Models of Explicit and Implicit Memory: Linking Priming, Recognition, and Source Memory

Memory comes in many different forms. Understanding how these different forms of memory operate and are enabled in the brain is a crucial goal in psychological and neuroscientific research. A fundamental question is whether there are distinct forms of memory that function independently of other kinds (multiple-systems theory), or whether different forms of memory actually draw upon the same underlying memory (single-system theory). This project will conduct rigorous tests of an existing single-system model by a) applying it to a memory task thought to be challenging for the model to explain -- source memory, and b) comparing its predictions to those of multiple-systems versions of the model in eight new experiments that are designed to discriminate between the models.

In previous research, the applicant has found evidence for a single-system model of recognition memory and repetition priming, and this model was preferred to multiple-systems ones. Recognition memory involves being able to determine whether you have encountered something before, for instance, being able to say that you have seen the face of a perpetrator of a crime before. Repetition priming is a more indirect example of memory -- it is typically shown via improvements in performance when a stimulus is re-encountered; for example, when presented with a familiar face that you have seen recently, you will tend to identify it faster than a face that you have not seen for a long time. Recognition memory is a traditional example of explicit (conscious) memory, whereas repetition priming is a traditional example of implicit (unconscious) memory.

The new task to be explained by the single-system model is source memory, another explicit memory task, which involves remembering specific information related to a prior encounter. For instance, source memory would be shown if you were able to remember that you saw a particular face at the scene of a crime and not at the supermarket, or if you were able to correctly remember that the colour of the jacket of the perpetrator of a crime was red and not blue. It is important for the single-system model to also explain source memory because many theories propose that the memory used to make source memory decisions is distinct from the kind of memory that drives repetition priming. If the relationship between source memory and repetition priming could be shown to follow the predictions of the single-system model, this would have important implications for these alternative theories, and would be a crucial test of the breadth of single-system views.

Multivariate signal detection theory will be drawn upon in extending the model. The main assumption of the single-system model is that priming, recognition, and source are driven by the same memory signal. Specific predictions of the model will be tested experimentally. For example, the model predicts that repetition priming will be positively related to the accuracy of source memory judgments and also the confidence with which these judgments are made. It also predicts that this relationship is modulated by recognition memory. Furthermore, it predicts that experimental variables will tend to affect priming, recognition and source memory similarly, and so would be strongly challenged by experiments in which these measures are differentially affected. Experiments manipulating attention and modality of presentation at encoding will investigate this. The model will be fit to the experimental data and compared with the fit and predictions of multiple-systems models; these will incorporate competing theoretical assumptions to the single-system model and allow for various degrees of independence between the memory signals driving explicit and implicit memory.

This research project will use a mathematical modelling approach to test competing theories of how different forms of memory are related to one another. The main goal is to advance our understanding of human memory. In the proposed research, a new mathematical model of memory will be developed. This can be used by other researchers as a framework for understanding the relationship between different measures of memory. Although the primary impact of the research is likely to be academic in nature, there is significant potential for the project to have a broader societal impact across a number of groups, as detailed below.

General public
A website will be built with information of the project's aims, background, and findings, and this will be accessible to the general public. A series of public lectures are also planned in order to engage and inform the general public on topics directly related to the project. These include CogTalk (Plymouth University) and outreach events run by the British Psychological Society (e.g., Psychology in the Pub). One aim will be to increase the public's awareness of how mathematical models can be used to guide psychological research in memory and how they can be used to test and develop theories. Another will be to increase the public's appreciation of 1) the multifaceted nature of memory and the different forms it takes, and 2) the importance of using mathematical modelling to identify general principles underlying human memory functioning, which may help to inform how best to utilise memory in everyday life.

Practitioners
Amongst clinical health practitioners, there is a general consensus that certain forms of memory are spared in particular memory disorders such as amnesia and ageing. For instance, there is a view that implicit (unconscious) memory, but not explicit (conscious) memory, is spared in amnesia. Recent research has found evidence to the contrary, showing instead that implicit memory is impaired in amnesia (and ageing). This deficit is actually predicted by the single-system mathematical model of memory that will be used in the proposed research, and demonstrates that this model may ultimately be useful in the evaluation of patients with memory disorders.

Academics in other fields
The mathematical model developed in the project may be useful for researchers working in other research fields and disciplines. Indeed, development of the model draws heavily upon signal detection theory, a framework that was originally developed by engineers to detect a signal in background noise. The potential for broad application of signal-detection models is clear: It is currently widely used across diverse areas such as attention, perception, decision making, deception, and eyewitness testimony. The multivariate signal detection modelling framework produced by this research will be useful for more extended applications of signal detection theory in these fields.

Project Staff
The project will also contribute to the development of the cross-transferable skills-base of the Researcher Co-Investigator working on the project. They will receive training in mathematical modelling, a tool which is used in a variety of domains and industries--from the civil service to the private sector--to understand and predict the behaviour of systems. Finally, the project will contribute to the expertise and skills-base of the Principal Investigator, allowing them to build upon the modelling foundations that were successfully laid in previous ESRC funded research.