This article was downloaded by: [98.82.76.88] On: 26 February 2013, At: 09:10 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Cognitive Psychology Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/pecp21

Working memory across the lifespan: A cross- sectional approach Tracy Packiam Alloway a & Ross G. Alloway b a University of North Florida, USA b University of Edinburgh, UK Version of record first published: 20 Feb 2013.

To cite this article: Tracy Packiam Alloway & Ross G. Alloway (2013): Working memory across the lifespan: A cross- sectional approach, Journal of Cognitive Psychology, 25:1, 84-93

To link to this article: http://dx.doi.org/10.1080/20445911.2012.748027

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

http://www.tandfonline.com/loi/pecp21

http://dx.doi.org/10.1080/20445911.2012.748027

http://www.tandfonline.com/page/terms-and-conditions

Working memory across the lifespan: A cross-sectional approach

Tracy Packiam Alloway1 and Ross G. Alloway2

1University of North Florida, USA 2University of Edinburgh, UK

The aim of the present study was to extend previous lifespan research to a wide age range (5 to 80 year olds) and investigate any potential differences in the development and decline of working memory functions. To that end, measures of both verbal and visuo-spatial working memory were included in a cross-sectional study. The findings indicated that there is considerable growth in childhood!on average 23 standard points; with performance peaking in 30-year olds. There was relatively little change in working memory performance in older adults, with 70 to 80 year olds performing at comparable levels to teenagers (13!19 year olds) in verbal working memory tests. Confirmatory factor analyses suggest that working memory skills across the lifespan are driven by domain differences (i.e., verbal or visuo-spatial stimuli), rather than functional differences (maintenance and manipulation of information).

Keywords: Working Memory; Lifespan; Development.

Working Memory capacity as a function of in- dividual differences throughout the lifespan has been found to mediate outcomes in reasoning tasks (Salthouse, 1993), IQ scores (Conway et al., 2002), academic attainment (Alloway & Alloway, 2010), reading (Siegel, 1994), and even building Lego blocks (Morrell and Park, 1993). There are several different theoretical models of working memory, but a common element is that it com- prises a higher-order skill related to the ability to allocate attentional resources despite distraction or interference (e.g., Baddeley, 1996; Cowan, 2006; Engle, Tuholski, Laughlin, & Conway, 1999). A major issue of debate is whether the verbal and visuo-spatial subsystems of working memory can be differentiated across the lifespan. One sugges- tion is that working memory is a domain-general component that manages both verbal and visuo- spatial information (Baddeley, 1996; see Engle, Kane, & Tulhoski, 1999, for a review), while another view that is that working memory resources are

separated into verbal and visuo-spatial constructs (Shah & Miyake, 1996).

One way to distinguish between these two views is to explore any possible changes from childhood to older adults. This issue has been studied in developmental populations using con- firmatory factor analyses to test several theore- tical models (Alloway, Gathercole, & Pickering, 2006; also Bayliss, Jarrold, Gunn, & Baddeley, 2003). The best fitting model for 4- to 12-year-olds was one with a common factor that captured the shared variance for manipulating verbal and visuo-spatial information. However, a different pattern seems to emerge in the adolescent years, where working memory capacity appears to be supported by 2 separate pools of domain-specific resources for verbal and visuo-spatial information (Jarvis & Gathercole, 2003; also Miyake, Friedman, Rettinger, Shah, & Hegarty, 2001).

Research on this issue in adults has focused on whether verbal and visuo-spatial working memory

Correspondence should be addressed to: Dr Tracy Packiam Alloway, Department of Psychology, University of North Florida, 1 UNF Drive, Jacksonville, FL. E-mail: [email protected]

Journal of Cognitive Psychology, 2013 Vol. 25, No. 1, 84!93, http://dx.doi.org/10.1080/20445911.2012.748027

# 2013 Taylor & Francis

D ow

nl oa

de d

by [9

8. 82

.7 6.

88 ] a

t 0 9:

10 2

6 Fe

br ua

ry 2

01 3

http://dx.doi.org/10.1080/20445911.2012.748027

skills decline at the same rate. Park et al. (2002) found that the rate of decline is equivalent in young and older adults (see also Salthouse, Kausler, & Saults, 1988). However, Jenkins, Myerson, Joerding, and Hale (2000) found a different pattern. Reasons for why these different patterns exist appear to lie in the methodology used to calculate the scores: Park et al. trans- formed the raw scores into z-scores, while Jenkins et al. equated the working memory demands and then investigated the absolute differences in performance (see Park & Payer, 2006 for further discussion).

There are a few lifespan studies looking specifically at working memory. For example Park et al. (2002) examined working memory in adults from 20 to 80; Swanson (1999) also looked at Working Memory across the lifespan, spanning ages 6 to 76 years. However, the majority of such research has utilized verbal working memory measures and did not include visuo-spatial work- ing memory measures. There is evidence to suggest that visuo-spatial memory may be more age- sensitive than verbal memory (e.g., Jenkins et al., 2000; also Johnson, Logie, & Brockmole, 2010) and the exclusion of such measures may mean that we have not had a complete picture of Working Memory across the lifespan. Thus, the aim of the present study was to extend this research in a cross-sectional study with individuals aged from 5 to 80 years old. The large age-range allowed us to investigate whether verbal and visuo-spatial working memory develop and decline at similar rates across the lifespan.

In the present study, working memory ability was measured by complex span tasks that required the individual to engage in the main- tenance and manipulation of information. Ac- cording to domain-general accounts of working memory, the manipulation aspect of the task is controlled by a centralized component (i.e., the central executive or controlled attention) that requires a sustained focus of attention. In contrast, the maintenance aspect is supported by a domain-specific component (i.e., verbal or visuo-spatial stimuli). From a domain-specific perspective, the theoretical distinction lies in how working memory is used to manage verbal and visuo-spatial stimuli, and not between the maintenance and manipulations components of a working memory task. Speed of processing is commonly associated with working memory changes, both in development (Towse, Hitch, & Hutton, 2002) and in aging (Park et al., 2002;

Salthouse, 1994). In an attempt to circumvent the potential negative effects of time pressure on performance, participants in the present study were not told to respond as fast as they could, nor were they told that there was a time limit in their response.

METHOD

Participants

There were 1070 participants aged between 5 and 80 years old from a range of demographic back- grounds. The age bands from 5!19 were created based on previous research establishing theoreti- cal patterns in Working Memory growth (e.g., Alloway et al., 2006). The age bands from 20!80 were grouped by decade following research by Park et al. (2002). All participants were English- as-a-first-language speakers. The data collected for the norms include an equivalent proportion of males and females, as well as reflecting a range of ethnic diversity in the UK, including those from Pakistan, Bangladesh, China, Africa and the Caribbean. Previous research on Working Memory scores in development populations has reported that neither sex nor ethnicity leads to significant differences in test performance (Alloway et al., 2006). None of the participants reported having any clinical diagnoses of physical, sensory, or behavioral impairment; and none were receiving any medication.

Recruitment

The under-20-year-olds were invited to participate from a range of schools, colleges and universities selected to provide a nationally representative demographic sample. The schools for the 5 to 12- year-olds were selected on the basis of the national average of performance on national assessments in English, mathematics, and science that students sit in the final year of elementary education at the age of 10 or 11. Schools in the UK are ranked on the basis of a combined or ‘aggregate’ score achieved in the 3 tests!the maximum possible being 300 (published by the Department for Education and Skills, 2006). Schools selected for the normative sample represent a range of low, average, and high performance in the combined score of the national test results. Schools were located in both urban and rural settings.

WORKING MEMORYACROSS THE LIFESPAN 85

D ow

nl oa

de d

by [9

8. 82

.7 6.

88 ] a

t 0 9:

10 2

6 Fe

br ua

ry 2

01 3

Participating high schools and colleges were selected on the basis of the national average of performance on the General Certificate of Sec- ondary Education (GCSE), the British qualifica- tions taken by high school students at the age of 15 or 16. The national average of the proportion of students attaining 5 or more good GCSEs and equivalents at the end of Key Stage 4 including English and maths is 46% in the year 2006 (published by the Department for Education and Skills, 2006). Selected schools for the stan- dardization data represent a range of low (30%), average (45%), and high (60%) proportions of pupils’ scores. Schools were located in both urban and rural settings.

The selected universities reflect both the de- mographic and academic range in tertiary educa- tion, and are based on the 2006 Sunday Times league table of UK universities and higher educa- tion colleges. This was compiled using data from the Higher Education Statistics Agency (HESA), the Quality Assurance Agency for Higher Educa- tion, the national funding councils, head teachers, peers and the institutions themselves. Universities are ranked according to marks scored in 9 key performance areas, including areas such as teach- ing excellence, student satisfaction and research quality. The selected universities for the standar- dization data represent top, average, and low ranking tertiary institutions.

The adult participants were recruited from across the UK from a range of different educa- tional and demographic backgrounds. They were also from a range of different ethnic backgrounds. These participants were recruited through uni- versity volunteer databases based on community members, as well as through various advertising channels, such as news and magazine links.

Materials

Working memory was measured using an online version of a standardized memory assessment, the Automated Working Memory Assessment (AWMA; Alloway, 2007a). All test trials began with 2 items, and increased by 1 item in each block, until the participant was unable to recall 3 correct trials at a particular block. There were 4 trials in each block and the number of correct trials was scored for each participant. The move forward and discontinue rules, as well as the scoring, were automated by the program.

There were 2 verbal and 1 visuo-spatial work- ing memory tests. In the Backward Digit Recall test, the individual recalled a sequence of spoken digits in the reverse order. In Processing Letter Recall, the participant views a letter in red that stays on the computer screen for 1 second. Another letter in black immediately follows this on the screen. Participants verify whether the black letter was the same as the red letter by clicking on a box marked either ‘Yes’ or ‘No’ on the screen (manipulation). They then click on the red letters they saw in the correct sequence (maintenance). The number of correct responses for both the maintenance and manipulation com- ponents was scored.

Visual working memory was tested using a shape recall test. The participant views a colored shape in a 4x4 grid. The shape and grid disappear and another shape appears in the center of the computer screen. They verify whether those 2 shapes were the same color and shape by clicking on a box marked either ‘Yes’ or ‘No’ on the screen (manipulation). Then they have to remem- ber the location of the first shape on the 4"4 grid in the correct sequence (maintenance). The sti- muli in all working memory tests were randomized so no stimulus sequence was repeated to avoid potential practice effects. The number of correct responses for both the maintenance and manip- ulation components was scored.

Test-reliability of the AWMA was established in a random selection of the normative sample tested on two separate occasions, four weeks apart. The reliability coefficient for the verbal working memory tests was .86 and for the visuo-spatial working memory test, it was .84 (Alloway, 2007a).

Procedure

Participants completed the test online via a web link that was provided to them. The testing was completed in a single session, lasting up to 30 minutes depending on how far the volunteers progressed through the levels. No compensation was offered in exchange for participation.

RESULTS

Data screening

There was no missing data, as the test did not record partial completions. Only fully completed

86 ALLOWAYAND ALLOWAY

D ow

nl oa

de d

by [9

8. 82

.7 6.

88 ] a

t 0 9:

10 2

6 Fe

br ua

ry 2

01 3

tests were recorded. The data were screened for univariate and multivariate outliers. Univariate outliers on each of the working memory tests were defined as scores more than 3.5 standard deviations above or below the mean. 24 values out of the 5350 (B1%) in the dataset met this criterion and were replaced with a value corre- sponding to plus or minus 3.5 standard deviations as appropriate. Two multivariate outliers with a Mahanalobis d2 score (pB.001) were eliminated (a 7-year-old and an 11-year-old). The final data set for subsequent analyses consisted of 1068 participants.

Descriptive statistics

The mean standard scores were calculated for the sample as a whole for each measure as a function of age group (see Table 1). Improve- ments in performance were observed across the age groups. A MANOVA was performed on the standard scores of all measures, as a function of age groups. The probability value associated with Hotelling’s T-test is reported. There was a significant effect of age (F#62.05, pB.001, h2p#.35). Post-hoc pairwise comparisons found significant differences between some of the groups (pB.001, Bonferroni adjustment for multiple com- parisons; see Table 1).

Confirmatory factor analyses

The degree to which the data fitted alternative models of working memory was tested formally using confirmatory factor analysis (Bentler, 2001; Bentler & Wu, 1995). This method provides a means of testing the adequacy of competing theoretical accounts of the relationships between measures, with each model specified in terms of paths between observed variables and latent constructs, and between constructs. A commonly used index of goodness of fit for each model is the x2 statistic, which compares the degree to which the predicted covariances in the model differ from the observed covariances. A good fit is determined by small and nonsignificant x2 values. Because this statistic is sensitive to variances in sample sizes, with very large samples as in the present study even the best-fitting models fre- quently yield significant x2 values (Kline, 1998).

Model adequacy was therefore evaluated using additional global fit indices that are more

T A B L E 1

M e a n st a n d a rd

sc o re s fo r th e m e m o ry

te st s a s a fu n ct io n o f a g e g ro u p

N u m b er s

L et te rs

M ai n te n an

ce L et te r M an

ip u la ti o n

S h ap

es M ai n te n an

ce S h ap

es M an

ip u la ti o n

A ge

G ro u p

M ea n

S D

P o st h o c

M ea n

S D

P o st h o c

M ea n

S D

P o st h o c

M ea n

S D

P o st -h o c

M ea n

S D

P o st h o c

5 !8

(n #

15 5)

82 .1 1

9. 31

2 !1

0 78

.9 6

12 .5 8

2 !1

0 81

.4 0

8. 53

2 !1

0 81

.1 3

11 .6 7

2 !1

0 83

.2 0

9. 80

2 !1

0 9 !1

0 (n

# 15

2) 89

.4 9

10 .4 5

1; 4 !1

0 89

.8 1

14 .3 4

1; 3 !1

0 89

.6 9

12 .0 1

1; 4 !1

0 91

.5 9

14 .6 7

1; 3 !9

92 .3 8

14 .2 7

1; 3 !9

11 !1

2 (n

# 89

) 92

.5 0

9. 59

1; 4 !1

0 95

.6 9

12 .8 9

1 !2

; 4 !1

0 93

.5 9

11 .4 3

1; 4 !1

0 99

.7 8

13 .0 9

1 !2

; 4 !8

99 .6 4

13 .6 8

1 !2

; 5 !6

13 !1

9 (n

# 70

) 10

2. 14

12 .6 3

1 !3

; 5 !9

10 2. 01

11 .8 0

1 !3

; 6 !9

10 0. 19

12 .6 1

1 !3

; 5 !9

10 6. 72

13 .0 3

1 !3

; 10

10 5. 69

12 .8 9

1 !2

; 6

20 !2

9 (n

# 51

) 10

9. 18

14 .9 2

1 !4

10 7. 28

10 .8 3

1 !3

10 7. 60

13 .3 4

1 !4

11 0. 04

15 .3 6

1 !3

; 10

11 1. 61

15 .6 9

1 !3

; 8 !1

0 30

!3 9 (n

# 55

) 11

4. 26

8. 56

1 !4

; 7 !1

0 11

1. 77

3. 81

1 !4

; 10

11 3. 90

7. 67

1 !4

; 7, 9 !1

0 11

2. 73

9. 35

1 !3

; 8 !1

0 11

3. 70

12 .2 6

1 !4

; 8 !1

0 40

!4 9 (n

# 10

5) 10

7. 81

11 .7 5

1 !4

; 6

10 8. 30

7. 73

1 !5

10 7. 78

11 .3 2

1 !4

; 6

10 7. 55

8. 77

1 !3

; 10

10 7. 47

11 .0 0

1 !3

; 10

50 !5

9 (n

# 19

2) 10

7. 60

11 .2 5

1 !4

; 6

10 9. 11

6. 03

1 !4

10 8. 70

9. 64

1 !4

10 5. 48

10 .3 8

1 !3

; 6

10 4. 58

12 .0 0

1 !2

; 5 !6

60 !6

9 (n

# 16

5) 10

7. 47

11 .2 5

1 !4

; 6

10 7. 58

7. 41

1 !4

10 7. 58

11 .1 0

1 !4

; 6

10 4. 43

9. 36

1 !2

10 2. 76

10 .5 9

1 !2

; 5 !6

70 !8

0 (n

# 46

) 10

2. 79

10 .1 7

1 !3

10 3. 96

9. 10

1 !3

10 2. 71

11 .5 4

1 !3

; 6

98 .5 9

9. 63

1; 4 !7

97 .6 8

12 .3 4

1; 5 !7

N o te : P o st -h o c d at a in d ic at es

si gn

if ic an

t d if fe re n ce s b et w ee

n th e ta rg et

ag e an

d co rr es p o n d in g ag

e gr o u p s la b el le d in

C o lu m n 1.

WORKING MEMORYACROSS THE LIFESPAN 87

D ow

nl oa

de d

by [9

8. 82

.7 6.

88 ] a

t 0 9:

10 2

6 Fe

br ua

ry 2

01 3

sensitive to model specification than to sample size (Jaccard &Wan, 1996; Kline, 1998). Fit indices such as the Comparative Fit Index (CFI; Bentler, 1990), the Incremental Fit Index (IFI; Bollen, 1989), and the Normed Fit Index (NFI; Bentler & Bonett, 1980) provide a further measure of fit computed by comparing the hypothesized model against a null model in which the relations between the latent variables are not specified and conse- quently are set at 0. Fit indices with values equal to or higher than .90 demonstrate a marginal fit, and values above .95 indicate a good fit. Further assessment of the extent to which the specified model approximates to the true model is the root mean square error of approximation (RMSEA). A RMSEA value of .08 or lower is acceptable, and a value below .05 indicates a good fit (see McDonald & Ho, 2002).

In the series of models tested, paths between latent constructs were left free to co-vary (repre- sented diagrammatically as bi-directional links) in the absence of justifiable assumptions concerning direction of causality. Such models are known as measurement models. In each case, the level of significance of the path weights between each observed variable and its associated factor, and the correlations between all pairs of factors, was set at an alpha level of .05. The statistics and fit indices generated by each of the measurement models are summarized in Table 2. The follow- ing models were tested using EQS Structural Equation Modeling Software (Bentler, 1995).

We tested the following models on the whole sample. However, based on the distribution of test scores as a function of age group, one might expect age-related differences in the model fit. In order to account for this potential source of measurement variance of working memory, we calculated the unstandardized residuals for each test score to capture some of the influence of age. The input for each theoretical model was the correlation matrix for the 5 working memory test scores.

Model 1 was a one-factor model that did not distinguish between maintenance or manipulation components, or between verbal or visuo-spatial stimuli (see Figure 1). This one-factor model did not provide a good fit to the data; the x2 value was highly significant, and the fit indices were less than .90.

Model 2 discriminated between the mainte- nance and manipulation components. The first factor was associated with the recall scores from both the verbal and visuo-spatial measures (Processing Letter Recall and Shape Recall), while Factor 2 was associated with the manipula- tion scores for both these tests. As Backward Digit Recall involves both maintenance and mani- pulation of information, this observed variable loads on both Factors 1 and 2. This model was based on the domain-general view of working memory, with functional distinctions. The model is summarized in Figure 2, and the fit statistics are shown in Table 2. This model also did not provide a satisfactory fit of the data: the x2 value (pB.001) was highly significant, all fit indices were less than .90.

In Model 3, the first factor was associated with all verbal working memory measures including the manipulation component, and the second factor was associated with the visuo-spatial work- ing memory maintenance and manipulation scores. This model followed the Shah and Miyake (1996) view of domain-distinct components in working memory. The model is summarized in Figure 3, and the fit statistics summarized in Table 2. This model provided a good fit of the data. The x2

value is lower than the other models (although still highly significant, pB.001), and all fit indices are above .90. It is noteworthy that this model provides the best fit for data. These values suggest that verbal and visuo-spatial working memory skills are moderately related but represent dissoci- able constructs.

TABLE 2 Goodness of fit statistics for the different measurement models for each age band

Model Sample x2 df p CFI IFI NFI RMSEA

1 Unadjusted 2024.31 5 B .001 .69 .69 .69 .62 Age-adjusted 1093.25 5 B .001 .75 .75 .75 .45

2 Unadjusted 1395.23 3 B .001 .79 .79 .79 .66 Age-adjusted 1017.94 3 B .001 .77 .77 .77 .56

3 Unadjusted 130.37 4 B .001 .98 .98 .98 .17 Age-adjusted 84.5 4 B .001 .98 .98 .98 .14

88 ALLOWAYAND ALLOWAY

D ow

nl oa

de d

by [9

8. 82

.7 6.

88 ] a

t 0 9:

10 2

6 Fe

br ua

ry 2

01 3

DISCUSSION

Several patterns emerged in the present study. First, working memory performance seems to experience tremendous change during develop- ment: between 5 to 19 years of age, there was a mean increase of 23 standard points. Contrast this growth to two other 20-year-periods: between 20 to 39 years of age (a mean increase of 4 standard points); and between 50 to 69 years of age (a mean decline of 1 standard point). Another key pattern is that working memory scores were highest in 30-year olds. This finding extends previous re- search indicating that working memory perfor- mance reached maximum capacity in the teenage years (Alloway et al. 2006). A surprising finding is

that there was little decline in working memory capacity in the older adults: between the 50- to 80-year-olds, there was a mean drop of 6 standard points. People in their 60s performed at a similar level to those in their 20s. The difference in working memory scores between age bands in the older adults was markedly less than that of the children.

There are several possible explanations for why the decline in working memory performance in adults is not as steep as the increase during childhood. First, the working memory tasks used in the present study may have benefited from larger knowledge stores in adulthood, such as declarative or procedural knowledge. Although the current stimuli were intended to be relatively

Figure 1. Structural exquation model for one factor model of Working Memory (Model 1).

Figure 2. Structural equation model for two-factor model based on a separable components for maintaining and manipulating information (Model 2).

WORKING MEMORYACROSS THE LIFESPAN 89

D ow

nl oa

de d

by [9

8. 82

.7 6.

88 ] a

t 0 9:

10 2

6 Fe

br ua

ry 2

01 3

abstract (e.g., letters and numbers), future studies could compare working memory performance across the lifespan using stimuli that are more closely linked to long-term knowledge stores, such as sentence comprehension tasks (see Fedorenko, Gibson, & Rodhe, 2006), as well as novel stimuli that would not have an existing representation in long-term memory. This comparison could directly address the issue of whether decline in working memory during adulthood benefits from greater knowledge stores.

Another explanation for this age-related pat- tern in working memory performance could be due to a change in processing speed. It is well established that adults are slower as they age (Deary & Der, 2005; Salthouse, 1993). Indeed, Salthouse (1993) suggested that almost 80% of age-related variance in tasks of fluid cognition can be accounted for by speed. The absence of a timed component in the present study may have provided an unintended ‘boost’ to the working memory scores of the older adults, as they were not pressed to respond in a particular time frame.

A third possibility is based on the link between working memory and related executive function skills, such as shifting between tasks, planning and updating information, and inhibiting prepotent responses (Miyake et al., 2000). These executive function skills experience tremendous growth during childhood (Barkley, 1997) and stabilize during adulthood (Hull, Martin, Beier, Lane, & Hamilton, 2008). To our knowledge, no studies have directly compared the increase and decline of various executive function skills with working memory capacity; it is possible that there are similarities in the growth trajectory that could account for the working memory performance reported here. For example, older children im-

prove in their ability to shift between tasks and plan ahead and thus, demonstrate better performance in working memory tasks compared to their younger peers. Likewise, there is little growth in such skills during adulthood, which may explain the relatively stable pattern of working memory performance reported in this study.

Interestingly, there were marked differences in the nature of working memory decline. Verbal working memory skills remained robust and individuals in their 70s to 80s performed as well as those in their teenage years. In contrast, visuo- spatial working memory skills in the same age group (70!80 years) were at the same level of 9 and 10 year olds. This pattern fits with the notion of a differentiation of visuospatial and verbal cognitive resources as we age. For exam- ple, Jenkins, Myerson, Joerding, and Hale (2000) reported that older adults experienced greater difficulty than younger adults in learning visuo- spatial information compared to verbal material (see also Myerson, Hale, Rhee, & Jenkins, 1999). One possible explanation for this difference in decline in verbal and visuo-spatial working mem- ory performance lies in brain imaging research. There have been reports that there is less deter- ioration of left hemisphere function (associated with verbal skills) compared to right hemisphere function in older adults (Goldstein & Shelly, 1981; Reuter-Lorenz, 2000). Thus, it is possible that the performance difference in verbal and visuo- spatial working memory tests in the present study reflects such changes across the lifespan.

Another issue that was investigated in the present study was whether changes in work- ing memory performance across the lifespan reflect changes in capacity size or in processing resources (Salthouse, 1996). The inclusion of both

Figure 3. Structural equation model based on distinct verbal and visuo-spatial working memory components (Model 3). Note: Factor loadings for the age-adjusted standardized residuals are shown in italics.

90 ALLOWAYAND ALLOWAY

D ow

nl oa

de d

by [9

8. 82

.7 6.

88 ] a

t 0 9:

10 2

6 Fe

br ua

ry 2

01 3

maintenance (capacity) and manipulation (pro- cessing) scores allowed us to address this issue. The data from the confirmatory factor analysis suggest that working memory across the lifespan follows a pattern that is domain-distinct (verbal vs. visuo-spatial), rather than functionally-distinct (maintenance vs. manipulation). The measure- ment model that provided the best fit for the data indicated that we manage verbal and visuo- spatial stimuli with different cognitive resources from childhood to older adulthood. Although the model in the present study that provided the best fit was based on Shah and Miyake’s (1996) view of domain-distinct components in working memory, the data suggest that these two systems are not entirely independent, but share some overlap in function (see Kane et al., 2004).

While one caveat is that the domain distinction was based on a few measures (i.e., the visuo- spatial latent construct was measured by a single task with two scores), this pattern fits well with both behavioral and brain imaging research on working memory. For example, Park et al. (2002) also reported domain-distinct performances in their adult population for working memory tasks. Brain imaging research suggests that there are functional differences in the prefrontal cortex for managing verbal and visuo-spatial information (Fuster, 2008; see also Buschman, Siegel, Roy, & Miller, 2011).

One question is why other studies have re- ported functional differences in memory, rather than domain differences (e.g., Engle, Tuholski, et al., 1999). The explanation may lie in the tasks used to measure memory. The notion of func- tional distinction typically relates to the distinc- tion between the storage component of memory, also known as short-term memory, and a general- ized processing component (working memory). Further, in some working memory tasks the kind of information used in the manipulation aspect is unrelated from that of the maintenance part. For example, solving a mental math problem (manipulation) and remembering a word list interspersed between the math problems (main- tenance). This disparity could yield the reported functional distinction in working memory skills. In contrast, the present study did not include measures of short-term memory that capture only storage components. Also, in the working mem- ory tasks, the manipulation component could be considered as closely linked to the maintenance aspect of the task as the participant had to manipulate information in the context of the

stimulus that was presented as part of the main- tenance phase. For example, they were shown a shape or letter and had to decide whether it was the same stimulus as presented previously.

A strength of the present study is that it is one of the few studies to examine the development and decline of working memory from a large sample with a wide age range. However, there are some limitations. Principally, this was a cross- sectional study and thus, we cannot make claims relating to any longitudinal changes. Another issue was the selection of measures included in the present study. We restricted the measures to working memory measures and did not include measures of general ability, short-term memory, or processing speed. Nonetheless, the wide age range and the inclusion of both maintenance and manipulation scores associated with work- ing memory tasks can complement existing life- span research (e.g., Baltes & Linderberger, 1997; Kaufman, 2001; Lee et al., 2005; Park et al., 2002).

The finding of working memory ability as domain-distinct skills has important implications for a range of fields, including education. In education, there are varying patterns of strengths and weaknesses in the working memory profile of individuals with learning needs. Students with reading and language impairments tend to exhibit verbal working memory deficits, and rela- tive strengths in visuo-spatial working memory (Archibald & Alloway, 2008). Conversely, those with motor dyspraxia and ADHD perform below age-expected levels in visuo-spatial working mem- ory tests (Alloway, 2007b; Holmes, Gathercole, Place, Alloway, & Elliot, 2010). Understanding these domain-distinct working memory profiles can be useful in supporting and training learning (Alloway, 2010; 2012). On the other end of the lifespan spectrum, based on the differentiation of working memory decline, there is scope for using training that is targeted for specific areas of need (Ackerman, Kanfer, & Calderwood, 2010; Jaeggi, Buschkuehl, Jonides, & Perrig, 2008; Nouichi et al., 2012).

Original manuscript received April 2012 Revised manuscript received November 2012

First published online February 2013

REFERENCES

Ackerman, P. L., Kanfer, R., & Calderwood, C. (2010). Use it or lose it? Wii brain exercise practice and

WORKING MEMORYACROSS THE LIFESPAN 91

D ow

nl oa

de d

by [9

8. 82

.7 6.

88 ] a

t 0 9:

10 2

6 Fe

br ua

ry 2

01 3

reading for domain knowledge. Psychology and Aging, 25, 753!766.

Alloway, T. P. (2007a). Automated Working Memory Assessment. Pearson Assessment, London.

Alloway, T. P. (2007b). Working Memory, Reading and Mathematical Skills in Children with Developmen- tal Coordination Disorder. Journal of Experimental Child Psychology, 96, 20!36.

Alloway, T. P. (2010). Improving Working Memory: Supporting Students Learning’. London: Sage Press.

Alloway, T. P. (2012). Can interactive working memory training improving learning? Journal of Interactive Learning Research, 23, 1!11.

Alloway, T. P., & Alloway, R. G. (2010). Investigating the predictive roles of working memory and IQ in academic attainment. Journal of Experimental Child Psychology, 106, 20!29.

Alloway, T. P., Gathercole, S. E., Adams, A. M., Willis, C., Eaglen, R., & Lamont, E. (2005). Working memory and other cognitive skills as predictors of progress towards early learning goals at school entry. British Journal of Developmental Psychology, 23, 417!426.

Alloway, T. P., Gathercole, S. E., & Pickering, S. J. (2006). Verbal and visuo-spatial short-term and working memory in children: Are they separable? Child Development, 77, 1698!1716.

Alloway, T. P., Gathercole, S. E., Willis, C., & Adams, A. M. (2004). A structural analysis of working memory and related cognitive skills in early childhood. ournal of Experimental Child Psychology, 87, 85!106.

Archibald, L. M., & Alloway, T. P. (2008). Comparing Language Profiles: Children with Specific Language Impairment and Developmental Coordination Dis- order. International Journal of Communication and Language Disorders, 43, 165!180.

Baddeley, A. D. (1996). Exploring the central execu- tive. Quarterly Journal of Experimental Psychology, 49A, 5!28.

Baltes, P. B., & Lindenberger, U. (1997). Emergence of a powerful connection between sensory and cogni- tive functions across the adult lifespan: A new window to the study of cognitive aging? Psychology and Aging, 12, 12!21.

Barkley, R. A. (1997). Behavioral inhibition, sustained attention, and executive functions: Constructing a unifying theory of ADHD. Psychological Bulletin, 121, 65!94.

Bayliss, D. M., Jarrold, C., Gunn, D. M., & Baddeley, A. D. (2003). The complexities of complex span: Explaining individual differences in working memory in children and adults. Journal of Experimental Psychology: General, 132, 71!92.

Bentler, P. M. (1990). Comparative fit indexes in structur- al models. Psychological Bulletin, 107, 238!246.

Bentler, P. M. (2001). EQS structural equation model- ling software (version 6.0) [computer software]. USA: mvsoft.

Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, 88, 588!606.

Bentler, P. M., & Chou, C. P. (1987). Practical issues in structural modeling. Sociological Methods & Re- search, 16, 78!117.

Bentler, P. M., & Wu, E. J. C. (1995). EQS for Macintosh User’s Guide. Encino, CA: Multivariate Software, Inc.

Bollen, K. A. (1989). Structural Equations with Latent Variables. New York: John Wiley & Sons, Inc.

Buschman, T. J., Siegel, M., Roy, J. E., & Miller, E. K. (2011). Neural substrates of cognitive capacity limitations. Proceedings of the National Academy of Sciences, 108, 11252!11255.

Conway, A. R. A., Cowan, N., Bunting, M. F., Ther- riault, D. J., & Minkoff, S. R. B. (2002). A latent variable analysis of working memory capacity, short term memory capacity, processing speed, and gen- eral fluid intelligence. Intelligence, 30, 163!183.

Cowan, N. (2006). Within fluid cognition: Fluid proces- sing and fluid storage? Behavioral and Brain Sciences, 29, 129!130.

Deary, I. J., & Der, G. (2005). Reaction time explains IQ’s association with death. Psychological Science, 16, 64!69.

Engle, R. W., Kane, M. J., & Tuholski, S. W. (1999a). Individual differences in working memory capacity and what they tell us about controlled attention, general fluid intelligence, and functions of the prefrontal cortex. In A. Miyake & P. Shah (Eds.), Models of working memory: Mechanisms of active maintenance and executive control (pp. 102!134). New York: Cambridge Univ. Press.

Engle, R. W., Tuholski, S. W., Laughlin, J. E., & Conway, A. R. A. (1999b). Working memory, short-term memory, and general fluid intelligence: A latent variable approach. Journal of Experimental Psychology: General, 125, 309!331.

Fedorenko, E., Gibson, E., & Rohde, D. (2006). The Nature of Working Memory Capacity in Sentence Comprehension: Evidence Against Domain-Specific Working Memory Resources. Journal of Memory and Language, 54, 541!553.

Fuster, J. M. (2008). The Prefrontal Cortex (Fourth Edition). London: Academic Press.

Goldstein, G., & Shelly, C. (1981). Does the right hemisphere age more rapidly than the left? Journal of Clinical Neuropsychology, 3, 65!78.

Holmes, J., Gathercole, S. E., Place, M., Alloway, T. P., & Elliott, J. (2010). An assessment of the diagnostic utility of executive function assessments in the identification of ADHD in children. Child & Adolescent Mental Health, 15, 37!43.

Hull, R., Martin, R. C., Beier, M. E., Lane, D., & Hamilton, A. C. (2008). Executive function in older adults: A structural equation modeling approach. Neuropsychology, 22, 508!522.

Jaccard, J., & Wan, C. (1996). LISREL analyses of interaction effects in multiple regression. Newbury Park, CA: Sage.

Jaeggi, S. M., Buschkuehl, M., Jonides, J., & Perrig, W. J. (2008). Improving fluid intelligence with training on working memory. Proceedings of the National Academy of Science, 105, 6829!6833.

Jarvis, H. L., & Gathercole, S. E. (2003). Verbal and non-verbal working memory and achievements on national curriculum tests at 11 and 14 years of age. Educational and Child Psychology, 20, 123!140.

92 ALLOWAYAND ALLOWAY

D ow

nl oa

de d

by [9

8. 82

.7 6.

88 ] a

t 0 9:

10 2

6 Fe

br ua

ry 2

01 3

Jenkins, L., Myerson, J., Joerding, J. A., & Hale, S. (2000). Converging evidence that visuospatial cogni- tion is more age-sensitive than verbal cognition. Psychology and Aging, 15, 157!175.

Johnson, W., Logie, R. H., & Brockmole, J. R. (2010). Working memory tasks differ in factor structure across age cohorts: Implications for dedifferentia- tion. Intelligence, 38, 513!528.

Kane, M. J., Hambrick, D. Z., Tuholski, S. W., Wilhelm, O., Payne, T., & Engle, R. W. (2004). The generality of working memory capacity: A latent-variable approach to verbal and visuospatial memory span and reasoning. Journal of Experimental Psychology: General, 133, 189!217.

Kaufman, A. S. (2001). WAIS-III IQs, Horn’ s theory, and generational changes from young adulthood to old age. Intelligence, 29, 131!167.

Kline, R. B. (1998). Principles and practice of structural equation modelling. New York: Guilford.

Lee, J.-Y., et al., (2005). Intellect declines in healthy elderly subjects and cerebellum. Psychiatry & Clinical NeuroScience, 59, 45!51.

McDonald, R. P., & Ho, R. M. (2002). Principles and practice in reporting structural equation analyses. Psychological Methods, 7, 64!82.

Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. (2000). The unity and diversity of executive functions and their con- tributions to complex ‘‘frontal lobe’’ tasks: A latent variable analysis. Cognitive Psychology, 41, 49!100.

Miyake, A., Friedman, N., Rettinger, D. A., Shah, P., & Hegarty, M. (2001). How are visuospatial working memory, executive functioning, and spatial abili- ties related? A latent-variable analysis. Journal of Experimental Psychology: General, 130, 621!640.

Morrell, R. W., & Park, D. C. (1993). The effects of age, illustrations, and task variables on the performance of procedural assembly tasks. Psychology and aging, 8, 389!399.

Myerson, J., Hale, S., Rhee, S. H., & Jenkins, L. (1999). Selective interference with verbal and spatial work- ing memory in young and older adults. Journal of Gerontology, 54B, 161!164.

Nouichi, R., Taki, Y., Takeuchi, H., et al. (2012). Brain Training Game Improves Executive Functions and Processing Speed in the Elderly: A Randomized Controlled Trial. PLoS ONE, 71: e29676.

Park, D. C., Lautenschlager, G., Hedden, T., Davidson, N. S., Smith, A. D., & Smith, P. K. (2002). Models of visuospatial and verbal memory across the adult life span. Psychology and aging, 17, 299!320.

Park, D. C., & Payer, D. (2006). Working memory across the adult lifespan. In E. Bialystock & F. Craik (Eds.), Lifespan Cognition: Mechanisms of Change. Oxford University Press: New York.

Reuter-Lorenz, P. A. (2000). Cognitive neuropsychol- ogy of the aging brain. In D. C. Park & N. Schwarz (Eds.), Cognitive aging: A primer (pp. 93!114). Philadelphia: Psychology Press.

Salthouse, T. A. (1993). Speed mediation of adult age differences in cognition. Developmental Psychology, 29, 722!738.

Salthouse, T. A. (1994). ‘‘The aging of working mem- ory’’. Neuropsychology, 8, 535!543.

Salthouse, T. A. (1996). The processing-speed theory of adult age differences in cognition. Psychological Review, 103, 403!428.

Salthouse, T. A., Kausler, D. H., & Saults, J. S. (1988). Utilization of path analytic procedures to investigate the role of processing resources in cognitive aging. Psychology and Aging, 3, 158!166.

Siegel, L. (1994). Working Memory and Reading: A lifespan perspective. International Journal of Behavioral Development, 17, 109!124.

Shah, P., & Miyake, A. (1996). The separability of working memory resources for spatial thinking and language processing: An individual differences approach. Journal of Experimental Psychology: General, 125, 4!27.

Swanson, H. L. (1999). What develops in Working Memory? A lifespan perspective. Developmental Psychology, 35, 986!1000.

Towse, J. N., Hitch, G. J., & Hutton, U. (2002). On the nature of the relationship between processing activity and item retention in children. Journal of Experimental Child Psychology, 82, 156!184.

WORKING MEMORYACROSS THE LIFESPAN 93

D ow

nl oa

de d

by [9

8. 82

.7 6.

88 ] a

t 0 9:

10 2

6 Fe

br ua

ry 2

01 3

View publication stats

https://www.researchgate.net/publication/271673579