Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Different attentional mechanisms subserve the attentional blink and visual search Ghorashi, S.M. Shahab 2004

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

Item Metadata

Download

Media
831-ubc_2004-0460.pdf [ 1.19MB ]
Metadata
JSON: 831-1.0099774.json
JSON-LD: 831-1.0099774-ld.json
RDF/XML (Pretty): 831-1.0099774-rdf.xml
RDF/JSON: 831-1.0099774-rdf.json
Turtle: 831-1.0099774-turtle.txt
N-Triples: 831-1.0099774-rdf-ntriples.txt
Original Record: 831-1.0099774-source.json
Full Text
831-1.0099774-fulltext.txt
Citation
831-1.0099774.ris

Full Text

D I F F E R E N T A T T E N T I O N A L M E C H A N I S M S S U B S E R V E T H E A T T E N T I O N A L B L I N K A N D V I S U A L S E A R C H by S. M . S H A H A B G H O R A S H I M.D. Esfahan University of Medical Sciences, 1993 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L M E N T OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF M A S T E R OF A R T S in T H E F A C U L T Y OF G R A D U A T E S T U D I E S (Department of Psychology, Cognitive Science Programme) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y OF BRIT ISH C O L U M B I A August 2004 © S. M . S. Ghorashi, 2004 JUBCl THE UNIVERSITY OF BRITISH COLUMBIA FACULTY OF G R A D U A T E STUDIES Library Authorization In presenting this thesis in partial fulfillment 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. Shahab Ghorashi ' 05/08/04 Name of Author (please print) Date (dd/mm/yyyy) Title of Thesis: Different attentional mechanisms subserve the attentional blink and visual search Degree: MA Year: 2004 Department of Psychology The University of British Columbia Vancouver, BC Canada grad.ubc.ca/forms/?formlD=THS page 1 of 1 last updated: 5-Aug-04 Abstract Joseph et al. (1997) found that a search task that was performed efficiently as a single task was impaired under dual-task conditions. This led to the claim that visual search is not performed preattentively. But that claim is questionable because different criteria (efficiency of visual search and mean level of performance) were used to assess attentional requirements under single- and dual-task conditions. In four experiments, I found that the two measures are affected in different ways by attentional manipulations and, therefore, represent different attentional processes. This finding questions the validity of the evidence used by Joseph et al. to substantiate the claim that all forms of visual search require attention. Table of Contents Abstract : Table of Contents List of Figures Dedication Introduction Experiment l a Method Observers Apparatus and Stimuli Procedure Results and Discussion Experiment lb Method Results and Discussion Experiment 2 a Method Results and Discussion Experiment 2b Method Results and Discussion General Discussion Mean level and slope index different attentional processes Is there such a thing as preattentive processing? References Figure Captions iv List of Figures Figure 1. Schematic diagram of the display 22 Figure 2. Results of Experiment l a 23 Figure 3. Results of Experiment lb 24 Figure 4. Results of Experiment 2a 25 Figure 5. Results of Experiment 2b 26 (5%- wAom G$owe t/ie mecmoncf. 1 Introduction Is there such a thing as preattentive processing in visual search? Or, in the same vein, can some classes of stimuli be processed without attention? The conventional answer to these questions has been in the affirmative (e.g., Treisman & Gelade, 1980). A contrary view, however, has been expressed by Joseph, Chun, and Nakayama (1997) who claimed that there is no such thing as preattentive processing, at least in the context of visual search. This claim was based on the finding that performance on an ostensibly preattentive task suffered under dual-task conditions. Here, we show that Joseph et al.'s conclusion is questionable because two different criteria (efficiency of visual search and mean level of performance) were used to assess attentional requirements under single and dual-task conditions. We suggest that the two criteria represent separate attentional mechanisms that are affected in different ways by attentional manipulations. Visual search is something we do routinely everyday. We search for a car in a parking lot or for a familiar face in a crowd. In laboratory studies of visual search, observers are required to find a target (e.g., the letter T) hidden among distractors (e.g., randomly-rotated letters L) . A common measure of the efficiency of visual search is the time taken to find the target (response time, RT) as a function of the number of items in the display (set size). If R T remains invariant as set size is increased, the search is considered to be efficient. This often occurs when the target and the distractors differ along a simple attribute such as colour or orientation. In this case, the slope relating R T to set size is shallow or flat. Conventionally, these tasks are regarded as preattentive, which means that the stimuli do not require attentional resources to be processed. 2 In contrast, when the target and the distractors differ along two or more attributes, or i f the target is very similar to the distractors, R T increases with set size, resulting in a relatively steep search slope. In this case, the search is said to be inefficient and the search task is said to be attentive, because attentional resources are required to sweep the focus of attention across the items in the display. This conventional distinction between attentive and preattentive processing has been questioned by Joseph et al. (1997) who claimed that all visual search tasks require attention. They argued that i f it is really true that preattentive tasks do not require attention, then performance on these tasks should not depend on the availability of attentional resources. To test this claim they used a dual-task paradigm known as the attentional blink (AB) in which perception of the second of two rapidly sequential targets is impaired i f it is presented within about 500 ms of the first (Raymond, Shapiro, & Arnel l , 1992). It is commonly believed that this second-target deficit occurs because the available attentional resources are preempted by the first target, thus making them unavailable for the second target. Joseph et al. reasoned that no A B deficit should occur i f the second target consists of a stimulus that is processed preattentively and, therefore, does not depend on the availability of attentional resources. In the study of Joseph et al. (1997), the first task was to identify a white letter in a stream of black letters presented in rapid serial visual presentation (RSVP) , a task commonly regarded as attentionally demanding. The second task was to report the presence or absence of an oddly-oriented Gabor patch among uniformly-oriented patches. This task is regarded as preattentive because it yields flat or shallow search slopes when performed as a single task. Joseph et al. reasoned that the preattentive hypothesis would 3 be confirmed i f this oddball task did not exhibit an A B deficit. If, however, an A B deficit were in evidence, the preattentive hypothesis would be disconfirmed. Contrary to the preattentive hypothesis, the results revealed a substantial A B deficit: identification accuracy for the second target was severely impaired at the shorter inter-target lags, and improved rapidly as the lag was increased. The important implication of this result is that at short lags, when all attentional resources are deployed to the first target, processing of even a simple feature such as orientation suffers. Joseph et al. (1997) regarded this outcome as strong support for the hypothesis that all forms of visual search, even those involving simple primitive features, require attention. On the face of it, this is a reasonable conclusion, supported by the empirical evidence. A problem arises, however, when one considers the indices of attention used in Joseph et al. (1997)'s study. To confirm that the orientation-oddball task met the conventional criterion for being preattentive, Joseph et al. performed a preliminary experiment in which the oddball task yielded a flat search function when performed as a single task. In testing the hypothesis that this task did, in fact, require attention, Joseph et al. presented the oddball display as the second target in the R S V P stream, with set size fixed at 12 items. The ensuing A B deficit was interpreted as inconsistent with the preattentive hypothesis. This interpretation, however, is not entirely unambiguous because two different measures of performance were used to index the need for attention in the oddball task when it was performed under single- and under dual-task conditions. The slope of the search function was used under single-task conditions whereas the mean level of accuracy was used under dual-task conditions. A similar approach was taken by 4 VanRul len, Reddy, and Koch (2004) who concluded that visual search and dual tasks draw upon separate attentional resources. A n implicit assumption in the procedure adopted by Joseph et al. (1997) was that the two measures - slope and mean level -indexed the same attentional process. However, it is commonly accepted (e.g., Sternberg, 1969) that the slope and the intercept of the search function can vary independently of one another. This was pointedly noted by Woodman, Vogel , and Luck (2001) who suggested that the slope may index the efficiency of visual search, while the intercept may reflect events, such as stimulus encoding or response selection, that precede or follow the search process. This raises the possibility that the A B deficit in mean level of performance found by Joseph et al. may represent an impairment not in the search process itself but in other processing events that preceded or followed the search process. It goes without saying that this would invalidate Joseph et al.'s tacit assumption that in both the single- and the dual-task conditions the search task drew upon the same attentional resources. The objective of the present work was to resolve this ambiguity by measuring both the mean level and the slope of the search function at each inter-target lag. We reasoned that i f the two measures reflected the same attentional process, then they should be affected in similar ways by the unavailability of attentional resources. Specifically, at short lags, an A B deficit should be instantiated not only in the mean level of performance but also in the efficiency of visual search as indexed by a steeper search slope. On the other hand, i f the two measures are found not to covary, it would have to be concluded that they reflect different - or at least non-identical - attentional processes. 5 Experiments l a and 2a revealed that the slope and the mean level of performance were affected in different ways by the attentional manipulation and, therefore, were assumed to reflect different attentional processes. We explored the generality of this finding in Experiments lb and 2b by using an accuracy-related response measure. E X P E R I M E N T 1A Experiment l a was a close replication of Joseph et al.'s (1997) study with the exception that we measured the slope of the search function as well as the mean level of performance across a range of inter-target lags. A s in Joseph et al.'s study, the second target involved an orientation-oddball task. In addition, to obtain estimates of the search slopes, we used two different set sizes (6 and 12 items), at each of three inter-target lags. Method Observers Twelve University of Brit ish Columbia undergraduates participated for course credit. A l l reported normal or corrected-to-normal vision. Apparatus and Stimuli Stimuli were displayed on an N E C AccuSync 70 colour monitor controlled by an IBM-compatible microcomputer. The background was mid-grey and contained a black fixation cross that subtended 0.5° of visual angle at the centre of the screen. The stimuli consisted of black upper-case letters, a white upper-case letter, and a set of white line segments, each subtending 0.5 x 0.2° of visual angle, spaced regularly around a circle of 2.5° radius, centred at fixation. The line segments were tilted 45° either to the left or to the right. A l l letters subtended 0.6° of visual angle vertically. 6 Procedure A l l displays were viewed from a distance of 57 cm. A t the beginning of each trial the fixation cross was presented in the centre of the screen. Observers initiated each trial by pressing the spacebar, at which point the fixation cross disappeared and the R S V P sequence began after a random delay of 400-800 ms. The R S V P consisted of a sequence of black-letter distractors presented in the centre of the screen. The letters were drawn randomly without replacement from the English alphabet excepting Q. Each letter was displayed for 40 ms and was separated from the next letter by an inter-stimulus interval (ISI) of 50 ms, during which the display was blank. Thus, the stimulus-onset asynchrony (SOA) between successive items in the R S V P stream was 90 ms. The first target was a white letter inserted in the R S V P stream, and was preceded by between 5 and 10 distractors, at random. The second target was a search display comprising either 6 or 12 tilted line segments, as described above. On 50% of the trials, all line segments had the same orientation; in the remaining 50% of the trials, the search display contained one line segment in the opposite orientation (an orientation oddball). There were 3 inter-target lags. A t Lag 0, the second target was presented simultaneously with the first target, as in the study of Joseph et al. (1997). A t Lag 3, the second target was presented 180 ms after the first. A t Lag 8, the second target was presented 720 ms after the first. The R S V P stream continued to be displayed during the inter-target interval. The search display remained on the screen for 5 seconds, or until the observer's response, whichever came first. A n example of the display sequence is presented in Figure 1. 7 The observers' tasks were to identify the first target (i.e., the white letter) by pressing the corresponding key on the keyboard, and to indicate whether or not the search display contained an orientation oddball by pressing the left ( "y e s " ) o r the right ("no") shift-key on the keyboard. The observers were instructed to respond to the search display first, as quickly as possible, and then to identify the first target at their leisure. Each observer performed two 156-trials blocks, one block for a search display of set size 6 and another for set size 12, all counterbalanced for presence or absence of the oddball, the tilt of the target and distractors, and for the inter-target lag. The order of the two blocks was counterbalanced across observers. The first block began with 36 practice trials, and the second block was preceded with 12 practice trials. Results and Discussion In this and all subsequent experiments, estimates of second-target identification were based only on those trials in which the first target was identified correctly. This procedure is commonly used in A B studies on the grounds that, on trials in which the first target fails to be identified, the source of the error is unknown, thus its effect on second-target processing cannot be evaluated. The mean percentage of correct responses for the first target, averaged over set size and lag, was 91.1%. The mean percentage of correct responses for the second target, averaged over set size and lag, was 98.9%. The median R T for each observer was calculated for each of the three lags in the two set-size conditions. These scores, averaged over all observers, are illustrated in Figure 2. A n analysis of variance ( A N O V A ) performed on the scores in Figure 2 comprised two within-subject factors: set size (6 vs. 12) and lag (0, 2, and 8). The analysis revealed a significant effect of lag, 8 F(2,22) = 152.07, p_ < .001, M S E - 5489.91, but neither the effect of set size, F ( l , l l ) = 1.75, rj > .20, M S E = 15219.89, nor the interaction effect between set size and lag, F(2,22) < 1, was significant. A substantial A B deficit was obtained in the present experiment, evidenced by the finding that mean RTs were slowest at Lag 0 and fastest at Lag 8. This result is consistent with the findings of Joseph et al. (1997) who obtained an A B deficit with a similar search task using accuracy as the dependent measure. This result is also consistent with the findings of Jolicceur and Del l 'Acqua (1998) and Kawahara, D i Lol lo, and Enns (2001) who obtained pronounced A B deficits using R T as the dependent measure. It is worth noting, at least in passing, that the mean R T in the single-task condition of Joseph et al. corresponds closely to the mean R T obtained at Lag 8 in the present study (about 600 ms). Thus, the mean R T at Lag 8 provides a credible estimate of single-task performance in the present study. Perhaps the most notable aspect of the results is that the mean RTs and the slopes of the search functions did not covary, as would be expected i f they represented the same underlying process. On the contrary, the results show that the attentional manipulation affected the mean level of performance and the slope of the search function in clearly different ways. As seen in Figure 2, the mean R T decreased progressively as the inter-target lag was increased while the slope remained invariant. On the hypothesis that the mean level and the slope indexed the need for the same attentional resources, the slope should be steeper at the shorter lags, reflecting a loss of efficiency while attentional resources are unavailable. Instead, the search slopes were as shallow at Lag 0, when attentional resources were presumably deployed to the first target, as they were at Lag 8, 9 when all resources were presumably available for the second target. Indeed, consistent with the findings of Bravo andNakayama (1992), the functions in Figure 2 exhibit consistent - though non-significant - negative slopes, suggesting that the orientation-oddball target became more readily visible as the density of the background texPire was increased. This pattern of results strongly suggests that the mean level and the slope of the search function reflect different - or at least non-identical - attentional processes. Considered jointly, the study of Joseph et al. (1997) and the present study indicate that what is impaired in the A B is the overall level of performance, not the efficiency of the visual-search process as indexed by the slope of the search function. This casts doubt on Joseph et al.'s conclusion that all search tasks require attention because this conclusion was based not on the slope of the search function, which is commonly regarded as an index of the search process, but on the mean level of performance, a measure which may not index the search process itself. A s noted by Woodman et al. (2001), the mean level of performance may represent not the search process itself but other processing events, such as stimulus coding and response selection, that precede or follow the visual search. Before speculating on the nature of such other processing events, we first examine the generality of the present results by replicating Experiment l a using an accuracy-related measure instead of R T as the dependent variable. E X P E R I M E N T IB Experiment l b was a replication of Experiment l a except that an accuracy-related measure was used in place of RT . The display sequence was the same as in Experiment l a except that a masking stimulus was presented directly after the second target (i.e., after the search display). The exposure duration of the second target was varied dynamically 10 for each observer by a staircase threshold-seeking procedure (PEST; Taylor & Creelman, 1967) set to converge on a level of 80% correct responses. The dependent measure consisted of the critical duration of the second target at which the observers obtained 80% correct responses. Method Twelve undergraduate students at the University of Brit ish Columbia participated for class credit. Apparatus, stimuli and procedures in Experiment l b were the same as in Experiment l a with the following exceptions. As noted above, the exposure duration of the search display was varied dynamically for each observer by the P E S T procedure. Immediately after the search display, a high-contrast white-noise ring-shaped mask was displayed for 150 ms, covering the imaginary circle on which the search items were located. The observers were instructed to respond to the first target first and then to the second target. None of the responses was speeded. Results and Discussion The mean percentage of correct responses for the first target, averaged over set size and lag, was 91.2%. The mean critical duration of the second target ( D U R c ) for each of the three lags in the two set-size conditions are illustrated in Figure 3, averaged over all observers. A n A N O V A performed on the scores in Figure 3 comprised two within-subject factors: set size (6 vs. 12) and lag (0, 2, and 8). The analysis revealed a significant effect of lag, F(2,22) = 37.99, p_ < .001, M S E = 13857.31, but neither the effect of set size, F ( l , l 1) < 1, nor the interaction effect between set size and lag, F(2,22) = 1.52, p > .20, M S E = 6232.11, was significant. 11 The results of Experiment lb are entirely consistent with those of Experiment la, and confirm the critical finding that the A B deficit is in evidence in the mean level of performance but not in the slope of the search function, whether the dependent measure is RT or DURc. Notably, the finding that the slope of the search function does not reveal an A B deficit is inconsistent with the claim that all visual search tasks require attention. Furthermore, this finding is consistent with the suggestion that the mean level and the slope of the search function index distinct underlying attentional processes. Before reaching a definitive conclusion, however, an alternative account needs to be considered. It may be suggested that the slopes of the search functions in Experiments la and lb remained shallow even at the shortest lags, when all attentional resources were supposedly deployed to the first target, because the orientation-oddball task was too easy. That is, observers might have been able to perform the search task efficiently with what little attentional resources might have been available after the bulk had been deployed to the first target. This is equivalent to saying that, because the task was so easy, the search slopes in both Experiments la and lb remained below a floor level and could not, therefore, reflect the impact of the attentional manipulations. This option was ruled out in Experiments 2a and 2b by replicating Experiments la and lb using a task (finding a "T" among rotated "L"s) known to yield steep search slopes when performed as a single task (Wolfe, 1998). EXPERIMENT 2A Experiment 2a was a replication of Experiment la with the exception that the second target consisted of a search task in which observers reported the presence or absence of a tilted "T" displayed among rotated "L"s. This task is known to yield 12 relatively steep search slopes, unconstrained by a floor level. On Joseph et al. (1997)'s assumption that an A B deficit in mean level reflects the attentional requirement of the visual-search process, a reduction in the available resources should lead to a slower and less efficient search resulting in correspondingly steeper slopes. Specifically, in the A B paradigm search slopes should be relatively steep at short inter-target lags, while attentional resources are deployed to the first target, but less steep at the longer lags, when resources are again available for the second target. In contrast, i f the attentional manipulation is found to affect the mean level but not the slope of the search function, as was the case in Experiment la, the conclusion would gain credence that the two measures represent different attentional processes. Method The design of Experiment 2a was the same as that of Experiment la except that the search display consisted of rotated "L"s instead of the line segments, and in 50% of trials it also contained a tilted "T". The letters in the search display subtended 0.5° of visual angle vertically. If the "T" was present, it could be tilted 45° either to the left or to the right. The task was to report the presence or absence of the "T" among "L"s. Results and Discussion The mean percentage of correct responses for the first target, averaged over set size and lag, was 79.3%. The mean percentage of correct responses for the second target, averaged over set size and lag, was 98.3%. The median RT for each observer was calculated for each of the three lags in the two set-size conditions. These scores, averaged over all observers, are illustrated in Figure 4. An A N O V A performed on the scores in Figure 4 comprised two within-subject factors: set size (6 vs. 12) and lag (0, 2, 13 and 8). The analysis revealed significant effects of lag, F(2,22) = 84.00, rj < .001, M S E = 12575.24, and set size, F ( l , l l ) = 50.87, rj < .001, M S E = 76334.50. The interaction effect between lag and set size was not significant, F(2,22) < 1. The results in Figure 4 exhibit a substantial A B deficit in the mean level of performance, confirming the findings in Experiment l a . As was the case in Experiment l a , however, the slopes of the search functions remained invariant across lags. This buttresses the claim that what is impaired in the A B is the overall level of performance, not the efficiency of the visual-search process as indexed by the slope of the search function. Notably, all slopes were relatively steep, ruling out the possibility that they were constrained by a floor level. In parallel with Experiment l b , the generality of the present finding was examined in Experiment 2b using critical duration of the second target as the dependent measure. E X P E R I M E N T 2B Method The design of Experiment 2b was the same as that of Experiment l b except that the search display consisted of rotated " L " s instead of the line segments, and in 50% of trials, a tilted " T " . If the " T " was present, it could be tilted 45° either to the left or to the right. The task was to report the presence or absence of the " T " among " L " s . Results and Discussion The mean percentage of correct responses for the first target, averaged over set size and lag, was 78.8%. The mean critical duration of the second target ( D U R c ) for each of the three lags in the two set-size conditions are illustrated in Figure 5, averaged over all observers. A n A N O V A performed on the scores in Figure 5 comprised two within-14 subject factors: set size (6 vs. 12) and lag (0, 2, and 8). The analysis revealed significant effects of lag, F(2,22) = 23.48, p. < .001, M S E = 26538.70, and set size, F ( l , l 1) = 47.46, p. < .001, M S E - 35722.01. The interaction effect between set size and lag was not significant, F(2,22)< 1. The results of Experiment 2b are entirely consistent with those of the preceding three experiments. Whether the dependent measure is RT or D U R c , the A B deficit is in evidence in the mean level of performance but not in the slope of the search function. General Discussion The present study questioned the validity of the evidence used by Joseph et al. (1997) to substantiate the claim that all forms of visual search require attention. This claim was based on the finding that the mean level of performance in a search task was impaired under dual-task conditions even though the same task was performed efficiently (i.e., yielded a shallow search slope) when done as a single task. In a series of four experiments, we confirmed Joseph et al.'s findings with respect to mean level, but showed that the efficiency of the search process was unaffected by dual-tasking. This discrepancy between the mean-level and the slope measures can be understood if it is assumed that, as suggested by Woodman et al. (2001), the two measures represent different underlying processes. While the slope of the search function indexes the efficiency of the search process itself, the mean level may be related to processing events that precede or follow the visual-search process. In light of these considerations, the results of the present study cast doubt on Joseph et al.'s conclusion that the deficit in mean level of performance obtained in their experiment represented a corresponding deficit in the process of visual search. 15 A remarkable parallel should be noted between the present results and the results reported by Woodman et al. (2001) who studied the efficiency of visual search in relation to the degree of saturation of working memory. Their observers performed a visual-search task under single-task and dual-task conditions. Under dual-task conditions, the observers had to remember a pattern of coloured patches while performing the search task. The results showed that under dual-task conditions, when working memory was ful l , the mean level of performance on the search task was substantially impaired as compared to the single-task level. However, the efficiency of the search process, as indexed by the slope of the search function, was the same in the dual-task and in the single-task conditions. This pattern of results mirrors that obtained in the present study, and suggests that the memory manipulations of Woodman et al. and the present attentional manipulations may affect the process of visual search in similar ways. Mean level and slope index different attentional processes Collectively, the evidence strongly suggests that the A B and visual search are separable processes that engage different attentional mechanisms. On this basis, it is possible to explain why an A B deficit is obtained in the mean level of performance but not in the efficiency of visual search. In the present experiments, and in the experiment of Joseph et al. (1997), the A B deficit in mean level of performance can be explained on the basis of a processing bottleneck that causes the processing of the second target to be postponed until the processing of the first target has been completed (Chun & Potter, 1985; Jolicceur & Del l 'Acqua, 1998). While so delayed, the representation of the second target is held to be vulnerable to masking or overwriting by the trailing items. A s a 16 consequence, the overall level of performance - whether R T or accuracy - is impaired, thus giving rise to the A B deficit. Our account of why the efficiency of visual search, as indexed by the slope of the search function, is not impaired hinges on the premise that when the second target is masked, the target cannot be identified reliably because there is no viable representation on which to carry out the search process. B y the same token, on those trials in which the second target escapes masking, the search process can be carried out as efficiently as for an unspoiled representation. On this account, what is impaired in the A B is the mean level of performance, not the efficiency of visual search. We hasten to note that our present purpose is not to present a theory of visual search, but merely to point out that a successful visual search cannot be carried out i f the representation of the display has been obliterated by masking. Is there such a thing as preattentive processing? Based on the finding that the mean level of accuracy in a visual-search task was impaired under dual-task conditions, Joseph et al. (1997) concluded that all visual search tasks require attention and that, therefore, there is no such thing as preattentive processing in visual search. The present findings disconfirm this conclusion by showing that, unlike the mean level of performance, the efficiency of visual search does not suffer during the A B . Does this mean that, contrary to Joseph et al.'s conclusion, there is such a thing as preattentive processing? On the face of it, the present results could be regarded as providing an affirmative answer to this question. That is, i f a given search task can be performed as efficiently when attention is not available as when attention is ful ly available, then it is reasonable to 17 surmise that such a task does not require attentional resources (i.e., it is performed preattentively). Although compelling, this is not the only possible interpretation of the present pattern of results or, for that matter, that of Joseph et al. (1997). A n equally plausible interpretation is provided by the postponement account outlined above, namely, that the search task is always carried out when attentional resources are available. This follows from the assumption that at short inter-target lags, when attentional resources are deployed mainly to the first target, the search process is postponed until the necessary resources become available for the second target. In practice, this means that, in the present study and in the study of Joseph et al., the search process may have never been carried out in the absence of attentional resources. Thus, the present findings do not necessarily imply that all search tasks require attention, as claimed by Joseph et al. (1997), nor do they imply that efficient visual search is mediated by built-in analyzers that respond automatically to specific stimulus attributes, as claimed in the conventional view (e.g., Treisman & Gelade, 1980). Rather, these findings are explained naturally by a model in which efficient visual search is said to be mediated by versatile input filters that are dynamically reconfigured so as to handle incoming stimuli with maximum efficiency (Di Lo l lo , Kawahara, Zuvic, & Visser, 2001). Stimuli that fit the characteristics of the input filter are handled efficiently, yielding shallow or flat search slopes possibly indicative of parallel processing. Other stimuli are handled less efficiently and yield correspondingly steeper search slopes, suggestive of serial processing. It is clear from these considerations that, for different reasons, the issue of preattentive processing in visual search could not be addressed in the present study nor in 18 the study of Joseph et al. (1997). A more promising approach might be developed, based on a recent study by Woodman and Luck (2004). They found that, in contrast to an earlier study (Woodman et al., 2001), the efficiency of visual search was impaired (i.e., the slope of the search function was steeper) under memory-load conditions, provided that the information held in working memory belonged to the same category as the information in the search display (e.g., i f both involved spatial details). This working-memory manipulation could be used as a model to examine the issue of preattentive processing in studies involving attentional manipulations. This approach might involve a replication of the present experiments under conditions in which both targets belonged to the same class of stimuli (e.g., both involved the processing of spatial information). If an impairment in the efficiency of visual search were revealed in such a study, the hypothesis of preattentive processing would be disconfirmed, a conclusion that has been reached on separate grounds by D i Lol lo et al. (2001). Such a study, however, is beyond the scope of the present paper. References Bravo, M . J . , & Nakayama, K. (1992). The role of attention in different visual-search tasks. Perception & Psychophysics, 51, 465-472. Chun, M . M . , & Potter, M . C. (1995). A two-stage model for multiple target detection in rapid serial visual presentation. Journal of Experimental Psychology: Human Perception and Performance, 21, 109-127. D i Lol lo , V . , Kawahara, J . , Zuvic, S. M , & Visser, T. A . W. (2001). The preattentive emperor has no clothes: a dynamic redressing. Journal of Experimental Psychology: General, 130, 479-492. Jolicceur, P. & Del l 'Acqua, R. (1998). The demonstration of short-term consolidation. Cognitive Psychology, 36, 138-202. Joseph, J . S., Chun, M . M . , & Nakayama, K. (1997, June 19). Attentional requirements in a preattentive feature search task. Nature, 387, 805-808. Kawahara, J . , D i Lol lo, V . , & Enns, J . T. (2001). Attentional requirements in visual detection and identification. Journal of Experimental Psychology: Human Perception and Performance, 27, 969-984. Raymond, J . E., Shapiro, K. L., & Arnel l , K. M . (1992). Temporary suppression of visual processing in an R S V P task: A n attentional blink? Journal of Experimental Psychology: Human Perception and Performance, 18, 849-860. Sternberg, S. (1969). The discovery of processing stages: Extensions of Donders' method. In W. G . Koster (Ed.), Attention and Performance II Amsterdam: North Holland, pp. 276-315. Taylor, M . M . & Creelman, C. D. (1967). P E S T : Efficient estimates on probability functions. The Journal of the Acoustical Society of America, 41, 782-787. Treisman, A . , & Gelade, G. (1980). A feature integration theory of attention. Cognitive Psychology, 12, 97-136. VanRul len, R., Reddy, L., & Koch, C. (2004). Visual search and dual-tasks reveal two distinct attentional resources. Journal of Cognitive Neuroscience, 16, 4-14. Wolfe, J . M . (1998). Visual search. In Pashler, H. (Ed.), Attention. London, U.K.: University College London Press. Woodman, G . F., & Luck, S. J . (2004). Visual search is slowed when visuospatial working memory is occupied. Psychonomic Bulletin & Review, 11, 269-274. Woodman, G. F., Vogel , E. K., & Luck, S. J . (2001). Visual search remains efficient when visual working memory is full. Psychological Science, 12, 219-224. 21 Figure Captions Figure 1. Schematic diagram of the display sequence in all four Experiment. The display sequence did not include a mask in Experiments l a and 2a. In Experiments 2a and 2b the target consisted of a tilted letter T, and the distractors were randomly rotated letters L. Figure 2. Mean median response times, averaged over all observers, at each set size in Experiment l a , separately for each inter-target lag. Figure 3. Mean critical durations of the second target, averaged over all observers, at each set size in Experiment lb , separately for each inter-target lag. Figure 4. Mean median response times, averaged over all observers, at each set size in Experiment 2a, separately for each inter-target lag. Figure 5. Mean critical durations of the second target, averaged over all observers, at each set size in Experiment 2b, separately for each inter-target lag. 1 st Target 7^ 2 n d Target G . ! •!» aw m • 'i r t-• • % • i , I f « • J fl • • . ••' • b i P - • — Mask Figure 1 23 1000 r , 800 -Inter-target lag (ms) • - • 0 O—O 180 720 600 -400 12 Set Size Figure 2 24 Figure 3 25 1800 z„ 1600 [ 1400 1200 1000 -800 Inter-target lag (ms) 0 Set Size 12 Figure 4 26 o I I 6 12 Set Size Figure 5 

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.831.1-0099774/manifest

Comment

Related Items