Huge kudos to @bernaguler.bsky.social for spearheading this project from design to writing, and to Fatih Serin for his critical contributions throughout. It was a privilege to collaborate with such brilliant and dedicated colleagues.
We'd love to hear your thoughts!

These results challenge the long-standing view that prediction errors are the key triggers of event segmentation. Instead, they suggest that stable contextual structure is what shapes episodic memory.

In Experiment 4, we manipulated predictability: some context transitions were preceded by a counter, making them fully expected. Segmentation was equally strong, whether transitions were predictable or not.

In Experiments 1–3, transitions across stable contexts (e.g., task rule, object category) produced stronger segmentation than prediction errors without stable context. Prediction error alone wasn’t enough.

We directly compared these two views across four experiments by independently manipulating contextual stability and prediction error—something past studies often confounded.

Event segmentation—the process of turning continuous experience into discrete memory episodes—is thought to be triggered by prediction errors. But an alternative account suggests it’s transitions between stable contexts that matter more.

Also by former lab members:

Nursena Ataseven: Revealing the effect of regularities on representations in visual working memory.

Nursima Ünver: EEG decoding reveals a link between visual working memory fidelity and the magnitude of similarity-induced memory bias.

Correction on the first announcement: Berna Güler is presenting now in Pavilion. The poster (#481) is regarding the working memory's role on the temporal structuring of episodic memories. It is hidden among temporal processing posters. :)

Correction on the first announcement: Berna Güler is presenting now in Pavilion. The poster (#481) is regarding the working memory's role on the temporal structuring of episodic memories. It is hidden among temporal processing posters. :)

May 20, 2:45 - 6:45 PM, Pavilion
56.425 EEG decoding reveals a link between visual working memory fidelity and the magnitude of similarity-induced memory bias. Nursima Ünver, Rosanne Rademaker, Keisuke Fukuda

Also see posters of former lab members :)

May 20, 2:45 - 6:45 PM, Pavilion
56.430 Revealing the effect of regularities on representations in visual working memory. Nursena Ataseven, Sahcan Özdemir, Daniel Schneider, Wouter Kruijne, Elkan G. Akyürek

and a small monthly stipend. Explore our projects here: pure.sabanciuniv.edu/projects?fac.... General information here: pure.sabanciuniv.edu/announcement...