Memory development for children with Down syndrome

There has been a significant amount of research conducted into the working memory function of children with Down syndrome in recent years. The research has focused on the phonological loop and visuo-spatial scratchpad, rather than the function of the central executive. This may be because many children with Down syndrome would find the tasks used to assess central executive function rather difficult to understand and to carry out.

The research into the two short-term storage systems has consistently demonstrated that for children with Down syndrome:- [see [@jarrold_down_1999]

• Visuo-spatial short-term memory spans are significantly better than verbal short-term memory spans. For typically developing children the pattern is usually the opposite, with verbal short-term memory spans better than visuo-spatial ones.
• Visuo-spatial short-term memory spans are similar to the spans for typically developing children and other children with learning disabilities of similar non-verbal mental age.
• Verbal short-term memory spans are poorer than the spans for typically developing children and other children with learning disabilities of similar non-verbal mental age.

A specific impairment in verbal short-term memory

This research indicates that, for children and teenagers with Down syndrome, their verbal-short-term memory spans are smaller than those of children of similar non-verbal mental age, (both typically developing younger children and children of similar age with learning disabilities but not Down syndrome). In fact, as teenagers, individuals with Down syndrome usually have digit spans of 2, 3 or 4 digits while typical teenagers have digit spans of 6, 7 or 8 digits. This difference is significant and will limit the language and cognitive processing abilities of teenagers with Down syndrome in their everyday lives.

The reason for comparing the performance of children with Down syndrome with children matched for mental age measures, is that this information may help to explain the reasons for their difficulties. If they have similar skills to other children of the same mental age, then this would seem to be a pattern of general intellectual delay and it might be assumed that memory skills are in fact progressing as expected. However, if the memory skills are actually lagging behind other mental age measures then it suggests that they are being influenced by some specific difficulties that might be affected by intervention. A number of studies have been designed to try and find out why the verbal short-term memory spans of children with Down syndrome are specifically impaired and the findings are discussed in the next section, which provides a review of the factors thought to influence development in the working memory system.

What influences working memory development during childhood?

Many studies have demonstrated that the capacities of all three components of working memory increase during childhood. There has been more research aimed at understanding the capacity of development in the phonological loop than in the other two parts of the system. [@gathercole_cognitive_1999] [TODO: 16] Knowledge of the factors that have been found to influence the development of the working memory system in typically developing children may help to identify ways to improve the development of working memory for children and adults with Down syndrome.

A dynamic system

In reading the information on the factors which influence the components of working memory, it is important to remember that it is a dynamic system, developing over time from infancy to adulthood and is influenced by experience and use, as well as by development in other parts of the sensory, perceptual and cognitive systems in the brain.

Influences on verbal short-term span

In typically developing children, the number of verbal items which can be held in the phonological loop increases with age up to about 16 years, with the most rapid change taking place between 3 and 10-11 years. A number of factors are thought to influence the improvement in verbal short-term spans and this information may be important in providing pointers for possible ways of improving phonological loop function. [@gathercole_cognitive_1999]

Auditory discrimination skills and hearing

Any hearing or auditory discrimination difficulties which lead to poor acoustic information being available to the phonological loop would be likely to impair its function. [@gathercole_cognitive_1999] The hearing difficulties which are common in children and adults with Down syndrome may be significantly reducing the acoustic information available.

In recent experimental working memory studies with children and teenagers with Down syndrome, variation in hearing loss has not explained the limited spans of the children. [@jarrold_short-term_2001] However, even if hearing loss at the time of the study does not relate to variation in performance, hearing loss during preschool years could have had an influence in limiting the auditory information necessary to create an accurate sound pattern for words.

Through their practical work with children with Down syndrome and their families, the authors regularly collect examples to support this view i.e. the 3 year old who uses the same sign for ‘man’ and ‘van’; the same sign for ‘red’ and ‘bread’; the same sign for ‘sauce’ and ‘horse’; though pointing correctly to examples of the two different items. These examples all suggest that the children are hearing only approximations of the words, i.e. “an”, “ed”, “orse”.

Longitudinal studies are needed, which follow children’s development from infancy and monitor their hearing difficulties over time, in order to identify the role that hearing loss plays in all aspects of speech, language and memory development. The levels of hearing loss associated with the conductive loss experienced by the majority of preschoolers with Down syndrome is in the region of 35-45dB. This level of loss will make it difficult for the child to hear the discriminating consonants for words which rhyme, as illustrated by the examples above. If the child cannot hear discriminating consonants, then they will not store accurate representations of the sound pattern of the word (the phonological pattern) in short or long term memory.

In addition, the child will not have accurate sound patterns to support accurate production of the words in speech as they begins to talk. The possible significance of the articulation and phonological difficulties (the unclear speech patterns) that are common in children with Down syndrome do not seem to have been considered in working memory research. If a child cannot hear the sounds or say the sounds clearly, they will have significant difficulty in building a store of accurate and stable phonological (sound) patterns for words. [@dodd_speech_2001]. This, in turn, will impair function in the phonological loop as, for familiar words, there is evidence that long-term memory for their phonological patterns influences short-term spans (see below).

One study of teenagers and young adults with learning difficulties has demonstrated links between auditory discrimination skills, speech discrimination, speech and language skills and persistent middle ear dysfunction for those with, but not for those without, Down syndrome. [@marcell_relationships_1995] The researchers suggest that lower auditory acuity and slower processing speeds both contributed to the poorer performance of the individuals with Down syndrome. In the practical experience of the authors, the minority group of children with Down syndrome who have not had any glue ear or other identified hearing loss have significantly better language comprehension and production skills by 5 years of age and much clearer articulation and phonology than the majority of children with Down syndrome. It would be interesting to know if their verbal short-term memory spans are also better than average.

The construction and maintenance of the memory trace in the phonological loop

The ability of the phonological loop to construct and to maintain a memory trace will effect verbal short-term memory skills. [@gathercole_working_1993] [TODO: 11] This refers to the basic function of the phonological loop itself. Does it construct a good representation of the sound pattern of the word and does it maintain this representation for an adequate length of time? This may vary between individuals; some may have phonological loops that create more accurate and detailed representations of the sound patterns of words than others. Similarly, the speed with which this trace decays may vary, with some people losing the information faster than others. The most recent review of the research into phonological loop function in Down syndrome by Jarrold and Baddeley [TODO: 7] concludes that the evidence indicates that the basic function of the loop may be impaired in individuals with Down syndrome and also recommends that research into rates of decay is needed.

There could be two explanations for the apparent impairment in phonological loop function in individuals with Down syndrome. Firstly, it could reflect a biological disadvantage - less effective function in the brain - although it should be remembered that brain functions usually show dynamic development, influenced by input and activity. Secondly, it could reflect the long term effects of the basic speech and language difficulties leading to poorer performance in the phonological loop.

Several studies have established that there is a link between phonological loop function and vocabulary development in typically developing children, and some authors stress that the phonological loop determines the rate of new vocabulary learning [@gathercole_cognitive_1999],[TODO: 18] but others suggest the opposite effect - that more vocabulary use by the child develops the phonological loop.[TODO: 19,20] The first argument, that phonological loop function determines vocabulary development, is based on the view that the phonological loop is essential for holding the phonological (sound) pattern of the word in order for a stable, long term representation to be constructed. Construction of an adequate, stable long-term representation of the phonological pattern is thought to require many exposures to a word, over time.[TODO: 11] The second argument, that vocabulary use develops phonological loop function, is based on the view that, as vocabulary size increases during childhood, the child will develop more detailed and accurate representations of words. Initially, the brain will store an approximate sound pattern for the whole word, but over time this will become more precise, with each speech sound (phoneme) in the word represented more accurately. This process is likely to take place over at least 6-7 years in typical development and to be affected by the learning of grammar and improved phonology (the ability to make all the sounds in words clear when speaking) as children begin to talk in sentences. Specialists in language development argue that control over grammar and phonology are linked and develop together.[TODO: 21, 22]

It is likely that both arguments are correct and that both effects are at work over time - illustrating the dynamic nature of development in this system.

The familiarity of the material

There is evidence from a number of studies that the familiarity of the material to be held in the phonological loop influences span scores. [@gathercole_cognitive_1999] Therefore spans will be larger for words or numbers known to the individual than for unfamiliar words or nonwords. This has implications for individuals with Down syndrome as their vocabularies are likely to be small relative to their age and relative to their non-verbal abilities. In addition, even the words that they do know have probably not been listened to or spoken as frequently as they would have been by a typically developing person.

It is likely that the more often a word has been heard, the better the phonological information about that word in long term memory would be, and the better the temporary trace produced in the phonological loop. Similarly, the more often the word has been said by an individual, the quicker they may be in producing the word under test conditions. In addition, the more often the individual has said the word, the more often they may have received feedback to help them improve the accuracy of their stored phonological ‘template’ for the word.

Reconstruction

Linked to the last point, if the words to be recalled are familiar, then the listener may be able to reconstruct a partially decayed trace of the word from their stored knowledge. [@gathercole_cognitive_1999] Some research with children with Down syndrome suggests that it takes them longer to access the phonological codes that they have stored in their long-term memories.[TODO: 23],[TODO: 24] If this is the case they might not be able to make such efficient use of reconstruction processes when faced with verbal short-term memory tasks. However, there is evidence that they are able to make some use of long-term stored phonological knowledge. In a study which used nonword repetition tasks with individuals with Down syndrome they found it easier to remember and recall those that were more like real words, showing that, like other children and adults, their learned knowledge of real words was helping them.[TODO: 25]

The ability to retain the order of the items

All tests of verbal short-term memory require the individual to remember the correct order of the words. Anyone having difficulty with retaining the order information will have smaller span scores. Research with individuals with Down syndrome does not suggest that they have any particular difficulty in retaining the order of the items. [@chapman_language_2017]

The ability to retrieve and repeat the information

Listening to and repeating words, non-words or digits, requires both storage of the information and the ability to retrieve it. It also requires the individuals to speak the word. The evidence in memory research all indicates that storage and retrieval are separate processes. Information may be stored adequately but there may be impairments in the retrieval process. In addition, someone with speech-motor difficulties may take longer to plan and organise the spoken output of the words, losing some of the information in the process.

There has been some investigation of these issues in relation to individuals with Down syndrome. The time between the researcher saying the last word and the individual starting to repeat the words back is known as the latency period. In some studies this latency period is longer in individuals with Down syndrome, suggesting that it may take them longer to either process the information or organise the spoken response, or both. [@seung_digit_2000] However, other studies have indicated latency periods which are not longer when compared with appropriate comparison groups. This is an area needing more research.

Rehearsal

At about 7 years of age, children begin to spontaneously rehearse information that has to be remembered either aloud or silently (sub-vocal rehearsal). The speed of this rehearsal process is related to the child’s speech rate. Therefore the increase in speech rate that occurs as children get older explains a large part of the increase seen in verbal short-term memory spans. [@gathercole_working_1993] [TODO: 11, 23]

Some of the research with individuals (children and teenagers) with Down syndrome indicates that they are able to use rehearsal strategies even though they may not have reached an overall mental age of 7 years. [@broadley_working_1995] [TODO: 30] There is evidence from training studies that children and teenagers with Down syndrome can be taught to rehearse information when engaged in memory tasks.[TODO: 29-31] However, their slow speech rate for age will be limiting the benefits of this rehearsal.

Influences on central executive function

To date, there has been no research on the central executive function of individuals with Down syndrome. However, in typical developing individuals, increases in processing efficiency and increases in attentional capacity are thought to explain the increases in central executive capacity. [see [@gathercole_cognitive_1999]

Increases in processing efficiency

The central executive is a limited capacity processing system and its ability to support complex tasks such as processing language during conversations, reading with comprehension or doing mental arithmetic will be influenced by how well the component parts of the tasks are practised or automatised. For example, in reading, if word recognition is well established and words are therefore recognised rapidly without the need to actively decode them, the reader can use more of their working memory capacity to read the text for meaning. Similarly, if multiplication tables are known, these number ‘facts’ to support mental calculations will be available without having to ‘work them out’ and there will be more working memory capacity available to support the rest of the calculation task. Practise of tasks leads to automatisation. For typically developing children, mental and motor skills tend to be practised daily as they are being mastered, so that they become automatised. When a child is acquiring skills more slowly, and each of the skills is more difficult for them to learn and to carry out, then they do not get the level of practise that the typically developing child would have. This lack of practice will lead to a delay in automatising skills.

The result of having to process words for meaning (and to store them) will be to overload the working memory system when trying to carry out tasks such as reading aloud and reading for meaning at the same time. The first author has had the experience of asking a teenager to do this in the course of a reading assessment, and the young man read the passage aloud correctly but could not answer the comprehension questions. However, when allowed to then read the passage over to himself again, he could answer the comprehension questions. This young man had significant speech production difficulties which included a stutter. It seemed that in order to read an unfamiliar passage aloud, accurately and intelligibly, he needed to use all the immediate processing capacity available to him. The author knew, from his classroom performance, that he should have been able to comprehend the material. This experience was a very clear demonstration of the demands of many cognitive tasks on processing capacity. Tasks that typically developing individuals can carry out simultaneously, may create difficulties for children and adults with limited working memory capacities. Teachers may conclude that the tasks are beyond their abilities, but they may not be, if taken one stage at a time.

Increases in attentional skills

The role of attentional skills in increasing working memory capacity has not been extensively explored and is an area that requires research for individuals with Down syndrome. It may be that it is not only the ability to sustain attention that is important, but the ability to switch attention between tasks or manage several aspects of a task simultaneously. Practise and experience of handling more complex tasks with support is important. It may also be important to encourage a child to record each step in the process rather than try to hold them in memory while tackling the next step.

The importance of memory development for children with Down syndrome

The phonological loop and learning to talk

The evidence from research with typically developing children suggests that the phonological loop plays a critical role in learning a spoken language, as it holds the sound pattern for the word to enable the child to store it and to link it to the meaning of the word. If the child cannot hold a clear sound pattern then language learning will be made more difficult. The examples of the children who seem to be unable to discriminate between rhyming words such as ‘van’ and ‘man’ or ‘dolly’ and ‘lolly’ illustrate this point. Children with Down syndrome learn to understand spoken words much more slowly than typically developing children and impairment in the phonological loop may be part of the reason for this. Children with Down syndrome also have difficulty in learning the grammar to understand and build sentences. Research with typically developing children suggests this is also affected by the function of the phonological loop. [see [@gathercole_working_1993]

A number of studies have tried to explore the links between measures of verbal-short term memory and measures of spoken language skills for children with Down syndrome. They report the same findings as the studies of typically developing children. Measures of verbal short-term memory skills seem to be closely related to the expressive language skills of children and teenagers with Down syndrome - that is, more strongly linked with their spoken language ability than their language comprehension. This finding might be expected, as spoken language will be improved by having better phonological information to support speech. [@seung_digit_2000] Precise phonological information may not be as important for comprehension of language, especially as contextual cues in the situation often help to support the meaning of spoken language. In a recent study with 33 children and adolescents with Down syndrome, a nonword repetition measure of verbal short-term memory was significantly related to the children’s vocabulary comprehension, grammar comprehension and reading abilities.[TODO: 32] In a follow up study of the same young people 5 years later, the measures of verbal-short term memory collected on the first occasion were significantly related to the progress made in receptive and expressive vocabulary.[TODO: 33]

In summary, there is increasing evidence that the functioning of the phonological loop plays a causal role in language learning and in spoken language development in children with Down syndrome, as it does for other children. The implication of this is that if phonological loop functioning can be improved for children with Down syndrome, this will improve their speech and language skills.

Reading, language and memory

Research with typically developing children has shown that learning to read influences the ongoing development of working memory skills. [@ellis_early_1988][TODO: 35] As children begin to learn to read, those who have better verbal and visuo-spatial short-term memory skills learn to read faster. Over the following school years, those children who are better readers show bigger gains in short-term memory measures. Research by the authors and their colleagues has demonstrated the same positive effects of reading instruction on language comprehension and short-term memory development in children with Down syndrome.[TODO: 36]

Table 1. Mean matrices, language and memory scores for readers and non-readers in 1991 and 1995 (raw scores) [@laws_influence_1995] Reproduced with permission

Table 2. Age equivalent scores for 1995 BPVS and TROG measures for readers and non-readers [@laws_influence_1995] Reproduced with permission

The children in this study had similar abilities on all the measures at the start of the study, as illustrated in [Table 1]. After 4 years, the non-verbal abilities of the children (Matrices scores) were still no different. Both groups had made the same amount of progress with non-verbal reasoning. However, the children who had received regular, daily reading instruction had made much faster progress in speech, language and memory development. The gains are equivalent to some two years of typical developmental progress - that is, the ‘reading’ children are now two years ahead of their peers who did not receive the same level of reading instruction (see [Table 2]).

The children in this study had all taken part in memory training in the first year of the study. The data in [Table 3] illustrates that the 2 groups of children benefited equally from memory training - see their scores to date. However, those in reading instruction continued to show increase in their memory spans over time, while the training gains were lost for the other children. Memory training is discussed further in the next section.

Table 3. Mean auditory and visual memory scores for readers and non-readers (s.d.s in brackets) [@laws_influence_1995] Reproduced with permission

This is an example of the dynamic nature of cognitive development and the reciprocal effects of gains in different parts of the system. It would seem that progress in either speech and language skills, or short-term memory skills or reading will each affect the other two in a positive manner.

Research with typically developing children has shown that training them in phonological awareness boosts non-word repetition scores. [@gillam_phonological_1996]. Phonological awareness training helps children to discriminate speech sounds and to be able to hear them when they are parts of words. For example, to be able to identify words that rhyme, to break words into their component sounds (cat = c-a-t) and to build words from their component sounds (t-a-p = tap). All this happens as children learn about phonics as they learn to read and spell in the classroom. Presumably this improves verbal short-term memory as a result of sharpening up the children’s ability to discriminate sounds and to discriminate words, particularly if they are very similar such as ‘man’ and ‘van’. The implication of this is that all work that helps children with Down syndrome to improve their speech sound discrimination skills, right from infancy, will not only help them to develop spoken language but will also improve working memory function.

Memory training for children with Down syndrome

It will be clear from all the evidence discussed above that there are potentially many ways of helping children with Down syndrome to improve their working memory function but few of these have been evaluated. However, one strategy in particular has been tried and that is the training of a rehearsal strategy. Since rehearsal of information definitely helps to retain it, several research groups have set up training studies to evaluate rehearsal training for children, teenagers and adults with Down syndrome

Evidence from training studies

The nature of short-term memory skills for children with Down syndrome

The authors and colleagues began to investigate working memory in children with Down syndrome, and the effects of memory training, in 1991.

In the first study 62 children, from 4 to 18 years of age, completed a battery of memory tests designed to explore their short-term memory skills and some of the children then took part in memory training. [@broadley_working_1995] The information collected at the start of the study illustrated that, while the visual and verbal short-term memory spans of the children with Down syndrome were small, their short-term memories appeared to function in the same way as in typically developing children.

For example, their performance on verbal tasks indicates that they were using phonological coding. [Table 4] illustrates that the children’s spans for longer, three syllable words were shorter, and [Table 5] illustrates that they found similar sounding words more difficult than dissimilar words.

The word-length effect, the poorer recall of longer words, is taken to reflect the fact that words that have longer phonological patterns will take up more space in the limited capacity phonological loop than short words, so that fewer can be held. The acoustic similarity effect is thought to be due to greater difficulty in discriminating between similar sounding words, as the memory traces in the loop decay.

Table 4. Mean span scores for 1, 2 and 3 syllable words as a function of presentation modality and age [@broadley_working_1995]] Reproduced with permission

Table 5. Mean span scores for acoustically similar (AS) and dissimilar (AD) words as a function of modality of presentation and age [@broadley_working_1995] Reproduced with permission

Table 6. Mean digit span scores from the British Ability Scores as a function of modality of presentation and age [@broadley_working_1995] Reproduced with permission

For the word tasks in this study there were no differences between verbal and visual spans. The visual spans in these tables are spans for spoken words with picture support for each word that the children had to recall. The auditory span tasks required the children to listen to spoken words and repeat them.

[Table 6] illustrates the results for digit span - recall of random numbers. On this task the children did show better spans in the visual than the auditory presentation mode. That is, they did significantly better when they saw the written numerals as well as hearing them, but not when they saw pictures to support words.

A possible explanation of the benefit of visual support for digits but not words is that written digits are very precise and constant images which the children have learned to link to spoken number names. The use of pictures may provide a less efficient aid: while the picture represents the word, the actual picture used (of a dog, or tree, for example) will be new to the child. Printed words might be a more effective support than pictures as they, like numbers, are also precise and constant images for words.

In Tables 4, 5 and 6, the data illustrates that the children’s short-term memory skills are improving with age, albeit slowly. However, the word-length effects and acoustic similarity effects did not change with age, suggesting evidence of phonological coding of the information, but no evidence of a rehearsal effect. Rehearsal starts at about 7 years in typically developing children.

A further interesting finding was the demonstration of a visual similarity effect, that is, the children found that lists of items that looked similar e.g. (ball, wheel, apple, orange), (spade, screw, fork, rake), (television, case, gate, box) were more difficult to remember than lists of items with different shapes (see [Table 7]). This finding suggests that the visual memory mechanisms are being used to support verbal memory. The lack of a verbal memory advantage on any word task, also suggests the children are using visual coding. On all these tasks, typically developing children show better verbal than visual spans. In typical children this is interpreted as evidence for verbal recoding of visual information, i.e. pictures would be named and the names rehearsed to aid visual memory. The authors of the study of the children with Down syndrome suggest that these children may recode verbal information to visual information to aid their poor phonological recording of spoken words. They also suggest that phonologically based training - training speech sound discrimination, and word discrimination - should be evaluated.

Table 7. Mean word span for visually similar (Vis Simm) and dissimilar (Vis Diss) pictures as a function of age [@broadley_working_1995] Reproduced with permission

The benefits of memory training

Several training studies have been carried out in recent years, and have reported similar results. They have all attempted to teach rehearsal strategies to individuals with Down syndrome. In one study, an organisational strategy was also taught; this will be discussed later in the [practical section].

Training studies have all used visual materials, pictures and words, during the training phase, except the most recent one, which is using only auditory (spoken) input. [@conners_memory_2001] Typical techniques used to train rehearsal are explained in the [activity section] of this module.

The evidence from all the studies is that training can improve both verbal and visual short-term memory spans for children, teenagers and adults. [@broadley_working_1995; @laws_effects_1996; @comblain_working_1994] [TODO: 39] The effects for visual short-term memory were larger than the effects for verbal short-term memory. However, it is not clear that the effects of a short period of training last more than a matter of months. There could be several reasons for this, indicated by data collected by the authors and colleagues.[TODO: 36],[TODO: 39],[TODO: 40]

In their studies, the memory training effects were more likely to be sustained when the children with Down syndrome were in mainstream schools and were learning to read. Reading instruction seems to be particularly important. In fact, in the large training study conducted by the authors and colleagues over a four year period, the children’s reading progress was the only significant predictor of memory gains. [@laws_long-term_1995]

[Table 8] illustrates the auditory and visual short-term memory spans for the children who received training and the comparison group, three years after the training. The visual memory spans of the trained group were still significantly better, but not the verbal spans. [Table 9] illustrates the difference between readers and non-readers on the memory measures in 1995. Readers are those children who can read some words on the British Ability Scales Word Reading Scale. Non-readers could not read any of the words. Unfortunately, the majority of the readers were in mainstream placement and the majority of the non-readers in special schools, so the benefits could also be due to being in a better spoken language environment.

Table 8. Auditory and visual short-term memory spans of the trained and comparison groups [@laws_long-term_1995]

Table 9. A

Auditory and visual short-term memory spans of readers and non-readers [@laws_long-term_1995]

There are several possible influences in mainstream schools which may be helping. Firstly, the children have Teaching Assistants (LSAs) and these LSAs actually carried out the memory training so they may have continued practising the activities and consolidating the skills after the training period ended. They may have been alert to the use of the memory skills in classroom learning activities. Secondly, the children in mainstream classrooms receive more intensive reading and writing instruction than their peers in special school and they read and write as they record their work in all areas of the curriculum. Even the children who are not independent readers or writers are helped to record their work in written form by their LSAs. This means that even the children who do not have independent reading and writing skills are exposed to reading and writing daily, and may show benefits from this for speech and language development and memory development. Thirdly, the children with Down syndrome in mainstream schools are in a more stimulating and normal spoken language environment than will exist in many special schools, where most children have speech and language delays.

The effects of inclusion in mainstream schools

Two studies [@laws_language_2000] [TODO: 42] which evaluate the effects of inclusion for children with Down syndrome provide evidence of the benefits for speech and language development, with both studies reporting gains of more than two years on language measures.[TODO: 42] One of the studies also reports significant verbal short-term memory gains for the children in the mainstream environment.[TODO: 41]

In a recent paper comparing 22 children with Down syndrome in special schools in a county with almost no inclusion, with 22 in mainstream schools in a neighbouring area of high inclusion, the researcher reports significantly better memory spans for the included children (see [Table 10]). The children in the mainstream schools also had significantly better language skills than the children in special schools even though the two groups did not differ on measures of general non-verbal mental abilities.

Table 10. Auditory and visual short-term memory spans of children in special schools and mainstream education [@laws_language_2000]

The differences in grammar comprehension, and visual and verbal short-term memory, between the children in the two school systems remained significant even when age and receptive vocabulary scores were controlled for in the statistical analysis.

Reading ability may have again been a significant factor in the gains achieved in mainstream education, as 20 of the 22 included children could achieve at least a minimal score on the reading assessment, but only 3 of the 22 special school children.

A recent small study reported the positive effects of a computer memory training programme, used over six weeks with children with Down syndrome and other children with learning difficulties in the same schools. Both groups showed steady improvement on the memory tasks on the computer and gains on standardised assessments of memory skills from the British Ability Scales. [@solomon_effectiveness_2000] The children with Down syndrome in the mainstream schools in this study made faster progress than those in special schools even in a six week period. This was possibly because they had more consistent LSA support for the daily training sessions.

These studies emphasise that the benefits of inclusion are significant and produce greater gains for the cognitive development than have been demonstrated for any specific therapy or intervention strategies to date.

Summary

Memory research has demonstrated that:

• Children with Down syndrome have a specific impairment in verbal short-term memory development
• While this may be partly due to hearing, auditory discrimination, and articulation or speech-rate difficulties, there is also evidence of basic impairment in phonological loop function
• The phonological loop is thought to be essential for spoken language learning as it is thought to hold the phonological pattern (sound pattern) of the word as the child learns the meaning of a new word. The phonological pattern of the word will then be stored in long term memory, with its meaning, and be available to guide the spoken production of the word
• Impairment in the phonological loop may explain some of the short-term memory span difficulties which face children with Down syndrome when they are trying to remember two or more items. However, it may also explain at least some of the pattern of speech and language delays and difficulties experienced by most children with Down syndrome, right from infancy
• Activities to improve the function of the phonological loop should be considered as part of memory training activities, as well as activities to improve attention and to increase memory spans

Memory training studies have demonstrated that:

• Children with Down syndrome are using phonological coding but probably not rehearsal strategies as they get older, unless they receive specific training
• Both verbal and visual short-term spans can be improved by teaching children rehearsal strategies
• Visual short-term memory spans tend to improve with training more than verbal spans
• Visual information may be used to support verbal recall tasks and visual recoding of verbal information may be used to aid memory
• The beneficial effects of a few weeks of memory training do not last over time if training or support for memory skills is not maintained
• Children with Down syndrome who are readers have significantly better verbal and visual short-term memory spans. The readers in these studies may only have limited independent reading abilities yet they show substantial memory gains suggesting that being in reading instruction is the important experience. There are many possible reasons for the benefits of reading for memory.

1. Reading, phonics, spelling and writing activities may improve children’s auditory discrimination of speech sounds and their phonological awareness for sounds in words. In turn, this may lead to more accurate phonological representations in the phonological loop.
2. Knowledge and storage of printed words (the visual patterns of words) may provide visual cues to support the storage and retrieval of spoken words (the sound patterns of words) during a memory task. The printed word pattern may prompt recall of spoken words as whole units or provide cues to the specific sounds within the words. Reading activities provide spoken language practice for the children.
3. When their learning support assistants help them to write sentences and record work, the children read aloud and practice grammatically complete sentences at a stage when they are not using such sentences in their spontaneous spoken language. This may help their memory for longer sequences of information.
4. When children with Down syndrome read aloud, their clarity of spoken production improves. This effect is seen even in beginning readers of 3 years of age. Being able to say words more clearly may help to store more accurate phonological patterns for the words in long-term memory.

• Children with Down syndrome who are being educated in fully inclusive schools have significantly better verbal and visual short-term memory spans. The reasons for this could be the positive effects of being in a better spoken language environment, with more speech and language stimulation from other children, greater demands on the children to communicate and higher levels of expectation for language comprehension in the classroom. It could also be the effects of more intensive reading instruction and literacy use in mainstream classrooms, compared with special education classrooms.

This evidence indicates that even if children with Down syndrome have some degree of impairment in basic phonological loop function their short-term memory skills can be improved. More longitudinal research in schools is needed to try and identify exactly how the short-term memory skills are being affected by literacy teaching or by the experience of inclusion.

Implications for activities to improve working memory function

1. Phonological loop impairment: The part of the working memory system that seems to be the most impaired for children with Down syndrome is the phonological loop. This phonological loop impairment will make processing of single words and speech difficult as well as reducing the ability to remember two or more items.
2. Reduced memory spans for lists of items: This results in verbal short-term memory skills that are specifically delayed relative to the children’s non-verbal abilities and to their visual short-term memory skills.
3. Limited central executive capacity: There has been no specific research into central executive function as the tasks used are too difficult for most children with Down syndrome indicating very limited processing capacities. However, research with typically developing children suggests that increasing attentional skills and increasing processing efficiency should help in addition to increasing the storage capacity in the two storage systems.

These findings suggest 3 types of intervention strategies may help to develop the working memory system:

1. Activities to improve phonological loop function
2. Activities to improve remembering of lists of items
3. Activities to improve attention and to increase processing capacity

Suggestions for practical activities are provided for each of these aspects of working memory function in the next section.