Hemispheric asymmetries in visual search

LATERALITY, 2012, 17 (6), 711
726

Hemispheric asymmetries in visual search William Poynter and Candice Roberts Psychology Department, Western Carolina University, Cullowhee, NC, USA We conducted two visual search experiments, and found that target-detection accuracy and speed were better when the target was projected to the right hemisphere in the feature search condition and better when the target was projected to the left hemisphere in the feature-conjunction search condition. We propose that the highly efficient, so-called parallel search performance characteristic of feature search is enabled by a broadly distributed, global view of the visual field, and the right hemisphere is more efficient than the left in such global processing. On the other hand, the less-efficient performance characteristic of conjunction search (demonstrated by the set-size effect) involves serial shifts of focused attention, and the left hemisphere is more efficient than the right in such localised attentional processing. We suggest that hemispheric asymmetries observed in visual search are related to the attentional demands of the task, and that we adjust our attentional distribution to fit task difficulty. When the target is very distinct, a global, lowresolution attentional distribution is sufficient, and enables parallel search; but a localised, narrow-aperture attentional distribution is sometimes necessary to find targets that either require feature binding, or are very similar to other ‘‘distractor’’ stimuli in terms of compositional attributes.

Keywords: Visual search; Attention; Hemispheric asymmetries.

Feature Integration theory (Treisman, 1998, 2006; Treisman & Gelade, 1980) provides a conceptual framework for the perceptual processes involved in visual search for features and feature conjunctions. Features are considered the compositional attributes of visual objects (e.g., orientation, colour, size, shape, texture), and feature conjunctions are the local combination of these attributes that define the object as a whole (e.g., combination of a rectangular shape that is red in colour and vertically oriented). According to the theory a parallel (simultaneous) search of objects in the visual field is sufficient to detect a feature target (e.g., a red circle in a field of green Address correspondence to: William Poynter, Psychology Department, 301 Killian Building, Western Carolina University, Cullowhee, NC 28723, USA. E-mail: [email protected] # 2012 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business http://www.psypress.com/laterality http://dx.doi.org/10.1080/1357650X.2011.626558

712

POYNTER AND ROBERTS

circles)*the red circle perceptually ‘‘pops out’’ of the field, without the need for item-by-item examination. On the other hand, a sequential search of individual objects is required to detect a feature conjunction target (e.g., a red circle among green circles and red and green squares)*identifying the presence of the red circle requires perceptual binding of colour and shape, which in turn requires localised attentional focus on individual objects in the field. In fact many studies have provided evidence that feature detection can occur with a simultaneous scan of the field while detection of feature conjunction targets requires sequential scanning (e.g., Treisman, 2006; Treisman & Gelade, 1980; Treisman & Gormican, 1988; Treisman & Souther, 1985). Studies employing what we will label the classic feature- and conjunction search test methods present a display filled with distractor items and one target stimulus (presented in a fraction of the trials, often 0.5). The set size (i.e., number of distractors) is varied across trials. Typically it is found that when the target can be discriminated from distractors based on a distinguishing feature (e.g., red colour), detection time is unaffected by set size, which is considered evidence that the field of stimuli can be scanned effectively in parallel. But when the target is a conjunction of features (e.g., shape and colour), detection time increases with set size, presumably because attention must be focused on stimuli one at a time in order to perceptually bind the object’s features together, thus enabling the identification of the target object. The idea that localised attention resources are required for feature conjunction binding is an explicit assumption of feature integration theory and seems to be a widely accepted assumption in the visual search literature, but the role of attention in feature search is controversial. Feature integration theory (e.g., Treisman & Gelade, 1980) originally proposed that feature targets ‘‘pop out’’ of a field of distractors because their presence is recorded in neurological feature maps (e.g., vertical line map, horizontal line map), and these maps can be scanned in parallel and pre-attentively. However, more recent descriptions refer to the role of divided attentional mechanisms and activity of a variable-aperture attentional window that can encompass both finely localised details as well as global views of the visual field (Treisman, 1999, 2006). It seems clear that a parallel scan of the visual field could not be accomplished if localised attention to individual stimuli was required, since this necessarily involves serial processing. But a spatially distributed (what we will call ‘‘global’’) spotlight of attention could enable parallel processing of stimuli in feature search. And in fact several studies have provided evidence that attentional resources are active in feature search. Joseph, Chun, and Nakayama (1997) found that detection of a feature target (presence of a distinct line orientation) was substantially degraded by engaging subjects in an attentionally demanding dual task, which suggests that the dual task was drawing needed attentional resources away from the feature search task. It has been argued that, since this particular study used

HEMISPHERIC ASYMMETRIES IN VISUAL SEARCH

713

an ‘‘attentional blink’’ dual task, the results might indicate the effect of attentional demands on memory functions, not perception (see Hyun, Woodman, & Luck, 2009, p. 22). However, other behavioural studies, using different methods to engage attention, have provided evidence that attentional resources are involved in feature detection (Kim & Cave, 1995; Nothdurft, 1999; Theeuwes, Kramer, & Atchley, 1999; Theeuwes, Van de Burg, & Belopolsky, 2008). Additional evidence for an attentional requirement in feature search comes from studies showing that pre-cueing the location of targets improves search performance for both features and conjunctions (Carrasco & Yeshurun, 1998; Cheal & Lyon, 1992). It could also be argued that finding a target in a classic feature search task is a matter of detecting an ‘‘oddball’’ feature in the display*a task of detecting a spatial discontinuity in colour, texture, shape, size, etc. But this task also seems to require a spatially distributed visual analysis, since only by comparing neighbouring stimuli could one detect spatial contrast in display features. So here we propose that while conjunction search seems to involve localised focus of attention, feature search might involve a more spatially distributed (global) attention in order to enable parallel processing. And if so, it follows that there might be a right hemisphere (RH) advantage in feature search and a left hemisphere (LH) advantage in conjunction search. Numerous studies have provided evidence that the RH is more efficient than the LH in processing global/low spatial frequency properties of visual patterns, whereas the LH is more efficient than the RH in processing local/high spatial frequency properties (Christman, Kitterle, & Hellige, 1991; Delis, Robertson, & Efron, 1986; Kitterle, Christman, & Hellige, 1990; Lamb, Robertson, & Knight, 1989; Michimata & Hellige, 1987; Robertson, Lamb, & Zaidel, 1993; Van Kleeck, 1989; Weissman & Woldorff, 2005). And functional neuroimaging studies (e.g., Fink et al., 1996; Foxe, McCourt, & Javitt, 2003; Weissman & Woldorff, 2005) also support the view of a RH (specifically temporo-parietal regions) bias towards global, spatially distributed attention and a LH bias towards local, narrowly focused attention. Additionally, a number of studies provide evidence that different modes of processing favour different distributions of attention (see Chong & Treisman, 2005; Robertson, Lamb, Egly, & Kerth, 1993). So here we suggest that the right hemisphere should be more efficient than the left in a classic feature search task based on the idea that it either involves spatially distributed attention enabling parallel scan of individual features, or alternatively enabling detection of feature contrast, which is a spatially distributed property of the stimulus pattern. Conversely, we propose that the left hemisphere should be better at classic feature-conjunction search because it is better equipped to utilise focused attention to facilitate local processing (which seems to be necessary for feature binding).

714

POYNTER AND ROBERTS

A number of previous studies have investigated hemispheric asymmetries in visual search. Polich (1984) found a consistent LH advantage in response time and accuracy when participants judged whether all of the elements within an array were the same (all Xs) or whether one was different from the rest (O). The observation that error rates and response times in this study increased with the number of display items (4, 9, or 16) seems to indicate that serial attention shifts and therefore localised processing was involved; if so, the results seem consistent with the hypothesis of a LH advantage in conjunction search. The author suggested that the left hemisphere obtains an advantage for this task because of LH superiority in performing fine-grained feature analysis, and this idea is similar to the notion of LH superiority in localised/high spatial frequency (and therefore ‘‘fine-grained’’) processing. Kingstone, Enns, Mangun, and Gazzaniga (1995) found that, in split-brain participants, performing guided search (i.e., a search that is limited to a relevant subset of localised display items) is a LH function, suggesting that the left hemisphere is better at utilising selective attention to benefit the serial, localised processing necessary for efficient conjunction search. It is also noteworthy that the control condition in the Kingstone et al. (1995) study showed an overall LH advantage in the conjunction search task. Polich, DeFrancesco, Garon, and Cohen (1990) found a RH advantage when the displays were constructed to elicit a gestalt perception*i.e., the perception of a single ‘‘different’’ item standing out in contrast against a uniform background of ‘‘same’’ distractor items (a sort of figure
ground gestalt). A LH advantage occurred when the display elements were less uniform and organised, thereby disabling a gestalt percept and requiring more local analysis. This finding is consistent with our hypothesis of a RH advantage when a spatially distributed (global) analysis of the display is sufficient for target detection, and a LH advantage when localised, focused attention is required. Pedersen and Polich (2001) conducted a study in which participants were asked to scan visual matrices containing up to 64 cells to determine whether the matrices contained an odd or even number of filled cells. Right hemisphere presentations yielded overall shorter RT and lower error rate, and the RH advantage was stronger for the larger matrix sizes. And Palmer and Tzeng (1990), using classic search tasks similar to Treisman and Gormican (1988), found that while left hemisphere presentations showed linearly increasing error rates with set size, right hemisphere presentations produced flat or even decreasing error rates with increasing set size. Their interpretation was that ‘‘the left hemisphere therefore appears to maintain a strictly sequential item-by-item search strategy, while the right hemisphere seems to break the stimulus array into groups of items,

HEMISPHERIC ASYMMETRIES IN VISUAL SEARCH

715

processing each group in parallel’’ (p. 55). These results are consistent with the idea of a variable-aperture attentional spotlight that can be adjusted to fit the difficulty of the search task*a large aperture, diffusely distributed attentional spotlight (which is more effectively deployed by the RH) might enable parallel search for a highly distinct ‘‘pop-out’’ feature target (e.g., red target in a field of green distractors), but a localised, highly focused attentional beam (more effectively deployed by the LH) might be necessary to identify a target defined either by feature conjunction, or a subtle difference in feature value (e.g., colour saturation). Results of studies investigating hemispheric asymmetries in visual search are not entirely consistent, in part because asymmetries depend on task demands, stimulus characteristics, and participant variables (Evert, McGlinchey-Berroth, Verfaellie, & Milberg, 2003; Michael & Ojeda, 2005). In the present study we will test for hemispheric asymmetries in classic feature and conjunction search tasks, using a repeated-measures method that at least partially controls for stimulus and participant factors by using the same participants across experimental conditions, and using the same number, colour, size, and spatial distribution of stimuli in both tasks. Our prediction is that there will be a RH advantage in our feature search task because the target and distractors are highly distinct, and therefore a global display analysis is sufficient for detecting the ‘‘pop-out’’ target. On the other hand, we predict a LH advantage in conjunction search because the target and distractors are less distinct, and therefore item-by-item, localised attentional focus will be required, which is more effectively deployed by the LH. Two experiments were conducted: the first presented displays for a limited time (150 ms), while the second left the display visible until the participant responded. The second method was a partial replication of Experiment 1 with different participants, and helped to ensure that in the conjunction search trials, the participants had enough time to perform an adequate serial search for targets.

METHOD Participants All participants signed informed consent documents prior to participating. In both experiments all participants were right-handed university students between 18 and 49 years of age: Experiment 1: males  19; females 42; Experiment 2 (feature search): males  14; females 22; Experiment 2 (conjunction search): males 6; females  9.

716

POYNTER AND ROBERTS

Materials and apparatus1 The experiments were conducted in a computer laboratory on a university campus. A Visual Basic 6.0 computer application realised the experimental method. The method was executed on Gateway desktop computer hardware. Stimuli (see Figure 1) were displayed on 19-inch LCD displays using 12801024 resolution (viewing distance 81 cm). Participants performed a visual task in which they searched for avertical line target among horizontal line stimuli (distractors), the number of which varied across trials (5, 20, or 40).2 Half the 72 trials in each of the two trial blocks (feature search and conjunction search blocks) presented a single target. Target location was distributed equally between left and right visual fields, and the number of trials presenting 5, 20, or 40 distractors was equalised across conditions. Display elements were either horizontal or vertical lines (0.56 degrees long, 0.11 degrees wide) presented on a dark grey background (2 cd/m2). The RGB levels of the two line colours (orange and green) were chosen to realise colours of approximate equal luminance (80 cd/m2) The target and distractor locations were randomly distributed to a set of 81 display locations that formed a 9 by 9 matrix measuring approximately 11 degrees square. Since the maximum number of display elements on any trial was 40, distributing display elements randomly across this location matrix presented an irregular, scattered pattern of stimuli, and at the same time prevented overlapping features that an unconstrained random spatial distribution method might have produced. Target distances from fixation ranged from approximately 1.0 to 7 degrees. Response Time (RT) required for target detection and Response Accuracy were the dependent variables.

Figure 1. Example of Visual Displays presented in Feature search (left image) and Conjunction Search (right image). In the actual experiments, targets were orange in colour for both tasks. In Feature search, the distractors were orange and green horizontal lines. In Conjunction search, distractors were a mix of green vertical lines and both orange and green horizontal lines. 1 2

Except where stated otherwise, the method described applies to both Experiments 1 and 2. In Experiment 2, 20 distractors were presented on each trial.

HEMISPHERIC ASYMMETRIES IN VISUAL SEARCH

717

Procedure3 After the participant was seated in a comfortable posture at the test workstation, we positioned the keyboard to enable easy reach. Participants were asked to maintain their present posture as best they could during the course of the experiment, as it was important to maintain a constant display viewing distance. An overhead projection system was employed to illustrate the display stimuli and time sequence of trials. The feature search and conjunction search trial blocks of Experiment 1 were counterbalanced across participants (i.e., roughly half the participants performed the feature task first and the other half performed the conjunction task first).4 The procedure for each task was explained just prior to beginning of each block. In the case of feature search, participants were informed that at the beginning of each trial that they would see a yellow fixation symbol appear at the centre of the screen, on which they should visually fixate. Two seconds following the appearance of the fixation symbol, the stimulus display was presented for 150 ms in Experiment 1 (unlimited viewing time in Experiment 2). For the feature search condition, participants were told that their task was to determine whether a vertical line flashed on the screen among a group of horizontal lines and that, since their reaction time and accuracy would be measured, they should respond as rapidly as they could, while being reasonably sure of their answer. They were told the various lines might be either green or orange in colour. The ‘‘1’’ and ‘‘2’’ keys on the numeric pad of a standard PC keyboard were used to convey answers of ‘‘yes’’ and ‘‘no’’, respectively. Participants used the right hand to respond. The procedure for the conjunction search trial block was the same as for feature search, except that participants were told that the task was to search for an orange vertical line among distractors that might include green vertical lines and both red and green horizontal lines, When the experimenter judged that the participant was comfortable with the task for each trial block and the procedure, they completed a set of 10 practice trials, after which they completed the experimental trials.

RESULTS Experiment 1 A repeated-measures analysis was employed, with Visual Field (LVF, RVF) and Task (feature search and conjunction search) being the repeatedmeasures factors. As expected, Task had a strong effect on both RT, 3

Except where stated otherwise, the procedure described applies to both Experiments 1 and 2. Different participant samples performed the feature and conjunction search tasks in Experiment 2. 4

718

POYNTER AND ROBERTS

F(1, 60) 103.2, pB.001, h2 .63, and Accuracy, F(1, 60) 187.8, p B0.001, h2 .76, indicating that responses were slower and less accurate for conjunction search. More interesting was the significant interaction between Visual Field and Task factors for both RT, F(1, 60)  6.0, p .017, h2 .09, and Accuracy, F(1, 60)  23.0, pB.001, h2 .28. Performance for feature search was better when targets appeared in the LVF, and performance for conjunction search was better when targets appeared in the RVF. Figures 2 and 3 illustrate the effect of visual field on response accuracy and response time for these two tasks. A total of 77% of participants were more accurate when feature targets were projected to the right hemisphere, and 74% were more accurate when conjunction targets were projected to the left hemisphere. Looking at each search task separately, we also examined the effects of Set Size (5, 20, 40) on our dependent variables. A widely replicated finding is that Set Size in a feature search task does not affect search performance, presumably because the stimulus set can be searched in parallel. On the other hand, quality of performance in conjunction search typically decreases with set size, presumably because the stimulus set must be searched serially. These typical results were replicated in the present study (see Figures 2 and 3). Neither accuracy nor RT were effected by Set Size in feature search, but both variables were effected in conjunction search*Accuracy: F(2, 120) 25.2, p.001; RT: F(2, 120) 5.4, p .01.

Figure 2. Visual Field and Set Size effects on target-detection accuracy for Feature and Conjunction search tasks.

HEMISPHERIC ASYMMETRIES IN VISUAL SEARCH

719

Figure 3. Visual Field and Set Size effects on target-detection response time for Feature and Conjunction search tasks.

To more clearly show the effect size of visual field on search performance, we looked at the accuracy data for each search task separately, using Visual Field as the sole independent variable. In feature search, visual field accounted for 29% of the variance, F(1, 60) 26.0, p.001; h2 .29, and for conjunction search, 13% of the variance, F(1, 60) 9.2, p.004; h2 .13. We also examined the RT and accuracy data in combination, by dividing the average accuracy score for each participant (proportion correct) by the average RT (seconds), for left and right visual fields. This metric is a measure of the search efficiency. Figure 4 illustrates this metric for each task and visual field. The statistical analysis of this metric showed the same main effect of Search Task, F(1, 60) 268.8, p B.001, and Search TaskVF interaction, F(1, 60) 44.7, p B.001, with the later interaction again showing the different performance asymmetry for feature and conjunction search: feature search efficiency was better for right hemisphere presentations (LVF), and conjunction search efficiency was better for left hemisphere presentations (RVF).

Experiment 2 It could be argued that, in the conjunction search trials, the participants did not have enough time to perform an adequate serial search for targets, and in fact, target detection accuracy was somewhat low in this task (see Figure 2). We therefore conducted a partial replication of Experiment 1 (with new

720

Figure 4.

POYNTER AND ROBERTS

Search efficiency in Feature and Conjunction search.

participants), in which the stimulus patterns were displayed until the participant responded. Arguably this is a weaker test of hemispheric asymmetry in search, since extending the display exposure might have provided time for the participant to move visual fixation to the target location after its initial exposure to either right or left hemisphere. However, feature search performance presented the same visual field effect as in Experiment 1, F(1, 35) 8.1, p.007, with average accuracy of 97.3% for LVF and 93.5 for RVF. Response times were also significantly faster for LVF than RVF, F(1, 35) 4.3, p .05, with average RT of 482 ms for LVF and 514 ms for RVF. The trend for conjunction search was also the same as in Experiment 1, but here the effect of visual field was only marginally significant, F(1, 14)  3.7, p .07, with average accuracy of 81.2% for LVF and 90.5% for RVF. As in Experiment 1, response times were slower in the left versus right visual field (LVF 850 ms; RVF 767 ms), although this difference was not statistically significant.

DISCUSSION The present studies provide evidence that classic visual search tasks do exhibit asymmetric performance, with feature search showing a right hemisphere advantage and conjunction search showing a left hemisphere

HEMISPHERIC ASYMMETRIES IN VISUAL SEARCH

721

advantage. We attribute the feature search asymmetry to a RH advantage in global processing, which has been demonstrated in numerous studies (e.g., Christman et al., 1991; Delis et al., 1986; Kitterle et al., 1990; Lamb et al., 1989; Michimata & Hellige, 1987; Robertson et al., 1993; Van Kleeck, 1989; Weissman & Woldorff, 2005). Performance in our feature search condition was not affected by set size, which is considered evidence for ‘‘parallel’’ search (i.e., the ability to examine the entire field of stimuli at the same time). The participant’s task was to detect a single vertical line within a field of horizontal line distractors*detection of an ‘‘oddball’’ feature sticking out against a background of uniform distractor items. This was a relatively easy task, evidenced by the high accuracy scores and rapid search times. The target perceptually ‘‘popped out’’ of the visual field because it was very different from distractors and no fine-grained analysis was required to detect it. A low-resolution, global scan of the stimulus field was therefore sufficient. Similar to Palmer and Tzeng (1990), we suggest that visual search is supported by a variable-aperture attentional spotlight that can be adjusted to fit the difficulty of the search task*a large-aperture, diffusely distributed attentional spotlight (which is more effectively deployed by the RH) enables parallel search for a highly distinct ‘‘pop-out’’ feature target (in our experiment a vertical line target in a field of horizontal distractors), but a localised, focused attentional beam (more effectively deployed by the LH) is sometimes required to find a target defined either by feature conjunctions (in our experiment colour and orientation), or a subtle difference in feature value (e.g., slight differences in colour saturation). While it seems widely accepted that conjunction search involves serial shifts of localised attention, it is controversial whether global, spatially distributed attention is utilised in feature search. But if attentional resources are involved in search for ‘‘popout’’ features, it seems reasonable to assume that these resources might be distributed globally in order to enable highly efficient, parallel processing. And a number of studies have presented evidence that different modes of processing (global vs local) favour different distributions of attention (broadly distributed versus locally focused) (see Chong & Treisman, 2005; Robertson et al., 1993). Functional neuroimaging studies (e.g., Fink et al., 1996; Weissman & Wodorff, 2005) also support the view of a right hemisphere bias towards global attention and a left hemisphere bias towards local attention. Feature integration theory (Treisman & Gelade, 1980) proposes that feature targets ‘‘pop out’’ of the visual field because they are represented by neurological maps that can be scanned pre-attentively (see also Wolfe, 2010). But a number of studies provide evidence that attentional resources are active in feature search (Carrasco & Yeshurun, 1998; Cheal & Lyon, 1992; Joseph et al., 1997; Kim & Cave, 1995; Nothdurft, 1999; Theeuwes et al., 2008).

722

POYNTER AND ROBERTS

We attribute the LH advantage we found in our feature conjunction condition to a LH advantage in local processing (Christman et al., 1991; Delis et al., 1986; Kitterle et al., 1990; Lamb et al., 1989; Michimata & Hellige, 1987; Robertson et al., 1993; Van Kleeck, 1989; Weissman & Woldorff, 2005). The view that conjunction search requires localised attentional focus is an explicit assumption of feature integration theory and seems widely accepted in the literature. This assumption of item-by-item, serial processing is consistent with the typical finding of a positive relationship between search time and set size in this task (which we replicated in Experiment 1). While the pattern of asymmetry we found has been observed in other studies (Kingstone et al., 1995; Pedersen & Polich, 2001; Polich, 1984; Polich et al., 1990), results depend on task demands, stimulus characteristics, and participant variables (Michael & Ojeda, 2005; Evert et al., 2003; Giesbrecht & Kingstone, 2004; Husain, Shapiro, Marton, & Kennard, 1997). From our perspective one would not always expect feature search to show a RH advantage and conjunction search to show a LH advantage, because the nature of attentional demand is a key factor determining the pattern of asymmetry. A feature search task in which the target is very similar to the distractors (e.g., a 5-degree tilt line target in a field of 10-degree tilt distractors, or an orange target within a field of reddish-orange distractors) might require localised attentional focus; that is, a high-resolution, localised analysis might be necessary to distinguish the subtle difference in feature attributes. And this would likely produce a measurable set size effect and perhaps a LH advantage, even though targets are not defined by feature conjunction, and no feature binding is required. Likewise a conjunction search task in which the target is highly salient (perhaps a familiar face among strangers) might be expected to produce a RH advantage because a low-resolution, global scan of the stimulus field is sufficient to detect the target, even though the target seems to be defined by a combination of features, which would typically require localised attentional focus. So stimulus characteristics, task demands, and participant variables might affect the attentional requirements of the search, which in turn might affect the pattern of search asymmetry. Current theories of attention suggest that selective attention involves several distinct processes, which are controlled by different cortical networks*disengaging spatial attention from one location, orienting/shifting attention to another location, and selectively engaging attention once shifted (see Fan, McCandliss, Sommer, Raz, & Posner, 2002). Therefore asymmetries in search might be expected to vary with those task demands and stimulus characteristics that differentially affect these separate attentional networks. Additionally, individual differences in attentional function might be expected to affect hemispheric differences in search. In fact, a recent study (Poynter, Ingram, & Minor, 2010) found that individuals

HEMISPHERIC ASYMMETRIES IN VISUAL SEARCH

723

who self-reported relatively high levels of attention problems exhibited significantly worse RH performance compared to those with low levels of attention problems. And this finding relates to other studies indicating that patients with RH damage tend to exhibit greater feature conjunction search deficits than those with LH damage (Giesbrecht & Kingstone, 2004; Husain et al., 1997). In some cases it can be argued that the global/local processing distinction is difficult to disentangle from that of low/high spatial frequency processing. The reason is that spatial frequency of the visual stimuli and the global/local nature of the processing task can be difficult to manipulate independently. For example, in tasks that use hierarchical stimuli (e.g., a large letter H, the contours of which are composed of small letter Ts), identifying the global property of the display (letter H) involves low spatial frequency processing because the letter H is relatively large. On the other hand, identifying the local property of the display (the little Ts composing the large H) involves high spatial frequency processing because the letters are small. But in the present study the global/local dimension and low/high spatial frequency dimension were not confounded. The stimuli used in both tasks were identical in terms of spatial frequency characteristics, yet opposite patterns of asymmetry were observed. Another possible explanation for the RH advantage in feature search comes from studies showing that the RH is better at making relatively fine discriminations in stimulus orientation (e.g., discriminating between 30 and 35 degree tilt; see Corballis, Funnell, & Gazzaniga, 2002). But in the present study the difference in orientation between target and distractors was highly distinct (90 degrees), which at least one study has reported results in a LH advantage, not the RH advantage we found (Umilta et al., 1974). Finally we can be relatively confident that the asymmetries we observed in Experiment 1 were not due to individual differences, because the same participants performed both tasks. In summary, we suggest that the asymmetries we observed may be due to a RH advantage in global processing and a LH advantage in local processing. Whether parallel processing in classic feature search involves globally distributed attentional resources is controversial; however, recent discussions of visual search theory (Treisman, 2006) propose the activity of an attentional window that can take on different apertures in order to encompass finely localised details as well as global views of the visual field. And this idea seems to fit well with our view that we adjust our attentional distribution to fit the nature of the search task. When the target stimulus is very distinct, a globally distributed, low-resolution attentional analysis is sufficient, and enables parallel search. But a localised, narrow-aperture attention distribution is sometimes necessary to find targets that either require feature binding, or are very similar to distractor stimuli in terms of compositional attributes. Predicting the dynamics of search in real-world

724

POYNTER AND ROBERTS

scenes is more complicated, in part because the distinction between stimuli that pop-out versus those that require focused attention is not always clear. In real-world environments that we are very familiar with and navigate through in daily life, experience tells us what objects to expect to see and where they will be (the sun in the sky, the red house on the corner, the dog in the yard, etc.). So we can extract a great deal of visual information from the scene without localised attentional focus. Wolfe (2010) presents a very interesting discussion of this topic. Here we offer our speculation that in realworld search individuals might often scan the scene first with a global attentional distribution that provides orientation, enables detection of highsalience stimuli, and areas-of-interest for subsequent search with a more localised attentional focus that enables fine-grained visual analysis. In future studies we would like to investigate asymmetries in visual search using real-world visual scenes, and methods that enable us to directly measure the utilisation of global versus local attentional mechanisms. Additionally we are interested in determining factors that predict individual differences in hemispheric bias, and whether participants can be trained to use global and local search strategies to enhance search performance. Manuscript received 18 August 2010 Revised manuscript received 9 August 2011 First published online 30 November 2011

REFERENCES Carrasco, M., & Yeshurun, Y. (1998). The contribution of covert attention to the set-size and eccentricity effects in visual search. J. Experimental Psychology: Human Perception and Performance, 24, 673
692. Cheal, M. L., & Lyon, D. R. (1992). Benefits from attention depend on the target type in location precued discrimination. Acta Psychologica, 81, 243
267. Chong, C. H., & Treisman, A. (2005). Attentional spread in the statistical processing of visual displays. Perception and Psychophysics, 67, 1
13. Christman, S., Kitterle, F. L., & Hellige, J. B. (1991). Hemispheric asymmetries in the processing of absolute versus relative spatial frequency. Brain and Cognition, 16, 62
73. Corballis, P., Funnell, M., & Gazzaniga, M. (2002). Hemispheric asymmetries for simple visual judgments in the split brain. Neuropsychologia, 40, 401
410. Evert, D. L., McGlinchey-Berroth, R., Verfaellie, M., & Milberg, W. P. (2003). Hemispheric asymmetries for selective attention apparent only with increased task demands in healthy participants. Brain and Cognition, 53, 34
41. Delis, D. C., Robertson, L. C., & Efron, R. (1986). Hemispheric specialization of memory for visual hierarchical stimuli. Neuropsychologia, 24, 205
214. Fan, J., McCandliss, B., Sommer, T., Raz, A., & Posner, M.I. (2002). Testing the efficiency and independence of attentional networks. Journal of Cognitive Neuroscience, 14(3), 340
347. Fink, G. R., Halligan, P. W., Marshall, J. C., Frith, C. D., Frackowiak, R. S. J., & Dolan, R. J. (1996). Where in the brain does visual attention select the forest and the trees? Nature, 382, 626
628.

HEMISPHERIC ASYMMETRIES IN VISUAL SEARCH

725

Foxe, J. J., McCourt, M. E., & Javitt, D. C. (2003). Parietal control of visuospatial attention: Line bisection judgments evaluated with high-density electrical mapping and source analysis. NeuroImage, 19, 710
726. Giesbrecht, B., & Kingstone, A. (2004). Right hemisphere involvement in the attentional blink: evidence from a split-brain patient. Brain and Cognition, 55, 303
306. Husain, M., Shapiro, K., Martin, J., & Kennard, C. (1997). Abnormal temporal dynamics of visual attention in spatial neglect patients. Nature, 385, 154
156. Hyun, L., Woodman, G. F., & Luck, S. J. (2009). The role of attention in the binding of surface features to locations. Visual Cognition, 17, 10
24. Joseph, J., Chun, M., & Nakayama, K. (1997). Attentional requirements in a ‘preattentive’ feature search task. Nature, 379(6567), 805
807. Kingstone, A., Enns, J. T., Mangun, G. R., & Gazzaniga, M. S. (1995). Guided search is a lefthemisphere process in split-brain patients. Psychological Science, 6, 118
121. Kim, M-S., & Cave, K. R. (1995). Spatial attention in visual search for features and feature conjunctions. Psychological Science, 6, 376
380. Kitterle, F. L., Christman, S., & Hellige, J. B. (1990). Hemispheric differences are found in the identification, but not the detection, of low vs. high spatial frequencies. Perception and Psychophysics, 48, 297
306. Lamb, M. R., Robertson, L. C., & Knight, R. T. (1989). Attention and interference in the processing of global and local information: Effects of unilateral temporal-parietal junction lesions. Neuropsychologia, 27(4), 471
483. Michael, G. A., & Ojeda, N. (2005). Visual field asymmetries in selective attention: Evidence from a modified search paradigm. Neuroscience Letters, 388, 65
70. Michimata, C., & Hellige, J. (1987). Effects of blurring and stimulus size on the lateralized processing of nonverbal stimuli. Neuropsychologia, 25, 397
407. Nothdurft, H-C. (1999). Focal attention in visual search. Vision Research, 39, 2305
2310. Palmer, T., & Tzeng, O. J. L. (1990). Cerebral asymmetry in visual attention. Brain and Cognition, 13, 46
58. Pedersen, H. K., & Polich, J. (2001). Hemispheric differences for visual matrix processing: Stimulus size and spatial frequency effects. Brain and Cognition, 47, 525
538. Polich, J. (1984). Hemispheric patterns in visual search. Brain and Cognition, 3, 128
139. Polich, J., DeFrancesco, D. P., Garon, J. F., & Cohen, W. (1990). Hemispheric differences in visual search of simple line arrays. Psychological Research, 52, 54
61. Poynter, W., Ingram, P., & Minor, S. (2010). Visual field asymmetries in attention vary with selfreported attention deficits. Brain and Cognition, 72, 355
361. Robertson, L. C., Lamb, M. R., Egley, R., & Kerth, L. (1993). Spatial attention and cuing to global and local levels of hierarchical structure. Journal of Experimental Psychology: Human Perception and Performance, 19, 471
487. Robertson, L. C., Lamb, M. R., & Zaidel, E. (1993). Interhemispheric relations in processing hierarchical patterns: Evidence from normal and commissurotomized subjects. Neuropsychology, 7, 325
342. Theeuwes, J., Kramer, A. F., & Atchley, P. (1999). Attentional effects on preattentive vision: Spatial precues affect the detection of simple features. Journal of Experimental Psychology: Human Perception & Performance, 25, 341
347. Theeuwes, J., Van der Burg, E., & Belopolsky, A. (2008). Detecting the presence of a singleton involves focal attention. Psychonomic Bulletin & Review, 15, 555
560. Treisman, A. (1998). Feature binding, attention and object perception. Philosophical Transactions of the Royal Society of London B, 353, 1295
1306. Treisman, A. (1999). Solution to the binding problem: Progress through controversy and convergence. Neuron, 24, 105
110.

726

POYNTER AND ROBERTS

Treisman, A. (2006). How the deployment of attention determines what we see. Visual Cognition, 14, 411
443. Treisman, A. M., & Gelade, G. (1980). A feature-integration theory of attention. Cognitive Psychology, 12, 97
136. Treisman, A., & Gormican, S. (1988). Feaure analysis in early vision: Evidence from search asymmetries. Psychological Review, 95, 15
48. Treisman, A., & Souther, J. (1985). Search asymmetry: A diagnostic for preattentive processing of separable features. Journal of Experimental Psychology, General, 114, 285
310. Van Kleeck, M. H. (1989). Hemispheric differences in global versus local processing of hierarchical visual stimuli by normal subjects: New data and a meta-analysis of previous studies. Neuropsychologia, 27(9), 1165
1178. Weissman, D. H., & Woldorff, M. G. (2005). Hemispheric asymmetries for different components of global/local attention occur in distinct temporo-parietal loci. Cerebral Cortex, 15, 870
876. Wolfe, J. (2010). Visual search. Current Biology, 20, 346
349. Umilta, C., Rizzolatti, G., Marzi, C. A., Zamboni, G., Franzini, C., Camarda, R., et al. (1974). Hemispheric differences in the discrimination of line orientation. Neuropsychologia, 12, 165
174.

Copyright of Laterality is the property of Psychology Press (UK) and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.