Demos: Multiple-object tracking of distinct objects

The following demos are examples of several tracking conditions used in Makovski & Jiang's study on "Feature binding in attentive tracking of distinct objects", to appear in the special issue on binding edited by James Brockmole for the journal Visual Cognition. The demos used colored digits.

Task: Participants were asked to track 4 objects from a field of 8 objects. The target objects were cued by white outline frames. At the end of the motion period, all objects changed into white disks. Subjects had to click on the four target disks. After their response, the correctly clicked balls are highlighted in green. (In the demo here the target balls are highlighted in green after a pause). Subjects were not required to remember the identities of the objects and we never probed their memory of each object's color or digit identity.

The actual displays in the experiment subtended 21x21 degrees, several times bigger than what's shown in the movie. The motion speed shown in the movies is an approximation, not exact value, to the speed used in the experiment.

Condition 1: Homogenous. All objects were identical in color and digit identity. Click here to view Homogenous.mov(File size: about 800KB)

Condition 2: Color-distinct. All objects were identical in digit identity, but unique in color. Click here to view Color-unique.mov

Condition 3: Digit-distinct. All objects were identical in color, but unique in digit identity. Click here to view Digit-unique.mov

Condition 4: Both-distinct. All objects were unique in color and in digit identity. Click here to view Both-unique.mov

Condition 5: Conjunction-unique. All objects were unique in the combination of color and digit. However, they were not unique in color alone or digit alone. A given target, such as a red 3, shares with one distractor its color and another distractor its digit identity. Click here to view Conjunction-unique.mov

Condition 6: Conjunction-paired. All four targets were unique, and all four distractors were unique, but each target was the same as one of the distractors, leading to a pairing between a target and a distractor. Click here to view Both-paired.mov

Note:Two other conditions, color-paired and digit-paired, were not shown here. Experimental results showed that participants were more accurate in tracking the color-unique, digit-unique, or both-unique conditions than the homogenous condition. However, conjunction-unique and both-paired were tracked less accurately than the homogenous condition. Thus, the advantage provided by uniqueness in object identities is confined to that of a single feature. Feature-to-feature binding is highly limited (or absent) in attentive tracking.

Control experiments confirmed these findings using novel shapes or orientations (rather than digits). The novel shapes and orientations used can be found on this figure.

For more information, contact Tal Makovski at tal.makovski@gmail.com.