Abstract
Spatial and featural attention have typically been studied separately. However, we know from scene research that search for particular items can be focused on likely locations, like a mug on a counter. These object-locations associations are learned through a lifetime of experience, but in our research we wanted to determine if these object-locations associations could be learned and lead to attentional effects within a single experimental session through statistical learning. In 4 experiments, we assessed the effect of object-location associations on visual attention. In the experiments, a visual cue indicated the target object in each trial. Critically, we created specific positional regularities of objects. In Experiment 1-3, each target object appeared at one location on 80% of trials (high probable location, HPL) and at another location on 20% of trials (low probable location, LPL). In Experiment 4, each target could appear at all 4 locations with different probabilities (40%; 25%; 25%; 10%). Importantly, target location was counterbalanced so overall target probability was equal at all locations. Experiment 1 showed targets at HPLs led to faster RTs than targets at LPLs. Experiment 2 showed that the effects in Experiment 1 were not due to recency, through an equal probability testing period. With a probe technique, Experiment 3 further showed that positional regularity influenced early attention selection independent of later object recognition. With a more complex object-location manipulations, Experiment 4 revealed that instead of only prioritizing the highest probability location, the visual system can pick up more complicated probability manipulations between locations and prioritize accordingly. With an awareness test, Experiment 4 also showed the effect of positional regularities occurs regardless of whether participants have explicit knowledge. These results demonstrate object-location associations can be formed, and guide attention, on the timescale of typical statistical learning.