Abstract
The human memory system balances two complementary functions: generalization, which involves extracting regularities across experiences, and specificity, which involves retaining the unique details of individual events. Understanding how this balance develops can benefit from a neurocomputational perspective in which three core computations, pattern categorization, pattern separation, and pattern completion, differentially support generalization and memory specificity. Guided by this framework, we present a dataset from 180 children aged 4–8 years who completed four age-appropriate memory tasks that provide behavioral indices of these computations. Children additionally completed three standardized intelligence subtests. Alongside the dataset, we describe a semi-automated data-cleaning pipeline designed for developmental behavioral data, which can serve as a template for identifying and resolving common data-quality issues in studies with young children. In sum, this dataset provides a resource for investigating the development of memory computations supporting adaptive functioning across early and middle childhood.