Top-down control of attention by stereoscopic depth

Abstract

Stereoscopic depth has a mixed record as a guiding attribute in visual attention. Visual search can be efficient if the target lies at a unique depth; whereas automatic segmentation of search arrays into different depth planes does not appear to be pre-attentive. These prior findings describe bottom-up, stimulus-driven depth guidance. Here, we ask about the top-down selection of depth information. To assess the ability to direct attention to specific depth planes, Experiment 1 used the centroid judgment paradigm which permits quantitative measures of selective processing of items of different depths or colors. Experiment 1 showed that a subset of observers could deploy specific attention filters for each of eight depth planes, suggesting that at least some observers can direct attention to a specific depth plane quite precisely. Experiment 2 used eight depth planes in a visual search experiment. Observers were encouraged to guide their attention to far or near depth planes with an informative but imperfect cue. The benefits of this probabilistic cue were small. However, this may not be a specific problem with guidance by stereoscopic depth. Equivalently poor results were obtained with color. To check and prove that depth guidance in search is possible, Experiment 3 presented items in only two depth planes. In this case, information about the target depth plane allows observers to search more efficiently, replicating earlier work. We conclude that top-down guidance by stereoscopic depth is possible but that it is hard to apply the full range of our stereoscopic ability in search.

Keywords: Centroid estimation; Depth; Stereoscopic disparity; Visual attention; Visual search.

Figure 1.

The centroid paradigm: exemplar procedure and analysis

Figure 2

Exemplar stimuli used in centroid estimation experiment (cross fusion), the red round dot is the centroid of three square dots that lie at the target depth

Figure 3

Cartoon of the experimental procedure: (a) Fixation frame, 1s. (b) Stimulus, 300 ms. (c) Response display shown until a response is made. (d) Feedback display that shows the stimulus, the target centroid, the observer’s response, and the error.

Figure 4.

Screening test of temporal stereopsis: Judge whether three red dots are coplanar. You can cross fuse to see stereo. This is an example of a non-coplanar case.

Figure 5

Examples of temporal screening for two observers. Data from two 1-Up/3-Down staircase measurements of the threshold for detecting if three dots lie at the same depth. The last six reversals are marked with yellow (failed) and purple (passed) asterisks and are averaged to obtain the threshold.

Figure 6

(a) Data-drivenness and (b) Efficiency from attention filter estimation (methods taken from Sun et al. (2016)). Error bars indicate ±1 SEM.

Figure 7

Estimated attention filters for depth. Lines with different colors represent filter weights of different target depths, as shown in the legend. (a-e) filter weights estimation for observers 1–5, and (f) the average. Error bars indicate 95% confidence intervals.

Figure 8

Errors in centroid estimation for different target depths. Error bars indicate ±1 SEM.

Figure 9

Regression analysis between filter weights and estimation errors

Figure 10

(a) Configuration in multi-depth experiment (b) Multi-color as a control

Figure 11

Probabilistic cues. (Multi-depth: 1 for near, 8 for far. Multi-color: 1 for red, 8 for yellow.)

Figure 12

Simulation on effects of probabilistic cues. Simulated slopes of functions are given next to condition names.

Figure 13

Results of Experiment 2: RT × set size functions for multi-depth and multi-color conditions. Slopes of functions are given parenthetical next to condition names. Error bars indicate ±1 SEM.

Figure 14

Target Present Slopes of Experiment 2: Cued vs. NoCue of multi-depth experiment (One data point represents one observer of each cue condition (Front, Back, Middle). The figure legend indicates tokens for different observers. The diagonal line represents the 1:1 ratio of cued vs. Nocue conditions while the dashed line represents the 5:2 ratio).

Figure 15

Results of Experiment 3: RT × set size functions of Two-depth vs. One-depth condition. Slopes of functions are given parenthetical next to condition names. Error bars indicate ±1 SEM.

Figure 16

Target Present Slopes of Experiment 3: Two-depth vs. One-depth (One data point represents one observer. The diagonal line represents the 1:1 slopes ratio of OneDepth vs. TwoDepth conditions, while the dashed line represents the 2:1 ratio).