Abstract
One can more quickly find a visual search target when it appears in a previously-encountered context than when it appears in a novel context. Previous studies suggest that this Contextual Cueing effect represents a form of associative learning of the relationship between repeated contexts and the target’s location. However, it is unclear what kind of training schedule is optimal for acquisition and maintenance of Contextual Cueing. Specifically, researchers have yet to determine whether the number of repetitions and the relative spacing of repetitions influence the magnitude and persistence of the reaction time advantage for repeated contexts, relative to novel contexts. To explore this, we compared three training schedules in a within-subjects design. Participants searched for a letter T among Ls and indicated the T’s orientation using arrow keys. Thirty-six displays repeated throughout the 24 training blocks, with 12 displays appearing once in each training block (standard schedule; 24 repetitions total); 12 appearing once in each of the first seven blocks and once in the last training block (massed schedule; eight repetitions total); and 12 appearing in blocks 1, 2, 4, 6, 12, 16, 20, and 24 (spaced schedule; eight repetitions total). In a subsequent testing phase, all displays appeared once in each of the four testing blocks, along with 36 novel displays. The only training schedule that yielded a reliably larger Contextual Cueing effect than others was the standard schedule with the most repetitions. This pattern of results (Experiment 1, N = 32) persisted when participants were told in advance that displays would repeat (Experiment 2, N = 31). It also held when participants were tested 24 hours after training (Experiment 3, N = 32). Thus, the number of repetitions, but not the spacing of the repetitions, modulates the associative learning that occurs in Contextual Cueing.