iiinova — where your digital self keeps evolving

Dominika Varga

@dkvarga.bsky.social

438 followers 247 following 12 posts

Postdoctoral Researcher in the Vaghi Lab - Birkbeck Studying how individual brain circuits support learning and goal directed behaviour Previously at Sussex Episodic Memory Group

Posts Replies Media

Dominika Varga @dkvarga.bsky.social · 04/09/2025

In Exp1, when people hadn't seen any clips pre-scanning, hippocampal activity was not modulated by the expectedness of actions based on schema knowledge alone. However, in Exp2 and 3, when actions mismatched memories of the clips seen pre-scanning, hippocampal activity increased. (cont.)

Hippocampal mismatch response is limited to signaling episodic memory–based mismatches. In Experiment 1, expectations are based only on schematic knowledge, in Experiment 2, expectations are based on both episodic memory and schematic knowledge, and in Experiment 3, expectations are based only on episodic memory. Bar charts showing the average response to Typical and Atypical target actions in the bilateral hippocampus in the three experiments (strip plots show individual participants’ averaged parameter estimates for the target actions estimated from a GLM). Error bars represent 95% CIs.

110

Dominika Varga @dkvarga.bsky.social · 04/09/2025

Inside an fMRI scanner, people viewed clips of everyday events (e.g., doing the laundry), which could be Typical (putting clothes) or Atypical (putting flowers). To test what prior expectations the hippocampus uses to detect mismatches, we manipulated clip familiarity pre-scanning across experiments

Experimental paradigm. (A) Still frames showing example moments from the two alternative versions of the “laundry” video clip (watch the clips here: https://e01.eventmemory.org/ExampleVid_paper.html). The two versions showed a nearly identical sequence of actions, except for the target action that was either Typical or Atypical. (B) Before scanning, participants either did not watch any clips (Experiment 1), watched the Typical version of each clip (Experiment 2), or watched the Atypical version of each clip (Experiment 3). Across three experiments, all participants watched half of the clips in the Typical and the other in the Atypical version during fMRI. By manipulating prescan familiarity with the clips, target actions in each experiment violated different types of expectations. After scanning, participants were asked to describe what happened in all video clips watched inside the scanner, focusing on the actions the actors performed, cued by the first frame of each clip.

100

Dominika Varga @dkvarga.bsky.social · 04/09/2025

In daily life, we rely on our internal models of the world to predict what happens next—e.g., anticipating someone’s actions while they do laundry. 👇 But sometimes our prior expectations don’t match reality. These mismatches are important for updating our models and learning. How do we detect them?

110