國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
4. Experiment 3: Can Binocular Disparity Affect Spatial Attention?
Experiments 3 and 4 are designed to test the hypothesis in different ways.
Experiment 3 tests whether the effect of binocular disparity can override the effect of spatial attention. The hypothesis is that attention can bias multistable motion perception by making the attended areas look closer in depth. Accordingly, if there is a clear binocular depth cue (by manipulating binocular disparity) to define the depth relationship between occluders and moving lines, then the effect of spatial attention on biasing multistable perception should be eliminated or overridden. In particular, the larger the binocular disparity is manipulated, the more the effect of attention will be overridden. In other words, the effect of spatial attention on biasing multistable motion perception should decrease or disappear with increasing binocular disparity.
The manipulation-check task of spatial attention used in Experiment 1b and Experiment 2 will also verify whether participants allocate their attention according to the instructions.
‧
Another 20 participants were recruited using the same standards described for Experiment 1a. The number of participants increased in this experiment because Experiment 2 showed that much data (nearly half) would be deleted and not included in further analysis under the criterion of positive index of attention. Increasing participants can increase statistical power against data deleting.
Design
The experiment is a 2 (attention) × 4 (binocular disparity) completely within- participant design. Participants’ attention was manipulated by instructing them to attend to the four occluders or attend to the four moving lines in different blocks, as in Experiments 1 and 2. The order of the two attention conditions was counterbalanced within participants. Binocular disparity was manipulated at four different levels (0, 3, 6, and 9 pixels, showed in Figure 14) with a stereoscope. The larger disparity means that the moving lines were seen as more behind the occluders. Zero disparity means that the occluders and the moving lines were at the same depth. In addition, participants’ intention was controlled to be the same (by instructing them to hold the
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
perception of coherent motion) throughout the experiment, as in Experiment 1b.
Participants had to report their motion perception (coherent or separate) by key-pressing during 3000 ms to 6500 ms trials. The average percentage of time perceiving coherent motion was measured as a dependent variable. In addition, at the end of each trial, participants had to respond to a probe—the lighting of an occluder or a moving line—by pressing a key as fast as possible.
(a) (b) (c) (d)
Figure 14. The diagram of the binocular disparities manipulated in Experiment 3: (a)
binocular disparity 0, (b) binocular disparity 3, (c) binocular disparity 6, and (d) binocular disparity 9.
Materials
The diamond stimulus used here was the same as in Experiment 2.
‧
The apparatus used here was identical to that in Experiment 2, including a View Sonic 19-inch CRT computer monitor, a chinrest, an observation box, a stereoscope, and two numerical keyboards.
Procedures
At the beginning of the experiment, the experimenter instructed participants to adjust the stereoscope and a depth-judgment task was executed in order to determine whether the participants can see three-dimensional pictures correctly, as in experiment 2. Only the participants who pass this task can continue.
The nature of multistable figures was then explained to participants. The experiment can proceed only after confirming that each of the two motion interpretations of the diamond stimulus can be perceived, as in Experiment 1a.
A practice block was then executed. Participants were trained to allocate their attention on the four occluders or the four lines and respond to their motion perception repeatedly until they were well practiced and familiar with the entire procedure.
At the beginning of each block, the experimenter told participants to allocate their attention on the four occluders or the four lines throughout the block.
Participants pressed the left-hand “0” key to start each trial when they were well
‧
prepared (fixate on the central cross and allocate their attention on demanded areas).
Participants had to report their motion perception by key-pressing, as in Experiment 1b.
The manipulation-check task of spatial attention was also added in this experiment, as in Experiment 1b. A probe was shown at 3000 ms, 3500 ms, 4000 ms, 4500 ms, 5000 ms, 5500ms , 6000ms, or 6500ms after the trial began, and participants had to additionally respond by pressing the numeral “4” key with their left hands as fast as possible. The eight probe-showing times were the same in all blocks, but the sequence was different (by randomizing). After responding to the probe, the trial ended and a new trial began. Since the recording of motion perception was terminated while the probe was presenting, the lighting of the occluder or line would not influence the multistable motion-perception process.
The experiment contained eight blocks: four attending-to-occluders blocks and four attending-to-moving-lines blocks. The sequence of the eight blocks was counterbalanced between participants and split over two days. Each block had eight trials in which the four levels of binocular disparity were presented twice, respectively, in a random sequence. In total, each participant completed 8 × 8 = 64 experimental trials in two days.
‧
Mean percentage of time perceiving coherent motion under the 2 (attention) × 4 (binocular disparity) conditions in eight blocks is plotted in Figure 15a. Two-way ANOVA shows that the main effect of attention is significant (F(1,19) = 4.50, p <
0.05, partial η2 = 0.191). The percentage of time perceiving coherent motion is higher in the attending-to-occluders condition (70.76%) than in the attending-to-moving-lines condition (63.73%), which is consistent with the hypothesis.
The main effect of binocular disparity is also significant (F(3,57) = 51.42, p < 0.001, partial η2 = 0.730), implicating that the manipulation of binocular disparity is valid.
However, the interaction of attention and binocular disparity did not reached a significant level (F(3,57) = 0.72, p = 0.541). In addition, the two-way ANOVA of RT to the probe in eight blocks also shows that the interaction is not significant, as predicted (F(1,19) = 1.71, p = 0.207), implicating that participants did not fully follow the instructions to allocate their attention. Next, only the data of participants who followed instructions and allocated their attention were selected for further analysis.
The method of selection was the same as in the Experiment 2: the eight blocks were divided into four sections (Section 1: Blocks 1 and 2; Section 2: Blocks 3 and 4, and so on). Each section contained an attending-to-occluders block and an
‧
attending-to-moving-lines block to check whether the index of attention (as described in Experiment 2) in each section is positive. Only the sections in which the index of attention is positive—implying that participants allocated their attention according to the instructions—were selected for further analysis.
There were 40 selected sections (total: 4 section × 20 participants = 80 sections), which contained 24 attending-to-occluders-first and attending-to-moving-lines-next sections and 16 attending-to-moving-lines-first and attending-to-occluders-next sections.
The mean percentage of time perceiving coherent motion under the 2 (attention)
× 4 (binocular disparity) conditions in the selected sections is plotted in Figure 15b.
Two-way ANOVA shows that only the main effect of binocular disparity is significant (F(3,57) = 41.62, p < 0.001, partial η2 = 0.687) and consistent with the
prediction. However, the main effect of attention (F(1,19) = 3.16, p = 0.092, partial η2 = 0.142) and the effect of interaction (F(3,57) = 1.59, p = 0.201, partial η2 =
0.077) are not significant. The simple main effect shows that the effect of attention is significant only at binocular disparity 0 (F(1,19) = 4.76, p < 0.05) and not significant at binocular disparity 3 (F(1,19) = 0.1, p = 0.760), 6 (F(1,19) = 2.21, p = 0.154), or 9 (F(1,19) = 1.73, p = 0.204). This evidence supports the prediction that the effect of attention decreases at non-zero binocular disparity due to the effect of binocular
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
disparity weakening the effect of attention.
(a)
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Figure 15. Results of Experiment 3. (a) Means and 1 standard errors of the percentage
of time that the diamond stimulus was perceived as coherent motion under the 2 (attention) × 4 (binocular disparity) conditions in eight blocks. (b) Means and 1 standard errors of the percentage of time that the diamond stimulus was perceived as coherent motion under the 2 (attention) × 4 (binocular disparity) conditions in the selected sections.
‧
The resulting pattern of the selected sections shows that the effect of attention decreases with increasing binocular disparity (the difference of the percentage of time perceiving coherent motion under the two attention conditions decreased from 16.37% and 2.14% to 7.62% and 5.79% with increasing binocular disparity). The simple main effect shows that the effect of attention is significant only at binocular disparity 0 and not significant at the other three large binocular disparities (power = 0.865, which is estimated by Cohen’s (1988) medium effect size f = 0.25). This result is consistent with the hypothesis and implies that binocular disparity weakens the effect of attention.
The main problem of the results is that the effect of attention seems to be mostly blocked at binocular disparity 3 but still appears slightly at binocular disparities 6 and 9. A possible reason is that other factors interfered with the entire mechanism at binocular disparities 6 and 9. For example, in the attending-to-moving-lines condition, the occlusion effect (the tendency to see the lines as behind the occluders completed into a diamond and moving coherently) of the occluders may decrease with increasing binocular disparity. This is because it is easier to ignore the occluders and attend to lines alone when binocular disparity is large, so
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
the occlusion effect decreases. Accordingly, in the attending-to-moving-lines condition, the percentage of time perceiving coherent motion perception at binocular disparity 6 or 9 is not as great as that in the attending-to-occluders condition as predicted. To solve this problem, the manipulation of binocular disparity will be changed into the manipulation of monocular depth cues in Experiment 4.
‧
國立 政 治 大 學