Spatiotopic working memory measured with oculomotor responses
Poster Presentation 56.405: Tuesday, May 19, 2026, 2:45 – 6:45 pm, Pavilion
Session: Visual Working Memory: Spatial
Schedule of Events | Search Abstracts | Symposia | Talk Sessions | Poster Sessions
Sylvie R. Bass1, Amelia H. Harrison1, Thomas C. Sprague1; 1University of California, Santa Barbara, Santa Barbara, CA
Every saccade shifts visual input, requiring the brain to update spatial representations to keep perception stable. Previous studies have shown that spatiotopic memory – memory for object locations relative to the world – degrades with each saccade away from the encoding gaze position (Golomb & Kanwisher, 2012; Shafer-Skelton & Golomb, 2018). However, using a mouse click to report a remembered object location may incur performance costs due to the need to convert spatiotopic representations across oculomotor movements into motor plans. To investigate whether the previously-seen degradation in recall precision reflects fundamental limitations in spatiotopic updating rather than modality conversion costs, we conducted an experiment in which participants reported a remembered spatiotopic location with an eye movement, rather than mouse click or touchscreen reach, after zero through three saccades, with the final gaze position matching or differing from the initial gaze position at encoding (where retinotopic and spatiotopic locations match). Our results replicate previous studies showing that spatiotopic memory error increases systematically with more saccades. However, there are observed benefits to spatial memory accuracy by returning to the initial encoding location with the final saccade endpoint. This suggests that the degradation arises from inherent limitations in spatial updating mechanisms, not from response modality constraints. Thus, the cumulative error may be attributed to the cost of converting spatial locations from our native retinotopic coordinate system into spatiotopic coordinates. When the two coordinate systems realign by returning gaze position to the encoding location, the preserved retinotopic memory trace may bypass remapping costs. Future work will seek to explore how both domain-general navigation processes and spatiotopic and retinotopic coordinate systems act in tandem to impact visual working memory performance when navigating visual space.