Multiple-object tracking task normally involves tracking a subset of moving objects (e.g., 4 out of 8) during a certain amount of time. We adopted this task to study implicit relational learning in adult humans.
In this task, participants are told to track 4 objects cued at the beginning of the trial. At the end of the trial, a single object is highlighted and the participants had to respond whether they tracked this object (a yes-no task). Unbeknown to the participants, the position of the object on the final display was predictive of the correct choice. For example, if the cued object was located to the narrow strip, then the correct answer was always "yes" and if it was located next to the wide strip, then the correct answer was always "no" (or vice versa). In the control condition, the location of the cued object was not predictive of the correct choice.
We found that none of the participants were aware that the location of the object relative to the strips could be used for making the choice. Yet, the participants in the experimental condition were significantly more accurate than the participants in the control condition (mean of 63.0% versus 49.3%). Moreover, the participants in the experimental condition maintained higher accuracy when presented with the strips of the novel width that they have not seen before (mean of 67.4% to the familiar trials versus 71.6% to the novel trials).
These results indicate a robust implicit relational learning in the task that requires sustained attention to the objects.We have began working on this project in Fall 2010, and many students have contributed to its development.
Comparative Perception and Attention Laboratory