From chunks to schemas: Learning in the Hebb repetition paradigm

The short-term and long-term memory systems are thought to interact via chunks – long-term memory representations that support maintenance of information for a short term. The creation and use of chunks is examined in the Hebb repetition paradigm, where participants repeatedly study and retrieve the same list of serially presented memoranda (a so-called Hebb list), which ultimately become chunked. Studies conducted to date underscore the inflexible nature of thus created chunks, which always encompass the exact series of memoranda. However, the long-term memory system is thought to often rely on schemas – flexibly adjusted representations that allow for generalization and transfer of learning. Here we examined whether the Hebb repetition paradigm can be used to induce the acquisition of schemas and their use to support short-term memory performance. We adapted the Hebb repetition paradigm so that participants learned and retrieved lists of category exemplars. Crucially, what was repeated here in the Hebb list was the order of categories from which those exemplars were taken, while the memoranda themselves changed from one list presentation to another. We demonstrated that in this task participants were able to derive the schema describing category order in the Hebb list and use it to support short-term memory performance for a novel set of schema-consistent memoranda. We conclude that chunks and schemas can be seen as similar types of representations that jointly provide an interface between short-term and long-term memory.