Practitioner Review: Short-term and working

memory impairments in neurodevelopmental

disorders: diagnosis and remedial support

Susan E. Gathercole and Tracy Packiam Alloway

University of Durham, UK

Background: This article provides an introduction to current models of working and short-term

memory, their links with learning, and diagnosis of impairments. The memory impairments associated

with a range of neurodevelopmental disorders (Down’s syndrome, Williams syndrome, Specific Lan-

guage Impairment, and attentional deficits) are discussed. Methods: Methods of alleviating the ad-

verse consequences of working and short-term memory impairments for learning are

porting learning during the childhood years. The

purpose of this review is to summarise current the-

oretical understanding of both working memory and

short-term memory, to characterise short-term and

working memory skills impairments associated with

a range of neurodevelopmental disorders, and to

identify methods of remedial support appropriate for

children with these impairments.

Working memory, short-term memory, and

learning

capacity of working memory is limited, and the

imposition of either excess storage or processing

demands in the course of an ongoing cognitive

activity will lead to catastrophic loss of information

from this temporary memory system.

An individual’s working memory capacity is usu-

ally assessed using complex memory span tasks in

which the participant is required to process and

store increasing amounts of information until the

point at which recall errors are made. The first

ensures that each sentence is processed. Other

variants of the complex memory span paradigm in-

clude listening span (in which the sentences are

presented auditorily; Daneman & Carpenter, 1980),

counting span (in which the task is to count the

number of target objects in each of a series of dis-

plays and subsequently recall their tally numbers;

Case, Kurland, & Goldberg, 1982), and operation

span (in which participants perform mathematical

calculations and finally attempt to recall unrelated

Alloway, Willis, & Adams, 2004a). Performance im-

proves steadily until the teenage years, when it starts

levelling off. By the age of 15, levels close to those of

adults are reached. At any particular age, the degree

of variation in working memory capacity is sizeable,

spanning many years of average development.

The key finding from the Daneman and Carpenter

(1980) study was that measures of verbal complex

memory span measures were strongly associated

with a range of measures of learning ability. High

memory span tasks from other measures of cognitive

abilities such as IQ is that performance is not limited

by the difficulty of the processing activities per se,as

these can be performed accurately in isolation. In-

stead, the crucial limiting factor is the sheer amount

of information that can be processed and stored at

the same time.

Journal of Child Psychology and Psychiatry 47:1 (2006), pp 4–15 doi:10.1111/j.1469-7610.2005.01446.x

 Association for Child Psychology and Psychiatry, 2005.

Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA

1998; Gathercole & Pickering, 2000; Gathercole,

Pickering, Knight, & Stegmann, 2004b; Swanson,

1994), mathematics (e.g., Bull & Scerif, 2001;

Gathercole & Pickering, 2000; Gathercole et al.,

2004b; Mayringer & Wimmer, 2000; Siegel & Ryan,

1989), and language comprehension (e.g., Cain,

Oakhill, & Bryant, 2004; Nation, Adams, Bowyer-

Crain, & Snowling, 1999; Seigneuric, Ehrlich, Oak-

hill, & Yuill, 2000). Complex memory span scores are

strongly associated with teacher ratings of children’s

memory are found in children with learning diffi-

culties in literacy and mathematics (e.g., Gathercole

& Pickering, 2001; Geary, Hoard, Byrd-Craven, &

DeSoto, 2004; Siegel & Ryan, 1989; Swanson, 1993)

that are extremely rare in samples of children with-

out learning difficulties (Pickering & Gathercole,

2004). Our own recent findings indicate that im-

pairments of working memory are more character-

istic of children whose learning difficulties extend

across the domains of literacy and mathematics than

is not mediated by either verbal abilities or IQ

(Gathercole et al., 2004a; see also, Cain et al., 2004).

Although most research has focused on verbal

working memory, recent evidence indicates that

there is a separate working memory capacity for

visuo-spatial material that is also important for

learning. In a study of 11- and 14-year-old children,

we found that both verbal and visuo-spatial working

memory scores each had strong and distinct links

with children’s attainments in national curriculum

2004; Case et al., 1982; Cowan, 2001; Engle, Kane,

& Tuholski, 1999; Just & Carpenter, 1992). One

common feature shared by the various models is that

working memory is supported by a limited capacity

resource that some theorists link with attention (e.g.,

Barrouillet et al., 2004; Cowan, 2001). Working

memory is related to, but distinguishable from,

short-term memory. Whereas working memory in-

volves both the storage and processing of informa-

tion, short-term memory systems are specialised

There is, however, a strong and highly specific link

between verbal short-term memory and the learning

of the sound patterns of new words in both the native

language over the early childhood year, and in sec-

ond language learning at all ages (e.g., Gathercole,

Hitch, Service, & Martin, 1997; Service & Craik,

1993; Service & Kohonen, 1995). Children with poor

verbal memory skills have specific impairments in

the process of learning the phonological structures of

new vocabulary items, and so acquire new vocabu-

Although the mechanisms underlying the link

between impairments of working memory and

learning difficulties are not as yet fully understood,

recent evidence suggests that it arises from working

memory constraints operating at the point of learn-

ing. In an observational study of young children with

marked impairments of working memory, it was

found that they made frequent errors in activities

with heavy working memory demands such as

remembering complex instructions, concurrent

gress, the child has to succeed in many different

structured learning activities designed to accumu-

late across time the body of knowledge and skills

that they need in areas of the curriculum such as

literacy and mathematics. If he or she frequently

fails in individual learning situations due to an

Assessment Criterion Assessment Criterion

Inclusionary:

Verbal STM Forwards

digit span

>1SD below mean At least one further measure:

e.g., word recall, nonword recall

>1SD below mean

Verbal WM Backwards

digit span

>1SD below mean At least one further measure:

e.g., listening span, counting span

>1SD below mean

be used for initial screening purposes. Forwards di-

git span involves the auditory presentation of ran-

dom digit sequences for immediate spoken recall,

and yields a reasonable initial assessment of verbal

short-term memory. However, as the processing load

of this test is minimal, it does not tap working

memory. The second memory test, backwards digit

span, does provide a useful measure of working

memory capacity. The requirement to recall the dig-

its in reverse sequence imposes a substantial pro-

yield a single memory score. Because the two

measures assess separate memory systems as out-

lined in the previous section, these scores do not

provide an adequate basis for distinguishing be-

tween impairments of verbal short-term memory and

verbal working memory if they fall in the ‘impaired’

range. However, tables in the manuals of some

standardised ability tests can be used to calculate

whether there are significant differences between

forwards and backwards digit span. The finding of a

short-term and working memory respectively, these

measures need to be supplemented by further

assessment of the relevant memory system before a

diagnosis of impairment is made. The most signific-

ant shortcoming of both measures is that they em-

ploy numerical stimuli, which can provide

misleading results for children with number-based

processing problems; they may perform poorly on

the task as a consequence of their lack of facility with

the stimuli rather than as a result of memory prob-

on either of the two digit span measures. One

measure that is strongly associated with verbal

short-term memory deficits is nonword repetition, in

which the child attempts to repeat single multi-syl-

labic nonwords. Suitable standardised measures

include the Children’s Test of Nonword Repetition

(Gathercole & Baddeley, 1996) and Nonword

Repetition Task (Campbell, Dollaghan, Needleman,

& Janosky, 1997).

Multiple assessments of both verbal short-term

assessments that correspond to systematic errors in

phonological production should be disregarded.

Hearing impairments may also contribute to low

test scores on measures of verbal short-term and

hearing impairments will depend on the severity and

frequency range of the impairments as well as the

specific acoustic contrasts sampled in the stimuli, it

is difficult to predict the consequences of the

identify whether hearing problems are a possible

source of low memory test scores. For individuals

with a confirmed hearing impairment, non-auditory

methods of assessing memory, such as visual pre-

sentation of memory items in either printed word

form or pictorial form in cases of assessments of the

phonological loop, can be used. However, no such

tests are as yet standardised for use with children. A

further problem is that because non-auditory pre-

sentation formats such as these require subvocal

working memory is often considered as a higher

executive function associated with the engagement

of the frontal lobes, impairments of working memory

within both developmental and adult neuropsycho-

logy have been found to be largely independent of

other classic executive functions such as shifting,

planning, inhibition, and verbal fluency (e.g., Penn-

ington & Ozonoff, 1996). A child with an impairment

of working memory would therefore not necessarily

be expected to exhibit deficits in these other execut-

Working memory and neurodevelopmental

disorders

In this section, memory profiles characteristic of four

neurodevelopmental disorders are considered. In

each case, the disorders are associated with

impairments of either short-term or working mem-

culties with language. Groups with Down syndrome

have been consistently found to be impaired on

measures of verbal short-term memory relative to

control groups composed either of individuals with

moderate learning difficulties of mixed aetiology, or

of younger typically developing children matched for

mental age (e.g., Broadley, MacDonald, & Buckley,

1995; Jarrold, Baddeley, & Hewes, 1999; Laws,

1998; Mackenzie & Hulme, 1987). Their visuo-spa-

tial short-term memory function, on the other hand,

actively maintaining phonological representations in

short-term memory and preventing them from rapid

decay. Consistent with this view, groups with Down

syndrome have been shown to benefit from training

of an overt cumulative rehearsal strategy, in which

participants are required to rehearse aloud increas-

ing amounts of material in the course of a short-term

memory task (Broadley & MacDonald, 1993; Com-

blain, 1996; Laws, McDonald, & Buckley, 1996).

Another possibility is that the low verbal memory

Because intact phonological loop function is im-

portant for long-term learning the sound structures

of new words (Baddeley et al., 1998), it is likely that

individuals with Down syndrome who have deficits of

verbal short-term memory will be slower to acquire

new vocabulary than would be expected on the basis

of their general cognitive abilities. Additional focused

support for vocabulary learning for such individuals

may therefore be of particular value in enhancing

their language development.

struction, drawing and visual processing (e.g., Bel-

lugi, Wang, & Jernigan, 1994; Farran, Jarrold, &

Gathercole, 2001). In marked contrast, expressive

language abilities in WS are much stronger, and can

include excellent pragmatic use of language and good

vocabulary knowledge (e.g., Bellugi et al., 1994).

Williams syndrome is associated with a sub-

stantial deficit of visuo-spatial short-term memory

that mirrors the more general impairment of spatial

cognition. Individuals with William syndrome typic-

term memory supports long-term learning within

this modality, particularly in learning new spatial

routes (Hanley, Young, & Pearson, 1991). The very

poor visuo-spatial memory skills of individuals with

Williams syndrome may therefore be a contributor to

Specific Language Impairment (SLI) is a relatively

common developmental language disorder, with an

estimated incidence of around 7% of the population

(Tomblin et al., 1997). It is characterised by delayed

SLI is associated with impairments of both verbal

short-term memory and working memory. A hall-

mark of SLI is that affected children have severe

deficits in one measure associated with verbal short-

term memory that exceed even the criterial language

impairments – nonword repetition (Dollaghan, &

Campbell, 1998; Edwards & Lahey, 1998; Ellis

Weismer et al., 2000; Gathercole & Baddeley, 1990;

Montgomery, 1995). This nonword repetition deficit

is highly heritable (Bishop et al., 1996, 1999; SLI

children with SLI have deficits of output phonology

and articulation, and some researchers believe that

they also have subtle auditory processing problems

(Tallal, Merzenich, Miller, & Jenkins, 1998). In such

cases, further testing following the procedures out-

lined in Table 1 is recommended.

It is now evident that children with SLI also have

impairments of verbal working memory. They per-

form very poorly on measures of verbal complex

memory (Ellis Weismer, Evans, & Hesketh, 1999;

exhibiting hyperactive behaviours. An unresolved

issue relates to children with inattentive profiles and

whether the impairment of working memory function

results from a primary deficit in working memory or

in intermittent failures to attend to working memory

tasks. The classroom behaviour that we have ob-

served to characterise children with working memory

deficits – of normal social skills combined with ex-

treme reservation in group activities and frequent

failures to complete learning activities (Gathercole

behaviour, working memory function is not unex-

pectedly poor. It does, however, remain possible that

children with attentional problems that are of an

inattentive nature may have impairments of working

memory. If such impairments are diagnosed, remed-

ial support following the principles outlined in the

section below is recommended.

(Johnston, Johnson & Gray, 1987) and in

individuals with Down syndrome (Broadley & Mac-

Donald, 1993; Comblain, 1996; Laws et al., 1996).

Gains in verbal short-term memory scores are,

however, relatively modest following rehearsal

training, and in some cases have been found not to

persist long after training. For this reason, it may be

useful to complement direct training of verbal short-

term memory with a learning support programme

that targets directly the process of vocabulary

recall and recognition of the novel sound structures.

To ensure effective learning, new words should be re-

presented and memory for them tested at increasing

intervals of days and subsequently weeks, to op-

timise the process of consolidation of learning.

Impairments of working memory result in several

characteristic kinds of classroom failures. The

examples below are taken from our observational

study of three 5- and 6-year-old children with very

fairly lengthy and did not represent a routine class-

room activity. Here are three examples of this kind of

failure.

On one occasion, the teacher gave the following

instruction to David: ‘Put your sheets on the green

table, put your arrow cards in the packet, put your

pencil away and come and sit on the carpet.’ David

failed to put his sheet on the green table. Teacher

asked David if he could remember where he was

supposed to put it; he couldn’t, and needed

Finally, Philip was asked to go back and put an n

in the word bean. He went back and asked the

classroom assistant what he had been asked to do.

examples of the child failing to keep track of position

in complex tasks such as writing were observed.

David’s teacher decided that the children should

write He had 36 barrels of gunpowder. The sentence

was repeated until the children appeared to

David then got stuck after writing several letters of

the word gunpowder, attempted and failed to get the

teacher’s attention to help him, and then forgot that

the word needed completing.

A further example of a place-keeping error was

provided by Philip. The teacher wrote on the board

Monday 11th November and, underneath, The Mar-

ket, which was the title of the piece of work. Philip

lost his place in the laborious attempt to copy the

words down letter by letter, writing moNemarket.It

storage of information. Many of these activities in-

volved counting. Although all three children were

capable of counting accurately in the context of a

simple task, many classroom activities combined

counting with other cognitive processes. One fre-

quent activity in literacy sessions involved counting

the numbers of words in a sentence, often prior to

writing the sentence down. Joshua was unable to

recall the sentence, isolate each word and count it

without assistance from the teacher. A group activity

concurrent and fairly demanding processing activity.

Further failures were observed in activities that

involved the detection of target items in spoken or

written text. These tasks imposed significant pro-

cessing demands (analysis and comprehension of

spoken language, or text reading) in conjunction

with the storage of multiple items. For example, the

children in Joshua’s class were asked to identify the

rhyming words in a text read aloud by the teacher.

They had to wait until all four lines had been read

There are two principal approaches to alleviating the

learning difficulties that result from impairments of

working memory. The first involves attempting to

remediate working memory directly. As yet there is

no accepted method of doing this, although some

studies have boosted children’s performance on

working memory tasks by training the children in the

use of rehearsal (Turley-Ames & Whitfield, 2003),

using the approach described in the previous sec-

tion. What is not known is the extent to which chil-

An alternative approach to enhancing learning in

children with impairments of working memory, and

the one recommended here, is to minimise the

memory-related failures in classroom-based learn-

ing activities frequently experienced by such chil-

dren, via effective management of working memory

loads (Gathercole & Alloway, 2004). We are cur-

rently planning a controlled intervention study that

adopts this approach, adopting the principles out-

lined below and summarised in Table 2. Further

memory simply forgetting what they had to do next,

leading to failure to complete many learning activ-

ities (Gathercole et al., in press). Children’s memory

Problem Solutions

Child forgets the task • Give brief and simple instructions, broken down into separate steps if

task is very complex.

• Check the child can remember the instructions. Repeat instructions if

necessary.

Child cannot meet combined processing

• For activities involving sentences, reduce sentence length, reduce syntactic

Practitioner Review: Memory deficits

of instructions. For tasks that take place over an

extended period of time, reminding the child of cru-

cial information for that particular phase of the task

rather than repetition of the original instruction is

likely to be most useful. Finally, one of the best ways

to ensure that the child has not forgotten crucial

information is to ask them to repeat it back.

Second, in activities that require the child to both

process and store information, working memory

demands and hence task failures will be reduced if

variety of ways, such as simplifying the vocabulary,

and using common rather than more unusual

words. In addition, the syntax of the sentence can be

simplified, by encouraging the child to use simple

structures such as active subject–verb–object con-

structions rather than sentences with a complex

clausal structure. The sentences can also be re-

duced in length. A child with poor working memory

skills working with short sentences, relatively unfa-

miliar words and easy syntactic forms is much more

providing memory support. External memory aids

such as useful spellings displayed on the teacher’s

board or the classroom walls and number lines are

widely used in classrooms. In our observational

study, however, we found that children with poor

working memory function often chose not to use

such devices, but gravitated instead towards lower-

level strategies with lower processing requirements

and reduced general efficiency. In order to encourage

children’s use of memory aids, it may be necessary to

losing their place in writing activities. Reducing the

processing load and opportunity for error in spelling

individual words will increase the child’s success in

completing the sentence as a whole. However, read-

ing off information from spellings of key words on the

teacher’s board was itself observed to be a source of

error in low memory children in our study, with

children commonly losing their place within the

word. Making available spellings of key words on the

child’s own desk rather than a distant class board

Impairments of working memory and of verbal short-

term memory are associated with a variety of neuro-

developmental disorders. In order to minimise the

adverse consequences for learning and educational

progress that result from these impairments, early

diagnosis followed by remedial support that targets

relevant domains of learning is strongly recommended.

Acknowledgments