Abstract
Our previous work (Frank et al., 2014, Human Brain Mapping) showed that search for visual feature conjunctions (e.g., conjunctions of color and location) becomes efficient after a few days of training. However, it is unclear what role attention plays in the learning of complex feature conjunctions. Does learning occur only if participants pay attention to the to-be-learned target/distractor stimuli? And if so, does the amount of learning correlate with the amount of attention deployed? To answer these questions, we used an attentionally demanding RSVP task at the screen center with pre-adjusted levels of difficulties, while search stimuli (red-green bisected disks or T/L) were passively presented in the periphery. In Experiment 1 participants performed eight days of training on the RSVP task and two dual-task sessions (RSVP and active search on the conjunction stimuli) before and after training. Results showed no significant improvement on the search task, suggesting that attention is necessary for learning of feature conjunctions. In Experiment 2, participants were trained on the same dual-task paradigm as in the test sessions of Experiment 1 for ten days. Two pairs of search stimuli were used (red-green bisected disks and T/L). Each target/distractor pair was presented during either a low- or high- attentional load central RSVP task. Results indicate that participants learned both pairs of conjunction stimuli and the magnitude of learning did not differ between low-load and high-load central task conditions. This suggests that learning effects do not increase with more attention deployed to the conjunction stimuli, and that a low level of attentional allocation is sufficient for learning to take place. In sum, our results suggest that attention is necessary for the learning of visual feature conjunctions, while the amount of attention allocated beyond some minimal level does not impact the efficiency of learning.
Meeting abstract presented at VSS 2016