Demos of several Multiple Object Tracking experiments

In these demos it helps if you keep your eyes fixed on the red squares or circles in the middle (the “fixation mark”).  In the first demo you will see 8 small squares.  Four of them will flash on and off.  Then they will all move randomly for 10 seconds.  Try to keep track of which ones had flashed (the “targets”).  (At the end of the trial you will see the correct answer indicated by circles around the targets).  Surprisingly most people have no trouble doing this although they are not sure how they do it.

Basic MOT Demo

In the first demo, the small square objects did not collide with each other because they were kept apart by a “force field”.  This makes it a bit easier to track.  But people can also track quite well if the targets are allowed to pass over or behind each other.  You can keep them separate as long as you can tell which one was in front.

Demo with self-occlusions

This ability to track multiple objects is quite robust.  You can even track them when they briefly (and completely) disappear from view behind an opaque strip.

Demo with occluding strips

This is true even if the edges of the occluding strips are invisible

Virtual occluding strips

Keeping track of the 4 targets requires that you keep track of their history – you need to trace them back to the start when they were visibly different.  So if you have to keep track of them as individual objects, you should be able to keep track of any special labels that were assigned to them at the beginning of the trial.  For example you should be able to remember if they had distinct numbers on them or if they started from each of the 4 corners.  Try to keep track of the targets in this demo and also remember the number that was on each.

Keeping track of target IDs

Or remember which corner each of them was in.

Keeping track of targets’ original location

You may find (as we found) that keeping track of the ID of each target is much harder than keeping track of them as targets, even though the latter presupposes the former.

For more details on the ID effect check out the report at:  ID Paper in Visual Cognition

Attentional Highbeams Demos at Yale Vision Science