Schedule of Events | Search Abstracts | Invited Symposia | Symposia | Poster Sessions | Data Blitz

Event Segmentation Patterns of Real-World Navigation from a First-Person Perspective

Poster Session B - Sunday, March 8, 2026, 8:00 – 10:00 am PDT, Fairview/Kitsilano Ballroom

Sydney Josifek¹ (), Luis Garcia², Uros Topalovic³, Mauricio Vallejo Martelo³, Matthias Stangl³, Tyler Davis¹, Martina Hollearn¹, Kiersten Olson¹, Amanda Holt¹, Justin Campbell¹, Lensky Augustin¹, Aydin Tasevac¹, Wyatt Wilson¹, Dawn Eliashiv¹, Nick Hasulak⁴, Sonja Hiller³, Itzhak Fried³, Nanthia Suthana³, Cory S. Inman¹; ¹University of Utah, ²University of Southern California, ³University of California Los Angeles, ⁴Phoenix Research Consulting

Event segmentation theory proposes that people divide continuous experience into discrete events using mental event models in working memory. These models integrate perceptual, semantic, and episodic information to predict what will happen next. Event boundaries—where one meaningful event ends and another begins—are key points in working and long-term memory, marked by increased processing and recall. Prior research shows high agreement among individuals on when boundaries occur, but often uses edited, third-person footage, which may not reflect real-world experience. Using first-person perspective videos, we hypothesized that segmentation agreement and alignment would be strongest and most immediate for highly salient physical context shifts and highly unpredictable events. We analyzed segmentation patterns using 38 unedited, first-person videos recorded by five epilepsy patients with intracranial EEG devices as they navigated UCLA’s medical campus, offering a more naturalistic view of everyday experience. A total of 430 undergraduates completed an event segmentation task by pressing a button when they believed one meaningful event had ended and another had begun. Results showed that physical boundaries—such as spatial thresholds and context changes—had the highest agreement and temporal alignment, partially supporting our hypothesis. Event segmentation theory suggests that information experienced around these physical boundaries should demonstrate increased recollection. In a secondary analysis, segmentation patterns from the patients who completed event segmentation tasks on their videos significantly overlapped with those of the undergraduate viewers but showed minor differences suggesting that similar but not identical event segmentation processes are used when viewing novel experiences and familiar, past lived experiences.

Topic Area: PERCEPTION & ACTION: Other