Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Spatial attention and metacontrast unmasking : integration of the two solitudes Lamenza, Ernesto A. 1998

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

Item Metadata

Download

Media
831-ubc_2001-0575.pdf [ 2.31MB ]
Metadata
JSON: 831-1.0089998.json
JSON-LD: 831-1.0089998-ld.json
RDF/XML (Pretty): 831-1.0089998-rdf.xml
RDF/JSON: 831-1.0089998-rdf.json
Turtle: 831-1.0089998-turtle.txt
N-Triples: 831-1.0089998-rdf-ntriples.txt
Original Record: 831-1.0089998-source.json
Full Text
831-1.0089998-fulltext.txt
Citation
831-1.0089998.ris

Full Text

Spatial attention and metacontrast unmasking: Integration of the two solitudes by . Ernesto A . Lamenza B . A . , Carleton University, 1995 A THESIS S U B M I T T E D I N 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 O F 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) We accept this thesis as conforming toJh§ rejjuired standard -T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A September 1998 © Ernesto A . Lamenza, 1.998 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 DE-6 (2/88) Abstract This thesis claims that metacontrast unmasking is influenced by attentional orienting towards the target location. This view is contrary to Breitmeyer, Rudd and Dunn (1981), who proposed that metacontrast unmasking is the product of inhibition of the primary mask's transient signal by the sustained signal of a secondary mask. A series of experiments demonstrate the thesis using a task in which observers discriminated the missing corner of a target diamond. Experiments 1 and 2 replicated metacontrast masking and unmasking, respectively, experiment 3 illustrated that contour proximity had no influences on unmasking, contrary to dual-channel inhibition theory. Experiments 4 and 5 indicated that metacontrast unmasking was influenced by spatial orienting. We propose an addition of attention to dual-channel theory as it is incomplete with regards to metacontrast unmasking. T A B L E O F C O N T E N T S Abstract ''. i i List of Tables v List of Figures vi Attention and metacontrast unmasking: Integration of the two solitudes 1 The first solitude: Metacontrast Metacontrast masking 1 Metacontrast unmasking 1 Explanation for metacontrast masking 2 Explanation for metacontrast unmasking 3 Electro-physiological evidence 3 Psychophysical evidence 6 The second solitude: Spatial Attention 6 Nakayama and Mackeben's sustained-transient model of attention 7 Experimental Hypothesis 8 Experiment 1 9 Experiment 2 11 Discussion ; 13 Comparison between experiments 1 and 2 14 Experiment 3 15 Discussion 16 Experiment 4 - Unpredictable location 19 Discussion 22 Experiment 5 - Predictable location 23 Discussion 26 Combined analysis of experiments 4 and 5 26 i i i Discussion 27 General Discussion 27 Integration of the two solitudes of attention and metacontrast 31 References 33 iv List of Tables Table 1. Properties of Sustained and Transient Cells 36 Table 2. Results of Experiment 1: Raw Data (% mean and SD) 37 Table 3. Results of Experiment 2: Raw Data (% mean and SD) 38 Table 4. Results from Experiment 3: Raw Data (% mean and SD) 39 Table 5: Results from Experiment 4: Raw Data (% mean and SD) 40 Table 6: Results from Experiment 5: Raw Data (% mean and SD) 41 V List of Figures Figure 1. Graphical illustration of metacontrast masking results and stimuli 42 Figure 2. Graphical illustration of metacontrast unmasking results and stimuli 43 Figures 3a and 3b. Results from Hoffman, Stone and Sherman 44 Figure 4. Results from Nakayama and Mackeben 45 Figure 5. Stimuli used in all experiments 46 Figure 6. Results from experiment 1 . .47 Figure 7. Results from experiment 2 48 Figure 8. Combined results from experiments 1 and 2 49 Figure 9. Results from experiment 3 50 Figure 10. Results from Breitmeyer, Rudd and Dunn 51 Figure 11. Stimuli position for experiment 4 52 Figure 12. Overall results from experiment 4 53 Figure 13. Masking and unmasking results from experiment 4 54 Figure 14. T only and inattentional masking results from experiment 4 55 Figure 15. Stimuli position for experiment 5 56 Figure 16. Overall results from experiment 5 57 Figure 17. Masking and unmasking results from experiment 5 58 Figure 18. T only and inattentional masking results from experiment 5 59 Figure 19. Combined analysis curve of experiments 4 and 5 60 Figure 20. Results from Mii l ler and Findlay's experiment 1 61 Figure 21. Results from Mii l ler and Findlay's experiment 2 62 Figure 22. Results from Mii l ler and Findlay's experiment 4 63 vi Attention and metacontrast unmasking: Integration of the two solitudes. Is attention powerful enough to produce an increase in the visibility of a masked object? Previous research has not considered spatial attention as a mechanism that can influence low level visual processes. Indeed, the two fields of attention and low level vision are analogous to the two Canadian solitudes of French Canada and English Canada. We however wish to bridge these research areas by proposing a model that can encompass both solitudes. The first solitude: Metacontrast Metacontrast masking. Metacontrast is a term to describe "... the reduction in the visibility of one briefly presented stimulus, the target, by a spatially adjacent and temporally succeeding, briefly presented second stimulus, the mask. A s such metacontrast is a form of backward masking in so far as the masking stimulus exerts a retroactive effect on the target stimulus." (Breitmeyer, 1984, pp. 4). Our stimuli consisted of a target diamond with a missing left or right corner (T), and a diamond frame as a mask ( M l ) , which fit snugly around the target (Figure 1). Target-mask spatial separation is crucial. The strongest masking of the target occurs when the outer contour of the target and the inner contour of the mask are adjacent. Increases in spatial separation produce a proportional and linear decrease in target masking. Temporal considerations are also of the utmost importance. Metacontrast masking occurs when the mask occurs 5 0 - 100 msec after the target (Figure 1). The typical U shaped curve shown in figure 1 is the signature of metacontrast masking. Insert Figure 1 about here Metacontrast unmasking. Metacontrast deals not only with masking, but also with a phenomena called unmasking. Unmasking involves the increase in the visibility of a target, 1 which would have otherwise been masked. Unmasking requires that a second mask (M2) occur along with the target (T) and mask ( M l ) . Typically T and M l are fixed at a specific time interval, which produces maximum masking. M 2 in turn is varied temporally before and after the T - M l arrangement (Figure 2). The end result is an inverted U-shaped curve, which is typical of unmasking. Insert Figure 2 about here Explanation for metacontrast masking. Metacontrast masking is a rich topic, which has produced innumerable theories (See Breitmeyer, 1984 for a brief review). Only three theories of metacontrast masking are applicable here (Breitmeyer and Ganz, 1976; Matin, 1975; Weisstein, Ozog, Szoc, 1975) because they all use the same framework to explain metacontrast masking: The sustained-transient channel dichotomy. A s indicated in Breitmeyer and Ganz (1976), electro-physiological research has revealed two classes of cells, sustained and transient cells. The properties of these cells are listed in Table 1. Interactions occur between sustained and transient cells due to the connections existing between them (Hoffman, Stone & Sherman, 1972; Singer & Bedworth, 1973). It is these interactions that produce metacontrast masking. For masking, Breitmeyer and Ganz (1976) argue for a form of interchannel inhibition whereby the transient signal from the mask inhibits the sustained signal from the target; this theory is called the dual channel interaction theory. Specifically, Breitmeyer and Ganz (1976) argue that a brief visual stimulus presented to an observer excites both sustained and transient neurons, which in turn produce separate sustained and transient responses. The transient neurons are first to respond with a brief burst of activity followed by the sustained neurons. This temporal difference is due to the shorter latency of the transient neurons in comparison to the longer latency of the sustained neurons. Thus when the mask is presented after the target, the sustained and transient responses 2 coincide temporally with the M l transient response overwhelming the T sustained response, thus producing metacontrast masking. Breitmeyer and Ganz suggest that optimal masking should occur when the mask is presented 50 to 100 msec after the target. This 50 to 100 msec figure comes from neurophysiological evidence that suggests transient activity precedes sustained activity by 50 to 100 msec (Dow 1974; Ikeda & Wright, 1975). Insert Table 1 about here Explanation for metacontrast unmasking. The task of science is to explain natural phenomena. Occam's Razor says the most parsimonious explanations are the best. Therefore it should come as no surprise that Breitmeyer attempted to explain metacontrast unmasking using the theory developed for metacontrast masking. Recall that metacontrast masking is said to be the product of transient-on-sustained inhibition, where the transient signal of M l inhibits the sustained signal o f T. Breitmeyer (1978) and Breitmeyer, Rudd, and Dunn (1981) argued that metacontrast unmasking was a case of sustained-on-transient inhibition. There were two lines of evidence used in support of sustained-on-transient inhibition, electro-physiological and psychophysical. Electrophysiological evidence. Hoffman et al. (1972) and Singer and Bedworth (1973) tentatively argued for a form of inhibition described as sustained-on-transient inhibition. Hoffman et al. (1972) performed electro-physiological recordings of dorsal L G N cat cells. One aspect of the electro-physiological recordings involved the convergence of sustained and transient responses from the retina onto dorsal L G N cells. The experiment is as follows: A combined total of 21 sustained and transient dorsal L G N ganglion cells were tested. The object was to test the recovery of responsiveness of each cell by measuring the number of times a cell discharged when presented with a series of 10 paired shocks. The two shock stimulus consisted 3 of a conditioning (first) shock which was varied in fixed steps from 1 volt to 3.2, 7.5 and 10 volts. The intensity of the test (second) shock was at or just above threshold. The ISI between stimuli was varied from 0 to 100 msec. Figure 3a indicates the response of a d L G N transient cell to the series of 10 paired shocks. The y-axis is the number of times the cell responded, (up to 10) and the x-axis is the ISI between shocks. The following are apparent from the graph: a) A s the voltage is increased, the latency of the transient cell decreases; b) the initial number of responses decreases. A s ISI is increased along with the voltage, the cell responsiveness decreases and begins at a shorter latency. Figure 3b is a frequency threshold histogram of all the d L G N cells tested. The x-axis is threshold and the y-axis illustrates the number of cells. The histogram illustrates the number and type of cells which discharge related to the voltage threshold. The clear bars indicate sustained cells and the darkened bars indicate transient cells. A s indicated, sustained cells fire when the voltage threshold is above 3.2 volts. Hoffman et al. contend that sustained cells start to discharge once the conditioning (first) shock voltage is increased to 7.5 volts and 10 volts. Hoffman et al. argue that the drop in the transient cell responsiveness is due to inhibition by sustained cells, because the voltage threshold is now high enough to produce a discharge from the sustained cells. What is also important is that sustained and transient d L G N cells receive inputs from sustained and transient cells, (i.e., the two different afferents are not independent of each other) which allows sustained-on-transient inhibition and transient-on-sustained inhibition, at least at the level of the d L G N . Several counterpoints occur from Hoffman et al.'s study. First, the time course of the recovery of responsiveness of each cell varied considerably. Second, inhibition was apparent only when the test shock was close to threshold for some cells. Third, some cells responded too unreliably to the optic chiasmic stimulation. Fourth, the increased reduction in responsiveness with increased voltage could be due to nothing more than the increase in the voltage to the cell and not due to sustained inhibition. 4 Insert Figure 3 about here Singer and Bedworth (1973) also investigated sustained and transient L G N cat cells via electro-physiological recordings. The aim of their investigation was to further clarify Hoffman et al.'s results of inhibitory interactions between sustained and transient cells in the L G N . Intracellular and extracellular recordings were performed and intrageniculate inhibition was measured via intracellular IPSPs. Mesencephalic reticular formation stimulation was used, and it reduces L G N inhibition. Thus for the extracellular recordings, L G N responses obtained with and without conditioning stimulation of the M R F was used in an assessment of L G N inhibition. In order to conclude the existence of sustained-on-transient inhibition, then the authors argued that an IPSP within the cellular recording of a transient or sustained cell should be visible whenever the stimulus intensity exceeded the higher threshold of the sustained cells. Please recall that the sustained and transient afferent connections interact at the L G N , thus allowing for sustained-on-transient and transient-on-sustained inhibition at the L G N , as indicated by Hoffman et al. (1972). Singer and Bedworth only occasionally found an IPSP within the cellular recordings. A previous paper co-authored by Singer (Singer, Poppel, Creutzfeldt; 1972) suggests that sustained inhibition from prolonged light stimuli could be used as evidence for the existence of sustained inhibition. Using the evidence from Singer, Poppel, Creutzfeldt (1972), and the current evidence, Singer and Bedworth (1973), conclude that only indirect evidence exists for sustained-on-transient inhibition. The above papers on d L G N inhibition in the cat indicate that sustained inhibition is not easily disentangled from the overall transient response. This is due to the higher threshold of the sustained cells, their smaller receptive fields and the response amplitude which is not as large as for the transient cells. Indeed the evidence indicates that the sustained cell response is embedded 5 within the transient cell response making it difficult to discern a specific sustained response. A s such, one can only conclude that there is weak evidence for sustained-on-transient inhibition. Psychophysical evidence. Breitmeyer et al. (1981) argued for reciprocal inhibition, i.e., masking was due to transient-on-sustained interaction (the transient response from M l inhibited the sustained response of T) and unmasking then is due to sustained-on-transient interaction (the sustained response of M 2 inhibits the transient response of M l ) . The typical masking paradigm could not detect the existence of the sustained-on-transient interaction because it is weaker than transient-on-sustained inhibition. Therefore Breitmeyer et al. (1981) used the unmasking paradigm to investigate the spatial and temporal parameters of sustained-on-transient inhibition. Breitmeyer et al's (1981) experiment 2 kept T and M l fixed at an S O A of 45 msec and varied the M 2 S O A in 30 msec steps from -180 to +180 msec. The results revealed an increase in the visibility of the target at the negative SOAs (Figure 2) with respect to M 2 . The authors argued that the long response persistence of sustained channels (up to 200 msec, Cleland, Levick and Sanderson, 1973; Meyer and Maguire, 1977) could have played a role in the extended unmasking interval obtained at the negative SOAs . The combined evidence suggests that unmasking is due to the sustained signal from M 2 inhibiting the transient signal from M l . Recall that it is the transient signal of M l , which causes masking. Because of the longer latency of the sustained channels, M 2 must be presented before the T - M l stimuli in order to produce unmasking. The properties of metacontrast unmasking are the same as those for masking: Increases in M 2 - M l spatial separation produce marked decreases in unmasking (Breitmeyer et al., 1981); stimuli orientation and location must be the same. The second solitude: Spatial Attention. Attention is a difficult concept to define. Part of the problem lies in the difficulty of measuring attention. Because of this, attention can be conceived of as a metaphor for a form of 6 mental energy, which cannot be directly perceived. Rather its effects can be discerned upon human behavior. For example, an increase in the speed of mental processing is observed when attention is allocated to a location. (Posner, Nissen, Ogden, 1978). The observer must orient their attention towards the source of sensory input in order to obtain this increase in processing speed (Posner, 1980). Posner (1980) describes two types of orienting: Covert orienting which involves a shift in attention without any head/eye/body orientation (Coren, Ward, Enns, 1994) and overt orienting which is a shift in attention with an accompanying head/eye/body orientation. Optimal increases in mental processing speeds occur when cues are used. There are two types of cues: Exogenous and endogenous cues (Wright & Ward, 1994). Exogenous cues are capable of producing reflexive responses, such that the observer involuntarily focuses attention towards the cued location in space and are exemplified by abrupt onsets near the target location which directly indicate target location. Endogenous cues convey target location such that the observer must interpret their meaning so as to determine the target location. A n example of an endogenous cue is a centrally presented arrow, pointing either left or right. Endogenous and exogenous cue effects differ as a function of cue target S O A . Exogenous cues tend to produce peak performance at cue-target S O A s of approximately 100 msec (Wright & Ward, 1994; Nakayama & Mackeben, 1989) whereas endogenous cues tend to produce more gradual performance improvements from 0 to 300 msec with sustained performance after 300 msec (Wright & Ward, 1994). Thus researchers must be aware of differences in performance due to the interaction between cue type and temporal parameters. Nakayama and Mackeben's sustained-transient model of attention. Recent work on the separable components of visual attention yields some interesting results applicable to metacontrast unmasking. Nakayama and Mackeben (1989) investigated the transient and sustained components of attention using a visual search paradigm and a cue. The authors investigated the properties of these two components using a percentage correct performance 7 measure. Experiment 3 used a visual search task, where an exogenous cue (outline of a square) indicated the target location within an 8 x 8-element matrix. Subjects fixed their gaze upon a fixation marker that was present as long as the cue and target were present. Subjects were then presented with the cue followed by the target within the 8 x 8 matrix. A 250 msec mask terminated the trial. The cue-target S O A was varied up to 500 msec, and the target duration was varied for three durations, 33, 83 and 117 msec. The cue always remained on during the presentation of the stimulus array. The authors discovered an initial monotonic rise in percentage correct from a cue-target S O A of 0 to approximately 50 msec, with maximum performance at a cue-target S O A of approximately 50 msec. Performance stayed at a high level at a cue-target S O A of approximately another 50 to 100 msec and then was dependent on stimulus duration. The shorter the stimulus duration, the greater the drop in performance and the lower the overall performance. Figure 4 illustrates the results obtained from Nakayama and Mackeben's experiment3. Insert Figure 4 about here The results obtained by Nakayama and Mackeben are interesting for two reasons: 1) They are consistent with results obtained with attentional cueing, i.e. an increase in performance when a 100% predictive cue is presented before the target; 2) the graphic similarity of Nakayama and Mackeben's results (Figure 4) with that of the metacontrast unmasking results (Figure 2). In both cases, there is an increase in performance when the cue is presented before the target. Thus the temporal parameters suggest some commonality between attention and metacontrast. Is there a connection between these two solitudes? Experimental Hypothesis Breitmeyer et al. argued that the increase in performance seen in their unmasking 8 experiment when a preceding stimulus directly indicated the target location arose from inter-channel interactions. Nakayama and Mackeben argue that the increase in performance seen in their experiment when a preceding stimulus directly indicated the target location arose from transient attention. The key for Nakayama and Mackeben is that a cue is used to indicate the target position before the appearance of the target. Breitmeyer et al.'s unmasking experiments presented the M 2 stimulus before the target T. Note that all the stimuli used in Breitmeyer et al.'s experiments were presented at a single location. Thus one may hypothesize that M 2 acts as a cue. If so, then a test of Breitmeyer and Ganz's dual channel theory is necessary. Experiment 1 Experiment 1 was a replication of Breitmeyer et al.'s (1981) masking study. Breitmeyer et al. used subjective rating scales as a measure of subject performance. While acceptable, there is no way of knowing whether the ratings obtained for masking are of the same unit scale size as the ratings obtained for unmasking, making comparisons between the two scales difficult. Therefore, I w i l l use a percentage correct measure as a measure of subject performance. Not only w i l l this allow direct comparison between the replications of masking and unmasking, it is also the same performance measure used by Nakayama and Mackeben (1989), thus allowing for consistency in performance measures between their attentional results and this thesis' masking/unmasking experiments. Method Observers A total of 12 U B C students participated in this experiment. A l l observers in all the experiments had normal or corrected-to-normal vision. A l l observers in experiment 1 received course credit for their participation. Apparatus and stimuli 9 A Power Macintosh computer running VScope software was used in all the experiments. The stimuli were black diamond shapes, and were presented on a white computer screen. Figure 5 illustrates all the stimuli used in all the experiments. The Michaelson contrast ratio for the stimuli was 0.935. A l l stimuli were presented for 15 msec. The mean viewing distance was 63.3 cm. The target (T) was 59.7' of visual angle, measured along the vertical extent and missing the left or right corner (3.9'). The second stimulus ( M l : 109.7'), had an open center and fit snugly around the target. The spatial separation between T and M l was 1.9'. The spatial separation for all the stimuli for all the experiments was measured along a diagonal offset 45° from the central vertical axis. Insert Figure 5 about here Procedure On each trial, the observer discriminated whether the left or right corner of T was missing. Subjects were told to fixate upon the center of the screen and to discriminate whether the left or right corner of T was missing. If the left corner was missing, the ' z ' key was pressed. If the right corner was missing the 7 ' key was pressed. The temporal relation between T and M l was varied across 10 differentially spaced SOAs , as follows: -135 msec, -90, -60, -45, 0, +45, +60, +90, +135, and +225 msec. Thus the M l S O A was the independent variable and the observer's discrimination performance, measured as the percentage correct number of responses, was the dependent variable. A l l stimuli were presented at the center of the screen. The intertrial interval was 675 msec. Trial by trial feedback was given at the end of each trial in the center of the screen. Subjects had 60 practice trials followed by 8 blocks of 60 trials each. A t the end of each block a percentage incorrect score was presented and at the end of 8 blocks a percentage incorrect score was given for the entire experiment. 10 Results Experiment 1 replicated the result from Breitmeyer et al.'s experiment 1, which used subjective ratings as their performance measure. A s shown in Figure 6, there was a large difference in performance between the 0 and 45 msec S O A conditions (87.5% vs. 62.4%) with worst overall performance at an S O A of +45 msec. Table 2 lists the means for the masking replication. The +45 msec S O A was the condition where maximum masking of the target occurred. Stimulus onset asynchrony was analyzed in a 1 factor A N O V A with repeated measures on the S O A factor. The main effect of S O A was significant, F(8,88) = 19.7, p < 0. Insert Figure 6 about here Insert Table 2 about here Experiment 2 Experiment 2 was a replication of Breitmeyer et al.'s unmasking study. Percentage correct was used as a performance measure rather than the subjective rating scales used by Breitmeyer et al. Method Observers A total of 13 U B C students participated in experiment 2. The data from four observers were dropped due to chance responding. A l l observers received course credit for their participation. Apparatus and stimuli The apparatus was the same as in experiment 1. However, a new stimulus, M 2 , was 11 added. Figure 5 indicates the M 2 stimulus. M 2 (186.6') had an open center and in turn fit snugly around M l . The spatial separation between the inner contour o f M 2 and the outer contour of M l was 1 pixel (1.9') and between the inner contour of M 2 and the outer contour of T was 8 pixels (15.2'). Procedure A s in experiment 1, on each trial the observer discriminated whether the left or right corner of T was missing. The data from experiment 1 indicated maximum masking at the +45 msec T - M l S O A , therefore T and M l were presented at a fixed +45 msec S O A . T - M l alone was the baseline/control condition. M 2 in turn was presented at 10 different S O A s with respect to T: -300 msec, -150, -90, -45, 0, +45, +90, +135, +195, and +345 msec. The M 2 S O A was the independent variable and the observer's discrimination performance was the dependent variable. The inter-trial interval was 900 msec. A l l other aspects of experiment 2 were the same as experiment 1. Results Experiment 2 replicates the unmasking result from Breitmeyer's experiment 2. A s shown in Figure 7, the highest level of performance occurred at -90 msec (83.3%). This particular S O A produced the highest target visibility, thus maximum unmasking occurred here. Note the horizontal line at 62.4%. This is the control condition from experiment 1, where maximum masking occurred at an S O A of +45 msec. Table 3 lists the means for the unmasking replication. Stimulus onset asynchrony was analyzed in a 1 factor A N O V A with repeated measures on the S O A factor. The main effect of S O A with a correction for non-sphericity was significant, F(9,72) = 4 .2 ,E<-01. Insert Figure 7 about here 12 Insert Table 3 about here Discussion Experiment 1 Experiment 1 was a successful replication of Breitmeyer et al.'s (1981) masking study. The results are graphically displayed in Figure 6. The maximum and minimum performance obtained was at the 0 and +45 msec SOAs respectively (90.2% and 63%). The difference between the two was 27.2 %. Experiment 2 Experiment 2 successfully replicated Breitmeyer et al.'s (1981) unmasking study. The results are graphically displayed in Figure 7. The maximum amount of unmasking was at the -90 msec S O A (83.3%) and the minimum obtained was.at the +90 msec S O A (69.7%). The horizontal arrow is the baseline S O A condition ( T - M l S O A o f 45 msec) from experiment 1. Several interesting points can be noted from the unmasking graph. First, minimum performance occurs at +90 msec. A t this point, the stimulus sequence is T followed by M l followed by M 2 with all stimuli separated by a 45 msec S O A . However, absolute minimum unmasking would be expected at the +45 msec S O A because, with M 2 and M l are presented simultaneously 45 msec S O A after T, the combined effect o f M 2 and M l would produce the greatest amount of masking than at any other S O A . But the standard error bars indicate no significant difference between the two points; therefore any difference between the +45 and the +90 msec S O A could be due to error. I f error is not a factor, then one could argue that at the +90 msec S O A , M 2 alone disrupts the processing of the T and M l stimuli, thus producing greater masking than when M 2 and M l are presented simultaneously. M 2 could either interrupt the processing o f T and M l or it integrates with T and M l . Overall, this is an example of backward masking of T by M 2 . 13 The last three positive S O A s (+135, +195, +345 msec) are nearly equal in performance and suggest a common effect between them. Performance is higher than for the +45 or +90 msec S O A . One possibility is the previous explanation, namely M 2 is an abrupt onset which disrupts the processing of the M l stimuli. However, in this case, the opposite effect is the case, namely an increase in the visibility of T, i.e., unmasking. Here one could argue that M 2 interrupts the processing of M l , i.e., this is an example of backward masking o f M l by M 2 . Comparison between experiment 1 and 2 Figure 8 compares the results from experiments 1 and 2. Two things are apparent. First the amount of unmasking is not the same as the amount of masking. Breitmeyer et al.'s subjective ratings were not able to pick this up because the scale for the masking and unmasking experiments were different, thus comparisons are not possible with subjective ratings. The difference between the masking and unmasking results indicates that unmasking is weaker than masking. This correlates well with the weaker effect associated with sustained-on-transient inhibition. Alternatively, this suggests a different mechanism for unmasking, namely attention, as indicated in this thesis. This argument w i l l be discussed fully in the general discussion. Insert Figure 8 about here The second point is that the minimum performance for masking (at the +45 msec SOA) not equivalent with the minimum performance for unmasking (at the +90 msec SOA) . Rather, the entire unmasking curve is shifted up. The addition of M 2 produces an increase in the visibility of T across the entire range. Thus unmasking does not simply occur at the negative SOAs but at the positive ones, as indicated previously. This thesis w i l l discuss unmasking only the negative SOAs . Again, as I indicated previously, the suggested reason for unmasking at the positive SOAs is a combination of backward masking o f T - M l and M l alone by M 2 . Or, the is 14 difference between the two curves could be simply due to error. Regardless, this thesis w i l l discuss unmasking at the negative SOAs under the assumption that a fuller and better explanation occurs when one attempts to answer the separate parts o f a problem than the whole problem. Experiment 3 Experiment 3 was a test of Breitmeyer and Ganz's dual-channel theory. Please recall our argument that M 2 is a cue. Breitmeyer et al.'s (1981) experiment 4 indicates that an increase in the spatial separation between M 2 and M l produced a resultant decrease in unmasking. If metacontrast unmasking is specifically due to sustained-transient channel interactions, then increasing the spatial separation between M 2 and M l should reduce the amount o f unmasking. However, I f M 2 acts as a cue, then an increase in the spatial separation should have a negligible effect on unmasking. Thus experiment 3 varied the spatial separation between the inner contour of M 2 and the outer contour of M l . Method Observers Experiment 3 tested whether spatial separation affected metacontrast unmasking. A total of 20 U B C students participated in experiment 3. The data from four observers were dropped due to chance responding. A l l observers received course credit or $5 for their participation. Apparatus and stimuli The apparatus was the same as in experiment 2. However a new stimulus was added, M 2 -large. Thus experiment 3 had two variants o f the same stimulus, M 2 (186.6') and the new M 2 large (248.3'). The spatial separation between M 2 large and M l was 10 pixels (19.6') and between M 2 large and T was 17 pixels (32.6'). Procedure The procedure was the same as in experiment 2. M 2 and M 2 large were varied across the 15 following 9 SOAs: -720 msec, -360, -180, -90, 0, +90, +180, +360, and +720 msec. A new S O A was added, separate from the other 9 S O A s and distinct from the previous unmasking experiments. The 10th S O A was the baseline/control condition ( T - M l with M l presented 45 msec after T). Neither M 2 nor M 2 large was presented to the observer in this particular S O A . Experiment 3 therefore obtained simultaneous measures of masking and unmasking, contrary to the previous unmasking experiment that used the baseline masking results o f experiment 1 as a control. Results Figure 9 graphically illustrates the obtained results. Table 4 lists the obtained means for experiment 3. A 2 (Spatial separation: 1.9' vs. 19.6') x 10 (SOA) A N O V A was used to analyze the data, with repeated measures on both factors. Only the S O A factor was significant, [F(9, 135) = 2.53, p. = .01]. Insert Figure 9 about here Insert Table 4 about here Discussion The rationale for experiment 3 tested dual-channel theory by increasing the spatial separation between M 2 and M l . Dual-channel theory suggests a radical decrease in unmasking. The results disconfirmed dual-channel theory and suggest that M 2 is a cue, orienting the subject towards the position of T. Arguments for and against dual-channel theory. Dual-channel theory uses sustained-on-transient inhibition to explain unmasking at long to moderate negative S O A s . Here, the sustained 16 signal from M 2 interferes with the transient signal from M l to produce unmasking. Breitmeyer et al. (1981) illustrated this argument by increasing the spatial separation between M 2 and M l from 0' to 68' and found a dramatic decrease in unmasking at 8.5' with no unmasking at 34' (Figure 10). Breitmeyer et al. (1981) also argued for another form of channel interaction called transient-on-transient inhibition. This is a form o f intrachannel inhibition which occurs at short, negative SOAs (up to 60 msec; Winters & Hamasaki, 1975, 1976) and across large spatial ranges (0' to 68'; Breitmeyer et al., 1981). Thus according to Breitmeyer et al. (1981) dual-channel theory can predict two forms of channel inhibition: 1) Sustained-on-transient interchannel inhibition at long to moderate negative SOAs , 2) transient-on-transient intrachannel inhibition at short negative SOAs . However, our results indicate no significant drop in unmasking at long to moderate negative S O A s when M 2 - M 1 spatial separation increased from 1.9' to 19.6'. This result is contrary to Breitmeyer et al.'s sustained-on-transient explanation of metacontrast unmasking. Wi th regard to transient-on-transient inhibition, our results indicate that it is not a factor as we obtained maximum unmasking at an S O A of -90 msec. However this result does not disconfirm transient-on-transient inhibition as the M 2 - M 1 S O A was -135 msec (M2-T S O A o f -90 msec + T - M l S O A of 45 msec). This is outside of the range of transient-on-transient inhibition. Thus the results are inconclusive as regards to transient-on-transient inhibition. Insert Figure 10 about here Breitmeyer (1978) and Breitmeyer et al. (experiment 3, 1981) kept M 2 on continuously when testing for unmasking, whereas we presented M 2 briefly. Continuous presentation of M 2 produces a continuous sustained signal (only transient signal present is at the initial presentation of M2) to inhibit the signals of M l , whereas a brief presentation of M 2 produces both transient and sustained signals. Thus our manipulation contained sustained and transient signals, whereas 17 Breitmeyer (1978) and Breitmeyer et al.'s (experiment 3, 1981) manipulation, after the initial abrupt onset of M 2 , produced only a continuous M2-sustained signal. One might argue that our results do not disconfirm dual-channel theory because our experiment has both sustained-on-transient and transient-on-transient inhibition. But i f transient-on-transient inhibition only works at short negative S O A s then how does one explain the unmasking at the longer SOAs? According to dual channel theory, only sustained-on-transient inhibition occurs at long to moderate S O A s even when both sustained and transient signals are present. Thus any increase in spatial separation between M 2 and M l at long to moderate S O A s should produce a resultant decrease in unmasking. Breitmeyer et al. (1981) suggested that sustained-on-transient inhibition has a spatial range o f up to 17'. The spatial separation between M2-large and M l was 19.6' and between M 2 and M l was 1.9'. Thus the spatial range for experiment 3 was 19.6' - 1.9' = 17.7' and we still obtained unmasking at long to moderate SOAs. Thus we maintain the argument that unmasking is not due to sustained-on-transient inhibition. Some other mechanism is responsible for unmasking. We propose that this mechanism is not low-level channel interactions but rather involves M 2 acting as a cue. Failure to replicate Breitmeyer et al.'s spatial separation experiments. One interesting question is why our spatial separation results did not replicate Breitmeyer et al.'s spatial separation results from experiments 3 and 4. Three reasons are apparent: a) Breitmeyer et al. used subjective ratings as a performance measure, whereas we used percentage correct; b) Breitmeyer et al. only used two S O A s whereas we used 10 SOAs; c) Breitmeyer et al. used two subjects for experiment 3 and only one for experiment 4 whereas we used 16. Our replication of Breitmeyer et al. produced had more statistical power, and the performance measure allowed for comparison between experiments. Also the range of SOAs used in our experiments gave a broader measure of the range of unmasking with variations in spatial separation. Thus any failure to replicate Breitmeyer et al. is due to the lack of power and small effect range for Breitmeyer et 18 al.'s experiments. What kind of cue is M 2 ? If M 2 is a cue, then it must share some of the properties of a cue. There are two types of cues, each one having different properties: Endogenous cues and exogenous cues (Wright & Ward, 1994). Exogenous cues are capable of producing reflexive responses, such that the observer involuntarily focuses attention towards the cued location in space. These cues are exemplified by abrupt onsets near the target location which directly indicate target location. Endogenous cues convey target location such that the observer must interpret their meaning so as to determine the target location. A n example of an endogenous cue is a centrally presented arrow, pointing either left or right. Endogenous and exogenous cues also differ as a function of cue target S O A . Exogenous cues tend to produce peak performance at cue-target SOAs of approximately 100 msec (Wright & Ward, 1994; Nakayama & Mackeben, 1989) whereas endogenous cues tend to produce more gradual performance improvements from 0 to 300 msec with sustained performance after 300 msec (Wright & Ward, 1994). Our experiment 3 indicates that unmasking occurs within a 720 msec time frame. Thus M 2 could be an endogenous cue. But M 2 is an exogenous cue because it directly indicates T location. Thus the question is: What type of cue is M 2 ? Experiment 4 - Unpredictable location The goal of this experiment was to randomly vary the spatial location of the stimuli such that M 2 validity was at chance (50%) by having each stimulus appear at one of two locations. Jonides (1980) indicates that an endogenous cue's effectiveness is directly proportional to its validity, i.e., the lesser an endogenous cue's validity, the lesser its effectiveness. If M 2 is a purely endogenous cue, then unmasking should not occur when M 2 validity is at chance. Method Observers Experiment 4 tested whether unpredictable spatial location affected metacontrast 19 unmasking. A total of 13 U B C students participated in experiment 4. Three students were dropped due to chance responding. A l l students received course credit for their participation. Apparatus and stimuli The apparatus and stimuli were the same as in experiment 2. Procedure The procedure was the same as per experiment 2 except for the following difference: The stimuli were randomly presented either 75 pixels to the left or 75 pixels to the right of the center of the screen. A n y one of the stimuli were randomly presented on either side of the center of the screen. Thus there were 2 possible locations for T, M l and M 2 , resulting in 8 combinations. The added factor of missing corner for T produced a total of 16 possible stimuli combinations. A s such, there were four stimulus sequences ( T - M 1 - M 2 , T - M 2 , T - M l , T). Thus T, M l , M 2 appeared in every trial of experiment 4. The total distance between the two stimulus centers was 150 pixels (288'). Figure 11 illustrates the position of the stimuli with respect to the 4 different stimulus sequences. Insert Figure 11 about here Results Metacontrast results. A total of four stimuli sequences were used: T - M 1 - M 2 , T - M l , T-M 2 , T. The graph of the results of these four sequences is in Figure 12. The data were analyzed in two separate groups, T - M 1 - M 2 and T - M l (metacontrast results), T - M 2 and T (inattentional masking results). Figure 13 illustrates the results for the T - M 1 - M 2 (unmasking) and T - M l (masking) sequences and table 5 indicates the obtained means. The graph of the T - M 1 - M 2 and T - M l sequences does not show the typical inverted U-shaped unmasking curve. A 2 (Stimuli sequence: T - M 1 - M 2 , T - M l ) x 10 (SOA) A N O V A was used to analyze the T - M 1 - M 2 20 (unmasking) and the T - M l (masking) sequences, with repeated measures on both factors. Only stimulus sequence was significant [F(l,9) = 9.65, p = .013]. The interaction was not significant. Thus the results indicate that there is a difference between the unmasking and masking curves. Insert Figure 12 about here Insert Figure 13 about here Insert Table 5 about here Inattentional masking results. Figure 14 graphically illustrates the results of the T - M 2 and T sequence. Note the strong masking effect of M 2 . This can be labeled as inattentional masking (Enns & D i Lol lo , 1996), which occurs when spatial attention is not fully focused on the target location. Typical metacontrast is dependent on the proximity o f the stimuli contours and on retinal eccentricity. For foveal presentations, very small contour separations dramatically decrease or even eliminate metacontrast masking, whereas for non-foveal presentations, metacontrast masking can be obtained with spatial separations as large as 2° (Breitmeyer, 1984). Enns and D i Lol lo (1996) discovered that masking could be obtained using a four-dot masking stimulus when visual attention was distributed across a wide range of possible target locations whereas masking with a four-dot mask could not be obtained when visual attention was concentrated at one target location. A 2 (Stimuli sequence: T - M 2 , T) x 10 (SOA) A N O V A was used to analyze the T - M 2 (inattentional masking) and the T (target only) sequences, with repeated measures on both factors. A significant interaction was obtained between the two 21 factors [F(l,9) = 8.48, p_ < .001]. A main effects analysis revealed only the T - M 2 (inattentional masking) sequence as significant [F(l,9) = 10.865, p_ < .0001]. This result confirms Enns and D i Lol lo ' s (1996) result of masking of unattended objects. Alerting effect. T-tests were performed on the non-positive S O A s of the T - M l (masking) sequence to test for any significant difference from chance. The 5 non-positive SOAs were significant at the .005 level [t(9) = 3.25, p_ < .005]. For this manipulation, M 2 occurred opposite the T - M l sequence. Thus the presence of M 2 before T produces a slight increase in performance. Insert Figure 14 about here Discussion. In this experiment, we hypothesized no unmasking because M 2 validity was at chance and its value as a cue would be ni l . In line with this expectation, the interaction was non-significant as was the S O A factor. However the stimulus sequence factor was significant indicating a difference between the unmasking and masking curves. This interesting result reflects the fact that there is a performance increase even when M 2 is not predictive of T location. If there is an increase in performance when M 2 is not an endogenous cue, then this increase in performance is not due to any endogenous effects. A l l that is left is an exogenous component. Thus by subtraction, the small increase in performance seen in experiment 4 for the unmasking curve is the exogenous component of M 2 . Jonides (1981) suggests that an exogenous cue can have exogenous and endogenous components. In one experiment (experiment 2) Jonides had observers ignore the presentation of an exogenous cue which occurred in the periphery of an 8 location circular visual search display. There was a positive R T difference (98 msec) between valid and invalid trials which suggested that the observers could not ignore the presentation of the exogenous cue, even when told to do so. Jonides considered this to be a reflexive component 22 in that an exogenous cue wi l l automatically pull one's attention toward its location regardless of an observer's intent. Observers also were told to attend to the exogenous cue, even though its validity was 12.5% (1 target location among 8 possible locations). Again another positive R T difference (95 msec) occurred. However, this positive R T difference was smaller than when the exogenous cue validity was at 70% (337 msec, experiment 1). Jonides suggested that an exogenous cue has endogenous and exogenous components, whereby the endogenous component is affected by the validity of the cue and the exogenous component reflexively pulls the observer's attention towards the location of the cue. Thus for the unpredictable location experiment, the validity of M 2 (50%) was not low enough to eliminate the endogenous component, and the mild unmasking obtained was due to smaller endogenous remnant. Alerting effect. T-tests performed on the non-positive SOAs of the masking sequence for experiment 4 and 5 revealed that all non-positive SOAs were significantly different from chance. I would argue that these results are due to an alerting effect. In experiment 4, for the unmasking sequence at the negative SOAs , M 2 occurred in every trial before T and M l and alerted the observer as to the future presentation of T. For the masking sequence, M 2 occurred opposite fixation from the masking sequence. In experiment 5, M 2 only occurred on the unmasking sequence. The masking sequence did not contain M 2 , i.e., the masking sequence did not have the M 2 stimulus present on the opposite side of fixation, as in experiment 4. However t-tests performed on the non-positive SOAs of the masking sequence for experiment 5 revealed them as significantly different from chance. The one explanation for this would be an expectation of an occurrence of M 2 . A n observer in an experiment 5 masking trial, which does not have M 2 , would still expect M 2 to occur. Thus the expectation of M 2 even when it does not occur produces an alerting effect. Experiment 5 - Predictable location The goal of this experiment was to produce unmasking with the typical inverted U -23 shaped curve between two locations. It was a control experiment to experiment 4. A s such, the design o f this experiment was the same as for experiment 4 except that the va l id i ty o f M 2 was 100%. Thus whenever M 2 appeared, it was a lways at the same loca t ion as the remain ing s t imul i . A g a i n , Jonides ' (1980) conc lus ion w i t h regards to cue va l id i ty was employed , i.e., the greater an exogenous cue 's va l id i ty , the greater its effectiveness. Thus i f M 2 is a purely exogenous cue, then unmask ing should occur w h e n M 2 va l id i ty is 100%. M e t h o d Observers Exper imen t 5 tested whether predictable spatial loca t ion affected metacontrast unmasking . A total o f 11 U B C students participated i n experiment 4. One student was dropped due to chance responding. A l l students received course credit for their par t ic ipat ion. Apparatus and s t imul i The apparatus and s t imul i were the same as i n experiment 4. Procedure The procedure was the same as per experiment 4 except for the f o l l o w i n g difference: The combina t ion o f s t imulus presentation was manipulated such that the s t imul i were presented only on the left or r ight o f f ixat ion. N o one st imulus was presented contralateral to the other s t imul i . Contrary to experiment 4, T , M l , M 2 d i d not appear o n every t r ia l o f experiment 5. There were four s t imulus sequences ( T - M 1 - M 2 , T - M 2 , T - M l , T ) presented i n two locat ions (left, right) result ing i n 8 possible combinat ions. The addi t ion o f mi s s ing target corner (left, right) produced 16 possible s t imul i combinat ions . Figure 15 indicates the pos i t ion o f the s t imul i w i t h respect to the four different s t imul i sequences. Insert F igure 15 about here 24 Results Metacontrast results. A s in experiment 4, four stimulus sequences were used: T - M 1 - M 2 , T - M l , T - M 2 , T. The graph of the results of these four sequences is in Figure 16. The data were again analyzed in two separate groups, T - M 1 - M 2 and T - M l (metacontrast results), T - M 2 and T (inattentional masking results). Figure 17 graphically illustrates the T - M 1 - M 2 and T 7 M 1 results and table 7 indicates the obtained means. Visual inspection of the graph indicates the typical inverted U-shaped curve found in unmasking. A 2 (Stimuli sequence: T - M 1 - M 2 , T - M l ) x 10 (SOA) A N O V A was used to analyze the T - M 1 - M 2 and the T - M l sequences with repeated measures on both factors. Only the interaction of the T - M 1 - M 2 and T - M l stimuli sequences were significant (F(9,81)=4.1, p=.0). Insert Figure 16 about here Insert Figure 17 about here Insert Table 7 about here Inattentional masking. Figure 18 illustrates the results of the T - M 2 and T sequences for experiment 5. Again please note the strong inattentional masking effect of M 2 on T. A 2 (Stimulus sequence: T - M 2 , T) x 10 (SOA) A N O V A was used to analyze the T - M 2 and T stimuli sequences with repeated measures on both factors. A significant interaction was found between the stimulus sequence and S O A factors [F(l,9) = 4.1, p_ < .001]. A main effects analysis revealed only the T - M 2 (inattentional masking) sequence as significant [F (1,9) = 10.46, p_ < .0001]. This 25 result again confirms the Enns and D i Lol lo (1996) result of masking of unattended objects. Table 8 indicates the obtained means. Insert Figure 18 about here Insert Table 8 about here Alerting effect. T-tests were performed on the non-positive S O A s of the T - M l (masking) sequence to test for any significant difference from chance. The 5 non-positive SOAs were significant at the .005 level [t(9) = 3.25, p_ < .005]. For this particular sequence, M2 was not presented with the T-Ml sequence. Discussion. In this experiment we hypothesized the typical inverted U-shaped curve of metacontrast unmasking because M 2 validity was at 100%. In line with expectation, there was a significant S O A x Stimulus sequence interaction which produce a significant main effect for the unmasking curve. Thus unmasking can occur even when the stimuli are presented at two locations as long as M 2 is fully predictive of T location. Thus for this experiment, M 2 is an exogenous cue directly indicating the location of T. Combined analysis of experiments 4 and 5 Rational and results The aim of the combined analysis was to: 1) Subtract any alerting effects from both experiments, 2) See i f exogenous and endogenous components of M 2 were significantly different from each other, 3) graphically indicate the exogenous and endogenous components of M 2 . Figure 19 is a composite graph of the combined unpredictable and predictable manipulations. Each curve is the difference between the unmasking and masking curves from each experiment. 26 Overall, this figure indicates the difference in performance between the two experiments. No significant interaction was found between the two curves across all S O A s and across the non-positive SOAs . Only the predictable curve was significant across all S O A s [F(9,162)=3.6, p_ = 0.0] and across the non-positive SOAs [F(4,72)=3.9, p = .06]. Insert Figure 19 about here Discussion Because of the alerting effect, we subtracted the masking curve data from the unmasking curve data, for both experiments. This allowed us to exclude any effect of alerting from the data. This also allowed us to isolate the endogenous and exogenous effects. Both these curves are illustrated in figure 19. Only the predictability curve was significant. According to our argument, this curve illustrates both the endogenous and exogenous effects of M 2 , whereas the unpredictable curve indicates the exogenous effects of M 2 . The unpredictable curve was not significant and this suggests that exogenous attention is not strong enough to produce a significant increase in visibility of T. Thus a combined exogenous and endogenous cueing effects is necessary to produce the increase in visibility found in metacontrast unmasking. General Discussion Experiments 1 and 2 were replications of Breitmeyer et al.'s masking and unmasking experiments. A percentage correct dependent variable was used rather than Breitmeyer et al.'s subjective ratings. In doing so, we discovered that the amount of unmasking was less than the amount of masking. Experiment 3 tested dual-channel interaction theory by varying the spatial separation between M 2 and M l . Breitmeyer et al. (experiment 3, 1981) suggest that sustained-transient inhibition has a spatial range of 17'. We varied the spatial separation between M 2 and M l such that there was a spatial range of 17.7', and we found statistically significant unmasking. 27 Breitmeyer et al.'s spatial separation experiment used subjective ratings. It also had too few subjects and only two SOAs , resulting in low power. Thus we rejected sustained-on-transient inhibition as the mechanism for metacontrast unmasking. Using the results from this experiment, we proceeded with the hypothesis that M 2 was a cue. Experiments 4 and 5 attempted to discern whether M 2 was an exogenous or endogenous cue. Experiment 4 varied the validity of M 2 such that it was only 50% predictive of T location, where the stimuli were presented at 2 locations. Here, unmasking occurred, but without the typical inverted U-shaped curve. Experiment 5 was the control for Experiment 4 and we discovered unmasking with the typical inverted U-shaped curve. From Experiments 4 and 5 we concluded the following: 1) The unmasking found in Experiment 4 was due to an endogenous component, 2) A s such, M 2 was an exogenous cue, with exogenous and endogenous effects. The combined analysis was a purely statistical analysis of experiments 4 and 5. We subtracted the masking curve from the unmasking curve for each experiment and plotted the resulting values (Figure 19). We hypothesized that the separate endogenous and exogenous components could be verified statistically in a between-within design, which i f true, would strengthen our argument. Unfortunately, only the curve from the predictable location experiment was statistically significant and this curve contained both exogenous and endogenous components. This suggests that the exogenous component, as illustrated in the unpredictable curve, is not strong enough on its own, to produce unmasking. Both exogenous and endogenous components are necessary. Our argument has focused on M 2 as a cue, and from our previous experiments, we conclude that M 2 is an exogenous cue, with both exogenous and endogenous components. Is it possible for an exogenous cue to have both endogenous and exogenous components? A s previously indicated in the section dealing with experiment 4, Jonides (1981) suggests that an exogenous cue produces exogenous and endogenous orienting, whereby the endogenous component is related to cue validity and the exogenous component is related to how the cue is 28 able to automatically attract attention. Is there any other evidence to fortify our claim of M 2 as an exogenous cue? One line of evidence comes from Mii l le r and Findlay (1988) who investigated exogenous and endogenous cueing for single and multiple displays with a percentage correct performance measure. I w i l l briefly describe their experiments and then indicate their conclusions. The observers' task was twofold, to discern 1) the location of the target, 2) whether the orientation of a target was the same as an initial comparison target. Single and multiple displays were used where the multiple displays contained four possible target locations. Experiment 1 investigated the properties of exogenous cues whereas experiment 2 investigated those of endogenous cues. In both experiments, the difference in performance between valid and invalid trials was analyzed. The exogenous cue was a 50 msec brightening of the outline of one of four peripheral boxes for the multiple display and whereas for the single display, only the single box was outlined. The endogenous cue was a 50 msec central arrow pointing to one of four peripheral boxes for the multiple display and for the single display, it pointed to a single box. For experiment 1 using the exogenous cues, Mii l ler and Findlay found an initial increase in performance within an S O A o f 50 to 150 msec, after which there was a decrease in performance. This effect was obtained for both single and multiple displays (Figure 20). For experiment 2 with endogenous cues, there was a slower increase in performance in comparison to exogenous cues, such that for valid trials, performance increased within an S O A of 100 to 300 msec, and then remained at this constant level. This was true for single and multiple displays (Figure 21). Insert Figure 20 about here 29 Insert Figure 21 about here One problem with experiments 1 and 2 was that the S O A s and observers were different, thus direct comparison between the effects of exogenous and endogenous was not possible. Mii l ler and Findlay rectified this in experiment 4 by having the same observers and SOAs for both exogenous and endogenous cues. The exogenous cue (50 msec brightening of box outline) and endogenous cue (50 msec central arrow) were the same as in experiments 1 and 2. The three variables were type o f cue (exogenous/endogenous), cue validity (valid/invalid), and S O A duration (100, 400, 700 msec). The experiment was the same as in experiments 1 and 2. The results indicate that exogenous cues produced optimum accuracy at the 100 msec S O A with a decline towards the 400 msec S O A followed by constant performance. Endogenous cues showed increased performance from the 100 to the 400 msec S O A followed by constant performance (Figure 22). Insert Figure 22 about here One of the results which stands out in Mii l ler and Findlay's paper is the increase in performance at short S O A s (e.g., 100 msec S O A , experiment 4) for exogenous cues. For endogenous cues, there was a gradual increase in performance with a peak at moderate SOAs (e.g., 400 msec S O A , experiment 4). Mii l ler and Findlay argue that the physical properties of an exogenous cue trigger an automatic orienting component, possibly t 30 This automatic facilitation of the cued location then passes over into the controlled facilitation component, which can be thought of as an active maintenance of attention. This in turn is seen in the gradual increase in performance with the endogenous cue. What is important for the controlled component is that its construction is initiated through the development of an expectancy for the likely target position on the basis of spatial information (e.g., location). Thus for this thesis, we argue that M 2 is an exogenous cue with automatic and controlled orienting effects. When M 2 is fully predictive of T location, an expectancy for T location is produced. This in turn allows the controlled component to develop fully, which leads to an increase in visibility. When M 2 is not predictive of T location, then attention cannot be automatically engaged towards T location. Thus only the controlled orienting component develops. The results obtained with endogenous cues indicate that the development of an expectancy is time-consuming and thus the initial automatic orienting component is necessary to allow for the development of the controlled process. Thus for our predictable experiment (experiment 5), maximal M 2 predictability produced maximal unmasking due to an automatic orienting towards T location, followed by a controlled orienting which was constructed during the time attention was being reflexively oriented towards T location. Or to use the terminology o f this thesis, maximal metacontrast unmasking occurs when M 2 is fully predictive of T location, and is composed of exogenous and endogenous orienting components. Integration of the two solitudes of attention and metacontrast Please recall that Breitmeyer et al's (1981) view of metacontrast unmasking was as an extension of metacontrast masking, or to put it another way, as an extension of dual-channel theory. Our model is quite different. We argue that the predictions made by dual-channel theory are incorrect with regard to metacontrast unmasking. Mii l le r and Findlay's (1988) two component model of spatial attention serves to complement dual-channel theory. A s previously 31 indicated, Mii l le r and Findlay (1988) suggest that transient channel activation triggers the automatic orienting component. However the strongest case for our argument of complementarity comes from Breitmeyer himself. Breitmeyer and Ganz (1976) suggested that the transient channels may provide a pathway for an effect of attention. The authors propose a role for attention within dual-channel theory whereby figural and location information enters via parallel channels; the sustained channels transmit figural information whereas transient channels transmit information about location or rapid changes in location. Breitmeyer and Ganz (1976) suggested that the retinotopic organization of the superior colliculus and pulvinar structures and the suggested location information transmitted by the transient channels direct selective attention to locations in space. Therefore, the transient channels are said to be an early warning system, transmitting information to the colliculus structure. This neural structure then initiates selective attentional orienting responses to spatial locations, which are then analyzed by information transmitted through the sustained channels. Breitmeyer and Ganz predict that maximum performance for pattern recognition should occur when attention is concentrated at a single spatial location. Other papers which suggest the transient pathway as a channel for activation of attention is Jonides (1981) and Yantis and Jonides (1984). Thus the two theories are important and do complement each other, leading to an integrated theory. Future research would deal with the role played by both low and high level processes and their effects on our perceptions. 32 References Breitmeyer, B . G . , (1978). Disinhibition in metacontrast masking of Vernier acuity targets: Sustained channels inhibit transient channels. Vis ion Research, 18, 1401-1405. Breitmeyer, B . G . , (1984). Visual masking: A n integrative approach. Cambridge: Oxford. Breitmeyer, B . G . , Ganz, L . , (1976). Implications of sustained and transient channels for theories of visual pattern masking, saccadic suppression, and information processing. Psychological Review, 83, 1-36. Breitmeyer, B . G . , Rudd, M . , Dunn, K . , (1981). Metacontrast investigations of sustained-transient channel inhibitory interactions. Journal of experimental psychology: Human perception and performance, 4, 770-779. Cleland, B . G . , Levick, W.R. , Sanderson, K . J . , (1973). Properties of sustained and transient cells in the cat retina. Journal of Physiology, 228, 649-680. Coren, S., Ward, L . M . , Enns, J.T., (1994). Sensation and perception. (4 t h ed.). New York: Harcourt Brace. Dow, B . M . , (1974). Functional classes of cells and their laminar distribution in monkey visual cortex. Journal of Neurophysiology, 37, 927-946. Enns, J.T, D i Lol lo , V . , (1996). Object Substitution: A new form of masking in unattended visual locations. Psychological Science, Hoffman, K . P . , Stone, J., Sherman, S . M . , (1972). Relay of receptive-field properties in dorsal lateral geniculate nucleus of the cat. Journal of Neurophysiology, 35, 518-531. Ikeda, H . , Wright, M . J . , (1975). Spatial and temporal properties of "sustained" and "transient" neurones in area 17 of the cat's visual cortex. Experimental Brain Research, 22, 363-383. Jonides, J., (1980). Toward a model of the mind's eye movement. Canadian Journal of Psychology, 34, 103-112. 33 Jonides, J., (1981). Voluntary versus automatic control over the mind's eye's movement. In J.B. Long & A . D . Baddeley (Eds.), Attention & Performance. V o l . 9. Hillsdale, N . J . : Erlbaum. Matin, E . , (1975). The two-transient (masking) paradigm. Psychological Review. 82(6), 451-461. Meyer, G .E . , Maguire, M . W . , (1977). Spatial frequency and the mediation of short-term visual storage. Science, 198, 524-525. Mii l ler , H . J., Findlay, J. M . , (1988). The effect of visual attention on peripheral discrimination thresholds in single and multiple element displays. Acta Psychologica. 69, 129-155. Nakayama, K . , Mackeben, M . , (1989). Sustained and transient components of focal visual attention. Vis ion Research, 29, 1631-1647. Posner, M. I . , Nissen, M . J . , Ogden, W . C . , (1978). Attended and unattended processings modes: The role of set for spatial location. In Pick, H . L . and Saltzman, I.J. (Eds.), Modes of Perceiving and Processing Information. Hillsdale, N . J . : Lawrence Erlbaum Associates Posner, M. I . , (1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32, 3-25. Singer, W. , Bedworth, N . , (1973). Inhibitory interaction between X and Y units in the cat lateral geniculate nucleus. Brain Research, 49, 291-307. Singer, W. , P Poppel, E . , Creutzfeldt, O.D., (1972). Inhibitory interaction in the cat's lateral geniculate nucleus. Experimental Brain Research, 14, 210-226. Weisstein, N . , Ozog, G. , Szoc, R., (1975). A comparison and elaboration of two models of metacontrast. Psychological Review, 82, 325-343. Winters, R .W. , Hamasaki, D.I., (1975). Peripheral inhibition in sustained and transient on-center ganglion cells in cat retina. Experientia, 31, 305-306. 34 Winters, R . W . , Hamasaki, D.I., (1976). Temporal characteristics of peripheral inhibition of sustained and transient ganglion cells in cat retina. Vis ion Research, 16, 37-45. Wright, R . D . , Ward, L . M . , (1994). Shifts of visual attention: A n historical and methodological overview. Canadian Journal of Experimental Psychology, 48, 151-166. Yantis, S., Jonides, J., (1984). Abrupt visual onsets and selective attention: evidence from visual search. Journal of Experimental Psychology: Human Peception and Performance, 10, 601-621. 35 Table 1. Properties of sustained and transient cells as described by Breitmeyer & Ganz (1976). Cel l type Transient Sustained •Short response latency 'Longer response latency & persistence & persistence •Higher response amplitude 'Lower response amplitude •Large receptive fields 'Smal l receptive fields •Heavily concentrated in parafovea 'Heavi ly concentrated in fovea and peripheral retinal regions Respond to: Respond to: •Movement/flicker 'St imuli presented for prolonged periods/low temporal rates •Stimuli onsets and offsets 'Stationary stimuli •Low to intermediate spatial 'Intermediate to high spatial frequencies frequencies 36 Table 2. Accuracy means and standard deviations from experiment 1 (masking replication) at different SOAs . S O A (msec) Mean & std. Dev. (%) -135 86.4 (9.4) -90 89.1 (11.8) -60 89 (8.5) -45 90.2 (8.2) 0 92.7 (7.8) +45 63 (12.5) +60 66 (15) +90 70.4 (13.2) +135 84.6 (11.9) 37 Table 3. Accuracy means and standard deviations from experiment 2 (unmasking replication) at different SOAs . S O A (msec) Mean & std. Dev. (%) -300 75.5 (11.5) -150 79.4 (10.1) -90 83.3 (7.6) -45 81.2 (7.1) 0 75.9 (10.4) +45 71.3 (9.6) +90 69.7 (6.9) +135 75 (10.9) +195 74.8 (9.9) +345 . 74.8 (7.2) 38 Table 4. Obtained means and standard deviations from experiment 3. Mean & std. Dev. (%) M 1 - M 2 Spatial separation T - M l Control 1.9' \9JL Masking S O A (msec) +45 60.7 (12.8) -720 61.7 (11) 59.1 (13.4) -360 63.8 (10.3) 63.8 (11.1) -180 64.9 (9.6) 64.8 (12.1) -90 71.1 (11.2) 68.2 (12.1) 0 58.1 (16.6) 58.3 (11.8) +90 62.2 (14.1) 59.9 (13.2) +180 61.7 (11.8) 64.3 (12.7) +360 59.6 (15.7) 63.6 (13.8) +720 64.6 (11.9) 63 (10) 39 Table 5. Accuracy means and standard deviations from experiment 4 (unpredictable) for unmasking and masking sequences respectively S O A (msec) T - M 1 - M 2 (unmasking) T - M l (masking) -300 61.7(14.8) 55 (12.6) -150 66.7(15.2) 53.3 (18.5) -90 64.2(17.1) 60 (17.5) -45 65.8(12.7) 55.8(11.8) 0 62.5 (11.3) 59.2(10.7) +45 65.8(11.4) 60 (15.1) +60 57.5 (9.2) 67.5 (16.4) +90 57.5(17.3) 56.7 (19.6) +150 60.8(10.4) 59.2 (12.7) +300 60 (18.3) 58.3 (14.7) 40 Table 6. Accuracy means and standard deviations from experiment 5 (predictable) for unmasking and masking sequences respectively S O A (msec) T - M 1 - M 2 (unmasking) T - M l (masking) -300 67.8 (19.2) 59.2(16.4) -150 78 (12.7) 53.3 (14.3) -90 69.7(10.9) 56.7(11.7) -45 65.8(15.9) 57.5 (17.8) 0 50.8(6.1) 59.2(18.2) +45 54.7(10 ) 60.8 (17.1) +60 52.2(11 ) 51.7(18.3) +90 55.6(14.3) 60 (16.6) +150 50.8 (10.7) 62.5(15.3) +300 62.2(11.4) 56.7 (20.3) 41 Figure Caption Figure 1. Graphical illustration of performance obtained in a metacontrast masking paradigm. Y -axis is percentage correct number of responses and x-axis is S O A . Target 'T ' is positioned at 0 S O A . M l is varied at SOAs from -300 msec to +300 msec. Note decrease in performance when M l is presented after T. Results are actual data from this thesis. 100 r 90 >> O 1 o o < 80 70 60 50 M l T I •135 -90 -60 0 +60 +90 +135 -45 +45 Stimulus Onset Asynchrony (ms) M • M l 4> T - M l at 0 S O A T - M l at 0 S O A 42 Figure Caption Figure 2. Graphical illustration of performance obtained in a metacontrast unmasking paradigm. Note decrease in performance when M l is presented after T. Control condition is T and M l presented at +45 msec S O A . Arrow at ordinate indicates performance for control condition. Results are actual data from this thesis. 100 r 90 a 80 % 70 60 50 •0 T M l _ i i i L 7?--150 -45 +45 +135 " + 3 4 5 -300 .90 0 +90 +195 Stimulus Onset Asynchrony (ms) M2 M2 T - M l - M2 at 0 SOA T - M l - M2 at 0 SOA 43 Figure Caption Figure 3. Results from Hoffman, Stone and Sherman demonstrating the convergence of inhibitory influence i n d L G N . Figure 3a illustrates the recovery of response of a transient cell. Y -axis is number of times the cell responded to the pair stimulus, x-axis is ISI.. • = 1 volt, T = 3.2 volt, g = 7.5 volt, • = 10 volt. Figure 3b illustrates the number and type of cells which discharged in the experiment. Shaded bars = transient cells, clear bars = sustained cells. Figure 3a. 44 Figure Caption Figure 4. Diagrammatic illustration of results from Nakayama and Mackeben's experiment 3. Performance as a function of cue lead time for three different stimulus durations of 117 msec, 83 msec and 33 msec. Black and white arrows indicate performance at 33 msec without cueing and with sustained cueing respectively. Diagram from Nakayama and Mackeben (1989). •4 iHmtwt | »cu»« ten* I wwtr feuw tfttfttrwg taction j cv§* I 3 durst tor* jm*Z3 1 0 0 M M • i y i n n i , • , j i 100 300 600 cu« ttttf t»w* {mwcl 45 Figure Caption Figure 5. Stimuli used in all experiments. Stimuli Visual angle @ 63.3 c Target right ^ J 59.7' Target left ^ J M l O 59.7' 109.7* M2 C M 186-6' M2 Large f 248-3' 46 Figure Caption Figure 6. Results from experiment 1. Metacontrast masking of T by M l . 50-l—J 1 1—i 1 L _ I • -135 -90 -60 0 +60 +90 +135 -45 +45 Stimulus Onset Asynchrony (ms) 47 Figure Caption Figure 7. Results from experiment 2. Metacontrast unmasking of T by M 2 . Horizontal line is the baseline S O A condition from experiment 1, which is the +45 msec S O A . This particular S O A produced maximum masking in experiment 1. 48 Figure Caption Figure 8. Combined results from experiments 1 and 2. Illustrates differences in a) baseline performance and b) overall effect. Masking -135 -100 -50 +50 +100 +135 -300 -150-100-50 0 +50+100 +200 +345 M l Time (SOASmsec) • T O Time (SOA \ msec) A 1 1 1 M2 T M 1 M2 49 Figure 9. Results from experiment 3 Figure Caption - spatial separation. Masking control is at +45 msec. Spatial Separation O 1.9' M2 - M l • 19.6' M2 large-Ml ^ 45msec Masking Control -j 100 -720 „ ^ -180 0 +180 -360 -90 +90 +360 Stimulus Onset Asynchrony (ms) +720 50 Figure Caption Figure 10. Metacontrast masking and unmasking magnitude as a function of M 1 - M 2 spatial separation. For masking, M l served as a target, M 2 as the mask. For unmasking, M 2 was presented continuously. From Breitmeyer et al. (1981). 51 Figure Caption Figure 11. Illustrates the position of the stimuli with respect to the 4 different stimulus sequences for experiment 4 - unpredictable location. Masking sequence Unmasking sequence + o + M l r Variable SOA + o + + IvL / + + Target only sequence Inattentional masking Bequence 4L M l msec Fixed SOA o + + Target r VarUbleSOA + M . / + + 52 Figure Caption Figure 12. Results from experiment 4 - unpredictable stimuli location. Each of the four stimulus sequences is represented: T - M 1 - M 2 (unmasking), T - M l (masking), T - M 2 (attentional masking), T. Control condition is the T - M l sequence. • Inattentional Masking • TOnly 0 Unmasking O Masking -300 -90 0 +60 +150 Stimulus Onset Asynchrony (ms) 53 Figure Caption Figure 13. Results from experiment 4 - unpredictable stimuli location - using only the T - M 1 - M 2 (unmasking), T - M l (masking) sequences. Control condition is the T - M l sequence. O O < loo r 90 80 70 60 50 • Unmasking O Masking r-150 -45 +45+90 +300 -300 -90 0 +60 +150 Stimulus Onset Asynchrony (ms) 54 Figure 14. Results from experiment 4 (inattentional masking), T sequences. Figure Caption - unpredictable stimuli location - using only the T - M 2 • Inattentional Masking • TOnly -300 -90 0 +60 +150 Stimulus Onset Asynchrony (ms) 55 Figure Caption Figure 15. Indicates the position of the stimuli with respect to the four different stimuli sequences for experiment 5 - predictable location. Masking sequence Unmasking sequence + o + • + M l msec Hxed SOA r Variable SOA + / + Taiyet only sequence + • + Tari V A msec Fixed SO A rget r Variable SOA + / Inattentional masking sequence o + + • + + 56 Figure Caption Figure 16. Results from experiment 5 - predictable stimuli location. Each of the four stimulus sequences is represented: T - M 1 - M 2 (unmasking), T - M l (masking), T - M 2 (attentional masking), T. Control condition is the T - M l sequence. 100 '—s 90 $ >> 80 o r \ < 70 60 50 • Inattentional Masking • TOnly ® Unmasking O Masking ™ n " 1 5 ° o n " 4 5 n + 4 5 + 9 0 " + 3 0 ° -300 -90 0 +60 +150 Stimulus Onset Asynchrony (ms) 57 Figure Caption Figure 17. Results from experiment 5 - predictable stimuli location - using only the T - M 1 - M 2 (unmasking), T - M l (masking) sequences. Control condition is the T - M l sequence. 100 90 U 80 h 70 60 U 50 • Unmasking O Masking -150 -45 +45+90 "+300 -300 -90 0 +60 +150 Stimulus Onset Asynchrony (ms) 58 Figure Caption Figure 18. Results from experiment 5 - predictable stimuli location. Only two stimulus sequences are represented: T - M 2 (inattentional masking), T. -300 -90 0 +60 +150 Stimulus Onset Asynchrony (ms) 59 Figure Caption Figure 19. Graphical illustration of exogenous and endogenous orienting effects of M 2 . Each curve is the difference between unmasking and masking curves for each location condition. - 3 0 0 - 9 0 0 + 6 0 + 1 5 0 Stimulus Onset Asynchrony (ms) 60 Figure Caption Figure 20. Results from Mii l le r and Findlay's experiment 1. Top graph contains SOAs from 50 to 200 msec, bottom graph contains S O A s from 100 to 500 msec. Open symbols are single displays, closed symbols are multiple displays., Va l id trials: triangles, baselines: circles, invalid trials: squares. 1 r—I 1 1 1 58 75 10B 125 1SB 2BB SOP (msec) tea i S B zee 3 0 0 « 0 0 see SOfl (msec) 61 Figure Caption Figure 21. Results from Mii l ler and Findlay's experiment 2. Open symbols are single displays, closed symbols are multiple displays. V a l i d trials: triangles, baselines: circles, invalid trials: squares. 62 Figure Caption Figure 22. Results from Mii l le r and Findlay's experiment 4. Open symbols are single displays, closed symbols are multiple displays. V a l i d trials: triangles, baselines: circles, invalid trials: squares. 63 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0089998/manifest

Comment

Related Items