Retrieval as a fast route to memory consolidation

Human long-term memory has an astonishing capacity to store information about our own past experiences (episodic memory) as well as general knowledge (semantic memory). One of the big challenges in psychology is to understand what makes a memory last, i.e. why some experiences are easily remembered after days, weeks or even years, while other experiences fade and are ultimately forgotten.

The aim of this project is to investigate the role that active remembering plays in stabilizing newly acquired memories. Previous research has shown that remembering can protect memories against forgetting, such that experiences that are repeatedly retrieved after initial acquisition become more robust and more accessible in the long term. For example, in a recent study participants learned English-Swahili vocabulary pairs (e.g. "tomato - nyanya") and were then given the opportunity to practice the pairs, either by extra study ("tomato - nyanya") or by extra testing ("tomato - ?"). After one week, the vocabulary practiced by extra testing was retained 50% better than the vocabulary practiced by extra study. This striking memory improvement through retrieval ("testing effect") stands in sharp contrast to common classroom practice, where tests are typically used to measure learning success, rather than to enhance learning success. Interestingly, the study on vocabulary learning found that even university students are largely unaware of the memory-boosting effects of retrieval.

Surprisingly little is known about why retrieval has such a strong beneficial effect on memory. This project is aimed at investigating the neurocognitive mechanisms underlying the testing effect. All experiments are based on the central idea that memory traces change each time they are actively retrieved. Previous research has shown that the brain repeatedly "replays" newly acquired information during sleep or periods of awake rest after learning. Such offline reactivation seems crucial for the long-term stabilization (consolidation) of new memories. The current project will test the novel hypothesis that active retrieval mimics the changes that occur via offline consolidation, including making memory representations more semantic, less context-dependent, and more robust to interference. The experiments use an innovative combination of established and novel behavioural methods, as well as state-of-the-art pattern analysis of neuroimaging data, the latter being particularly well suited to track online the reactivation of individual memories in the brain.

The basic design used in all experiments is similar to the above vocabulary study. Participants will be given lists of materials to study once. They then receive i) no further practice, ii) practice by extra study, or iii) practice by extra testing. Retention of the materials will be assessed either immediately or after one week. The first set of experiments will explore whether visual and verbal materials equally benefit from testing. The second set of experiments will behaviourally test the central hypothesis that retrieval produces consolidation-like effects on memories. If retrieval mimics consolidation, it should have beneficial effects on some aspects of the learned information (e.g. semantic content), but detrimental effects on other aspects (e.g. contextual information). The final set of experiments uses functional magnetic resonance imaging (fMRI) to directly observe how the neural patterns that represent an individual memory change with repeated recall, and how these neural changes predict enhanced memory on a delayed test. Together, the results from these studies will advance our understanding of how active retrieval shapes and strengthens long-term memories. From an applied perspective, the results will help to identify the conditions under which repeated testing is useful in practice (e.g. in educational settings, or to boost memory in ageing populations).

The proposed work addresses the central question in memory research on how repeated testing changes newly acquired memory traces and leads to enhanced long-term retention. The project is basic experimental in nature, and its primary impact will thus be on the field of basic psychological and neuroscientific research on long-term memory. Academics in these fields will benefit from the innovative framework of this project, which introduces a new theoretical perspective on retrieval-mediated learning, generating novel predictions and findings to inform theories of human memory. The project also uses an innovative methodological approach that has the potential to methodologically advance the field.

Second, the research on retrieval-mediated learning has a high potential to be applied to educational contexts. Therefore, the second group expected to benefit from the project are those working in the educational sector, be it policy makers in the Department for Education, or teachers of students at all age classes. Prior research has shown that teachers and students are mostly unaware of the memory-enhancing effects of testing. The educational sector will thus globally benefit from enhanced awareness of this research. More specifically, the project addresses questions of how testing improves memory, and which types of learning materials can benefit from this improvement. This information is highly relevant for teachers and policy makers alike, particularly with respect to the development of educational materials.

Third, there has always been a broad interest in memory research from the general public. This interest has increased in recent years, partly as a consequence of concerns about the health and wellbeing of a constantly ageing population frequently reporting with memory impairments. The project, although not directly addressing age-related memory changes is generally concerned with strategies to improve memory, a topic that can be expected to attract the interest of this broad audience, and clearly fits within the "Lifelong Health and Wellbeing" priority area of the UK research councils. The findings from this project will therefore be disseminated to the general public (see 'Pathways to Impact'), providing accessible information about mnemonic strategies and specifically about how retrieval-mediated learning can protect newly learned information against forgetting.

Fourth, the project is expected to have a significant impact on the postdoctoral researcher's and the PI's scientific work. The PI is an early career scientist who began her first permanent academic position at the University of Birmingham in autumn 2013. The project is thus expected to establish important pathways to impact for the PI's future scientific work, with the objective to initiate novel research networks with scientists in related areas, setting the stage for future collaborative projects (e.g. together with researchers in applied fields like educational psychology). The post-doctoral research assistant (PDRA) will be involved in a wide range of academic and impact activities, providing extensive scientific and methodological training and improvement of the PDRA's communication skills.