A long tradition of research in cognitive psychology has allowed for the development of computational models of declarative memory that aim to capture how individuals acquire, retain, and forget information over time. These models use behavioral measures, such as the accuracy and latency of retrieval attempts, to estimate parameters of internal memory processes, allowing them to predict future memory retrieval performance. Interpreting learners’ responses in a retrieval task through the lens of a cognitive model provides a sensitive, personalized measure of memory performance with both educational and clinical applications.

In the educational domain, we can employ cognitive memory models in computerized adaptive learning applications that help learners memorize information, such as the MemoryLab learning system (previously SlimStampen). This system for adaptive retrieval practice continually adjusts the scheduling of items based on the memory performance of the individual student, allowing them to learn vocabulary items and other paired associates more effectively than with traditional methods. More generally, model-based insights in a student’s memory enable a more principled assessment of students’ mastery of the learning material.

In clinical settings, adaptive retrieval practice can also provide an accessible, low-cost, and non-invasive measure of patients’ memory performance. By measuring the speed and accuracy of memory retrievals, we can quantify individual differences in forgetting speed through the parameters of the memory model with high reliability and sensitivity. Clinical applications of this technique include identifying and tracking memory decline, as well as tracking changes in memory performance that result from interventions, disease or injury.