threshold account as well as the novel perceptual learning account. We find relatively weak support for the expedited search account.     

Top-up search and the attentional blink: A two-stage account of the preview effect in search

An inefficient visual search task can be facilitated if half the distractor items are presented as a preview prior to the presentation of the remaining distractor items and target. This benefit in search is termed the preview effect. Recent research has found that a preview effect can still occur if the previewed items disappear before reappearing again just before the search items (the “top-up” procedure). In this paper we investigate the attentional demands of processing during the preview and the top-up periods. Experiment 1 found that if attention is withdrawn from the top-up stage then no preview effect occurs. Likewise if attention is withdrawn from the initial preview period then the preview effect is reduced (Experiment 2). The data suggest that the preview effect is dependent on attention being paid both to the initial display and also to the re-presentation of the old display before the search display appears. The data counter accounts of preview search in terms of automatic attention capture by new items or by inhibition of return. We discuss alternative accounts of the results, and in particular suggest an amalgamation of a temporal grouping and a visual marking account of preview search.     

Eye movements and time-based selection: Where do the eyes go in preview search?

In visual search tasks, presenting one set of distractors (previewing them) before a second set which contains the target, improves search efficiency compared to when all items appear simultaneously. It has been proposed that this preview benefit reflects an attentional bias against old information and toward new information. Here we tested directly whether there was such a bias by measuring eye movement behavior. The main findings were that fixations were biased against, and overall dwell times were shorter on, old stimuli during search in the preview condition. In addition, the initial onset of search was delayed in the preview condition and saccades made during the preview period did not disrupt the ability to prioritize new items. The data demonstrate directly that preview search results in an attentional bias toward new items and against old items.     

Contingent visual marking by transients

Preview search is a phenomenon in which a set of new items can be searched with seemingly no interference from items already present in the display. The preview effect has been shown to occur only when the presentation of the new items is accompanied by a luminance change (Donk & Theeuwes, 2001). In a series of experiments, we extend the type of transients that can lead to a preview benefit to offsets and motion, and confirm Donk and Theeuwes's finding that equiluminant color changes do not lead to a preview effect. Like Donk and Theeuwes, we find that preview search does not occur when only the old items undergo a transient change, suggesting that the processes responsible for preview search are triggered when the new items undergo a change detectable by the magnocellular system. In addition, we find that irrelevant transients interfere with preview search only when they match the current attentional set (e.g., luminance change or motion). Results suggest that preview search is not the automatic capture of attention by transients, but rather is contingent on top-down control settings.     

Selective learning of spatial configuration and object identity in visual search

To conduct an efficient visual search, visual attention must be guided to a target appropriately. Previous studies have suggested that attention can be quickly guided to a target when the spatial configurations of search objects or the object identities have been repeated. This phenomenon is termed contextual cuing. In this study, we investigated the effect of learning spatial configurations, object identities, and a combination of both configurations and identities on visual search. The results indicated that participants could learn the contexts of spatial configurations, but not of object identities, even when both configurations and identities were completely correlated (Experiment 1). On the other hand, when only object identities were repeated, an effect of identity learning could be observed (Experiment 2). Furthermore, an additive effect of configuration learning and identity learning was observed when, in some trials, each context was the relevant cue for predicting the target (Experiment 3). Participants could learn only the context that was associated with target location (Experiment 4). These findings indicate that when multiple contexts are redundant, contextual learning occurs selectively, depending on the predictability of the target location.     

Investigating cue competition in contextual cuing of visual search

A fundamental principle of learning is that predictive cues or signals compete with each other to gain control over behavior. Associative and propositional reasoning theories of learning provide radically different accounts of cue competition. Propositional accounts predict that under conditions that do not afford or warrant the use of higher order reasoning processes, cue competition should not be observed. We tested this prediction in 2 contextual cuing experiments, using a visual search task in which patterns of distractor elements predict the location of a target object. Blocking designs were used in which 2 sets of predictive distractors were trained in compound, with 1 set trained independently. There was no evidence of cue competition in either experiment. In fact, in Experiment 2, we found evidence for augmentation of learning. The findings are contrasted with the predictions of an error-driven associative model of contextual cuing (Brady & Chun, 2007).     

Fast color grouping and slow color inhibition: evidence for distinct temporal windows for separate processes in preview search

The authors report 4 experiments that examined color grouping and negative carryover effects in preview search via a probe detection task (J. J. Braithwaite, G. W. Humphreys, & J. Hodsoll, 2003). In Experiment 1, there was evidence of a negative color carryover from the preview to new items, using both search and probe detection measures. There was also a negative bias against probes on old items that carried the majority color in the preview. With a short preview duration (150 ms) carryover effects to new items were greatly reduced, but probe detection remained biased against the majority color in the old items. Experiments 2 and 4 showed that the color bias effects on old items could be reduced when these items had to be prioritized relative to being ignored. Experiment 3 tested and rejected the idea that variations in the probability of whether minority or majority colors were probed were crucial. These results show that the time course of color carryover effects can be separated from effects of early color grouping in the preview display: Color grouping is fast, and inhibitory color carryover effects are slow.