The interplay between short-term memory and long-term memory.

Lead Research Organisation: University of Cambridge

Abstract

Short-term memory (memory over a few seconds) and long-term memory are neurally and cognitively separate, but they almost always operate in concert. Short-term memory never operates entirely on its own. Even a task as simple as remembering a sequence of digits and repeating them back immediately is influenced by information stored in long-term memory. For example, the digits are easier to remember if they are similar to numbers we already have in long-term memory, such as a telephone number. Exactly how does this happen? How do short- and long-term memory interact with each other? This is the question this project aims to answer. We use a combination of behavioural experiments with both adults and children, neuroimaging, and computational modelling. The aim is to develop a computational model of the interaction between short- and long-term memory.

Technical Summary

The primary function of memory is not simply to remember the past, but to construct representations that can guide our actions in the future. Our perceptual and cognitive processes are always influenced by our prior experience. This programme investigates the way experience (information in long-term memory (LTM)) influences short-term memory (STM), and the way in which new experiences are combined to construct effective representations of prior knowledge in LTM. Although long-and short-term memory are neurally and cognitively separate short- and long-term memory systems, the two necessarily operate in concert. How does this interaction occur? From a theoretical standpoint most of the work takes place within a Bayesian framework. Long term knowledge can be viewed as a set of priors which must be continually updated in the light of new information. In turn, these priors influence the interpretation of new information arriving in STM.
We investigate this by means of a combination of behavioural experiments, neuroimaging, and computational modelling.
We conduct behavioural studies of sequence learning and of the way in which ’chunks’ of information in long term memory help short-term recall. For example, this addresses the classic issue of how chunking influences the capacity of STM.
Our computational modelling work focusses on the challenging task of modelling the learning of representations of sequences. The target data-set for most of this work is an extensive body of literature on the Hebb repetition task. We also model the development of chunks in LTM and how this interacts with STM, and have developed a novel Bayesian model of recency effects in visual STM.
In order to understand the neural mechanisms underlying the interaction between STM and LTM conduct a complementary series of neuroimaging studies. These examine the neural representation of serial order in short-term memory and how those representations develop and are transferred into long-term memory.
The programme also encompasses research on training working memory conducted in collaboration with S. Gathercole, and a further strand on visual word recognition in collaboration with S. Kinoshita of Macquarie University.

Publications

10 25 50

publication icon
Gathercole, S.E. (2019) Working memory training involves learning new skills in Journal of Memory and Language

publication icon
Kinoshita S (2018) The semantic Stroop effect is controlled by endogenous attention. in Journal of experimental psychology. Learning, memory, and cognition

publication icon
Kinoshita S (2017) Evidence accumulation in the integrated and primed Stroop tasks. in Memory & cognition

publication icon
Kinoshita S (2017) The magic of words reconsidered: Investigating the automaticity of reading color-neutral words in the Stroop task. in Journal of experimental psychology. Learning, memory, and cognition

publication icon
Kinoshita S (2018) Orthographic and phonological priming effects in the same-different task. in Journal of experimental psychology. Human perception and performance

publication icon
Norris D (2018) Commentary on "Interaction in Spoken Word Recognition Models". in Frontiers in psychology