Short-Term Memory Research Paper

2. Historical Origin Of The Construct ‘Short-Term Memory’

Suggestions of a distinction between two types of memory, one concerned with temporary retention, and the other having the function of a storehouse for materials which have been laid out of sight, date back at least to John Locke’s ‘Essay Concerning Human Understanding’ (1700). William James (1895) revived the distinction, suggesting the existence of a limited capacity ‘primary memory,’ embracing the present and the immediate past, and subserving consciousness. Psychological research in the nineteenth century and in the first half of the twentieth century was, however, mainly concerned with the diverse factors affecting learning and retention, in the context of a basically unitary view of human memory. It was only in the 1950s that short-term memory became the object of systematic behavioral studies in normal subjects (Baddeley 1976). In the late 1960s the division of human memory into a short-and a long-term system became a current view (Atkinson and Shiffrin 1968).

3. Functional Architecture Of Short-Term Memory

3.1 Evidence From Normal Subjects

Three main behavioral phenomena suggest the existence of a discrete limited-capacity system, concerned with short-term retention (Baddeley 1976, Baddeley 1986, Glanzer 1972). (a) Accuracy in recalling a short list of stimuli (e.g., trigrams of consonants or of words) decreases dramatically in a few seconds if the subjects’ repetition of the memory material (rehearsal) is prevented by a distracting activity such as counting backwards by threes. (b) In the immediate free recall of a sequence of events, such as words, the final five to six stimuli on the list are recalled better that the preceding ones. This ‘recency effect’ vanishes after a few seconds of distracting activity, and is minimally affected by factors such as age, rate of presentation of the stimuli, and word frequency. (c) In the immediate serial recall of verbal material (memory span), the subject’s performance is affected by factors such as phonological similarity and word length, with the effects of semantic factors being comparatively minor. Each of these phenomena subsequently proved to be considerably more complex than initially thought. They were also interpreted as compatible with a unitary, single-system, view of human memory. This account proved untenable, however, mainly on the basis of neuropsychological evidence.

These empirical observations illustrate the main characteristics of ‘short-term memory’: a retention system with limited capacity, where the memory trace, in the time range of seconds, shows a decay, which may be prevented through rehearsal. Material stored in short-term memory has a specific representational format, which, in the case of the extensively-investigated verbal domain, involves phonological codes, separately from lexical-semantic representations stored in long-term memory. The latter contribute, however, to immediate retention, e.g., in verbal span tasks.

The functional architecture of phonological short-term memory has been investigated in detail using effects which break down storage and rehearsal subcomponents. The effect of phonological similarity, whereby the immediate serial recall of auditory and visual verbal material is poorer for sequences of phonologically similar stimuli than for dissimilar ones, reflects the coding which takes place in the phonological short-term store. The effect of word length, whereby the immediate serial recall of auditory and visual verbal material is poorer for sequences of long words than for short ones, is held to reflect the activity of the process of rehearsal, abolished by ‘articulatory suppression,’ i.e., a continuous uttering of an irrelevant speech sound. Suppression, while disrupting rehearsal, also reduces immediate memory performance in span tasks. The interaction between phonological similarity, input modality and articulatory suppression, with suppression abolishing phonological similarity only when the stimuli are presented visually, suggests that rehearsal participates in the process of conveying visual verbal material to the phonological short-term store. Finally, some phonological judgments (such as rhyme and stress assignment) are held to involve the articulatory components of rehearsal, because the performance of normal subjects is selectively impaired by suppression (Burani et al. 1991).

In the nonverbal domain a similar distinction is drawn between visual and spatial short-and long-term memory systems, with the relevant representational format being in terms of the shape or spatial location of the stimulus (Baddeley 1986, Della Sala and Logie 1993). Also in the visuo-spatial domain, similarity, recency and interference effects have been observed. Visuo-spatial short-term memory is likely to comprise storage and rehearsal (pictorial and spatial) components. In the case of spatial locations, rehearsal may be conceived in terms of planned movements (e.g., ocular, manual reaching, locomotion) towards a target coded in a spatial reference frame, (e.g., egocentric).

3.2 Evidence From Brain-Injured Patients

Studies in patients with localized brain damage provide unequivocal evidence that supports the independence of short and long-term memory systems, conjuring up a double dissociation of deficits. In patients with ‘global amnesia,’ which may be characterized as a selective impairment of the declarative or explicit component of long-term memory verbal and visuo-spatial short-term memory are unimpaired, with immediate serial span, the recency effect in immediate free recall, short-term forgetting being within normal limits. This functional dissociation suggests a serial organization of the two systems, with temporary retention being a necessary condition for long-term storage (Atkinson and Shiffrin 1968).

Since the late 1960s selective impairments of short-term memory have been reported (Vallar and Papagno 1995, Vallar and Shallice 1990). The more extensively investigated area concerns auditory-verbal (phonological) short-term memory. Patients show a disproportionate reduction of the auditory-verbal span to an average of less than three items (digits, letters, or words); the recency effect in immediate free recall of auditory-verbal lists of words is abolished; short-term forgetting is abnormally rapid. The disorder is modality-specific, with the level of performance being better when the material is presented visually (Fig. 1). This input-related dissociation has two main implications: (a) discrete phonological and visual short-term memory components exist; (b) in the input– output processing chain, the phonological short-term store should be envisaged as an input system, rather than an output buffer store. This division argues against a monolithic view of the system as a single store, which is amodal, i.e., not specific for the different sensory modalities (see Atkinson and Shiffrin 1968). Additional support to this input locus of the system comes from the observation that in some of these patients speech production is entirely preserved.

Patients with defective phonological short-term memory may show unimpaired long-term verbal and visuo-spatial learning (Fig. 2). This observation further corroborates the complete independence of short-and long-term memory systems, but is incompatible with a serial organization, in which defective short-term memory entails a long-term memory impairment (see Atkinson and Shiffrin 1968), suggesting a parallel architecture instead. After early perceptual analysis, information may enter short or long-term memory, either of which may be selectively disrupted by brain damage. The learning abilities of these patients are, however, dramatically impaired when the phonological material to be learned does not possess pre-existing lexical-semantic representations in long-term memory. This is the case of pronounceable meaningless sequences of letters, such as nonwords, or words of a language unknown to the subject (Fig. 2).This phonological learning deficit indicates that, within a specific representational domain, temporary retention in short-term memory is necessary for stable long-term learning and retention, as predicted by serial models.

Other brain-injured patients show deficits of short-term retention in the visual and spatial domain, even though these patterns of impairment have been explored less extensively than the phonological disorder. Immediate retention of short sequences of spatial locations, as assessed by a block tapping task (a spatial analogue of digit span), may be selectively defective, with verbal memory being unaffected, in both its short and long-term components. In other patients, the deficit may involve the visual (shape) component of short-term memory, while impairments in the short-term visual recognition of unfamiliar faces, objects, voices, colours have been also described. The disorder of patients with defective visual imagery (as assessed, for instance, by tasks requiring colour or size comparisons) may be interpreted in terms of an impaired visual short-term memory store. A deficit of visuo-spatial short-term memory may also disrupt long-term learning of unfamiliar non-verbal material, as assessed by recognition memory for unfamiliar faces and objects. This extends to the visuo-spatial domain the conclusion that long-term acquisition requires short-term storage.

The process of rehearsal has been traditionally described as an activity which, through rote repetition, refreshes the short-term memory trace, preventing its decay. The precise characteristics of rehearsal have been elucidated in more detail in recent years, particularly in the domain of phonological memory. Rehearsal may be conceived in terms of the recoding of the memory trace from input (auditory-verbal) to output-related (articulatory) representations and vice-versa. More specifically, rehearsal of verbal material has long been regarded as ‘articulatory’ in nature, involving output-related verbal processes— such as the motor programming of speech production in a phonological assembly system or output buffer— the actual articulation of the material to be rehearsed (subvocal rehearsal), or both.

Anarthric subjects, who are unable to utter any articulated speech sound due to congenital disorders or brain damage acquired in adult age, may, nonetheless, show a preserved immediate memory, including verbal rehearsal. This suggests that the process is ‘central’ and does not require the activity of the peripheral musculature. Brain-damaged patients with a selective impairment of rehearsal (Vallar et al. 1997) show a defective immediate verbal span, as do patients with damage to the phonological short-term store. Both components of phonological memory contribute to immediate retention, though the phonological store provides a major retention capacity. The ability to perform phonological judgments is disrupted, however, by damage to the rehearsal process, but not to the phonological store. By contrast, a defective rehearsal leaves the recency effect in immediate free recall of auditory lists of words largely unimpaired. This recency effect is largely reduced or absent in patients with damage to the phonological short-term store.

4. The Uses Of Short-Term Memory

Is there a use for a system providing the temporary retention of a limited amount of stimuli, besides infrequent situations such as the following? A friend tells us an unfamiliar eight-digit number, which we have to dial on a telephone placed on the other side of a large street and we have no paper and pencil to write it down. The answer is positive. Short-term retention contributes to the stable acquisition of new information in long-term memory. More specifically, phonological short-term memory plays an important role in learning new vocabulary and participates in the processes of speech comprehension and production.

4.1 Long-Term Learning

The observation that patients with defective auditory verbal span are also impaired in learning unfamiliar pronounceable letter sequences gives rise to the possibility that phonological memory may contribute to a relevant aspect of language development, the acquisition of vocabulary (Fig. 2). Similarly, subjects with a developmental deficit of phonological memory are impaired in vocabulary acquisition and in non-word learning. An opposite pattern is provided by subjects with a congenital cognitive impairment which selectively spares phonological short-term memory. Acquisition of vocabulary, foreign languages and nonword learning are also preserved. Converging evidence from different subject populations supports this view. Correlational studies in children have shown that the capacity of phonological memory is a main predictor of the subsequent acquisition of vocabulary, both in the native and in a second language. In normal adult subjects, the variables which disrupt immediate memory span (phonological similarity, item length, articulatory suppression) also impair the acquisition of non-words.

Polyglots have a greater capacity of phonological memory, compared to nonpolyglots, and a better ability to learn novel words. Phonological short-term memory may be considered as a learning device for the acquisition of novel phonological representations, and the building up of the phonological lexicon (Baddeley et al. 1998). A few observations in brain-damaged patients suggest a similar role for visuo-spatial short-term memory in the acquisition of new visual information, such as unfamiliar faces and objects.

4.2 Language Processing

The idea that short-term retention contributes to speech comprehension dates back to the 1960s. Phonological memory may withhold incoming auditory- verbal strings, while syntactic and lexical-semantic analyses are performed. Patients with defective phonological memory show a preserved comprehension of individual words, as well as many sentential materials, and a normal ability to decide whether or not sentences are grammatically correct. This may reflect, on the one hand, the operation of on-line lexical-semantic processes, heuristics and pragmatics, and, on the other, the complete independence of syntactic and lexical-semantic processes from phonological memory. Patients are however impaired by ‘complex’ sentences, where ‘complexity’ refers to a number of non-mutually-exclusive factors, such as: (a) a high speed of material presentation, which prevents the immediate build-up of an unambiguous cognitive representation; (b) word order conveying meaningcrucial information (e.g., in sentences in which a semantic anomaly is introduced by a change in the linear arrangement of words: ‘The world divides the equator into two hemispheres, the southern and the northern’); (c) extralinguistic presuppositions biasing the interpretation of the spoken message.

Under such conditions, adequate interpretation may require backtracking to the verbatim (phonological) representation of the sentence, temporarily held in phonological memory. This provides a ‘backup’ or ‘mnemonic window’ resource for performing supplementary cognitive operations necessary for comprehension (Vallar and Shallice 1990).

5. Neural Architecture Of Short-Term Memory

5.1 Phonological Short-Term Memory

Anatomoclinical correlation studies in brain-damaged patients with a selective impairment of the auditory-verbal span indicate that the inferior parietal lobule (supramarginal gyrus) of the left hemisphere, at the temporoparietal junction, represents the main neural correlate of the ‘store’ component of phonological short-term memory (Vallar and Papagno 1995). The frontal premotor regions in the left hemisphere and other structures such as the insula are the major neural correlates of the ‘rehearsal’ component, even though the available anatomoclinical data are more limited (Vallar et al. 1997).

Functional neuroimaging studies in normal subjects concur with this pathological evidence, to suggest a left-hemisphere-based network. Activation in the left supramarginal gyrus [Brodmann’s area (BA) 40] is associated with the ‘store’ component of short-term phonological memory, activation in the left frontal premotor BA 44 (Broca’s area) and BA 6, and in the left insula, with the ‘rehearsal’ component (Paulesu et al. 1996). In these studies, in line with the behavioral evidence from normal subjects and patients, an immediate verbal span task activates both the inferior parietal region (phonological short-term store) and the premotor cortex (rehearsal process) in the left hemisphere. Conversely, rhyme judgements selectively activate the left premotor regions, whose damage, in turn, disrupts the patients’ ability to perform this task. These activation and lesion-based data support, from an anatomofunctional perspective, the behavioral distinction between a ‘storage’ component and a ‘rehearsal process’ in phonological short-term memory. Furthermore, they qualify ‘rehearsal’ as a process which makes use of components also concerned with the planning (i.e., programming in the left premotor cortex) of articulated speech. Seen in this perspective, phonological memory may be regarded as a component of the language system. Finally, connectionist modelling of this architecture is currently being developed (Burgess and Hitch 1999).

5.2 Visual And Spatial Short-Term Memory

Studies in brain-damaged patients suggest an association between damage to the posterior regions of the right hemisphere and defective spatial short-term memory, as assessed by tasks requiring the reproduction of sequences of spatial locations. There is also evidence suggesting that damage to the posterior regions of the left hemisphere brings about disorders of visual short-term memory, such as defective immediate retention of sequences of visual stimuli, (e.g. lines), and impaired recognition of more than one visual stimulus at a time (defective simultaneous form perception) (Vallar and Papagno 1995).

Neuroimaging activation studies in humans support the distinction between primarily spatial (a location, a ‘where’ component) and visual short-term memory systems (Smith et al. 1995). The main neural correlates of spatial memory for locations include the occipital extra-striate, posterior–inferior parietal, dorsolateral premotor and prefrontal cortices. Short-term visual recognition memory is associated with activations in a network including the posterior–inferior parietal, temporal and ventral premotor and prefrontal cortices. Right hemisphere regions may play a more relevant role in spatial memory for locations, while left hemisphere regions are more involved in visual memory for objects. These patterns of activation indicate an association of the ‘dorsal visual stream’ with spatial short-term memory for locations and of the ‘ventral visual stream’ with short-term visual recognition memory.

6. Conclusion

Behavioural observations and neuroanatomical evidence from normal subjects and brain-injured patients concur in suggesting that ‘short-term memory’ should be conceived as a multiple-component system with specific functional properties and discrete neural correlates. These systems secure the retention of a limited amount of material in the time range of seconds and contribute to relevant aspects of cognition, such as long-term learning.

Bibliography:

1. Atkinson R C, Shiffrin R M 1968 Human memory: a proposed system and its control processes. In: Spence K W, Taylor Spence J (eds.) The Psychology of Learning and Motivation. Advances in Research and Theory. Academic Press, New York
2. Baddeley A D 1976 The Psychology of Memory. Basic Books, New York
3. Baddeley A D 1986 Working Memory. Clarendon Press, Oxford, UK
4. Baddeley A D, Gathercole S, Papagno C 1998 The phonological loop as a language learning device. Psychological Review. 105: 158–73
5. Burani C, Vallar G, Bottini G 1991 Articulatory coding and phonological judgements on written words and pictures: the role of the phonological output buff European Journal of Cognitive Psychology 3: 379–98
6. Burgess N, Hitch G J 1999 Memory for serial order: A network model of the phonological loop and its timing. Psychological Review 106: 551–81
7. Della Sala S, Logie R H 1993 When working memory does not work: the role of working memory in neuropsychology. In: Boller F, Grafman J (eds.) Handbook of Neuropsychology. Elsevier, Amsterdam
8. Glanzer M 1972 Storage mechanisms in recall. In: Bower G H (ed.) The Psychology of Learning and Motivation. Advances in Research and Theory. Academic Press, New York
9. James W 1895 The Principles of Psychology. Holt, New York
10. Miller G A 1956 The magic number seven, plus or minus two: some limits to our capacity for processing information. Psychological Review 63: 81–93
11. Paulesu E, Frith U, Snowling M, Gallagher A, Morton J, Frackowiak R S J, Frith C D 1996 Is developmental dyslexia a disconnection syndrome? Evidence from PET scanning. Brain 119: 143–57
12. Smith E E, Jonides J, Koeppe R A, Awh E, Schumacher E H, Minoshima S 1995 Spatial versus object working memory: PET investigations. Journal of Cognitive Neuroscience 7: 337–56
13. Vallar G, Di Betta A M, Silveri M C 1997 The phonological short-term store-rehearsal system: Patterns of impairment and neural correlates. Neuropsychologia 35: 795–812
14. Vallar G, Papagno C 1995 Neuropsychological impairments of short-term memory. In: Baddeley A D, Wilson B A, Watts F (eds.) Handbook of Memory Disorders. Wiley, Chichester, UK
15. Vallar G, Shallice T (eds.) 1990 Neuropsychological Impairments of Short-term Memory. Cambridge University Press, Cambridge, UK