Right Place, Right Time: Spatiotemporal Predictions Guide Attention in Dynamic Visual Search

Most visual search research relies on static displays, yet real-world search occurs in dynamic, unfolding environments where both the location and timing of relevant items are variable. This study investigates whether humans can learn and exploit spatiotemporal regularities—predictable patterns of when and where targets appear—to guide attention in a noisy, dynamic search task.

The researchers designed a dynamic search task where participants searched for targets (vertical lines) that faded in and out among distractors (tilted lines) over several seconds. In each trial, some targets appeared at predictable times and within specific spatial quadrants, while others were unpredictable. Across four experiments, the authors measured behavioral performance (accuracy and reaction time) and eye movements to determine if participants could learn these patterns and use them to proactively allocate attention.

Participants consistently demonstrated superior performance for spatiotemporally predictable targets compared to unpredictable ones. Eye-tracking data confirmed that this performance boost was driven by proactive attentional guidance, as participants fixated the target-relevant quadrant earlier when the target was predictable. The results suggest that these benefits are supported by both long-term memory (as the effects persisted even after trials with no regularities) and short-term, single-trial priming (as repeating a pattern once improved performance on otherwise unpredictable targets). This indicates that the brain integrates temporal and spatial information into dynamic priority maps to navigate complex visual scenes.

This research bridges the gap between traditional static visual search models and the reality of dynamic, time-dependent environments. By demonstrating that temporal regularities are a critical, learnable dimension for attentional guidance, the study provides a foundation for more ecologically valid models of human perception and attention in the real world.