Content uploaded by Jeremy W Grabbe
Author content
All content in this area was uploaded by Jeremy W Grabbe
Content may be subject to copyright.
This article was downloaded by: [Jeremy Grabbe]
On: 21 September 2011, At: 11:34
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Activities, Adaptation & Aging
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/waaa20
Sudoku and Working Memory
Performance for Older Adults
Jeremy W. Grabbe
a
a
Psychology Department, State University of New York, Plattsburgh,
Hultsch, NY
Available online: 20 Sep 2011
To cite this article: Jeremy W. Grabbe (2011): Sudoku and Working Memory Performance for Older
Adults, Activities, Adaptation & Aging, 35:3, 241-254
To link to this article: http://dx.doi.org/10.1080/01924788.2011.596748
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, re-distribution, re-selling, 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.
Activities, Adaptation & Aging, 35:241–254, 2011
Copyright © Taylor & Francis Group, LLC
ISSN: 0192-4788 print/1544-4368 online
DOI: 10.1080/01924788.2011.596748
Sudoku and Working Memory Performance
for Older Adults
JEREMY W. GRABBE
Psychology Department, State University of New York, Plattsburgh, Hultsch, NY
In recent years interest has increased toward mental exercise as a
way to promote healthy cognitive aging. Consistent findings have
shown that declines in working memory performance are associ-
ated with aging. Sudoku is a popular puzzle game that has task
demands similar to working memory processes. Younger and older
adults completed a battery of tests and solved Sudoku puzzles. The
results showed that Sudoku performance had a significant rela-
tionship to working memory. This suggests that Sudoku has the
potential to become a new focus in the study of mental exercise
and cognitive aging.
KEYWORDS Sudoku, working memory, compensation
Can mental exercise enhance cognitive abilities that are associated with
working memory? Studies have shown that cognitive engagement preserves
cognitive functioning in older adults (Hultsch, Hertzog, Small, & Dixon,
1999). Studies of plasticity and aging have shown that older adults can
improve performance and reduce age-related differences in cognitive ability
(Hultsch et al., 1999), or at least reduce the rate of decline. Much of the work
on plasticity focuses on mental exercise. An important step in the study of
mental exercise is examining for a possible connection between the form of
mental exercise and a cognitive domain. If a form of mental exercise has a
relationship to a specific domain of cognition then the next step is to exam-
ine if this mental exercise can be used to offset or slow age-related changes
in cognition as noted in the use-it-or-lose-it hypothesis (Schooler, 2007).
Recently, Ackerman, Kanfer, and Calderwood (2010) found that mental exer-
cise on a Nintendo Wii resulted in some improvements in domain-specific
cognitive performance. This study examined if the popular game Sudoku
Received 30 November 2010; accepted 27 May 2011.
Address correspondence to Dr. Jeremy W. Grabbe, Psychology Department, SUNY
Plattsburgh, 101 Broad Street, Plattsburgh, NY 12901. E-mail: jgrab001@plattsburgh.edu
241
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
242 J. W. Grabbe
has a relationship with measures of fluid intelligence abilities such as work-
ing memory. Age-related declines in working memory have been observed
in numerous studies (Allain et al., 2007; Zeintl & Kliegel, 2007).
Working memory involves executive function such as planning and
supervision of attention. Key processes of working memory involve the
maintenance and manipulation of visual information. Working memory has
been characterized as a facet of fluid intelligence (Blair, 2006). Fluid intelli-
gence consists of areas of cognition that are innate and not learned such as
problem solving and inductive reasoning (Klauer, Willmes, & Phye, 2002).
Fluid intelligence, although innate, is largely thought to be responsive to
training (Bond, Wolf-Wilets, Fiedler, & Burr, 2000). Mental exercise is a form
of cognitive training that seeks to ameliorate age-related declines in cogni-
tive performance. Some studies have found advantages for mental exercise
(Paggi & Hayslip, 1999; Bond et al., 2000) while others have cast doubt on
mental exercise’s effectiveness (Salthouse, 2006). One possible suggestion
to reconcile these findings is that mental exercise may be domain-specific
(Salthouse, 2006). Although working memory does show some plasticity, the
extent of the plasticity is unknown. To examine new possibilities of answer-
ing this question I looked at a task (Sudoku) as a domain-specific (working
memory) form of mental exercise.
Studies examining working memory in older adults have found that
older adults have diminished working memory performance (Emery, Hale,
& Myerson, 2008). Given the pivotal role working memory plays in activities
of daily living (ADL) and other tasks; it is of particular concern if Sudoku
presents a possible way in which to exercise working memory, thus possibly
reducing the age-related declines in working-memory performance. Given
that one of the most glaring age-related differences in cognition comes
from older adults’ poorer performance on working memory tasks (Zeintl
& Kliegel, 2007), the study of how mental exercise may relate to working
memory is of great importance.
Sudoku, a popular Japanese game involving the correct placement of
nine nonrepeating digits, has recently become popular in the United States.
Sudoku has great potential for those wishing to study mental exercise, par-
ticularly mental exercise related to working memory. One criticism of mental
exercise is that it can be monotonous. A factor in the popularity of Sudoku
is that many find it very enjoyable. Many of the forms of mental exercise
that may be regarded as fun may not possess attributes that relate to the
functions of working memory. Sudoku is exciting to the field of mental
exercise because the cognitive demands involved in playing Sudoku require
the maintenance of several numbers in memory at once. Also involved is the
use of logic, planning, and manipulations (e.g., mental rotation, rehearsal,
planning, etc.). These mental processes fall w ithin the domain of working
memory. The added bonus is that Sudoku is enjoyable and popular, which
makes people more interested in using Sudoku as a form of mental exercise.
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
Sudoku and Working Memory 243
Studies of age and mental exercise have examined games such as chess
(Roring & Charness, 2007). In this study Sudoku performance was evalu-
ated and compared to a battery of working-memory measures. This study
sought to take the crucial step of assessment for any potential form of men-
tal exercise—to determine if the exercise has a relationship with a specific
domain of cognition (working memory). Because the task demands required
to play Sudoku are similar to working-memory processes, it was predicted
that Sudoku-playing performance and working-memory test performance
will share a significant relationship.
METHOD
Participants
Forty-seven participants were recruited for this experiment. Twenty-eight
younger adults (mean age = 29.6 years, SD = 8.8 years; 5 males, 23 females)
who were undergraduates at SUNY Plattsburgh at Queensbury volunteered
for course credit. Nineteen older adults (mean age = 72.2 years, SD = 7.3
years; 7 males, 12 females) were recruited from the community through
social groups as well as through a subject pool from previous research not
related to this study. The mean years of education for younger adults was
14.9 while older adults reported a mean of 13.5. A self-report of familiarity
with Sudoku for younger adults indicated that 16 had never played, 5 had
played less than three times in their lives, and 6 played at least once a month.
Thirteen older adults reported never playing Sudoku, three reported playing
less than three times in their lives, and three reported playing at least one
game per week. This finding showed that both age groups were comparable
in ratio of novices-to-experienced Sudoku players. The low rate of people
highly familiar with Sudoku made an analysis of expertise unwarranted.
Procedure
Participants completed informed consents and then were presented with a
series of cognitive tests and Sudoku puzzles. Specific instructions for each
task were given at the time of task administration. Once the subjects com-
pleted the cognitive task they went on to play Sudoku. All participants were
instructed on how to play regardless of their prior experience with the
game. After receiving instructions on how to play, participants completed
an instructional Sudoku during which the experimenter assisted the partic-
ipants when asked. During this time the instructor would also point out
errors. When the participant completed the instructional Sudoku they then
began the regular Sudoku trials. Participants were instructed to never guess
at any point. If they believed that they had to guess in order to complete
the game, the participant was required to stop. Participants completed the
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
244 J. W. Grabbe
puzzles in pencil and were allowed to erase errors only if that error was
not caused by another error. That is, if the participant made an error (e.g.,
placed a 4 in a row that already had a 4 in it) and discovered that this error
was made because of a previous error they had made and had not corrected,
they would have to stop.
Participants started with easy and then moved on to medium-difficulty
Sudoku puzzles. Performance on Sudoku puzzles was recorded by the
number of hits and number of errors. Hits were recorded when a num-
ber was correctly placed; an error was when a number was incorrectly
placed. Participants completed the study in one session lasting approxi-
mately 90 minutes in a laboratory at the Plattsburgh campus. Regular breaks
were given to prevent fatigue.
Materials
Materials were from a test battery used by Hedden and Yoon (2006). These
tests represent a broad and exhaustive array of measures of working memory
and cognition.
P
LUS-MINUS TASK
The plus-minus task consisted of three separate trials that involved the pre-
sentation of random two-digit numbers. The participants’ task was to add
three to the target number on the first block of trials. On the second block
of trials participants had to subtract three from the target number. The
third block involved alternating between adding three to the target num-
ber on one trial and subtracting three from the target number on the next
trial. Participants had to keep track of whether they were to add or sub-
tract on each trial. No external cues were given to participants. Participants
responded by pressing one of four keys that spatially corresponded to four
possible answers presented on the screen below the target number.
L
ETTER MEMORY
The letter memory task involved the presentation of a string of letters in
which the last four letters must be recalled. Participants wrote down their
response. The four different lengths of letter strings were 5, 7, 9, and 11.
Each particular length was presented three times for a total of 12 lists. Letters
were presented on a computer screen one at a time for 2,000 milliseconds
per letter. Then the word recall appeared and participants wrote down the
last four letters.
B
ACKWARD DIGIT SPAN
For the backward digit span (BDS) task participants heard the experimenter
speak a series of digits. Digits were spoken at a rate of one per second.
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
Sudoku and Working Memory 245
Participants then had to repeat the digits they just heard in the opposite
order. Blocks consisted of two trials, and each trial consisted of two to eight
digits. When the participant failed to successfully repeat the digits in both
trials of the same block the task was terminated. The number of correct trials
was the dependent variable.
S
TROOP TASK
A traditional Stroop task, adapted from Hedden and Yoon (2006), was per-
formed on a computer. Participants had to press one of four corresponding
keys on a keyboard to respond to the color of a stimulus: red, green, blue,
or yellow. In 72 neutral trials, only asterisks were shown. In addition, in 72
word trials a word appeared (BLUE, RED, GREEN, or YELLOW). In 60 of
these word trials the color would be incongruent; the remaining 12 trials
were congruent (e.g., BLUE was colored blue). The dependent variable was
reaction time and accuracy.
D
IGIT SYMBOL SUBSTITUTION
Participants completed the digit symbol substitution task from the Wechsler
Adult Intelligence Scale-Revised (WAIS-R). Participants had 90 seconds to
match symbols under corresponding digits. The dependent variable was the
number of correct substitutions.
L
ETTER COMPARISON
The letter comparison task consisted of presenting on a computer pairs of
letter strings to the participant. The participants’ task was to determine if
the letter strings, which were 3, 6, or 9 letters in length, matched or were
different. Participants pressed one of two keys to respond if the strings
were different or the same. The dependent variable was derived by tak-
ing the number of correct trials and subtracting the number of incorrect
trials.
V
OCABULARY
Participants’ vocabulary was assessed using the WAIS-R vocabulary subscale.
S
EMANTIC FLUENCY
For the semantic fluency task participants had 60 seconds to write down as
many animals as they could. Participants responded by writing down the
names of animals on an answer sheet.
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
246 J. W. Grabbe
SUDOKU
Participants had two easy Sudoku puzzles to complete that were timed for a
maximum of 5 minutes each. Medium-difficulty Sudoku puzzles were timed
for a maximum of 5 minutes each. Participants also completed one easy and
one medium untimed Sudoku puzzle.
RESULTS
Data were subjected to a canonical correlation analysis. The significance
level was set at p < 0.05 two-tailed. The correlation matrix is presented in
Table 1. There was no missing data because all 47 participants completed
all measures of the experiment.
Comparison of Sudoku and Working Memory
Digit-symbol performance had a strong relation to Sudoku performance
(see Table 1). Digit symbol was negatively correlated with easy error,
r(47) =−.395, p < .05; easy timed error, r(47) =−.320, p < .05; and untimed
error, r(47) =−.301, p < .05. A strong positive correlation existed for digit
symbol with medium hits, r(47) = .355, p < .05; easy timed hits, r(47) = .344,
p < .05; and untimed hits, r(47) = .317, p < .05.
Looking at the relation between Sudoku and traditional measures,
a negative correlation existed between percentage of errors in untimed
Sudoku puzzles and percentage of correct responses in a traditional
Stroop task: r(47) =−.473, p < .01 (see Table 2). Increased reaction
time for the Stroop task had a significant positive correlation with
mean medium hits for timed Sudoku puzzles, r(47) = .290, p < .05.
A strong association between Sudoku and Stroop-task performance was
found. Significant negative correlations were found between accuracy
and error for easy Sudoku puzzles, r(47) =−.495, p < .05; medium
error, r(47) =−.460, p < .01; easy timed error, r(47) =−.303, p < .05;
and medium timed error, r(47) =−.360, p < .05. This demonstrated
that increased accuracy on the Stroop task was correlated with fewer
errors on Sudoku. Curiously, there was a significant negative correla-
tion between Stroop-task accuracy and medium timed hits, r(47) =−.347,
p < .05. This may indicate the possibility that participants’ sacrificed speed
to retain accuracy on the Stroop task (this would explain why accurate
people had less errors on a timed puzzle but not enough time to mark
many correct hits). Furthermore, higher levels of error in easy Sudoku puz-
zles, r(47) = .330, p < .05, and medium timed Sudoku puzzles, r(47) = .290,
p < .05, were positively correlated with increased Stroop-task reaction time.
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
TABLE 1 Sudoku Performance, Age, Vocabulary, Digit Symbol, Semantic Fluency, Backward Digit, and Letter Memory C orrelations
Age
Digit
symbol
Semantic
fluency
Backward
digit
Letter
memory
5 digits
Letter
memory
7 digits
Letter
memory
9 digits
Letter
memory
11 digits
Mean easy hits 0.064 0.283 0.010 −0.107 0.301
∗
0.098 0.080 0.173
Mean easy error 0.462
∗∗
−0.395
∗∗
−0.431
∗∗
−0.176 −0.063 −0.128 −0.255 −0.033
Mean medium error 0.482
∗∗
−0.267 −0.403
∗∗
−0.153 −0.123 −0.241 −0.259 −0.052
Mean medium hits −0.072 0.355
∗
0.198 −0.057 0.358
∗
0.218 0.109 0.142
Mean easy timed hits −0.056 0.344
∗
0.125 0.205 0.189 0.329
∗
0.220 0.243
Mean easy timed error 0.281 −0.320
∗
−0.267 −0.082 0.049 0.019 −0.086 0.021
Mean untimed error 0.499
∗∗
−0.301
∗
−0.423
∗∗
−0.288 −0.157 −0.300
∗
−0.371
∗
−0.071
Mean untimed hits −0.013 0.317
∗
0.106 −0.183 0.337
∗
0.111 0.087 0.125
Mean medium timed hits 0.281 −0.027 −0.212 0.004 0.151 −0.043 −0.165 −0.001
Mean medium timed error 0.344
∗
−0.176 −0.311
∗
0.124 −0.110 −0.082 −0.079 0.009
∗
p < 0.05.
∗∗
p < 0.01.
247
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
248 J. W. Grabbe
TABLE 2 Sudoku Performance and Letter Memory Correlations
Stroop
accuracy Stroop RT
Plus-Minus
task RT
Plus-Minus task
accuracy
Mean easy hits 0.081 0.122 −0.038 −0.120
Mean easy error −.495
∗∗
.330
∗
.518
∗∗
−0.232
Mean medium error −.460
∗∗
0.244 .505
∗∗
−0.236
Mean medium hits 0.084 −0.027 −0.097 −0.099
Mean easy timed hits 0.051 0.015 −0.166 0.108
Mean easy timed error −.303
∗
0.275 .339
∗
−0.269
Mean untimed error −.473
∗∗
0.220 .547
∗∗
−0.163
Mean untimed hits 0.137 0.044 −0.061 −0.144
Mean medium timed hits −.347
∗
.290
∗
0.195 −0.089
Mean medium timed error −.360
∗
0.188 .353
∗
−0.212
∗
p < 0.05.
∗∗
p < 0.01. RT = Reaction Time.
Significant positive correlations were also identified between reaction
time for the plus-minus task and mean error for Sudoku puzzles: easy error,
r(47) = .518, p < .01; medium error, r(47) = .505, p < .01; easy timed error,
r(47) = .339, p < .05; untimed error, r(47) = .547, p < .01; and medium timed
error, r(47) = .353, p < .05. This suggests that poorer reaction-time perfor-
mance was associated with high error rates. However, the error rates on
the plus-minus task were not related to the error rates for Sudoku puz-
zles. Because four of the five Sudoku measures were timed (e.g., easy
timed error) or untimed (e.g., medium error) combined with time that
was correlated with the plus-minus task reaction time, this might indicate
poorer performance when under time constraints. Speed has been a fac-
tor in age-related performance (Salthouse, 1984). Processing speed appears
to be a factor in this finding. There was a significant negative correlation
between semantic fluency and mean error for both easy, r(47) =−.431,
p < .05, and medium, r(47) =−.403, p < .05. Semantic fluency also had
a negative correlation with untimed error rate, r(47) =−.423, p < .05, and
with medium timed error, r(47) =−.311, p < .05. This demonstrated that
as we see an increase in fluency the error rate decreases. This suggests
that accurate Sudoku performance involves the ability to inhibit contra-
vening information. Semantic fluency measures the inhibition executive
function of working memory (Hedden & Yoon, 2006), therefore, part of
playing Sudoku involves knowing where not to misplace a digit. This is
very interesting in terms of relating working memory performance to work-
ing memory process. A significant positive correlation exists between letter
memory (5 digits) and mean easy hits, r(47) = .301, p < .05; with mean
medium hits, r(47) = .358, p < .05; and with untimed hits, r(47) = .337,
p < .05. Examining letter memory (7 digits) had a significant positive
correlation with easy timed hits, r(47) = .329, p < .05, and a significant neg-
ative correlation with untimed error,
r(47) =−.300, p < .05. Letter memory
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
Sudoku and Working Memory 249
(9 digits) had a negative correlation with untimed errors for Sudoku puzzles,
r(47) =−.371, p < .05.
The results of letter comparison and Sudoku were interesting because of
the parallel in task demands between the two. There was also a significant
positive correlation between reaction time on letter comparison (3 digits)
and error for easy Sudoku puzzles (see Table 3), r(47) = .328, p < .05, as
well as for untimed error, r(47) = .374, p < .05. However, the error rates
for Sudoku were not significantly correlated to error rates for letter compar-
ison (3 digits). There was a significant negative correlation between mean
medium overall hits for Sudoku and letter comparison (9 digits) reaction
time, r(47) =−.320, p < .05. Another significant negative correlation was
found for mean easy timed hits for Sudoku and letter comparison (9 digits)
reaction time, r(47) =−.319, p < .05. These results are interesting because a
greater level of accuracy on Sudoku is associated with the ability to scan and
compare nine digits. All Sudoku rows, columns, and three-by-three boxes
contain exactly nine digits. The ability to quickly process nine digits for
error would facilitate better performance on Sudoku puzzles. Individuals
who would not be able to scan nine digits easily would have a more difficult
time making hits.
From the r esults, so far we see that speed (plus-minus task), inhibi-
tion (semantic fluency), and span (letter comparison) correlate to Sudoku
performance. These results demonstrate that there is a relationship between
Sudoku and working memory, which was the goal of this study. If Sudoku
could become a form of mental exercise for the aging, what is the role of
age in Sudoku and working memory? In the following section the effects of
age are reported.
Age-Related Findings
Although age did correspond to significantly higher levels of error on some
Sudoku puzzles, it was not related to the number of hits. There were several
significant positive correlations between age and error rate for Sudoku (see
Table 1). When looking at measures of working memory it was revealed
that older adults performed more poorly on measures of working memory
(see Table 4). A trend also showed poorer performance on the more difficult
letter memory tasks as age increased. Age showed a strong relationship to
increases in reaction time and decreases in accuracy for the Stroop task
and the plus-minus task. Age showed no significant effect on accuracy for
the letter comparison task, despite a significant increase in reaction time
associated with age. Although some of the correlations were not as strong,
they still demonstrated a previous relationship between Sudoku and working
memory. These age effects also provide enough evidence to warrant future
studies examining aging and expertise in Sudoku.
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
TABLE 3 Sudoku Performance, Stroop Task, and Plus-Minus Task Correlations
Letter
comparison
3RT
Letter
comparison
3 accuracy
Letter
comparison
6RT
Letter
comparison
6 accuracy
Letter
comparison
9RT
Letter
comparison
9 accuracy
Mean easy hits 0.006 0.188 −0.045 0.114 −0.152 0.146
Mean easy error .328
∗
0.087 0.164 −0.123 0.289 −0.017
Mean medium error 0.289 0.114 0.143 −0.105 0.199 −0.018
Mean medium hits −0.090 0.190 0.072 0.011 −.320
∗
0.156
Mean easy timed hits −0.137 0.085 −0.131 0.074 −.319
∗
−0.053
Mean easy timed error 0.142 0.025 0.078 −0.102 0.206 −0.128
Mean untimed error .374
∗
0.151 0.226 −0.109 0.248 0.106
Mean untimed hits −0.013 0.226 0.007 0.098 −0.195 0.216
Mean medium timed hits 0.211 0.027 0.133 −0.092 0.104 0.003
Mean medium timed error 0.136 0.023 0.041 −0.106 0.077 −0.158
∗
p < 0.05. RT = Reaction Time.
250
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
TABLE 4 Age Comparison Among Measures of Working Memory
Vocabulary
Semantic
fluency
Backwards
digit
Letter
memory
5 digits
Letter
memory
7 digits
Letter
memory
9 digits
Letter
memory
11 digits
Plus-minus
task RT
Plus-minus
task
accuracy
Age −.292
∗
−.581
∗∗
−0.237 −0.127 −0.249 −.488
∗∗
−.361
∗
.663
∗∗
−.355
∗
Letter
comparison
3RT
Letter
comparison
3 accuracy
Letter
comparison
6RT
Letter
comparison
6 accuracy
Letter
comparison
9RT
Letter
comparison
9 accuracy
Stroop
accuracy Stroop RT
Age .745
∗∗
−0.017 0.266 −0.214 .709
∗∗
−0.099 −.568
∗∗
.675
∗∗
∗
p < 0.05.
∗∗
p < 0.01.
251
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
252 J. W. Grabbe
DISCUSSION
This study examined whether Sudoku has a relationship with traditional
measures of working memory. Furthermore, the role that age plays in
this possible relationship was studied. The results show that Sudoku has
a significant relationship with various measures of working memory. This
relationship is centered around a consistent finding that increased error on
Sudoku is r elated to poorer performance on measures of working memory.
Age is also associated with poorer working memory and Sudoku perfor-
mance. The current body of research on mental exercise has called attention
to a domain-specific focus of mental exercise (Hertzog, Kramer, Wilson, &
Lindenberger, 2009; Paggi & Hayslip, 1999; Salthouse, 2006; Schooler, 2007).
This study completed the crucial step in bringing a new tool to the focus
of mental exercise by demonstrating that the tool (Sudoku) does have some
form of relationship to working memory.
A relevant issue in the study of mental exercise is that not all forms
of mental exercises are universal preservative of all facets of cognition but,
rather, are domain specific (Salthouse, 2006). The data suggest a domain-
specific relationship for Sudoku. It is not only a fluid-intelligence domain
but also one that has been established to have plasticity (Bond et al., 2000)
and is known to decline with age (Emery et al., 2008; Holtzer, Stern, &
Rakitin, 2004). This brings a popular and fun game to the attention of mental
exercise, which is related to a critical component of cognition: working
memory. There is now the potential as well as the need to examine if Sudoku
can be used as a tool to promote working memory performance in older
adults.
The aim of this study is also one of its limitations. This study did not
establish that Sudoku improves working memory performance. The goal
was to examine if a relationship exists between Sudoku and working mem-
ory. The results show that a relationship does exist, but what the specific
relationship looks like requires further research. One aspect of this study
is that Sudoku errors were more often related to poorer performance on
measures of working memory than Sudoku hits. This may give some insight
for future studies in that perhaps a monitoring aspect of executive func-
tion has a role in this relationship. Another limitation of this study was
that the effect of expertise was not studied. Again, the precursory nature
of this study prohibited a closer examination of expertise. Furthermore,
although there were some people who had significant experience playing
Sudoku, there were not enough expert subjects to allow the study of exper-
tise. A focus for future research should be the role of expertise in Sudoku.
The study of expertise has often examined the limits of transfer and speci-
ficity (Karbach & Kray, 2009). Future studies of expertise and Sudoku could
elucidate on the possibility of transfer of Sudoku expertise and working
memory.
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
Sudoku and Working Memory 253
The results of this study now inject a relationship between working
memory and a popular puzzle game into the mental exercise literature. It is
essential to establish a relationship between mental exercise and a specific
cognitive domain (Karbach & Kray, 2009; Schooler, 2007), which is what has
been done in this study. The next logical step for future studies is to exam-
ine if extensive training in Sudoku can lead to improved working memory
performance.
The findings are novel for the study of mental exercise in that various
measures of working memory (digit symbol, plus-minus task, Stroop task,
letter memory, and letter comparison) are related to a widely popular puzzle.
Sudoku is a new element in the study of mental exercise, which can help
to expand the range of interest researchers can devote to the interaction
between aging and mental exercise. Given the need to promote measures to
preserve cognitive performance in older adults, this study shows an encour-
aging possibility for a popular puzzle to become a new focus in the study
of mental exercise.
REFERENCES
Ackerman, P. L., Kanfer, R., & Calderwood, C. (2010). Use it or lose it? Wii brain
exercise practice and reading for domain knowledge. Psychology and Aging,
25, 753–766.
Allain, P., Berrut, G., Etcharry-Bouyx, F., Barré, J., Dubas, F., & Le Gall, D.
(2007). Executive functions in normal aging: An examination of script sequenc-
ing, script sorting, and script monitoring. The Journals of Gerontology, 62B,
187–190.
Blair, C. (2006). How similar are fluid cognition and general intelligence? A devel-
opmental neuroscience perspective on fluid cognition as an aspect of human
cognitive ability. Behavioral and Brain Sciences, 29, 109–125.
Bond, G. E., Wolf-Wilets, V., Fiedler, F. E., & Burr, R. L. (2000). Computer-aided
cognitive training of the aged: A pilot study. Clinical Gerontologist, 22, 19–42.
Emery, L., Hale, S., & Myerson, J. (2008). Age differences in proactive interference,
working memory, and reasoning. Psychology and Aging, 23, 634–645.
Hedden, T., & Yoon, C. (2006). Individual differences in executive processing predict
susceptibility to interference in verbal working memory. Neuropsychology, 20 ,
511–528.
Hertzog, C., Kramer, A. F., Wilson, R. S., & Lindenberger, U. (2009). Enrichment
effects on adult cognitive development: Can the functional capacity of older
adults be persevered and enhanced? Psychological Science in the Public Interest,
9, 1–66.
Holtzer, R., Stern, Y., & Rakitin, B. C. (2004). Age-related differences in executive
control of working memory. Memory & Cognition, 32, 1333–1345.
Hultsch, D. F., Hertzog, C., Small, B. J., & Dixon, R. A. (1999). Use it or lose it:
Engaged lifestyle as a buffer of cognitive decline in aging. Psychology and
Aging, 14, 245–226.
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
254 J. W. Grabbe
Karbach, J., & Kray, J. (2009). How useful is executive control training? Age differ-
ences in near and far transfer of task-switching training. Developmental Science,
12 , 978–990.
Klauer, K. J., Willmes, K., & Phye, G. D. (2002). Inducing inductive reasoning: Does
it transfer to fluid intelligence? Contemporary Educational Psychology, 27, 1–25.
Paggi, K., & Hayslip, B. (1999). Mental aerobics: Exercises for the mind in later life.
Educational Gerontology, 25, 1–12.
Roring, R. W., & Charness, N. (2007). A multilevel model analysis of expertise in
chess across the life span. Psychology and Aging, 22, 291–299.
Salthouse, T. A. (1984). Effects of age and skill in typing. Journal of Experimental
Psychology: General, 113, 345–371.
Salthouse, T. A. (2006). Mental exercise and mental aging. Perspectives on
Psychological Science, 1, 68–87.
Schooler, C. (2007). Use it—and keep it, longer, probably. Perspectives on
Psychological Science, 2, 24–29.
Zeintl, M., & Kliegel, M. (2007). How do verbal distracters influence age-related
operation span performance? A manipulation of inhibitory control demands.
Experimental Aging Research, 33, 163–175.
Downloaded by [Jeremy Grabbe] at 11:34 21 September 2011
