Selective attention by contextual spatiotemporal regularities

Undergraduate Just-In-Time Abstract

Poster Presentation: Tuesday, May 20, 2025, 2:45 – 6:45 pm, Banyan Breezeway
Session: Undergraduate Just-In-Time 2

Kaiki Chiu1 (), William Narhi-Martinez1, Irene Echeverria-Altuna1, Pranava Dhar1, Anna (Kia) Nobre1; 1Yale University

Our current behaviour is often guided by long-term memories triggered by present contexts. Extensive research has established how learned spatial regularities proactively guide attention, yet the impact of spatiotemporal regularities in long-term memory remains less understood. Building on the work of Cravo et al. (2017), who demonstrated through behavioural and EEG results that contextual temporal associations can guide attention, we investigated the dynamics of learning spatiotemporal regularities and their enduring influence on attention. To probe how spatiotemporal regularities are acquired and maintained over time, we designed a paradigm in which participants completed a discrimination task within repeated contexts (across eight real-world scenes) during a one-hour session. During each trial, participants differentiated between a target object and multiple foil objects (50/50 probability), which appeared at varying locations within each scene and at distinct time points (early or late) relative to scene onset. Four scenes contained contextual spatiotemporal regularities (predictable; targets appeared at consistent locations and times within each scene), while the other four contained no spatiotemporal regularities (unpredictable). The assignment of target/foil objects and predictable/unpredictable scenes was counterbalanced across participants, and all scenes were presented in random order within each of the five completed blocks. Behavioural results revealed that over time, participants became progressively faster at identifying targets in predictable compared to unpredictable scenes, showing evidence that context-specific spatiotemporal associations were gradually learned and used to speed performance. An explicit post-task memory test further indicated that these performance enhancements did not depend on participants’ awareness of the spatiotemporal regularities within each context, as no significant interactions emerged between explicit knowledge of spatiotemporal regularities and behavioural benefits. These findings provide evidence that contextually cued spatiotemporal expectations can be learned to guide attention and flexibly deployed to optimise behaviour.