The problems with short term memory
According to the multi-store model of memory, short term memory is a unitary and inflexible concept that has the simple properties of phonological encoding, a capacity of 7 plus or minus 2 items, and a duration of up to about 30 seconds. Whilst this is acceptable at a most basic level, it does not allow for patients such as KF who had a damaged short term memory (he had a digit span of just 2 items) but he was able to remember as much visual information in his short term memory as a non-brain damaged person. It also does not explain how people are able to perform two tasks at the same time, such as remembering a string of digits at the same time as performing a verbal resigning task (Baddeley & Hitch, 1974). Clearly short term memory must be a far more complex system than previously thought, and it must have some kind of control system that links separate verbal and visual components together.Structure of the working memory model
There are three main components to the original 1974 version of the working memory model. These are the central executive which acts like a conductor or manager allocating processing resources and coordinating the activity of the two slave systems, the phonological loop and the visuo-spatial sketch pad. The phonological loop is the inner ear and inner voice and is responsible for processing phonological information. The visuo-spatial sketch pad is the inner eye and processes visual and spatial information. Having separate slave systems explains how people are able to perform two tasks at the same time (one visuo-spatial and one phonological), and accounts for patients such as KF who have a reduced digit span (damage to the phonological loop) but normal visual short term memory.
Key features of the working memory model
The central executive is responsible for allocating processing resources to the two slave systems. It is the most important component of working memory and its disfunction is often implicated in many neurological disorders.
The phonological loop is auditory short term memory. It is a loop because it can store and repeat sound (either the inner voice or inner ear) for a limited period of around 20-30 seconds. It has two subcomponents, phonological store and articulatory control process. The phonological store is the inner ear and contains snippets of sounds either from the outside world or recalled from long term memory. The articulatory control process is the inner voice which essentially is our own voice inside our heads - imagine working out a maths problem and hearing your own voice speaking and manipulating the numbers over and over again.
The visuo-spatial sketch pad is visual short term memory, or the inner eye. It can manipulate images in two and three dimensions, for example you can recall a friend’s face from long-term memory in two dimensions and you can imagine walking around the kitchen in your house in three dimensions.Evidence for the individual components of working memory
One key strength for the working memory model is the wealth of research support for its various components.
The existence of the phonological loop was demonstrated by Baddeley et al (1975) with their study of the word length effect. If there is a phonological loop that lasts for about 20-30 seconds then it would make sense that it could store more single syllable words (short words) than it could multi-syllabic words (long words) as it takes longer to say multiple syllables than it does a single syllable. Badeley et al found that participants could indeed remember more single syllable words than multi-syllablic ones, and they called this the word length effect. However when they were prevented from rehearsing the words (by giving them an interference task that creates articulatory suppression, such as counting backwards in threes) the word length effect disappeared as they were unable to use their phonological loops for remembering the words because it was engaged in the articulatory suppression task. This provides strong evidence that the phonological loop exists.
Baddeley & Hitch (1976) demonstrated the existence of the visuo-spatial sketch pad by using a dual task paradigm. In this they asked participants to do two tasks that would utilise the visuo-spatial sketch pad either separately or at the same time. One task was to use a pointer to track a point of light moving around a screen, and the other task was to imagine a capital ‘F’ and mentally move round the edge of the letter classifying the angles as either belonging to its top or bottom. They found that participants were able to perform the tasks separately without any difficulty, but when they did them together their performance at both was impaired. If they asked them to do one of the two tasks at that same time as a verbal task then they were as good at the visual task as when performing it alone. This shows that when undertaking two visual tasks they compete for the limited resources of the visuo-spatial sketch pad and provides strong evidence of its existence.
Brain imaging studies have also been very useful in showing the existence of individual components of working memory. Paulescu et al (1993) used positron emission tomography (a PET scan) to record brain activity when participants were performing either verbal tasks (using the phonological loop) or visual tasks (using the visuo-spatial sketch pad), and found that very different areas of the brain were active during each task. This showed quite clearly that verbal and visual tasks are performed by different brain regions and therefore suggests that the phonological loop and visuo-spatial sketch pad are physically separate from each other as the working memory model suggests.Evaluating the working memory model
The working memory model has many strengths over the multi-store model. It describes short term memory as a collection of active processing mechanisms that work in two modalities (verbal and visual) rather than a single store that simply rehearses verbal information. This allows it to explain memory function in patients with impairments to one component of working memory but with normal function in the other components, such as KF.
There is considerable research evidence for the existence of the individual components of working memory (see section above).
The major weakness of the working memory model is the central executive which, ironically, is its most important component. There is relatively little research evidence for its existence and it is by its nature very hard to directly investigate, although some recent research into patients with Alzheimer’s disease (e.g. Baddeley et al, 1991) has identified how an impaired central executive may affect memory function. The problem with researching the central executive is that it cannot itself be measured directly, and instead its function has to be inferred from performance at verbal and visual tasks. When we want to measure the functioning of a component we give it a task designed to reduce its performance, such as occupying the phonological loop with an articulatory suppression task. If participants are able to perform the task then researchers conclude that the central executive is working properly, but if they perform the task poorly then they conclude that the phonological loop is impaired. In other words it is impossible to design a task that effectively tests the central executive without also affecting its slave systems, and so the central executive may never be adequately investigated and it will always remain a concept with limited support - Richardson (1984) calls this the circular argument.