Keara M. Ginell¹ (), David E. Osher¹; ¹The Ohio State University

Attention serves as a fundamental mechanism for selecting the most valuable sensory information. Attentional control encompasses three modes: goal-driven (top-down), stimulus-driven (bottom-up), and experience-driven (memory-guided). Memory-guided attention reflects how prior experiences shape attentional biases; individuals can attend to specific locations based on learned patterns rather than immediate physical salience or explicit goals. Previous research has used search tasks with repeating experimental parameters (e.g. contextual or probabilistic cueing) to probe memory-guided attention towards static cues. However, these approaches do not investigate attention operating dynamically over time. We developed a novel modification of the serial reaction time task to investigate memory-guided attention in space and time. In each trial, a gabor stimulus—tilted 2° left/right of vertical—appeared briefly in one of four screen locations. Participants reported the gabor tilt (left/right). The experiment alternated between random blocks, where targets appeared in random locations, and repeating blocks, where targets reliably appeared in a sequence of locations across trials. Participants were not informed of the repeating sequence beforehand. We found that participants used statistical learning to guide attention to specific locations at specific times, performing faster and more accurately during repeating blocks than random blocks. This was evident in group-level analyses and at the individual level. Altering the repeating sequence resulted in slower responses and more errors, indicating perseveration. In a post-task, participants reported the next stimulus location after viewing two sequential stimuli from the repeating sequence. We observed significant correlations between this post-task performance and main-task effect sizes; participants whose reaction times and accuracy benefitted most from memory-guided spatiotemporal attention performed best on the post-task. This work provides evidence for memory-guided spatiotemporal attention. Future experiments will use this paradigm to investigate neural underpinnings of memory-guided attention in space and time, and learned spatiotemporal distractor suppression.