The Human Memory System
This blog post is a slightly modified version of an essay that I recently submitted as part of my Master of Education (Educational Psychology) studies.
The ability of the human brain to process, store and retrieve information has been the subject of much research and debate by cognitive psychologists over a long period of time. Whilst the terminology and functioning of the components of the human memory system (also known as our human cognitive architecture or HCA) have changed based on the findings of research, it is widely accepted that our memory system consists of a sensory memory that receives information from our surrounds, a working memory to process this information and also to retrieve information from our storage area known as the long-term memory. This essay will discuss the research and findings about how the human memory system operates and furthermore, the application of these findings to learning and instruction.
As far back as 1890, the human memory system was proposed by William James to be a dual-system comprising of a primary memory (or conscious awareness) and a secondary memory (containing lasting memories). However, it was not until the late 1950’s that evidence and acceptance for the division of the memory into multiple systems began to emerge. Around this time, research by Peterson and Peterson found that unfamiliar information could only be held for a matter of seconds before being forgotten. In addition, Brown proposed that memory traces decay over time and his experiments also demonstrated forgetting occurring over short time periods. This in turn led to the proposal that human memory be separated into short-term and long-term systems. These findings added to the earlier research of Miller (1956) who found that there were limitations to the capacity of information that can be processed by the human memory system, in this case he discovered that only about seven (plus or minus two) pieces of new information could be held at any time.
During the 1960’s, short-term memory (STM) and long-term memory (LTM), also referred to as the long-term store (LTS), were conceptualised as separate systems and this was reflected in the various models that began to emerge. Of these, Atkinson and Shiffrin’s Modal model (1968) became the most influential depiction of the human memory system. This model “assumes that information comes in from the environment through a parallel series of sensory memory systems into a limited-capacity short-term store (STS), which forms a crucial bottle-neck between perception and LTM. The STS was also assumed to be necessary for recall, and to act as a limited-capacity working memory” (Baddeley, 2000, p.81).
Despite the influence of the Modal model, two shortcomings became apparent in the early 1970’s. The first the assumption was that if information was held in the STS for a sustained amount of time, there would be an increased likelihood that it would be transferred to the LTS. This did not account for any processing of the information and was contested by Craik and Lockhart in 1972, who incidentally, were not in favour of multi-store models, instead conceptualising memory as being “tied to levels of perceptual processing” in what they referred to as the primary memory. This resulted in the development of their framework of levels of processing known as Type 1 or surface processing and Type 2 or deep processing. The second shortcoming of the Modal model was that the structure of the model itself implies that a person with a damaged STS would therefore experience problems processing information and also long-term learning, however further studies on people with impaired STS found that this was not the case.
While the Modal model proposes a single STS, Baddeley and Hitch (in 1974), conceptualised a working memory comprising of three components – two slave systems known as the phonological loop and the visuo-spatial sketchpad both of which are controlled by the central executive – that replace the STS. The term working memory has largely been adopted in preference of the term short-term memory and better reflects the processing activities carried out by this part of the human memory system. The phonological loop is somewhat similar to the conceptualisation of the STS and is comprised of a phonological store that holds “acoustic or speech-based information for 1 or 2 seconds” (Baddeley, 1992, p. 558) and an articulatory control process that, with repetition, circulates information from the phonological store via an inner voice. In addition, the process also converts “visually presented material such as words or nameable pictures” into a form that can be registered by the phonological store (Baddeley, 1992, p.558). The role of the visuo-spatial sketch pad is to process visual as well as spatial information. Coordinating the activities of both slaves systems is carried out by the central executive however, unlike the phonological store and visuo-spatial sketchpad, the empirical evidence demonstrating the existence of a central executive has not been found.
This model of working memory was modified by Baddeley in 2000 in order to address two concerns arising from the original model. The first concern was in relation to the integration of the working memory components (due to each component coding information differently) and the second was around how the working memory communicates with long-term memory. To address these concerns a fourth component known as the episodic buffer was added to the model was assumed to be “a limited capacity temporary store that forms an interface between a range of systems all having different basic memory codes. It is assumed to do so by having a multi-dimensional coding system” (Baddeley et. al. 2010, p.229). However, rather than the active link between the subsystems, it was found that the episodic buffer was a more passive store of bound information and not responsible for binding coded information.
It must also be mentioned that information enters the working memory from one of two sources, either from our sensory memory through our interaction with the world around us or it is retrieved from our long-term memory.
Whilst the working memory is responsible for the active processing of information, the long-term memory is the storage area of the human memory system. Our long-term memory “consists of a large, relatively permanent store of information” (Sweller, 2004, p.11). It was the work of De Groot who found that chess grand masters were able to defeat novice players because they held vast numbers of board configurations in their long-term memory. This was demonstrated when grand masters were able to accurately reproduce real game board configurations compared to novice players. This finding was confirmed by Chase and Simon (1973) who found that grand masters could reproduce mid-play board configurations with fewer referrals back to the mid-play board. Interestingly, Chase and Simon also found that grand masters performed worse than novice players when attempting to reproduce random chess board configurations because they were attempting to apply actual configurations to a haphazard setting.
Information is stored in the long-term memory in knowledge structures known as schemas. Also known as mental models, schemas “permit us to treat a large number of information elements as a single element” (Clark et. al, 2006). The reason that chess grand masters performed better than novice players in reconstructing actual board configurations is because they have many more board configuration schemas stored in their long-term memory that they can access. The number of schemas held is what differentiates experts from novices therefore, the focus of any instruction should be the formation and construction of schemas in the long-term memory.
Implications for Learning and Instruction
The study of the human memory system and its components has provided extensive evidence about how humans process and store new and existing pieces of information. This knowledge is essential when it comes to designing instructional activities and account for the processing and storage capabilities of the human memory system. When learning something new, there are three types of cognitive load: intrinsic which is the inherent level of complexity of the content, germane which allow cognitive resources to be put towards learning and extraneous which are irrelevant elements that actually impose extra mental processing. These forms of cognitive load are additive, therefore in order for instruction to be effective and permit transfer to long-term memory, they should not exceed working memory capacity.
Cognitive load theory is “a universal set of instructional principles and evidence-based guidelines that offer the most efficient methods to design and deliver instructional environments in ways that best utilise the limited capacity of working memory” (Clark et. al, 2006, p.342). Examples of these principles include: the worked example effect – giving novice learners worked solutions of unfamiliar problems to study, the split-attention effect – reducing the need to integrate multiple sources of information in order for it to be understood, the modality effect – presenting information via both the visual and auditory channels and the redundancy effect – not presenting the same information via both the visual and auditory channels. Applying these principles to instructional design will facilitate improved learning outcomes because they incorporate the findings of research into the functioning of the human memory system.
Baddeley, A. D. (1992). Working memory. Science, 255(5044), 556-559.
Baddeley, A. D. (2000). Short-Term and Working Memory. In Tulving, E., & Craik, F. I. M. (Eds) The Oxford Handbook of Memory, 77-92, Oxford University Press.
Baddeley, A. D., Allen, R. J., & Hitch, G. J. (2010). Investigating the episodic buffer. Psychologica Belgica, 50(3&4), 223-243.
Clark, R., Nguyen, F., & Sweller, J. (2006). Efficiency in Learning, San Francisco: John Wiley & Sons Inc.
Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11, 671-684.
Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. The Psychological Review, 63(2), 81-97.
Ricker, T. J., Vergauwe, E., & Cowan, N. (2014). Decay theory of immediate memory: From Brown (1958) to today (2014). The Quarterly Journal of Experimental Psychology, 1-27.
Sweller, J. (2004). Instructional design consequences of an analogy between evolution by natural selection and human cognitive architecture. Instructional Science, 32, 9-31.