UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The role of inter-item symmetry in visual search for shape Roggeveen, Alexa B. 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-0401.pdf [ 1.21MB ]
Metadata
JSON: 831-1.0091107.json
JSON-LD: 831-1.0091107-ld.json
RDF/XML (Pretty): 831-1.0091107-rdf.xml
RDF/JSON: 831-1.0091107-rdf.json
Turtle: 831-1.0091107-turtle.txt
N-Triples: 831-1.0091107-rdf-ntriples.txt
Original Record: 831-1.0091107-source.json
Full Text
831-1.0091107-fulltext.txt
Citation
831-1.0091107.ris

Full Text

THE R O L E OF INTER-ITEM S Y M M E T R Y IN V I S U A L S E A R C H FOR SHAPE By A L E X A B. R O G G E V E E N B.A., Johns Hopkins University, 2001 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF M A S T E R OF ARTS In THE F A C U L T Y OF G R A D U A T E STUDIES D E P A R T M E N T OF P S Y C H O L O G Y We accept this thesjs-^S^onforming to the required standard  THE UN4VERSITY OF BRITISH C O L U M B I A A U G U S T 2003 © Alexa B. Roggeveen, 2003  In presenting  this thesis in partial fulfilment  of the  requirements for an advanced  degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department  or  by  his  or her representatives.  It is  understood  that  copying or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department The University of British Columbia Vancouver, Canada Date  DE-6 (2788)  Abstract Does visual search involve a serial inspection of individual display items (Feature Integration Theory) or are there perceptual processes that group and segregate items from one another prior to their consideration as a possible target (Resemblance Theory)? For example, for targets defined by motion and shape, there is strong support for grouping processes (Kingstone & Bischof, 1999). The present study looked for evidence of grouping based on shape symmetry. Participants searched for target shapes among distractors that were mirror images over either the vertical or horizontal axis. The results indicated: (1) symmetry between items strongly influences search, (2) search was influenced by grouping among target and distractor items in the display, and (3) symmetry was influential in between-distractor grouping only when displays were 'cortically magnified' in order to equate the salience of symmetry across display locations. These results confirm that static shapes are grouped on the basis of their symmetrical similarity to one another, prior to their explicit identification as being either 'target' or 'distractor'. Thus, as with items in motion, static items are grouped and segregated prior to consideration as a target, in accordance with Resemblance Theory.  n  T A B L E OF CONTENTS Abstract  ii  Table of Contents  iii  List of Figures  iv  CHAPTER I  Introduction  1  C H A P T E R II  Experiment 1: Target-Distractor Relations - P's  5  2.1 2.2 2.3  5 5 6  C H A P T E R III  C H A P T E R IV  CHAPTER V  C H A P T E R IV  Introduction Method Results  Experiment 2: Target-Distractor Relations - F's  8  3.1 3.2 3.3  8 8 9  Introduction Method • Results  Experiment 3: Distractor-Distractor Relations - F's  10  4.1 4.2 4.3  10 10 10  Introduction Method Results  Experiment 4: Distractor-Distractor Relations - Scaling  12  5.1 5.2 5.3  Introduction Method Results  12 13 14  General Discussion  15  Bibliography Appendix I  20 Figures  23  iii  FIGURES Figure 1. Example display, Experiment 1  23  Figure 2. Stimuli, Experiment 1  ,  24  Figure 3. Experiment 1, Mean Correct Reaction Time and Proportion Correct  25  Figure 4. Stimuli, Experiment 2  26  Figure 5. Experiment 2, Mean Correct Reaction Time and Proportion Correct  .27  Figure 6. Stimuli, Experiment 3  28  t  Figure 7. Experiment 3, Mean Correct Reaction Time and Proportion Correct.  29  Figure 8. Experiment 4 display - Cortical Scaling  30  Figure 9. Experiment 4, Mean Correct Reaction Time and Proportion Correct  31  IV  Chapter I  Introduction  In the real world, different items compete for our visual attention. How does one select one item (target) from among the many other items (distractors)? Treisman (1986; Treisman & Gelade, 1980) hypothesized that a target defined by a single feature (e.g. a red car among blue cars) would "pop out" from the distractors because single features were processed in parallel without attention. However, a target defined by a conjunction of features (e.g. a small red car among small blue cars and large red cars) would require a serial inspection of items in order to identify the correct conjunction of features, because attention is needed to conjoin features.  '  .  Subsequent studies, however, were problematic for feature integration theory. Investigators reported that parallel, rather than serial, search could occur for a variety of feature conjunctions (Enns & Rensink, 1990; Nakayama & Silverman, 1986; McLeod, Driver, and Crisp, 1988; Triesman, 1988). These new findings were accommodated within feature integration theory by proposing that searching may involve the inhibition or excitation of a shared target feature (Triesman & Sato, 1990; Wolfe, Cave, and Franzel, 1989; Wolfe, 1994). For example, to find a small red bike among small blue bikes and large red bikes, one might inhibit all the large bikes, causing the small red bikes to pop out from among the small blue bikes. Duncan (1995) noted that while such a solution was theoretically plausible, serial models could not accommodate findings that demonstrated the importance of grouping of items in a search display. This point was driven home by a detailed consideration of a study by Driver, McLeod & Dienes (1992), who found that a common direction of motion among target and distractor items in the search display allowed search to proceed  1  more efficiently than when items in the display moved out of phase with one another. Because motion phase is, by definition, concerned with the relationship of items in the display to one another, serial search theories that emphasize attentional allocation on an item-by-item basis do not seem to fit these results. On the other hand, a theory of visual search that emphasizes interactions among items within a display is uniquely poised to explain the effects of motion phase on visual search performance. This is precisely the emphasis adopted by Duncan and Humphreys' (1989, 1992) resemblance theory of visual search. According to resemblance theory, perceptual grouping is critical to understanding visual search performance, and grouping by motion is, in particular, crucial for understanding the Driver et al. (1992) search results. Kingstone and Bischof (1998) undertook an explicit test of this hypothesis. They began by replicating the basic Driver et al. paradigm. However, in addition they also moved a random set of dots in phase or counter phase, in either the target or distractor item motion directions. In this way they were able to test whether a manipulation of item grouping based on motion direction actually influenced search. Their results showed that grouping by motion had a dramatic effect on search, and as such gave strong support in favor of resemblance theory. Similar support for resemblance theory has been advanced by other visual search studies employing motion (e.g., Watson and Humphreys, 1999). Of course, resemblance theory is grounded in the notion that search depends on interactions between stimuli based on features that are shared between the items even when they are not in motion. The aim of the present study was to put this idea to the test.  2  In this study, we tested whether inter-item grouping would occur in a static search display consisting of simple shapes. The potential grouping factor we chose to examine was mirror-image symmetry between items. Shapes that are identical to one another except for being reflected across the vertical axis (e.g., p vs. q) are generally perceived to be more similar to each other than items that are identical except for a reflection across the horizontal axis (e.g., p vs. b) (e.g. Cairns & Steward, 1970, Hershenson & Ryder, 1982, Bagnara, Boles, Simion and Umilta, 1983, Blough & Franklin, 1985). Similarity is indexed in these studies by the degree to which shapes are confused or in the length of time that is required for their discrimination. According to resemblance theory, the fact that visual items that are related by vertical symmetry are easily confused suggests that their similarity might also be the basis of item grouping in a visual search task. Items that are vertically symmetrical to one another, because they are perceived to be more similar, should therefore group more readily than items that are symmetrical to one another about the horizontal axis. Following this line of reasoning, the primary questions of this study were whether between-item symmetry would influence visual search (1) when the symmetry was between target and distractor items, and (2) when the symmetry was only between distractor items. Four experiments were conducted to address these questions. Experiments 1 and 2 examined a visual, search task in which targets were symmetrical to distractors over either a vertical or horizontal axis. The results showed that the axis of symmetry had a strong influence on search. Experiment 3 used a similar design, but instead of varying the symmetry relations among targets and distractors, the symmetry relations among two  3  different distractors were examined. The results from this experiment revealed no significant influences of symmetry. In Experiment 4, the same distractor relations were tested with search displays that were cortically magnified (i.e. stimuli size was increased with eccentricity), in an effort to ensure that the symmetry relations among items were equally salient at all visual field locations. The results showed that search for a target among distractors that were mirror images of one another was more difficult than search for the same target among items related by vertical symmetry. Therefore, these results provide strong support for the perceptual grouping of items on the basis of symmetry, prior to their identification as either the target or a distractor.  4  Chapter II  Experiment I: Targe t-Distractor Relations - P's  Experiment 1 was designed to determine whether similarity between targets and distractors based on axis of symmetry affects performance in visual search. Targets were symmetrical to distractors over either a vertical or a horizontal axis. As vertical symmetry between items causes a greater perception of similarity than horizontal symmetry, it was anticipated that greater target-distractor similarity would cause more perceptual grouping, making the search process less efficient. Method Participants. Twelve participants (4 male and 8 female), aged 18 to 28 from the University of British Columbia undergraduate subject pool participated in this experiment. Eleven participants were right handed; one was left-handed. A l l participants reported normal or corrected-to-normal vision. Participants received one unit of course credit for one hour of participation. Stimuli. Stimuli were presented on a Macintosh iMac computer, with 800 x 600 screen resolution (95 Hz), using VScope software (Enns & Rensink, 1991). In order to explore the target-distractor relationship, the stimuli were identical to one another in shape, but differed in their orientation. (See Figure 2.) The target (a lower-case letter p) was symmetrical to distractors over either a vertical axis (a lower-case letter q) or a horizontal axis (a lower-case letter b). The stimuli were rotated 90 degrees for half of the participants to control for effects due to familiarity with letters. When rotated, the symmetry relationship between the target and distractors shifted: a rotated lower-case p was now symmetrical to a rotated lower-case q over the horizontal axis, and symmetrical  5  to a rotated lower-case b over a vertical axis. A l l stimuli subtended a visual angle of 0.804 by 0.452 degrees. Procedure. Participants performed a typical visual search task, in which they were presented with an array of items, and were asked to determine whether a specific target was present (indicated by the "/" key on a standard keyboard) or absent (indicated by the "z" key). Participants performed 10 blocks of 60 trials (600 trials). The display remained on the screen until participants made a response. After a response, feedback was given for 472.5 ms, in the form of a plus sign (+) if the response was correct, and a minus sign (-) if the response was incorrect. The number of items in the display varied between 8, 16, and 24 elements . Trials were mixed such that the target-distractor symmetry relationship and display size were both randomly assigned on each trial. Results Results are presented in Figure 3. They clearly show that target-distractor grouping was influenced by symmetry. When the target was symmetrical to the distractors over a vertical axis, inverse efficiency scores were higher than when the symmetry was across the horizontal axis. To arrive at this conclusion, a 2x2x3 analysis of variance (ANOVA) was performed on inverse efficiency score data. Axis of symmetry (vertical or horizontal), orientation of the items in the display (upright or rotated), and display size (8, 16, 24) were the within-subject independent factors involved in the analysis. Reaction time (RT) and accuracy were the dependent variables. Because the measures of RT and accuracy reflected the same patterns, for the purposes of presentation the two measures were combined in the form of inverse efficiency scores (Townsend & Ashby, 1983). This  6  involves forming a ratio of correct RT over proportion correct, for each observer and condition, which creates a compact easily interpretable index of performance. Assuming a linear relationship between RT and accuracy, inverse efficiency scores are especially useful when error rates are variable across conditions. They are interpreted in the same manner as RT data, because if accuracy is perfect, RT remains the same; as accuracy decreases, RT is modified to accommodate for lesser performance. The stimuli used in this experiment did not translate effectively to an experiment where distractor symmetry was manipulated. To use these stimuli, one would pair p's and b's to create horizontal symmetry among distractors, and p's and q's to create vertical symmetry among distractors. Attempts to create a target that both shares enough features with the distractors to keep target-distractor similarity high enough to make search demanding, yet not be related to either of the distractors around an axis of symmetry were unsuccessful. Thus, we were unable to test, using these stimuli, whether the same perceptual grouping between items based on inter-item symmetry extends to distractordistractor relationship. These results, then, led us to Experiment 2.  7  Chapter III  Experiment 2: Target-Distractor Relations -F's  In order to ensure that our results were not stimulus-specific, and to use stimuli that were better suited to a distractor-distractor experiment, a second experiment exploring target-distractor relations was conducted, using the same design as in Experiment 1. These stimuli were able to be manipulated for an experiment examining distractor-distractor relations in a manner that allowed search to be both difficult enough to not be a pop-out search, and to not be related around an axis of symmetry. Method Participants. Eighteen participants (7 male and 11 female), aged 19 to 30 from the University of British Columbia undergraduate subject pool participated in this experiment. Sixteen were right-handed; two were left-handed. A l l participants reported normal or corrected-to-normal vision. Participants received one unit of course credit for one hour of participation. Stimuli. Stimuli were presented using the same technology as in Experiment 1. A n upper case letter F, rotated 90 degrees to the left, was used as the target. Distractors were upper case F's, rotated 90 degrees to the left, and then flipped over either the horizontal or vertical axis, (see Figure 4). To ensure that effects due to letter familiarity were not an issue, only rotated items were used in this experiment, because if items were upright, only the target, and not the distractors, would be considered a Standard English letter. Additionally, because sideways orientation significantly more difficult than upright orientation in Experiment 1, rotation would ensure that the task was difficult enough to not be a pop-out search. Stimuli subtended a visual angle of 0.452 by 0.804 degrees. Procedure. The procedure in Experiment 2 was identical to that of Experiment 1.  8  Results Results are shown in Figure 5. As in Experiment 1, the results show that targetdistractor grouping is influenced by axis of symmetry. Correct reaction time is significantly higher and proportion correct significantly lower in the condition where the target and distractors are symmetrical about a vertical axis. A 2x3 A N O V A was performed on the correct reaction time data and the proportion correct data. No two-way interaction was found between axis of symmetry and display size for correct reaction time (F(2, 34) = 1.137 p > .3). However, a main effect of the axis of symmetry was Sj  demonstrated for correct reaction time ( F ( l , 17) = 4.748, p < .05), as well as for display size for correct reaction time (F(2, 17) = 100.550, /? < .001). A two-way interaction between axis of symmetry and display size was found for proportion correct (F(2, 34) = 8.213,/? < .05). Also, a main effect was found for the axis of symmetry (F(\, 17) = 37.59,/? < .001), as for display size (F(2, 17) = 31.86,/? < .001). These results replicate Experiment 1, again showing that target-distractor relationships based on symmetry influence visual search results. This experiment also confirms that these results hold for two different sets of shapes.  9  .Chapter IV  Experiment 3: Distractor-Distractor Relations -F's  Having replicated the results of Experiment 1 using stimuli conducive to a distractor-distractor study, we designed an experiment to look at how the distractordistractor relationship may be influenced by inter-item symmetry. We hypothesized that that perceptual grouping between distractors would be greater when distractors were symmetrical to one another over a vertical axis, making search easier than if the distractors were symmetrical over a horizontal axis. Method Participants. Sixteen participants (2 male and 14 female), aged 19 to 40 from the University of British Columbia undergraduate subject pool participated in this experiment. A l l participants reported normal or corrected-to-normal vision. A l l were right-handed. Participants received one unit of course credit for one hour of participation. Stimuli. The technology used in this experiment is identical to what was used in Experiment 1. The symmetrical relationship was varied between the distractors in the display, unlike Experiments 1 and 2, where the symmetry was between the target and distractors. The target, again, was an F that had been rotated 90 degrees. Distractors were "hybrid" F's in the rotated orientation. (See Figure 6.) The "hybrid" F's were symmetrical to one another around either the horizontal or vertical axis. Stimuli subtended a visual angle of 0.754 by 0.804 degrees. Procedure. The procedure in Experiment 3 was identical to that of Experiment 1. Results Results are shown in Figure 7. There was no significant difference in performance on a visual search task when the relationship between distractors varies according to axis  10  of symmetry. A 2x3 A N O V A was performed on the correct reaction time and proportion correct data. No significant effect was found in the two-way interaction between axis of symmetry and display size (F(2, 30) = .014,/? > .50). There was also no significant main effect of either correct reaction time (F(l, 15) = .076, p > .50) for the axis of symmetry. There was a significant main effect for display size in correct reaction time (F(2, 15) = 89.467,/?<.001) In the proportion correct data, there was no significant interaction between axis of' symmetry and display size (F(2, 30) = .953,p > .50). There was also no significant main effect of axis of symmetry (F(\, 15) = .210, p > .50). However, there was a significant main effect for display size in proportion correct (F(2, 15) = 13.493,/) < .001).  11  Chapter 5  Experiment 4: Distractor-Distractor Relations - Scaling  Why did we not see an effect of distractor-distractor symmetry in Experiment 3? According to resemblance theory, one would anticipate that greater similarity between distractors would lead to greater perceptual grouping - and faster and more accurate performance. However, sensitivity to symmetry drops as eccentricity increases (as items become further away from fixation), so grouping of the distractors may have been impaired by an inability to perceive the symmetrical relationship in the periphery (Gurnsey, Herbert & Kenemy, 1998). When a scaling function is applied, however, symmetry detection at greater eccentricities is facilitated (Sally & Gurnsey, 2001). Cortical scaling, or cortical magnification, is an increase in size according to a linear function as an item is located farther away from fixation. Scaling is intended to minimize the effects of increased receptive field size in the areas of the retina outside of the fovea, the effect of which accounts for lesser resolution in the periphery. Cortical magnification has also been found to effectively eliminate issues of retinal field eccentricity in visual search (Carrasco & Frieder, 1997). This was demonstrated in an experiment which showed that the eccentricity effect, or less efficient search as the target is farther away from the center of the display, was eliminated using cortical magnification (Carrasco, Evert, Chang & Katz, 1995). Thus, in Experiment 4 we cortically scaled the displays used in Experiment 3, making items farther from fixation larger in a linear fashion. Presumably, removing any issues of an inability to perceive symmetry relations in the periphery would allow  12  symmetry relations to have an influence on search in the manner predicted by resemblance theory. Method Participants. Thirty-six participants (12 male and 24 female), aged 18 to 27 from the University of British Columbia undergraduate subject pool participated in this experiment. Thirty-one were right handed, and 5 were left handed. A l l participants reported normal or corrected-to-normal vision. Participants received one unit of course credit for one hour of participation. Stimuli. Stimuli were presented on a Macintosh eMac computer, with 800 x 600 screen (112 Hz) resolution, using VScope software (Enns & Rensink, 1991). Similar to Experiment 3, a rotated F was used as the target, and rotated "hybrid" F's, which were symmetrical to one another around either the horizontal or vertical axis, were used as the distractors. However, unlike in Experiment 3, distractors increased in size in a linear fashion as they were presented farther away from the central fixation. (See Figure 8.) Stimuli subtended visual angles of 0.603 by 0.804 degrees, 1.457 by 1.859 degrees, or 3.869 by 3.717 degrees, depending upon their distance from the central fixation. Procedure. The procedure was the same as in Experiments 1, 2, and 3, with some important differences. Participants were instructed to fixate on a central fixation cross, which appeared for 240 ms, and then, when the search display appeared, to look for the target, a rotated F, among rotated "hybrid" Fs. Display sizes varied between 6, 12, and 18 items. The target could appear at any of three distances away from the central fixation cross. The location of distractors were varied randomly among the three possible distances away from fixation between trials.  13  Results Results are shown in Figure 9. With the use of cortical magnification, axis of symmetry had an effect on search performance. The task was much more difficult than in Experiments 1, 2, and 3: the correct reaction times ranged from 1300 to 2300 ms. The mean correct reaction times for trials where the distractors were related across the vertical axis were significantly higher than those trials where the distractors were related across the horizontal axis, showing that vertically symmetric distractors were more difficult to search among than horizontally symmetrical distractors. A 2x3 A N O V A was performed on the correct reaction time and proportion correct data. In the correct reaction time data, a significant two-way interaction was found between axis of symmetry and display size (F(2, 70) = 4.152,/? > .0198). A significant main effect was found for the axis of symmetry for correct reaction time (F(\, 35) = 29.952,/? < .001). There was also a significant main effect for display size for correct reaction time (F(2, 35) = 233.783,/? < .001). A significant two-way interaction was also found between axis of symmetry and display size for proportion correct (F(2, 70) = 3.437,/? < .05). A main effect approaching significance for the axis of symmetry for proportion correct (F(l, 35) = 3.91 \,p > .05). A main effect was also evident for display size for proportion correct (F(2, 35) = 54.432,/? < .001).  14  Chapter 6  General Discussion  From these results, there are clear answers to our two original questions concerning inter-item symmetry, and its role in visual search. First, between-object symmetry influences the process of comparison between the target image and the distractor items. Experiments 1 and 2 supported this, showing that search for a target was affected by the axis of symmetry it shared with distractors. Finding a target that is vertically symmetric to distractors is significantly more difficult than finding a target that is horizontally symmetric to distractors. This is consistent with the current understanding of symmetry and similarity: objects that are vertically symmetrical to one another are perceived to be more similar than objects that are horizontally symmetrical to one another. This result fits nicely into the framework of resemblance theory (Duncan & Humphreys, 1989, 1992). When targets are more similar to distractors, perceptual grouping occurs more easily, making search more difficult. When targets are more dissimilar from distractors, perceptual grouping occurs less easily, making search easier. Here, perceptual grouping occurred between vertically symmetric items more readily than between horizontally symmetric items. Our second question asked whether between-object symmetry also influences the process by which distractor items are segregated from the target. The answer is yes - but only when items are cortically magnified to ensure equal visibility at all locations in the display. In Experiment 3, the distractors were symmetrical to one another around either the vertical or horizontal axis. Perceptual grouping, such as we saw in Experiments 1 and 2, occurs more easily between items that are symmetrical to one another around a vertical  15  axis. Therefore, it would make sense that when the distractors were vertically symmetrical, it would be easier to find the target. Conversely, it would be more difficult to find the target when distractors were horizontally symmetric, because perceptual grouping would occur less easily. However, the results of Experiment 3 were varied from these predictions: we found no significant difference between the two symmetries. This may have occurred because symmetry relations are difficult to perceive in the periphery without magnification. The cortical magnification displays in Experiment 4 resulted in an outcome which was completely consistent with our understanding of symmetry and perceived similarity. It was more difficult to locate the target among distractors that are vertically symmetric to one another, rather than among distractors that are horizontally symmetric. These results translate directly to an interpretation by resemblance theory, since it was more difficult to find the target among distractors that were vertically symmetrical to one another (stronger grouping) than it was to find the target among distractors that were horizontally symmetrical to one another (weaker grouping). Static items vs. items in motion Studies exploring grouping among items in motion (e.g. Driver et al, 1992; Kingstone & Bischof, 1999; Watson & Humphreys, 1999) have offered some very clear support for resemblance theory. In resemblance theory, grouping between items in the display is understood as part of a continuum of two independent factors: (1) grouping between target and distractors and (2) grouping between distractors themselves. Both the Driver et al. and Kingstone & Bischof studies showed that coherent motion between distractors made search faster than when the target, and items which shared the same  16  motion as the target, moved coherently. Additionally, in the Kingstone study, when perceptual grouping was enhanced by the random dots in the display, distractor grouping continued to see an advantage over target-distractor grouping. Correct reaction time was faster than when the random dots moved with the target. Distractor grouping, therefore, affects search for items in motion to a greater degree than target-distractor grouping. However, in our study, the opposite relations for the strength of target and distractor grouping held true. Grouping between the target and distractors based on symmetry had a significant effect on both the speed and accuracy of search. The manipulation of similarity based on symmetry between targets and distractors had clear influence on search efficiency. Grouping between distractors, however, was not as strong. In Experiment 4, grouping between distractors based on inter-item symmetry was weaker than in Experiments 1 and 2. Global vs. local processing Results for Experiment 4 were counter to what one might expect from the predictions of resemblance theory, which state that when distractors are more similar to one another, search is both faster and more accurate. In our results, the opposite occurred: when the distractors were vertically symmetrical, search was slower. Some explanation may possibly be found in a close examination of the stimuli used, and their relationship to one another. Looking back at Figure 8, one can observe that the global shape of the target (an L on its side) is shared with half of the distractors present in the display. It is wellestablished that the perception of global shape is more accessible than local shape (e.g. Navon, 1977). The precedence of global shape in human visual perception allows for the possibility that the global shape of the items in the display served as a preliminary way to  17  group items in the search space. The similarity between the target and half of the distractors means that the global shape of the other half of the distractors is vertically or horizontally symmetric to the global shape of the target/Thus, the target, in sharing the global shape with half of the distractors, would group with those distractors - and, in turn, group more readily with those distractors which share a more similar global shape. In this case, that would be those distractors that were related around a vertical axis to one another. The symmetry relationship between the distractors is also, in part, between the target and half of the distractors. Therefore, the same symmetry effects that occurred in Experiments 1 and 2 are also occurring in Experiment 4, but are muted because of the heterogeneity of the distractors. Gestalt contributions To further understand the anomalous results of Experiment 4, it may be useful to turn to a theory of search display configuration, proposed by Banks and Prinzmetal (1976). Following up on an experiment performed by Banks, Bodinger and Illige (1974), they found, using what they termed to be "F-T's", that the configuration of the display had a large effect on search results. The targets used were identical to the distractors which were used in Experiments 3 and 4. When the F-T's were close together and away from the target, they had little effect on target detection. However, if they were in a cluster that included the target, they had a large effect on target detection. Our displays, while random, tended to display items in a group, rather than separated from one another, especially as display sizes increased. Banks and Prinzmetal then turned to the Gestalt principles of organization, and argued that the more that a target is perceptually grouped with "noise," or, in this ease,  18  distractors, the more difficult it becomes to differentiate the target from the noise with which it has been grouped. To put it simply, the target gets lost among the distractors. Perhaps in our Experiments 3 and 4, the level of noise in the vertical displays was greater because of greater similarity between the distractors. Other work considering Gestalt grouping in search may also have some relevance to this study. Giving more support to Banks and Prinzmetal's work, Treisman (1982) proposed that perceptual grouping is performed by preattentive mechanisms, showing that when items in a search display are grouped according to Gestalt principles, these groups, rather than the items themselves, were assessed as a singular chunk. Conclusions Our results show that perceptual grouping, such as is proposed in resemblance theory, does occur for static shapes. However, the grouping that occurs is not precisely analogous to the grouping that occurs due to coherent motion in a search display. Rather, target-distractor relationships appear to have precedence over distractor-distractor relationships when items in the display are related according to axis of symmetry. While valuable to acknowledge that some of these effects may occur due to global precedence and Gestalt grouping, it is still worth noting that target-distractor relations, when based on inter-item symmetry, follow the guidelines set forth by resemblance theory. Future research should include work that eliminates the potential effects of global'shape and unintentional spatial grouping.  19  Bibliography Bagnara, S., Boles, D.B., Simion, F., & Umilta, C. (1983). Symmetry and similarity effects in the comparison of visual patterns. Perception and Psychophysics, 34(6), 578-584 Banks, W.P., Bodinger, D., & Illige, M . (1974). Visual detection accuracy and targetnoise proximity. Bulletin of the Psychonomic Society, 2, 411-414 Banks, W.P. & Prinzmetal, W. (1976). Configurational effects in visual information processing. Perception and Psychophysics, 19 (4), 361-367. Blough, D.S. & Franklin,-J.J. (1985). Pigeon discrimination of letters and other forms in texture displays. Perception and Psychophysics, 38(6), 523-532. Cairns, N . U . & Steward, M.S. (1970). Young children's orientation of letters as function of axis of symmetry and stimulus alignment. Child Development, 41(4), 9931002. Carrasco, M . , Evert, D.L., Chang, I. & Katz, S.M. (1995). The eccentricity effect: target eccentricity affects performance on conjunction searches. Perception & Psychophysics, 55(8), 1241-1261.  ~  Carrasco, M . & Frieder, K. (1997). Cortical magnification neutralizes the eccentricity effect in visual search. Vision Research, .57(1), 63-82. Duncan, J. & Humphreys, G. (1989). Visual Search and Stimulus Similarity. Psychological Review, 96 (3), 433-458. Duncan, J. & Humphreys, G. (1992). Beyond the search surface: visual search and attentional engagement. Journal of Experimental Psychology: Human Perception and Performance, 18 (2), 578-588.  20  Driver, J., McLeod, P. & Dienes, Z. (1992). Modelling the role of parallel processing in visual search. Cognitive Psychology, 22, 255-271.  Eastwood, J.D., Smilek, D., & Merickle, P.M. (2001). Differential attentional guidance by unattended faces expression positive and negative emotion. Perception & Psychophysics, 63(6), 1004-1013. Enns, J.T. & Rensink, R.A. (1990). Influence of scene-based properties on visual search. Science, 247, 721-723. Enns, J.T. & Rensink, R.A. (1991). VSearch Color: Full-color visual search experiments on the Macintosh II. Behavior Research Methods: Instruments and Computers. 23(2), 265-272. Gurnsey, R., Herbert, A. & Kenemy, J. (1998). Bilateral symmetry embedded in noise is ' detected accurately only at fixation. Vision Research, 38, 3795-3803. Hershenson, M . & Ryder, J. (1982). Perceived symmetry and visual matching. American Journal of Psychology, 95 (4), 669-680. Kingstone, A. & Bischof, W. (1999). Perceptual grouping and motion coherence in visual search. Psychological Science, 10 (2), 151-156. McLeod, P., Driver, J., & Crisp, J. (1988). Visual search for conjunction of movement in form in parallel. Nature, 332, 154-155. Nakayama, K., & Silverman, G. (1986). Serial and parallel processing of visual feature conjunctions. Nature, 320, 264-265. Navon, D. (1977). Forest before trees: the precedence of global features in visual perception. Cognitive Psychology, 9(3), 353-383. '  21  Sally, S. & Gurnsey, R. (2001). Symmetry detection across the visual field. Spatial . Vision, 14 (2), 217'-234. Treisman, A. (1986). Features and objects in visual processing. Scientific American, 255, 106-115. Triesman, A. (1988). Features and objects: The fourteenth Bartlett memorial lecture. The Quarterly Journal of Experimental Psychology, 40A, 201-237. Treisman, A., & Gelade, G. (1980). A feature integration theory of attention. Cognitive Psychology, 12, 97-136. Triesman, A., & Sato, S. (1990). Conjunction search revisited. Journal of Experimental Psychology: Human Perception and Performance, 16, 459-478. Treisman, A., Sykes, M . , & Gelade, G. (1977). Selective attention and stimulus integration. In S. Dornic (Ed.), Attention and performance VI. Hillsdale, N J : Lawrence Earlbaum. 333-361. Watson, D.G., & Humphreys, G.W. (1999). Segmentation on the Basis of Linear and Local Rotational Motion: Motion Grouping in Visual Search. Journal of Human Perception and Performance, 25(1), 70-82. Wolfe, J.M. (1994). Guided search 2.0: A revised model of visual search. Psychonomic Bulletin & Review, 7(2), 202-238. Wolfe, J.M., Cave, K.R., & Franzel, S.L. (1989). Guided search: A n alternative to the feature integration model for visual search. Journal of Experimental Psychology: Human Perception and Performance, 75,419-433.  22  Figure 1: Example display, Experiment 1  b b  b b b b b b b b b b b P b b b b b b b b b b  23  Figure 2: Stimuli, Experiment 1 Upright Orientation  Sideways Orientation Target  ,  _•  •_ srtical axis  a  !  Distractor  Distractor  >  24  nonzoniai axis  Figure 3: Experiment 1, Mean Correct Reaction Time and Proportion Correct Experiment 1: Average Correct Reaction Time  1600 i  —  display size  Experiment 1: Average Proportion Correct 1  .6  1  i  8  16  1  1  1  r  24  8  16  24  display size  25  Figure 4: Stimuli, Experiment 2  26  Figure 5: Experiment 2, Mean Correct Reaction Time and Proportion Correct  Experiment 2: Mean Correct Reaction Time  1600 1500 1400 '1300  1200 1100  c o  % 1000 CD  900•  800  vertical axis  O horizontal axis  700  i 16  24  display size  Experiment 2: Mean Proportion Correct  .98 .96  TJ .94 CD 1_ !_  O O c o '•e o a. o  .92  •  .86 .84  vertical axis  O-horizontal axis "T"  16  8  display size  27  24  Figure 6: Stimuli, Experiment 3  Vertical axis  Hnriynntal avis  28  lure 7: Experiment 3, Mean Reaction Time and Proportion Correct  Experiment 3: Mean Correct Reaction Time  1600 1500 1400 1300 E  1200  c o  1100  TO  1000 900  •  vertical axis  O horizontal axis  800 700  —r  —r 16  -  8  24  display size Experiment 3: Proportion Correct  .995 \.99 h o  to O c g t o  .985 H .98  O  .975 .97 h  •  vertical axis  o  horizontal axis  .965  8  16 display size  29  24  Figure 8: Experiment 4 display - Cortical Scaling  30  Figure 9: Experiment 4, Mean Correct Reaction Time and Proportion Correct Experiment 4: Mean Correct Reaction Time  2600 1  :  2400 h  1200 6  12  18  display size  Experiment 4: Proportion Correct  .99 h  .87  u  1  1  T  6  12.  18  display size  31  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0091107/manifest

Comment

Related Items