Schema-driven prediction effects on episodic memory across the lifespan

Javier Ortiz-Tudela, Gözem Turan, Martina G. Vilas, Lucía Melloni, Yee Lee Shing

TL;DR

This study explores how our brains use "schemas"—mental templates of how the world works—to prioritize what we remember. By testing children, young adults, and older adults with visual narratives, the researchers found a shifting "U-shaped" relationship between surprise and memory: while children and young adults are primed to remember unexpected "prediction errors," older adults lean more heavily on remembering information that fits their established expectations.

What Problem Does This Paper Address?

The "predictive processing" framework suggests the brain isn't a passive recorder but an active prediction machine. We constantly project future states based on internal models (schemas). When reality deviates from our prediction, it creates a Prediction Error (PE).

Previous research suggests a U-shaped relationship between PE and memory: we remember things best when they perfectly match our expectations (schema integration) or when they completely defy them (novelty/surprise). However, the "meh" middle ground—events that are neither expected nor surprising—is often forgotten. This paper investigates how this mnemonic priority changes across the human lifespan. Since children are still building their world models and older adults have highly stable ones, do they use these two memory-boosting mechanisms differently?

Methods

The researchers recruited three groups: children (8-10), young adults (18-25), and older adults (65-70). They used a novel experimental paradigm involving visual narratives (comic strips).

  1. Encoding Phase: Participants viewed 60 five-panel comic strips. The first four panels established an action (e.g., someone throwing a ball). The fifth panel (the ending) was manipulated into three conditions:
  2. Expected: The action completes as predicted (low PE).
  3. Unexpected: A surprising interruption occurs (high PE).
  4. Neutral: The action remains unresolved/ambiguous (intermediate PE).
  5. Distractor Phase: 10 minutes of math problems to clear short-term memory.
  6. Retrieval Phase: Four tasks tested different memory facets:
  7. Cue Recognition: Identifying the first panel (peripheral details).
  8. Object Recall: Naming the central object in the story.
  9. Outcome Memory: Remembering if the goal was achieved.
  10. Ending Recognition: Picking the correct ending from three options.

The team used Linear Mixed Models (LMM) and polynomial analysis to distinguish between linear trends (favoring surprise) and quadratic trends (the U-shape).

Datasets

The study utilized a custom-designed set of 80 five-panel comic strips depicting simple interpersonal actions.

  • 60 strips were used for encoding (20 per condition).
  • 20 strips were used as "foils" (new items) for the recognition tasks.
  • The experiment was conducted online via PsychoPy and hosted on Pavlovia, allowing for a diverse but controlled testing environment.

Key Insights and Ideas

  • The Lifespan Shift: The most significant "aha" moment is the double dissociation between age groups. Children and young adults show a "left-tilted" U-shape, meaning they prioritize the Unexpected. Older adults show a "right-tilted" U-shape, prioritizing the Expected.
  • Children as Information Sponges: In childhood, schemas are unstable. The brain prioritizes Prediction Errors because they signal the most "learnable" information needed to update their developing world models.
  • Older Adults as Pattern Matchers: As we age, the mechanism for encoding surprise (likely linked to the hippocampus) may decline, while the ability to integrate information into existing, robust schemas remains a primary strength.
  • Holistic Encoding: Surprise doesn't just help you remember the "shocking" ending; it actually boosts memory for the entire episode, including peripheral details present at the start of the story (cue recognition).

Key Results

  • Encoding Performance: All groups performed at near-ceiling levels (~97-99% accuracy) in identifying the story outcomes during the initial viewing, ensuring they were paying attention.
  • Cue Recognition: Children and young adults showed a strong positive linear trend—they were significantly better at recognizing the start of a story if it ended unexpectedly. Older adults showed no such modulation.
  • Ending Recognition:
    • Children: Remembered unexpected endings best, followed by expected, with neutral being the worst (Left-tilted U).
  • Young Adults: Similar to children, but with a slightly reduced boost for expected endings.
  • Older Adults: Showed the opposite; they had significantly better memory for Expected endings than for unexpected ones (Right-tilted U).
  • The "Neutral" Valley: Across almost all tasks and ages, the "Neutral" (intermediate PE) endings were remembered the worst, confirming that the brain filters out information that is neither predictable nor novel.

Limitations and Future Work

  • Task Difficulty: The "Object Recall" and "Outcome Memory" tasks suffered from floor effects (they were too hard), which may have masked some PE effects in those specific categories.
  • Schema Complexity: The researchers noted that the comic strip schemas might have been too simple for adults. Future work could use more complex, multi-layered narratives to see if adult schema-integration benefits become even more pronounced.
  • The "Neutral" Problem: Defining a truly "neutral" baseline is a known challenge in cognitive science. The authors used "unfinished" endings, but acknowledge that a lack of closure itself can sometimes impact memory.
  • Neural Mechanisms: While the paper discusses the likely roles of the hippocampus and medial temporal lobe, future fMRI studies using this paradigm could pinpoint exactly where the age-related shift in PE-processing occurs in the brain.

Summary updated

Abstract

The predictive processing framework posits that one of the main functions of the brain is to anticipate the incoming information. Internal models facilitate interactions with the world by predicting future states against which actual evidence is compared. The difference between predicted and actual states, the prediction error (PE), signals novel information. However, how PE affects cognitive processing downstream is not fully understood: one such aspect pertains to how PE influences episodic memories, and whether those effect on memory differ across the lifespan. We examine the relationship between PE and episodic memory in children, young and older adults. We use a novel paradigm whereby rich visual narratives are used to build action schemas that enable probing different mnemonic aspects. To create different levels of PE, we manipulate the story endings to be either expected, neutral or unexpected with respect to the unfolded action. We show that (i) expected endings are better encoded than neutral endings and (ii) unexpected endings improve the encoding of mismatching events and other aspects of the narrative. These effects are differentially modulated across the lifespan with PE-driven encoding being more prominent in children and young adults and with schema integration playing a larger role on memory encoding in older adults. These results highlight the role of predictions by enriching past experiences and informing future anticipations.This article is part of the theme issue 'Elements of episodic memory: lessons from 40 years of research'.
BibTeX
@inproceedings{ortiztudela2024schema,
  author = {Javier Ortiz-Tudela and Gözem Turan and Martina G. Vilas and Lucía Melloni and Yee Lee Shing},
  title = {{Schema-driven prediction effects on episodic memory across the lifespan}},
  booktitle = {Philosophical},
  year = {2024},
  doi = {10.1098/rstb.2023.0401},
}
Figure