RESEARCH ARTICLE

The impact of leisure activities on older adults’

cognitive function, physical function, and

mental health

Giovanni SalaID
1
, Daniela Jopp

2
, Fernand Gobet

3
, Madoka Ogawa

4
, Yoshiko Ishioka

5
,

Yukie Masui
4
, Hiroki Inagaki

4
, Takeshi NakagawaID

6
, Saori Yasumoto

7
, Tatsuro Ishizaki

4
,

Yasumichi Arai
8
, Kazunori Ikebe

9
, Kei Kamide

10
, Yasuyuki Gondo

7*

1 Institute for Comprehensive Medical Science, Fujita Health University, Toyoake, Japan, 2 Institute of

Psychology, University of Lausanne, Lausanne, Switzerland, 3 Centre for Philosophy of Natural and Social

Science, London School of Economics and Political Science, London, United Kingdom, 4 Tokyo Metropolitan

Institute of Gerontology, Tokyo, Japan, 5 Graduate School of Science and Technology, Keio University,

Yokohama, Japan, 6 National Center for Geriatrics and Gerontology, Aichi, Japan, 7 Graduate School of

Human Sciences, University of Osaka, Osaka, Japan, 8 Center for Supercentenarian Medical Research,

Keio University, Tokyo, Japan, 9 Graduate School of Dentistry, Osaka University, Osaka, Japan,

10 Graduate School of Medicine, Osaka University, Osaka, Japan

* [email protected]

Abstract

Engagement in leisure activities has been claimed to be highly beneficial in the elderly. Prac-

ticing such activities is supposed to help older adults to preserve cognitive function, physical

function, and mental health, and thus to contribute to successful aging. We used structural

equation modeling (SEM) to analyze the impact of leisure activities on these constructs in a

large sample of Japanese older adults (N = 809; age range 72–74). The model exhibited an

excellent fit (CFI = 1); engaging in leisure activities was positively associated with all the

three successful aging indicators. These findings corroborate previous research carried out

in Western countries and extend its validity to the population of Eastern older adults. Albeit

correlational in nature, these results suggest that active engagement in leisure activities can

help older adults to maintain cognitive, physical, and mental health. Future research will clar-

ify whether there is a causal relationship between engagement in leisure activities and suc-

cessful aging.

Introduction

Leisure activities (hereafter LA) can be defined as activities people engage in during free time

[1]. Engagement in LAs has been found to be positively associated with cognitive function,

physical function, and mental health in late adulthood and in the elderly. The possible protec-

tive effects of LA engagement against aging-related decline have thus been the object of investi-

gation in the last two decades.

Of these three outcomes, preserved cognitive function has received most attention and the

link with LA engagement in the elderly is well established [2–5]. Three possible explanations

have been proposed for the observed relationship between cognitive function and LA engage-

ment. First, practicing mentally challenging activities (e.g., music, board games, video games,

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 1 / 13

a1111111111

a1111111111

a1111111111

a1111111111

a1111111111

OPEN ACCESS

Citation: Sala G, Jopp D, Gobet F, Ogawa M,

Ishioka Y, Masui Y, et al. (2019) The impact of

leisure activities on older adults’ cognitive function,

physical function, and mental health. PLoS ONE 14

(11): e0225006. https://doi.org/10.1371/journal.

pone.0225006

Editor: Stephen D. Ginsberg, Nathan S Kline

Institute, UNITED STATES

Received: June 25, 2019

Accepted: October 25, 2019

Published: November 8, 2019

Copyright: © 2019 Sala et al. This is an open
access article distributed under the terms of the

Creative Commons Attribution License, which

permits unrestricted use, distribution, and

reproduction in any medium, provided the original

author and source are credited.

Data Availability Statement: According to the

approval by the Institutional Review Board of

Osaka University Graduate School of Dentistry

(approval number H22-E9), our dataset cannot be

made available for others. If the other researcher

requests to use our dataset, he/she should be

approved as a co-researcher by our IRB. If

someone requires to use the dataset, please

contact the data access committee at first:

[email protected] In order to reproduce the

main models, we provide the covariance matrices

and R codes (https://osf.io/csvwt/).

and brain training) may enhance overall cognitive function [6]. However, this idea has

received little adequate empirical support [7–9] in the general population. Second, people

exhibiting superior overall cognitive function may be more likely to engage in LAs that are

cognitive demanding. This hypothesis has been corroborated by numerous studies in the field

of chess and music [10,11]. Finally, engaging in intellectually demanding LAs may slow down

cognitive decline. This idea relies on the so-called “use it or lose it” hypothesis according to

which engaging in intellectually demanding activities helps to preserve cognitive function in

the elderly [12,13]. This hypothesis has received some support by studies implementing dual-

change approaches to test for causality between LA engagement and preserved cognitive func-

tion [14,15].

Compared to the link between LA engagement and cognitive function, the impact of LA

engagement on physical function has been less studied [16]. Most research has focused on the

adverse effects of illness and injuries on LA engagement. Reduced LA engagement has, in

turn, detrimental effects on indicators of mental health such as well-being and life satisfaction

[17]. The field has thus paid more attention to how physical function influences LA engage-

ment rather than vice-versa. Another line of research has examined the effects of physical

activity and LAs (as independent variables) on cognitive function [12,18].

Finally, LA engagement seems to be related to mental health as well. Mental health aspects

such as well-being and life satisfaction have been found to positively correlate with LA engage-

ment in several studies (for a review, see [17]). Studies implementing a longitudinal design

have confirmed these findings [19]. Nonetheless, the amount of robust experimental evidence

is still modest [20].

Considering its positive effects for key dimensions of functioning in older age, LA engage-

ment seems to have an important role for successful aging. In their seminal article, Rowe and
Kahn [21] introduced the concept of successful aging in opposition to usual aging. While usual
aging emphasizes the non-pathological aging (e.g., absence of disease), successful aging cap-

tures an optimal aging process. Successful aging is, according to Rowe and Kahn, best charac-

terized by the concurrent presence of three dimensions, namely high cognitive and physical

function, low probability of disease and disability, and active engagement in life. The latter

captures the involvement in productive and social activities, giving those activities a similar

importance as health and functioning for successful aging. The importance of activities for suc-

cessful aging has also been stressed very early on in other seminal theories, including activity

theory by Havighurst, Neugarten, and Tobin [22]. Such more multidimensional conceptuali-

zations of successful aging have resulted in complementing prior approaches by including

mental-health related dimensions such as well-being [17,23–25]. Here, we investigate the rela-

tionship between engagement in LAs and fundamental dimensions of successful aging such as

cognitive function, physical function, and mental health.

The present investigation

This study examines the impact of engagement in LAs on cognitive function, physical func-

tion, and mental health in a group (N = 809) of Japanese older adults. As most of the studies in
the field are based on samples from Western countries (e.g., US and Germany), little is known

about this issue in non-Western countries. The present study aims to fill this gap by adding

new information from a country of the Far East (i.e., Japan). Moreover, prior studies usually

suffer from limitations such as relying on subjective measures of cognitive and physical func-

tion [26,27] and dichotomization of continuous variables [27,28]. Also, the use of multiple-

indicator latent variables, which is necessary to reduce measurement error and, hence, pro-

duce more reliable estimates, has been sporadic [2]. The present study employs objective

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 2 / 13

Funding: This work was supported by the Japan

Society for the Promotion of Science [17F17313 to

GS; 17H02633 to YG], the Ministry of Education,

Culture, Sports, Science, and Technology of Japan

[21330152 to YG; 26310104 to YG], the Grant for

promoting Human Sciences research by Osaka

university Graduate school of Human Sciences and

the International Joint Research Promotion

Program at Osaka. The funders had no role in

study design, data collection and analysis, decision

to publish, or preparation of the manuscript.

Competing interests: The authors have declared

that no competing interests exist.

measures of functioning as well as multiple indicators allowing complex modeling and con-

trols for measurement error.

Furthermore, previous studies of both Western and Eastern populations have focused on

the association between LA engagement and other components of successful aging (e.g., cogni-

tive function) one at a time. By contrast, here we study the relationship between LA engage-

ment and cognitive function, physical function, and mental health in a single model. This

approach allows us to examine not only the direct effects of LA engagement on indicators of

successful aging, but also how cognitive function, physical function, and mental health influ-

ences each other. Recent research into the field has in fact shown that mutual relationships

occur between cognitive function, physical function, and mental health [17,29–31]. However,

these bi-directional effects have never been studied in relationship with LA engagement.

We aim to address the above limitations with Structural Equation Modeling (SEM). An

SEM model allows us to analyze the impact of LA engagement on all the variables of interest–

cognitive function, physical function, and mental health–simultaneously (i.e., in a single SEM

model). Such a model rules out potential confounds (e.g., Type I error due to a missing covari-

ate) and takes into account the bidirectional paths between the endogenous latent variables.

This way, the model estimates the unique (i.e., not confounded by other correlated variables)

impact of LA engagement on the other dimensions of successful aging. To the best of our

knowledge, no such study has been carried out in this field so far. Also, SEM models can esti-

mate latent constructs of interest (i.e., cognitive function, physical function, and mental

health), which are psychometrically more precise and reliable than observed indicators (i.e.,

test and questionnaires scores). Finally, while the studies in the field have usually assessed

physical function with self-report questionnaires [19], the indicators we use are objective

measures.

To sum up, the study aims (a) to test the previous claims about the relationship between LA

engagement and successful aging by employing a more robust and comprehensive modeling

approach; (b) to extend the current empirical evidence–which is mostly on cognitive func-

tion–by examining the impact of LA engagement on less studied dimensions of successful

aging such as physical function and mental health; (c) to quantify both direct and indirect

effects of LA engagement; and (d) to extend the relatively small amount of data concerning the

role played by LA engagement in the successful aging dimensions of Eastern populations.

Materials and methods

Participants

The study included a total of 809 Japanese participants (381 men and 428 women). The age

range was 72 to 74. The data were retrieved from the second wave of the Septuagenarians,

Octogenarians, Nonagenarians Investigation with Centenarians (SONIC) survey. These partic-

ular study cohort and wave were selected because they reported the most extensive and

detailed LA-engagement questionnaire.

The SONIC is an ongoing survey whose main purpose is to identify the correlates of healthy

aging. It includes both urban and rural (ratio 2:1) community-dwelling older adults. The par-

ticipants were recruited from residential registries and contacted by postal mail. They gave

their informed consent on site prior to starting the survey. For all the details about the SONIC

survey, see [32].

Variables

Leisure-activity (LA) engagement. This questionnaire included 158 yes/no items regard-

ing the participant’s engagement in as many activities. The questionnaire was based on the

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 3 / 13

questionnaires presented in Karp et al. and Jopp and Hertzog’s studies [4,33], and it was

extended by adding common activities among Japanese older adults (e.g., playing shogi, prac-

ticing tai chi, and going to a public bath [onsen]). The questionnaire showed good internal

consistency (Cronbach’s α = .88). The list of the activities is reported in the Supplemental
materials available online. A latent factor representing LA engagement was extracted from the

questionnaire with the ltm R package and used in the analyses [34].

Cognitive function. Three measures of cognitive ability were used to estimate a latent var-

iable representing cognitive function: the Japanese version of the Montreal Cognitive Assess-

ment (MoCA-J; [35]); the number series completion task from the Brief Test of Adult

Cognition by Telephone as a measure for reasoning skills [36]; and the recall subtest of the Alz-

heimer’s Disease Assessment Scale (ADAS; [37]). For all tests, a specific score was obtained,

with higher values indicating higher capacity.

Physical function. Two measures of physical function from the Short Physical Perfor-

mance Battery (SPPB; [38]) were administered: the participants’ gait speed and chair stand

test. For both tests, performance indicators were captured in seconds, with fewer seconds to

complete the task indicating better health functioning. The third measure was the 10-second

open-close stepping test [39], for which a smaller number of steps to complete the task indi-

cated better functional health.

Mental health. Three measures of mental health were administered: the Japanese versions

of the WHO-5 well-being index questionnaire [40], positive section of the Positive and Nega-

tive Affect Scales (PANAS; [41]), and Satisfaction With Life Scale [42]. In all the indicators,

higher numbers indicate better mental health.

Covariates. We included gender (male, female), education and wealth as covariates to

control for their potential effect on LA engagement or its link to the successful aging outcomes.

Education consisted of three levels indicating the highest degree achieved by the participant

(1 = primary/middle school, 2 = high school, and 3 = university/college education). Wealth

described the participant’s economic situation and included five levels (from 1 = difficult eco-

nomic situation to 5 = very good economic situation). These two variables were added to

assure that the effect of LA engagement was not confounded with SES-related variables such as

education and wealth status. Finally, gender was used as a grouping variable to test for mea-

surement invariance across males and females. The rationale of this choice is that SES variables

(such as education and wealth) may not always be equally good predictors of the constructs of

interests between males and females. For example, it is reasonable to suppose that, in the fifties,

intellectually gifted males were more likely to advance in their studies than equally intellectu-

ally gifted females.

Data preparation and analysis

Variables transformation and outlier treatment. All the continuous variables were nor-

malized and standardized. Normalization reduces the inflation of absolute measures of fit (e.g.,

χ2) and incidence of outliers in the dataset. Also, normalization tends to linearize the relation-
ship among multiple measures of the same construct [43], and thus reduces biases due to

potential non-linear relationships between indicators. The normalization was run with the

bestNormalize R package [44].

The normalized variables were inspected for possible outliers. A value was considered as

an outlier if it fell outside of the following range [Q1–2.2�IQR; Q3 + 2.2�IQR], where Q1,

Q3, and IQR were the first quartile, the third quartile, and the interquartile range of the vari-

able, respectively [45]. Only two outliers were detected (both in the MoCA-J scores) and

winsorized.

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 4 / 13

Power calculation. The statistical power for not-close fit hypothesis testing was estimated

[43]. Assuming the null RMSEA = .050, alternative RMSEA = .010, and α = .050, the statistical
power was very high (> 99%). The number of participants recruited was thus more than ade-

quate for global fit testing and rejection of false models. These analyses were run with the sem-

Tools R package [46].

SEM modeling

The lavaan R package [47] was used to run all the SEM models. An equality constraint between

two factors (physical function and mental health) was added to identify the models. We built

one model including all the variables. LA engagement, subjective wealth, and education were

predictors of the three latent variables cognition, physical function, and mental health in the

regression equations. Also, since some studies have provided evidence of a bidirectional rela-

tionship between physical function, cognitive function, and mental health [17,29,30], feedback

loops were included between the latent variables. Due to the inclusion of two ordinal variables

(education and subjective wealth), we chose the unweighted weighted least squares estimator

(ULS). (An additional set of analyses was carried out with an alternative estimator [WLSM]

that does not assume multivariate normality.) Gender was used for multiple-group analysis.

Results

The descriptive statistics (untransformed means and standard deviations) are reported in

Table 1.

The SEM model in which LA engagement predicted cognitive function, physical function,

and mental health as concurrent successful aging outcomes, and controlling for SES (i.e., edu-

cation and wealth), proved to have an excellent fit: χ2(79) = 70.34, RMSEA = 0.000,
SMRM = 0.037, CFI = 1.000. Comparative fit testing showed that this model outperformed the

homologous model without the grouping variable (i.e., gender; CFI = 1.000). Measurement

invariance analysis showed that the weak (metric) invariance was not rejected (p = .356). The
strong (i.e., scalar) invariance hypothesis was rejected (p < .001), suggesting the presence of a
differential additive response style [43] between genders. The model structure is in Fig 1.

The latent variables and regression coefficients are summarized in Table 2. LA engagement

was positively related to all the latent factors in both male and female participants (all ps �
.003). The size of the effects was moderate (Std.Est ranging between 0.185 and 0.306; Table 2),

Table 1. The descriptive statistics of the variables.

Males (N = 381) Females (N = 428) Total (N = 809)
Variable Mean SD Mean SD t-value p-value Mean SD Range
LA Engagement 21.64 10.45 22.47 10.49 -1.12 .263 22.08 10.48 0–65

MoCA-J 23.41 2.92 23.95 2.96 -2.58 .010 23.69 2.95 12–30

ADAS Recall 14.91 3.71 16.61 3.74 -6.49 .000 15.81 3.82 5–26

Number Series 0.11 1.08 -0.21 1.05 4.29 .000 -0.06 1.07 -2.59–2.00

Chair Stand Test (s) 10.88 3.37 10.14 2.96 3.36 .001 10.49 3.18 3.78–32.57

Gait Speed (s) 2.47 0.57 2.41 0.51 1.63 .104 2.44 0.54 1.36–5.69

Stepping Test 25.19 6.42 26.40 5.75 -2.82 .005 25.82 6.10 6–46

WHO5 15.63 5.09 16.17 4.85 -1.54 .123 15.91 4.97 0–25

Positive Affect 10.38 2.76 11.29 2.57 -4.83 .000 10.86 2.70 3–15

Life Satisfaction 23.24 5.72 23.03 5.38 0.52 .602 23.13 5.54 5–35

Education 2.17 0.73 1.98 0.67 3.78 .000 2.07 0.71 1–3

Wealth 2.94 0.81 2.91 0.79 0.55 .584 2.92 0.80 1–5

https://doi.org/10.1371/journal.pone.0225006.t001

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 5 / 13

yet comparable or even superior to the one of education and wealth. Also, the results showed a

statistically significant reciprocal effect between cognitive function and physical function (Std.

Est ranging between 0.113 and 0.201). No significant effect was observed between mental

health and physical function or cognitive function. Finally, no meaningful difference was

observed with the WLSM estimator (see the OSF link). Fig 2 highlights the statistically signifi-

cant paths in the SEM model.

Discussion

This paper quantified the impact of LA engagement on measures of successful aging including

cognitive function, physical function, and mental health in a sample of Japanese older adults.

LA engagement appears to be positively related to all these three constructs in both men and

women (all ps � .003). The implementation of a comprehensive SEM model allowing to con-
sider different successful aging indicators concurrently, the excellent goodness of fit, the high

statistical power, and the implementation of latent constructs based on tasks measuring objec-

tive performance (cognitive function and physical function) and reliable self-reported ques-

tionnaires (mental health) make the findings of the present study more reliable than most of

previous research.

Substantive findings

Overall, the present results corroborate the idea that LA engagement contributes to explaining

the individual differences in cognitive function, physical function, and mental health. LA

engagement was positively associated with successful aging indicators. The results are in line

with the “use it or lose it” theoretical framework and previous findings in Western populations

[3,13,48]. Thus, the findings of the present study support the position that leading an active

lifestyle, here assessed by engagement in leisure activities, is a universal and culture-indepen-

dent means contributing to successful aging, invariant across different countries and cultures.

Fig 1. The SEM model. The squares and rectangles represent observed variables. The rectangles represent indicators

for the latent variables (circles). The arrows represent the paths.

https://doi.org/10.1371/journal.pone.0225006.g001

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 6 / 13

Inspection of the regression standardized path coefficients (Std.Est column in the Table 2)

indicates that the impact of LA engagement on the three latent variables is statistically signifi-

cant and comparable (or even superior; range of Std.Est 0.185–0.306) to the one exerted by

education and self-perceived wealth. Furthermore, the impact of LA engagement is systemati-

cally superior to the one exerted by the three endogenous variables on each other (range

-0.096–0.201). Nonetheless, the size of these effects is somewhat smaller than the ones reported

in those few studies implementing an SEM approach (e.g., 0.490; [3]). This discrepancy is

probably due to the inclusion of three endogenous latent variables connected to each other

with feedback loops in a single non-recursive model. Such a model controls for the potential

confounding effects of one latent variable on the others. Part of the variance that it would be

Table 2. The results of the SEM model.

Males Females
Latent variables

Estimate Std.Err p-value Std.Est Estimate Std.Err p-value Std.Est
Cognitive Function

MoCA-J 1.000 0.672 1.000 0.596

ADAS Recall 0.854 0.104 .000 0.587 0.920 0.182 .000 0.552

Number Series 0.590 0.081 .000 0.405 0.603 0.126 .000 0.353

Physical Function
Chair Stand Test 1.000 0.589 1.000 0.646

Gait Speed 0.867 0.104 .000 0.498 0.836 0.122 .000 0.547

Stepping Test 1.522 0.190 .000 0.854 1.082 0.164 .000 0.732

Mental Health
WHO5 1.000 0.727 1.000 0.696

Positive Affect 1.103 0.115 .000 0.808 1.079 0.150 .000 0.764

Life Satisfaction 0.997 0.102 .000 0.719 0.882 0.115 .000 0.617

Regressions R2 R2

Cognitive Function .427 .204
LA Engagement 0.219 0.063 .000 0.306 0.163 0.054 .003 0.253

Physical Function 0.228 0.047 .000 0.201 0.160 0.042 .000 0.174

Mental Health 0.015 0.048 .745 0.017 -0.082 0.038 .034 -0.096

Education 0.267 0.088 .002 0.292 0.131 0.081 .108 0.150

Wealth 0.081 0.049 .096 0.103 0.041 0.064 .527 0.055

Physical Function .199 .132
LA Engagement 0.134 0.045 .003 0.212 0.172 0.046 .000 0.246

Cognitive Function 0.133 0.036 .000 0.151 0.123 0.030 .000 0.113

Mental Health
a
-0.003 0.026 .896 -0.004 0.031 0.027 .251 0.034

Education 0.030 0.062 .621 0.038 0.011 0.071 .878 0.011

Wealth 0.093 0.049 .057 0.127 -0.003 0.055 .952 -0.004

Mental Health .214 .114
LA Engagement 0.203 0.053 .000 0.259 0.140 0.045 .002 0.185

Cognitive Function 0.002 0.030 .945 0.002 -0.070 0.025 .005 -0.059

Physical Function
a
-0.003 0.026 .896 -0.004 0.033 0.024 .165 0.031

Education 0.008 0.075 .915 0.008 0.008 0.072 .912 0.008

Wealth 0.297 0.058 .000 0.327 0.192 0.057 .000 0.220

Note. Estimate = unstandardized path coefficient; Std.Err = standard error; p-value = significance level; Std.Est = standardized path coefficient.
a

An equality constraint was applied to identify the model.

https://doi.org/10.1371/journal.pone.0225006.t002

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 7 / 13

intercepted by LA engagement is instead absorbed by the paths connecting the three latent var-

iables. Simply put, the model controls, for example, that the positive relationship between LA

engagement and cognitive function does not only stem from better mental health or physical

function. That being said, overall, LA engagement appears to contribute significantly to

explaining unique variance in individual differences in important aspects of successful aging.

Some indirect effects were observed for LA engagement, specifically when considering cog-

nitive and physical function. In particular, cognitive function and physical function influenced

each other in a feedback loop, which supports recent evidence showing that the relationship

between cognition and physical health is bidirectional [29,30]. LA engagement thus seems to

exert an indirect effect on cognitive function mediated by physical function and vice versa. On

this regard, the benefits of physical activity on cognitive function can be attributed to an ame-

liorated overall health condition (e.g., brain oxygenation and stimulation of neurogenesis

[49]). Why better cognition should lead to improved physical health is less obvious. A possible

explanatory mechanism could be an overall better self-regulation enabling the participant to

be more physically active [29]. However, due to the correlational nature of the present study,

no causal inference in either direction can be made and, thus, none of these hypotheses can be

tested.

By contrast, mental health was not associated with either of the other two latent variables.

In particular, the absence of any link between physical function and mental health appears to

contradict previous findings in the field [17,50]. Possibly, this discrepancy is explained by the

use of subjective measures of physical function in prior studies. In fact, while self-reported

measures of physical function and mental health may reflect individuals’ perception of their

Fig 2. Representation of the significant paths in the SEM model. The numbers indicate the standardized path

coefficients in males and females, respectively. Non-significant paths in either males or females and the indicators of

the latent variables (all significant) are omitted for the sake of clarity.

https://doi.org/10.1371/journal.pone.0225006.g002

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 8 / 13

overall health condition, objective measures of physical function may not be correlated with

subjective measures of mental health. It is further of note that we used exclusively balance and

gait measures as indicators of physical function, which may also have contributed to the lack

of relation between our physical and mental health latent construct. That being said, the topic

is certainly worth of further investigation.

Finally, there is no clear evidence that the effect of LA engagement on the three latent con-

structs differed between males and females. The relevant standardized path coefficients were

relatively homogeneous when comparing the patterns across genders. A notable difference

between the models for males and females was, however, the amount of explained variance,

which was nearly twice as large in males compared to females. This is probably because the

covariates were associated with greater factor loadings in males than in females. Basically, con-

sidering gender through multi-group testing improved the goodness of fit but had no notable

impact on the relationship between LA engagement and the endogenous variables (i.e., cogni-

tive function, physical function, and mental health).

Recommendations for future research

In this study, we have considered LA engagement as a single trait. This assumption is upheld

by the high internal consistency of our LA questionnaire (Cronbach’s α = .88) and is in line
with substantial findings in the field [3]. Nonetheless, it is possible that engagement in particu-

lar types of activities (e.g., playing strategy games, technology use) is more strongly linked to

specific constructs such as cognitive function [10,11,33,48]. In future studies, it is thus recom-

mendable to test whether overall LA engagement subsumes specific types of activities impact-

ing differently on the examined constructs, or whether specific activities are more important

for certain successful aging aspects than others. The main drawback of this approach is that

requires larger sample sizes than the ones usually included in surveys. Reducing the number of

activities in the questionnaire is an alternative [2] but it has the serious shortcoming of

decreasing the overall reliability of the measure. In fact, some individuals may practice rela-

tively unusual LAs that would be necessarily excluded in a shorter questionnaire.

Another point of interest is the frequency of LA engagement. The sheer number of activities

practiced by older adults is certainly a good proxy for their engagement in LA, but it is not nec-

essarily the best possible one. How often an activity is practiced may play a significant role too.

A further improvement may thus be, for instance, the use of a Likert scale to assess the fre-

quency with which the participants engage in the activities they practice [3,33]. It is worth not-

ing, however, that including frequency may not meaningfully enhance the predictive power of

the measure. In fact, previous research suggests that most health benefits of an active life-style

in the elderly occur after only a moderate amount of engagement [51].

Finally, a fundamental caveat needs to be mentioned. Our analysis is correlational in nature,

and thus cannot establish any direction of causality between LA engagement and measures of

successful aging. In our opinion, there are three possible explanations for our results: (a) LA

engagement causes improvements in cognitive function, physical function, and mental health;
(b) people with superior cognitive function, physical function, and mental health, are more

likely to engage in LA; and (c) LA and the three constructs influence each other, that is, LA

engagement positively affects measures of successful aging which, in turn, promotes LA

engagement. This latter possibility is, in our opinion, the most likely explanation. That being

said, studies implementing an LA intervention are necessary to test this hypothesis. Specifi-

cally, a longitudinal non-recursive SEM model would allow researchers to assess this presumed

“virtuous circle” between LA engagement and measures of successful aging by imposing a

feedback loop between the variables at different time points.

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 9 / 13

Conclusions

The present study reported an SEM model examining the relationship between LA engage-

ment and three essential dimensions of successful aging (i.e., cognitive function, physical func-

tion, and mental health) in a large sample of Japanese older adults. In line with substantial

research into the field, the results confirmed the link between LA engagement and cognitive

function. However, the size of the effect was meaningfully smaller than the one reported in

previous studies. Similar effects were found for physical function and mental health.

The investigation significantly extends our knowledge in the field. First, thanks to a more

comprehensive modeling approach, the study provides a more reliable estimate of the impact

of LA engagement on cognitive function. Second, due to the use of multiple and objective

physical indicators, it adds much-needed evidence of a link between LA engagement and pres-

ervation of good physical function in the elderly. Similar considerations apply to the influence

of LA engagement in older adults’ mental health. Third, the findings suggest that the role

played by LA engagement is cultural-independent. Finally, our study sheds some light on

mechanisms of LA engagement that have not been much (if not at all) investigated so far, such

as the bidirectional effects of practicing LAs on cognitive function and physical function.

Acknowledgments

We gratefully thank all the participants for their time and effort. We also thank Fred Oswald

and Yves Rosseel for their assistance in the statistical analysis.

Author Contributions

Conceptualization: Giovanni Sala, Yasuyuki Gondo.

Data curation: Giovanni Sala, Madoka Ogawa, Yoshiko Ishioka, Yukie Masui, Hiroki Inagaki,

Takeshi Nakagawa, Saori Yasumoto, Tatsuro Ishizaki, Yasumichi Arai, Kazunori Ikebe, Kei

Kamide, Yasuyuki Gondo.

Formal analysis: Giovanni Sala.

Funding acquisition: Yasuyuki Gondo.

Investigation: Madoka Ogawa, Yoshiko Ishioka, Yukie Masui, Hiroki Inagaki, Takeshi Naka-

gawa, Saori Yasumoto, Tatsuro Ishizaki, Yasumichi Arai, Kazunori Ikebe, Kei Kamide,

Yasuyuki Gondo.

Methodology: Yasuyuki Gondo.

Project administration: Yukie Masui, Yasuyuki Gondo.

Software: Yasuyuki Gondo.

Writing – original draft: Giovanni Sala, Daniela Jopp, Fernand Gobet, Yasuyuki Gondo.

References
1. Pressman SD, Matthews KA, Cohen S, Martire LM, Scheier M, Baum A, et al. Association of enjoyable

leisure activities with psychological and physical well-being. Psychosom Med. 2009; 71: 725–732.

https://doi.org/10.1097/PSY.0b013e3181ad7978 PMID: 19592515

2. Gow AJ, Pattie A, Deary IJ. Lifecourse activity participation from early, mid, and later adulthood as

determinants of cognitive aging: The lothian birth cohort 1921. Journals Gerontol—Ser B Psychol Sci

Soc Sci. 2017; 72: 25–37. https://doi.org/10.1093/geronb/gbw124 PMID: 27974473

3. Jopp D, Hertzog C. Activities, self-referent memory beliefs, and cognitive performance: Evidence for

direct and mediated relations. Psychol Aging. 2007; 22: 811–825. https://doi.org/10.1037/0882-7974.

22.4.811 PMID: 18179299

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 10 / 13

4. Karp A, Paillard-Borg S, Wang HX, Silverstein M, Winblad B, Fratiglioni L. Mental, physical and social

components in leisure activities equally contribute to decrease dementia risk. Dement Geriatr Cogn Dis-

ord. 2006; 21: 65–73. https://doi.org/10.1159/000089919 PMID: 16319455

5. Yates LA, Ziser S, Spector A, Orrell M. Cognitive leisure activities and future risk of cognitive impairment

and dementia: Systematic review and meta-analysis. Int Psychogeriatrics. 2016; 28: 1791–1806.

https://doi.org/10.1017/S1041610216001137 PMID: 27502691

6. Strobach T, Karbach J. Cognitive training: An overview of features and applications. New York:

Springer; 2016. https://doi.org/10.1007/978-3-319-42662-4

7. Melby-Lervåg M, Redick TS, Hulme C. Working memory training does not improve performance on
measures of intelligence or other measures of “far transfer”: Evidence from a meta-analytic review. Per-

spect Psychol Sci. 2016; 11: 512–534. https://doi.org/10.1177/1745691616635612 PMID: 27474138

8. Sala G, Aksayli ND, Tatlidil KS, Tatsumi T, Gondo Y, Gobet F. Near and far transfer in cognitive training:

A second-order meta-analysis. Collabra Psychol. 2019; 5: 18. https://doi.org/10.1525/collabra.203

9. Simons DJ, Boot WR, Charness N, Gathercole SE, Chabris CF, Hambrick DZ, et al. Do “Brain-Training”

programs work? Psychol Sci Public Interes. 2016; 17: 103–186. https://doi.org/10.1177/

1529100616661983 PMID: 27697851

10. Burgoyne AP, Sala G, Gobet F, Macnamara BN, Campitelli G, Hambrick DZ. The relationship between

cognitive ability and chess skill: A comprehensive meta-analysis. Intelligence. 2016; 59: 72–83. https://

doi.org/10.1016/j.intell.2016.08.002

11. Sala G, Burgoyne AP, Macnamara BN, Hambrick DZ, Campitelli G, Gobet F. Checking the “Academic

Selection” argument. Chess players outperform non-chess players in cognitive skills related to intelli-

gence: A meta-analysis. Intelligence. 2017; 61: 130–139. https://doi.org/10.1016/j.intell.2017.01.013

12. Kramer AF, Bherer L, Colcombe SJ, Dong W, Greenough WT. Environmental influences on cognitive

and brain plasticity during aging. Journals Gerontol Ser A Biol Sci Med Sci. 2004; 59: M940–M957.

https://doi.org/10.1093/gerona/59.9.m940 PMID: 15472160

13. Salthouse TA. Mental exercise and mental aging: Evaluating the validity of the “Use It or Lose It” hypoth-

esis. Perspect Psychol Sci. 2006; 1: 68–87. https://doi.org/10.1111/j.1745-6916.2006.00005.x PMID:

26151186

14. Ghisletta P, Bickel JF, Lövdén M. Does activity engagement protect against cognitive decline in old

age? Methodological and analytical considerations. Journals Gerontol—Ser B Psychol Sci Soc Sci.

2006; 61: P253–P261. https://doi.org/10.1093/geronb/61.5.P253 PMID: 16960228

15. Lövdén M, Ghisletta P, Lindenberger U. Social participation attenuates decline in perceptual speed in

old and very old age. Psychol Aging. 2005; 20: 423–434. https://doi.org/10.1037/0882-7974.20.3.423

PMID: 16248702

16. Fuertes E, Carsin AE, Antó JM, Bono R, Corsico AG, Demoly P, et al. Leisure-time vigorous physical

activity is associated with better lung function: The prospective ECRHS study. Thorax. 2018; 73: 376–

384. https://doi.org/10.1136/thoraxjnl-2017-210947 PMID: 29306902

17. Paggi ME, Jopp D, Hertzog C. The importance of leisure activities in the relationship between physical

health and well-being in a life span sample. Gerontology. 2016; 62: 450–458. https://doi.org/10.1159/

000444415 PMID: 26974682

18. Colcombe S, Kramer AF. Fitness effects on the cognitive function of older adults: A meta-analytic study.

Psychol Sci. 2003; 14: 125–130. https://doi.org/10.1111/1467-9280.t01-1-01430 PMID: 12661673

19. Menec VH. The relation between everyday activities and successful aging: A 6-year longitudinal study.

Journals Gerontol—Ser B Psychol Sci Soc Sci. 2003; 58B: S74–S82. https://doi.org/10.1093/geronb/

58.2.S74 PMID: 12646596

20. Stern C, Konno R. Cognitive leisure activities and their role in preventing dementia: A systematic review.

Int J Evid Based Healthc. 2010; 8: 2–17. https://doi.org/10.1111/j.1744-1609.2010.00150.x PMID:

20923507

21. Rowe JW, Kahn RL. Human aging: usual and successful. (physiological changes associated with

aging). Science (80-). 1987; 237: 143–149. https://doi.org/10.1126/science.3299702 PMID: 3299702

22. Havighurst R, Neugarten B, Tobin S. Disengagement and patterns of aging. In: Neugarten B, editor.

Middle Age and Aging. Chicago: University of Chicago Press; 1968. pp. 161–172.

23. Chao SF. Changes in leisure activities and dimensions of depressive symptoms in later life: A 12-year

follow-up. Gerontologist. 2016; 56: 397–407. https://doi.org/10.1093/geront/gnu052 PMID: 24986219

24. Jopp DS, Wozniak D, Damarin AK, De Feo M, Jung S, Jeswani S. How could lay perspectives on suc-

cessful aging complement scientific theory? Findings from a U.S. and a German life-span sample. Ger-

ontologist. 2015; 55. https://doi.org/10.1093/geront/gnu059 PMID: 24958719

25. Pruchno RA, Wilson-genderson M, Cartwright F. A two-factor model of successful aging. J Gerontol

Psychol Sci. 2010; 65: 671–679. https://doi.org/10.1093/geronb/gbq051 PMID: 20624759

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 11 / 13

26. Hsu HC, Jones BL. Multiple trajectories of successful aging of older and younger cohorts. Gerontologist.

2012; 52: 843–856. https://doi.org/10.1093/geront/gns005 PMID: 22403165

27. Iwasa H, Yoshida Y, Kai I, Suzuki T, Kim H, Yoshida H. Leisure activities and cognitive function in

elderly community-dwelling individuals in Japan: A 5-year prospective cohort study. J Psychosom Res.

2012; 72: 159–164. https://doi.org/10.1016/j.jpsychores.2011.10.002 PMID: 22281459

28. Dodge HH, Kita Y, Takechi H, Hayakawa T, Ganguli M, Ueshima H. Healthy cognitive aging and leisure

activities among the oldest old in Japan: Takashima study. Journals Gerontol—Ser A Biol Sci Med Sci.

2008; 63: 1193–1200. https://doi.org/10.1093/gerona/63.11.1193 PMID: 19038834

29. Best JR, Nagamatsu LS, Liu-Ambrose T. Improvements to executive function during exercise training

predict maintenance of physical activity over the following year. Front Hum Neurosci. 2014; 8: 353.

https://doi.org/10.3389/fnhum.2014.00353 PMID: 24904387

30. Daly M, McMinn D, Allan JL. A bidirectional relationship between physical activity and executive function

in older adults. Front Hum Neurosci. 2015; 8: 1044. https://doi.org/10.3389/fnhum.2014.01044 PMID:

25628552

31. Kobayashi-Cuya KE, Sakurai R, Suzuki H, Ogawa S, Takebayashi T, Fujiwara Y. Observational evi-

dence of the association between handgrip strength, hand dexterity, and cognitive performance in com-

munity-dwelling older adults: A systematic review. J Epidemiol. 2018; 28: 373–381. https://doi.org/10.

2188/jea.JE20170041 PMID: 29526916

32. Gondo Y, Masui Y, Kamide K, Ikebe K, Arai Y, Ishizaki T. SONIC Study: A Longitudinal Cohort Study of

the Older People as Part of a Centenarian Study. Encyclopedia of Geropsychology. Singapore:

Springer Singapore; 2016. pp. 1–10. https://doi.org/10.1007/978-981-287-080-3_182–1

33. Jopp DS, Hertzog C. Assessing adult leisure activities: An extension of a self-report activity question-

naire. Psychol Assess. 2010; 22: 108–120. https://doi.org/10.1037/a0017662 PMID: 20230157

34. Rizopoulos D. ltm: An R Package for Latent Variable Modeling. Journal of Statistical Software. 2006.

35. Suzuki H, Kawai H, Hirano H, Yoshida H, Ihara K, Kim H, et al. One-year change in the Japanese ver-

sion of the Montreal Cognitive Assessment performance and related predictors in community-dwelling

older adults. J Am Geriatr Soc. 2015; 63: 1874–1879. https://doi.org/10.1111/jgs.13595 PMID:

26313522

36. Lachman ME, Agrigoroaei S, Tun PA, Weaver SL. Monitoring cognitive functioning: psychometric prop-

erties of the brief test of adult cognition by telephone. Assessment. 2014; 21: 404–417. https://doi.org/

10.1177/1073191113508807 PMID: 24322011

37. Rosen WG, Mohs RC, Davis KL. A new rating scale for Alzheimer’s disease. Am J Psychiatry. 1984;

141: 1356–1364. https://doi.org/10.1176/ajp.141.11.1356 PMID: 6496779

38. Puthoff ML. Outcome measures in cardiopulmonary physical therapy: Short physical performance bat-

tery. Cardiopulm Phys Ther J. 2008; 19: 17–22. PMID: 20467494

39. Kobayashi K, Hiiragi Y, Maruyama H. Development of the “10-second Open-Close Stepping Test”

(OCS-10) and fundamental study of its measurement values through a comparison of healthy young

people and community-dwelling elderly. J Phys Ther Sci. 2012; 24: 747–749. https://doi.org/10.1589/

jpts.24.747

40. Awata S, Bech P, Koizumi Y, Seki T, Kuriyama S, Hozawa A, et al. Validity and utility of the Japanese

version of the WHO-Five Well-Being Index in the context of detecting suicidal ideation in elderly commu-

nity residents. Int Psychogeriatrics. 2007; 19: 77–88. https://doi.org/10.1017/S1041610206004212

PMID: 16970832

41. Nakahara J. Deliberation on factor structure on affective well-being scale and the development of its

short version. Japanese J Gerontol. 2011; 4: 434–442.

42. Diener E, Emmons RA, Larsen RJ, Griffin S. The Satisfaction With Life Scale. J Pers Assess. 1985; 49:

71–75. https://doi.org/10.1207/s15327752jpa4901_13 PMID: 16367493

43. Kline., Kline R. RB (2015). P and practice of structural equation modeling. G publications. B. Principles

and practice of structural equation modeling. 4th ed. Guilford publications. New York: The Guildford

Press; 2015.

44. Peterson R. Estimating normalization transformations with bestNormalize. 2017.

45. Hoaglin DC, Iglewicz B. Fine-tuning some resistant rules for outlier labeling. J Am Stat Assoc. 1987; 82:

1147–1149. https://doi.org/10.1080/01621459.1987.10478551

46. Terrence J, Pornprasertmanit S, Alexander S, Rosseel Y. semTools: Useful tools for structural equation

modeling. R package version 0.5–1. 2018.

47. Rosseel Y. lavaan: An R Package for Structural Equation Modeling. J Stat Softw. 2015; 48: 1–36.

https://doi.org/10.18637/jss.v048.i02

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 12 / 13

48. Hultsch DF, Hertzog C, Small BJ, Dixon RA. Use it or lose it: Engaged lifestyle as a buffer of cognitive

decline in aging? Psychol Aging. 1999; 14: 245–263. https://doi.org/10.1037//0882-7974.14.2.245

PMID: 10403712

49. Kempermann G, Fabel K, Ehninger D, Babu H, Leal-Galicia P, Garthe A, et al. Why and how physical

activity promotes experience-induced brain plasticity. Front Neurosci. 2010; 4. https://doi.org/10.3389/

fnins.2010.00189 PMID: 21151782

50. Tsaousis I, Nikolaou I, Serdaris N, Judge TA. Do the core self-evaluations moderate the relationship

between subjective well-being and physical and psychological health? Pers Individ Dif. 2007; 42: 1441–

1452. https://doi.org/10.1016/j.paid.2006.10.025

51. Andersen LB, Schnohr P, Schroll M, Hein HO. All-Cause mortality associated with physical activity dur-

ing leisure time, work, sports, and cycling to work. Arch Intern Med. 2000; 160: 1621–1628. https://doi.

org/10.1001/archinte.160.11.1621 PMID: 10847255

Leisure activities and successful aging

PLOS ONE | https://doi.org/10.1371/journal.pone.0225006 November 8, 2019 13 / 13

Copyright of PLoS ONE is the property of Public Library of Science and its content may not
be copied or emailed to multiple sites or posted to a listserv without the copyright holder’s
express written permission. However, users may print, download, or email articles for
individual use.