The efficacy of working memory training in improving
crystallized intelligence
Tracy Packiam Alloway1 & Ross G. Alloway2
1

Psychology Department, University of Stirling, Stirling FK9 4LA, UK; 2English Literature

Department, University of Edinburgh, Edinburgh EH8 9JZ, UK). These authors contributed

Nature Precedings : hdl:10101/npre.2009.3697.1 : Posted 28 Aug 2009

equally to this work.

Crystallized intelligence (Gc) is thought to reflect skills acquired through knowledge and
experience and is related to verbal ability, language development1 and academic success2.
Gc, together with fluid intelligence (Gf), are constructs of general intelligence3. While Gc
involves learning, knowledge and skills, Gf refers to our ability in tests of problem-solving,
pattern matching, and reasoning. Although there is evidence that Gf can be improved
through memory training in adults4, the efficacy of memory training in improving acquired
skills, such as Gc and academic attainment, has yet to be established. Furthermore,
evidence of transfer effects from gains made in the trained tasks is sparse5. Here we
demonstrate improvements in Gc and academic attainment using working memory
training. Participants in the Training group displayed superior performance in all
measures of cognitive assessments post-training compared to the Control group, who
received knowledge-based training. While previous studies have indicated that gains in
intelligence are due to improvements in test-taking skills6, this study demonstrates that it is
possible to improve crystallized skills through working memory training. Considering the
fundamental importance of Gc in acquiring and using knowledge and its predictive power
for a large variety of intellectual tasks, these findings may be highly relevant to improving
educational outcomes in those who are struggling.

Working memory capacity is thought to be as a fluid cognitive skill7 that is closely linked
with fluid intelligence (Gf)8. There is substantial evidence that working memory and Gf share
neural substrates, such as the prefrontal and parietal cortices9,10. While some psychologists
suggest that the two constructs are so highly correlated that they could be considered as
isomorphic properties11, Gf and working memory do appear to be dissociable12,13. Given both the
neural basis and the psychometric evidence for the close relationship between working memory
and Gf, training of one neural circuit might led to benefits in another shared domain. Indeed,

Nature Precedings : hdl:10101/npre.2009.3697.1 : Posted 28 Aug 2009

recent evidence suggests that memory training results in gains in fluid intelligence4, though this
finding is controversial14.

In contrast to Gf, crystallized intelligence (Gc) reflects acquired skills and knowledge.
Accordingly, different neural substrates are associated with Gc: it is more closely linked to brain
regions that involve the storage and usage of long-term memories, such as the hippocampus15.
Would training fluid cognitive skills, such as working memory, result in improvements in
acquired skills, such as Gc and academic attainment? To assess the potential gains in acquired
skills as the result of working memory training, we used a paradigm to train working memory in
the context of specific tasks that reflect acquired skills16. The training program consisted of three
games with up to 30 levels in each game and the participant has to successfully answer 8 out of
10 trials in each level to move forward to the next level. If the participant struggles, the program
adapts and moves to an easier level. All three games require the individual to simultaneously
process and remember information for a brief period.

The task in the first game was to scan a 4x4 grid with stimuli as quickly as possible and
remember the location of the target stimulus. The first level began with letters and became
progressively harder so the participant had to remember highly familiar word endings, and then
complete words using those word endings. As performance improved, the amount of information
on the grid increased and the time to respond decreased. In the second game, the task was to

process letter rotations starting with simple rotation (is the letter facing up or down?) to more
complex rotations (mirror image). The memory component in this game was to remember the
location of a red dot that appeared next to the letter. As performance improved, the complexity of
the rotations increased as did the number of dot locations to be remembered. The task of the last
game was to solve math problems while remembering the solutions in the correct sequence. At
the easiest level, the participant had to solve one problem and remember one solution. As
performance improved the complexity of the problems increased (e.g., from single-digit addition

Nature Precedings : hdl:10101/npre.2009.3697.1 : Posted 28 Aug 2009

and subtraction to multiplication to double-digit addition and subtraction), and remember up to
six solutions in the correct sequence.

To determine the efficacy of memory training in improving Gc, we randomly allocated
participants into one of two groups. Those in the Training group used the memory training three
times a week and completed 75 trials on average for all three memory games (25 trials per game)
over an 8-week period, lasting 30 minutes per session. Those in the Control group received
targeted educational support at school three times a week over an 8-week period for
approximately 25 total sessions. These sessions lasted 30 minutes each time and focused on
acquired skills relevant for attainment. All participants were pre-tested on measures of Gc,
academic attainment, and working memory; and then post-tested on the same measures. It is
important to note that the Training and Control groups did not differ with respect to crystallized
intelligence, working memory, or academic attainment in the pre-training assessment. As the
working memory training involved tasks that were distinct from the test measures, we postulated
that any observed gains in Gc and academic attainment could be explained by the training
program, rather than practice effects or test-taking skills.

To examine the gains as a function of memory training, we subtracted the pre-test scores
from the post-test scores and compared the difference in scores as a function of group (Training
vs. Targeted educational support). In Figure 1, scores below 0, as marked by the line, indicate
that the group performed worse at the post-test. Scores above 0 indicate improvements that the
group made after 10 weeks. There are marked differences in the gains made between the
Training and the Control groups. The superior performance of the Training group compared to
the Control group was confirmed in all the cognitive measures: Gc (U=8.5, p=.02), academic

Nature Precedings : hdl:10101/npre.2009.3697.1 : Posted 28 Aug 2009

attainment (U=12.5, p=.04), and working memory (U=8.5, p=.04).

Are these gains meaningful? Yes: participants in the Training group made on average an
increase of almost 10 standard points in the measure of Gc. There are two lines of evidence to
suggest that participants would not have achieved this gain without memory training. First, the
control group, who did not participate in the training program, showed no improvement in Gc
despite receiving targeted educational support that was tailored to improve the knowledge and
skills. The second line of evidence that demonstrates participants do not improve in their Gc
without training comes from a recent study on participants with learning difficulties17. They were
assessed on measures of fluid and crystallized intelligence, working memory, and academic
attainment at two time points. In between the testing periods, all participants received targeted
educational support from their schools for two years. However, retest scores indicated that none
of the participants showed any significant gains on the measures of acquired skills. One
possibility is that if fluid cognitive abilities such as working memory are deficient, the ability to
acquire knowledge and related skills is limited. This fits with the idea that working memory
functions like a bottleneck for learning in individual learning episodes required to increase
knowledge18. It is reasonable to suggest that without memory training, those with learning
difficulties struggle to ‘catch up’ with their peers19.

It is worth considering why memory training improved performance. We look first at the
superior performance of the Training group in the memory test, which can be explained by the
inherent properties of the working memory training program. Participants were required to
process some information, continually update representations in their ‘mental workspace’, and
then recall the information in the correct sequence20. As the surface features of the stimuli in the
working memory training were different from that of the memory test and the stimuli in the
memory test were randomized, the gains made in the memory test are unlikely due to a practice

Nature Precedings : hdl:10101/npre.2009.3697.1 : Posted 28 Aug 2009

effect. Rather, these gains can be explained by an increase in either capacity or attentional
control via training21, which facilitated superior recall.

The key finding is that this increase in working memory capacity was not restricted to
improvements in fluid skills but transferred to acquired skills as demonstrated by gains in Gc and
academic attainment. As the Control groups did not demonstrate gains in these tests, it is
unlikely that the training-related gains are due to practice effects. These transfer effects that
emerged are likely due to the nature of the working memory training. Not only did the training
focus on the ability to remember and process information, it integrated this skill with knowledge
and skills necessary for academic success.

While longitudinal research is necessary to explore the maintenance effects of the gains,
this study represents an exciting first step in understanding more about the underlying
relationship between fluid cognitive skills such as working memory and acquired skills like
crystallized intelligence. Although they do not share the same neural substrates, working
memory does appear to impact our ability to acquire knowledge. This view is consistent with
working memory is fundamental to general intelligence, predicting as much as 70% of variance
in these skills22. Not only are working memory tests powerful predictors of ability, our study
demonstrates that training working memory can improve this ability. This finding is significant
because it demonstrates that Gc is not resistant to change and can be improved without training

test-taking skills. There are tremendous implications for this that relate not only to education, but
in professional environments and vulnerable populations associated with low levels of

Nature Precedings : hdl:10101/npre.2009.3697.1 : Posted 28 Aug 2009

crystallized intelligence, such as those with learning disabilities, as well as juvenile delinquents.

Methods
Participants: Fifteen students with learning disabilities participated in this study. None of
the participants had any physical, sensory, or behavioral impairments. Eight children participated
in the Training condition (boys=86%; M age= 12.9 yr, SD=1 yr), while the remaining seven
children formed the Control group (boys=88%; M age= 13 yr, SD=0.4). They received targeted
learning support through an Individual Education Program (IEP) in schools for the duration of

Nature Precedings : hdl:10101/npre.2009.3697.1 : Posted 28 Aug 2009

the training period (8 weeks). The Training and Control groups did not differ significantly with
respect to age; t(13)<1.
Measures: All of the following measures were administered pre- and post-intervention for
both the Training and Control groups. Raw scores were converted into standard scores where
100 is the mean and 85 is the standard deviation. Crystallized intelligence was assessed using the
Vocabulary subtest from Wechsler Abbreviated Scales of Intelligence23. Standardized procedure
was followed. The Training and Control groups did not differ significantly with respect to the
crystallized intelligence score pre-training; t(13)=1.3; p=22.
Academic attainment was measured using the Numerical Operations test from the
Wechsler Objective Numerical Dimensions24. It consists of 10 four-item tests. The first set is
designed to assess the ability to write dictated numerals. The subsequent sets refer to
computational problems addition, subtraction, multiplication and division. Standardized
procedure was followed. The Training and Control groups did not differ significantly with
respect to the academic attainment score pre-training; t(13)<1.
Working memory was tested using a Letter recall test where the participant was shown a
letter on the computer screen, immediately followed by another letter. They had to verify
whether the letters were the same and then remember the target letters in the correct sequence.
The stimuli were randomized so no stimulus sequence was repeated to avoid potential practice

effects. The Training and Control groups did not differ significantly with respect to the working
memory score pre-training; t(13)=1.8; p=10.

1. Kline, P. in The new psychometrics: Science, psychology and measurement (Routledge,
London, 1998).
2. Deary, I. J., Strand, S., Smith, P. & Fernandes, C. Intelligence and educational achievement.
Nature Precedings : hdl:10101/npre.2009.3697.1 : Posted 28 Aug 2009

Intelligence. 35, 13-21 (2007).
3. Cattell, R. B. in Abilities: Their structure, growth, and action (Houghton Mifflin, New York,
1971).
4. Jaeggi, S. M., Buschkuehl, M., Jonides, J. & Perrig, W. J. Improving fluid intelligence with
training on working memory. Proc Natl Acad Sci USA. 105, 6829–6833 (2008).
5. Healy, A. F., Wohldmann, E. L., Sutton, E. M. & Bourne, L. E. Specificity effects in training
and transfer of speeded responses. J Exp Psychol Learn Mem Cognit. 32, 534–546 (2006).
6. Bors, D. A. & Vigneau, F. The effect of practice on Raven’s Advanced Progressive Matrices.
Learn Individ Differ. 13, 291–312 (2003).
7. Blair, C. How similar are fluid cognition and general intelligence? A developmental
neuroscience perspective on fluid cognition as an aspect of human cognitive ability. Behavioral
and Brain Sciences, 29, 109-160 (2006).
8. Kyllonen, P. C. & Christal, R. E. Reasoning ability is (little more than) working-memory
capacity. Intelligence. 14, 389–433 (1990).
9. Gray, J. R., Chabris, C. F. & Braver, T. S. Neural mechanisms of general fluid intelligence.
Nat Neurosci. 6, 316–322 (2003).

10. Kane, MJ, Engle RW (2002) The role of prefrontal cortex in working-memory capacity,
executive attention, and general fluid intelligence: An individual-differences perspective.
Psychon Bull Rev 9:637–671.
11. Stauffer, J. M., Ree, M. J., & Carretta, T. R. Cognitive components tests are not much more
than g: An extension of Kyllonen's analyses. Journal of General Psychology, 123, 193−205
(1996).

Nature Precedings : hdl:10101/npre.2009.3697.1 : Posted 28 Aug 2009

12. Ackerman, P. L., Beier, M. E. & Boyle, M. O. Working memory and intelligence: The same
or different constructs? Psychol Bull. 131, 30–60 (2005).
13. Conway, A. R., Kane, M. J. & Engle, R. W. Working memory capacity and its relation to
general intelligence. Trends Cognit Sci. 7, 547–552 (2003).
14. Moody, D. E. Can intelligence be increased by training on a task of working memory? Proc
Natl Acad Sci USA. 37, 327-328.
15. Geary, D. C. in The origin of mind: Evolution of brain, cognition, and general intelligence.
American Psychological Association, Washington, DC: 2005).
16. Alloway, T. P. & Alloway, R. G. JungleMemory Training Program (Memosyne Ltd, UK,
2008).
17. Alloway, T. P. Working memory, but not IQ, predicts subsequent learning in children with
learning difficulties. European Journal of Psychological Assessment. 25, 92-98 (2009).
18. Gathercole, S. E., Alloway, T. P., Willis, C. & Adams, A.M. Working memory in children
with reading disabilities. Journal of Experimental Child Psychology. 93, 265-281 (2006).
19. Alloway, T. P., Gathercole, S. E, Kirkwood, H. J. & Elliott, J. E. Evaluating the validity of
the Automated Working Memory Assessment. Educational Psychology. 7, 725-734 (2008).
20. Baddeley, A. Working memory: Looking back and looking forward. Nature Reviews
Neuroscience. 4, 829-839 (2003).

21. Cowan, N., & Alloway, T.P. in Development of Memory in Childhood, 2nd edition (eds
Cowan N. & Courage, M.) 303-342 (Psychology Press, Hove, 2008).
22. Jarrold, C. & Towse, J. N. Individual differences in working memory. Neuroscience. 139,
39-50.
23. Weschler D. Weschler Abbreviated Scale of Intelligence. (Psychological Corporation,

Nature Precedings : hdl:10101/npre.2009.3697.1 : Posted 28 Aug 2009

London, 1999).
24. Wechsler, D. in Wechsler Objective Numerical Dimensions. (Pearson Assessment, London,
1996).

Author Contributions. T.P.A. and R.G.A contributed equally to the draft of the article.

Correspondence and requests for materials should be addressed to Tracy Alloway (t.p.alloway@stir.ac.uk).

Fig. 1. Shown are the differences in standard scores pre- and post-training for the
Control and Training groups.

Natura Pracadings hdb10101/nara.2003.9087.1 Pasled 28 Aug 2003

Nature Precedings : hdl:10101/npre.2009.3697.1 : Posted 28 Aug 2009

Mean

Mice Difference
Sattainment difference
Owitterence

°
SS

fo
Control
Training

Group

Error bars: 05% Cl