Bridging Bayesian and representational theories of memory to predict memory bias.

Abstract

Understanding how people integrate gist and item-specific information is central to explaining how memory changes over time. We examined how these representations interact in visual memory by integrating a Bayesian framework with representational models of individual item and gist memory and tested the model on its capacity to predict people's gist-based memory distortions. To this end, participants completed separate tasks designed to measure gist and item-specific memory. We used data from both tasks and independent measures of people's stimulus representations to quantify memory fidelity for individual items and gist. We then substituted these parameters into our model to generate predictions of people's memory errors in a third task where people's memory for individual items was biased by category-level structure. Our model predicted entire distributions of people's memory errors, including the magnitude and direction of memory biases, as well as fine-grained individual differences in people's error distributions, such as their skew and variance, at different offsets. Our findings highlight the power of combining normative Bayesian with algorithmic representational modeling approaches to understand how people integrate noisy memory representations at different levels of abstraction. (PsycInfo Database Record (c) 2026 APA, all rights reserved).