空間注意力經由深度影響模稜運動知覺 - 政大學術集成

全文

(1)The Effect of Spatial Attention on Multistable Motion Perception via the Depth Mechanism. By Hua-Chun Sun (孫華君). 政治大. B.S., National Chengchi University, 2007. THESIS. 學. ‧ 國. 立. ‧. sit. io. al. in. er. Nat. National Chengchi University. y. Submitted to the Department of Psychology of. n. v i n Cofhthe RequirementsUfor the Degree of Partial Fulfillment engchi Master of Science. Thesis Advisor: Shwu-Lih Huang, Ph.D. (黃淑麗博士) July, 2010.

(2) 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 2. i n U. v.

(3) Abstract. Many studies have found that fixating or directing spatial attention to different regions can bias the perception of the Necker cube, but whether this effect of spatial attention is due to attended areas perceived as being closer have yet to be examined. This issue was directly investigated in this study. The stimulus used was the diamond stimulus,. 政治大. containing four occluders and four moving lines that can be perceived as coherent or. 立. separate motions. The results of Experiment 1 show that coherent motion was. ‧ 國. 學. perceived more often under the attending-to-occluders condition than under the. ‧. attending-to-moving-lines condition, indicating that spatial attention can bias. Nat. io. sit. y. multistable perception. The results of Experiment 2 show that the mean probability of. er. reporting lines behind occluders in small binocular disparities was significantly higher. al. n. v i n under the attending-to-occludersCcondition h e n gthanc under h i Uthe attending-to-lines condition, indicating that spatial attention can make attended areas look slightly closer. The results of Experiments 3 and 4 show that the effect of spatial attention on biasing multistable perception was weakened when there were binocular or monocular depth cues to define the depth relationship between the occluders and the lines. These results are all consistent with the notion that spatial attention can bias multistable perception through affecting depth perception, making attended areas look closer.. 3.


(5) Table of Contents. 1. Introduction………………………………………………………………………..9 1.1 Multistable Figures: Definitions, Properties, Examples, and Categories……...9 1.2 Theoretical Accounts of Multistable Figure Perception...................................13 1.3 Effect of Fixation and Spatial Attention……………………………………...22. 政治大. 1.4 Aspects of Spatial Attention that may Affect Depth Perception......................29. 立. 1.5 Experimental Stimulus……………………………………………………….33. ‧ 國. 學. 1.6 Purpose, Questions, and Hypothesis of the Current Study……………….......36. ‧. 2. Experiment 1: Can Spatial Attention Bias Multistable Motion Perception?……..41. Nat. io. sit. y. 2.1 Experiment 1a: The Relationship between Spatial Attention and Intention:. er. Dependent or Independent Mechanisms?.........................................................43. al. n. v i n Ch 2.1.1 Methods……....………….……...……...………................................43 engchi U 2.1.2 Results….…………………………………………………….……...46 2.1.3 Discussion…………………………………………………………...48. 2.2 Experiment 1b: Adding a Manipulation-Check Task to Spatial Attention….49 2.2.1 Methods……....………….……...……...………................................50 2.2.2 Results….…………………………………………………….……...53 2.2.3 Discussion…………………………………………………………...57. 5.

(6) 3. Experiment 2: The Effect of Attention: Can Spatial Attention Alter Perceived Depth?....................................................................................................................59 3.1 Methods……………………………………...……………...........................61 3.2 Results….……………………………………………………...…….……...66 3.3 Discussion……………………………………………...…………………...71 4. Experiment 3: Can Binocular Disparity Affect Spatial Attention?.................…...73. 政治大. 4.1 Methods……....………….……...……...………...........................................74. 立. 4.2 Results….…………………………………………………….…………......78. ‧ 國. 學. 4.3 Discussion…………………………………………………………………..82. ‧. 5. Experiment 4: Can Monocular Depth Cues Block the Effect of Spatial. Nat. io. sit. y. Attention?...............................................................................................................85. er. 5.1 Methods……………………………………………...……...........................87. al. n. v i n Ch 5.2 Results…………………………………………………................................90 engchi U 5.3 Discussion…………………………………………………..........................95. 6. General Discussion………………………………………………………….........97 7. References……………………………………………………………………....109 8. Acknowledgment……………………………………………………………….117. 6.

(7) List of Figures. Figure 1. This figure shows the following examples of multistable figures…………10 Figure 2. The experimental paradigm of Kornmeier and Bach (2005)………………17 Figure 3. The two corners of the Necker cube and their correspondent perception when attending or fixating on them………………………………………..23. 政治大. Figure 4. One of the partially biased Necker cubes used by Peterson Gibson (1991).27. 立. Figure 5. Procedures of the two experiments in Tsal and Kolbet’s (1985) study……29. ‧ 國. 學. Figure 6. The diamond stimulus……………………………………………………...33. ‧. Figure 7. Percentage of coherent motion perception in one-minute trials under. Nat. io. sit. y. different kinds of binocular disparity………………………………..…...35. er. Figure 8. The diagram of the hypothesis……………………………………..………37. al. n. v i n Figure 9. Results of ExperimentC 1a…………………………………………………..47 hengchi U. Figure 10. Results of Experiment 1b…………………..……………………………..54 Figure 11. The mean percentage of time perceiving coherent motion under the 2 (group) × 2 (attention) conditions……………………….………...……56 Figure 12. The flow chart of the depth-judgment trials in Experiment 2……..…..….65 Figure 13. Results of Experiment 2…………………………………………………..70 Figure 14. The diagram of the binocular disparities manipulated in Experiment 3.....75. 7.

(8) Figure 15. Results of Experiment 3………………..…………………………………81 Figure 16. Experimental stimuli used in Experiment 4…………………..…………..88 Figure 17. Results of Experiment 4…………………………………………..……....93 Figure 18. Theoretical framework of the study……………………………………..102. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 8. i n U. v.

(9) 1. Introduction. 1.1 Multistable Figures: Definitions, Properties, Examples, and Categories. A multistable figure is an ambiguous visual stimulus that can form at least two. 政治大. markedly different perceptual interpretations. When looking at a multistable figure,. 立. observers can perceive just one interpretation at one time. Some people have called. ‧ 國. 學. this characteristic of multistable figures the "property of exclusivity” (Leopold &. ‧. Logothetis, 1999). In other words, observers’ perceptual systems seem unable to. Nat. io. sit. y. fixate on a single stable interpretation of a multistable figure. Instead, their perception. er. fluctuates, alternating between different interpretations during a period of continuous. al. n. v i n CSuzuki viewing (Toppino & Long, 2005; 2000). h e n&gPeterson, chi U. The Necker cube is one of the famous examples of multistable figures, shown in Figure 1a, which can be seen as the front face down to the left or the front face up to the right. Other well-known examples include Rubin’s faces/vase figure, which can be interpreted as a vase or two faces, and Boring’s young girl/old woman figure, which can be interpreted as a young girl or an old woman. They are shown in Figures lb and lc, respectively.. 9.

(10) (a). (b). 學. (d). 政治大. ‧ 國. 立. (c). (e). (f). Figure 1. This figure shows the following examples of multistable figures: (a) Necker. ‧. cube; (b) Rubin’s face/vases figure; (c) Boring’s young girl/old woman figure; (d). sit. y. Nat. io. n. al. er. Schroeder staircase; (e) duck/rabbit figure; and (f) equilateral triangles (Toppino & Long, 2005; Long & Toppino, 2004).. Ch. engchi. i n U. v. Garcia-Perez (1992) classified multistable figures into five types of percept-changing properties. The first type is the figure-ground reversible figure, such as the Rubin’s vase/face figure and the diamond stimulus, which will be introduced later. Reversals of this type of figure are related to figure-ground organization changes. The second type is the perspective-reversible figure, such as the Necker cube 10.

(11) and the Schroeder staircase (Figure 1d). Reversals of this type of figure are related to fluctuations in perspective (i.e., change in perceived depth). The third type is the meaning-ambiguous figure, such as the young girl/old woman figure and the duck/rabbit figure (Figure 1e). The perceptual instability of this type of figure is associated with changes in meaning (Long & Toppino, 2004). The fourth type is the orientation-reversible figure, such as the equilateral. 政治大. triangles in Figure 1f. The multistability of this type of figure involves the assignment. 立. of reference frames for the description of the shapes.. ‧ 國. 學. The last type is the stereokinesis moving pattern, such as a two-dimensional. ‧. ellipse rotating in the frontal plane, which can be perceived as a disc oscillating in. Nat. io. sit. y. three-dimensional space or an elongated egg slanted in three-dimensional space and. er. describing a circular trajectory in the frontal plane. It is the analysis of motion that. n. al. i n C causes the interpretation change (Garcia-Perez, h e n g c1992). hi U. v. To simplify the classification of multistable figures, it is feasible to classify them in terms of whether the perception change is related to the alternation of depth perception. Based on this definition, all figure-ground reversible figures, perspective-reversible figures, and stereokinesis moving patterns belong to the depth-reversible figure category. Conversely, meaning-ambiguous figures and orientation-reversible figures do not belong to this category because their reversals. 11.

(12) have nothing to do with depth reversal. This study is aimed at the investigation of depth-reversible figures. Since the multistable figure itself does not change at all, the perceptual interpretation of a multistable figure changing with time indicates that perceptual processing must be affected by some internal factors, which may be related to perceptual processing mechanisms (Pitts, Gavin, & Nerger, 2008). Therefore, many. 政治大. researchers use multistable figures to determine what factors affect our perceptual. 立. processing and how to affect the processing. In other words, the operation of. ‧ 國. 學. perceptual mechanisms might be understood better by studying the nature of. ‧. multistable figure perception (Toppino, 2003).. Nat. io. sit. y. On the other hand, many researchers believe that ambiguity is the hallmark of. er. retinal stimulus in almost all visual perception, so the visual system must solve the. al. n. v i n C h effectively (Long ambiguity to adapt to the environment e n g c h i U & Toppino, 2004; Peterson & Gibson, 1991). The ambiguity comes from the inverse problem of the visual system. That is, the perception of three-dimensional environments is underdetermined by the two-dimensional image on the retina. In general, our visual system usually uses a heuristic process, producing the most likely interpretation to solve the inverse problem (Palmer, 1999). However, when viewing a multistable figure, the visual system produces more than one interpretation. Instead of fixating on a single, stable. 12.

(13) interpretation, the observers’ perception alternates between different interpretations of a multistable figure during a period of continuous viewing, called instability. Some psychologists think that the perception of multistable figures is just like the perception of other stimuli in a normal environment; they are all ambiguous (Slotnick & Yantis, 2005; Peterson & Gibson, 1991), and they all involve the observers’ perception of form, except that the ambiguity of multistable figures is available to consciousness. 政治大. while the ambiguity of other stimuli is not (Peterson & Gibson, 1991). In other words,. 立. the problem of ambiguity is solved in the initial period of the perception of normal. ‧ 國. 學. stimuli, so the ambiguity does not enter the consciousness. However, in the perception. ‧. of multistable figures, the ambiguity comes into the consciousness dramatically due to. Nat. io. sit. y. the instability of multistable figure perception. Nevertheless, the underlying processes. er. are believed to be similar to the perception of normal stimuli (Long & Toppino, 2004;. n. al. i n C Peterson & Gibson, 1991; Leopoldh & Logothetis, 1999). engchi U. v. 1.2 Theoretical Accounts of Multistable Figure Perception. Different studies have found different factors that can influence multistable figure perception, such as adaptation, fixation, and intention. Consequently, a variety. 13.

(14) of theories have been proposed to account for multistable figure perception. Some psychologists have classified these theories into two types of approaches according to the nature of the assumed underlying mechanisms—bottom-up theory and top-down theory (e.g., Kornmeier & Bach, 2005; Toppino & Long, 2005; Kornmeier, Hein, & Bach, 2009). The former emphasizes the influence of bottom-up processes and the latter emphasizes the influence of top-down processes that influence multistable. 政治大. figure perception. These two types of approaches have been supported by respective. 立. evidence.. ‧. ‧ 國. 學. Bottom-up theory. Nat. io. sit. y. In bottom-up theory, the proposed explanation of multistability is in terms of. er. relatively passive, automatic, sensory, stimulus-driven, and bottom-up mechanisms.. al. n. v i n Creflect Figure reversals are assumed to and recovery of neural channels h e nthegfatigue chi U that underlie the alternate perceptions of a multistable figure (Toppino & Long, 2005; Kornmeier & Bach, 2005; Toppino & Long, 1987; Kornmeier, Hein, & Bach, 2009). When viewing a multistable figure for a long time, one set of channels underlying one percept of a multistable figure is adapted or fatigued. At the same time, the second set of channels underlying the other percept becomes more and more dominant, causing a reversal. Afterwards, as the second set of channels becomes fatigued with extended. 14.

(15) viewing and the first set of channels recovers from fatigue, the percept will reverse again accordingly. There is a great deal of evidence supporting bottom-up theory. A reliable finding of multistable figure literature that has been replicated many times is as follows: with prolonged viewing of a multistable figure, reversals become more frequent. Bottom-up theorists attribute this phenomenon to each channel fatiguing. 政治大. more quickly with increased viewing time. Therefore, reversals take place more. 立. quickly, and the channels do not fully recover before the next reversal begins. ‧ 國. 學. (Toppino & Long, 2005).. ‧. Toppino and Long’s (1987) study supported bottom-up theory more strongly.. Nat. io. sit. y. The multistable figure they used was a rotating Necker cube. When viewing this. er. stimulus, the direction of rotation can be perceived as to the left or to the right. They. al. n. v i n C h adaptation periodUand a two-minute test period. divided one trial into a two-minute engchi They found that if the rotating Necker cube’s location or size presented in the. adaptation period was the same as in the test period, the reversal rate was high during the test period, reflecting neural-channel fatigue. On the contrary, the reversal rate was low (as a no-adaptation control condition) when the rotating cube’s location or size presented in the two periods was different. This is because the different location or size of stimulus in the two periods is related to the processes of different neural. 15.

(16) channels. Thus, neural-channel fatigue did not occur. A more striking piece of evidence was as follows: when the stimulus presented in the test period changed into two side-by-side rotating cubes, the reversing rate of the two cubes was different. The cube at the same location in the adaptation period reversed more rapidly than the cube at a different location in the adaptation period. These localization characteristics of fatigue are consistent with bottom-up theory and incompatible with top-down theory.. 政治大. Another study supporting bottom-up theory is Kornmeier and Bach’s (2005). 立. (2005). They recorded event-related potentials (ERPs) when presenting Necker cubes. ‧ 國. 學. intermittently with a temporal regime (800ms on, 400ms off, shown in Figure 2).. ‧. Participants have to report whether their perception reversed or not by comparing any. Nat. io. sit. y. given stimulus with the preceding one. According to their report, the ERP data can be. er. divided into two conditions for comparison—perceptual reversal and non-reversal.. al. n. v i n C h ERP difference They found that the earliest significant e n g c h i U associated with perceptual reversal peaks around 120 ms after stimulus onset, and it is most prominent at the occipital cortex. This result shows that perceptual reversals can be initiated during very early visual processing stages, implicating bottom-up processing.. 16.

(17) Figure 2. The experimental paradigm of Kornmeier and Bach (2005).. 治政 However, there still are some findings that 大 cannot be explained well by 立 ‧ 國. 學. bottom-up theory. For example, Long et al. (1983) separated the viewing sessions of multistable figures across four weeks and found that the reversal rate increased across. ‧. the four weeks. This result is difficult to explain using the bottom-up fatigue and. sit. y. Nat. io. n. al. er. recovery mechanism because neural-channel fatigue seems unlikely to last for weeks.. i n U. v. Instead, this can be explained by some higher-order processes such as learning. Ch. engchi. (Toppino & Long, 2005). Many researchers have used this viewpoint to investigate the mechanism underlying multistable figure perception, and they are called top-down theorists.. Top-down theory According to top-down theory, multistability is explained in terms of higher-level cognitive processes that exert a top-down influence on the interpretation 17.

(18) of ambiguous retinal stimulation. Top-down theory conceptualizes perception as requiring problem-solving or decisional processes that are restrained by limited attentional resources and influenced by experience (Long & Toppino, 2004). They attribute reversals to fluctuations of attention (e.g., Kawabata & Mori, 1992), alternating decision processes (e.g., Rock, 1975), or successive testing of perceptual hypotheses (e.g., Gregory, 1974). In summary, top-down theory attributes the reversal. 政治大. of multistable figure perception and the property of exclusivity to higher-order. 立. cognitive processes, such as learning, expectation, experience, the viewer's intentions,. ‧ 國. 學. cognitive state, and rival demands on attentional resources (Kornmeier & Bach, 2005;. ‧. Long & Toppino, 2004).. Nat. io. sit. y. Contrary to bottom-up theorists, top-down theorists attribute the increasing rate. er. with increased viewing time to learning. That is, observers become more facile in. al. n. v i n C h of each representation accessing the internal representation of a multistable figure engchi U with practice (Toppino & Long, 2005). Furthermore, many researchers have found that observers can exert intentional control (or volitional control) during multistable figure perception (Suzuki & Peterson, 2000; Kornmeier, Hein, & Bach, 2009). For instance, observers can hold a particular percept of the figure or switch between each percept as rapidly as possible (Peterson & Gibson, 1991; Struber & Stadler, 1999). This kind of intentional control is assumed to be produced by top-down priming or. 18.

(19) activation of the desired representation (Toppino, 2003; Toppino & Long, 2005; Suzuki & Peterson, 2000). Shifting in the locus of spatial attention is also a top-down influence on multistable figure perception (Tsal & Kolbet, 1985). Slotnick and Yantis (2005) used event-related functional MRI to compare neural activities during voluntary shifts of spatial attention and voluntary shifts of the perception of the Necker cube. They found. 政治大. this two brain processing is associated with common activity in the posterior parietal. 立. cortex, implying that voluntary shifts of spatial attention are usually accompanied by. ‧ 國. 學. shifts of the perception. This result implies that spatial attention, a top-down influence,. ‧. can mediate perceptual multistability.. Nat. io. sit. y. Knowledge of multistability is also an important factor that can influence. er. multistable figure perception. For example, if the observer does not know that the. al. n. v i n C h and perceivesUit just in one way, few reversals figure has more than one interpretation engchi will be reported (Long & Toppino, 2004). This evidence supports top-down theory.. A hybrid model The best interpretation of the findings of the two approaches is that both bottom-up and top-down processes have influence on multistable figure perception. One of the most striking pieces of evidence of this contention is the effects of. 19.

(20) pre-exposure to an unambiguous figure before the multistable figure presentation (Toppino & Long, 2005). Long et al. (1992) manipulated the duration of pre-exposure to an unambiguous figure (such as a rotating cube) from zero to 150 seconds. Then a 30-second test period followed immediately, and observers had to report their perception of the ambiguous (multistable) figure (such as a rotating Necker cube) that corresponded to the pre-exposure stimulus. They found the same-bias effect (i.e.,. 政治大. initial perception of the ambiguous figure tended toward the same configuration as the. 立. prior unambiguous figure) in short pre-exposure duration trials (less than 5 seconds). ‧ 國. 學. and a reverse-bias effect (i.e., initial perception tended toward the opposite. ‧. configuration as the prior unambiguous figure) in long pre-exposure duration trials. Nat. io. sit. y. (more than two minutes). This is because the short pre-exposure duration of the. er. unambiguous figure induces a top-down priming effect toward the same. al. n. v i n C hambiguous figure isUperceived as a bias toward the representation. Thus, the followed engchi. same representation. On the contrary, a long pre-exposure duration of an unambiguous figure induces a bottom-up adaptation and fatigue effect. That is, the neural channels that were not adapted by the unambiguous figure presented previously became more dominant than the adapted channels, causing the followed ambiguous figure to be perceived the other way (Toppino & Long, 2005). Consequently, many researchers (e.g., Hochberg, 1968; Long et al., 1992;. 20.

(21) Palmer & Bucher, 1981; Toppino & Long, 1987) have adopted the opinion that multistable figure perception is influenced by both bottom-up and top-down processes. This opinion is called hybrid theory. In the hybrid theoretical framework, one of the most important issues is clarifying how the bottom-up and top-down processes interrelate, contributing to the multistable figure perception together (Toppino & Long, 2005; Pitts, Gavin, & Nerger, 2008; Kornmeier, Hein, & Bach, 2009).. 政治大. For example, Toppino (2003) used the Necker cube as a stimulus to investigate. 立. whether intentional control (top-down) over perception could be explained in terms of. ‧ 國. 學. intentionally selecting appropriate focal features within the stimulus for primary. ‧. processing, which is called “focal-feature processing” (bottom-up). He mainly independent. variables—observers’. intention. (to. hold. io. sit. two. y. Nat. manipulated. er. down-to-the-left or up-to-the-right orientation of the cube) and fixation (fixating at the. al. n. v i n Cthe bottom-left or top-right corner of viewing the Necker cube. He found h ecube) n gwhen chi U that these two factors have additive effects on perception. Even when the cube was too small to exhibit the fixation effect, intention still can bias perception of the cube. Another experiment by Meng and Tong (2004) also used the Necker cube as a stimulus and saw similar findings, with the manipulation of intention called “selective attention.” Hence, it is likely that intentional control is independent of focal-feature processing—perhaps through top-down activation or priming of perceptual. 21.

(22) representations—as previously described. The effect of focal-feature processing will be discussed below.. 1.3 Effect of Fixation and Spatial Attention. Effect of fixation. 立. 政治大. Many studies have found that fixating on different locations within a Necker. ‧ 國. 學. cube tends to favor one or the other perceptual response. There is a tendency to. ‧. perceive the vertex of the Necker cube that is nearest the visual fixation point as the. Nat. io. sit. y. frontal face (Inui, Tanaka, Okada, Nishizawa, Katayama, & Konishi, 2000; Kawabata,. er. Yamagami, Noakl, 1978). For example, fixating at its right-up corner tends to favor. al. n. v i n C hface down to the left, the perception of the cube as front e n g c h i U and fixating at the left-down corner tends to favor the front face up to the right perception (Kawabata, Yamagami, & Noakl, 1978; Long & Toppino, 2004), as showed in Figure 3. This is because “the fixated area tends to be seen as being relatively closer to the observer” (p. 1287) (Toppino, 2003). In Necker’s view, “the foveated portion of the figure was naturally supposed [by the observer] to be nearer and foremost” (p. 337) (Necker, 1832, 1964; Long & Toppino, 2004). Kawabata, Yamagami, and Noakl (1978) thought that the. 22.

(23) degree of clarity caused by the difference of visual resolution and attention is an important cue for the depth perception of the Necker cube. That is, areas near the visual fixation point are seen clearly, and the clear areas are seen as being in the front.. 立. 政治大. ‧. ‧ 國. 學. Figure 3. The two corners of the Necker cube and their correspondent perception. Nat. er. io. sit. y. when attending or fixating on them.. al. n. v i n Garcia-Perez (1992) alsoCassumed perceptions of multistable figures are h e n that gchi U biased depending on where observer fixates in a figure. The explanation is that because fixations near a focal area for one of the percepts will bring out fine-detail (high spatial-frequency) information relevant to that percept, it is easier to build that perceptual interpretation. This opinion can also be explained by the notion that fixation areas are seen as being closer because the clearer the image of an object is, the nearer this object is perceived to be.. 23.

(24) Toppino (2003) summarized these effects of fixation on multistable figure perception described above and proposed the focal-feature hypothesis. Its main assumption is that “different focal regions or areas within a multistable figure favor different global interpretations” (p. 1286). Furthermore, “which interpretation is perceived is assumed to depend on which focal area is selected for enhanced processing, regardless of whether that selectivity reflects the locus of fixation or the focus of attention alone” (p. 1286).. 立. 政治大. We should notice that the locus of fixation is not the same as the focus of. ‧ 國. 學. attention. Even though these two coincide with each other most of the time, attention. ‧. can be shifted independently of eye movements (Posner, 1980; Palmer, 1999; Tsal &. Nat. io. sit. y. Kolbet, 1985). For example, we can fixate on one place while attending to another. er. place; this is also called covert attentional shift (Suzuki & Peterson, 2000; Theeuwes,. al. n. v i n C h Therefore, the effect 1992; Leopold & Logothetis, 1999). e n g c h i U of fixation and the effect of attention should be differentiated. In addition, there is another issue in the focal-feature hypothesis that needs to be clarified: what is the “enhanced processing” for the selected focal area that biases the interpretation of multistable figures? Toppino (2003) did not clearly point out why “different focal areas within a multistable figure favor different global interpretations” (p. 1286). One of the most likely reasons is that focal areas tend to be seen as closer or. 24.

(25) nearer to the observer, as Toppino and Necker (1832, 1964) assumed.. Effect of spatial attention Kawabata (1986, 1987) directed observers’ attention to an angle at a vertex of a briefly presented (500 ms) Necker cube. He found that the attended angle was perceived as the front part of the cube and other parts were interpreted to match this. 政治大. interpretation. However, this study had two main disadvantages with respect to. 立. directing observers’ attention. First, in every trial, the vertex that needed attending. ‧ 國. 學. was always presented at the fixation point. This design can allow observers to attend. ‧. to the vertex easily and spontaneously, but the effect of the attention would be. Nat. io. sit. y. confounded with the effect of fixation. Second, the author used two heavy lines to. er. indicate the angle (one of the three angles of the attended vertex) to which observers. al. n. v i n Cithmay change the nature should give attention. However, e n g c h i U of the stimulus itself. Xu. and Franconeri’s (2010) study had similar findings and challenges. They found that people are more likely perceiving the cued side of the Necker cube as the closer side. The problem is that Xu and Franconeri used bright lines on the corners of the cube as a cue to direct people’s exogenous attention. However, showing the cue may affect perception. In summary, the effect of attention found in these studies may be confounded with the effect of fixation or the effect of the cue itself.. 25.

(26) Peterson and Gibson (1991) used the Necker cube to investigate whether people can direct spatial attention into different sub-regions of an object while ignoring other sub-regions. They manipulated observers’ fixation location, spatial attention location (on the biased region or unbiased region, as showed in Figure 4), and intention (to hold either the horizontal or the vertical line in front at the attended intersection) and recorded their perception (whether the horizontal or the vertical line look forward) in. 政治大. 30-second trials. They found that the effect of the bias region shows only when the. 立. bias region is attended regardless of fixation location. They propose that the. ‧ 國. 學. processing of the stimulus is facilitated in the attended location, and it is attenuated in. ‧. the unattended location. Thus, when attention is directed to the biased region, the. Nat. io. sit. y. processing of the depth information—occlusion and shading—is facilitated, and it. er. strongly affects the perception of the stimulus. On the other hand, when attention is. al. n. v i n directed to the unbiased region,C thehprocessing of theU e n g c h i depth information of the biased region is attenuated. At this time, intention can exert its effect through top-down activation of the desired representation. However, due to spatial attention locations mentioned here that are different from the attended corners mentioned previously (such as in Figure 3), this study provides no evidence about whether attended locations look closer.. 26.

(27) Figure 4. One of the partially biased Necker cubes used by Peterson Gibson (1991). 1 is the biased region on which occlusion and shading define the horizontal line as. 政治大. being in front of the vertical line. 2 is the unbiased region.. 立. ‧ 國. 學. Tsal and Kolbet (1985) used meaning-ambiguous figures, such as the. ‧. duck/rabbit figure, to investigate whether the perception of an ambiguous figure. Nat. io. sit. y. results from focusing attention on a focal area that contains features significant for. er. this percept. In their Experiment 1 (see Figure 5, left), observers were instructed to. al. n. v i n interpretationCthrough h e n gthecblock. h i UAfter the. maintain a given. meaning-ambiguous. figure was presented briefly, a letter may have been presented at one of the two focal areas. Observers had to respond if they saw a letter only when they maintained the given interpretation successfully. They found that the letter detection was faster when the letter appeared in the focal area of the perceived interpretation versus the focal area of the alternative one. In their Experiment 2 (see Figure 5, right), the observers’ attention was directed by a letter shortly before the presentation of the figure. They. 27.

(28) found that the perception of the figure was more frequently associated with the letter-presented focal area than with the alternative focal area. They suggested that maintaining different interpretations of the same ambiguous figure is mediated by focusing attention on different focal areas of the figure. Because different features of the figure support different overall interpretations (for example, the right part of the duck/rabbit figure in Figure 5 looks more like a beak than a pair of ears), attending to. 政治大. a feature may cause the correspondent interpretation to become dominant (Tsal &. 立. Kolbet, 1985). Nevertheless, this explanation may be appropriate only for. ‧ 國. 學. meaning-ambiguous figures. For depth-reversible figures such as the Necker cube, the. ‧. right-up corner looks just like the left-down corner; they have similar features. Nat. io. sit. y. (including a horizontal line and a vertical line), so it is difficult for them to support. er. different overall interpretations. There may be other reasons that cause spatial. al. n. v i n C hfigure perception. One attention to affect depth-reversible e n g c h i U of the most likely reasons is that spatial attention can affect depth perception, making the attended areas look closer, like the effect of fixation mentioned above. Thus, in the current study, experiments are designed to test this notion. Before discussing this, some relative studies need to be reviewed in order to evaluate the possibility that attention can affect depth perception, making the attended areas look closer.. 28.

(29) 立. 政治大. ‧ 國. 學. Figure 5. Procedures of the two experiments in Tsal and Kolbet’s (1985) study.. ‧. io. sit. y. Nat. n. al. er. 1.4 Aspects of Spatial Attention that may Affect Depth Perception. Ch. engchi. i n U. v. The evidence indicating that spatial attention may affect depth perception can be classified into two categories. The first category comprises the evidence that attention can modulate very early visual processes. For example, Tootell, Hadjikhani, Hall, Marrett, Vanduffel, Vaughan et al. (1998) used functional magnetic resonance imaging (fMRI) to label cortical activity due to visual spatial attention. They found that directing attention to a stimulus location results in increased activity not only in. 29.

(30) visual areas V4, V3, V3a, and V2 (robust attentional modulations) but also in V1 (smaller attentional modulations). Similarly, McAdams and Maunsell (1999) trained two monkeys attending to a stimulus inside or outside the receptive fields of recorded neurons in V4 and V1. They found that attention enhanced not only the responses of V4 neurons (median 26 % increase) but also the selectivity of V1 neurons (median 8 % increase).. 政治大. Event-related potential (ERPs) studies have found that modulations of spatial. 立. attention first occur in the extrastriate cortex at a latency of 80-100ms after stimulus. ‧ 國. 學. onset. In addition, a delayed attentional modulation of V1 begins around 130-160ms,. ‧. indicating a mechanism that delays feedback from higher areas in modulating neural. Nat. io. sit. y. activity in V1 (Martinez et al., 2001; Noesselt et al., 2002; Di Russo, Martinez, and. er. Hillyard, 2003). Lamme and Spekreijse (2000) proposed that the delayed feedback. al. n. v i n Ch activity in V1 may improve figure-ground e n g csegregation h i U and enhance the stimulus salience in attended areas (Hopfinger, Luck, & Hillyard, 2004).. Vecera, Flevaris, and Filapek (2003, 2004) directed participants’ attention by placing an unpredictable peripheral cue inside a figure-ground display so that exogenous attention was directed to one of the possible figural regions. Participants then had to indicate which of two probe stimuli matched one of the regions that appeared in the figure-ground display. They found that exogenous attention can. 30.

(31) influence figure-ground assignment: participants are faster to match the cued region from memory than they are the uncued region. On the other hand, in general, participants are faster to match the convex region than the concave region from memory. Convexity is a Gestalt figure-ground cue, so a convex region is more likely to be seen as figure than the concave region. However, this difference of reaction time (RT) between the convex and concave regions is reduced when attention is directed to. 政治大. the concave region. These results suggest that figure-ground processes are not entirely. 立. completed prior to the operation of spatial attention.. ‧ 國. 學. To sum up the evidence in the first category, attention can modulate very early. ‧. visual processes, including neural activity in V1 and figure-ground processes. Thus, it. Nat. sit er. io. perception.. y. is possible for attention to modulate follow-up visual processes, such as depth. al. n. v i n C h the evidence Uthat spatial attention can alter The second category comprises engchi. subjective perception of a stimulus appearance. One example is that both exogenous/involuntary/transient spatial attention and endogenous/voluntary/sustained spatial attention can increase perceived stimulus contrast, causing the apparent contrast of cued stimulus to be higher than that of uncued stimulus (Carrasco, 2006; Carrasco, Ling, & Read, 2004; Carrasco, Fuller, & Ling, 2008; Liu, Abrams, & Carrasco, 2009; Read & Carrasco, 2003; Treue, 2004; Ling, Carrasco, Lipson, Roche,. 31.

(32) Little, & Jones, 2007). Another example is that both exogenous and endogenous spatial attention can increase perceived spatial resolution, causing visual acuity to increase at cued locations in the Landolt gap resolution task (Carrasco, Williams, & Yeshurun, 2002; Montagna, Pestilli, & Carrasco, 2009; Yeshurun & Carrasco, 1999). Still another example is that exogenous spatial attention can increase perceived spatial frequency, causing spatial frequency in peripheral cued locations to be higher than in. 政治大. neutral cue conditions (Gobell & Carrasco, 2005). Exogenous spatial attention can. 立. also increase the perceived gap size of a Landolt square, the size of the moving. ‧ 國. 學. random dot patterns (Gobell & Carrasco, 2005; Anton-Erxleben, Henrich, & Treue,. ‧. 2007), as well as increasing color saturation (Fuller & Carrasco, 2006; Fuller, Ling, &. io. sit. y. Nat. Carrasco, 2004).. er. To sum up the evidence in the second category, attention can increase perceived. al. n. v i n C hspatial frequency,Uand size. It is worth noting that stimulus contrast, spatial resolution, engchi. these properties are relative to depth perception. For instance, if a stimulus is nearer to us, the stimulus will be more salient and acute (due to increased contrast, spatial resolution, and spatial frequency), and the size will be larger. Hence, it is possible for spatial attention to influence depth perception, making attended areas look closer. In order to test this hypothesis, the property of the selected stimulus must be able to reflect the effect of depth perception on it. Therefore, the diamond stimulus is chosen.. 32.

(33) 1.5 Experimental Stimulus. The multistable stimulus used in this study is the diamond stimulus, introduced by Lorenceau and Shiffrar (1992). The diamond stimulus contains a diamond outline moved in a circular trajectory with the four corners hidden by four occluders, as shown in Figure 6. When looking at this stimulus, observers can perceive the single. 政治大. coherent motion behind the occluders (Figure 6a) or the separate motions of the line. 立. segments (i.e., each of the four line segments moves sinusoidally in the direction. ‧ 國. 學. shown in Figure 6b) (McDermott & Adelson, 2001).. ‧. n. er. io. sit. y. Nat. al. Ch. engchi. (a). i n U. v. (b). Figure 6. The diamond stimulus, which can be perceived as (a) coherent motion or (b) separate motion.. The diamond stimulus is appropriate for this study for two main reasons. First, under Garcia-Perez’s classification framework of multistable figures, the diamond. 33.

(34) stimulus is a figure-ground reversible figure. That is, when the four moving lines are seen as the ground, the moving lines are seen as connecting to each other, moving coherently behind the occluders. However, when they are seen as a figure, they are not seen as being behind the occluders and connected to each other. Thus, separate motion will be perceived. Most of the previous studies that found the effect of spatial attention or fixation used the Necker cube as the stimulus, which belongs to. 政治大. perspective-reversible figure category. Therefore, it is unclear whether spatial. 立. attention can bias the perception of figure-ground reversible figures. However, if. ‧ 國. 學. spatial attention is exerted through the depth mechanism, the effect of attention found. ‧. in perspective-reversible figures should also present in figure-ground reversible. Nat. io. sit. y. figures because they all belong to the depth-reversible figure category.. er. The second reason for using the diamond stimulus is that the depth relationship. al. n. v i n between occluders and movingClines h e can h i U to investigate the depth n gbec manipulated effect of spatial attention. Kuo and Huang (2005) have found that the depth relationship between the occluders and moving lines can influence the perception of its motion. For example, if the occluders are in front of the moving lines, the percentage of coherent motion perception in one minute is higher than that under the conditions of occluders being perceived as behind or on the same plane as the moving lines. Conversely, if the moving lines are in front of the occluders, the percentage of. 34.

(35) coherent motion perception is lower than that under the conditions of occluders being perceived as behind or on the same plane as the moving lines. The results are shown in Figure 7.. 立. (a). 政治大 (b). (c). ‧ 國. 學. Figure 7. Percentage of coherent motion perception in one-minute trials under. ‧. different kinds of binocular disparity of the occluders and moving lines. Coherent. Nat. io. sit. y. motion perception was greater when the occluders were in front of the moving lines (a). er. and less when the moving lines were more in front of the occluders (c). When the. al. n. v i n occluders and the moving linesC were plane, Coherent motion perception h eonntheg same chi U fell between the other two conditions (b) (Kuo & Huang, 2005).. Since the motion perception of the diamond stimulus can be affected by the depth relationship between the occluders and the moving lines, it is suitable to be used to investigate whether spatial attention can bias multistable motion perception by making the attended areas look closer in depth. If the answer is yes, motion perception. 35.

(36) should be different under the two attention conditions. That is, when observers attend to the four occluders, coherent motion perception should be reported more than the attending-to-lines condition. This is because the attended occluders are seen as closer and the moving lines are seen as connected to each other behind the occluders. In contrast, when observers attended to the four moving lines, separate motion perception should be reported more than under the attending-to-occluders condition. 政治大. because the moving lines are seen as closer, not occluded, and moving separately.. 立. ‧. ‧ 國. 學. 1.6 Purpose, Questions, and Hypothesis of the Current Study. io. sit. y. Nat. er. Based on previous studies, directing fixation or spatial attention to different. al. n. v i n C hthe perception of it.UMany researchers believe this regions of a Necker cube can bias engchi is because fixated or attended regions look closer. However, no study has examined this assumption directly. In this study, the multistable stimulus used to examine this assumption is the diamond stimulus, which belongs to the depth-reversible figure category, which includes the Necker cube. In addition, the mechanisms of fixation and spatial attention are different and should be differentiated. In the research concerned with attention, few studies have mentioned whether. 36.

(37) spatial attention can affect depth perception. Many studies have found that spatial attention can influence many properties of stimulus perception, which may relate to depth perception. Thus, it is possible for spatial attention to affect multistable perception through the depth mechanism. The purpose of this study is to investigate the effect of spatial attention on depth-reversible figures perception and the possible underlying mechanism of this. 政治大. effect. The hypothesis is that spatial attention can bias the multistable motion. 立. perception of the diamond stimulus by making the attended areas look closer in depth,. ‧ 國. 學. as shown in Figure 8.. ‧. Nat. y. Multistable motion perception (depth-reversible figure category). io. n. al. sit. Depth perception. er. Attention. i n U. v. Figure 8. The diagram of the hypothesis. Attention affects multistable motion. Ch. engchi. perception by making the attended areas look closer in depth.. In order to test this hypothesis, four main questions need to be considered serially. The first is whether spatial attention can bias multistable motion perception. If this is supported, we can further check to determine whether the direction of bias coincides with the prediction that the attended areas look closer. These two questions will be verified in Experiment 1. 37.

(38) However, even if the results of Experiment 1 are consistent with the prediction of our hypothesis, this will not directly confirm that spatial attention is biasing multistable motion perception by affecting depth perception. Accordingly, in Experiment 2, this question will be verified directly. If the results of Experiment 2 support the notion that attention can affect depth perception to make attended areas look closer, it is more likely that the effect of spatial attention on multistable motion. 政治大. perception found in Experiment 1 occurs through the mechanism of making the. 立. attended areas look closer in depth.. ‧ 國. 學. Experiments 3 and 4 are designed to test the hypothesis from another. ‧. perspective. Binocular and monocular depth cues are manipulated in Experiments 3. Nat. io. sit. y. and 4, respectively. The question to be clarified is whether the effect of spatial. er. attention on biasing multistable motion perception will decrease or disappear when. al. n. v i n Ch depth cues of the stimulus are well-defined. reasoning is that if attention can bias e n gThe chi U multistable motion perception by making the attended areas look closer in depth, then, if there are clear, well-defined depth cues (regardless of binocular or monocular cues) to define the depth relationship between occluders and moving lines, the effect of attention should be eliminated or overridden. In other words, the effect of attention on multistable motion perception will decrease or disappear. In addition, we should note that intention is also a key factor that can influence. 38.

(39) multistable figure perception greatly, as mentioned previously. For example, observers can hold or switch to a particular percept of a multistable figure (Toppino & Long, 2005). Suzuki and Peterson (2000) suggested that although observers can view a multistable display passively without actively intending to see one or the other alternative perception, uncontrolled intentions could seriously confound experimental results. In particular, observers’ expectations or supposition of the experimental. 政治大. purpose may affect their intention. Therefore, it would be advantageous to control the. 立. observers’ intention, or the effect of attention may not be revealed. Therefore, the. ‧ 國. 學. variable of intention will be controlled in all experiments in the current study.. ‧. n. er. io. sit. y. Nat. al. Ch. engchi. 39. i n U. v.


(41) 2. Experiment 1: Can Spatial Attention Bias Multistable Motion Perception?. Experiment 1 is designed to examine whether spatial attention can bias multistable motion perception. Participants’ attention was manipulated by requesting they attend to the four occluders or the four moving lines while simultaneously remaining fixated on the center cross. Using this method, the effect of fixation can be. 治政 eliminated and the pure effect of spatial attention can be 大revealed. If attention can bias 立 ‧ 國. 學. multistable perception by making the attended areas look closer, then, when participants attend to the four occluders, the occluders will look closer, as if they are. ‧. in front of the moving lines. Moreover, previous research has found that when the. sit. y. Nat. io. n. al. er. occluders are in front of the moving lines, coherent motion perception increases (Kuo. i n U. v. & Huang, 2005). Therefore, coherent motion perception should increase under the. Ch. engchi. attending-to-occluders condition. Similarly, when participants attend to the four moving lines, the four moving lines will look closer, as if they are in front of the occluders. Previous research has found that when the moving lines are in front of the occluders, coherent motion perception decreases. Therefore, coherent motion perception should decrease under the attending-to-moving-lines condition. In Experiment 1a, participants’ intention was also manipulated as an independent variable for controlling purposes by requesting that participants attempt 41.

(42) to hold the perception of coherent or separate motion. The prediction is that participants can exert intention control to hold coherent motion or separate motion, so the percentage of time perceiving coherent motion should be higher in “hold coherent” condition than “hold separate” condition. On the other hand, investigating the mechanisms between attention and intention is also explores in Experiment1a. Toppino (2003) found that observers’. 政治大. intention (to hold the down-to-the-left or up-to-the-right orientation of the Necker. 立. cube) and fixation (fixating at the bottom-left or top-right corner of the cube) are two. ‧ 國. 學. independent mechanisms that have additive effects on perception. Hence, if the. ‧. mechanism of attention is similar to fixation, which works with intention. Nat. io. sit. y. independently, then there should be no interaction effect between attention and. er. intention. Moreover, in this study’s further experiments, the control of intention can. al. n. v i n Chold remain constant (for example, all h ecoherent h i U if the effect of attention under n g c motion) the two intention condition is similar.. 42.

(43) 2.1 Experiment 1a: The Relationship between Spatial Attention and Intention: Dependent or Independent Mechanisms?. 2.1.1 Methods. Participants. 政治大. Twelve participants who had normal or adjusted to normal vision were recruited. 立. in this experiment. Their age range was from 18 to 24 years. After finishing this. ‧ 國. 學. experiment, they received a 0.5 bonus point in the “Methods of Psychological. ‧. Experiment” course or 50 NT dollars.. er. io. sit. y. Nat. Design. al. n. v i n C h × 2 (attention) The experiment is a 2 (intention) e n g c h i U mixed factorial design. The. first factor is a between-participant factor and the second is a within-participant factor. Participants’ attention was manipulated by instructing them to attend to the four occluders in one block and the four moving lines in the other block. The order of the two attention conditions was counterbalanced across participants. At the same time, their intention was also manipulated through the experiment by instructing half of the participants to hold coherent motion perception and the other half to hold separate. 43.

(44) motion perception. Twelve participants were randomly assigned to each of the two intention conditions. Participants had to report their motion perception (coherent or separate) by key-pressing during two one-minute trials in each block. The average percentage of time perceiving coherent motion was measured as a dependent variable.. Materials. 立. 政治大. The visual stimulus used here was the diamond stimulus mentioned above,. ‧ 國. 學. containing four gray occluders, four white moving lines, and a fixation on the center.. ‧. It subtended 7.43º × 7.51º of the visual angle on the monitor programmed by using. Nat. n. al. er. io. sit. y. Macromedia Flash MX 2004.. Apparatus. Ch. engchi. i n U. v. A View Sonic 17-inch CRT monitor with a refresh rate 75 Hz and screen resolution of 1024 × 768 pixels was used to present the stimulus. Participants’ heads were restrained with a chinrest so their observation distance could be kept at 100 cm to the middle of screen. In order to isolate other light sources in the environment, participants viewed the stimuli through an observation box. Two numerical keyboards were used for key-pressing at each hand to start each trial and record responses.. 44.

(45) Procedures At the beginning of the experiment, the nature of the diamond stimulus was explained to the participants. The experiment could proceed only after it is confirmed that each of the two motion interpretations of the diamond stimulus can be perceived. Each block contained two one-minute-trials. Before the beginning of each block, there were two 30-second practice trials to let participants practice the procedures of their. 政治大. particular intention and attention conditions.. 立. In each trial, the diamond stimulus mentioned previously was presented for one. ‧ 國. 學. minute on a black background. There was a fixation cross (+) located on the center of. ‧. the stimulus for participants to keep their fixation on. Participants had to report their. Nat. io. sit. y. motion perception by key-pressing during one-minute trials. Pressing the numeral “1”. er. key indicated that they perceived coherent motion perception, and pressing the. al. n. v i n C h motion perception. numeral “3” key indicated separate e n g c h i U The average percentage of time perceiving coherent motion was measured as the dependent variable.. 45.

(46) 2.1.2 Results. Mean percentage of time perceiving coherent motion under the 2 (attention) × 2 (intention) conditions is plotted in Figure 9. Two-way ANOVA show that both the main effect of attention and intention are significant. The percentage of time perceiving coherent motion is higher in the attending-to-occluders condition (47.15%). 政治大. than in the attending-to-moving-lines condition (41.26%) (F(1,10) = 5.49, p < 0.05,. 立. partial η2 = 0.354), which is consistent with the hypothesis that attention can bias. ‧ 國. 學. multistable motion perception. Further, the percentage of time perceiving coherent. ‧. motion in the “hold coherent” condition (52.84%) is higher than that in the “hold. Nat. io. sit. y. separate” condition (35.57%) (F(1,10) = 11.31, p < 0.01, partial η2 = 0.531), which. er. is consistent with the intention control effect found previously. There is no significant. al. n. v i n Cintention interaction between attention and = 0.17, p = 0.689). The effects of h e n g(F(1,10) chi U attention and intention are additive, implicating that they are independent mechanisms on multistable motion perception. The effect of attention is similar to the effect of fixation found previously (e.g., Toppino, 2003) in two aspects. First, all have no interaction with the effect of intention. This implies that the mechanisms of attention and intention are independent. Second, the effect of intention on affecting multistable perception seems more powerful than the effects of fixation and attention (The mean. 46.

(47) difference of the percentage of time perceiving coherent motion under the two attention conditions is 5.89%; the mean difference between the two intention conditions is 17.27%).. 立. 政治大. ‧. ‧ 國. 學. Nat. io. sit. y. Figure 9. Results of Experiment 1a. Means and 1 standard errors of the percentage of. n. al. er. time that the diamond stimulus was perceived coherent motion are plotted under the. Ch. engchi. 2 (attention) × 2 (intention) conditions.. 47. i n U. v.

(48) 2.1.3 Discussion Results of this experiment show that spatial attention can bias multistable motion perception, and the type of bias is consistent with the prediction of the hypothesis. That is, the percentage of time perceiving coherent motion is higher in the attending-to-occluders condition than the attending-to-moving-lines condition. This result is consistent with that of attention biased multistable motion perception by. 政治大. making attended areas look closer, but it does not directly support this notion. This. 立. notion will be directly investigated in Experiment 2.. ‧ 國. 學. The effect of attention is comparatively smaller than the effect of intention. One. ‧. reason might be that the trial duration (1 minute) was too long to keep the. Nat. io. sit. y. participant’s attention on demanded areas, weakening the effect of attention. Thus, in. er. Experiment 1b, the trial duration will be reduced, and a manipulation check task of. al. n. v i n C h whether participants attention will be executed to examine e n g c h i U allocate their attention on demanded areas. In addition, there was no interaction between attention and intention (power = 0.725, which is estimated by Cohen’s (1988) medium effect size f = 0.25), implicating that attentional control and intentional control are independent mechanisms in influencing multistable motion perception. Therefore, the control of intention can be held constant—all coherent motion—in further experiments of this study because the effects of attention under the two intention conditions are similar.. 48.

(49) 2.2 Experiment 1b: Adding a Manipulation-Check Task to Spatial Attention. In Experiment 1a, participants’ attention was manipulated by instructing them to attend to the four occluders or attend to the four moving lines. However, it is still not clear whether participants obeyed these instructions. Hence, in Experiment 1b, a. 政治大. manipulation-check task of spatial attention was added. Participants had to respond to. 立. a probe (the lightening of an occluder or a moving line) presented at the end of each. ‧ 國. 學. trial as fast as possible. The reaction time to the probe was analyzed to verify whether. ‧. the manipulation of spatial attention was valid. The prediction is that if participants. Nat. io. sit. y. allocated their attention according to the instructions, RT to the lightening of the. n. al. er. occluder and line should be different. The RT should be shorter to for the lightening. Ch. engchi. of attended areas than for the unattended areas.. 49. i n U. v.

(50) 2.2.1 Methods. Participants Another nine participants were recruited with the same standards as described for Experiment 1a.. Design. 立. 政治大. This experiment is a one factor (attention) within-participant design.. ‧ 國. 學. Participants’ attention was manipulated by instructing them to attend to the four. ‧. occluders and attend to the four moving lines in different blocks. Simultaneously, all. Nat. io. sit. y. participants were instructed to hold coherent motion perception through the. er. experiment. In addition, at the end of each trial, participants had to respond to a. al. n. v i n C h or a moving line—by probe—the lightening of an occluder e n g c h i U pressing a key as fast as possible. The response time (RT) to the probe was measured as a dependent variable. In addition, participants had to report their motion perception (coherent or separate) by key-pressing until a probe presented. The average percentage of time in perceiving coherent motion was measured as another dependent variable.. 50.

(51) Materials The stimuli used in this experiment were similar to that in Experiment 1a, except that the four gray occluders were changed into hollow occluders to make the two probe conditions similar (that is, the light of an occluder outline or the light of a line).. Apparatus. 立. 政治大. The apparatus used here was identical to that in Experiment 1a.. ‧. ‧ 國. 學. Procedures. Nat. io. sit. y. The procedure was similar to Experiment 1a, except that participants had to. er. additionally respond to a probe that was shown 3, 4, 5, 6, 7, or 8 seconds after the trial. al. n. v i n began by pressing the numeralC“4” the left hand. The six probe-showing h ekeynwith gchi U. times were the same in all blocks, but the sequence was different (by randomizing) so participants could not predict when the probe would present itself. After responding to the probe, the trial ended and a new trial began. Since the recording of motion perception was terminated during probe presenting, the lightening of the occluder or line would not influence the multistable motion perception process. This experiment had four blocks—two attending-to-occluders blocks and two. 51.

(52) attending-to-moving lines blocks. The order of the four blocks was counterbalanced within participants. Each block had six trials corresponding to the six probe-showing times. In total, each participant has to do 4 × 6 = 24 experimental trials.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 52. i n U. v.

(53) 2.2.2 Results. The trials in which the RT to the probe was greater than 1500 ms were deleted and not included in further analysis. One trial was deleted for two participants, and another participant’s entire data were deleted because the deleted trials were too much (six trials of total 24 trials). Thus, the data for further analysis contained only eight participants’ non-deleted data.. 立. 政治大. The mean percentage of time perceiving coherent motion and RT to the. ‧ 國. 學. lightening of the occluder or line under the two attention conditions are plotted in. ‧. Figure 10. One-way ANOVA shows that the effect of attention is significant. The. Nat. io. sit. y. percentage of time perceiving coherent motion is higher in the attending-to-occluders. er. condition (67.28%) than in the attending-to-moving lines (49.17%) (F(1,7) = 24.87 , p. al. n. v i n C his consistent with U < 0.01, partial η = 0.780), which e n g c h i the hypothesis that attention can 2. bias multistable motion perception. The type of bias was also consistent with the prediction that the attended areas look closer. On the other hand, a two-way ANOVA of RT shows that the interaction of the RT to the probe under the two attention conditions and the two probe conditions is significant (F(1,7) = 6.35 , p < 0.05, partial η2 = 0.476), and the pattern of interaction in Figure 10 implies that the participants allocated their attention on demanded areas. That is, the value of subtracting the RT of. 53.

(54) lightening line from the RT of occluder lightening under the attending-to-lines condition is larger than that under the attending-to-occluders condition. The main effect of the two attention conditions (F(1,7) = 0.91 , p = 0.371) and the two probe conditions (F(1,7) = 1.38 , p = 0.278) are not significant.. 立. 政治大. ‧. ‧ 國. 學 er. io. sit. y. Nat. al. n. v i n C h1b. Mean percentage Figure 10. Results of Experiment e n g c h i U of time perceiving coherent motion and RT to the lightening of an occluder or a line under the two attention conditions.. In a further analysis of the data of RT to the probe, the eight participants can be divided into two groups depending on whether their individual patterns of interaction imply that the participants allocated their attention on demanded areas, which is. 54.

(55) defined by “the index of attention.” That is, subtract the RT of line lighting from the RT of occluder lighting under the attending-to-lines condition and under the attending-to-occluders condition, then subtract the later from the former; this value is the index of attention. If the index of attention is positive, the participant is assigned to the "positive interaction" group, implying he/she does allocate his/her attention on demanded areas. Conversely, participants are assigned to the "negative interaction". 政治大. group if the index of attention is negative.. 立. The mean percentage of time perceiving coherent motion under the 2 (group) ×. ‧ 國. 學. 2 (attention) conditions are plotted in Figure 11. There are seven participants in the. ‧. “positive interaction” group and only one participant in the “negative interaction”. Nat. io. sit. y. group. Two-way ANOVA shows that the effect of attention (F(1,6) = 6.26, p < 0.05,. er. partial η2 = 0.510) and interaction is significant (F(1,6) = 6.67, p < 0.05, partial η2 =. al. n. v i n C hthat the effect of attention 0.527). Simple main effect shows e n g c h i U is significant only in the “positive interaction” group (F(1,6) = 51.72, p < 0.001) but not significant at the. “negative interaction” group (F(1,6) = 0, p = 0.965). This evidence implies that the effect of attention will show, as predicted, only when participants allocate their attention on demanded areas (in the “positive interaction” group). Hence, the index of attention can be used in later experiments to filter out the data that does not meet the criterion of allocating attention on demanded areas.. 55.

(56) 政治大. 立. ‧ 國. 學. Figure 11. The mean percentage of time perceiving coherent motion under the 2. ‧. (group) × 2 (attention) conditions.. n. er. io. sit. y. Nat. al. Ch. engchi. 56. i n U. v.

(57) 2.2.3 Discussion. This experiment repetitively verified the main result of Experiment 1a that the percentage of time for perceiving coherent motion was higher in the attending-to-occluders than in the attending-to-moving-lines condition. This result is consistent with the prediction that spatial attention can bias multistable motion. 政治大. perception by making attended areas look closer. Furthermore, after reducing. 立. Experiment 1a’s trial duration from 1 minute to 3 to 8 seconds, the effect of attention. ‧ 國. 學. increased dramatically. The difference of the percentage of time perceiving coherent. ‧. motion under the two attention conditions increased from 5.89% to 18.11%. This. Nat. io. sit. y. implies that participants can keep their attention on demanded areas better during a. er. shorter duration trial than in a longer one. The manipulation check task of attention. al. n. v i n C h allocated their participants e n g c h i U attention. also showed that most. according to the. instructions. On the other hand, the increasing effect of attention brings another interesting issue worth investigating in the future: is the effect of intention still more powerful than the effect of attention after reducing the trial duration? Because intention was not manipulated in this experiment, this question cannot be answered here. However, even if the results of Experiments 1a and 1b were consistent with the. 57.

(58) hypothesis that attention can bias multistable motion perception, it is still not certain whether the effect of attention is actually affected by the depth perception mechanism, making attended areas look closer. Accordingly, in Experiment 2, this question will be verified directly.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 58. i n U. v.

(59) 3. Experiment 2: The Effect of Attention: Can Spatial Attention Alter Perceived Depth?. Experiment 2 is a depth-judgment task designed to investigate whether spatial attention can affect depth perception, making attended areas look closer. This experiment has nothing to do with multistable motion perception, but the diamond. 政治大. stimulus is still used here in order to determine whether spatial attention affects the. 立. stimulus used in Experiment 1b. In other words, the test will indicate whether. ‧ 國. 學. attending to the four occluders can make the occluders look closer and whether. ‧. attending to the four lines can make the lines look closer.. Nat. io. sit. y. In the depth-judgment task, nine different levels of binocular disparity of the. er. occluders and the lines are manipulated. Participants have to judge which one is. al. n. v i n C h or the four lines,Uafter the diamond stimulus is farther in depth, the four occluders engchi briefly shown. Participants’ attention is also manipulated in different blocks, as in Experiment 1b. The point of subjective equality (PSE, on which the participants have a 50% probability of reporting that the lines are more behind the occluders) will be measured as the dependent variable, and it is predicted that the PSE under the attending-to-occluders. condition. will. be. different. from. that. under. the. attending-to-lines condition. If attention can make the attended areas look closer, then. 59.

(60) when participants attend to the four occluders, the occluders will look closer. Thus, participants will perceive occluders and lines in the same plane only when the lines are defined by binocular disparity as in front of occluders, so the PSE should be smaller than zero. In the same way, the four lines will look closer when attending to the four lines. Thus, participants will perceive occluders and lines in the same plane only when the occluders are defined by binocular disparity as in front of lines, so the. 政治大. PSE should be larger than zero. Therefore, the PSE under the condition of attending to. 立. the four lines will be larger than under the condition of attending to the four occluders.. ‧ 國. 學. A manipulation-check task of spatial attention (as in Experiment 1b) will also be. ‧. executed to verify whether the manipulation of attention is valid.. n. er. io. sit. y. Nat. al. Ch. engchi. 60. i n U. v.