Implicit relational learning in a multiple-object tracking task

We studied implicit relational learning by embedding contextual relational information into a multiple-object tracking task. In two experiments, participants were instructed to track two or four out of eight moving objects and report at the end of the trial whether a single cued object was among those they tracked (yes/no task). The stimulus display also contained two background strips of different width. In the informative condition, the location of the cued object predicted the correct choice: If the answer was "yes", then the cued object was always located next to the narrower strip; otherwise, it was always located next to the wider strip (or vice versa). In the random condition, the location of the object did not predict the correct choice. Participants in the informative condition consistently displayed lower tracking accuracy than in the random condition, possibly due to attentional demands introduced by implicit relational task. At the same time, participants in the informative condition demonstrated no awareness of the task structure; instead, their reports were consistent with the attempts to track moving objects. Our task can provide a suitable model for studying implicit relational learning in adult participants that is essential for establishing generality of factors affecting relational learning.