Working Memory Model (Edexcel A-Level Psychology): Revision Notes
Working Memory Model
Introduction to the working memory model
Alan Baddeley and Graham Hitch first proposed the working memory model in 1974 as an alternative to the multi-store model of memory. Atkinson and Shiffrin (1968) had used the term 'working memory' to describe the short-term store, but Baddeley and Hitch identified problems with this approach. The multi-store model was overly simplistic and focused too heavily on rehearsal as the key to learning. Baddeley and Hitch wanted to understand short-term memory as a more complex and active system.
The working memory model describes how we temporarily store and manipulate information whilst we are using it. We rely on working memory for many everyday tasks, including:
- Remembering telephone numbers and lists
- Understanding sentences and sequences of words
- Mental calculation and reasoning
- Following instructions whilst performing a task
Working memory has limited capacity and is easily disrupted by distraction or by overload (such as complicated calculations or trying to remember too many items).
The model structure
Baddeley and Hitch proposed that working memory consists of three main components:
- A central executive that manages and controls the memory system
- Two slave systems that process specific types of information:
- The phonological loop for verbal information
- The visuospatial sketchpad for visual and spatial information
Working memory is viewed as a limited capacity system that can only handle a restricted amount of information temporarily whilst it is being manipulated or processed.
Worked Example: Mental Multiplication
If you multiply in your head, you need to temporarily store the initial numbers (15 and 32) whilst retrieving multiplication knowledge. You likely perform two calculations ( and ) and must store these intermediate results before adding them together. All these processes are performed by working memory.
The central executive
The central executive was originally described as a limited capacity component involved in general processing. It was essentially seen as a homunculus (a little man) with a supervisory role, deciding how the two slave systems should be used. The central executive has limited capacity but can handle different types of sensory information (modality free).
In the early development of the theory, the role of the central executive was unclear. Later refinements described it more precisely as an attentional controller with the capacity to focus, divide and switch attention between different tasks or information streams.
Psychology as a science
The central executive is largely a theoretical concept with limited direct experimental support. Because it is abstract, it cannot be directly tested and does not meet the criterion of being scientific in that empirical data cannot easily be gathered. This represents a limitation when evaluating the working memory model from a scientific perspective.
The phonological loop
The phonological loop is a slave system that deals with the temporary storage of verbal information. The phonological loop was initially believed to have two components:
- The articulatory rehearsal system - allows subvocalisation or refreshing of verbal information
- The phonological store - holds a limited amount of verbal information for a few seconds, which can be extended through rehearsal
Evidence for the phonological loop
The phonological store can explain the phonological similarity effect: it is more difficult to remember similar sounding words and letters (man, cad, mat, cap, can) compared to words and letters that sound different from one another (pen, sup, cow, day, hot). However, this effect was not found when remembering words that had semantic similarity (huge, long, wide, tall, large) or words that were semantically unrelated (thin, wet, old, late, strong). This demonstrates that the phonological store relies on acoustic encoding for storage (Baddeley, 1966a).
The articulatory rehearsal system was used to explain the word length effect: short monosyllabic words (cat, rug, hat) were recalled more successfully than longer polysyllabic words (intelligence, alligator, hippopotamus). Longer words filled up the limited capacity of the articulatory rehearsal system, resulting in words positioned earlier in the list being lost - the longer the word, the more capacity was used up and forgetting was more likely.
This also explained why recall deteriorated when rehearsal was prevented through articulatory suppression (repeating the word 'the' whilst learning a word list).
Research on language acquisition and impairment
Subsequent research into the phonological loop has provided understanding of language acquisition. Researching an Italian woman (VP) with a phonological impairment, Baddeley found that she was unable to retain any vocabulary learned from a different language, suggesting that the phonological loop may have evolved for language acquisition (Baddeley et al., 1988).
Further research using children with Specific Language Impairment (SLI) demonstrated that they found it more difficult to recall non-words (slime, pell, sep) than words (pen, set, dot, cap, vet), and this correlated to the size of their vocabulary. This finding suggested that the phonological loop was necessary for language acquisition and that deficits in this component of working memory resulted in difficulty learning and comprehending novel language (Gathercole and Baddeley, 1996).
Non-word repetition tasks are now a standard and widely used test for SLI as an indicator of Specific Language Impairment.
Psychology as a science
Much of the research into working memory is experimental and laboratory based, involving the testing of specific hypotheses concerning the nature of short-term memory that have testable outcomes (e.g. word recall accuracy). This research meets many of the criteria of being scientific because there is an emphasis on control, objectivity and replicability.
Visuospatial sketchpad
The visuospatial sketchpad (VSSP) is a slave system that temporarily holds and manipulates verbal and spatial (position/location) information. The VSSP can deal with visuospatial information either directly through observing images or by retrieving visuospatial information from long-term memory. The role of the VSSP is to maintain and integrate visual and spatial information from these different channels using a visual code.
We rely on visuospatial information in our long-term memory to remember routes home or navigate familiar environments.
Research on visual and spatial components
Recent research has attempted to distinguish between the visual and spatial components of the VSSP using tasks that test memory span. Spatial span has been tested using the Corsi block tapping task, where participants are presented with a series of blocks on a screen that light up in a sequence that they have to repeat. Error frequency increases with the number in the sequence, suggesting a limited capacity to spatial memory.
Research Evidence: Klauer and Zhao (2004)
Research by Klauer and Zhao (2004) found that:
- Visual memory tasks were more disrupted by visual interference
- Spatial tasks were more disrupted by spatial interference
This offers evidence for separate components to the visuospatial sketchpad.
This separation of components was further supported by Darling et al. (2007), where 72 non-student participants were presented with a series of white squares positioned randomly on a black screen. In one of the squares was the letter 'p'. Participants had to recall either the appearance (font) of the letter 'p' or its location within a square on the screen. Participants then experienced either spatial interference (tapping keyboard keys in a figure of eight) or visual disruption (a visual array of black and white flickering dots on the screen) before recalling the information.
The researchers found that the spatial interference task disrupted spatial memory but not memory for appearance, and that the visual disruption task affected memory for appearance but not for location. This provides evidence for separate visual and spatial memory systems. However, visual memory often has to contend with an array of visual stimuli rather than just one category of visual information (the letter 'p'), so this experiment does not reflect everyday visual processing of images.
Evaluation
Evidence for separate visuospatial and phonological subsystems comes from both experimental research and neurophysiological evidence.
Neurophysiological evidence
Williams syndrome is a rare condition where individuals show normal language ability but impaired visual and spatial ability. Individuals with this condition are affected by the same phonological factors, such as word length and word similarity, as the general population, but perform poorly on Corsi block tapping tests. This offers clinical evidence for separate visuospatial and phonological subsystems.
Interestingly, children with Williams syndrome were also found to have problems comprehending sentences with spatial prepositions (words that describe the position of an object in relation to another object, such as behind, underneath, against), suggesting an association between visuospatial memory and language acquisition (Phillips et al., 2004).
Further neurological evidence comes from the single case study of KF (Shallice and Warrington, 1974) who suffered short-term memory impairment following a motorbike accident that damaged the parietal lobe of his brain. KF had a digit span of one, suggesting a gross impairment in his phonological store, but his visual memory was intact. In contrast, Henry Molaison suffered from a gross impairment in his spatial memory with a relatively unaffected short-term memory for verbal information. This supports the proposal that working memory has two subsystems that process verbal and visuospatial information relatively independently.
Neuropsychological case studies offer an insight into memory function but are limited to unique individuals with specific impairments, so care should be taken when generalising these findings.
Evidence from neuroimaging
Neuroimaging has also offered evidence for the localisation of the different subcomponents of working memory in the brain. Paulesu et al. (1993) demonstrated that different regions of the brain were activated when undertaking tasks that employed the phonological store and the articulatory rehearsal system. Using a PET scan, they found that Broca's area (an area of the left frontal lobe associated with language production) was activated during a subvocal rehearsal task (remembering words) and the supramarginal gyrus (an area of the parietal lobe associated with language perception) was activated when the phonological store was being used. This research provides evidence for the phonological loop and its separate subcomponents.
However, the exact location of the central executive has been difficult to find as it is largely diffuse across the cortex.
Experimental evidence
Dual task experiments require participants to perform two tasks simultaneously that involve one or more slave systems of working memory. Baddeley and Hitch (1976) conducted an experiment where participants had to simultaneously use a pointer to track the location of a moving light on a screen whilst imagining the capital letter 'F' and mentally tracking the edges of the letter, then verbally saying whether the angles they imagined were at the top or bottom of the image.
Dual Task Experiment Results:
Participants could easily complete each task separately, but had difficulty performing the tasks simultaneously. This shows how two visual tasks both compete for the limited resources of the visuospatial sketchpad, resulting in impairment in performance.
However, when participants were asked to perform the visual tasks whilst undertaking a verbal task at the same time, performance was not affected because one task used the visuospatial sketchpad and the other task used the phonological loop.
Dual task experiments offer support for separate visual and verbal slave systems because performance is affected by whether the tasks compete for the limited resources of the same or different slave systems.
Research into separate visual and spatial memory systems
Recent research into the visuospatial sketchpad has been concerned with distinguishing between the visual and spatial components. Klauer and Zhao (2004) found that visual memory tasks were more disrupted by visual interference and spatial tasks more disrupted by spatial interference, offering evidence for separate components to the visuospatial sketchpad.
Alzheimer's disease and the role of the central executive
Evidence for the coordination role of the central executive is far less extensive than research into the subsystems. Nevertheless, research with clinical patients suffering from Alzheimer's disease (a neurological degenerative disease that impairs cognitive functioning, causing memory loss and impairments in thinking and language) has shown decreased central executive function as the disease has progressed.
Alzheimer's Research: Baddeley et al. (1991)
Baddeley et al. (1991) conducted a series of dual task experiments on young, elderly and Alzheimer's patients using verbal and visual tasks together or separately.
Key Finding: The performance of the Alzheimer group did not differ from the other groups when performing a visual or verbal task but showed impairment when trying to do them together.
According to Baddeley et al., the central executive is responsible for the coordination of the subsystems, so this impairment in performance demonstrates problems with executive functioning.
The episodic buffer
A limitation with the working memory model was its inability to explain why we could store only a limited number of word sequences in the phonological loop but could account for longer sentence sequences (up to 15 to 20 sentence units). It seemed that word sequences in the form of sentences could be bound together by meaning and grammar that could not be explained by the limited capacity of the phonological loop alone, and this somehow related to information held in long-term memory.
A further problem with the original model was that it did not explain how verbal and visual encoding could be combined. The model could not explain how the subcomponents could interface with each other or with long-term memory.
Baddeley addressed this in 2000 with the addition of a fourth component called the episodic buffer. The episodic buffer was proposed as a limited capacity storage system that could integrate information between the subcomponents and feed into and retrieve information from long-term memory.
Individual differences
It can be claimed that the working memory model ignores individual differences in capacity and functioning of each subsystem, or at least fails to explore these individual differences. Yet we know that some people have better short-term memory than others and, in many cases, poor working memory has been associated with dyslexia and Specific Language Impairment.
Remember!
Key Points to Remember:
-
The working memory model (Baddeley and Hitch, 1974) describes short-term memory as a complex, active system with multiple components rather than a single store.
-
The model consists of four components: the central executive (attentional controller), phonological loop (verbal information), visuospatial sketchpad (visual and spatial information), and episodic buffer (interfaces with long-term memory).
-
Strong experimental and neurophysiological evidence supports the existence of separate phonological and visuospatial slave systems, including dual-task studies, brain imaging research, and clinical case studies.
-
The phonological loop explains effects such as phonological similarity and word length, and plays a role in language acquisition, as demonstrated by research with Specific Language Impairment.
-
The episodic buffer was added in 2000 to explain how working memory interfaces with long-term memory and how different types of information are integrated.