Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
University of Groningen
Physical activity and depressive symptomsStavrakakis, Nikolaos
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.
Document VersionPublisher's PDF, also known as Version of record
Publication date:2015
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):Stavrakakis, N. (2015). Physical activity and depressive symptoms: is a healthy body necessary for ahealthy mind?. [S.n.].
CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.
Download date: 26-05-2021
Physical Activity and Depressive Symptoms
Is a healthy body necessary for a healthy mind?
Nikolaos Stavrakakis
Physical activity and depressive symptoms
Is a healthy body necessary for a healthy mind?
PhD thesis
to obtain the degree of PhD at the
University of Groningen on the authority of the
Rector Magnificus Prof. E. Sterken and in accordance with
the decision by the College of Deans.
This thesis will be defended in public on
Monday 9 March 2015 at 16.15 hours
by
Nikolaos Stavrakakis
born on 23 June 1981 in Cholargos, Greece
The infrastructure for the TRacking Adolescents' Individual Lives Survey (www.TRAILS.nl) has been financially supported by grants from the Netherlands Organization for Scientific Research NWO (Medical Research Council program grant GB-MW 940-38-011; ZonMW Brainpower grant 100-001-004; ZonMw Risk Behaviour and Dependence grants 60-60600-97-118; ZonMw Culture and Health grant 261-98-710; Social Sciences Council medium-sized investment grants GB-MaGW 480-01-006 and GB-MaGW 480-07-001; Social Sciences Council project grants GB-MaGW 452-04-314 and GB-MaGW 452-06-004; NWO large-sized investment grant 175.010.2003.005; NWO Longitudinal Survey and Panel Funding 481-08-013; NWO Vici 016.130.002; NWO Gravitation 024.001.003), the Dutch Ministry of Justice (WODC), the European Science Foundation (EuroSTRESS project FP-006), Biobanking and Biomolecular Resources Research Infrastructure BBMRI-NL (CP 32), the Gratama foundation, the Jan Dekker foundation, the participating universities, and Accare Centre for Child and Adolescent Psychiatry.
Financial support for the publication of this dissertation was kindly provided by the research school of Behavioural and Cognitive Neurosciences (BCN), the Rijksuniversiteit Groningen and University Medical Center Groningen.
Design and printing METROPOLIS Graphic Arts S.A., Athens, Greece
© 2015 Nikolaos Stavrakakis, Athens, Greece
All rights reserved. No part of this dissertation may be reproduced, stored or transmitted in any form or by any means without the written permission of the author.
ISBN (print): 978-90-367-7600-4ISBN (digital): 978-90-367-7599-1
Physical Activity and Depressive SymptomsIs a healthy body necessary for a healthy mind?
PhD thesis
This thesis will be defended in public on
Monday 9 March 2015 at 16.15 hours
by
Nikolaos Stavrakakis
born on 23 June 1981in Cholargos, Greece
to obtain the degree of PhD at theUniversity of Groningenon the authority of the
Rector Magnificus Prof. E. Sterkenand in accordance with
the decision by the College of Deans.
Physical activity and depressive symptoms
Is a healthy body necessary for a healthy mind?
PhD thesis
to obtain the degree of PhD at the
University of Groningen on the authority of the
Rector Magnificus Prof. E. Sterken and in accordance with
the decision by the College of Deans.
This thesis will be defended in public on
Monday 9 March 2015 at 16.15 hours
by
Nikolaos Stavrakakis
born on 23 June 1981 in Cholargos, Greece
SupervisorsProf A.J. Oldehinkel
Prof P. de Jonge
Co-supervisorDr. A.M. Roest
Assessment committeeProf A. Steptoe
Prof B.W. Penninx
Prof S.A. Reijneveld
ParanymphsBertus F. Jeronimus
Anna-Roos E. Zandstra
Table of Contents
CHAPTER I General Introduction 9
CHAPTER II Bidirectional Prospective Associations between Physical 27 Activity and Depressive Symptoms. The TRAILS Study
CHAPTER III Physical Activity and Onset of Depression in Adolescents: 43 A Prospective Study in the General Population Cohort TRAILS
CHAPTER IV Plasticity Genes do not Modify Associations between Physical 55 Activity and Depressive Symptoms
CHAPTER V Competitive Sport Participation, Sport Competence and 73 Depressive Symptoms in Adolescent Boys and Girls
CHAPTER VI Relative-Age Effects in Adolescents: The TRAILS Study 99
CHAPTER VII The Dynamic Relationship between Physical Activity and 127 Mood in the Daily Life of Depressed and Non-Depressed Individuals: a Within-Subject Time-Series Analysis
CHAPTER VIII Summary and General Discussion 147
Summary in English 171
Summary in Dutch 177
Acknowledgements 183
About the Author 187
List of Publications 191
Recent TRAILS Dissertations 195
“Mens sana in corpore sano”(A healthy mind existing in a healthy body)
Juvenal, Satire X
General Introduction
CHAPTER
13GENERAL INTRODUCTION
GENERAL INTRODUCTION
Physical activity (PA) is considered to be beneficial to general health. However, the benefits
of PA on mental health, and especially depression are less clear. This thesis explores in
depth the relationship between PA and depressive symptoms, in adolescents and adults.
Definition of Physical Activity and Depression
It is important to clarify some of the terms that will be used in this thesis, namely, PA, physi-
cal fitness (PF)1, exercise, sports and depression.
Physical Activity, Physical Fitness, Exercise and Sports
The main two distinctive groups are PA and PF, while exercise and sports are subgroups of
PA. PF is defined as “the ability to carry out daily tasks with vigor and alertness, without
undue fatigue and with ample energy to enjoy leisure-time pursuits and to meet unforeseen
emergencies” (Caspersen et al., 1985; see p. 128). Hence, PF reflects health attributes (for
example, cardiovascular endurance, muscular endurance and strength, flexibility), as well as
skills related to athletic ability (accuracy, spatio-temporal coordination, etc.). PA is defined
as “any bodily movement produced by skeletal muscles that results in energy expenditure”
(Caspersen et al., 1985; see p. 126). The energy expenditure is commonly measured in kilo-
calories (kcal) or kilojoules (kJ). PA can be classified in many different ways. The simplest
way to classify PA is by distinguishing between activity while working and activity during
leisure-time (voluntary activity)2. Further classifications include light, moderate, and inten-
sive activities, team activities, and outdoor activities. The intensity of each activity can be
expressed in the Metabolic Equivalent of Tasks (MET). The MET score assesses the energy
cost of PA and is defined as “the ratio of the associated metabolic rate for the specific activ-
ity divided by the resting metabolic rate (RMR)3” (Ainsworth et al., 1993; see page 72)4.
Although PA and exercise share common characteristics and have been used interchange-
ably in the literature, exercise is a distinct subcomponent of PA. Exercise is defined as “any
physical activity that is planned, structured, repetitive, and purposive in the sense that
improvement or maintenance of one or more components of physical fitness is (the) objec-
tive” (Caspersen et al., 1985; see page 128). The term ‘sports’ does not have a universal
definition, but the majority of definitions on what constitutes a sport tend to converge on a
competitive element. The Oxford dictionary defines sport as: “an activity involving physical
exertion and skill in which an individual or team competes against another or others for
entertainment”5. Thus, the term ‘sports’ is also a subcomponent of PA and shares com-
mon traits with exercise.
CHAPTER I14
Depression
Depression has been defined in diverse ways in the literature (Abela & Hankin, 2008) and is
considered a largely heterogeneous disorder (Hasler, Drevets, Manji, & Charney, 2004; Lux &
Kendler, 2010). Generally, depression is characterized as a state of low mood or a loss of
interest, which affects an individual’s well-being, cognition and behavior (Joorman, 2009).
According to DSM-V criteria for major depressive disorder (MDD), at least one of the symp-
toms of depressed mood and loss of interest or pleasure has to be present for 2 weeks. At
least five of the following nine additional symptoms should also persist almost every day
for at least 2 weeks: a) appetite problems, that is, weight loss or weight gain, or decrease
or increase in appetite; b) sleep difficulties, such as hyper-insomnia or insomnia; c) fatigue
or loss of energy; d) feelings of worthlessness; e) psychomotor agitation; f) cognitive prob-
lems, such as diminished ability to concentrate and indecisiveness; and g) suicidal ideation
or suicide attempts (American Psychiatric Association, 2013). The prevalence rates of MDD
are higher than of all other psychiatric disorders, with approximately 19% of the Dutch
population experiencing a clinically significant episode at least once in their lives6 (de Graaf,
ten Have, van Gool, & van Dorsselaer, 2012).
Humans Evolved to Move
PA has been at the core of every free-living organism’s evolutionary history. The rela-
tionship between caloric acquisition (energy consumed as food) and caloric expenditure,
i.e., PA, has exerted an “ongoing adaptive pressure that affected (the) selection of genes
related to the cardio-respiratory and musculoskeletal systems as well as the internal
metabolism processes of our progenitors” (Eaton & Eaton, 2003; see p. 153). This can be
seen in the evolution of the human species, where the environmental pressures that have
acted in order to produce our unique human nature have taken place immediately after the
‘split’ between our pre-human ancestors with those of the modern day apes (pongids),
an estimated 5-7 million years ago (Cordain et al., 1998; Eaton & Konner, 1985). The ap-
pearance of the Australopithecines7, approx. 4 million years ago8 coincided with gradual
environmental changes in the African continent, which resulted in a ‘drying’ up of tropical
forests to a more open savanna (Behrensmeyer & Cooke, 1985; deMonocal & Bloemen-
dal, 1995; Foley, 1987). The early hominins developed specific structural and functional
characteristics, which helped them exploit their changing environment in ways that their
pongid ancestors could not (see for example: Leonard & Robertson, 1997). The first such
change was the ability to walk straight (bipedalism9), which was still relatively primitive
compared to the abilities of (much) later hominins (i.e., Homo habilis, Homo erectus and
Homo sapiens; McΗenry, 1994).
15GENERAL INTRODUCTION
Bipedalism conferred clear advantages in this new environment, such as facilitated visual
detection of food, water and predators (Shipman, 1986), as well as enabling weapon or tool
use while walking (Cordain et al., 1997; Lovejoy, 1988). The biggest advantage, however, of
bipedalism is considered to be the development of a more complex eccrine sweat gland sys-
tem, which in turn allowed our early ancestors to walk or run over larger distances in hot
climates without overheating (Cordain et al., 1997; Wheeler, 1991)10. Bipedalism paved the
way for further changes that are seen with the rise of the genus Homo. Homo habilis had a
body size similar to the Australopithecines, but with a markedly increased cranial capacity11
(Cordain et al., 1997; Lieberman, 2011; Raichlen & Polk, 2013)12. Homo habilis was succeeded
by Homo erectus who showed more similarities in body size, physique and cranial capacity
to modern humans13 (Brown, Harris, Leakey, & Walker, 1985; McΗenry, 1994; Ruff, 1991). It
is estimated that Homo erectus had better aerobic capacities than the pongids and early
hominins (Australopithecines)(Cordain et al., 1997).
High levels of sustained activity are primarily dependent on an efficient cardiovascular sys-
tem, which can be inferred from the structure of the thoracic cage and an efficient system for
heat dissipation (Aiello & Wheeler, 1995). In Homo the upper part of the rib is able to enlarge
the thorax during inhalation (Aiello & Wheeler, 1995; Cordain et al., 1997) while in Australo-
pithecines and earlier species the rib cages can be narrowed during respiration in order to
accommodate the muscle groups responsible for arboreal locomotion, namely the pectoral
girdle (Aiello & Wheeler, 1995; Schmid, 1991). Since ventilation of the lungs depends on the
movement of the diaphragm, the ability of Homo to enlarge the thorax suggests that they
might have had a more efficient ventilation system and in turn a more efficient cardiovascu-
lar system for prolonged activities compared to pongids and Australopithecines (Aiello &
Wheeler, 1995). Moreover, Homo erectus also showed a relative decrease in maximal pelvic
breadth relative to stature (Ruff, 1991), which would favor better heat dissipation during
intensive physical activities (such as running) even though the absolute body size had in-
creased from earlier hominids (Cordain et al., 1997).
Because of bipedalism and efficient thermoregulation, the increased encephalization14 of
our closest ancestors progressed simultaneously with more complex behaviors and physi-
ological changes (see for example: Bingham, 200015; Dunbar, 2003), such as a larger daily
range for searching for resources16 (Leonard, 2010) and an increase in daily total energy
expenditure17 as well as an increase in size and height (Cordain et al., 1998). Larger brains
inevitably require more energy18 and through improvements in aerobic capacity of early
Homo, larger areas could be forested to obtain high-energy diets. Raichlen and Polk (2013)
have argued that increases in PA and aerobic capacity might have influenced the encepha-
lization of modern humans directly19. This is in accordance with the hypothesis that early
Homo, in order to compete with other scavengers for food, evolved a high specialization in
CHAPTER I16
long-distance running (Bramble & Lieberman, 2004). This adaptation would have ensured
a high quality food supply, which would allow for the development of a larger brain. Re-
gardless of what came first (larger brains or higher activity), it is clear that PA has played
a large part in the evolution of modern humans.
PA is not an important aspect of modern humans’ daily lives, in terms of food security, and
the relationship between energy consumption and energy expenditure of modern humans
has been abrogated (Cordain et al., 1997; Eaton & Eaton, 2003; Katzmarzyk, 2010). The radi-
cal advances of new technologies, the global and organized domestication of animals and
the facilitation of commerce and trading have made food sources more abundant and the
exertion required to obtain food supplies is lower than in any other period in human history.
The estimated ratio of energy consumed over energy expended due to PA (energy efficiency
ratio) is 2.25 (i.e., 1kJ of energy expenditure to consume 2.25kJ of food) for a 57-kg Stone
Age human, and 3.66 for a 64-kg contemporary human (Eaton & Eaton, 2003). Essentially,
modern humans spend less energy and have a higher energy input (i.e., consume more food)
than their predecessors.
Additionally, humans living in modern societies are considerably less physically active than
modern hunter-gatherers (Katzmarzyk, 2010). Studies investigating PF and PA of Eskimo
communities (Igloolik, northern Canada), which have transitioned to a more Western life-
style, show a reduction of PF and PA over time (from 1970 to 1990) in all age groups (Rode &
Shephard, 1984; Rode & Shephard, 1994). Furthermore, a recent study comparing the energy
expenditure of humans living a modern lifestyle and the Old Order Amish population, who
reject modern technology, show large differences, with humans living a modern lifestyle be-
ing much more sedentary than the Amish (Bassett, Schneider, & Huntington, 2004).
In sum, even though PA has played a large part in human evolution, it has been reduced
markedly in modern societies. To illustrate this point, the World Health Organization (WHO)
recommends at least 150 minutes of moderate or light activity or at least 75 minutes of vig-
orous activity per week, but recent reports suggest that the average individual20 does not
adhere to this recommendation (Biddle & Goudas, 1996; Brawley & Rodgers, 1993; Desha et
al., 2007; Sagatun et al., 2007; WHO, 201321). Our genetic evolution is slower than the cultural
changes that have taken place over the years and, as a result, our genes have remained
adapted to the environmental circumstances of the past (Gould, 1980; p. 83). The health
implications of this mismatch between human gene evolution (‘hunter-gatherer genes’22)
and human culture (‘21st century lifestyle’) are numerous and well documented (Cordain et
al., 1997; Cordain et al., 1998; Eaton & Eaton, 2003; Katzmarzyk, Church, Craig, & Bouchard,
2009; Katzmarzyk, 2010).
17GENERAL INTRODUCTION
Overview of Physical Activity and Health Research
From times dating as far back as Hippocrates (460 – c. 370 BC), PA has been considered a key
factor for good health. Galen (129 – c. 210 AD) structured his medical theory around three
concepts, the “naturals” (i.e., physiology), the “non-naturals” (i.e., health) and the “contra-
naturals” (i.e., pathology). He primarily focused on the uses and abuses of the “six things non-
natural”23, which he thought were instrumental for health. These included: a) air; b) diet; c)
sleep and waking; d) exercise (motion) and rest; e) excretions and retentions; and f) passions
of the mind. Galen suggested that practicing the non-naturals (also known as the “six laws of
health”) in moderation would sustain health and prevent illness, whereas if these laws were
not followed, or followed in excess, disease would be the result (Berryman, 1989; Galen, 1991).
The modern scientific field of PA research began in the early 1950s in London. Morris et al.
(1953) investigated the mortality rates of London Transport Executive and Post-Office em-
ployees, and observed lower mortality rates from coronary problems in physically active
employees (bus conductors and post-men) than in their sedentary colleagues (bus-drivers
and telephone switchboard operators)24. Since then, research on the potential health ben-
efits of PA has been flourishing. Early life PA has been shown to promote the formation of
dense, well-mineralized bones (Burr, Ruff, & Thompson, 1990), as well as influence the bone
structural geometry (Larsen, 1997). Moreover, an inverse association exists between PA and
all-cause mortality (Paffenbarger, Hyde, Wing, & Hsieh, 1986; Powell & Blair, 1994; Salonen,
Puska, Kottke, Tuomilehto, & Nissinen, 1983), cardiac events such as myocardial infarction
(Balady, 2002; Myers et al., 2002), hypertension (Fagard, 2001), obesity25 (Prentice & Jebb,
2000), and type II diabetes (Hardman & Stensel, 2003). Furthermore, PA has been associated
with benefits to the immune system26 (Pedersen & Hoffman-Goetz, 2000; Shephard & Shek,
1994) and brain function; since it increases the blood and oxygen flow in the brain (Dishman
et al., 2006). Finally, PA has been positively associated with neurogenesis, gliogenesis and
neuroprotection in some areas of the animal and human brain (van Praag, Christie, Sejnowski,
& Gage, 1999; van Praag, Kempermann, & Gage, 1999; van Praag, 2008), as well as with cogni-
tive improvements in young children, adolescents (Chaddock, Hillman, Buck, & Cohen, 2011;
Hillman, Erickson, & Kramer, 2008), and older adults (Colcombe & Kramer, 2003; Colcombe et
al., 2004). The hypothesized mechanism for these benefits is the activation of neurotrophins
and other growth factors, such as Brain Derived Neurotrophic Factor (BDNF).
The recognition that PA has widespread benefits on (physical) health has led to the explora-
tion of whether the benefits of PA can also extend to well-being27 and mental health28. The
first epidemiological studies showed encouraging results. Stephens (1988), in a large study
of over 55,000 individuals, concluded that there was a clear positive relationship between
PA and well-being, positive mood and lower levels of anxiety and depression. Several later
CHAPTER I18
studies have also shown a positive relationship between PA and mental health (Deslandes
et al., 2009; Paluska & Schwenk, 2000).
Depression imposes a burden on the health of individuals and work place productivity. The
poor physical and social functioning of patients with a depressive disorder or sub-clinical de-
pressive symptoms has been shown to be comparable to (or even worse) than the function-
ing of patients with major chronic medical conditions (Merikangas et al., 2007; Wells et al.,
1989). Studies focusing on the impact of depression on the workplace (Kessler et al., 2006),
in terms of annual salary equivalent costs, have estimated that the cost of depression-
related loss of productivity in the USA is approximately $36.5 million per year. This is likely
to be an underestimation of the true cost to society.
The effectiveness of depression treatments such as anti-depressant medication and be-
havioral therapies is still being widely debated (Cuijpers, van Straten, Bohlmeijer, Hollon, &
Andersson, 2010; Cuijpers, Andersson, Donker, & van Straten, 2011; Khin, Chen, Yang, Yang,
& Laughren, 2011; Kirsch et al., 2008; Turner, Matthews, Linardatos, Tell, & Rosenthal, 2008).
Therefore, there is a need for alternative effective treatments for depression, especially low
cost treatments, with PA being a primary candidate. The UK National Institute for Clinical Ex-
cellence (NICE) guideline for depression recommends structured, supervised exercise three
times a week (45 to 60 minutes) for 10 to 12 weeks for mild depression (National Institute
for Health and Clinical Excellence, 2009). Although these Exercise Referral Schemes have
been implemented, the empirical evidence of their effectiveness on depression is not robust.
For several decades now, researchers have examined the potential use of PA as an effective
treatment for depression (Mead et al., 2009). Recent systematic reviews and meta-analyses
of randomized controlled studies (RCTs) have shown mixed results. Krogh et al. (2010) con-
cluded their meta-analysis by stating that there is a short-term benefit of exercise in clinically
depressed patients, but very little evidence of a long-term benefit. Another meta-analysis of
RCTs by Rimer et al. (2012) concluded that exercise seems to improve depressive symptoms in
clinically depressed patients when compared to no treatment or a control condition. However,
when only methodologically high-quality studies were included, the effect of exercise became
much smaller. Mead et al. (2009) and Cooney et al. (2013) reached a similar conclusion: when
only studies fulfilling three strict methodological criteria (i.e., allocation concealment, blinded
outcome assessment and intention to treat trials) were analyzed, the effect sizes observed
became smaller and not statistically significant. A recent large intervention study (Chalder et
al., 2012) did not find any evidence that PA is beneficial for depressive symptoms29. However,
intervention studies are only valuable in delineating whether a specific treatment is effective
for depression, and do not provide information on the numerous factors that might influence
the risk of developing depression and the progression of the disorder.
19GENERAL INTRODUCTION
Since the effect of PA on depressive symptoms has been reported frequently but seems
to be moderate to weak and may not even exist at all in some populations, there is a need
for a better theoretical understanding of this relationship, in order to derive the best ways
to maximize the potential benefit of PA in treating depression. Epidemiological studies in
representative populations can be useful in elucidating multifactorial causes of the disorder
and can help to provide a better theoretical framework of the relationship between PA and
depressive symptoms.
Research Gaps
Most studies conducted so far have focused on adults. Studies on adolescents have been
lacking; only recently has there been an increase in studies investigating the relationship
between PA and depressive symptoms in adolescents. Adolescence is a period of radical
physical, cognitive, and emotional change, characterized by a high incidence of depression
(Hankin et al., 1998). Youth depression and adult depression are potentially different due
to age-related differences in cognition, emotion, behavior, and physical development; so
adult models of depression cannot be generalized automatically to children and adolescents
(Abela & Hankin, 2008). Therefore, it is important to better understand the development and
trajectory of depressive symptoms and their relationship with PA during adolescence.
Observational cross-sectional and prospective studies have shown mixed results in the
relationship between PA and depressive symptoms in adolescents. In a meta-analysis of
observational studies, Biddle and Asare (2011) showed that PA has a potentially beneficial
effect on depressive symptoms, but the associations were weak and most studies suffered
from serious methodological limitations. Johnson and Taliaferro (2011) reached a similar
conclusion in their meta-analysis. Small to moderate protective effects of PA or sport par-
ticipation on depressive symptoms were observed, but methodological limitations plagued
the majority of these studies.
Besides obvious methodological problems related to this type of research, such as discrep-
ant measures of depressive symptoms and PA and an over-reliance on cross-sectional de-
signs (Johnson & Taliaferro, 2011), other important issues need to be addressed. One issue is
the direction of the association. It is possible that PA helps to alleviate depressive symptoms
(protective hypothesis), but also that depressive symptomatology hinders PA participation
(inhibition hypothesis). Perhaps both, i.e., bidirectional associations, are true. Consideration
of potentially moderating factors such as genetic or psychosocial variables is also vital, as
these might make individuals more prone to benefit from PA, or be more adversely impacted
by physical inactivity.
CHAPTER I20
Aim and Objectives of this Thesis
The overall aim of this thesis is to provide a deeper understanding of the association be-
tween PA and depressive symptoms in adolescents and adults. The study in chapter II in-
vestigates the direction of the relationship between PA and depressive symptoms, while the
study in chapter III investigates the hypothesis that early engagement in PA in adolescence
might prevent an onset of a major depressive episode in early adulthood, and explores the
role of specific characteristics of PA, i.e., the nature, frequency, intensity and duration.
The second part of the thesis focuses on possible moderators of the relationship between
PA and depressive symptoms. Chapter IV examines genetic factors that might influence the
extent to which adolescents benefit from PA, i.e., genes that work on important neuromodu-
lators (such as serotonin and dopamine) and neurotrophins (BDNF). These neuromodulators
and neurotrophins are thought to play an important role in the development and course of
depression and also to be affected by PA. Psychosocial moderators and their influence on
the relationship between PA and depressive symptoms are examined in chapter V and VI.
The study in chapter V investigates whether peer-nominated sport competence and gen-
der moderate the association of sport participation with current depressive symptoms and
symptom changes over time. Chapter VI concerns the relative-age position of adolescents
among their classmates, which might affect the way they are perceived on their athletic
abilities by teachers, and their levels of depressive symptoms.
The final part considers individual differences in the association between PA and mood in
adults. Chapter VII evaluates individual differences in the relationship between PA, as as-
sessed with accelerometers, and repeated daily assessments of positive and negative af-
fect in depressed and non-depressed adults. In this study, the direction of the relationship
between PA and daily mood is investigated at the individual rather than at the group level.
In chapter VIII the main results of the research undertaken are summarized and discussed
further, along with clinical implications and possible directions for future research.
Summary of Samples Used
The TRAILS Sample
Five out of the six empirical studies in this thesis are based on data from the Dutch TRack-
ing Adolescents’ Individual Lives Survey (TRAILS). TRAILS is a prospective population-based
cohort of Northern Dutch adolescents. The general aim of TRAILS was to delineate etiologies
and trajectories of (mental) health problems in the Dutch population. Data collection for the
baseline measurement (T1) started in 2001 and finished in 2002. Data collection for the second
(T2), third (T3) and fourth (T4) measurement waves took place at intervals of approximately
2.5 years. For this thesis, data have been used from all four measurement waves (T1 to T4). At
21GENERAL INTRODUCTION
T1, 2230 adolescents agreed to participate. The response rates at T2, T3 and T4 were 96.4%
(N = 2149; 51% girls, mean age = 13.7, SD = 0.5), 81.4% (N = 1816; 52.3% girls, mean age = 16.3,
SD = 0.7) and 84.3% (N = 1881; 52.3% girls, mean age = 19.1, SD = 0.6) respectively. Detailed in-
formation on the profile, design and procedures of TRAILS have been described elsewhere (de
Winter et al., 2005; Huisman et al., 2008; Nederhof et al., 2012; Ormel et al., 2012).
The MOOVD Sample
The final study of this thesis is based on data from the ongoing ‘Mood and movement in
daily life’ (MOOVD) study, which was set up to investigate the dynamic temporal relationship
between PA and mood in daily life, as well as the role of several biomarkers therein. Data col-
lection of this study started in 2012, and the data from the first 20 individuals to participate
were used (N depressed = 10; 30% males, mean age = 36.4, SD = 10.3 & N non-depressed = 10;
30% males, mean age = 36.7 and SD = 7.9). Participants were intensively monitored in their
natural environment for 30 days, by means of electronic diaries, saliva sampling and contin-
uous actigraphy. Participants were pair-matched on gender, smoking status, age, and Body
Mass Index (BMI). Further details on inclusion and exclusion criteria, procedures, measures,
and ambulatory sampling can be found in Chapter VII.
Notes 1 PF will not be explored further in this thesis. The definition of PF is provided for a better understanding of
the distinct terms related to PA that are sometimes used in the literature. 2 Sometimes the energy expended during sleep, although very low, is still considered part of PA (Caspers-
en, Powell, & Christenson, 1985). In this thesis the energy expended during sleep was not considered as part of PA.
3 RMR is the energy expended during quiet sitting. 4 1MET ≡ 1kcal/kg*h ≡ 4.184kJ/kg*h. 5 See: oxforddictionaries.com/definition/english/sport?q = sport 6 Slightly lower than the prevalence of depression in the USA, which is around 20% (Kessler & Wang, 2009). 7 The Australopithecines are classified as hominins. The terminology has changed recently, but according
to the newest terminology used there are two main distinctions: a) Hominids (hominidae), which are a taxonomic family of all modern and extinct Great Apes including humans; while b) Hominins (Homininae), is a subcategory of the Hominids, including all modern humans, extinct human species and all our im-mediate ancestors (Australopithecines, Homo, Paranthropus and Ardipithecus).
8 Australopithecines walked upright (bipedalism; McΗenry, 1994). The first evidence of bipedalism comes with Australopithecus anamensis, which is assumed to be an ancestor of the better known Australo-pithecus afarensis (Cordain, Gotshall, Eaton, & Eaton, 1998; Leakey, Feibel, McDougall, & Walker, 1995). Although the Australopithecines retained ape like upper limb structural features (McΗenry, 1994; Stern &
CHAPTER I22
Susman, 1983), some evidence from both fossilized footprints and leg bone/pelvic structure suggest that they might have walked with a relatively similar mechanical efficiency to that of contemporary humans (Lovejoy, 1988).
9 It is still disputed whether bipedalism is as efficient, in relation to energy expenditure, as walking in quad-rupeds, but this is not something that will be explored in further detail here (see for example: Halsey & White, 2012).
10 At the same time, the evolution of attenuated body hair size (‘naked’ skin) also improved the evaporative cooling efficiency of the cutaneous sweat glands (Wheeler, 1992).
11 All increases in cranial capacity are discussed in relation to body size. It should be noted, that cranial capacity does not necessarily indicate higher intelligence, since brain structure/volume is more important than actual size.
12 The estimated brain size, relative to body size, of Australopithecines did not exceed 350-600cc, a size much smaller than later Homo species (around 700-1500cc; see figure 4, p. 6784 in McΗenry, 1994) and quite similar in size to modern great apes (around 400cc in chimpanzees and gorillas; Schoenemann, 2006). There is one exception to this, namely Homo floresiensis (Brown et al., 2004) with an estimated brain size equal or smaller than that of chimpanzees. However, it should be noted that the study of brain sizes (paleoneurology) in extinct hominid species is hindered by the incomplete fossil record and as a result these brain sizes are approximations.
13 At the same time of the development of a larger brain, the gut in Homo species started shrinking, simi-larly to what is observed in other carnivores, i.e., a compensation occurring at the expense of the size of the gut to accommodate a more metabolically active brain (Aiello & Wheeler, 1995; Cordain, Gotshall, & Eaton, 1997). Because the gut is not encased by bones, as is the case with the brain, the decrease in size of the gut has been estimated through differences in the anatomy (especially the abdomen and rib cages) of two early species of hominids, namely A. afarensis and Homo ergaster (Aiello & Wheeler, 1995).
14 I.e., developing larger brains.15 The refined ability of humans to throw projectiles in combination with human complex social abilities
(social cooperation) is also another hypothesized factor affecting our current development (for an inter-esting overview of this hypothesis see: Bingham, 1999 and Bingham, 2000).
16 Which was aided by bipedalism and the ability to run large distances (at moderate pace) more efficiently than many quadrupeds (for an interesting discussion see: Bramble & Lieberman, 2004).
17 The estimated energy expenditure of early Homo and modern humans was larger than that of non-human primates (Leonard & Robertson, 1997).
18 The human brain uses around 20% of the body’s energy consumption (Lassen, 1959), a percentage much higher than in other nonhuman primates (Armstrong, 1985; Armstrong, 1990) and other mammals (Arm-strong, 1983).
19 The authors base their hypothesis on the link between proximate mechanisms such as neurotrophins, PA and brain size by: a) reviewing intra and inter-specific and artificial selection studies that suggest that by improving endurance capacity these proximal mechanisms will be altered; and b) studying what is termed ‘athletic species’ (for more information see: Raichlen & Polk, 2013).
20 For example, fewer than 60% of American and Norwegian adolescents adhere to these guidelines (De-sha, Ziviani, Nicholson, Martin, & Darnell, 2007; Sagatun, Sogaard, Bjertness, Selmer, & Heyerdahl, 2007).
21 who.int/dietphysicalactivity/factsheet_inactivity/en/index.html22 For approximately 99% of our existence as a species we have been hunters and food gatherers (Astrand,
1994).23 The works of Hippocrates and especially his three books on regimen, heavily influenced Galen.
23GENERAL INTRODUCTION
24 However these associations might be explained (better) by differences in time spent sitting rather than the activity itself (Hamilton, Hamilton, & Zderic, 2007).
25 However, the relationship between PA and obesity is very complex (Biddle & Mutrie, 2008; p. 22).26 Overtraining, however, may also impair the immune system (Shephard & Shek, 1994).27 Since Hippocrates and Galen, anecdotal evidence or philosophical theories have been proposed on the
benefits of PA on quality of life and well-being. Nonetheless, it should be noted that the terms ‘well-be-ing’ and ‘quality of life’ are rather complex, with various different ways to measure ‘well-being’ and ‘qual-ity of life’, whilst there are different research groups focusing on different aspects of the two (Biddle & Mutrie, 2008).
28 It is important to note here that the distinction between physical and mental health seems to be arbi-trary, since ‘mental health’ problems are multifactorial and probably caused by a dysfunction in both psychological and neurobiological processes (Kendler, 2012). However, this discussion is out of the scope of this thesis and the scientific nomenclature is used for clarity.
29 This study also investigated the cost-effectiveness of a PA intervention and their findings raised doubts on whether PA is a cost-effective treatment for depression.
CHAPTER I24
REFERENCES
Abela, J. R. Z., & Hankin, B. L. (2008). Handbook of depression in children and adolescents. New York: The Guildford Press.
Aiello, L. C., & Wheeler, P. (1995). The expensive-tissue hypothesis - the brain and the digestive-system in human and primate evolution. Current Anthropology, 36(2), 199-221.
Ainsworth, B. E., Haskell, W. L., Leon, A. S., Jacobs, D. R., Montoye, H. J., Sallis, J. F., et al. (1993). Compen-dium of physical activities - classification of energy costs of human physical activities. Medicine and Science in Sports and Exercise, 25(1), 71-80.
American Psychiatric Association (2013). Diagnostic and statistical manual of mental disorders (DSM-5) (5th ed.). Arlington, VA: American Psychiatric Association.
Armstrong, E. (1983). Relative brain size and metabolism in mammals. Science, 220(4603), 1302-1304.
Armstrong, E. (1985). Relative brain size in monkeys and prosimians. American Journal of Physical Anthro-pology, 66(3), 263-273.
Armstrong, E. (1990). Brains, bodies and metabolism. Brain Behavior and Evolution, 36(2-3), 166-176.
Astrand, P. O. (1994). Physical activity and fitness: Evolutionary perspective and trends for the future. In C. Bouchard, R. J. Shephard, & T. Stephens (Eds.), Physical activity, fitness and health (pp. 98-105). Cham-paign, IL: Human Kinetics.
Balady, G. J. (2002). Survival of the fittest - more evidence. New England Journal of Medicine, 346(11), 852-854.
Bassett, D. R., Schneider, P. L., & Huntington, G. E. (2004). Physical activity in an Old Order Amish community. Medicine and Science in Sports and Exercise, 36(1), 79-85.
Behrensmeyer, A. K., & Cooke, H. S. (1985). Paleoenvironments, stratigraphy and taphonomy in the African Pliocene and Early Pleistocene. In E. Delson (Ed.), Ancestors: The hard evidence (pp. 60-62). New York: Wiley-Liss.
Berryman, J. W. (1989). The tradition of the 6 things non-natural - exercise and medicine from Hippocrates through antebellum America. Exercise and Sport Sciences Reviews/series, 17, 515-557.
Biddle, S. J. H., & Mutrie, N. (2008). Psychology of physical activity: Determinants, well-being and interven-tions (2nd ed.). New York, USA: Routledge.
Biddle, S. J. H., & Goudas, M. (1996). Analysis of children’s physical activity and its association with adult encouragement and social cognitive variables. Journal of School Health, 66(2), 75-78.
Biddle, S. J. H., & Asare, M. (2011). Physical activity and mental health in children and adolescents: A review of reviews. British Journal of Sports Medicine, 45(11), 886-895.
Bingham, P. M. (1999). Human uniqueness: A general theory. Quarterly Review of Biology, 74(2), 133-169.
Bingham, P. M. (2000). Human evolution and human history: A complete theory. Evolutionary Anthropology, 9(6), 248-257.
Bramble, D. M., & Lieberman, D. E. (2004). Endurance running and the evolution of Homo. Nature, 432(7015), 345-352.
Brawley, L. R., & Rodgers, W. M. (1993). Social-psychological aspects of fitness promotion. In P. Seraganian (Ed.), Exercise psychology: The influence of physical exercise on psychological processes (pp. 254-298). New York: Wiley.
Brown, F., Harris, J., Leakey, R., & Walker, A. (1985). Early Homo-erectus skeleton from West Lake Turkana, Kenya. Nature, 316(6031), 788-792.
Brown, P., Sutikna, T., Morwood, M. J., Soejono, R. P., Jatmiko, Saptomo, E. W., et al. (2004). A new small-bodied hominin from the Late Pleistocene of Flores, Indonesia. Nature, 431(7012), 1055-1061.
25GENERAL INTRODUCTION
Burr, D. B., Ruff, C. B., & Thompson, D. D. (1990). Patterns of skeletal histologic change through time - com-parison of an archaic Native-American population with modern populations. Anatomical Record, 226(3), 307-313.
Caspersen, C. J., Powell, K. E., & Christenson, G. M. (1985). Physical-activity, exercise, and physical-fitness: Definitions and distinctions for health-related research. Public Health Reports, 100(2), 126-131.
Chaddock, L., Hillman, C. H., Buck, S. M., & Cohen, N. J. (2011). Aerobic fitness and executive control of rela-tional memory in preadolescent children. Medicine and Science in Sports and Exercise, 43(2), 344-349.
Chalder, M., Wiles, N. J., Campbell, J., Hollinghurst, S. P., Haase, A. M., Taylor, A. H., et al. (2012). Facilitated physical activity as a treatment for depressed adults: Randomised controlled trial. British Medical Jour-nal (Clinical Research Ed.), 344, e2758.
Colcombe, S. J., & Kramer, A. F. (2003). Fitness effects on the cognitive function of older adults: A meta-analytic study. Psychological Science, 14(2), 125-130.
Colcombe, S. J., Kramer, A. F., Erickson, K. I., Scalf, P., McAuley, E., Cohen, N. J., et al. (2004). Cardiovascular fitness, cortical plasticity, and aging. Proceedings of the National Academy of Sciences of the United States of America, 101(9), 3316-3321.
Cooney, G. M., Dwan, K., Greig, C. A., Lawlor, D. A., Rimer, J., Waugh, F. R., et al. (2013). Exercise for depres-sion. Cochrane Database of Systematic Reviews, (9), CD004366.
Cordain, L., Gotshall, R. W., Eaton, S. B., & Eaton, S. B. (1998). Physical activity, energy expenditure and fit-ness: An evolutionary perspective. International Journal of Sports Medicine, 19(5), 328-335.
Cordain, L., Gotshall, R. W., & Eaton, S. B. (1997). Evolutionary aspects of exercise. Nutrition and Fitness: Evolutionary Aspects, Children’s Health, Programs and Policies, 81, 49-60.
Cuijpers, P., van Straten, A., Bohlmeijer, E., Hollon, S. D., & Andersson, G. (2010). The effects of psychothera-py for adult depression are overestimated: A meta-analysis of study quality and effect size. Psychologi-cal Medicine, 40(2), 211-223.
Cuijpers, P., Andersson, G., Donker, T., & van Straten, A. (2011). Psychological treatment of depression: Re-sults of a series of meta-analyses. Nordic Journal of Psychiatry, 65(6), 354-364.
de Graaf, R., ten Have, M., van Gool, C., & van Dorsselaer, S. (2012). Prevalence of mental disorders and trends from 1996 to 2009. Results from the Netherlands Mental Health Survey and Incidence Study-2. Social Psychiatry and Psychiatric Epidemiology, 47(2), 203-213.
de Winter, A. F., Oldehinkel, A. J., Veenstra, R., Brunnekreef, J. A., Verhulst, F. C., & Ormel, J. (2005). Evalua-tion of non-response bias in mental health determinants and outcomes in a large sample of pre-adoles-cents. European Journal of Epidemiology, 20(2), 173-181.
deMonocal, P. B., & Bloemendal, J. (1995). Plio-Pleistocene climatic variability in subtropical Africa and the paleoenvironment of hominid evolution: A combined data-model approach. In E. S. Vrba, G. H. Denton, T. C. Patridge, & L. H. Burckle (Eds.), Paleoclimate and evolution with an emphasis on human origins (pp. 262-288). New Haven: Yale University Press.
Desha, L. N., Ziviani, J. M., Nicholson, J. M., Martin, G., & Darnell, R. E. (2007). Physical activity and depres-sive symptoms in American adolescents. Journal of Sport & Exercise Psychology, 29(4), 534-543.
Deslandes, A., Moraes, H., Ferreira, C., Veiga, H., Silveira, H., Mouta, R., et al. (2009). Exercise and mental health: Many reasons to move. Neuropsychobiology, 59(4), 191-198.
Dishman, R. K., Berthoud, H. R., Booth, F. W., Cotman, C. W., Edgerton, V. R., Fleshner, M. R., et al. (2006). Neurobiology of exercise. Obesity, 14(3), 345-356.
Dunbar, R. I. M. (2003). The social brain: Mind, language, and society in evolutionary perspective. Annual Review of Anthropology, 32, 163-181.
Eaton, S. B., & Eaton, S. B. (2003). An evolutionary perspective on human physical activity: Implications for health. Comparative Biochemistry and Physiology A-Molecular & Integrative Physiology, 136(1), 153-159.
CHAPTER I26
Eaton, S. B., & Konner, M. (1985). Paleolithic nutrition - a consideration of its nature and current implications. New England Journal of Medicine, 312(5), 283-289.
Fagard, R. H. (2001). Exercise characteristics and the blood pressure response to dynamic physical training. Medicine and Science in Sports and Exercise, 33(6), S484-S492.
Foley, R. (1987). Another unique species: Patterns in human evolutionary ecology. Harlow: Longman.
Galen, A. (1991). On the therapeutic method (R. J. Hankinson Trans.). Oxford: Clarendon Press.
Gould, S. J. (1980). The panda’s thumb. W.W. Norton & Company.
Halsey, L. G., & White, C. R. (2012). Comparative energetics of mammalian locomotion: Humans are not differ-ent. Journal of Human Evolution, 63(5), 718-22. doi:10.1016/j.Jhevol.2012.07.008.Epub 2012 Sep 7
Hamilton, M. T., Hamilton, D. G., & Zderic, T. W. (2007). Role of low energy expenditure and sitting in obesity, metabolic syndrome, type 2 diabetes, and cardiovascular disease. Diabetes, 56(11), 2655-2667.
Hankin, B. L., Abramson, L. Y., Moffitt, T. E., Silva, P. A., McGee, R., & Angell, K. E. (1998). Development of de-pression from preadolescence to young adulthood: Emerging gender differences in a 10-year longitudinal study. Journal of Abnormal Psychology, 107(1), 128-140.
Hardman, A. E., & Stensel, D. J. (2003). Physical activity and health: The evidence explained. London: Rout-ledge.
Hasler, G., Drevets, W. C., Manji, H. K., & Charney, D. S. (2004). Discovering endophenotypes for major depres-sion. Neuropsychopharmacology, 29(10), 1765-1781.
Hillman, C. H., Erickson, K. I., & Kramer, A. F. (2008). Be smart, exercise your heart: Exercise effects on brain and cognition. Nature Reviews Neuroscience, 9(1), 58-65.
Huisman, M., Oldehinkel, A. J., de Winter, A. F., Minderaa, R. B., de Bildt, A., Huizink, A. C., et al. (2008). Cohort profile: The Dutch ‘TRacking Adolescents’ Individual Lives’ Survey’; TRAILS. International Journal of Epidemiology, 37(6), 1227-1235.
Johnson, K. E., & Taliaferro, L. A. (2011). Relationships between physical activity and depressive symptoms among middle and older adolescents: A review of the research literature. Journal for Specialists in Pedi-atric Nursing: JSPN, 16(4), 235-251.
Joorman, J. (2009). Cognitive aspects of depression. In I. H. Gotlib, & C. L. Hammen (Eds.), Handbook of depression (2nd ed., pp. 298-321). New York: The Guildford Press.
Katzmarzyk, P. T., Church, T. S., Craig, C. L., & Bouchard, C. (2009). Sitting time and mortality from all causes, cardiovascular disease, and cancer. Medicine and Science in Sports and Exercise, 41(5), 998-1005.
Katzmarzyk, P. T. (2010). Physical activity, sedentary behavior, and health: Paradigm paralysis or paradigm shift? Diabetes, 59(11), 2717-2725.
Kendler, K. S. (2012). The dappled nature of causes of psychiatric illness: Replacing the organic-functional/hardware-software dichotomy with empirically based pluralism. Molecular Psychiatry, 17(4), 377-388.
Kessler, R. C., & Wang, P. S. (2009). Epidemiology of depression. In I. H. Gotlib, & C. L. Hammen (Eds.), Hand-book of depression (2nd ed., pp. 5-22). New York: The Guildford Press.
Kessler, R. C., Akiskal, H. S., Ames, M., Birnbaum, H., Greenberg, P., Hirschfeld, R. M. A., et al. (2006). Preva-lence and effects of mood disorders on work performance in a nationally representative sample of U.S. workers. American Journal of Psychiatry, 163(9), 1561-1568.
Khin, N. A., Chen, Y., Yang, Y., Yang, P., & Laughren, T. P. (2011). Exploratory analyses of efficacy data from major depressive disorder trials submitted to the US Food and Drug Administration in support of new drug applications. Journal of Clinical Psychiatry, 72(4), 464-472.
Kirsch, I., Deacon, B. J., Huedo-Medina, T. B., Scoboria, A., Moore, T. J., & Johnson, B. T. (2008). Initial severity and antidepressant benefits: A meta-analysis of data submitted to the Food and Drug Administration. Plos Medicine, 5(2), 260-268.
27GENERAL INTRODUCTION
Krogh, J., Nordentoft, M., Sterne, J. A., & Lawlor, D. A. (2010). The effect of exercise in clinically depressed adults: Systematic review and meta-analysis of randomized controlled trials. The Journal of Clinical Psychiatry, 72(4), 529-538.
Larsen, C. S. (1997). Bioarcheology. Interpreting behavior from the human skeleton. Cambridge: Cambridge University Press.
Lassen, N. A. (1959). Cerebral blood flow and oxygen consumption in man. Physiological Reviews, 39(2), 183-238.
Leakey, M. G., Feibel, C. S., McDougall, I., & Walker, A. (1995). New 4-million-year-old hominid species from Kanapoi and Allia Bay, Kenya. Nature, 376(6541), 565-571.
Leonard, W. R., & Robertson, M. L. (1997). Comparative primate energetics and hominid evolution. American Journal of Physical Anthropology, 102(2), 265-281.
Leonard, W. R. (2010). Size counts: Evolutionary perspectives on physical activity and body size from early hominids to modern humans. Journal of Physical Activity & Health, 7, S284-S298.
Lieberman, D. E. (2011). The evolution of the human head. Cambridge, MA: Harvard University Press.
Lovejoy, C. O. (1988). Evolution of human walking. Scientific American, 259(5), 118-125.
Lux, V., & Kendler, K. S. (2010). Deconstructing major depression: A validation study of the DSM-IV symp-tomatic criteria. Psychological Medicine, 40(10), 1679-1690.
McΗenry, H. M. (1994). Tempo and mode in human-evolution. Proceedings of the National Academy of Sci-ences of the United States of America, 91(15), 6780-6786.
Mead, G. E., Morley, W., Campbell, P., Greig, C. A., McMurdo, M., & Lawlor, D. A. (2009). Exercise for depres-sion. Cochrane Database of Systematic Reviews (Online), (3), CD004366.
Merikangas, K. R., Ames, M., Cui, L., Stang, P. E., Ustun, T. B., Von Korff, M., et al. (2007). The impact of comor-bidity of mental and physical conditions on role disability in the US adult household population. Archives of General Psychiatry, 64(10), 1180-1188.
Morris, Jn., Heady, J. A., Raffle, P. A., Roberts, C. Q., & Parks, J. W. (1953). Coronary heart-disease and physi-cal activity of work. Lancet, 265(6796), 1111-20.
Myers, J., Prakash, M., Froelicher, V., Do, D., Partington, S., & Atwood, J. E. (2002). Exercise capacity and mortality among men referred for exercise testing. New England Journal of Medicine, 346(11), 793-801.
National Institute for Health and Clinical Excellence (2009). Depression: The treatment and management of depression in adults. London: British Psychological Society and Royal College of Psychiatrists.
Nederhof, E., Jorg, F., Raven, D., Veenstra, R., Verhulst, F. C., Ormel, J., et al. (2012). Benefits of extensive recruitment effort persist during follow-ups and are consistent across age group and survey method. The TRAILS study. BMC Medical Research Methodology, 12(1), 93.
Ormel, J., Oldehinkel, A. J., Sijtsema, J., van Oort, F., Raven, D., Veenstra, R., et al. (2012). The TRacking Adolescents’ Individual Lives Survey (TRAILS): Design, current status, and selected findings. Journal of the American Academy of Child and Adolescent Psychiatry, 51(10), 1020-1036.
Paffenbarger, R. S., Hyde, R. T., Wing, A. L., & Hsieh, C. C. (1986). Physical-activity, all-cause mortality, and longevity of college alumni. New England Journal of Medicine, 314(10), 605-613.
Paluska, S. A., & Schwenk, T. L. (2000). Physical activity and mental health: Current concepts. Sports Medi-cine, 29(3), 167-180.
Pedersen, B. K., & Hoffman-Goetz, L. (2000). Exercise and the immune system: Regulation, integration, and adaptation. Physiological Reviews, 80(3), 1055-1081.
Powell, K. E., & Blair, S. N. (1994). The public-health burdens of sedentary living habits - theoretical but realistic estimates. Medicine and Science in Sports and Exercise, 26(7), 851-856.
CHAPTER I28
Prentice, A. M., & Jebb, S. A. (2000). Physical activity and weight control in adults. In C. Bouchard (Ed.), Physical activity and obesity (pp. 247-261). Champaign, IL: Human Kinetics.
Raichlen, D. A., & Polk, J. D. (2013). Linking brains and brawn: Exercise and the evolution of human neurobiol-ogy. Proceedings of the Royal Society B-Biological Sciences, 280(1750), 20122250.
Rimer, J., Dwan, K., Lawlor, D. A., Greig, C. A., McMurdo, M., Morley, W., et al. (2012). Exercise for depression. Cochrane Database of Systematic Reviews, (7), CD004366.
Rode, A., & Shephard, R. J. (1984). Growth, development and acculturation - a 10 year comparison of Cana-dian Inuit children. Human Biology, 56(2), 217-230.
Rode, A., & Shephard, R. J. (1994). Growth and fitness of Canadian Inuit - secular trends, 1970-1990. Ameri-can Journal of Human Biology, 6(4), 525-541.
Ruff, C. B. (1991). Climate and body shape in hominid evolution. Journal of Human Evolution, 21(2), 81-105.
Sagatun, A., Sogaard, A. J., Bjertness, E., Selmer, R., & Heyerdahl, S. (2007). The association between week-ly hours of physical activity and mental health: A three-year follow-up study of 15-16-year-old students in the city of Oslo, Norway. BMC Public Health, 7, 155.
Salonen, J. T., Puska, P., Kottke, T. E., Tuomilehto, J., & Nissinen, A. (1983). Decline in mortality from coronary heart-disease in Finland from 1969 to 1979. British Medical Journal, 286(6381), 1857-1860.
Schmid, P. (1991). The trunk of the Australopithecines. In Y. Coppens, & B. Senut (Eds.), Origine(s) de la bipi-die chez les hominides. Paris: Edition du CNRS.
Schoenemann, P. T. (2006). Evolution of the size and functional areas of the human brain. Annual Review of Anthropology, 35, 379-406.
Shephard, R. J., & Shek, P. N. (1994). Potential impact of physical-activity and sport on the immune-system - a brief review. British Journal of Sports Medicine, 28(4), 247-255.
Shipman, P. (1986). Scavenging or hunting in early hominids - theoretical framework and tests. American Anthropologist, 88(1), 27-43.
Stephens, T. (1988). Physical activity and mental health in the United States and Canada: Evidence from four population surveys. Preventive Medicine, 17(1), 35-47.
Stern, J. T., & Susman, R. L. (1983). The locomotor anatomy of Australopithecus afarensis. American Jour-nal of Physical Anthropology, 60(3), 279-317.
Turner, E. H., Matthews, A. M., Linardatos, E., Tell, R. A., & Rosenthal, R. (2008). Selective publication of antidepressant trials and its influence on apparent efficacy. New England Journal of Medicine, 358(3), 252-260.
van Praag, H., Christie, B. R., Sejnowski, T. J., & Gage, F. H. (1999). Running enhances neurogenesis, learn-ing, and long-term potentiation in mice. Proceedings of the National Academy of Sciences of the United States of America, 96(23), 13427-13431.
van Praag, H., Kempermann, G., & Gage, F. H. (1999). Running increases cell proliferation and neurogenesis in the adult mouse dentate gyrus. Nature Neuroscience, 2(3), 266-270.
van Praag, H. (2008). Neurogenesis and exercise: Past and future directions. Neuromolecular Medicine, 10(2), 128-140.
Wells, K. B., Stewart, A., Hays, R. D., Burnam, M. A., Rogers, W., Daniels, M., et al. (1989). The functioning and well-being of depressed-patients - results from the Medical Outcomes Study. Jama-Journal of the American Medical Association, 262(7), 914-919.
Wheeler, P. E. (1991). The influence of bipedalism on the energy and water budgets of early hominids. Jour-nal of Human Evolution, 21(2), 117-136.
Wheeler, P. E. (1992). The influence of the loss of functional body hair on the water budgets of early homi-nids. Journal of Human Evolution, 23(5), 379-388.
Bidirectional Prospective Associations between Physical Activity and Depressive Symptoms. The TRAILS Study.
“All truly great thoughts are conceived by walking.”
Friedrich Nietzsche, Twilight of the Idols, Or, How to Philosophize with the Hammer, 1889
“If you would get exercise, go in search of the springs of life.”Henry David Thoreau, Walking, 1862
Stavrakakis N, de Jonge P, Ormel J & Oldehinkel AJ
J Adolesc Health 2012; 50(5):503-8
CHAPTER
CHAPTER II30
ABSTRACT
Purpose: Low levels of physical activity have been shown to be associated with depression
in adults. The few studies that focused on adolescents yielded mixed and inconsistent re-
sults. Efforts to examine the direction of this relationship have been inconclusive up to now.
The aims of this study were therefore to investigate (1) the direction of the inverse associa-
tion between physical activity and depressive symptoms over time and (2) whether these
associations are specific to particular clusters of depressive symptoms in adolescents.
Methods: Depressive symptoms and physical activity were assessed in a population sample
of adolescents (N = 2230), who were measured at three waves between age 11 and age 16.
Depressive symptoms were measured by the Affective Problems scale of the Youth Self-Re-
port (YSR) and Child Behavior Checklist (CBCL), while physical activity was operationalized
as the amount of time spent on physical exercise. Structural Equation Modeling was used to
examine bidirectional effects of physical activity and depressive symptoms over time.
Results: We found significant cross-lagged paths from prior physical activity to later de-
pression as well as from prior depression to later physical activity (beta values = -0.033 to
-0.047). After subdividing depression into affective and somatic symptoms, the affective
symptoms were reciprocally related to physical activity, while the paths between somatic
symptoms and physical activity did not reach statistical significance.
Conclusion: An inverse bidirectional association between physical activity and general de-
pressive symptoms was observed. This association was restricted to affective symptoms.
31PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
INTRODUCTION
Associations between physical activity (PA) and health have been well documented. Stud-
ies have shown that exercise has beneficial effects on medical conditions such as diabetes
and cardiovascular diseases (Hu et al., 2004; Wessel et al., 2004), and also lowers the risk
of developing a variety of cancers (Howard et al., 2008; Bardia et al., 2006). Associations
of PA with mental health, however, have been harder to elucidate. There is some evidence
suggesting that PA is inversely related to depression in adults, but it has been difficult to
establish causality and the strength of the associations varied from study to study (Harris,
Cronkite, & Moos, 2006; Cooper-Patrick, Ford, Mead, Chang, & Klag, 1997).
In observational studies, causality cannot be established, but the temporal order of changes
in PA and depression may provide clues about the direction of the effects. It has been very
difficult to determine temporality so far, however, due to the cross-sectional nature of most
studies. Weyerer (1992) showed a cross-sectional association between low levels of PA and
depression, but could not find a prospective association at 5-year follow-up. Since then, only
a few studies have assessed the direction of the association between depression and PA,
and those were relatively unsuccessful in giving a clear answer (Birkeland, Torsheim, & Wold,
2009; Jerstad, Boutelle, Ness, & Stice, 2010). Intervention trials have yielded ambiguous re-
sults as well. Whereas a number of studies indicated that PA reduces feelings of tension,
anxiety and anger (Babyak et al., 2000), others could not find such clear effects (Salmon, 2001;
Haboush, Floyd, Caron, LaSota, & Alvarez, 2006). In a recent review, Mead and colleagues
(2009) showed that, when only trials with blinded outcome assessments, intention to treat
and allocation concealment were included in the analysis, effect sizes were weaker and not
statistically significant. In adolescents, although cross-sectional and intervention studies
have shown a consistent inverse association between PA and depressive symptoms, results
from prospective studies have been inconsistent and will be briefly described later.
Why are there so many mixed results regarding associations between PA and depression?
Part of the discrepancies can be attributed to methodological limitations such as small
sample sizes and, for prospective studies, short follow-up periods. Another reason might be
that depression is a heterogeneous disorder, consisting of both affective and somatic symp-
toms. Instruments that are used to assess depressive symptoms vary widely with respect
to the symptom profiles they represent, where some instruments do not include somatic
symptoms, while other instruments contain multiple questions on somatic symptoms. It has
been shown that somatic and affective symptoms are differentially associated with cardiac
autonomic (de Jonge et al., 2006) and HPA axis function (Bosch et al., 2009). Since these two
subgroups of symptoms show distinct patterns of association with various physiological
measures, they may be differentially related to PA as well.
CHAPTER II32
Despite these discrepancies, several, not necessarily mutually exclusive, hypotheses have
been proposed to explain the inverse association between PA and depression:
1) Protection hypothesis (direct causal effect): PA decreases depressive symptoms through
biological (elevation of endorphins, serotonin or endocannabinoids) or social (social con-
tact and self-esteem) mechanisms (Allgower, Wardle, & Steptoe, 2001).
2) Inhibition hypothesis (reverse causality): depression negatively influences patterns of PA
(Patten, Williams, Lavorato, & Eliasziw, 2009) through symptoms such as lack of energy,
anhedonia, low mood and social withdrawal (Struder & Weicker, 2001).
3) Common Cause hypothesis: PA and depression share risk factors, for instance genetic or
familial factors such as socioeconomic status (SES), neighborhood and parental rearing
style (Stubbe, de Moor, Boomsma, & de Geus, 2007).
In the present study, associations between PA and depressive symptoms were studied in
a large prospective population cohort of Dutch adolescents. Studying this association in
adolescents is important for at least two reasons. First, evidence regarding the association
between PA and depression is even more ambiguous in adolescents than in adults. While
cross-sectional and intervention studies in adolescents have shown moderate to strong as-
sociations between PA and depression respectively (Monshouwer, ten Have, van Poppel,
Kemper, & Vollebergh, 2009; Kirkcaldy, Shephard, & Siefen, 2002), longitudinal studies have
yielded mixed results varying from strong effects (Patten et al., 2009; Motl, Birnbaum, Ku-
bik, & Dishman, 2004) to weak (Jerstad et al., 2010; Sagatun, Sogaard, Bjertness, Selmer,
& Heyerdahl, 2007) or even no effects at all (Birkeland et al., 2009; Ströhle et al., 2007).
Second, a substantial number of adolescents evince sub-threshold depressive symptoms
(Pine, Cohen, Cohen, & Brook, 1999). Since adolescent depressive symptoms are thought to
predict full-blown depressive disorders later in life (Pine et al., 1999), researching depressive
symptoms in adolescence may help to improve the understanding of the etiology of depres-
sive disorders and help design effective interventions for treating affective disorders.
The first aim of this study was to investigate patterns in the association between PA and
depressive symptoms, in order to determine whether changes in PA precede, follow or co-
occur with changes in depressive symptoms in healthy adolescents. A longitudinal design
is the only way to illuminate such patterns. The second aim was to examine the association
between PA and two subgroups of depressive symptoms, that is, affective symptoms such
as depressed mood and loss of interest, and somatic symptoms such as sleep disturbances
and lack of energy. To the best of our knowledge, PA has not been investigated in relation to
these symptom groups before.
33PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
METHODS
Design
Data for this study were collected as part of the Tracking Adolescents' Individual Lives Sur-
vey (TRAILS), a prospective cohort of Northern Dutch adolescents. The study was approved
by the Dutch Central Committee on Research Involving Human Subjects. A detailed descrip-
tion of its objectives and main design, as well as of the sample selection procedure and non-
response, can be found elsewhere (de Winter et al., 2005).
Data collection took part in three assessment waves (T1, T2, T3), which ran from March
2001 to July 2002; September 2003 to December 2004; and September 2005 to August 2007
respectively. Parental written informed consent was obtained after the procedures had been
fully explained. Adolescents gave written informed assent at the second and third assess-
ment wave.
Participants
Of 2935 adolescents initially approached, 2230 (76%; girls = 51%, mean age = 11.11, SD = 0.55)
took part in the first assessment wave (T1). The response rate at the second wave (T2) was
96.4% (N = 2149; 51% girls, mean age = 13.65, SD = 0.53), while the response rate was 81.4%
(N = 1816; 52.3% girls, mean age = 16.27, SD = 0.73) at the third assessment wave (T3).
Measures
Depressive Symptoms
Depressive symptoms were assessed by the Affective Problems scale of the Youth Self Re-
port (YSR; Achenbach, 1991b), which was completed by the adolescents at school. Parents
filled out the parent version of the YSR, the Child Behavior Checklist (CBCL; Achenbach,
1991a) at home. The mean scores of the YSR and CBCL scales were used in the analyses.
The YSR and CBCL are composed of a list of problems that are scored on a three-point
scale (0 = never or not at all true, 1 = sometimes true and 2 = very often or very true). The
Affective Problems scale contains 13 items covering depressive symptoms according to the
DSM-IV (Achenbach, Dumenci, & Rescorla, 2003), that is, sadness, loss of pleasure, crying,
self-harm, suicidal ideation, feelings of worthlessness, guilt, loss of energy, overtiredness,
eating problems and sleeping problems. Because a previous study in the same sample had
shown that omission of one sleep item (‘I sleep more than most kids’) increased the internal
consistency of the scale (Bosch et al., 2009), this item was excluded. Scores on the remain-
ing twelve items were averaged to construct a total depressive symptoms scale with an
internal consistency (Cronbach’s alpha) of 0.75 at T1, of 0.84 at T2 and 0.89 at T3. In addition,
CHAPTER II34
we constructed a somatic symptoms subscale and an affective symptoms subscale, as de-
scribed by Bosch et al. (2009). The affective symptom subscale included loss of pleasure,
crying, self-harm, suicidal ideation, feelings of worthlessness, feelings of guilt and sadness,
while the somatic symptoms subscale included lack of appetite, overtiredness, reduced
sleep, trouble sleeping and lack of energy. The internal consistency for the affective symp-
toms scale was 0.72 at T1, 0.82 at T2 and 0.85 at T3, while the consistencies of the somatic
symptoms scale were 0.57 at T1, 0.69 at T2 and 0.77 at T3.
Physical Activity
At T1, PA was assessed by the question: “How often do you perform physical exercise (for
example, swimming, playing football, horse-riding)?”. The question could be answered on
a five-point scale (0 = never, 1 = once per week, 2 = two to three times per week, 3 = four to
five times per week, and 4 = five to six times per week). At T2 and T3, PA was measured by
two questions assessing the time per week the individual engaged in PA during the summer
(question 1) and the winter (question 2): “How many days in an average week in the summer/
winter do you take part in physical activities?”. These questions were rated on an eight-
point scale (0 = never, to 7 = seven days per week). The questions on summer and winter were
averaged and mean scores were used in the analysis. To achieve a similar metric for the PA
measures at all measurement waves, the T1 data (ranging from 0-4) were recoded into an
eight-point scale using the monotonic R7 transform recommended by Little (in press).
Covariates
Gender and SES were included as covariates. SES was calculated by averaging five stan-
dardized variables (professional occupation and educational attainment for both father and
mother, and household income). Three SES groups were created, where the lowest 25%
of scores were categorized as ‘low SES’, the highest 25% as ‘high SES’ and the remaining
scores were grouped as ‘intermediate SES’.
Analysis
Imputation of Missing Data
A detailed account of attrition rates within the TRAILS study can be found elsewhere (Huis-
man et al., 2008). We performed multiple imputations to complete the dataset (i.e., N = 2230
at all three assessment waves) using STATA version 10 (STATA corp., College station, Texas)
and created five different datasets (Rubin, 1996), which were subsequently imported into
Mplus 5 (Muthén & Muthén, 2007) and used in the analyses.
35PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
Cross-Lagged Path Model
Cross-sectional and prospective associations between PA and depressive symptoms were in-
vestigated by means of Structural Equation Modeling (SEM), using Mplus 5 (Muthén & Muth-
én, 2007). Since the depression variables (both total depressive symptoms score and scores
on the two subscales) were positively skewed and slightly kurtotic, they were log trans-
formed prior to analysis. A cross-lagged path model was used to investigate cross-sectional
and prospective associations between PA and depressive symptoms. Cross-lagged panel
designs take into account the time precedence and control for multivariate dependencies of
the antecedent predictor variables. Therefore, besides the cross-lagged paths (interpreted as
a linkage of the level of one variable at the first wave with a relative change in another vari-
able in the subsequent assessment wave) between depression and PA and vice versa (Figure
1, path c), autoregressive paths (interpreted as relative stability over time) within depression
and PA (Figure 1, path b), and cross-sectional covariances (interpreted as correlations at T1
and as correlated change at T2 and T3) between PA and depression (Figure 1, path a) were
also estimated. Correlated change and cross-lagged paths reflect longitudinal relationships.
In order to test the predictive relationships between PA and depressive symptoms we devel-
oped two cross-lagged models. The first model concerned associations between PA and the
total depressive symptoms score, while the second model concerned associations between
b
b
b
b
c c
c c
a a
DepressionT1
PA T1
DepressionT2
PA T2
DepressionT3
PA T3
a
b
b
Figure 1. Model showing Cross-Sectional Paths (a), Autoregressive (b), and Cross-Lagged Longitudinal Paths (c).
PA = physical activity, T1 = first measurement wave, T2 = second measurement wave and T3 = third measurement wave.
CHAPTER II36
PA and the subscales of somatic and affective symptoms (adjusted for each other). Both mod-
els were corrected for gender and SES. As a first step, all cross-lagged and cross-sectional
paths were constrained to be equal over time, and the fit indices of the constrained model
were compared to those of the unconstrained model. If the fit indices were adequate and
not significantly worse than those of the unconstrained model, the more parsimonious, i.e.,
constrained, model was used. The significance of the cross-lagged paths was established by
testing if removal of the paths led to a deterioration of the model fit. To test whether the two
subgroups of depressive symptoms (somatic and affective symptoms) differed with regard to
their cross-sectional and cross-lagged association with PA, we compared a model where the
paths between somatic symptoms and PA and affective symptoms and PA respectively, were
constrained to be equal with a model where these paths were allowed to differ.
A good model fit was defined when the Comparative Fit Index (CFI) and the Tucker-Lew-
is Index (TLI) were greater than 0.95 while the Root Mean Square Error of Approximation
(RMSEA) was lower than 0.05. Ideally, the χ2 should be non-significant (p>0.05) as well, but
larger samples increase the likelihood of obtaining significant p-values (Bentler, 1990).
RESULTS
Descriptive Statistics
Descriptive statistics of the main variables used in this study are shown in Table 1. Depres-
sive symptoms increased slightly over time, mainly due to a rise in somatic symptoms. PA
remained stable over time.
Table 1. Means (Standard Deviations) of the Main Variables.
T1 (mean age 11.1)
T2 (mean age 13.7)
T3 (mean age 16.3)
Total Depressive Symptoms
0.25 (0.20) 0.25 (0.23) 0.31 (0.29)
Affective Symptoms 0.21 (0.20) 0.19 (0.23) 0.23 (0.28)
Somatic Symptoms 0.32 (0.27) 0.33 (0.30) 0.42 (0.36)
Physical Activity 3.38 (2.09) 3.35 (1.73) 3.45 (1.88)
N = 2230 for all measurements since missing data were imputed. T1 = first measurement wave, T2 = second measurement wave and T3 = third measurement wave.
37PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
Relationships between Depression and PA
Preliminary analyses revealed that cross-lagged paths and correlated changes were con-
stant over time: the model fit parameters of the models where these paths were constrained
to be equal across time were adequate and the χ2-values of the constrained models were not
significantly worse than the χ2-values of the unconstrained model (χ2difference = 5.555, df = 3,
p>0.05 for the model involving total depressive symptoms; χ2difference = 6.365, df = 6, p>0.05
for the model involving depression subgroups). Because the constrained models were more
parsimonious we used these models in all subsequent analyses.
Relationships between the total depressive symptoms score and PA are shown in Figure 2 (over-
all model fit, χ2 = 13.254, df = 6, p<0.05). Fit indices showed an excellent fit of model (CFI = 0.997,
TLI = 0.988, RMSEA = 0.022), which was significantly better than the fit of the model without
cross-lagged paths from PA to depression (χ2difference = 8.103, df = 1, p<0.05) and the model with-
out cross-lagged paths from depression to PA (χ2difference = 11.556, df = 1, p<0.05). As expected
from these results, the cross-lagged paths were all significant. The autoregressive estimate of
depression was higher than the autoregressive estimate of PA. The cross-sectional association
between PA and depression was significant at T1 but not at later assessment waves.
Figure 2. Associations between Depression (Total Depressive Symptoms Score) and PA over time.
.586* (.015)
.229* (.023)
.556* (.020)
.272* (.023)
-.039* (.016)
-.033* (.010)
-.047* (.015)
-.042* (.017)
-.046 (.026)
-.039 (.022)
DepressionT1
PA T1
DepressionT2
PA T2
DepressionT3
PA T3
-.078* (.022)
.120* (.024)
.135* (.025)
PA = physical activity, T1 = first measurement wave, T2 = second measurement wave and T3 = third measurement wave.The path coefficients reflect model estimated standardized beta values. Statistical insignificant paths shown are indicated by broken lines. The effects are adjusted for gender and socioeconomic status. * p<0.05.
CHAPTER II38
The second model, investigating the relationship between somatic symptoms, affective
symptoms and PA over time, is shown in Figure 3. Indices of approximate fit showed an
excellent fit of model variables (χ2 = 24.170, df = 13, p<0.05, CFI = 0.998, TLI = 0.992, RM-
SEA = 0.020). While cross-lagged relationships between PA and somatic symptoms did not
reach statistical significance, affective symptoms both predicted and were predicted by PA.
This is reflected in the tests comparing the models with and without these cross-lagged
paths: whereas removal of the cross-lagged to and from somatic symptoms did not worsen
the model fit significantly (χ2difference = 2.404, df = 2, p>0.05), removal of the paths to and from
affective symptoms led to a significant decrease of the model fit (χ2difference = 26.755, df = 2,
p<0.05).
.479* (.019)
.228* (.024)
.448* (.023)
.406* (.024)
.272* (.024)
.457* (.035)
.002 (.018)
-.062* (.016)
-.015 (.010)
-.049* (.018)
-.021 (.014)
-.048* (.017)
-.002 (.018)
-.044* (.012)
-.035 (.019)
-.038 (.028)
-.032 (.017)
-.034 (.025)
Somatic T1
PA T1
Affective T1
Somatic T2
PA T2
Affective T2
Somatic T3
PA T3
Affective T3
-.046* (.022)
-.088* (.022)
Figure 3. Association between Somatic Symptoms, Affective Symptoms and PA over time.
PA = physical activity, T1 = first measurement wave, T2 = second measurement wave and T3 = third measurement wave.The path coefficients reflect model estimated standardized beta values. Statistically insignificant paths are indicated by broken lines. The effects are adjusted for gender and socioeconomic status. * p<0.05. Paths between somatic and affective symptoms although defined in the model (cross-sectional, autoregressive between T1 and T3, and cross-lagged) were omitted from this figure for clarity purposes.
39PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
DISCUSSION
The results of this study indicate a bidirectional cross-lagged association between depres-
sive symptoms and PA, in which PA precedes a decrease in depressive symptoms and vice
versa. Quite surprisingly, the inverse associations between PA and depressive symptoms
concerned affective symptoms such as depressed mood, loss of pleasure and low self-worth
in particular. Somatic symptoms such as sleep disturbances, eating problems and lack of en-
ergy were not associated with PA. Interestingly, we found that PA remained stable over time,
a finding which is inconsistent with some previous studies, which reported a steady decrease
of PA over time (Riddoch et al., 2004).
Limitations and Strengths of this Study
This study has a number of notable strengths. First, we used a large, population-based
sample of adolescents. Studying adolescents has two main advantages: a) the probability
of confounding by prior depressive episodes is low at this age, and b) the prevalence of so-
matic conditions that may hamper engagement in PA is still relatively low in adolescents
compared to older people. Second, the longitudinal design with three measurements across
6 years made it possible to study changes in PA and depressed mood over a long period of
time. Third, the model fit was excellent, modeling the relationship between PA and depres-
sive symptoms in our data with great accuracy. A final strength is the fact that we subdi-
vided depressive symptoms into somatic and affective symptoms, which, to the best of our
knowledge has not been done before in this context. Depression is a heterogeneous disorder,
with different symptoms observed in different individuals. The use of subgroups does justice
to this heterogeneity and may help to understand individual differences in the association
between PA and depression.
There are also limitations. First, all of the measures used were self-reports and the mea-
surement of PA relied on a single question, unlike validated PA questionnaires (e.g., IPAQ),
which include a series of detailed questions of PA involvement. Self-reports tend to be less
reliable than more objective measures of PA, such as VO2 max, heart rate variability and
calculation of Metabolic Equivalent of Tasks by, for instance, accelerometers. In addition,
at the first measurement wave of this study, PA was assessed slightly different than at the
last two waves, in which separate questions for winter and summer, and more response cat-
egories were used. Furthermore, we did not measure the nature (aerobic-anaerobic, social-
isolated etc.), intensity, or duration of PA, information which may prove useful for a better
understanding of the relationship between PA and depressive symptoms. Concerning the
nature of PA, for instance, aerobic exercises have been reported to have an effect on depres-
sive symptoms while anaerobic exercises do not (Goodwin, 2003). Similarly, differences in the
CHAPTER II40
intensity and the duration of the activity may be related to depressive symptoms (Galper,
Trivedi, Barlow, Dunn, & Kampert, 2006). A final limitation is that the reliability of the so-
matic symptoms subscale was relatively low, which prevents strong associations with other
variables.
The associations observed were bidirectional. The more times per week adolescents en-
gaged in PA, the more likely they were to report a reduction in depressive symptoms. The
opposite also held true: the more depressive symptoms adolescents exhibited at some point
in time, the more likely they were to report a reduction in PA later on. These findings are in
accordance with recent reviews in adults which showed that regular PA decreases the risk
for developing depression (Teychenne, Ball, & Salmon, 2008) and that baseline depression
might play an important role in the development of an inactive lifestyle or decreased level
of PA (Roshanaei-Moghaddam, Katon, & Russo, 2009). In adolescents, longitudinal stud-
ies have yielded mixed results. Whereas Jerstad et al. (2010) demonstrated a bidirectional
relationship between exercise and depression, Birkerland and colleagues (2009) could not.
The bidirectional association between PA and depression found in our study provide support
for both the Protection hypothesis and Inhibition hypothesis, which were mentioned earlier
in the Introduction.
We also showed that PA is differentially associated with subgroups of depressive symp-
toms; the effects are stronger for affective than for somatic depressive symptoms. This lack
of association between PA and somatic symptoms is unexpected. Although it seems plau-
sible that adolescents who exercise more sleep better, eat better and have more energy,
this was not supported by our data. A reason for this lack of association (between PA and
somatic depressive symptoms) may be the before-mentioned low reliability of the somatic
symptoms scale, which prevents strong relationships. The low reliability implies that there
is heterogeneity within the somatic symptom domain. Whether specific somatic symptoms
are differentially related to PA was beyond the scope of this paper and remains to be inves-
tigated in the future.
The inverse associations found between PA and depression were statistically significant
but were relatively weak. This is generally in accordance with previous studies, which rarely
yielded strong effects. This could be due to the measures of PA used, which usually reflect
only a small portion of activities and do not take into account the nature of the activities
(e.g., voluntary, social, competitive or isolated), their duration and their intensity. PA may be
beneficial to mental health only after an intensity/duration threshold is reached, which could
have weakened the associations. Another likely reason for the weak associations is the fact
that both depressed mood and PA are affected by multiple non-shared variables and there-
fore cannot be expected to explain much variance in each other. Finally, it might be possible
that PA and depressive symptoms are not associated in all individuals in the same way; due
41PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
to genetic or personality differences, some adolescents may benefit from PA while others do
not. An elucidation of the underlying mechanisms might provide clues about which individu-
als may benefit most from PA.
CONCLUSION
The current prospective study investigated the inverse bidirectional associations between
physical activity and depressive symptoms, in particular the two sub-clusters of depres-
sive symptoms (somatic vs. affective), in a population cohort of Dutch adolescents. Physical
activity has been shown to be related to depressive symptoms and vice versa but only in
relation to affective symptoms. Further research is needed to investigate whether individual
symptoms are differentially associated with physical activity, in order to improve the under-
standing of this complex relation.
CHAPTER II42
REFERENCES
Achenbach, T. M. (1991a). Manual for the child behavior checklist/ 4-18 and 1991 profile. University of Ver-mont, Burlington.
Achenbach, T. M. (1991b). Manual for the youth self report and 1991 profile. University of Vermont, Burlington.
Achenbach, T. M., Dumenci, L., & Rescorla, L. A. (2003). DSM-oriented and empirically based approaches to constructing scales from the same item pools. Journal of Clinical Child and Adolescent Psychology, 32(3), 328-340.
Allgower, A., Wardle, J., & Steptoe, A. (2001). Depressive symptoms, social support, and personal health behaviors in young men and women. Health Psychology, 20(3), 223-227.
Babyak, M., Blumenthal, J. A., Herman, S., Khatri, P., Doraiswamy, M., Moore, K., et al. (2000). Exercise treat-ment for major depression: Maintenance of therapeutic benefit at 10 months. Psychosomatic Medicine, 62(5), 633-638.
Bardia, A., Hartmann, L. C., Vachon, C. M., Vierkant, R. A., Wang, A. H., Olson, J. E., et al. (2006). Recreational physical activity and risk of postmenopausal breast cancer based on hormone receptor status. Archives of Internal Medicine, 166(22), 2478-2483.
Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107(2), 238-246.
Birkeland, M. S., Torsheim, T., & Wold, B. (2009). A longitudinal study of the relationship between leisure-time physical activity and depressed mood among adolescents. Psychology of Sport and Exercise, 10(1), 25-34.
Bosch, N. M., Riese, H., Dietrich, A., Ormel, J., Verhulst, F. C., & Oldehinkel, A. J. (2009). Preadolescents’ so-matic and cognitive-affective depressive symptoms are differentially related to cardiac autonomic func-tion and cortisol: The TRAILS study. Psychosomatic Medicine, 71(9), 944-950.
Cooper-Patrick, L., Ford, D. E., Mead, L. A., Chang, P. P., & Klag, M. J. (1997). Exercise and depression in midlife: A prospective study. American Journal of Public Health, 87(4), 670-673.
de Jonge, P., Ormel, J., van den Brink, R. H., van Melle, J. P., Spijkerman, T. A., Kuijper, A., et al. (2006). Symp-tom dimensions of depression following myocardial infarction and their relationship with somatic health status and cardiovascular prognosis. The American Journal of Psychiatry, 163(1), 138-144.
de Winter, A. F., Oldehinkel, A. J., Veenstra, R., Brunnekreef, J. A., Verhulst, F. C., & Ormel, J. (2005). Evalua-tion of non-response bias in mental health determinants and outcomes in a large sample of pre-adoles-cents. European Journal of Epidemiology, 20(2), 173-181.
Galper, D. I., Trivedi, M. H., Barlow, C. E., Dunn, A. L., & Kampert, J. B. (2006). Inverse association between physical inactivity and mental health in men and women. Medicine and Science in Sports and Exercise, 38(1), 173-178.
Goodwin, R. D. (2003). Association between physical activity and mental disorders among adults in the United States. Preventive Medicine, 36(6), 698-703.
Haboush, A., Floyd, M., Caron, J., LaSota, M., & Alvarez, K. (2006). Ballroom dance lessons for geriatric depression: An exploratory study. The Arts in Psychotherapy, 33(2), 89-97.
Harris, A. H., Cronkite, R., & Moos, R. (2006). Physical activity, exercise coping, and depression in a 10-year cohort study of depressed patients. Journal of Affective Disorders, 93(1-3), 79-85.
Howard, R. A., Freedman, D. M., Park, Y., Hollenbeck, A., Schatzkin, A., & Leitzmann, M. F. (2008). Physical activity, sedentary behavior, and the risk of colon and rectal cancer in the NIH-AARP Diet and Health Study. Cancer Causes & Control, 19(9), 939-953.
Hu, G., Lindstrom, J., Valle, T. T., Eriksson, J. G., Jousilahti, P., Silventoinen, K., et al. (2004). Physical activ-ity, body mass index, and risk of type 2 diabetes in patients with normal or impaired glucose regulation. Archives of Internal Medicine, 164(8), 892-896.
43PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
Huisman, M., Oldehinkel, A. J., de Winter, A. F., Minderaa, R. B., de Bildt, A., Huizink, A. C., et al. (2008). Cohort profile: The Dutch ‘TRacking Adolescents’ Individual Lives’ Survey’; TRAILS. International Journal of Epidemiology, 37(6), 1227-1235.
Jerstad, S. J., Boutelle, K. N., Ness, K. K., & Stice, E. (2010). Prospective reciprocal relations between physi-cal activity and depression in female adolescents. Journal of Consulting and Clinical Psychology, 78(2), 268-272.
Kirkcaldy, B. D., Shephard, R. J., & Siefen, R. G. (2002). The relationship between physical activity and self-image and problem behaviour among adolescents. Social Psychiatry and Psychiatric Epidemiology, 37(11), 544-550.
Little, T. D. (in press). Longitudinal structural equation modeling: Individual difference panel models. New York: Guildford press.
Mead, G. E., Morley, W., Campbell, P., Greig, C. A., McMurdo, M., & Lawlor, D. A. (2009). Exercise for depres-sion. Cochrane Database of Systematic Reviews (Online), (3), CD004366.
Monshouwer, K., ten Have, M., van Poppel, M., Kemper, H., & Vollebergh, W. (2009). Low physical activity in adolescence is associated with increased risk for mental health problems. Medicina Sportiva, 13(2), 74-81.
Motl, R. W., Birnbaum, A. S., Kubik, M. Y., & Dishman, R. K. (2004). Naturally occurring changes in physical activity are inversely related to depressive symptoms during early adolescence. Psychosomatic Medi-cine, 66(3), 336-342.
Muthén, L. K., & Muthén, B. O. (2007). Mplus user’s guide: Statistical analysis with latent variables (5th ed.). Los Angeles, CA: Muthén & Muthén.
Patten, S. B., Williams, J. V., Lavorato, D. H., & Eliasziw, M. (2009). A longitudinal community study of major depression and physical activity. General Hospital Psychiatry, 31(6), 571-575.
Pine, D. S., Cohen, E., Cohen, P., & Brook, J. (1999). Adolescent depressive symptoms as predictors of adult depression: Moodiness or mood disorder? The American Journal of Psychiatry, 156(1), 133-135.
Riddoch, C. J., Bo Andersen, L., Wedderkopp, N., Harro, M., Klasson-Heggebo, L., Sardinha, L. B., et al. (2004). Physical activity levels and patterns of 9- and 15-yr-old European children. Medicine and Science in Sports and Exercise, 36(1), 86-92.
Roshanaei-Moghaddam, B., Katon, W. J., & Russo, J. (2009). The longitudinal effects of depression on physi-cal activity. General Hospital Psychiatry, 31(4), 306-315.
Rubin, B. D. (1996). Multiple imputation after 18+ years. Journal of the American Statistical Association, 91(434), 473.
Sagatun, A., Sogaard, A. J., Bjertness, E., Selmer, R., & Heyerdahl, S. (2007). The association between week-ly hours of physical activity and mental health: A three-year follow-up study of 15-16-year-old students in the city of Oslo, Norway. BMC Public Health, 7, 155.
Salmon, P. (2001). Effects of physical exercise on anxiety, depression, and sensitivity to stress: A unifying theory. Clinical Psychology Review, 21(1), 33-61.
Ströhle, A., Hofler, M., Pfister, H., Muller, A. G., Hoyer, J., Wittchen, H. U., et al. (2007). Physical activity and prevalence and incidence of mental disorders in adolescents and young adults. Psychological Medicine, 37(11), 1657-1666.
Struder, H. K., & Weicker, H. (2001). Physiology and pathophysiology of the serotonergic system and its im-plications on mental and physical performance. Part II. International Journal of Sports Medicine, 22(7), 482-497.
Stubbe, J. H., de Moor, M. H., Boomsma, D. I., & de Geus, E. J. C. (2007). The association between exercise participation and well-being: A co-twin study. Preventive Medicine, 44(2), 148-152.
Teychenne, M., Ball, K., & Salmon, J. (2008). Physical activity and likelihood of depression in adults: A re-view. Preventive Medicine, 46(5), 397-411.
CHAPTER II44
Wessel, T. R., Arant, C. B., Olson, M. B., Johnson, B. D., Reis, S. E., Sharaf, B. L., et al. (2004). Relationship of physical fitness vs body mass index with coronary artery disease and cardiovascular events in women. JAMA: The Journal of the American Medical Association, 292(10), 1179-1187.
Weyerer, S. (1992). Physical inactivity and depression in the community. Evidence from the Upper Bavarian Field Study. International Journal of Sports Medicine, 13(6), 492-496.
Physical Activity and Onset of Depression in Adolescents: A Prospective Study in the General Population Cohort TRAILS
“Living with depression is like trying to keep your balance while you dance with a goat—it is perfectly sane to prefer a partner with a better sense of balance”
Andrew Solomon, The Noonday Demon, 2001
Stavrakakis N, Roest AM, Verhulst FC, Ormel J, de Jonge P & Oldehinkel AJ
J Psychiatr Res. 2013; 47(10):1304-8
CHAPTER
CHAPTER III46
ABSTRACT
Purpose: Although it has often been suggested that physical activity and depression are
intertwined, only few studies have investigated whether specific aspects of physical activ-
ity predict the incidence of major depression in adolescents from the general population.
Therefore the aim of this study was to investigate the effects of nature, frequency, duration
and intensity of physical activity during early adolescence on the onset of a major depressive
episode in early adulthood.
Methods: In a population sample of adolescents (N = 1396), various aspects of physical ac-
tivity were assessed at early adolescence (mean age = 13.02, SD = 0.61). Major depressive
episode onset was assessed using the Composite International Diagnostic Interview. A Cox
regression model was performed to investigate whether physical activity characteristics
and their interactions with gender predicted a major depressive episode onset up until mean
age 18.5 (SD = 0.61).
Results: The individual characteristics of physical activity (nature, frequency, duration and
intensity) or their interactions with gender did not predict a major depressive episode onset
(p-values>0.05).
Conclusion: So far, there is no prospective evidence that physical activity protects against
the development of adolescent depressive episodes in either boys or girls.
47PHYSICAL ACTIVITY AND ONSET OF DEPRESSION
INTRODUCTION
Physical activity (PA) has been shown to improve both physical and mental health, espe-
cially depression, in adults (Deslandes et al., 2009; Dunn, Trivedi, & O’Neal, 2001; Penedo &
Dahn, 2005). The possibility that PA can reduce the risk of depression is particularly relevant
in adolescence, a period of life characterized by a high incidence of depression (Hankin et al.,
1998). However, only few studies have investigated the association between PA and depres-
sion in adolescents, and the results have been inconsistent so far (for a detailed review see:
Biddle & Asare, 2011). A possible reason for these inconsistencies is that various aspects
of PA affect the development of depression in diverse ways (Allison et al., 2005; Johnson
& Taliaferro, 2011; Tao et al., 2007). To our knowledge, there are no studies on the specific
effects of nature, duration and frequency of PA on depression in adolescents. A few stud-
ies have investigated the effects of PA intensity on depressive symptoms in adolescents,
but their results are equivocal (Allison et al., 2005; Brand et al., 2010; Goldfield et al., 2011;
Johnson & Taliaferro, 2011; Tao et al., 2007). Allison et al. (2005) found that vigorous PA was
inversely associated with problems in social functioning, but it was not related to depres-
sion or anxiety after adjustment for age, gender and socioeconomic status (SES). Tao et al.
(2007) found that light to moderate PA protected against the development of depressive
symptoms, whereas high intensity PA was a risk factor for suicidal ideation and psychologi-
cal distress. However, Tao et al.’s (2007) findings might be explained by cultural differences,
since the population target was on Chinese adolescents. In contrast, Goldfield et al. (2011)
found that light to moderate intensity PA was not related to depressive symptoms, whereas
high intensity PA was associated with reduced depressive symptoms in boys. Goldfield’s
findings suggest that gender differences might be a source of heterogeneity in effects as
well (Goldfield et al., 2011). Finally, Brand et al. (2010) found that adolescent athletes who
exercised vigorously reported improved sleep and psychological functioning, including lower
depressed mood than non-athletes. Another reason for the lack of clarity regarding the in-
fluence PA may have on adolescent depression is that most associations tested were cross-
sectional (see reviews: Biddle & Asare, 2011 and Johnson & Taliaferro, 2011). One study
found an inverse association between childhood PA and depression in late adulthood (Jacka
et al., 2011) but this finding was based on retrospective reports of childhood PA and did not
involve specific aspects of PA. Therefore, it remains to be elucidated whether specific as-
pects of PA have the potential to prevent depressive episodes in adolescents, and whether
this preventive potential is equal for boys and girls.
CHAPTER III48
Aim
The aim of this study was to investigate the prospective association of the nature (leisure-
time PA), frequency, duration and intensity of PA during early adolescence with the onset
of major depression during adolescence, and gender differences therein. We hypothesized
that the amount of leisure-time PA, frequency, duration and intensity would be negatively
associated with the probability to develop a major depressive episode (MDE) and that the
associations would be stronger in boys than in girls.
METHODS
Design
This study is part of the Tracking Adolescents’ Individual Lives Survey (TRAILS), a Dutch
prospective cohort study of adolescents. The Dutch Central Committee on Research In-
volving Human subjects approved the study to be undertaken. Data collection for the first
measurement wave (T1) started in 2001, while for the remaining waves data collection took
place at intervals of approximately 2.5 years. Previous TRAILS publications (de Winter et al.,
2005; Huisman et al., 2008; Ormel et al., 2012) provide extensive description(s) of TRAILS
main design, data collection and sample selection. Informed consent was collected from the
parents at T1, while for the second (T2) and third wave (T3) informed consent was obtained
from both parents and participants. In the fourth wave (T4), the participants had reached the
age of 18 and parental consent was not requested (Ormel et al., 2012). In this study we used
data from T2 and T4.
Participants
At baseline, 2935 adolescents were requested to participate, of whom 2230 (response rate
76.0%; 51% girls, mean age = 11.1, SD = 0.55) agreed to do so. The response rates at T2 and
T4 were 96.4% (N = 2149; 51% girls, mean age = 13.65, SD = 0.53), and 84.3% (N = 1881; 52.3%
girls, mean age 19.1, SD = 0.60) respectively.
Procedure
During T2 questionnaires were sent through the mail to the parents or guardians, while
the adolescents completed questionnaires at school. During T4, the adolescents filled out
a web-based questionnaire, and were interviewed at home by a trained test assistant (for
more information on recruitment procedures for all waves see: Nederhof et al., 2012).
49PHYSICAL ACTIVITY AND ONSET OF DEPRESSION
Measures
Physical Activity
At T2, adolescents were asked to rate how often they engaged in PA during an average week:
“How many days in an average week do you take part in physical activities?”. These ques-
tions were rated on an eight-point scale ranging from 0 = never, up to 7 = 7 days per week.
Furthermore, they could indicate the frequency and amount of time spent on up to four
specific sports. Additional questions concerned the frequency and duration of walking or
cycling to school or work and walking or cycling in free time. The percentage of leisure time
PA (% LTPA) was calculated as a percentage of time spent in leisure time compared to the
time spent on walking and cycling to school or work. Frequency was defined as the number
of times the participant engaged in PA and duration (in minutes) of the PA per time, regard-
less of the nature of the PA. If participants reported walking or cycling both as a sport and in
their free time, we only used the sports information to calculate the frequency and duration
to avoid duplicate answers. Based on visual inspection of the duration data, participants
who reported spending more than 225 minutes on a single activity were considered outliers
and excluded from the analyses. The intensity of PA was expressed in metabolic equivalent
of task (MET) scores, which were assigned to each sport using the MET score 2011 compen-
dium (Ainsworth et al., 2011). Based on these scores we created two intensity measures:
the peak MET score (the highest MET score) and the mean MET score across all activities. In
case a participant did not report any sport, the MET scoring of cycling or walking in leisure
time was used.
Depression
Depression was operationalized as a DSM-IV MDE and assessed using the World Health Or-
ganization Composite International Diagnostic Interview (CIDI) version 3.0 (Kessler & Ustun,
2004), which was administered at T4. The CIDI is a structured diagnostic interview, which
yields the occurrence and age at onset of psychiatric disorders according to DSM-IV criteria,
and has been shown to have good reliability and validity (Kessler et al., 2004). Because we
aimed to investigate the effect of PA on the incidence of depression, adolescents who had
developed an MDE before T2 (n = 103) were excluded from the analyses.
Socioeconomic Status
A measure of SES was obtained by averaging five standardized variables, describing the
educational attainment and professional occupation of both parents and household income.
CHAPTER III50
Statistical Analysis
All analyses were conducted using the Statistical Package for the Social Sciences (SPSS Inc.,
Chicago, IL.) version 20. Two-tailed tests with an alpha-level of 0.05 were performed. Cox re-
gression analysis was used to investigate whether the T2 PA measures (%LTPA, frequency,
duration and intensity) predicted the onset of an MDE between T2 and T4. SES and gender
were included as covariates. The proportional hazards (PH) assumption was tested using
the Schoenfield residuals. In case this test was significant for any of the predictor variables
an extended Cox model was used to examine whether that variable interacted significantly
with the time variable. If the interaction was not significant a normal Cox regression model
was used. First, we tested a model including the main effects of the PA measures (%LTPA,
frequency, duration, intensity) and covariates as predictors. After that, we added interactions
of each of the PA measures with gender one by one. If the interaction term was significant,
it was kept in the model. Finally, we examined to what extent the exclusion of adolescents
due to outliers had influenced the results by repeating the analyses including these cases.
RESULTS
Descriptive Statistics
Of all T4 participants (n = 1881), 408 were excluded because they either did not complete the
CIDI interviews or had developed an MDE before T2. A further 18 participants were excluded
from the final analysis because the values reported for the average duration per activity
were considered outliers. Finally, 59 participants had missing values on the PA and SES
questions and therefore were also excluded from the analysis. The final sample consisted of
1396 participants (mean age at T2 = 13.02, SD = 0.61; mean age at T4 = 18.53, SD = 0.61).
739 (52.9%) girls and 657 boys, of whom 101 (13.6%) and 39 (5.9%) respectively, reported an
MDE onset between T2 and T4. Table 1 shows the total number of minutes spent on PA per
week, by gender, SES category, and MDE. The percentage spent on LTPA, the weekly fre-
quencies of PA, the duration in minutes spent per activity, and the mean and peak intensity
of activities for boys and girls are shown in table 2.
Prediction of Depression Onset by Physical Activity
The PA frequency variable violated the proportional hazards assumption (Schoenfield re-
sidual = 0.17, p<0.05), but the extended Cox regression model indicated that the interaction
effect of frequency and time was not significant (p-value = 0.18), so we performed a regular
Cox regression analysis. The results can be seen in Table 3. The main effect of gender was
significant (with adolescent girls being more likely to develop an MDE than boys) but neither
51PHYSICAL ACTIVITY AND ONSET OF DEPRESSION
Table 1. Minutes per week PA, by Gender, SES, and MDE.
Variables (N) Minutes per week of General PA (Median and Interquartile range)
GenderFemale (N = 739)Male (N = 657)
360 (210-735)410 (245-815)
SESLow SES (N = 262)Interm. SES (N = 701)High SES (N = 433)
350 (190-752)375 (225-775)420 (245-763)
MDE after T2Yes (N = 140)No (N = 1256)
380 (200-774)385 (226-765)
PA = physical activity, SES = socioeconomic status, MDE = major depressive episode.
Table 2. PA Nature (LTPA), Frequency, Duration and Intensity according to Gender and in the Total Sample.
Sample size LTPA (%)
PA Frequency (times per week)
PA Duration (time)
PA Intensity (peak)
PA Intensity(mean)
Males (n = 657)MedianInterq. Range
8873-95
11 8-15
4026-59
76-8
32-4
Females (n = 739)MedianInterq. Range
8467-94
108-14
3723-55
64-7
32-4
Total (n = 1396)MedianInterq. Range
8669-94
118-14
3824-58
75-7
32-4
PA = physical activity, %LTPA = percentage of time spent on leisure time physical activities.
CHAPTER III52
any of the PA measures nor their interactions with gender (p-value between 0.13 and 0.74)
significantly predicted MDE onset.
As mentioned earlier, 18 subjects were excluded from the analysis because their average
duration per activity was considered unrealistic (over 4 hours per activity). Including these
subjects in the analyses did not change the results (data upon request).
DISCUSSION
To our knowledge, this is the first study on the prospective relationship between specific
aspects of PA and the risk of developing an MDE in adolescents. Contrary to our expecta-
tions we did not find an effect of any of these aspects, nor interaction effects between PA
aspects and gender.
Studies in adults suggest that PA may help prevent depression (Farmer et al., 1988; Kritz-
Silverstein, Barrett-Connor, & Corbeau, 2001; Strawbridge, Deleger, Roberts, & Kaplan, 2002),
but only few studies investigated the preventive role of PA in adolescents. Adolescence is a
critical period with a high incidence of psychiatric disorders (Hankin et al., 1998). Prospective
studies with assessments before the onset can help to understand the mechanisms that
Table 3. Relationship of PA and MDE.
B SE p-value Hazard Ratio 95% CI
%LTPA -0.16 0.11 0.12 0.85 0.69-1.05
PA Frequency 0.08 0.12 0.50 1.09 0.86-1.38
PA Duration -0.02 0.10 0.81 0.98 0.81-1.18
PA MET Peak -0.20 0.13 0.11 0.82 0.64-1.05
PA MET Mean 0.21 0.16 0.19 1.23 0.90-1.68
Gender1 -0.84 0.19 <0.001 0.43 0.30-0.63
SES2 -0.07 0.11 0.53 0.93 0.75-1.16
n = 1396, all PA variables were z-transformed prior to analysis. PA = physical activity, %LTPA = percentage of leisure-time physical activity, MDE = major depressive episode, MET = metabolic equivalent of tasks, SES = socioeconomic status, SE = standard error, CI = confidence intervals.1 males are the reference group. 2 SES was used as a continuous variable in analyses.
53PHYSICAL ACTIVITY AND ONSET OF DEPRESSION
lead to the development of mental health problems. In a previous study involving the same
sample (Stavrakakis, de Jonge, Ormel, & Oldehinkel, 2012), we studied prospective relation-
ships between general levels of PA and depressive symptoms between age 11 and 16, and
found weak but significant negative associations, in both directions. In the present study, we
focused on onset of depressive disorder and specific aspects of PA, and we did not find evi-
dence of such effects. This might indicate that the association between PA and depressive
symptoms found is mainly caused by the potential of PA to reduce symptoms in those suf-
fering from depression (Mota-Pereira et al., 2011; Pilu et al., 2007; Silveira et al., 2013), or that
PA is more strongly associated with sub-threshold depressive symptoms than with clinical
depression. However, it is still possible that PA can prevent an onset of MDE in late adult-
hood, as reported by Jacka et al. (2011), even though we did not find an effect on adolescent
depression. Alternatively, the different results of these two studies might be due to meth-
odological differences between the two. Although Jacka et al.’s study (2011) had a larger
sample than ours (n = 2152 vs. n = 1396), they only used self-reports for defining depression
and not a validated and reliable measure such as the CIDI. Furthermore, they used a sample
of older persons (median age for males = 55; median age for females = 57) who were asked to
retrospectively calculate their PA levels at the age of 15. This might have introduced error in
their results, since activity levels across the lifespan have been reported to be only weakly
correlated (Friedman et al., 2008). In general, the relationship between PA and depression or
depressive symptoms is a complex one. Networks of genetic and biological factors, person-
ality, SES, and lifestyle habits interact and jointly affect depressive symptoms. In combina-
tion with specific other factors, PA might prevent the onset of a depressive episode.
Although there are indications that males and females benefit differently from different lev-
els of PA participation (Asztalos, De Bourdeaudhuij, & Cardon, 2010; Mikkelsen et al., 2010),
we did not observe an interaction between PA aspects and gender on MDE onset. For adults,
gender differences in depression risk have been suggested with regard to the intensity of
PA; women have been reported to benefit more from low intensity activities and men benefit
more from higher intensity activities. In our study, PA intensity did not prevent the onset of
depression in either gender.
It is important to elucidate the true role of PA on depression, since its effect may have been
overestimated in the literature (Daley & Jolly, 2012). Reviews have suggested that PA is
beneficial in alleviating depression (Dunn et al., 2001; Penedo & Dahn, 2005) but most of
the reviewed findings concerned cross-sectional associations in observational studies,
small samples of clinically depressed patients, short follow-up periods or non-clinical vol-
unteers in intervention studies (Daley & Jolly, 2012). In other words, far-reaching conclusions
have been drawn concerning the ability of PA to treat and prevent depression, while in fact
the empirical basis is less firm than often thought. Our study by no means excludes the
CHAPTER III54
possibility that PA can help to reduce depressive symptoms in healthy individuals as shown
previously (Brand et al., 2010; Sigfusdottir, Asgeirsdottir, Sigurdsson, & Gudjonsson, 2011;
Stavrakakis et al., 2012), but clearly suggests that PA cannot prevent or postpone the de-
velopment of incident depressive episodes in adolescents.
A strength of this study is its large sample size, which provides sufficient power to detect
relevant differences and avoid type II errors. A further advantage is the young age of the
sample; the probability of confounding by prior depressive episodes or somatic conditions is
relatively low in adolescence; and the (retrospective) dating of first MDE onsets is less liable
to recall bias than in older samples. Furthermore, the longitudinal design, with a follow-up
period of over 5 years, made it possible to investigate prospectively whether early PA en-
gagement may protect against MDE onset. Finally, the use of the CIDI interviews, adminis-
tered by trained test assistants, provided standardized measures of DSM-IV depression and
its onset.
However, the results of this study have to be interpreted with some caution due to sev-
eral limitations. First, all PA measures were based on self-reports. Although self-reports
can provide an accurate measure of an individual’s engagement in PA (Haskell, 2012; Mil-
ton, Clemes, & Bull, 2012), more objective measures such as accelerometers are preferred
when possible. Self-reports can lead to an over-estimation of the levels of PA (due to social
pressures) in some participants and therefore partly explain the lack of associations found.
Second, we did not use a validated questionnaire, such as the International Physical Ac-
tivity Questionnaire (IPAQ), and therefore cannot be certain on the trustworthiness of the
responses. Finally, the measure of the nature of PA (%LTPA) is quite difficult to be used
accurately in adolescence, since engagement in occupational PA is not that clear compared
to adulthood. This measure was broad and only measured approximately the percentage of
time spent on leisure-time compared to non-leisure time activities (commuting in this case).
Therefore, conclusions about differential associations of PA nature (leisure-time vs. occupa-
tional activities) on depression onset should be tentative.
CONCLUSION
We did not find any evidence that the nature, frequency, duration or intensity of PA influenced
the probability to develop a depressive episode in adolescence in neither boys nor girls.
55PHYSICAL ACTIVITY AND ONSET OF DEPRESSION
REFERENCES
Ainsworth, B. E., Haskell, W. L., Herrmann, S. D., Meckes, N., Bassett, D. R.,Jr, Tudor-Locke, C., et al. (2011). 2011 compendium of physical activities: A second update of codes and MET values. Medicine and Science in Sports and Exercise, 43(8), 1575-1581.
Allison, K. R., Adlaf, E. M., Irving, H. M., Hatch, J. L., Smith, T. F., Dwyer, J. J., et al. (2005). Relationship of vigorous physical activity to psychologic distress among adolescents. Journal of Adolescent Health, 37(2), 164-166.
Asztalos, M., De Bourdeaudhuij, I., & Cardon, G. (2010). The relationship between physical activity and men-tal health varies across activity intensity levels and dimensions of mental health among women and men. Public Health Nutrition, 13(8), 1207-1214.
Biddle, S. J. H., & Asare, M. (2011). Physical activity and mental health in children and adolescents: A review of reviews. British Journal of Sports Medicine, 45(11), 886-895.
Brand, S., Gerber, M., Beck, J., Hatzinger, M., Puehse, U., & Holsboer-Trachsler, E. (2010). High exercise levels are related to favorable sleep patterns and psychological functioning in adolescents: A comparison of athletes and controls. Journal of Adolescent Health, 46(2), 133-141.
Daley, A., & Jolly, K. (2012). Exercise to treat depression. British Medical Journal (Clinical Research Ed.), 344, e3181.
de Winter, A. F., Oldehinkel, A. J., Veenstra, R., Brunnekreef, J. A., Verhulst, F. C., & Ormel, J. (2005). Evalua-tion of non-response bias in mental health determinants and outcomes in a large sample of pre-adoles-cents. European Journal of Epidemiology, 20(2), 173-181.
Deslandes, A., Moraes, H., Ferreira, C., Veiga, H., Silveira, H., Mouta, R., et al. (2009). Exercise and mental health: Many reasons to move. Neuropsychobiology, 59(4), 191-198.
Dunn, A. L., Trivedi, M. H., & O’Neal, H. A. (2001). Physical activity dose-response effects on outcomes of depression and anxiety. Medicine and Science in Sports and Exercise, 33(6 Suppl), S587-97; discussion 609-10.
Farmer, M. E., Locke, B. Z., Moscicki, E. K., Dannenberg, A. L., Larson, D. B., & Radloff, L. S. (1988). Physical activity and depressive symptoms: The NHANES I epidemiologic follow-up study. American Journal of Epidemiology, 128(6), 1340-1351.
Friedman, H. S., Martin, L. R., Tucker, J. S., Criqui, M. H., Kern, M. L., & Reynolds, C. A. (2008). Stability of physical activity across the lifespan. Journal of Health Psychology, 13(8), 1092-1104.
Goldfield, G. S., Henderson, K., Buchholz, A., Obeid, N., Nguyen, H., & Flament, M. F. (2011). Physical activity and psychological adjustment in adolescents. Journal of Physical Activity & Health, 8(2), 157-163.
Hankin, B. L., Abramson, L. Y., Moffitt, T. E., Silva, P. A., McGee, R., & Angell, K. E. (1998). Development of de-pression from preadolescence to young adulthood: Emerging gender differences in a 10-year longitudinal study. Journal of Abnormal Psychology, 107(1), 128-140.
Haskell, W. L. (2012). Physical activity by self-report: A brief history and future issues. Journal of Physical Activity & Health, 9 Suppl 1, S5-10.
Huisman, M., Oldehinkel, A. J., de Winter, A. F., Minderaa, R. B., de Bildt, A., Huizink, A. C., et al. (2008). Cohort profile: The Dutch ‘TRacking Adolescents’ Individual Lives’ Survey’; TRAILS. International Journal of Epidemiology, 37(6), 1227-1235.
Jacka, F. N., Pasco, J. A., Williams, L. J., Leslie, E. R., Dodd, S., Nicholson, G. C., et al. (2011). Lower levels of physical activity in childhood associated with adult depression. Journal of Science and Medicine in Sport / Sports Medicine Australia, 14(3), 222-226.
Johnson, K. E., & Taliaferro, L. A. (2011). Relationships between physical activity and depressive symptoms among middle and older adolescents: A review of the research literature. Journal for Specialists in Pedi-atric Nursing: JSPN, 16(4), 235-251.
CHAPTER III56
Kessler, R. C., Abelson, J., Demler, O., Escobar, J. I., Gibbon, M., Guyer, M. E., et al. (2004). Clinical calibration of DSM-IV diagnoses in the World Mental Health (WMH) version of the World Health Organization (WHO) Composite International Diagnostic Interview (WMHCIDI). International Journal of Methods in Psychiat-ric Research, 13(2), 122-139.
Kessler, R. C., & Ustun, T. B. (2004). The World Mental Health (WMH) survey initiative version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI). International Journal of Methods in Psychiatric Research, 13(2), 93-121.
Kritz-Silverstein, D., Barrett-Connor, E., & Corbeau, C. (2001). Cross-sectional and prospective study of ex-ercise and depressed mood in the elderly: The Rancho Bernardo Study. American Journal of Epidemiol-ogy, 153(6), 596-603.
Mikkelsen, S. S., Tolstrup, J. S., Flachs, E. M., Mortensen, E. L., Schnohr, P., & Flensborg-Madsen, T. (2010). A cohort study of leisure time physical activity and depression. Preventive Medicine, 51(6), 471-475.
Milton, K., Clemes, S., & Bull, F. (2012). Can a single question provide an accurate measure of physical activ-ity? British Journal of Sports Medicine, 47(1), 44-48.
Mota-Pereira, J., Silverio, J., Carvalho, S., Ribeiro, J. C., Fonte, D., & Ramos, J. (2011). Moderate exercise im-proves depression parameters in treatment-resistant patients with major depressive disorder. Journal of Psychiatric Research, 45(8), 1005-1011.
Nederhof, E., Jorg, F., Raven, D., Veenstra, R., Verhulst, F. C., Ormel, J., et al. (2012). Benefits of extensive recruitment effort persist during follow-ups and are consistent across age group and survey method. The TRAILS study. BMC Medical Research Methodology, 12(1), 93.
Ormel, J., Oldehinkel, A. J., Sijtsema, J., van Oort, F., Raven, D., Veenstra, R., et al. (2012). The TRacking Adolescents’ Individual Lives Survey (TRAILS): Design, current status, and selected findings. Journal of the American Academy of Child and Adolescent Psychiatry, 51(10), 1020-1036.
Penedo, F. J., & Dahn, J. R. (2005). Exercise and well-being: A review of mental and physical health benefits associated with physical activity. Current Opinion in Psychiatry, 18(2), 189-193.
Pilu, A., Sorba, M., Hardoy, M. C., Floris, A. L., Mannu, F., Seruis, M. L., et al. (2007). Efficacy of physical activ-ity in the adjunctive treatment of major depressive disorders: Preliminary results. Clinical Practice and Epidemiology in Mental Health: CP & EMH, 3, 8.
Sigfusdottir, I. D., Asgeirsdottir, B. B., Sigurdsson, J. F., & Gudjonsson, G. H. (2011). Physical activity buffers the effects of family conflict on depressed mood: A study on adolescent girls and boys. Journal of Ado-lescence, 34(5), 895-902.
Silveira, H., Moraes, H., Oliveira, N., Coutinho, E. S., Laks, J., & Deslandes, A. (2013). Physical exercise and clinically depressed patients: A systematic review and meta-analysis. Neuropsychobiology, 67(2), 61-68.
Stavrakakis, N., de Jonge, P., Ormel, J., & Oldehinkel, A. J. (2012). Bidirectional prospective associations between physical activity and depressive symptoms. The TRAILS study. Journal of Adolescent Health, 50(5), 503-508.
Strawbridge, W. J., Deleger, S., Roberts, R. E., & Kaplan, G. A. (2002). Physical activity reduces the risk of subsequent depression for older adults. American Journal of Epidemiology, 156(4), 328-334.
Tao, F. B., Xu, M. L., Kim, S. D., Sun, Y., Su, P. Y., & Huang, K. (2007). Physical activity might not be the pro-tective factor for health risk behaviours and psychopathological symptoms in adolescents. Journal of Paediatrics and Child Health, 43(11), 762-767.
Plasticity Genes do not Modify Associations between Physical Activity and Depressive Symptoms
“Let us understand what our own selfish genes are up to, because we may then at least have a chance to upset their designs, something that no other species has ever aspired to do.”
Richard Dawkins, The Selfish Gene, 1989
Stavrakakis N, Oldehinkel AJ, Nederhof E, Oude Voshaar RC, Verhulst FC, Ormel J & de Jonge P
Health Psychol. 2013; 32(7):785-92
CHAPTER
CHAPTER IV58
ABSTRACT
Purpose: Physical activity is inversely associated with depression in adolescents but the
overall associations are fairly weak, suggesting individual differences in the strength of the
associations. The aim of this study was to investigate whether plasticity genes modify the
reciprocal prospective associations between physical activity and depressive symptoms
found previously.
Methods: In a prospective population-based study (N = 1196), physical activity and depres-
sive symptoms were assessed three times, around the ages of 11, 13.5 and 16. Structural
Equation Modeling was used to examine reciprocal effects of physical activity and depressive
symptoms over time. The plasticity genes examined were 5-HTTLPR, DRD2, DRD4, MAOA,
TPH1, 5-HTR2A, COMT, and BDNF. A cumulative gene plasticity index consisting of three
groups (low, intermediate and high) according to the number of plasticity alleles carried by
the adolescents was created. Using a multi-group approach we examined if the associations
between physical activity and depressive symptoms differed between the three cumulative
plasticity groups as well as between the individual polymorphisms.
Results: We found significant cross-sectional and cross-lagged paths from physical activ-
ity to depressive symptoms and vice versa. Neither the cumulative plasticity index nor the
individual polymorphisms modified the strengths of these associations.
Conclusion: Associations between adolescents’ physical activity and depressive symptoms
are not modified by plasticity genes.
59GENES, PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
INTRODUCTION
Individuals carrying specific genotypes have a higher sensitivity to environmental influences
than others (Belsky et al., 2009). Although research initially focused on the negative conse-
quences of these genotypes in the context of the classic diathesis-stress framework, more
recent work stresses the possibility that individuals who suffer most from negative experi-
ences might also benefit most from environmental support or the absence of adversity (Bel-
sky & Beaver, 2011; Ellis, Boyce, Belsky, Bakermans-Kranenburg, & van Ijzendoorn, 2011).
This ‘For better and for worse’ manner in which individuals are more responsive to environ-
mental influences suggests that individuals who carry specific genotypes might benefit more
from positive environmental influences (such as family or social support) in early life, while
they might also be more vulnerable to negative environmental influences (such as child mal-
treatment and negative life events) in developing psychopathology later on. Genes thought
to be responsible for this differential susceptibility have been labeled plasticity genes.
Belsky and Pluess (2009) proposed several genes to operate as plasticity genes including
the serotonin transporter gene (SLC6A4), the dopamine receptor 2 and 4 genes (DRD2 &
DRD4), the dopamine transporter gene (DAT1), the serotonin 2a receptor gene (5-HTR2A),
the catechol-O-methyl transferase gene (COMT), the monoamine oxidase gene (MAOA), and
the tryptophan hydroxylase 1 gene (TPH1). Although these genes are not the only likely
candidates, they have been shown most consistently to act as plasticity genes in the lit-
erature (for a review see Belsky & Pluess, 2009). Belsky and Pluess proposed a cumulative
genetic plasticity model, in which the susceptibility to particular environmental influences is
hypothesized to increase with increasing numbers of plasticity alleles (Belsky et al., 2009;
Belsky & Beaver, 2011).
In adults, inverse associations between physical activity (PA) and depressive symptoms have
been shown previously (Penedo & Dahn, 2005), supporting the idea that engaging in PA can
at least prevent the development of depressive symptoms. In adolescents, fewer studies
have investigated this relationship, and the results have been inconsistent so far (Johnson &
Taliaferro, 2011). A recent prospective study in older adults demonstrated reciprocal inverse
relationships between PA and depressive symptoms (Lindwall, Larsman, & Hagger, 2011).
They found that there was a stronger relationship between PA predicting later depressive
symptoms than between depressive symptoms predicting later engagement in PA. We also
have reported reciprocal prospective associations between PA and depressive symptoms in
a cohort of Dutch adolescents (Stavrakakis et al., 2012). Associations were measured using
a cross-lagged path analysis model involving three measurements. The associations found
were significant but rather weak, indicating a high amount of unexplained variance. Part of
this unexplained variance may be due to individual differences in both the potential benefits
CHAPTER IV60
of PA in alleviating depressive symptoms and the negative effects of depressive symptoms
on PA levels. A better understanding of these individual differences is needed in order to
design effective interventions for treating affective disorders.
We hypothesized that the plasticity genes might underlie part of the assumed individual
differences in the reciprocal associations between PA and depressive symptoms. In their
review, Belsky and Pluess (2009), proposed a theoretical framework in which they suggest
that these individuals might have lower thresholds for experiencing pleasure and/or displea-
sure, which might result in differential susceptibility to environmental influences. Although
this is speculative, two major neurobiological mediating mechanisms have been identified,
namely the serotonergic (experience of displeasure) and dopaminergic systems (reward sen-
sitivity and sensation seeking). These two systems have been implicated in the aetiology
of depression (Lapin & Oxenkrug, 1969; van Praag & Korf, 1970) but also have been shown
to be influenced by PA (Dishman et al., 2006). Alterations in serotonin and noradrenaline
levels have been suggested to be the main reason for the beneficial effects of PA on stress-
resilience and mood (Dishman et al., 2006). Evidence for an effect of PA on the dopaminergic
system comes mainly from animal research where treadmill running increased the release
(Meeusen, Piacentini, & De Meirleir, 2001) and turnover of dopamine (Hattori, Naoi, & Nishino,
1994) in the striatum of rats. Increasing dopamine levels is known to be an effective way
to improve mood (Lemke, Brecht, Koester, & Reichmann, 2006). Therefore, these plastic-
ity genes might modify the reciprocal association between PA and depressive symptoms
through their differential action on the serotonergic and dopaminergic systems.
To date, only three studies have tried to elucidate whether specific genotypes may be as-
sociated with the effects of PA on depressive symptoms in humans. Rethorst, Landers, Na-
goshi and Ross (2010) showed that after 30 minutes of exercise carriers of a long allele of
the serotonin-transporter-linked polymorphic region (5-HTTLPR) exhibited greater reduc-
tion in depressive symptoms than individuals with two short alleles. In a cross-sectional
study, Rethorst and colleagues (2011) demonstrated that the 5-HTTLPR gene modified the
relationship between PA and depressive symptoms in a sample of 170 students. Finally,
Mata, Thompson and Gotlib (2010) conducted a study in adolescent girls, and found that
PA involvement reduced depressive symptoms in carriers of the Brain Derived Neurotrophic
Factor (BDNF) met allele, but not in carriers of the val/val polymorphism. BDNF is a growth
factor that is encoded by the BDNF gene and is abundant throughout the human body. Bel-
sky et al. did not include the BDNF gene in their list of plasticity genes. However, because
BDNF seems to be related to both PA (Knaepen et al., 2010; Goekint et al., 2010) and depres-
sive symptoms (Hashimoto, 2010; Verhagen et al., 2010), the BDNF gene was considered a
plausible modifier of the reciprocal association between PA and depressive symptoms, and
therefore was included in our study as well.
61GENES, PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
The aim of the present study was to investigate whether plasticity genes modified the
reciprocal prospective associations between PA and depressive symptoms for several rea-
sons. First, it is important to extensively test robustness of gene environment interac-
tions (GxE), as this field seems to be plagued by non-confirmations. For example, the GxE
between the short 5-HTTLPR allele and adversity on depression first reported by Caspi et
al. (2003) has not been confirmed in two meta-analyses (Risch et al., 2009; Munafo et al.,
2009), but has been justified in a later qualitative review (Caspi et al., 2010) and in the most
recent meta-analysis (Karg et al., 2011). Thus, for the advancement in the field of GxE re-
garding PA and depressive symptoms, confirmation studies are of utmost importance. Sec-
ond, previous investigations focused only on the unidirectional effects of PA on depressive
symptoms (Mata et al., 2010; Rethorst et al., 2010; Rethorst et al., 2011). We extend this
line of research by investigating possible GxE in the reciprocal association between PA and
depressive symptoms using previously developed cross-lagged models in a multi-group
analysis. Third, we extend the GxE literature regarding PA and depressive symptoms by
using an index of cumulative plasticity (Belsky & Pluess, 2009; Belsky et al., 2009; Belsky &
Beaver, 2011). Summarizing, in order to test if the associations between PA and depressive
symptoms were moderated by genetic polymorphisms, we performed multi-group analyses
with regard to both cumulative plasticity index and individual polymorphisms on cross-
lagged models of PA and depressive symptoms.
METHODS
Design
Data were used from the TRacking Adolescents’ Individual Lives Survey (TRAILS), a pro-
spective cohort (three assessment waves) of Dutch adolescents. Permission for the study
was obtained from the Dutch Central Committee on Research Involving Human Subjects.
Further information on TRAILS objectives, main design, data collection, sample selection
and non-response can be found elsewhere (de Winter et al., 2005; Huisman et al., 2008).
Parents gave informed consent in the first wave (T1), while for the second (T2) and third
wave (T3) informed consent was obtained from both the parents and the participants
themselves.
Procedure
During T1, parents or guardians were visited by TRAILS interviewers in their homes and were
administered an interview. In addition, they were also asked to complete a questionnaire.
During T2 and T3, parents had to complete a questionnaire, which was sent through the
CHAPTER IV62
mail. The adolescents filled out self-report questionnaires at school. Further information on
TRAILS procedures can be found elsewhere (de Winter et al., 2005; Huisman et al., 2008).
Participants
At T1, 2230 adolescents from a possible 2935 (51% females, mean age = 11.1, SD = 0.55)
agreed to take part in TRAILS. The response rates at T2 and T3 were 96.4% (N = 2149;
51% girls, mean age = 13.65, SD = 0.53) and 81.4% (N = 1816; 52.3% girls, mean age = 16.27,
SD = 0.73) respectively. For the present study only participants with complete genetic data
were included (N = 1196; 52.3% girls). Individuals that did not have both parents born in the
Netherlands were also excluded from the final analysis, since these SNPs can differ between
ethnic groups. This focus sample was still representative of a normal population of adoles-
cents and the problems observed therein.
Measures
Depressive Symptoms
Adolescents filled out the Affective Problems scale of the Youth Self Report (YSR; Achen-
bach, 1991b) for each assessment wave at school. Parents completed the parent version of
the YSR the Child Behavior Checklist (CBCL; Achenbach, 1991a) at home. The mean scores
of the YSR and CBCL scales were used in the analyses. The Affective Problem scale con-
tains 13 items, which correspond to DSM-IV depressive symptoms (van Lang et al., 2005).
These items include sadness, loss of pleasure, crying, self-harm, suicidal ideation, feelings
of worthlessness, guilt, loss of energy, overtiredness, eating problems, and sleeping prob-
lems; and are scored on a three-point scale (0 = never or not at all true, 1 = sometimes true
and 2 = very often or very true). One of the sleep items (‘I sleep more than most kids’) was ex-
cluded because exclusion of this item increased the internal consistency of the scale (Bosch
et al., 2009). Scores on the remaining twelve items were averaged in order to construct a
total depressive symptoms scale with an internal consistency (Cronbach’s alpha) of 0.74 at
T1, of 0.81 at T2 and 0.86 at T3.
Physical Activity
At T1 adolescents were asked to rate how many times per week they performed physi-
cal exercise (for example, swimming, playing football, horse-riding) on a five-point scale
(0 = never, 1 = once per week, 2 = two to three times per week, 3 = four to five times per week,
and 4 = five to six times per week). At T2 and T3, separate questions were asked for PA during
the summer and the winter: “How many days in an average week in the summer/winter do
you take part in physical activities?” and the responses were scored on an eight-point scale
63GENES, PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
(0 = never, to 7 = seven days per week). The answers to these two questions were averaged
and mean scores were used in the analysis. T1 responses (ranging from 0-4) were recoded
into an eight-point scale using the monotonic R7 transform recommended by Little in order
to achieve a similar metric for the PA measures at all measurement waves (Little, in press).
Plasticity Genes
The list of plasticity genes as proposed by Belsky and Pluess (2009) include DRD2, DRD4,
5-HTTLPR, MAOA, TPH1, 5-HTR2A, COMT and the DAT1. In our study DAT1 was excluded be-
cause it was not genotyped, while the BDNF gene was included for reasons explained in the
Introduction. Following the criteria set by Belsky and Pluess (2009), the A1 allele of DRD2,
the long version of DRD4 (from 7 repeats to 10 repeats), the short allele of 5-HTTLPR, the
2R/3R alleles of MAOA, the A allele of TPH1, the T allele of the 5-HTR2A, the val (G) allele of
COMT and the met allele of BDNF were defined as plasticity alleles.
DNA Extraction
Blood samples (n = 1190) or buccal swabs (Cythobrush®) (n = 275) were collected for DNA
extraction using a manual salting out procedure as described by Miller et al. (1988).
Genotyping of BDNF, DRD2, COMT, TPH1 and 5-HTR2A
The BDNF single nucleotide polymorphism (rs6265), DRD2/TaqIA (rs1800497), COMT/val-
158met (rs4680), TPH1 (rs1799913) and 5-HTR2A (rs6313) were genotyped on the Golden
Gate Illumina BeadStation 500 platform (Illumina, San Diego, California) according to the
manufacturer’s protocol. Call rates were: 81% for BDNF, 100% for DRD2, 95% for COMT,
100% for TPH1 and 100% for 5-HTR2A. All DNA samples could be amplified and concordance
between DNA replicates (n = 53) showed a 100% genotyping accuracy (Nederhof et al., 2010).
Data cleaning was completed according to recommended procedures (Steptoe, 2010). All
SNPs were well within Hardy-Weinberg equilibrium, with HWE p-values ranging between
0.42 and 0.52.
Genotyping Length Polymorphisms of 5-HTTLPR, DRD4 and MAOA
Genotyping of the length polymorphisms (LP) DRD4, MAOA, HTTLPR and SNP rs25531
(A/G SNP in L HTTLPR) was done at the Research lab for Multifactorial Diseases within the
Human Genetics department of the Radboud University Nijmegen Medical Centre in Nijme-
gen, the Netherlands. Genotyping of the HTTLPR polymorphism in the promoter region of
SLC6A4 (5-HTT, SERT) gene was performed by simple sequence length analysis. Call rate
was 91.6%. A custom-made TaqMan assay (Applied Biosystems) was utilized to genotype
the single nucleotide substitution (A to G) which is present in the HTTLPR long (l) allele
CHAPTER IV64
(rs25531). Call rate was 96.5%. Concordance between DNA replicates showed an accuracy of
100%. All lg alleles were recoded into s’, because it has been shown that this polymorphism
represents low serotonin expression comparable to the s’ allele, while la was recoded as l’
(Nederhof et al., 2010). The 48 bp direct repeat polymorphism in exon 3 of DRD4 was geno-
typed on the Illumina BeadStation 500 platform (Illumina Inc., San Diego, CA, USA). Three
percent blanks as well as duplicates between plates were taken along as quality controls
during genotyping. Determination of the length of the alleles was performed by direct analy-
sis on an automated capillary sequencer (ABI3730, Applied Biosystems, Nieuwerkerk a/d
IJssel, the Netherlands) using standard conditions. Call rate for DRD4 was 99.4%. The 30bp
variable number of tandem repeat polymorphism (called MAOA-LPR or MAOA-uVNTR) was
also genotyped on the Illumina BeadStation 500 platform (Illumina Inc., San Diego, CA, USA).
Three percent blanks as well as duplicates between plates were taken along as quality con-
trols during genotyping. Call rate was 100% for MAOA. All polymorphisms were well within
Hardy-Weinberg equilibrium (HWE p-values ranged from 0.77 to 0.87).
Cumulative Genetic Plasticity Index
Each polymorphism was assigned one point if a plasticity genotype was present and these
values were summed to create a cumulative index. Girls who were heterozygous for the
MAOA gene, which is located on the X chromosome, were categorized into the low activ-
ity (i.e., high plasticity) genotype as proposed by Belsky and Beaver (2011). Three groups
were created according to the number of plasticity genotypes present: a low group (high
plasticity in 0-3 genes), intermediate group (high plasticity in 4-5 genes) and a high group
(high plasticity in 6-8 genes). Initially the low group was defined as individuals who were
not carrying any plasticity genotypes, but because of insufficient power (n = 22) we merged
this group with the individuals carrying one to three plasticity genotypes. The three groups
were of unequal size, with the intermediate group having the most individuals. We expected
the intermediate group to be the most heterogeneous with regard to each of the individual
genes, therefore we deemed necessary to create relatively equally sized groups for the high
and lowest plasticity groups.
Covariates
Gender and socioeconomic status (SES) were included as covariates. Five standardized vari-
ables (professional occupation and educational attainment for both father and mother, and
household income) were averaged and used to calculate SES. The lowest 25% of scores were
considered as ‘low SES’, the highest 25% as ‘high SES’ and the remaining were grouped as
‘intermediate SES’.
65GENES, PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
Analysis
Imputation of Missing Data
A detailed account of attrition rates within the TRAILS study can be found elsewhere (Huis-
man et al., 2008). Multiple imputations were performed for all variables except for the miss-
ing genotyping data (N = 1196 at all three measurement waves) using STATA version 10
(STATA corp., College station, Texas). Five different datasets (Rubin, 1996) were created and
were subsequently imported into Mplus 5 (Muthén & Muthén, 2007) in order to be used in
the analysis.
Cross-Lagged Path Model
A cross-lagged path model (Structural Equation Modeling) using Mplus was created to in-
vestigate cross-sectional and prospective associations between PA and depressive symp-
toms. Using this model we estimated the direction of the prospective effect (cross-lagged)
between PA and depressive symptoms but also the autoregressive paths (relative stability
over time) and the cross-sectional covariances (interpreted as correlations at T1 and as
correlated change at T2 and T3). Correlated change and cross-lagged paths reflect longi-
tudinal relationships. Due to positive skewness of the depression variables, they were log
transformed prior to analysis. The cross-lagged path model was also adjusted for gender
and SES. A Comparative Fit Index (CFI) and Tucker-Lewis Index (TLI) greater than 0.95 and
a Root Mean Square Error of Approximation (RMSEA) lower than 0.05 were considered a
good model fit.
Cumulative Plasticity Index and Individual Genes Comparisons
A cross-lagged path model was specified for the three groups based on the cumulative
plasticity index. To test whether there was a group difference in the reciprocal associations
between PA and depressive symptoms, we compared a model where the cross-sectional
paths at T1 and the cross-lagged paths were constrained to be equal between groups with
a model where these paths were allowed to differ. To test whether the individual polymor-
phisms affect the reciprocal association between PA and depressive symptoms a similar de-
sign was used for each of the individual genes. A chi-square difference test was performed to
test whether the differences between the models reached statistical significance. To avoid
possible chance findings due to multiple testing a significant chi-square difference score at
the 0.01 level was considered a significant difference between groups.
CHAPTER IV66
RESULTS
Descriptive Statistics
Descriptive statistics of the main variables used in this study are shown in Table 1. Both depres-
sive symptoms and PA remained stable over time. Frequencies for the cumulative plasticity
index and the frequencies for the different genotypes in individual genes are shown in Table 2.
Relationships between Depression and PA
Relationships between the total depressive symptoms score and PA are shown in Figure
1 (overall model fit, χ2 = 6.445, df = 6, p>0.05). Fit indices showed an excellent fit of model
(CFI = 0.999, TLI = 0.999, RMSEA = 0.008), which was significantly better than the fit of the
model without cross-lagged paths from PA to depressive symptoms (χ2difference = 6.825,
df = 1, p<0.05) and the model without cross-lagged paths from depressive symptoms to PA
(χ2difference = 8.333, df = 1, p<0.05). As in our previous paper with more participants, all cross-
lagged paths were significant, indicating that depressive symptoms at any wave predict-
ed PA at the next and vice versa. The autoregressive estimate of depression was higher
than the autoregressive estimate of PA. The cross-sectional association (and correlational
change) between PA and depression was significant at all assessment waves.
Multi-Group Analysis
The multi-group analysis for the three different putative plasticity groups (low, intermediate
and high) is shown in Table 3. Because the middle group was expected to be more heteroge-
neous with regard to each of the individual genes, we first compared the highest group against
the lowest group (χ2unconstrained = 9.456, df = 12, χ2
constrained = 16.897, df = 16) and then compared
all groups with each other (χ2unconstrained = 16.709, df = 18, χ2
constrained = 24.728, df = 26). The fit
indices for both models were excellent (CLI = 0.998, TLI = 0.995, RMSEA = 0.015). As seen in
the table, the χ2difference test was not significant at the 0.01 level.
Table 1. Means (Standard Deviations) of the Main Variables (N = 1196).
T1(mean age 11.1)
T2(mean age 13.5)
T3 (mean age 16.3)
Total Depressive Symptoms
0.21 (0.19) 0.19 (0.21) 0.21 (0.25)
Physical Activity 3.52 (2.09) 3.38 (1.68) 3.51 (1.85)
T1 = first measurement wave, T2 = second measurement wave and T3 = third measurement wave.
67GENES, PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
Table 2. Frequencies of the Plasticity Index and Individual Genes.
Groups (N = 1196)
Low Group(0-3 plasticity alleles)
Intermediate Group (4-5 plasticity alleles)
High Group(6-8 plasticity alleles)
Plasticity Index 28.7% 52.6% 18.7%
Individual Genes Genotype Distribution (%)
BDNF Val/Val62.5%
Val/Met*34.4%
Met/Met*3.1%
DRD2 G/G 63%
A/G*33%
A/A*
4%
COMT A/A30.2%
A/G*
50.1%
G/G* 19.7%
DRD4 s/s63%
s/l*32.3%
l/l*4.7%
5-HTR2A C/C41%
C/T*44.6%
T/T*14.4%
5-HTTLPR l/l 26.2%
l/s*
50%
s/s*
23.8%
TPH1 C/C 36.5%
C/A*
48.7%
A/A*
14.8%
MAOA girls(n = 625)
High19.4%
Intermediate*55.2%
Low*25.4%
MAOA boys(n = 571)
High 37.1%
Low*
62.9%N/A
* Indicates which genotypes were used as the plasticity genotype(s).
CHAPTER IV68
Individual Polymorphisms
We carried out the same multi-group analysis approach for each of the individual polymor-
phisms, comparing the putative plasticity alleles against the non-plasticity alleles. Table 3
depicts the χ2difference test for each individual polymorphism. No significant differences were
found, just a trend (p = 0.03) for MAOA.
DISCUSSION
The results of this study demonstrate reciprocal associations between PA and depressive
symptoms consistent with previous studies (Stavrakakis et al., 2012; Lindwall et al., 2011)
in larger samples. This suggests that early engagement in PA precedes a decrease in later
depressive symptoms and vice versa.
This was the first study using a cumulative plasticity index as proposed by Belsky et al. as
a moderator in the association between PA and depressive symptoms. According to Bel-
sky and Pluess (2009), the cumulative plasticity index assumes an additive effect of these
genes, where the susceptibility to environmental influences is hypothesized to increase with
increasing numbers of plasticity alleles. These genes play a role, through the translation and
transcription of specific proteins, in the formation and degradation of serotonin and dopa-
Figure 1. Associations between Depressive Symptoms and PA over time.
.584* (.020)
.269* (.028)
.558* (.025)
.246* (.030)
-.035*(.015)
-.056*(.020)
-.055*(.026)
DepressionT1
PA T1
DepressionT2
PA T2
DepressionT3
PA T3
-.074*(.029)
-.049*(.024)
-.057*(.020)
-.048*(.020)
PA = physical activity, T1 = first measurement wave, T2 = second measurement wave and T3 = third measurement wave.The path coefficients reflect model estimated standardized beta values. The effects are adjusted for gender and socioeconomic status (SES). * p<0.05.
69GENES, PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
mine. Both of these neurotransmitters are affected by PA and physical inactivity. They also
have been implicated in the etiology of depression, therefore making them primary candi-
dates for modifying the reciprocal association between PA and depressive symptoms. How-
ever, in the current study, we did not find support for any modifying effects of the cumulative
plasticity index on the associations between PA and depressive symptoms.
We then proceeded in investigating the role of individual polymorphisms on the association
of PA and depressive symptoms. To date, there are only three published studies investigat-
ing modifying effects of specific individual polymorphisms on the association between PA
and depressive symptoms (Mata et al., 2010; Rethorst et al., 2010; Rethorst et al., 2011). The
two studies by Rethorst and colleagues investigated the 5-HTTLPR polymorphism, while
the study by Mata and colleagues investigated the BDNF polymorphism. All three studies
found an influence of specific genotypes (the carriers of met allele for BDNF and the carriers
of the short allele for 5-HTTLPR ) on the relationship between PA and depressive symptoms.
None of these findings were confirmed in our study. Both the BDNF and 5-HTTLPR geno-
types as well as the other individual genes that we investigated did not modify the reciprocal
association between PA and depressive symptoms in our sample.
Table 3. Multi-Group Analysis comparing the three different Cumulative Plasticity Groups (low, intermediate and high) and Individual Polymorphisms.
Polymorphism x2diff = x2
con. – x2uncon. dfdiff = dfcon-dfuncon
p-values
Cumulative Plasticity (high vs. low of 3 groups)
7.44 4 0.11
Cumulative Plasticity(3 groups)
8.02 8 0.43
Individual Polymorphisms x2diff = x2
con. – x2uncon. dfdiff = dfcon-dfuncon
p-values
BDNF 3.03 4 0.55
5-HTTLPR 7.12 4 0.13
DRD2 2.50 4 0.64
MAOA 10.75 4 0.03
5-HTR2A 3.64 4 0.46
COMT 5.82 4 0.21
DRD4 4.23 4 0.38
TPH1 8.41 4 0.08
CHAPTER IV70
This lack of modifying effect could be down to a number of reasons. First, perhaps these
plasticity genes are not influencing the relationship between PA and depressive symptoms
at all, contrary to our expectations. It is possible that other genes acting on different neu-
rotransmitter systems, such as the endorphin system (Hegadoren et al., 2009; Fox, 1999) or
the endocannabinoid system (Sparling et al., 2003), might be playing a more important role
in influencing the relationship between PA and depressive symptoms. Second, it is possible
that plasticity genes play a role on the association between PA and depressive symptoms
but the design chosen in this study was not appropriate to reveal these effects. This could
be either because of measurement error (see later section on strengths and limitations for
further information) or because the measurement waves were not close enough to capture
the processes involved in the interrelationships of plasticity, PA and depressive symptoms.
In animal research the ability to remain active (through free wheel running) is thought to be
an enriching environment with numerous benefits on behavior and physiology. In humans, PA
can be considered a supportive or an enriching experience, especially when individuals start
exercising from a very early age and find it a pleasurable experience, but that is still a matter
of debate. In our study, we did not find an effect of early engagement in exercise (at the age of
11) on depressive symptoms and vice versa in individuals carrying more plasticity genotypes
compared to individuals carrying less. However, in order to investigate this plasticity theory
in depth, the effects of even earlier (early childhood) exercise participation on depressive
symptoms should be investigated in individuals carrying these plasticity genotypes. Finally,
the exact functioning of certain polymorphisms is not entirely clear. For example, with re-
spect to the TPH1 gene previous studies have produced mixed results of which allele might
be acting as the plasticity allele. Belsky et al. (2009) assumed that carrying the A allele is
the putative plasticity genotype while other studies have suggested that the A allele might
be playing a protective role against MDD irrespective of positive environmental influences
(Viikki et al., 2010). MAOA is another example of this complexity. The MAOA gene is only
found at the X chromosome, which makes the grouping of females more troublesome than
the grouping for males.
Further research on the exact function of these genes on the neurotransmitter systems and
the effects of exercise on the human biology is needed in order to discover important genetic
and biological markers of the association between PA and depressive symptoms. Epistatic in-
teractions (gene-gene) must be also studied extensively. Although studies on epistasis might
be complicated and difficult to conduct, the results might be informative in relation to indi-
vidual differences behind the reciprocal associations between PA and depressive symptoms.
There are some notable strengths in this study. First, the sample size, which was consider-
ably larger than previous studies mentioned. Second, the longitudinal design (3 measure-
ments) allowed for the detection and investigation of changes between PA and depressive
71GENES, PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
symptoms over a long period of time (6 years). Finally, the model fit was excellent, model-
ing the reciprocal relationship between PA and depressive symptoms in our data with great
accuracy.
However, there are also limitations. First, most of the analysis relied on self-reports. Self-
reports are not as reliable as more objective measurements of PA, such as, calculation of
Metabolic Equivalent of Task (MET) in minutes by accelerometers. In addition, PA was as-
sessed by a single question, which was slightly different at T1 than at the last two waves,
unlike validated PA questionnaires (e.g., IPAQ) that include a series of detailed questions of
PA involvement. At T2 and T3 separate questions for winter and summer were asked and
more response categories were used compared to T1. Moreover, information on the nature
(leisure-time, voluntary or spontaneous), intensity or duration of PA were not collected. This
type of information may prove useful for a better understanding of the influence of specific
genes on the association between PA and depressive symptoms, since different types, in-
tensities and durations of PA have been shown to be differentially associated with depres-
sive symptoms. Furthermore, due to insufficient power, we merged the group without any
plasticity alleles (n = 22) with the group carrying one to three alleles, probably leading to
some information loss. Finally, it would be optimal to investigate changes in neurotrans-
mitter levels as a result of engaging in PA or increases/decreases in depressive symptoms
before investigating genetic modification, in order to advance the knowledge of how these
effects (PA and depressive symptoms) affect human biology.
CONCLUSION
Although gene-environment and studies on epistasis can produce conflicting results and
seem overly complicated, there is still optimism that these studies can shed light on impor-
tant genetic factors, which might explain individual differences in the associations between
PA and depressive symptoms. This elucidation of genetic differences between individuals is
necessary for designing effective treatments for affective disorders. However, at this time
we did not find support for the notion that plasticity genes may moderate the reciprocal as-
sociation between depressive symptoms and PA.
CHAPTER IV72
REFERENCES
Achenbach, T. M. (1991a). Manual for the child behavior checklist/ 4-18 and 1991 profile. University of Ver-mont, Burlington.
Achenbach, T. M. (1991b). Manual for the youth self report and 1991 profile. University of Vermont, Burlington.
Belsky, J., & Beaver, K. M. (2011). Cumulative-genetic plasticity, parenting and adolescent self-regulation. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 52(5), 619-626. doi:10.1111/j.1469-7610.2010.02327.x
Belsky, J., Jonassaint, C., Pluess, M., Stanton, M., Brummett, B., & Williams, R. (2009). Vulnerability genes or plasticity genes? Molecular Psychiatry, 14(8), 746-754. doi:10.1038/mp. 2009.44
Belsky, J., & Pluess, M. (2009). Beyond diathesis stress: Differential susceptibility to environmental influ-ences. Psychological Bulletin, 135(6), 885-908. doi:10.1037/a0017376
Bosch, N. M., Riese, H., Dietrich, A., Ormel, J., Verhulst, F. C., & Oldehinkel, A. J. (2009). Preadolescents’ somatic and cognitive-affective depressive symptoms are differentially related to cardiac auto-nomic function and cortisol: The TRAILS study. Psychosomatic Medicine, 71(9), 944-950. doi:10.1097/PSY.0b013e3181bc756b
Caspi, A., Hariri, A. R., Holmes, A., Uher, R., & Moffitt, T. E. (2010). Genetic sensitivity to the environment: The case of the serotonin transporter gene and its implications for studying complex diseases and traits. The American Journal of Psychiatry, 167(5), 509-527. doi:10.1176/appi.ajp. 2010.09101452
Caspi, A., Sugden, K., Moffitt, T. E., Taylor, A., Craig, I. W., Harrington, H., & Poulton, R. (2003). Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science, 301(5631), 386-389. doi:10.1126/science.1083968
de Winter, A. F., Oldehinkel, A. J., Veenstra, R., Brunnekreef, J. A., Verhulst, F. C., & Ormel, J. (2005). Evalua-tion of non-response bias in mental health determinants and outcomes in a large sample of pre-adoles-cents. European Journal of Epidemiology, 20(2), 173-181.
Dishman, R. K., Berthoud, H. R., Booth, F. W., Cotman, C. W., Edgerton, V. R., Fleshner, M. R., & Zigmond, M. J. (2006). Neurobiology of exercise. Obesity, 14(3), 345-356. doi:10.1038/oby.2006.46
Ellis, B. J., Boyce, W. T., Belsky, J., Bakermans-Kranenburg, M. J., & van Ijzendoorn, M. H. (2011). Differential susceptibility to the environment: An evolutionary–neurodevelopmental theory. Development and Psy-chopathology, 23(1), 7-28. doi:10.1017/S0954579410000611
Fox, K. R. (1999). The influence of physical activity on mental well-being. Public Health Nutrition, 2(3A), 411-418.
Goekint, M., Roelands, B., de Pauw, K., Knaepen, K., Bos, I., & Meeusen, R. (2010). Does a period of detrain-ing cause a decrease in serum brain-derived neurotrophic factor? Neuroscience Letters, 486(3), 146-149. doi:10.1016/j.neulet.2010.09.032
Hashimoto, K. (2010). Brain-derived neurotrophic factor as a biomarker for mood disorders: An historical overview and future directions. Psychiatry and Clinical Neurosciences, 64(4), 341-357. doi:10.1111/j.1440-1819.2010.02113.x
Hattori, S., Naoi, M., & Nishino, H. (1994). Striatal dopamine turnover during treadmill running in the rat: Relation to the speed of running. Brain Research Bulletin, 35(1), 41-49.
Hegadoren, K. M., O’Donnell, T., Lanius, R., Coupland, N. J., & Lacaze-Masmonteil, N. (2009). The role of beta-endorphin in the pathophysiology of major depression. Neuropeptides, 43(5), 341-353. doi:10.1016/j.npep. 2009.06.004
Huisman, M., Oldehinkel, A. J., de Winter, A. F., Minderaa, R. B., de Bildt, A., Huizink, A. C., & Ormel, J. (2008). Cohort profile: The Dutch ‘TRacking Adolescents’ Individual Lives’ Survey’; TRAILS. International Jour-nal of Epidemiology, 37(6), 1227-1235. doi:10.1093/ije/dym273
73GENES, PHYSICAL ACTIVITY AND DEPRESSIVE SYMPTOMS
Johnson, K. E., & Taliaferro, L. A. (2011). Relationships between physical activity and depressive symptoms among middle and older adolescents: A review of the research literature. Journal for Specialists in Pedi-atric Nursing: JSPN, 16(4), 235-251. doi:10.1111/j.1744-6155.2011.00301.x
Karg, K., Burmeister, M., Shedden, K., & Sen, S. (2011). The serotonin transporter promoter variant (5-HT-TLPR), stress, and depression meta-analysis revisited: Evidence of genetic moderation. Archives of Gen-eral Psychiatry, 68(5), 444-454. doi:10.1001/archgenpsychiatry.2010.189
Knaepen, K., Goekint, M., Heyman, E. M., & Meeusen, R. (2010). Neuroplasticity - exercise-induced response of peripheral brain-derived neurotrophic factor: A systematic review of experimental studies in human subjects. Sports Medicine, 40(9), 765-801. doi:10.2165/11534530-000000000-00000
Lapin, I. P., & Oxenkrug, G. F. (1969). Intensification of the central serotoninergic processes as a possible determinant of the thymoleptic effect. Lancet, 1(7586), 132-136.
Lemke, M. R., Brecht, H. M., Koester, J., & Reichmann, H. (2006). Effects of the dopamine agonist pramipex-ole on depression, anhedonia and motor functioning in Parkinson’s disease. Journal of the Neurological Sciences, 248(1-2), 266-270. doi:10.1016/j.jns.2006.05.024
Lindwall, M., Larsman, P., & Hagger, M. S. (2011). The reciprocal relationship between physical activity and depression in older European adults: A prospective cross-lagged panel design using SHARE data. Health Psychology, 30(4), 453-462. doi:10.1037/a0023268
Little, T. D. (in press). Longitudinal structural equation modeling: Individual difference panel models. New York: Guildford press.
Mata, J., Thompson, R. J., & Gotlib, I. H. (2010). BDNF genotype moderates the relation between physical activity and depressive symptoms. Health Psychology, 29(2), 130-133. doi:10.1037/a0017261
Meeusen, R., Piacentini, M. F., & De Meirleir, K. (2001). Brain microdialysis in exercise research. Sports Medi-cine, 31(14), 965-983.
Miller, S. A., Dykes, D. D., & Polesky, H. F. (1988). A simple salting out procedure for extracting DNA from hu-man nucleated cells. Nucleic Acids Research, 16(3), 1215.
Munafo, M. R., Durrant, C., Lewis, G., & Flint, J. (2009). Gene X environment interactions at the serotonin transporter locus. Biological Psychiatry, 65(3), 211-219. doi:10.1016/j.biopsych.2008.06.009
Muthén, L. K., & Muthén, B. O. (2007). Mplus user’s guide: Statistical analysis with latent variables (5th ed.). Los Angeles, CA: Muthén & Muthén.
Nederhof, E., Bouma, E. M., Oldehinkel, A. J., & Ormel, J. (2010). Interaction between childhood adversity, brain-derived neurotrophic factor val/met and serotonin transporter promoter polymorphism on depres-sion: The TRAILS study. Biological Psychiatry, 68(2), 209-212. doi:10.1016/j.biopsych.2010.04.006
Penedo, F. J., & Dahn, J. R. (2005). Exercise and well-being: A review of mental and physical health benefits associated with physical activity. Current Opinion in Psychiatry, 18(2), 189-193.
Rethorst, C. D., Landers, D. M., Nagoshi, C. T., & Ross, J. T. (2010). Efficacy of exercise in reducing depressive symptoms across 5-HTTLPR genotypes. Medicine and Science in Sports and Exercise, 42(11), 2141-2147. doi:10.1249/MSS.0b013e3181de7d51
Rethorst, C. D., Landers, D. M., Nagoshi, C. T., & Ross, J. T. (2011). The association of 5-HTTLPR genotype and depressive symptoms is moderated by physical activity. Journal of Psychiatric Research, 45(2), 185-189. doi:10.1016/j.jpsychires.2010.05.007
Risch, N., Herrell, R., Lehner, T., Liang, K. Y., Eaves, L., Hoh, J., & Merikangas, K. R. (2009). Interaction be-tween the serotonin transporter gene (5-HTTLPR), stressful life events, and risk of depression: A me-ta-analysis. JAMA: The Journal of the American Medical Association, 301(23), 2462-2471. doi:10.1001/jama.2009.878
CHAPTER IV74
Rubin, B. D. (1996). Multiple imputation after 18+ years. Journal of the American Statistical Association, 91(434), 473.
Sparling, P. B., Giuffrida, A., Piomelli, D., Rosskopf, L., & Dietrich, A. (2003). Exercise activates the endocan-nabinoid system. Neuroreport, 14(17), 2209-2211. doi:10.1097/01.wnr.0000097048.56589.47
Stavrakakis, N., de Jonge, P., Ormel, J., & Oldehinkel, A. J. (2012). Bidirectional prospective associations between physical activity and depressive symptoms. Τhe TRAILS study. Journal of Adolescent Health, 50(5), 503-8. doi:10.1016/j.jadohealth.2011.09.004
Steptoe, A., Nolte I. M., McCaffery, J. M., & Snieder, H. (2010). Candidate gene and genome-wide association studies in behavioral medicine. In A. Steptoe (Ed.), Handbook of behavioral medicine: Methods and ap-plications (1st ed., ). New York: Springer.
van Lang, N. D., Ferdinand, R. F., Oldehinkel, A. J., Ormel, J., & Verhulst, F. C. (2005). Concurrent validity of the DSM-IV scales affective problems and anxiety problems of the youth self-report. Behaviour Research and Therapy, 43(11), 1485-1494. doi:10.1016/j.brat.2004.11.005
van Praag, H. M., & Korf, J. (1970). L-tryptophan in depression. Lancet, 2(7673), 612.
Verhagen, M., van der Meij, A., van Deurzen, P. A., Janzing, J. G., Arias-Vasquez, A., Buitelaar, J. K., & Franke, B. (2010). Meta-analysis of the BDNF Val66Met polymorphism in major depressive disorder: Effects of gender and ethnicity. Molecular Psychiatry, 15(3), 260-271. doi:10.1038/mp. 2008.109
Viikki, M., Kampman, O., Illi, A., Setala-Soikkeli, E., Anttila, S., Huuhka, M., et al. (2010). TPH1 218A/C poly-morphism is associated with major depressive disorder and its treatment response. Neuroscience Let-ters, 468(1), 80-84. doi:10.1016/j.neulet.2009.10.069
Competitive Sport Participation, Sport Competence and Depressive Symptoms in Adolescent Boys and Girls
“Serious sport has nothing to do with fair play. It is bound up with hatred, jealousy, boastfulness, disregard of all rules and sadistic pleasure in witnessing violence: in other words it is war minus the shooting.”
George Orwell, in The Sporting Spirit, 1945
“Certification from one source or another seems to be the most important thing to people all over the world. A piece of paper from a school that says you’re smart, a pat on the head from your parents that says you’re good or some reinforcement from your peers that makes you think what you’re doing is worthwhile. People are just waiting around to get certified.”
Frank Zappa, in Oui interview, 1979
Stavrakakis N, Roest AM, Verhulst FC, Veenstra R, Ormel J, de Jonge P & Oldehinkel AJ
Under preparation
CHAPTER
CHAPTER V76
ABSTRACT
Purpose: Physical activity is inversely but moderately related to depressive symptoms in
adolescents. This might indicate that not all types of physical activity are associated with
depressive symptoms, or that the association is present only in a subgroup. We set out to
investigate: a) whether participation in sports with a clear competitive element is associ-
ated with current depressive symptoms and symptom changes over time in adolescents; b)
whether these associations are modified by gender and sport competence.
Methods: Data came from the Dutch TRacking Adolescents’ Individual Lives Survey
(TRAILS), a longitudinal population-based study in adolescents (N = 2149; girls = 51%, mean
age = 13.65, SD = 0.53). Depressive symptoms were measured by the Affective Problems
scale of the Youth Self-Report. Based on participants’ reported participation in sports, we
selected sports with a clear competitive element. Sport competence was assessed by peer
nominations.
Results: Participation in competitive sports was significantly associated with lower cur-
rent depressive symptom scores (p = 0.002) but not with positive changes in depressive
symptoms (p = 0.23). After adjustment for sport competence the association between sport
participation and current depressive symptoms was no longer significant (p = 0.41). No three-
way interactions between competitive sports, gender, and sport competence were observed.
Conclusion: Participation in competitive sports was related to lower current depressive
symptoms, which could be explained by sport competence, but not with changes in depres-
sive symptoms. Finally, we found no evidence that associations between sport participation
and depressive symptoms are modified by gender and sport competence.
77COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
INTRODUCTION
Physical activity (PA) has received significant attention over the last years as being poten-
tially beneficial in alleviating depressive symptoms (Vilhjalmsson & Thorlindsson, 1992;
Sanders et al., 2000; Salmon, 2001; Gore et al., 2001; Dunn et al., 2005; Rimer et al., 2012). Re-
views and meta-analyses of intervention studies investigating the relationship between PA
and depressive symptoms in adults have also shown that PA can reduce depressive symp-
toms, but their conclusions should be taken with caution because of the low methodological
quality of most studies included (Rimer et al., 2012; Cooney et al., 2013; Rosenbaum et al.,
2014). Two meta-analyses of intervention studies in adolescents also showed that PA was
inversely related to depressive symptoms but the effect sizes were small, and again there
was a lack of high methodological quality studies (Larun et al., 2006; Brown et al., 2013).
Because depressive symptoms in adolescence have been shown to predict the development
of a depressive disorder in adulthood (Hankin et al., 1998; Lewinsohn et al., 1999; Fergusson
et al., 2005; Monshouwer et al., 2012), it is pertinent to investigate the relationship between
PA and depressive symptoms at an early age. According to recent reviews, the majority of
observational studies in adolescents, regardless of design, show an inverse association be-
tween PA and depressive symptoms, with weak to moderate effect sizes (Biddle & Asare,
2011; Johnson & Taliaferro, 2011). Two possible reasons that might explain the relatively
weak associations observed in the relationship between PA and depressive symptoms were
investigated in the current study. First, the definition of PA is broad, which could mean that
not all types of PA are associated with depressive symptoms, resulting in an overall small-
sized association. Second, PA might be inversely associated with depressive symptoms only
in specific subgroups of individuals and not in others, also resulting in an overall modest
association. For example, a recent study showed that PA was inversely associated with de-
pressive symptoms in males but not in females (Edman et al., 2014).
PA is defined as any bodily muscle movement that leads to energy expenditure, and includes
a broad range of activities regardless of the environment (e.g., leisure-time vs. occupational)
or intensity of the activity (see: Caspersen et al., 1985). Sport participation is a more specific
definition relying on a competitive element of the activity. Previous research has shown an
inverse association between sport participation and depressive symptoms (Vilhjalmsson &
Thorlindsson, 1992; Sanders et al., 2000; Gore et al., 2001), with effect sizes similar to the
association between PA and depressive symptoms (see for a review: Johnson & Taliaferro,
2011). However, these studies did not define sports according to the competitive element
of the activity, but used an unspecified definition of sports, or categorized sports according
to their social aspect (club, individual, or team) (Vilhjalmsson & Thorlindsson, 1992; Sand-
ers et al., 2000; Gore et al., 2001). Because participation in sports is suggested to have an
CHAPTER V78
evolutionary origin (Lombardo, 2012), we hypothesized that in particular the competitive
element of sports might play a crucial role in the association with depressive symptoms.
According to Lombardo (2012), one of the best ways to achieve social status is through
sports and especially those with a highly competitive element, because sport competitions
in the modern world closely resemble hunting activities and primitive warfare. Competing
in sports gives an opportunity for males to rank each other according to their abilities and
therefore be able to select capable allies or avoid capable enemies. This ranking is assumed
to result in improved or diminished chances of being selected by females (Lombardo, 2012).
This can be seen in both ancient (e.g., Olympians in Classical Greece, see: Poliakoff, 1987
and Golden, 2008) and modern athletes, who tend to gain status fast, especially if they are
champions (Loy, 1972; Chase & Dummer, 1992; Holland & Andre, 1994; Buss, 2007; Chase &
Machida, 2011). This suggests that individuals who are athletically competent will be more
likely to achieve high status than their less competent peers.
The social status that the individual attains or maintains has also been hypothesized to
influence the development of mental disorders. Price’s Social Competition Hypothesis of De-
pression (Price et al., 1994) suggests that depressed mood is a result of inhibitory mecha-
nisms that force the individual to concede defeat and downplay their ambitions and desires
in situations where it is impossible or unlikely for them to succeed. This involuntary defense
mechanism, is hypothesized to help individuals to psychologically deal with ‘what would
otherwise be unacceptably low social status’ (Price et al., 1994; in abstract). Status can be
operationalized in many ways (Linton, 1936; Price, 1972; Price et al., 1994; Anderson et al.,
2001). In this study, we will focus on status in terms of achievements (Price, 1972), which
mainly depends on the ranking of individuals according to their competency levels as per-
ceived by their peers. We focus on peer-perceived sport competence because of the impor-
tance given to sport competence and sport participation in adolescence, especially in boys
(Gill, 1992; Kavussanu & Harnisch, 2000; Sallis et al., 2000).
There are reasons to assume that the above-described mechanisms apply to males in par-
ticular. Boys and men attach more importance to sport participation and are conspicuously
more competitive than girls and women (Price, 1988; Balafoutas et al., 2012; Deaner et al.,
2012), as their achievement status will directly influence their chances of being selected by
women (Fieder & Huber, 2012; Lombardo, 2012; Li et al., 2013). Furthermore, several reports
suggest that achievement status is more strongly associated with well-being in men than in
women (Brendgen et al., 2005; Tiffin et al., 2005; Oldehinkel et al., 2007). Taken together, this
indicates that the way competition for status influences social hierarchy formation and sub-
sequently the development of mental health problems may be different for boys and girls.
The aims of this study were to investigate: 1) whether sport participation is inversely asso-
79COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
ciated with current depressive symptoms and depressive symptom change in late adoles-
cence, and 2) whether these associations are modified by peer-perceived sport competence
and gender. We hypothesized that peer-perceived sport competence modifies the relation-
ship between sport participation and depressive symptoms, and that sport-competent boys
will benefit more from performing competitive sports than girls and less competent boys.
METHODS
Design
Data used for this study were part of a large population cohort of adolescents in the Neth-
erlands called TRacking Adolescents’ Individual Lives Survey (TRAILS). The aim of TRAILS
was to delineate the development of mental health problems from childhood to adulthood.
Approval of the study was obtained from the Dutch Committee on Research Involving Hu-
man Subjects. Data for the first measurement wave (T1) were collected in 2001, and for the
second (T2) and third (T3) measurement waves data collection took place between 2003
and 2007, with approximate intervals of 2.5 years between each wave. Informed consent was
collected from the parents at T1, while both parents and participants gave informed consent
at T2 and T3. Extensive descriptions of TRAILS design, data and sample collection can be
found elsewhere (de Winter et al., 2005; Huisman et al., 2008; Ormel et al., 2012). The pres-
ent study was based on data from T2 and T3.
Participants and Procedure
The total sample at T1 included 2230 adolescents (50.8% girls, mean age = 11.1, SD = 0.6).
The response rates at T2 and T3 were 96.4% (N = 2149; 51% girls, mean age = 13.7, SD = 0.5)
and 81.4% (N = 1816; 52.3% girls, mean age = 16.3, SD = 0.7) respectively (Ormel et al., 2012).
Detailed information on sampling procedures can be found elsewhere (de Winter et al., 2005;
Nederhof et al., 2012). During T2 and T3, parents received a questionnaire by mail. Adoles-
cents and their teachers were asked to complete questionnaires at school.
Outcome Variables
Depressive Symptoms
Depressive symptoms were assessed using the Affective Problem Scale of the Youth Self
Report (YSR; Achenbach & Rescorla, 2001). The Affective Problem Scale consists of 13 items,
which correspond to DSM-IV depressive symptoms (van Lang et al., 2005). These items are
scored on a three-point scale (0 = never or not at all true, 1 = sometimes true and 2 = very of-
ten or very true) and include questions on sadness, loss of pleasure, crying, self-harm, suicidal
CHAPTER V80
ideation, feelings of guilt and worthlessness, loss of energy, eating problems, overtiredness, and
sleeping problems (3 items). The internal consistency (Cronbach’s alpha) for the 13 items was
0.77 for T2 and 0.78 for T3. A mean score of all 13 items for T2 and T3 was used in the analyses.
Predictors
Sport Participation
According to the Oxford dictionary a sport is: “an activity involving physical exertion and skill
in which an individual or team competes against another or others for entertainment” (Ox-
ford dictionaries; http://oxforddictionaries.com/definition/english/sport?q=sport). Through-
out this manuscript we will use this definition of sports, because we were interested in the
competitive element of activities. During T2, adolescents were asked to indicate whether
they participate in any sports and what types of sports they participate in (maximum of four
sports). From these responses, we created a binary variable indicating whether at least one
of the activities reported, corresponded to the definition of sports we used. If none of the
activities met the criterion for a sport the participant was coded as a non-sporter. Activities
such as ‘running’, ‘walking’ or ‘dancing’ were not considered as sports per se, because these
activities do not necessarily imply direct competition with other individuals or teams. In con-
trast, activities such as “running in competitions” were considered a sport. For responses
that had an ambiguous competitive element (such as gymnastics and roller-blading), the
frequencies per week (more or less often than two times per week) were used to distinguish
between sporters and non-sporters. In total, 2089 adolescents had valid responses on the
sport participation measure. The 141 (6.3%) individuals that did not respond to the sport
participation and the general PA questions were considered missing. More information on
the sport participation measure can be found in the Appendix, Table F.
Peer Nominations of Sport Competence
At T2, peer nominations were assessed with regard to various characteristics, including
sport competence. These nominations were assessed in classes with at least three regular
TRAILS participants. Peer nominations were assessed in 172 classes of 34 schools in the
first year (72 school classes) and second year (100 school classes) of secondary education,
each including 7 to 30 participating adolescents per class (mean = 18.4, SD = 5.99). In total,
3312 students (1675 boys; 1637 girls) including 1007 regular TRAILS participants filled out
the peer nomination questionnaire (mean age = 13.7, SD = 0.66). For the purpose of this study,
we selected the question on sport competence (‘Who is good at sports?’), and used the per-
centage of nominations as a continuous measure in the analyses. Further details on the peer
nomination measures in TRAILS and the representativeness of the peer nominations sample
have been described elsewhere (Dijkstra et al., 2008).
81COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
Teacher Evaluations of Sport Competence
At T2, the teachers were asked to rate the athletic competence of each TRAILS participant
in their class. This question was rated on a five-point scale, with 1 = inadequate, 2 = more
or less adequate, 3 = adequate, 4 = good, and 5 = outstanding. Information on the teacher’s
evaluation of sport competence was collected for 1455 adolescents.
Socioeconomic Status
Socioeconomic status (SES) of the family was obtained by averaging five standardized vari-
ables measured at T1, i.e., professional occupation and educational attainment of both par-
ents/guardians, and household income. The internal consistency of the SES scale was 0.84.
The scale captures 61.2% of the variance in the five items. SES was used as a continuous
measure in the analysis. At T1, 16.8% of the TRAILS participants lived in families with a
disposable income annual up to € 13620, which is below the at-risk-of-poverty-threshold.
Further information on each of these variables and the aggregated SES can be found else-
where (Amone-P’Olak et al., 2010).
Analyses
Missing Data
Differences between responders and non-responders at T3, and between the adolescents
that completed the peer evaluations and those that did not, for all predictor variables were
estimated using independent t-tests and chi-square difference tests depending on whether
the variables in question were continuous or dichotomous. Non-responders at T3 were more
likely to be males (x2 = 15.7, df = 1, p = 0.001), participate in sports (x2 = 9.5, df = 1, p = 0.002),
report more depressive symptoms at T2 (t = -2.7, p = 0.01) and have lower SES (t = -10.1,
p = 0.001) than responders. There was no significant difference between responders and
non-responders at T3 regarding the peer evaluations of sport competence (t = -0.6, p = 0.53).
Adolescents who had completed the peer nominations were more likely to have higher SES
(t = -6.3, p = 0.001) and participate in sports (x2 = 8.6, df = 1, p = 0.004) than those who did not,
while they did not differ regarding gender (x2 = 0.7, df = 1, p = 0.42), depressive symptoms at T2
(t = 0.8, p = 0.40) and depressive symptoms at T3 (t = 0.5, p = 0.64).
Correlations
All analyses were performed using SPSS (version 20, SPSS Inc., Chicago, Illinois). The as-
sociations between the predictors, outcome variables and covariates were calculated using
Pearson’s correlations (r). For the sport participation and gender variables, the correlations
can be interpreted as point-biserial correlations because of the dichotomous nature of these
variables (Field, 2009).
CHAPTER V82
Linear Regressions
The hypotheses were tested by two-tailed tests with an alpha-level of 0.05. Because the
residuals of the outcome variable (i.e., depressive symptoms) were positively skewed, we
performed linear regressions with bootstrapping, that is, estimations of confidence intervals
based on repeated sample extractions from the original dataset, which do not rely on a priori
distributional assumptions (for more information see Field, 2009 p. 163; Hayes, 2009). Boot-
strapping of 1000 samples with Bias Corrected and Accelerated confidence intervals (CIs)
was performed for all analyses.
In order to test associations with current depressive symptoms, sport participation was
used as the predictor, and gender and SES as covariates. In order to predict depressive
symptom change, baseline (T2) depressive symptoms were included as an additional co-
variate. In total, 2054 participants were included in the cross-sectional analysis, while 1614
adolescents were included in the depressive symptom change analysis. We repeated these
analyses including sport competence. Because the peer perceptions of sport competence
were collected from a subsample of TRAILS, a total of 960 adolescents were included in the
cross-sectional analysis, while 785 adolescents were included in the depressive symptom
change analysis. In order to test interactions between sport participation, sport competence,
and gender, we created a model including all possible two-way and three-way interactions
among these variables, and removed any non-significant interactions from the model, pro-
viding the model was still hierarchical.
Finally, in order to examine to what extent the results were affected by the measures used,
the analyses were repeated using the teacher evaluations of sport competence and a broad-
er definition of sports. In this broader definition, all reported physical activities (including, for
example, walking in the park, dancing, or running) were categorized as a sport.
RESULTS
Descriptive Statistics
From a total of 2089 participants (51% girls) with valid responses in the sport participation mea-
sure, 1177 (56.3%) reported at least one competitive sport, while 912 (43.7%) did not take part in
any activity that fit our definition. Within the sample that participated in the peer nominations,
134 (13.3%) adolescents received no nominations for being good at sports from their peers, 391
(38.8%) received between 1 and 25% nominations, 245 (24.3%) between 26 and 50%, and 237
(23.6%) between 51 and 100%. Table 1 shows the frequencies of all predictors according to sport
participation. Depressive symptoms at T2 (mean age = 13.7) had a mean score of 0.27 (SD = 0.26)
while depressive symptoms at T3 (mean age = 16.3) had a mean score of 0.30 (SD = 0.27).
83COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
Correlations
As can be seen from table 2, for boys sport participation was significantly correlated with
depressive symptoms at both T2 and T3, whereas for girls sport participation was only sig-
nificantly correlated with depressive symptoms at T2. Sport competence was significantly
associated with T2 and T3 depressive symptoms and with sport participation for both boys
and girls (see also table A in appendix).
Linear Regressions
In the cross-sectional analysis, sport participation was significantly associated with depres-
sive symptoms at T2 after adjustment for gender and SES (B = -0.05, CIs = -0.07 to -0.02,
p = 0.002), but its effect was no longer significant after inclusion of sport competence (table
3). This model explained around 4% of the variance in the depressive symptoms. Neither
sport participation nor sport competence was significantly associated with change in de-
pressive symptoms between T2 and T3 (table 4). This model explained approximately 30%
of the variance, due to the inclusion of T2 depressive symptoms as a covariate.
The three-way interaction between sport participation, sport competence, and gender did
not predict current depressive symptoms significantly (SES-adjusted B = 0.06, CIs = -0.03 to
0.16, p-value = 0.17), and was therefore removed from the model. In the remaining model there
were no significant two-way interactions (B ranges = -0.04 to -0.00, p-value ranges = 0.31 to
0.95), hence a model with only main effects described the data best. The same was true for
the prediction of change in depressive symptoms between T2 and T3 (B ranges = -0.04 to
0.01, p-value ranges = 0.30 to 0.98).
Repeating the analyses with the teacher evaluations of sport competence (tables B and C
in the appendix) yielded similar results, with only one small difference: the cross-sectional
association between sport participation and depressive symptoms remained marginally sig-
nificant (B = -0.03, CIs = -0.06 to 0.001, p = 0.054) after adjustment for the teacher evaluations
of sport competence. When using a broader, non-competitive, definition of sport participa-
tion, sport participation remained a significant predictor of current depressive symptoms
after adjustment for (peer-perceived) sport competence as well. Apart from this, the results
were very similar to the results obtained when sports was defined as being competitive (see
appendix, tables D and E).
CHAPTER V84
Table 1. Gender, SES, and Peer-Perceived Sport Competence by Sport Participation.
Variable Sport Participation
Yes Percentage % No Percentage %
Gender (N = 2089)
BoysGirls
745 432
63.3 36.7
273 639
29.9 70.1
Indices of SES (N = 2058)
Maternal Education
Elementary 52 4.5 67 7.6
Lower-Track Secondary 332 28.9 293 33.3
Higher-Track Secondary 413 35.9 315 35.8
Senior Vocational 252 21.9 157 17.8
University 100 8.8 49 5.5
Paternal Education
Elementary 49 4.8 46 6.2
Lower-Track Secondary 259 25.1 214 28.8
Higher-Track Secondary 326 31.7 243 32.7
Senior Vocational 232 22.5 151 20.3
University 164 15.9 89 12
Maternal Occupation
Low Level 144 16.7 122 18.6
Intermediate Level 462 53.6 349 53.1
High Level 314 29.7 122 28.3
Paternal Occupation
Low Level 266 27.1 240 33.9
Intermediate Level 307 31.3 213 30
High Level 409 41.6 256 36.1
85COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
Variable Sport Participation
Yes Percentage % No Percentage %
Family Income
Low (less than €2,500 per month)
139 12.9 159 19.6
Intermediate (€2,500-5,500 per month)
597 55.6 482 59.4
High (more than € 5,500 per month)
338 31.5 171 21
Aggregate SES (N = 2058)a
Low SES 249 21.5 251 27.9
Intermediate SES 560 48.3 466 51.8
High SES 350 30.2 182 20.2
Percentage of Peer Nominations of Sport Competence (N = 1007)a
No Nominations (0%) 46 7.9 85 21.6
1 to 25% of Nominations 166 28.5 210 53.4
26 to 50% of Nominations 170 29.2 68 17.3
51 to 75% of Nominations 130 22.3 24 6.1
76 to 100% of Nominations 71 12.1 6 1.5
SES = socioeconomic status. a For SES and sports competence, categories were created for descriptive purposes; in the analyses continuous measures were used.
Table 1. continued.
CHAPTER V86
Table 2. Pearson Correlations according to Gender.
Variables 1. Sport Participation
2. Sport Competence
3. Depressive Symptoms T2
4. Depressive Symptoms T3
5. SES
1. Sport Participationa
— 0.36** -0.07* -0.05 0.15**
2. Sport Competence
0.40** — -0.12** -0.10** 0.12**
3. Depressive Symptoms T2
-0.12** -0.15** — 0.51** -0.05
4. Depressive Symptoms T3
-0.12** -0.14** 0.53** — -0.10**
5. SES 0.13** 0.11* -0.02 -0.02 —
The top half shows correlations for girls while the bottom half shows correlations for boys.SES = socioeconomic status, T2 = second measurement wave and T3 = third measurement wave. Two-tailed significance **p<0.01, *p<0.05. a Point-biserial correlation was conducted for sport participation because of the binary nature of this variable.
Table 3. Prediction of Depressive Symptoms at T2 (N = 960).
Variable Adjusted R2 B (SE) p-value CI (95%)
Sport Participation 0.04 -0.02 (0.02) 0.41 -0.06 to 0.02
Gender -0.05 (0.02) 0.01 -0.09 to -0.01
Sport Competence -0.04 (0.01) 0.001 -0.05 to -0.02
SES -0.01 (0.01) 0.41 -0.02 to 0.01
SES = socioeconomic status, CI = Bias Corrected and Accelerated confidence intervals, SE = standard error, T2 = second measurement wave. All continuous/interval predictors were z-transformed prior to analysis.
87COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
DISCUSSION
Three findings emerged from the current study. First, participation in sports was associated
with current depressive symptoms but not depressive symptom change. Second, when peer-
perceived sport competence was included in the model, the association between participa-
tion in competitive sports and current depressive symptoms was no longer statistically sig-
nificant. Finally, interactions between sport participation, perceived sport competence, and
gender were associated with neither current depressive symptoms nor depressive symptom
change in adolescence.
Participation in both competitive and more broadly defined sports was associated with
low current depressive symptom levels. This suggests that the association with depres-
sive symptoms cannot be ascribed to the competitive element of some physical activities.
The effect sizes were relatively weak and comparable with the ones observed in a previ-
ous study investigating the relationship between general PA and depressive symptoms in
the same sample (Stavrakakis et al., 2012). Therefore, our hypothesis that focusing on the
competitive element of sports might strengthen the association with depressive symp-
toms was not supported.
The reason why we evaluated this question was that it has been suggested that sports
have an evolutionary origin (Lombardo, 2012), and that the competitive element of activities
plays a role in subsequent hierarchy formations. The positive correlations observed in the
Table 4. Prediction of Depressive Symptoms at T3, adjusting for Depressive Symptoms at T2 (N = 785).
Variable Adjusted R2 B (SE) p-value CI (95%)
Sport Participation 0.30 0.02 (0.02) 0.28 -0.02 to 0.06
Gender -0.08 (0.02) 0.001 -0.11 to -0.05
Sport Competence -0.01 (0.01) 0.24 -0.03 to 0.01
SES -0.02 (0.01) 0.05 -0.03 to -0.00
Depressive Symptoms T2 0.14 (0.01) 0.001 0.12 to 0.16
SES = socioeconomic status, CI = Bias Corrected and Accelerated confidence intervals, SE = standard error, T2 = second measurement wave and T3 = third measurement wave. All continuous/interval predictors were z-transformed prior to analysis.
CHAPTER V88
current study between sport participation and peer-perceived sport competence allude to
this, i.e., peers regarded the adolescents participating in competitive sports as being more
competent than non-sporters. The perceived competence can be considered an indication of
social status, which has been linked to depressive symptoms (Price et al., 1994; Gilbert & Al-
lan, 1998; Oldehinkel et al., 2007). Consistent with these findings, peer-perceived sport com-
petence was cross-sectionally associated with depressive symptoms in the current study,
and explained most of the variance observed in this association. This suggests that being
perceived as not being very competent by peers might increase the likelihood of depressive
symptoms regardless of participation in competitive sports. Interestingly, while the asso-
ciation between participation in competitive sports and depressive symptoms became non-
significant after including peer-perceived sport competence in the model, the effect of more
broadly defined PA remained significant. Hence, the ranking achieved according to peer-per-
ceived sport competence may be a more important mediator of the effect of participation in
competitive sports than of the effect of PA in general, with regard to depressive symptoms.
Not only the measurement of sport participation, but also that of perceived sport compe-
tence affected the results: when sport competence was assessed by teacher reports, it ex-
plained less of the effect of participation in competitive sports on depressive symptoms
than when assessed by peer nominations. These diverging results for peer nominations and
teacher evaluations could be due to sample size differences: peer nominations were col-
lected in only part of the sample. Another explanation is that these two measures cover
different aspects of perceived sport competence, as suggested by the only moderate cor-
relations between the two. Perhaps teachers’ evaluations are a more accurate and objective
measure of adolescents’ athletic competence than perceptions of peers. On the other hand,
adolescents spend a lot of time with each other, so may have a better notion of their peers’
actual competence than teachers. Because we were particularly interested in the hierarchy
formation among adolescents, we consider peer nominations of sport competence as a bet-
ter measure than teacher evaluations within the context of this study. Additionally, we did
not know whether the peer perceptions of sport competence were based on school sports
only or also on observations made outside the school setting. Considering the discrepancy
between the peer perceptions and the teacher evaluations of sport competence observed
in this study, it is possible that the perceptions of the peers and the teachers were partly
based on different situations. It would be interesting for future studies to explore the influ-
ence of their teammates perceptions of sport competence on depressive symptoms, since
they might differ from the perceptions of their schoolmates.
Regardless of its definition, sport participation did not predict change in depressive symp-
toms; apparently its benefits, if any, were not strong enough to affect mood in the longer
term. The lack of prospective associations between sport participation and depressive
89COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
symptoms might indicate a reciprocal association between sport participation and depres-
sive symptoms. Indeed, previous studies have shown not only that physically active indi-
viduals will be less likely to suffer from depressive symptoms, but also that individuals
who suffer from depressive symptoms will be less physically active (Jerstad et al., 2010;
Lindwall et al., 2011; Stavrakakis et al., 2012). Therefore, any long-term effects of sport
participation on depressive symptoms might be counteracted by the fact that adolescents
with many depressive symptoms are less likely to participate in sports (reverse causality/
inhibition hypothesis).
The Social Competition Hypothesis (Price et al., 1994) provides an interesting theoretical
framework to predict depressive problems, and we tried to extend this theory to investi-
gate whether hierarchy formations through ritual agonistic behaviors could be a possible
explanation of the relationship between sport participation and depressive symptoms. We
expected that adolescents who participated in competitive sports would report fewer de-
pressive symptoms than non-sporters, and that the beneficial effects of sport participa-
tion would be especially strong for boys who were perceived as highly sport competent.
However, the lack of interactions between sport participation, gender and perceived sport
competence on depressive symptoms did not support this expectation. Perhaps other fac-
tors, such as the results of competitions (winning or losing), self-esteem and self-worth,
or biological mechanisms (e.g., neurotrophin levels) are more important moderators of the
relationship between sport participation and depressive symptoms than the external per-
ceptions of an individuals’ sport competence (Dishman et al., 2006; Buss, 2007; Laske et
al., 2010). A further point that needs to be acknowledged is that our assumption that peer-
perceived sport competence is an indication of social status is an oversimplification of a
complex construct. Social status is multidimensional and can be operationalized in diverse
ways (see: Linton, 1936; Gilbert et al., 1995; Anderson et al., 2001), and modern social hier-
archies do not depend as strictly on ritual agonistic fighting as suggested in Price’s Social
Competition Hypothesis. Social hierarchies can also be formed by how much social atten-
tion is given to an individual by others (for example see: Baumeister & Leary, 1995; Leary
& Baumeister, 2000), or through capacities that do not rely on physical abilities, such as
intelligence (Nettle, 2003). Having a high status in non-sport domains might counteract the
effect of a low peer-perceived sport competence status on depressive symptoms in adoles-
cents and especially in boys, regardless of their sport participation levels. Some support for
this counteracting of other social hierarchy dimensions has been shown in previous studies
(Seroczynski et al., 1997; Oldehinkel et al., 2007).
CHAPTER V90
Limitations
Our study has some limitations. First, some of the sports reported were uninformative with
regard to the competitive element. Activities such as ballroom dancing, horse riding, or gym-
nastics can be competitive in specific circumstances, but are not necessarily so. Because we
did not have information on the actual competitions that adolescents took part in, we used
the frequency of these hard-to-categorize activities, assuming that competitive engagement
requires a frequency of more than twice a week. This limitation could have led to an over-
estimation or underestimation of the proportion of competitive sporters. However, because
using a broader definition of sports in the analysis hardly altered the results, this potential
source of bias is probably not very influential. Second, peer evaluations of sport competence
were not assessed at T3. Therefore, we could not investigate whether adolescents who regu-
larly participated in sports improved their actual athletic competence and subsequently the
perceptions of their peers and how these changes might have affected depressive symptom
changes over time. Moreover, participants with missing data at T3 were more likely to be
males, to participate in sports and to have more depressive symptoms at T2. Missing data
were not accounted for in this study and it is possible that non-response might be directly
caused by the adolescents’ previous depressive symptoms. Therefore, our findings might not
be accurately generalizable to the population. Furthermore, we based our hypotheses on a
combination of two distinct evolutionary hypotheses, namely the evolution of sports and an
evolutionary perspective of the development of depressive symptoms. The measurements
used in TRAILS were not selected from an evolutionary perspective per se, and perhaps were
not optimal to test the full extent of these theories. Therefore the negative findings observed
in the current study might be explained by the less than optimal measures. Finally, we did
not take into account the frequency, duration or intensity of sport participation. It is possible
that only individuals who engage moderately or frequently in sports or PA experience fewer
depressive symptoms (Sanders et al., 2000; Dunn et al., 2005; Goldfield et al., 2011). Although
we cannot exclude this possibility, it has been previously shown that the frequency, duration
and intensity of PA did not predict the onset of a first major depressive episode in adolescents
(Stavrakakis et al., 2013), so a dose-response influence of PA or sport participation on depres-
sive symptom levels seems unlikely.
CONCLUSION
Sport participation was inversely related to current depressive symptoms. This association
might be explained by peer-perceived sport competence. However, we found no evidence
that the interactions between peer-perceived sport competence, sport participation and
gender predicted current depressive symptoms or depressive symptom change over time.
91COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
REFERENCES
Achenbach, T. M., & Rescorla, L. A. (2001). Manual for the ASEBA school-age forms and profiles. Burlington, VT: University of Vermont, Research center for children, youth and families.
Amone-P’Olak, K., Ormel, J., Oldehinkel, A. J., Reijneveld, S. A., Verhulst, F. C., & Burger, H. (2010). Socioeco-nomic position predicts specialty mental health service use independent of clinical severity: The TRAILS study. Journal of the American Academy of Child and Adolescent Psychiatry, 49(7), 647-655.
Anderson, C., John, O. P., Keltner, D., & Kring, A. M. (2001). Who attains social status? Effects of personality and physical attractiveness in social groups. Journal of Personality and Social Psychology, 81(1), 116-132.
Balafoutas, L., Kerschbamer, R., & Sutter, M. (2012). Distributional preferences and competitive behavior. Journal of Economic Behavior & Organization, 83(1), 125-135.
Baumeister, R. F., & Leary, M. R. (1995). The need to belong - desire for interpersonal attachments as a fundamental human-motivation. Psychological Bulletin, 117(3), 497-529.
Biddle, S. J. H., & Asare, M. (2011). Physical activity and mental health in children and adolescents: A review of reviews. British Journal of Sports Medicine, 45(11), 886-895.
Brendgen, M., Wanner, B., Morin, A. J. S., & Vitaro, F. (2005). Relations with parents and with peers, tempera-ment, and trajectories of depressed mood during early adolescence. Journal of Abnormal Child Psychol-ogy, 33(5), 579-594.
Brown, H. E., Pearson, N., Braithwaite, R. E., Brown, W. J., & Biddle, S. J. H. (2013). Physical activity inter-ventions and depression in children and adolescents: A systematic review and meta-analysis. Sports Medicine, 43(3), 195-206.
Buss, D. M. (2007). Evolutionary psychology: The new science of the mind (3rd ed.). New York: Allyn and Bacon.
Caspersen, C. J., Powell, K. E., & Christenson, G. M. (1985). Physical-activity, exercise, and physical-fitness: Definitions and distinctions for health-related research. Public Health Reports, 100(2), 126-131.
Chase, M. A., & Dummer, G. M. (1992). The role of sports as a social-status determinant for children. Re-search Quarterly for Exercise and Sport, 63(4), 418-424.
Chase, M. A., & Machida, M. (2011). The role of sport as a social status determinant for children: Thirty years later. Research Quarterly for Exercise and Sport, 82(4), 731-739.
Cooney, G. M., Dwan, K., Greig, C. A., Lawlor, D. A., Rimer, J., Waugh, F. R., et al. (2013). Exercise for depres-sion. Cochrane Database of Systematic Reviews, (9), CD004366.
de Winter, A. F., Oldehinkel, A. J., Veenstra, R., Brunnekreef, J. A., Verhulst, F. C., & Ormel, J. (2005). Evalua-tion of non-response bias in mental health determinants and outcomes in a large sample of pre-adoles-cents. European Journal of Epidemiology, 20(2), 173-181.
Deaner, R. O., Geary, D. C., Puts, D. A., Ham, S. A., Kruger, J., Fles, E., et al. (2012). A sex difference in the predisposition for physical competition: Males play sports much more than females even in the contem-porary US. Plos One, 7(11), e49168.
Dijkstra, J. K., Lindenberg, S., & Veenstra, R. (2008). Beyond the class norm: Bullying behavior of popular adolescents and its relation to peer acceptance and rejection. Journal of Abnormal Child Psychology, 36(8), 1289-1299.
Dishman, R. K., Hales, D. P., Pfeiffer, K. A., Felton, G., Saunders, R., Ward, D. S., et al. (2006). Physical self-concept and self-esteem mediate cross-sectional relations of physical activity and sport participation with depression symptoms among adolescent girls. Health Psychology, 25(3), 396-407.
Dunn, A. L., Trivedi, M. H., Kampert, J. B., Clark, C. G., & Chambliss, H. O. (2005). Exercise treatment for de-pression: Efficacy and dose response. American Journal of Preventive Medicine, 28(1), 1-8.
CHAPTER V92
Edman, J. L., Lynch, W. C., & Yates, A. (2014). The impact of exercise performance dissatisfaction and physi-cal exercise on symptoms of depression among college students: A gender comparison. Journal of Psy-chology, 148(1), 23-35.
Fergusson, D. M., Horwood, L. J., Ridder, E. M., & Beautrais, A. L. (2005). Subthreshold depression in adoles-cence and mental health outcomes in adulthood. Archives of General Psychiatry, 62(1), 66-72.
Fieder, M., & Huber, S. (2012). An evolutionary account of status, power, and career in modern societies. Human Nature, 23(2), 191-207.
Field, A. (2009). Discovering statistics using SPSS (3rd ed.). SAGE publications Ltd.
Gilbert, P., & Allan, S. (1998). The role of defeat and entrapment (arrested flight) in depression: An explora-tion of an evolutionary view. Psychological Medicine, 28(3), 585-598.
Gilbert, P., Price, J. S., & Allan, S. (1995). Social-comparison, social attractiveness and evolution - how might they be related. New Ideas in Psychology, 13(2), 149-165.
Gill, D. L. (1992). Gender and sport behaviour. Advances in sport psychology (pp. 143-160). Champaign, IL: Human Kinetics.
Golden, M. (2008). Greek sport and social status. Austin TX: University of Texas Press.
Goldfield, G. S., Henderson, K., Buchholz, A., Obeid, N., Nguyen, H., & Flament, M. F. (2011). Physical activity and psychological adjustment in adolescents. Journal of Physical Activity & Health, 8(2), 157-163.
Gore, S., Farrell, F., & Gordon, J. (2001). Sports involvement as protection against depressed mood. Journal of Research on Adolescence, 11(1), 119-130.
Hankin, B. L., Abramson, L. Y., Moffitt, T. E., Silva, P. A., McGee, R., & Angell, K. E. (1998). Development of de-pression from preadolescence to young adulthood: Emerging gender differences in a 10-year longitudinal study. Journal of Abnormal Psychology, 107(1), 128-140.
Hayes, A. F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Com-munication Monographs, 76(4), 408-420.
Holland, A., & Andre, T. (1994). Athletic participation and the social-status of adolescent males and females. Youth & Society, 25(3), 388-407.
Huisman, M., Oldehinkel, A. J., de Winter, A. F., Minderaa, R. B., de Bildt, A., Huizink, A. C., et al. (2008). Cohort profile: The Dutch ‘TRacking Adolescents’ Individual Lives’ Survey’; TRAILS. International Journal of Epidemiology, 37(6), 1227-1235.
Jerstad, S. J., Boutelle, K. N., Ness, K. K., & Stice, E. (2010). Prospective reciprocal relations between physi-cal activity and depression in female adolescents. Journal of Consulting and Clinical Psychology, 78(2), 268-272.
Johnson, K. E., & Taliaferro, L. A. (2011). Relationships between physical activity and depressive symptoms among middle and older adolescents: A review of the research literature. Journal for Specialists in Pedi-atric Nursing: JSPN, 16(4), 235-251.
Kavussanu, M., & Harnisch, D. L. (2000). Self-esteem in children: Do goal orientations matter? The British Journal of Educational Psychology, 70(2), 229-242.
Larun, L., Nordheim, L. V., Ekeland, E., Hagen, K. B., & Heian, F. (2006). Exercise in prevention and treatment of anxiety and depression among children and young people. The Cochrane Database of Systematic Reviews, (3), CD004691-CD004691.
Laske, C., Banschbach, S., Stransky, E., Bosch, S., Straten, G., Machann, J., et al. (2010). Exercise-induced normalization of decreased BDNF serum concentration in elderly women with remitted major depression. The International Journal of Neuropsychopharmacology, 13(5), 595-602.
Leary, M. R., & Baumeister, R. F. (2000). The nature and function of self-esteem: Sociometer theory (pp. 1-62). San Diego, CA, US: Academic Press.
93COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
Lewinsohn, P. M., Rohde, P., Klein, D. N., & Seeley, J. R. (1999). Natural course of adolescent major depressive disorder: I. Continuity into young adulthood. Journal of the American Academy of Child and Adolescent Psychiatry, 38(1), 56-63.
Li, N. P., Yong, J. C., Tov, W., Sng, O., Fletcher, G. J., Valentine, K. A., et al. (2013). Mate preferences do predict attraction and choices in the early stages of mate selection. Journal of Personality and Social Psychol-ogy, 105(5), 757-776.
Lindwall, M., Larsman, P., & Hagger, M. S. (2011). The reciprocal relationship between physical activity and depression in older European adults: A prospective cross-lagged panel design using SHARE data. Health Psychology, 30(4), 453-462.
Linton, R. (1936). The study of man: An introduction. Appleton Century Crofts, Inc.
Lombardo, M. P. (2012). On the evolution of sport. Evolutionary Psychology, 10(1), 1-28.
Loy, J. W. (1972). Social origins and occupational mobility patterns of a selected sample of American ath-letes. International Review for the Sociology of Sport, (7), 5-25.
Monshouwer, K., Smit, F., Ruiter, M., Ormel, J., Verhulst, F. C., Vollebergh, W., et al. (2012). Identifying target groups for the prevention of depression in early adolescence: The TRAILS study. Journal of Affective Disorders, 138(3), 287-294.
Nederhof, E., Jorg, F., Raven, D., Veenstra, R., Verhulst, F. C., Ormel, J., et al. (2012). Benefits of extensive recruitment effort persist during follow-ups and are consistent across age group and survey method. The TRAILS study. BMC Medical Research Methodology, 12(1), 93.
Nettle D. (2003). Intelligence and class mobility in the British population. British Journal of Psychology, 94(4), 551-61.
Oldehinkel, A. J., Rosmalen, J. G. M., Veenstra, R., Dijkstra, J. K., & Ormel, J. (2007). Being admired or being liked: Classroom social status and depressive problems in early adolescent girls and boys. Journal of Abnormal Child Psychology, 35(3), 417-427.
Ormel, J., Oldehinkel, A. J., Sijtsema, J., van Oort, F., Raven, D., Veenstra, R., et al. (2012). The TRacking Adolescents’ Individual Lives Survey (TRAILS): Design, current status, and selected findings. Journal of the American Academy of Child and Adolescent Psychiatry, 51(10), 1020-1036.
Poliakoff, M. B. (1987). Combat sports in the ancient world: Competition, violence and culture. New Haven, CT: Yale University Press.
Price, J. S. (1972). Genetic and phylogenetic aspects of mood variations. International Journal of Mental Health, (1), 124-144.
Price, J. S. (1988). Alternative channels for negotiating asymmetry in social relationships. In M. R. A. Chance (Ed.), Social fabrics of the mind (pp. 157-195). Erlbaum.
Price, J. S., Sloman, L., Gardner, R., Gilbert, P., & Rohde, P. (1994). The social competition hypothesis of de-pression. British Journal of Psychiatry, 164, 309-315.
Rimer, J., Dwan, K., Lawlor, D. A., Greig, C. A., McMurdo, M., Morley, W., et al. (2012). Exercise for depression. Cochrane Database of Systematic Reviews, (7), CD004366.
Rosenbaum, S., Tiedemann, A., Sherrington, C., Curtis, J., & Ward, P. (2014). Physical activity interventions for people with mental illness: A systematic review and meta-analysis. Journal of Clinical Psychiatry, 31, 1555-2101.
Sallis, J. F., Prochaska, J. J., & Taylor, W. C. (2000). A review of correlates of physical activity of children and adolescents. Medicine and Science in Sports and Exercise, 32(5), 963-975.
Salmon, P. (2001). Effects of physical exercise on anxiety, depression, and sensitivity to stress: A unifying theory. Clinical Psychology Review, 21(1), 33-61.
Sanders, C. E., Field, T. M., Diego, M., & Kaplan, M. (2000). Moderate involvement in sports is related to lower depression levels among adolescents. Adolescence, 35(140), 793-797.
CHAPTER V94
Seroczynski, A. D., Cole, D. A., & Maxwell, S. E. (1997). Cumulative and compensatory effects of competence and incompetence on depressive symptoms in children. Journal of Abnormal Psychology, 106(4), 586-597.
Stavrakakis, N., de Jonge, P., Ormel, J., & Oldehinkel, A. J. (2012). Bidirectional prospective associations be-tween physical activity and depressive symptoms. The TRAILS study. The Journal of Adolescent Health, 50(5), 503-508.
Stavrakakis, N., Roest, A. M., Verhulst, F. C., Ormel, J., de Jonge, P., & Oldehinkel, A. J. (2013). Physical activ-ity and onset of depression in adolescents: A prospective study in the general population cohort TRAILS. Journal of Psychiatric Research, 47(10), 1304-1308.
Tiffin, P. A., Pearce, M. S., & Parker, L. (2005). Social mobility over the lifecourse and self reported mental health at age 50: Prospective cohort study. Journal of Epidemiology and Community Health, 59(10), 870-872.
van Lang, N. D., Ferdinand, R. F., Oldehinkel, A. J., Ormel, J., & Verhulst, F. C. (2005). Concurrent validity of the DSM-IV scales affective problems and anxiety problems of the youth self-report. Behaviour Research and Therapy, 43(11), 1485-1494.
Vilhjalmsson, R., & Thorlindsson, T. (1992). The integrative and physiological-effects of sport participation – a study of adolescents. Sociological Quarterly, 33(4), 637-647.
95COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
SUPPLEMENTARY MATERIAL
Table A. Correlations in the Total Sample.
Variable 1. 2. 3. 4. 5. 6. 7. 8.
1. Competitive Sport —
2. Broader Definition of Sports
0.54** —
3. Peer Nomination of Sport Competence
0.42** 0.26** —
4. Teacher Evaluations of Sport Competence
0.22** 0.22** 0.27** —
5. Gender 0.33** 0.03 0.31** 0.01 —
6. Depresive Symptoms T2
-0.14** -0.10** -0.17** -0.10** -0.19** —
7. Depressive Symptoms T3
-0.14** -0.08** -0.18** -0.14** -0.24** 0.54** —
8. SES 0.12** 0.18** 0.09** 0.11** -0.03 -0.04 -0.06* —
SES = socioeconomic status, T2 = second measurement wave and T3 = third measurement wave. Shown here are Pearson correlations between the variables. Point-biserial correlations were conducted for sport participation (both competitive and broad definitions) and gender because of the binary nature of these variables. Because the teacher’s evaluations of sport competence are in an ordinal scale, Spearman Rho correlations were calculated. The correlations between the teacher evaluations of sport competence and dichotomous variables (both sport participation and gender) were converted to rank-biserial correlations using the following formula: rb = (2/n)*(Y1-Y0) where Y1 and Y0 are the mean scores for the group with 1 and 0 respectively (Kraemer, 1982).
REFERENCES
Kraemer, H. C. (1982). Biserial correlation. Encyclopaedia of Statistical Sciences, 1, 276-279.
CHAPTER V96
Table B. Prediction of Depressive Symptoms at T2 by Competitive Sport Participation, Teacher Evaluations of Sport Competence and Gender (N = 1396).
Step Variable Adjusted R2 B (SE) p-value CI (95%)
1 Sport Participation 0.04 -0.03 (0.02) 0.05 -0.06 to 0.001
Gender -0.07 (0.02) 0.001 -0.10 to -0.04
Sport Competence -0.03 (0.01) 0.002 -0.04 to -0.01
SES -0.004 (0.01) 0.63 -0.02 to 0.01
2 Sport Participation* Gender* Sport Competence
0.04 -0.01 (0.03) 0.81 -0.06 to 0.05
SES = socioeconomic status, CI = Bias Corrected and Accelerated confidence intervals, SE = standard error, T2 = second measurement wave.All continuous/interval predictors were z-transformed prior to analysis. Step 1: shows the main effects of the predictors without interactions (adjusted for SES) with current depressive symptoms. Step 2: shows the three-way interaction of sport participation, teacher evaluations of sport competence and gender (adjusted for SES) with current depressive symptoms. In this model the three-way interaction, all two-way interactions (not shown here) and main effects of the predictors (not shown here) were included.
97COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
Table C. Prediction of Depressive Symptoms at T3 by Competitive Sport Participation, Teacher Evaluations of Sport Competence and Gender, adjusting for Depressive Symptoms at T2 (N = 1134).
Step Variable Adjusted R2 B (SE) p-value CI (95%)
1 Sport Participation 0.33 -0.01 (0.02) 0.55 -0.04 to 0.02
Gender -0.09 (0.02) 0.001 -0.12 to -0.06
Sport Competence -0.02 (0.01) 0.003 -0.04 to -0.01
SES -0.02 (0.01) 0.02 -0.03 to -0.001
Depressive Symptoms T2 0.14 (0.01) 0.001 0.12 to 0.15
2 Sport Participation* Gender* Sport Competence
0.33 0.02 (0.04) 0.65 -0.05 to 0.08
SES = socioeconomic status, CI = Bias Corrected and Accelerated confidence intervals, SE = standard error, T2 = second measurement wave and T3 = third measurement wave.All continuous/interval predictors were z-transformed prior to analysis. Step 1: shows the main effects of the predictors without interactions (adjusted for SES and baseline depressive symptoms) with depressive symptom change. Step 2: shows the three-way interaction of competitive sport participation, teacher evaluations of sport competence and gender (adjusted for SES and baseline depressive symptoms) with depressive symptom change. The three-way interaction, all two-way interactions (not shown here) and main effects of the predictors (not shown here) were included in this model.
CHAPTER V98
Table D. Prediction of Depressive Symptoms at T2 by Sport Participation according to a Broader Definition, Peer Nominations of Sport Competence and Gender (N = 960).
Step Variable Adjusted R2 B (SE) p-value CI (95%)
1 Sport Participation 0.05 -0.06 (0.03) 0.01 -0.11 to -0.02
Sport Competence -0.03 (0.01) 0.001 -0.05 to -0.02
Gender -0.06 (0.02) 0.003 -0.09 to -0.01
SES -0.001 (0.01) 0.68 -0.02 to 0.01
2 Sport Participation* Gender* Sport Competence
0.05 0.11 (0.07) 0.13 -0.02 to 0.24
SES = socioeconomic status, CI = Bias Corrected and Accelerated confidence intervals, SE = standard error, T2 = second measurement wave.All continuous/interval predictors were z-transformed prior to analysis. Step 1: shows the main effects of sport participation, sport competence and gender (adjusted for SES) with current depressive symptoms. Step 2: shows the three-way interaction of sport participation, peer nominations of sport competence and gender (unadjusted) with current depressive symptoms. In this model the three-way interaction, all two-way interactions (not shown here) and main effects of the predictors (not shown here) were included.
99COMPETITIVE SPORT PARTICIPATION AND DEPRESSIVE SYMPTOMS
Table E. Prediction of Depressive Symptoms at T3 by Sport Participation according to a Broader Definition, Peer Nominations of Sport Competence and Gender, adjusting for Depressive Symptoms at T2 (N = 785).
Step Variable Adjusted R2 B (SE) p-value CI (95%)
1 Sport Participation 0.30 -0.02 (0.03) 0.42 -0.07 to 0.04
Sport Competence -0.01 (0.01) 0.52 -0.02 to 0.01
Gender -0.08 (0.02) 0.001 -.011 to -0.05
SES -0.01 (0.01) 0.12 -0.03 to 0.01
Depressive Symptoms T2 0.14 (0.01) 0.001 0.12 to 0.16
2 Sport Participation* Gender* Sport Competence
0.30 0.01 (0.06) 0.89 -0.11 to 0.13
SES = socioeconomic status, CI = Bias Corrected and Accelerated confidence intervals, SE = standard error, T2 = second measurement wave and T3 = third measurement wave.All continuous/interval predictors were z-transformed prior to analysis. Step 1: shows the main effects of sport participation (adjusted for baseline depressive symptoms and SES) with depressive symptom change. Step 2: shows the three-way interaction of sport participation, peer nominations of sport competence and gender (adjusted for baseline depressive symptoms at T2 and SES) with depressive symptom change. The three-way interaction, all two-way interactions (not shown here) and main effects of the predictors (not shown here) were included in this model.
Table F. All Sports (sport1, sport2, sport3, sport4) classified according to decision on Competitiveness, Non-Competitiveness and Uncertain Competitiveness.
Decision # of Sport Activities
Frequencies of Participants for each of Four Sports
Sport1 Sport2 Sport3 Sport4
Sum Competitive 48 1008 332 117 74
Sum Non-Competitive 17 647 541 280 135
Sum Ambiguous 65 59 27 15 9
Total 130 1714 900 412 218
Relative-Age Effects in Adolescents: The TRAILS Study
“I can honestly say that insecurity was something formerly unknown to me. I was always the best of my grade. The tallest, the fastest – I thought I was Superman. It turned out this was mainly because I was born in January, thus older than my peers”.
Gert Verhulst, in De Volkskrant interview, 2013
Jeronimus BF, Stavrakakis N, Veenstra R & Oldehinkel AJ
Submitted
CHAPTER
CHAPTER VI102
ABSTRACT
Purpose: This study examined intra-classroom age position (or relative-age) effects on ado-
lescents’ school progress and performance (as rated by teachers), physical development,
temperament, and depressive symptoms, all adjusted for age at the time of measurement.
Methods: Data were derived from three waves of Tracking Adolescents’ Individuals Lives
Survey (TRAILS) of 2230 Dutch adolescents (51% girls, baseline mean age = 11.1, SD = 0.6).
Results: Relative-age predicted school progress (grade retention ORs = 0.83, skipped grade
OR = 1.47, both p<0.001), but substantial effects in adolescents with a normative school tra-
jectory were absent, in contrast to most literature.
Conclusion: Adolescents who had repeated a grade showed inverse relative-age effects on
physical development and school performance, as well as on depressive symptoms, favoring
the relative young.
103RELATIVE-AGE EFFECTS IN ADOLESCENTS
INTRODUCTION
Relative-Age
In most countries, children at school are assorted in same-age groups based on the month
and year of birth (Gledhill, Ford, & Goodman, 2002; Verachtert, De Fraine, Onghena, & Ghes-
quiere, 2010). Consequently, within a single classroom children may differ in age by up to 11
months. Relatively older children have a slightly more developed physique and mind than
their younger classmates (Martin, Foels, Clanton, & Moon, 2004; Verachtert et al., 2010).
These physical and psychological advantages may become catalyzed into different de-
velopmental trajectories through favorable peer-contrast effects (Bedard & Dhuey, 2006;
Gladwell, 2008). We define developmental differences due to the age position driven peer
contrast effects as relative-age effects.
Taller and more mature children tend to have more prestige (Harris, 2009; Henrich & Gil-
White, 2001), which in turn affects friendship formation (Dijkstra, Cillessen, & Borch, 2013;
Newcomb, Bukowski, & Pattee, 1993), and enhances learning opportunities (Carruthers,
Laurence, & Stich, 2006). Relative-age has been associated with higher intelligence (Lawlor,
Clark, Ronalds, & Leon, 2006), school success (Cobley, McKenna, Baker, & Wattie, 2009),
identity formation (Allen, 2008), peer-perceived competence and leadership (Dhuey & Lip-
scomb, 2008), success in sports (Musch & Grondin, 2001), and positive self-perception and
self-esteem (Fenzel, 1992; Thompson, Barnsley, & Battle, 2004). Children’s internal working
models are based on self- and other representations (“looking glass self”), which emerge
and crystallize relatively early in development (Fraley, 2002; Harris, 1995): five-year olds
have already stable and clearly established classroom hierarchies (Boyce et al., 2012).
Self-Fulfilling Prophecy
Children’s relative-age position can become a self-fulfilling prophecy (“learning by being”), cata-
lyzed by reciprocal feedback loops between the developing phenotypes and their environments
(Allen, 2008; Pierson, Addona, & Yates, 2014), analogous to the corresponsive principle (Caspi &
Shiner, 2006) and Dick-Flynn model (Beam & Turkheimer, 2013; Flynn, 2009). Part of this effect
may be driven by external adult evaluations, based on unfavorable intra-cohort contrast effects
(Verachtert et al., 2010). Relatively old children are granted special opportunities for success,
such as extra coaching in sports (Gladwell, 2008; Musch & Grondin, 2001; Persico, Postlewaite,
& Silverman, 2004), whereas relatively young children meet lower expectations and have an
increased risk to repeat a grade (Cobley et al., 2009; Doornbos, 1971; Martin et al., 2004).
Though classroom hierarchies are established in childhood, their consequences may be
particularly salient in adolescence. Adolescents become increasingly able to influence their
CHAPTER VI104
environment (while parental socialization wanes), and select a rapidly expanding peer net-
work and a first romantic partner (Collins, Welsh, & Furman, 2009; Cyranowski, Frank,
Young, & Shear, 2000). Furthermore, earlier work associated being relatively young with
victimization (Muhlenweg, 2010), psychiatric problems (Goodman, Gledhill, & Ford, 2003;
Morrow et al., 2012), and suicide before age 20 (Thompson, Barnsley, & Dyck, 1999). Be-
cause unfavorable relative-age contrast effects modulate children’s self-perception and
self-esteem (Thompson et al., 2004), which are known risk factors for affective disor-
ders (Abela, Fishman, Cohen, & Young, 2012), being relatively young may increase risk
of depression, which has a high incidence in adolescence (Hankin et al., 1998; Oldehinkel,
Wittchen, & Schuster, 1999). Because low self-perception and self-esteem can have a per-
sistent impact on affect, the effect of relative-age might also be discernible in tempera-
mental negative affect.
Finally, relatively older children tend to play more sports – partly driven by selection effects
(Gladwell, 2008; Pierson et al., 2014), which may explain observed relative-age effects on
multiple indices of physical growth in adolescence (Sandercock et al., 2013). Timing of pu-
berty onset seems also influenced by environmental factors (up to 12% variance), including
experiences unshared by twins (Eaves et al., 2004) and neighborhood characteristics (Dear-
dorff et al., 2012). Hence, small relative-age effects on body mass and rate of maturation
seem speculative, but not impossible.
Ability Streaming
Though relative-age effects have been observed in multiple countries (Bedard & Dhuey,
2006), most literature is based on samples from the United States of America (USA) or United
Kingdom (UK) (Musch & Grondin, 2001; Verachtert et al., 2010). One of the challenges in iso-
lating relative-age effects is that their manifestation is contingent on mechanisms to group
children in classes, which differ over time and place (Dalton, 2012; Cascio and Schanzenbach,
2007). For example, the Dutch cohort under study was allocated over classes based upon
an annual birthdate cutoff (pre/post October), and all children in each grade attended all
courses together. In other systems children attend courses (e.g., language, math, or sports)
subdivided on base of ability, called ‘setting’ or ‘banding’, which may alleviate relative-age
effects (Pierson et al., 2014).
The influence of apparently innocuous institutional differences becomes manifest in interna-
tional comparisons, e.g., in the 2006 Program for International Student Assessment (PISA)
up to half of the Dutch (or Belgian, Austrian, or Czech) 15-year olds were in a different grade
than expected in a normative trajectory, compared to 12% in the USA, 1% in the UK, and
none in Japan or Finland (Dalton, 2012). Relative-age effects may thus manifest themselves
via grade progression, that is, the possibility to allocate children to a higher or lower grade
105RELATIVE-AGE EFFECTS IN ADOLESCENTS
than the normative one, based on their abilities (Doornbos, 1971; Dalton, 2012). We expect
that especially the (less qualified) relative young are retained, whereas the (highly qualified)
relative old are accelerated, which we call ability streaming.
The Present Study
A demonstration of persistent relative-age effects in adolescence, bestowed upon children
by an adult-imposed structuring of their worlds, could lead to renewed awareness and pre-
vention strategies of teachers, parents, and psychiatrists. We therefore aimed to quantify
associations between relative-age position and multiple outcome domains in early and
middle adolescence. Relative-age effects were defined as the effects of intra-cohort age
position adjusted for the actual age at the time of measurement. Only after adjustment for
the additional developmental time granted to the relatively old (a methodological artifact)
the alleged accumulating benefits and disadvantages driven by adolescents’ relative-age
position can be observed.
We hypothesized unfavorable outcomes, in multiple domains, for the relatively young com-
pared to the relatively old adolescents. More specifically, we expected to replicate relative-
age effects on school progress (H1), and tested whether relative-age predicted whether
adolescents had repeated or skipped a grade. For both the adolescents with a normative
progress and the group who repeated a grade we tested whether relative-age predicted
weight (H2a), pubertal status (H2b), school performance (H3), sport competence (H4), and
peer status in terms of peer rejection (H5a) and popularity (H5b). Innovatively, we tested
whether relative-age effects predicted depressive symptoms (H6a) and temperamental fear
and frustration (negative affect, H7a); as well as change in depressive symptoms (H6b) and
fear and frustration (H7b) between age 11 and 16. Because earlier work suggested associa-
tions between maternal socioeconomic status (SES) and month of birth (Bobak & Gjonca,
2001; Buckles & Hungerman, 2013; Martin, 2013), we ran all analyses without and with ad-
justment for family SES.
METHODS
Design and Participants
Data were collected as part of the TRacking Adolescents' Individual Lives Survey (TRAILS),
a large ongoing prospective cohort of Dutch adolescents followed to study the psychologi-
cal, social and physical development of children towards adulthood (Ormel et al., 2012). The
core aim was to unravel developmental pathways to psychological (ill) health. The study
was approved by the Dutch Central Committee on Research Involving Human Subjects. The
CHAPTER VI106
first measurement wave (T1) started in 2001, whereas data collection for the remaining
waves (T2 to T3) took place at intervals of approximately 2.5 years. Informed consent was
collected from the parents at T1, whereas for T2 and T3 informed consent was obtained from
both parents and adolescents. The TRAILS design, sample selection, and data collection are
described extensively elsewhere (de Winter et al., 2005; Huisman et al., 2008; Ormel et al.,
2012). Briefly, participants were selected from five municipalities in the North of the Nether-
lands. From a total of 2935 children, 2230 agreed to take part at T1 (response rate 76%; 51%
girls, mean age 11.1, SD = 0.6). The response rates were 96% at T2 (N = 2149; 51% girls, mean
age 13.7, SD = 0.5) and 81% at T3 (N = 1816; 52% girls, mean age 16.3, SD = 0.7). Non-response
associated slightly with low socioeconomic background, male gender, low IQ and school per-
formance, non-western ethnicity, and externalizing problems, but not with other emotional
and behavioral problems (de Winter et al., 2005). At T1, the parents or guardians were inter-
viewed at their homes, and handed in a previously sent questionnaire at that occasion. At T2
and T3, the questionnaire was sent to the parents or guardians by mail. At all three waves,
the adolescents and their teachers completed the questionnaires at school.
Measures
Relative-Age
As outlined, when the TRAILS children entered school, children born between the first of Oc-
tober and the 30th of September of the next year were allocated in the same age group in the
Netherlands. Consequently, in a normative situation, children born in September were the
youngest in a given grade (month 1), while children born in October were the oldest (month
12). This relative-age measure (1-12) was used as a continuous measure in our analysis.
School Progress
Information on school progress was collected from the schools before sample selection.
Four categories were distinguished: (1) normal progression, (2) children who repeated a
grade, (3) children who skipped a grade, and (4) children in special education. The groups of
adolescents who had repeated one or two grades were merged because only nine adoles-
cents had repeated a grade twice.
Physical Development
At T2, the length in centimeters (cm) and the weight in kilograms (kg, without shoes and
heavy clothing) were measured. Body Μass Ιndex (BMI) was calculated by dividing the
weight (kg) by the square of the height (m2).
107RELATIVE-AGE EFFECTS IN ADOLESCENTS
Pubertal Status
At T2, pubertal status was measured with the Pubertal Development Scale (PDS), previously
shown to be reliable, with a Chronbach’s alpha of 0.77 for boys and 0.81 for girls (Shirtcliff,
Dahl, & Pollak, 2009). The PDS assesses development on five (Tanner) characteristics, includ-
ing growth spurt in height, skin changes, body hair in both boys and girls, breast development
and menarche in girls, and voice change and facial hair growth in boys (Janssens et al., 2011).
Each item was rated on a four-point scale (0 = not yet started, 1 = just started, 2 = going on for
a while, 3 = passed that). In our analyses we used the mean of the four item scores.
School Performance
At T2, teachers provided school ratings on history, geography, math, and natural sciences.
Adolescent’s school performance was operationalized as the composite of these marks
rated on a five point scale, ranging from 1 = inadequate to 5 = outstanding. Cronbach’s alpha
was 0.86.
Sport Competence
At T2, teachers were asked to rate the sport competence of each adolescent on a five-point
scale (1 = inadequate, 2 = hardly adequate, 3 = adequate, 4 = good, and 5 = outstanding).
Peer Status (being popular or rejected)
At T2, social status was assessed using a sociometric nomination procedure in classrooms
with at least three TRAILS respondents (Oldehinkel, Rosmalen, Veenstra, Dijkstra, & Ormel,
2007). Adolescents could nominate an unlimited number of classmates on a total of 18 ques-
tions, covering a wide range of issues and behaviors. For the purpose of this study, we used
the proportions of the peer nominations for rejection (“being disliked”) and popularity (“being
someone others want to be associated with”), and contrasted these nominations against
the adolescents without this specific peer nomination (i.e., rejected/popular status adoles-
cents vs. all adolescents without this peer nomination).
Depressive Symptoms
Depressive symptoms were assessed at T1 and T3 with the Affective Problem scales of
the Youth Self Report (YSR; Achenbach, 1991b) and parent-reported Child Behavior Checklist
(CBCL; Achenbach, 1991a), which cover depressive symptoms according to DSM-IV criteria
with 13 items, including information on sadness, loss of pleasure, crying, self-harm, suicidal
ideation, feelings of worthlessness, guilt, loss of energy, overtiredness, eating problems and
sleeping problems (Achenbach, Dumenci, & Rescorla, 2003; Ferdinand, 2008). Both scales are
rated on a three-point scale ranging from 0 = never or not at all true to 2 = very often or very
CHAPTER VI108
true. Cronbach’s alpha for the YSR was 0.72 at T1 and 0.78 at T3 and for the CBCL 0.68 at T1
and 0.76 at T3. In the analyses we used the combined mean scores of the CBCL and YSR scales.
Temperamental Fear and Frustration
Temperamental negative affectivity was assessed at T1 and T3 with the Dutch parent ver-
sion (Hartman, 2000) of the revised Early Adolescent Temperament Questionnaire (EATQ-R;
Putnam, Ellis, & Rothbart, 2001), which is based on the temperamental model by Rothbart et
al. (2011; 2000). Earlier work in TRAILS showed the EATQ factor structure of the parent ver-
sion to be superior to the child version (Oldehinkel, Hartman, de Winter, Veenstra, & Ormel,
2004). Fear and frustration were measured with five questions rated on a five-point scale
(ranging from 1 = almost never to 5 = almost always true). Cronbach’s alpha was 0.63 (T1) and
0.66 (T3) for Fear and 0.74 (T1) and 0.75 (T3) for Frustration.
Socioeconomic Status
SES of the family of origin was the composite of five standardized variables measured at
T1, i.e., professional occupation and educational attainment of both parents/guardians, and
household income.
Plan of Analyses
Data cleaning, calculation of descriptives, and all analyses were performed in SPSS (version
20, SPSS Inc., Chicago, Illinois). To test whether relatively young or relatively old adolescents
were more likely to repeat or skip a grade (H1), we performed bootstrapped multinomial
regressions with school progress as outcome (1 = normal progression, 2 = repeated grade,
3 = skipped grade, and 4 = special education). Normal school progress was used as reference
category, and relative-age effects on school progress were expressed in odds ratios (ORs).
Hypotheses H2 to H7 were tested with partial Pearson’s correlations (rp) adjusted for age at
testing. In addition, we performed a series of linear regression analyses in which relative-age
predicted, respectively, length and weight (H2a), pubertal status (H2b), school performance
(H3), sport competence (H4), depressive symptoms (H6a), fear and frustration (H7a), and
change in depressive symptoms (H6b) or temperament (H7b), adjusted for age at testing.
To obtain change scores for depressive symptoms and temperamental fear and frustration
we subtracted T1 from T3 scores. To test relative-age effects on being rejected (H5a) and
popular (H5b), we applied binary logistic regression analyses adjusted for age at testing.
All hypotheses were tested in adolescents with a normal school progress and in those
who had repeated a grade (with the exception of H5 because of insufficient peer nomination
data). We lacked power to test effects in the group that skipped a grade (see appendix for
calculations). To ensure the robustness of our results, and because height, weight and BMI
109RELATIVE-AGE EFFECTS IN ADOLESCENTS
were non-normally distributed, we bootstrapped all regression analyses (k = 10000 with bias
corrected confidence intervals), see (Li, Wong, Lamoureux, & Wong, 2012). We calculated
the power for our regression analyses, and to enable comparison with other literature, we
converted some results to Cohen’s d (standardized effect sizes), based on formulas derived
from Borenstein et al. (2009), and Peterson and Brown (2005). To reduce family-wise alpha
inflation, we only interpreted correlations that were significant at p<0.01.
Finally, we performed two robustness checks. First, we repeated all analyses adjusted for
family SES. Second, to circumvent the strong positive association between relative-age
and biological age, we applied an alternative approach to adjust for relative-age, and com-
bined the relative old children from grade 7 (month 7-12) and relative young from grade 8
(month 1-6). In this control sample of 794 children with a normative school progress rel-
ative-age showed an inverse association with biological age, and was used to repeat our
analyses (H2-7).
RESULTS
Descriptive Statistics
Descriptive statistics of all variables used in this study are reported in table 1. The correlations
shown in table 2 indicate that relative old adolescents were still somewhat larger, heavier,
and in a more advanced pubertal development stage than the relative young adolescents.
School Progress
Relative-age effects on school progress are reported in table 3. For each additional month,
relative-age was associated with 17% lower odds of grade repetition and a 47% increased
odds of skipping a grade, see also figure 1. The relatively young quartile (July to Sept) was
almost four times more likely to repeat a grade (29.6% vs. 8.2%) and over twenty times less
likely to skip a grade (0.3% vs. 7%) than the relatively old quartile (Oct to Dec); all details
can be found in the appendix (table A1). Compared to adolescents with a normative school
progress adolescents who repeated a grade were on average 8 weeks younger and those
who skipped a grade were on average 14 weeks older (75% of the latter were relatively old).
As shown in table 3, no association was found for special education (d≈0.03; 95% CI = -0.06 to
0.01). Notably, after adjustment for SES, the effects became slightly stronger, and the effect
on special education significant (see appendix, table A3).
Adolescents with a Normative School Progress (75.4%, n = 1681)
Recall that all results in the following section were adjusted for age at testing. Partial cor-
relations and linear regression models (table 4) showed that relative-age effects predicted
CHAPTER VI110
temperamental frustration (d≈0.22); but this effect disappeared after adjustment for SES (see
appendix, table A2). We further observed that relative-age predicted social rejection (OR = 1.12,
p<0.001), but was unrelated to popularity (OR = 0.99, see appendix, table A4). This rather small
relative-age effect on rejection followed a u-shape, favoring the middle quartiles (9.3%, 6.1%,
6.5%, 9.8%, respectively; all details are given in the appendix, table A5). All other associations
were absent. Because relative-age correlated ( r = 0.53) with biological age (see table 2), we
calculated a control relative-age sample (see method section), which correlated ( r = -0.28)
with biological age (see table 2), but led to similar results (see table 4). In sum, there were no
substantial relative-age effects in the adolescents with a normative school progress.
Adolescents who had Repeated a Grade (16.9%, n = 377)
Though adolescents who had repeated a grade were inherently relative old compared to
their new peers (with a normative school progress), relative-age effects might still play a
role for adolescents who repeated a grade. Partial correlations, presented in table 4, showed
slightly lower intellectual ability and more depressive symptoms for the relatively old ado-
lescents who repeated a grade; a reversed relative-age effect. As shown in table 4, linear re-
gression models showed that relative older adolescents were heavier (d≈0.39), had a higher
BMI (d≈0.42), lower intellectual ability (d≈0.35), and reported more depressive symptoms
(d≈0.50). These results persisted after adjustment for SES (see appendix, table A2).
Perc
enta
ge
Relative-Age in Months
01Sept
02Aug
03Jul
04Jun
05May
06Apr
07March
08Feb
09Jan
10Dec
11Nov
12Oct
100%
80%
60%
40%
20%
0%
Figure 1. School Progress stratified over Relative-Age Positioning.
Special Education Normative Progress Skipped Grade Repeated GradeSchool Progress:
111RELATIVE-AGE EFFECTS IN ADOLESCENTS
Table 1. Descriptive Statistics of the Variables.
Variable Wave N Range Mean SD
Relative-Age 2230 1 to 12 6.20 3.44
Relative-Age (control) 794 1 to 12 8.02 2.99
Age 1 2230 10.01 to 12.58 11.11 0.55
Fear 1 1982 1 to 5 2.42 0.73
Frustration 1 1983 1 to 4.80 2.79 0.66
Depressive Symptoms 1 2024 0 to 2.31 0.48 0.36
SES 1 2188 -1.94 to 1.73 -0.05 0.80
Age in Years 2 2149 12.15 to 15.15 13.65 0.53
Length (cm) 2 2041 131 to 195 164.85 8.24
Weight (kg) 2 2030 29 to 134 52.84 11.08
BMI 2 2028 12.23 to 40.20 19.00 3.21
Physical Development 2 2087 1 to 20 9.34 3.38
Intellectual Development 2 1534 1 to 20 11.64 3.71
Social Status 2 1007 1 to 5 3.40 1.31
Sport Competence 2 1455 1 to 5 3.48 0.64
Age 3 1819 14.69 to 18.69 16.27 0.73
∆ Fear 1 to 3 1396 -3.57 to 4.12 0.05 1.04
∆ Frustration 1 to 3 1397 -4.07 to 3.11 0.02 0.99
∆ Depressive Symptoms 1 to 3 1343 -3.63 to 4.94 0.00 1.06
N = 2230 (50.8% women), T1 = baseline wave, T3 = third measurement wave, ∆ = change score between T1 and T3, BMI = body mass index, cm = centimeter, k = number of categories, kg = kilogram, N = number of participants, SD = standard deviation.
CHAPTER VI112
Table 2. Pearson Correlations among all Study Variables.
Wave 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.
1. Relative-Age —
2. Length 2 0.15*** —
3. Weight 2 0.13*** 0.62*** —
4. BMI 2 0.08*** 0.19*** 0.88*** —
5. Physical Dev. 2 0.15*** 0.46*** 0.43*** 0.29*** —
6. Intellectual Dev. 2 -0.01 0.02 -0.01 -0.03 0.03 —
7. Popular Status 2 0.00 -0.04 -0.08 -0.08 0.08 0.10* —
8. Rejected Status 2 0.06 0.04 0.08 0.10 0.06 0.02 — —
9. Sport Competence 2 -0.02 -0.05 -0.16*** -0.17*** -0.02 0.28*** 0.12* -0.06 —
10. Fear 1 -0.02 -0.03 0.03 0.06* 0.03 -0.04 0.10* 0.02 -0.03 —
11. Frustration 1 0.01 0.04 0.07** 0.07** 0.02 -0.05 -0.06 0.11** -0.05 0.31*** —
12. Depressive Sx 1 -0.01 -0.03 0.03 0.04 0.01 -0.02 -0.10* 0.17*** -0.12*** 0.29*** 0.32*** —
13. ΔFear 1-3 -0.01 -0.02 -0.02 0.01 0.03 -0.06* -0.02 -0.02 -0.03 -0.51*** -0.08** -0.07* —
14. ΔFrustration 1-3 0.03 -0.02 -0.02 -0.02 0.02 -0.07* 0.04 -0.06 -0.05 -0.11*** -0.49*** -0.08** 0.26*** —
15. Δ Depressive Sx 1-3 0.03 0.00 0.05 0.07** 0.11*** 0.00 0.06 -0.08 -0.04 -0.08** -0.08** -0.52*** 0.19*** 0.18***
16. Rel. Age (control) 1.00*** 0.02 0.07 0.08* 0.02 0.01 0.03 -0.07*** 0.01 -0.01 0.05 0.03 -0.04 0.04 0.01
17. Biological Age 0.53*** 0.15*** 0.14*** 0.09*** 0.16*** -0.04 -0.02 -0.02 -0.03 -0.07** -0.04 -0.05* 0.08** 0.06* 0.07* -0.28***
N = 2230 (50.8% women). Rel. Age = Relative-Age, T1 = baseline wave, T3 = third measurement wave, ∆ = change score between T1 and T3. Table 1 gives details (e.g., ages). For popular and rejected social status we report biserial correlations (e.g., being popular or rejected or not), because the scale was artificially dichotomous. Partial correlations adjusted for real age at testing, which show the relative-age effects, are presented in table 4. Significance ***p<0.001, **p<0.01, *p<0.05, two-tailed.
113RELATIVE-AGE EFFECTS IN ADOLESCENTS
Table 2. Pearson Correlations among all Study Variables.
Wave 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.
1. Relative-Age —
2. Length 2 0.15*** —
3. Weight 2 0.13*** 0.62*** —
4. BMI 2 0.08*** 0.19*** 0.88*** —
5. Physical Dev. 2 0.15*** 0.46*** 0.43*** 0.29*** —
6. Intellectual Dev. 2 -0.01 0.02 -0.01 -0.03 0.03 —
7. Popular Status 2 0.00 -0.04 -0.08 -0.08 0.08 0.10* —
8. Rejected Status 2 0.06 0.04 0.08 0.10 0.06 0.02 — —
9. Sport Competence 2 -0.02 -0.05 -0.16*** -0.17*** -0.02 0.28*** 0.12* -0.06 —
10. Fear 1 -0.02 -0.03 0.03 0.06* 0.03 -0.04 0.10* 0.02 -0.03 —
11. Frustration 1 0.01 0.04 0.07** 0.07** 0.02 -0.05 -0.06 0.11** -0.05 0.31*** —
12. Depressive Sx 1 -0.01 -0.03 0.03 0.04 0.01 -0.02 -0.10* 0.17*** -0.12*** 0.29*** 0.32*** —
13. ΔFear 1-3 -0.01 -0.02 -0.02 0.01 0.03 -0.06* -0.02 -0.02 -0.03 -0.51*** -0.08** -0.07* —
14. ΔFrustration 1-3 0.03 -0.02 -0.02 -0.02 0.02 -0.07* 0.04 -0.06 -0.05 -0.11*** -0.49*** -0.08** 0.26*** —
15. Δ Depressive Sx 1-3 0.03 0.00 0.05 0.07** 0.11*** 0.00 0.06 -0.08 -0.04 -0.08** -0.08** -0.52*** 0.19*** 0.18***
16. Rel. Age (control) 1.00*** 0.02 0.07 0.08* 0.02 0.01 0.03 -0.07*** 0.01 -0.01 0.05 0.03 -0.04 0.04 0.01
17. Biological Age 0.53*** 0.15*** 0.14*** 0.09*** 0.16*** -0.04 -0.02 -0.02 -0.03 -0.07** -0.04 -0.05* 0.08** 0.06* 0.07* -0.28***
CHAPTER VI114
DISCUSSION
In this study we tested effects of intra-classroom relative-age position on multiple domains
of functioning and well-being in adolescence. Three key observations merit further discussion.
First, we observed substantial relative-age effects on school progress; relative young adoles-
cents repeated a grade about four times more often than the relatively old, who in turn were
over 20 times more likely to skip a grade. These observations align with our first hypothesis
(H1), and replicate earlier studies (Cobley et al., 2009; Doornbos, 1971; Verachtert et al., 2010).
Second, in adolescents with a normative school progress (75.4%), no substantive relative-
age effects were observed. Although we observed a small effect on peer rejection (H5a;
which might be a chance finding), we could not replicate previously reported relative-age
effects on weight (H2a), pubertal status (H2b), school performance (H3), sport-competence
(H4), or peer status in terms of popularity (H5b). Neither did we observe substantial effects
on depressive symptoms (H6a), temperamental negative affectivity (H7a), and changes in
depressive symptoms (H6b) or negative affectivity (H7b) between age 11 and 16. Third, in
the subgroup of adolescents who had repeated a grade (16.9%), inverse relative-age effects
were observed; the relatively young were thinner (weight and BMI), had higher school marks,
and reported less depressive symptoms than their relatively older peers.
Relative-Age Effects
The absence of substantial relative-age effects in adolescents with a normative school prog-
ress is surprising, given the existing literature (Doornbos, 1971; Goodman et al., 2003; Martin
et al., 2004; Verachtert et al., 2010), which also covers adolescents between age 11 and
17 (Allen, 2008; Bell & Daniels, 1990; Musch & Grondin, 2001). Some studies reported that
Table 3. Relative-Age Effects on School Progress (Normative Development is reference).
Binary Outcome OR CI (95%)
Repeated Grade 0.83*** (0.80 to 0.86)
Skipped Grade 1.47*** (1.30 to 1.67)
Special Education 0.96 (0.91 to 1.01)
N = 2230 (50.8% women). Regression estimates are based on bootstrapping (k = 10000). CI = Bias Corrected and Accelerated confidence intervals, OR = Ratio of the probability (odds) of an event occurring in one group to the odds of it occurring in another group.***p<0.001, **p<0.01, *p<0.05, two-tailed.
115RELATIVE-AGE EFFECTS IN ADOLESCENTS
Normative School Progress
Repeated Class Control Sample (robustness check)
Variable Wave rp B CI (95%) rp B CI (95%) rp B CI (95%)
Weight (kg) 2 -.03 -0.10 -0.28 to 0.08 .09 0.57* 0.07 to 1.09 .01 0.05 -0.21 to 0.30
BMI 2 -.01 -0.01 -0.06 to 0.04 .11 0.19* 0.04 to 0.36 .00 0.04 -0.03 to 0.11
Pubertal Status
2 -.01 -0.04 -0.10 to 0.02 .09 0.07 -0.07 to 0.19 -.02 -0.03 -0.11 to 0.05
Intellectual Development
2 .04 0.02 -0.00 to 0.04 -.14* -0.05* -0.10 to -0.01 .05 0.04 -0.07 to 0.15
Sport Competence
2 .01 0.00 -0.02 to 0.02 -.02 -0.01 -0.05 to 0.04 .02 0.00 -0.00 to 0.00
Fear 1 -.02 0.01 -0.01 to 0.02 .10 0.04 0.00 to 0.08 -.03 -0.01*** -0.01 to -0.01
Frustration 1 .05* 0.01* 0.01 to 0.02 .05 0.02 -0.03 to 0.06 .03 0.01 -0.01 to 0.02
Depressive Symptoms
1 .02 0.00 -0.01 to 0.01 .18*** 0.07** 0.02 to 0.11 .03 0.00 -0.01 to 0.01
Δ Fear 1-3 -.05 -0.02 -0.04 to 0.00 -.11 -0.06 -0.13 to 0.02 -.06 -0.02*** -0.02 to -0.02
Δ Frustration 1-3 -.01 -0.00 -0.23 to 0.02 -.04 -0.03 -0.09 to 0.04 .02 0.01* 0.01 to 0.01
Δ Depressive Symptoms
1-3 .01 0.00 -0.02 to 0.02 -.08 -0.03 -0.09 to 0.00 -.01 -0.01*** -0.01 to -0.01
Table 4. Relative-Age Effects, adjusted for Actual Age, as predictor of Multiple Domains, for Adolescents with a Normative School Progress (n = 1681) and Adolescents who had Repeated a Grade (n = 377). Finally, the Control Sample (n = 794) comprised Relative Old Children from Grade 7 and Relative Young from Grade 8 with a Normative School Progress.
T1 = baseline wave, T3 = third measurement wave, ∆ = change between T1 (age 11) and T3 (Age 16), BMI = body mass index, CI = Bias Corrected and Accelerated confidence intervals, rp = partial correlations between relative-age and outcome, adjusted for real age at time of testing. Regression estimates were bootstrapped (k = 10000, bias corrected intervals), and indicate change in outcome per month in relative-age, after adjustment for age at testing. Note that for change variables we also adjusted for change in age between T1 and T3. Details on all measures and procedures can be found in the method section. All correlations between all variables are given in table 2, and SES-adjusted regression estimates in the appendix (table A2). Significance ***p<0.001, **p<0.01, *p<0.05, two-tailed.
CHAPTER VI116
relative-age effects reverse for relatively young adolescents who managed to stay in their
initial cohort (Dalton, 2012; McDonald, 2001), called the extended Akerlof/Kranton model
(Allen, 2008), but we neither replicated this. It has been commonly argued that relative-age
advantages erode when full maturity in the specific system has been reached, in contrast to
advantages conferred by genes (Allen, 2008). Some studies indeed reported this dissipation
of differences (Hutchison & Sharp, 1999; Verachtert et al., 2010).
Many studies reported relative-age effects on health, educational attainment, earnings, and
mortality that persisted into adulthood (Bedard & Dhuey, 2006; Cobley et al., 2009; Persico et
al., 2004). This suggests that relative-age advantages modulate personal and social devel-
opment such that perpetuating mechanisms “get under the skin”, e.g., via identity-formation
(Fenzel, 1992; Thompson et al., 2004), opportunity costs (Bedard & Dhuey, 2006; Flynn,
2009; Gladwell, 2008; Persico et al., 2004), and/or development of (soft) skills (Dhuey & Lip-
scomb, 2008; Musch & Grondin, 2001). We may interpret the substantial relative-age effects
on school progress along similar lines, an amplification of small differences.
Ability Streaming
The absence of substantive relative-age effects in adolescents with a normative school prog-
ress may be a Dutch cultural artifact. In our sample 20% of the adolescents at age 11 had
repeated or skipped a grade (another 6% went to special education), which is already more
than the 1% in the UK and 12% in the USA at age 15 (Dalton, 2012). Ability streaming might
have diluted the relative-age effects in the adolescents with a normative school progress. Al-
ternative (and convergent) explanations are that the Dutch environment is less competitive,
which is a known moderator of relative-age effects (Musch & Grondin, 2001), or that combined
classrooms (with multiple grades) dilute effects. Finally, it may be that positive spillover from
relative older peers result in a net zero effect for the relative young.
This ability streaming hypothesis aligns in a slightly unexpected way with the hypothesized
reciprocal feedback loops in which individuals shape (select/evoke) their environments to
their propensities, in analogy to the corresponsive principle (Caspi & Shiner, 2006) and Dick-
Flynn model (Beam & Turkheimer, 2013; Flynn, 2009). The corresponsive principle predicts
that small intrinsic differences can become amplified by intra-classroom relative-age position
effects, and this process may become catalyzed by strong ability streaming (via its outcome,
that is, a change of grade) which results in a qualitative different classroom environment.
We may understand the reversed relative-age effects in adolescents who repeated a grade
(lower school performance and more depressive symptoms for the relative old than the rela-
tive young) also in terms of the ability-streaming hypothesis; with increasing relative-age
low innate potential (or maturity) rather than chronological age (the cultural relative-age
117RELATIVE-AGE EFFECTS IN ADOLESCENTS
artifact) becomes the preeminent reason for grade retention. In addition, relatively young
retenders also lost their undesirable intra-cohort position (they became relatively old),
whereas the relatively old became conspicuously old (Doornbos, 1971). Negative long-term
effects of grade retention have been reported before (Wu, West, & Hughes, 2010).
Strengths and Limitations
Our results should be interpreted in light of the following strengths and limitations. A no-
table strength of this study is the large sample of adolescents from the general Dutch popu-
lation. We are quite confident that substantial relative-age effects are absent in our sample,
because our power calculations (in the appendix) for our specific regression analyses in-
dicated that we would observe effects that explained 1% of the variance in each outcome
variable in adolescents with normative school progress (≥d = 0.18) and 3% in the group who
repeated a grade (≥d = 0.34). Moreover, we repeated analyses with a control relative-age
sample, which, despite a negative association with biological age (r = -0.28 versus r = 0.53),
showed similar results. Nevertheless, our adjusted regression analyses lack the rigor of an
experimental design.
The biggest limitation is our lack of knowledge about why children repeated a grade. For
example, some studies reported that some high SES parents deliberately delay their relative
young child’s entry into school (“academic redshirting”) to create a favorable relative old age
position (Bedard & Dhuey, 2006; Graue & DiPerna, 2000; Musch & Grondin, 2001). This sug-
gests cumulative disadvantages for relatively young children from a low SES background
(Dhuey & Lipscomb, 2008). However, we did not observe more relatively young adolescents
from families in the low SES quartile (see appendix table A1), and our results remained un-
changed after adjustment for family SES. Notably, alternative explanations like season-of-
birth effects have been refuted in earlier work, because relative-age effects were observed
in both hemispheres (Bedard & Dhuey, 2006; Musch & Grondin, 2001) and remained after sta-
tistical adjustment for seasonal effects (Lawlor et al., 2006). However, future studies might
explore gender effects, because most studies suggest that relative-age effects are stronger
in males (Musch & Grondin, 2001; Cascio & Schanzenbach, 2007; Sandercock et al., 2013).
CONCLUSION
We studied the relative-age effect bestowed upon children by an adult-imposed structuring
of their worlds, a cultural artifact that may modulate the development of their innate abili-
ties. Our preeminent observation is that intra-classroom position influenced school prog-
ress, and rendered relatively young adolescents more likely to repeat a grade. Second, we
CHAPTER VI118
could not replicate substantial relative-age effects on physical and psychosocial develop-
ment and well-being in adolescents with a normative progress. We argued that this absence
of relative-age effects in adolescents with a normative school progress may reflect a “wash-
out effect” due to ability streaming. This ability streaming (grade retention) may distinguish
the Dutch sample from earlier studies in Britain and the USA. Third, in the subgroup of ado-
lescents who repeated a grade, reverse relative-age effects were observed; the relative old
adolescents performed worse in school and reported more depressive symptoms than their
relatively younger peers did. Future studies might explore in more detail the mechanisms by
which relative-age effects may get internalized, and models that can explain alleged cultural
differences in their effects.
119RELATIVE-AGE EFFECTS IN ADOLESCENTS
REFERENCES
Abela, J. R. Z., Fishman, M. B., Cohen, J. R., & Young, J. F. (2012). Personality predispositions to depression in children of affectively-ill parents: The buffering role of self-esteem. Journal of Clinical Child and Ado-lescent Psychology, 41(4), 391-401. doi:10.1080/15374416.2012.654463
Achenbach, T. M. (1991a). Manual for the child behavior Checklist/4-18 and 1991 profiles. Burlington, VT: Department of Psychiatry, University of Vermont.
Achenbach, T. M. (1991b). Manual for the youth self-report and 1991 profile. Burlington: University of Ver-mont.
Achenbach, T. M., Dumenci, L., & Rescorla, L. A. (2003). DSM-oriented and empirically based approaches to constructing scales from the same item pools. Journal of Clinical Child and Adolescent Psychology, 32(3), 328-340. doi:10.1207/S15374424JCCP3203_02
Allen, J. (2008). Relative age, identity and schooling: An extension of the Akerlof/Kranton model which solves a puzzle. The Journal of Socio-Economics, 37(1), 343-352. doi:http://dx.doi.org/10.1016/j.so-cec.2007.06.011
Beam, C. R., & Turkheimer, E. (2013). Phenotype-environment correlations in longitudinal twin models. De-velopment and Psychopathology, 25(1), 7-16. doi:10.1017/S0954579412000867
Bedard, K., & Dhuey, E. (2006). The persistence of early childhood maturity: International evidence of long-run age effects. The Quarterly Journal of Economics, 121(4), 1437-1472. doi:10.1093/qje/121.4.1437
Bell, J. F., & Daniels, S. (1990). Are summer-born children disadvantaged - the birth-date effect in education. Oxford Review of Education, 16(1), 67-80. doi:10.1080/0305498900160106
Bobak, M., & Gjonca, A. (2001). The seasonality of live birth is strongly influenced by sociodemographic fac-tors. Human Reproduction, 16(7), 1512-1517. doi:10.1093/humrep/16.7.1512
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. R. (2009). Introduction to meta-analysis. United Kingdom: John Wiley & Sons Ltd.
Boyce, W. T., Obradovic, J., Bush, N. R., Stamperdahl, J., Kim, Y. S., & Adler, N. (2012). Social stratification, classroom climate, and the behavioral adaptation of kindergarten children. Proceedings of the National Academy of Sciences, 109, 17168-17173.
Buckles, K. S., & Hungerman, D. M. (2013). Season of birth and later outcomes: Old questions, new answers. Review of Economics and Statistics, 95(3), 711-724. doi:10.1162/REST_a_00314
Carruthers, P., Laurence, S., & Stich, S. (Eds.). (2006). The innate mind: Culture and cognition (vol. 2). Lon-don: Oxford University Press.
Cascio, E., & Schanzenbach, D. W. (2007). First in the class? Age and the education production function. (NBER Working Paper No. 13663). Cambridge, MA: National Bureau of Economic Research.
Caspi, A., & Shiner, R. (2006). Personality development. In N. Eisenberg (Ed.), Handbook of child psychology (Vol. 3 ed., pp. 300). Hobokon, New Jersey: John Wiley & Sons, Inc.
Cobley, S., McKenna, J., Baker, J., & Wattie, N. (2009). How pervasive are relative age effects in secondary school education? Journal of Educational Psychology, 101(2), 520-528. doi:10.1037/a0013845
Collins, W. A., Welsh, D. P., & Furman, W. (2009). Adolescent romantic relationships. Annual Review of Psy-chology, 60(1), 631-652.
Cyranowski, J., Frank, E., Young, E., & Shear, M. (2000). Adolescent onset of the gender difference in lifetime rates of major depression - A theoretical model. Archives of General Psychiatry, 57(1), 21-27. doi:10.1001/archpsyc.57.1.21
Dalton, B. (2012). Grade level and science achievement: US performance in cross-national perspective. Comparative Education Review, 56(1), 125-154. doi:10.1086/660745
CHAPTER VI120
Deardorff, J., Fyfe, M., Ekwaru, J. P., Kushi, L., Greenspan, L., & Yen, I. (2012). Does neighborhood environ-ment influence girls’ pubertal onset? Findings from a cohort study. BMC Pediatrics, 12(1), 27.
de Winter, A. F., Oldehinkel, A. J., Veenstra, R., Brunnekreef, J. A., Verhulst, F. C., & Ormel, J. (2005). Evalua-tion of non-response bias in mental health determinants and outcomes in a large sample of pre-adoles-cents. European Journal of Epidemiology, 20(2), 173-181. doi:10.1007/s10654-004-4948-6
Dhuey, E., & Lipscomb, S. (2008). What makes a leader? Relative age and high school leadership. Economics of Education Review, 27(2), 173-183. doi:10.1016/j.econedurev.2006.08.005
Dijkstra, J. K., Cillessen, A. H. N., & Borch, C. (2013). Popularity and adolescent friendship networks: Selec-tion and influence dynamics. Developmental Psychology, 49(7), 1242–1252. doi:10.1037/a0030098
Doornbos, K. (1971). Geboortemaand en schoolsucces: Een onderwijskundig onderzoek naar de samen-hang tussen de leeftijdspositie binnen de jaargroep en het schoolsucces van leerlingen in het basis, voortgezet en buitengewoon onderwijs. [Birth month and school success: An educational study about the association between intra-class relative-age position and the education success of pupils in pri-mary and secondary education]. PhD thesis, Rijksuniversiteit Groningen.
Eaves, L., Silberg, J., Foley, D., Bulik, C., Erkanli, A., Angold, A., et al. (2004). Genetic and environmental influ-ences on the relative timing of pubertal change. Twin Research, 7(5), 471.
Fenzel, L. M. (1992). The effect of relative age on self-esteem, role strain, GPA, and anxiety. The Journal of Early Adolescence, 12(3), 253-266. doi:10.1177/0272431692012003002
Ferdinand, R. F. (2008). Validity of the CBCL/YSR DSM-IV scales anxiety problems and affective problems. Journal of Anxiety Disorders, 22(1), 126-134. doi:http://dx.doi.org/10.1016/j.janxdis.2007.01.008
Flynn, J. R. (2009). What is intelligence? (2nd ed.). Cambridge: Cambridge University Press.
Fraley, R. (2002). Attachment stability from infancy to adulthood: Meta-analysis and dynamic modeling of developmental mechanisms. Personality and Social Psychology Review, 6(2), 123-151. doi:10.1207/S15327957PSPR0602_03
Gladwell, M. (2008). Outliers: The story of success. USA: Little, Brown and Company.
Gledhill, J., Ford, T., & Goodman, R. (2002). Does season of birth matter? The relationship between age within the school year (season of birth) and educational difficulties among a representative general popu-lation sample of children and adolescents (aged 5-15) in Great Britain. Research in Education, 68(1), 41-47. doi:10.7227/RIE.68.4
Goodman, R., Gledhill, J., & Ford, T. (2003). Child psychiatric disorder and relative age within school year: Cross sectional survey of large population sample. British Medical Journal, 327(7413), 472-475. doi:10.1136/bmj.327.7413.472
Graue, M., & DiPerna, J. (2000). Redshirting and early retention: Who gets the “gift of time” and what are its outcomes? American Educational Research Journal, 37(2), 509-534. doi:10.2307/1163532
Hankin, B., Abramson, L., Moffitt, T., Silva, P., McGee, R., & Angell, K. (1998). Development of depression from preadolescence to young adulthood: Emerging gender differences in a 10-year longitudinal study. Journal of Abnormal Psychology, 107(1), 128-140. doi:10.1037//0021-843X.107.1.128
Harris, J. R. (1995). Where is the childs environment - A group socialization theory of development. Psycho-logical Review, 102(3), 458-489. doi:10.1037/0033-295X.102.3.458
Harris, J. R. (2009). The nurture assumption: Why children turn out the way they do (2nd ed.). New York, USA: Free Press, Simon & Schuster, Inc.
Hartman, C. A. (2000). Dutch translation of the early adolescent temperament questionnaire. Internal Re-port, Department of Psychiatry, University of Groningen, the Netherlands.
Henrich, J., & Gil-White, F. (2001). The evolution of prestige - freely conferred deference as a mechanism for enhancing the benefits of cultural transmission. Evolution and Human Behavior, 22(3), 165-196. doi:10.1016/S1090-5138(00)00071-4
121RELATIVE-AGE EFFECTS IN ADOLESCENTS
Huisman, M., Oldehinkel, A. J., de Winter, A. F., Minderaa, R. B., de Bildt, A., Huizink, A. C., et al. (2008). A cohort profile: The Dutch ‘TRacking Adolescents’ Individual Lives’ Survey’; TRAILS. International Journal of Epidemiology, 37(6), 1227-35. doi:10.1093/ije/dym273
Hutchison, D., & Sharp, C. (Sept, 1999). A lasting legacy? The persistence of season of birth effects. British Educational Research Association Conference, National Foundation for Educational Research, 18-10-2013. doi: http://www.leeds.ac.uk/educol/documents/000001095.htm
Janssens, K. A. M., Rosmalen, J. G. M., Ormel, J., Verhulst, F. C., Hunfeld, J. A. M., Mancl, L. A., et al. (2011). Pubertal status predicts back pain, overtiredness, and dizziness in American and Dutch adolescents. Pediatrics, 128(3), 553-559. doi:10.1542/peds.2010-2364
Lawlor, D. A., Clark, H., Ronalds, G., & Leon, D. A. (2006). Season of birth and childhood intelligence: Findings from the Aberdeen children of the 1950s cohort study. British Journal of Educational Psychology, 76, 481-499. doi:10.1348/000709905X49700
Li, X., Wong, W., Lamoureux, E. L., & Wong, T. Y. (2012). Are linear regression techniques appropriate for analysis when the dependent (outcome) variable is not normally distributed? Investigative Ophthalmol-ogy & Visual Science, 53(6), 3082-3083. doi:10.1167/iovs.12-9967
Martin, R. (2013). How we do it. The evolution and future of human reproduction. New York: Basic Books.
Martin, R., Foels, P., Clanton, G., & Moon, K. (2004). Season of birth is related to child retention rates, achievement, and rate of diagnosis of specific LD. Journal of Learning Disabilities, 37(4), 307-317. doi:10.1177/00222194040370040301
McDonald, G. (2001). Comparing school systems to explain enduring birth date effects. Compare: A Journal of Comparative and International Education, 31(3), 381-391. doi:10.1080/03057920120098509
Morrow, R. L., Garland, E. J., Wright, J. M., Maclure, M., Taylor, S., & Dormuth, C. R. (2012). Influence of relative age on diagnosis and treatment of attention-deficit/hyperactivity disorder in children. Canadian Medical Association Journal, 184(7), 755-762. doi:10.1503/cmaj.111619
Muhlenweg, A. M. (2010). Young and innocent international evidence on age effects within grades on vic-timization in elementary school. Economics Letters, 109(3), 157-160. doi:10.1016/j.econlet.2010.08.032
Musch, J., & Grondin, S. (2001). Unequal competition as an impediment to personal development: A review of the relative age effect in sport. Developmental Review, 21(2), 147-167. doi:10.1006/drev.2000.0516
Newcomb, A. F., Bukowski, W. M., & Pattee, L. (1993). Childrens peer relations – a meta-analytic review of popular, rejected, neglected, controversial, and average sociometric status. Psychological Bulletin, 113(1), 99-128. doi:10.1037//0033-2909.113.1.99
Oldehinkel, A. J., Hartman, C. A., de Winter, A. F., Veenstra, R., & Ormel, J. (2004). Temperament profiles associated with internalizing and externalizing problems in preadolescence. Development and Psycho-pathology, 16(2), 421-440. doi:10.1017/S0954579404044591
Oldehinkel, A. J., Wittchen, H., & Schuster, P. (1999). Prevalence, 20-month incidence and outcome of unipo-lar depressive disorders in a community sample of adolescents. Psychological Medicine, 29(3), 655-668. doi:10.1017/S0033291799008454
Oldehinkel, A. J., Rosmalen, J. G. M., Veenstra, R., Dijkstra, J. K., & Ormel, J. (2007). Being admired or being liked: Classroom social status and depressive problems in early adolescent girls and boys. Journal of Abnormal Child Psychology, 35(3), 417-427. doi:10.1007/s10802-007-9100-0
Ormel, J., Oldehinkel, A. J., Sijtsema, J., van Oort, F., Raven, D., Veenstra, R., & Verhulst, F. C. (2012). The TRacking Adolescents’ Individual Lives Survey (TRAILS): Design, current status, and selected findings. Journal of the American Academy of Child & Adolescent Psychiatry, 51(10), 1020-1036. doi:10.1016/j.jaac.2012.08.004
Persico, N., Postlewaite, A., & Silverman, D. (2004). The effect of adolescent experience on labor market outcomes: The case of height. National Bureau of Economic Research Working Paper Series, No. 10522.
CHAPTER VI122
Peterson, R. A., & Brown, S. P. (2005). On the use of beta coefficients in meta-analysis. Journal of Applied Psychology, 90(1), 175-181. doi:10.1037/0021-9010.90.1.175
Pierson, K., Addona, V., & Yates, P. (2014). A behavioural dynamic model of the relative age effect. Journal of Sports Sciences, 32(8), 776-784. doi:10.1080/02640414.2013.855804
Putnam, S. P., Ellis, S. K., & Rothbart, M. K. (2001). The structure of temperament from infancy through adolescence. In A. Eliasz, & A. Engleneiter (Eds.), Advances in research on temperament (pp. 165-182). Berlin: Pabst Scientist Publicer.
Rothbart, M. K. (2011). Becoming who we are. New York, NY: Guilford Press.
Rothbart, M. K., Ahadi, S. A., & Evans, D. E. (2000). Temperament and personality: Origins and outcomes. Journal of Personality and Social Psychology, 78(1), 122-135. doi:10.1037/0022-3514.78.1.122
Sandercock, G. R., Taylor, M. J., Voss, C., Ogunleye, A. A., Cohen, D. D., & Parry, D. A. (2013). Quantification of the relative age effect in three indices of physical performance. Journal of Strength & Conditioning Research, 27(12), 3293.
Shirtcliff, E. A., Dahl, R. E., & Pollak, S. D. (2009). Pubertal development: Correspondence between hormonal and physical development. Child Development, 80(2), 327-337. doi:10.1111/j.1467-8624.2009.01263.x
Thompson, A., Barnsley, R., & Battle, J. (2004). The relative age effect and the development of self-esteem. Educational Research, 46(3), 313-320. doi:10.1080/0013188042000277368
Thompson, A., Barnsley, R., & Dyck, R. (1999). A new factor in youth suicide: The relative age effect. Cana-dian Journal of Psychiatry-Revue Canadienne De Psychiatrie, 44(1), 82-85.
Verachtert, P., De Fraine, B., Onghena, P., & Ghesquiere, P. (2010). Season of birth and school suc-cess in the early years of primary education. Oxford Review of Education, 36(3), 285-306. doi:10.1080/03054981003629896
Wu, W., West, S. G., & Hughes, J. N. (2010). Effect of grade retention in first grade on psychosocial outcomes. Journal of Educational Psychology, 102(1), 135-152. doi:10.1037/a0016664
123RELATIVE-AGE EFFECTS IN ADOLESCENTS
SUPPLEMENTARY MATERIAL
Table A1. Frequencies for School Progress and Socioeconomic Status (SES) stratified over Relative-Age Distribution in Quartiles.
Relative Young
2nd Quartile
3rd Quartile
Relative Old
Relative-Age: 1 to 3 4 to 6 7 to 9 10 to 12
Birth Month: Sept, Aug, July
June, May, April
March, Feb, Jan
Dec, Nov, Oct
Total Sample
2230 100.0% 624 100% 569 100% 537 100% 500 100%
Normative Development
1681 75.4% 398 63.8% 441 77.5% 443 82.5% 399 79.8%
Repeated Grade
377 16.9% 185 29.6% 90 15.8% 61 11.4% 41 8.2%
Skipped Grade
48 2.2% 2 0.3% 4 0.7% 7 1.3% 35 7.0%
Special Education
124 5.5% 39 6.3% 34 6.0% 26 4.8% 25 5.0%
Low SES Quartile
547 25.0% 135 24.7% 142 26.0% 149 27.2% 121 22.1%
High SES Quartile
547 25.0% 150 27.4% 141 25.8% 139 25.4% 117 21.4%
CHAPTER VI124
Table A2. Relative-Age Effects on Multiple Domains, adjusted for Actual Age at testing and Socioeconomic Status (SES), for Adolescents with a Normative School Progress (n = 1681), and Adolescents who Repeated a Grade (n = 377).
Normal School ProgressRelative-Age Effect
Repeated GradeRelative-Age Effect
Variable Wave rp B CI (95%) rp B CI (95%)
Length (cm) 2 -.04 -0.10 (-0.24 to 0.04) -.01 0.09 (-0.25 to 0.42)
Weight (kg) 2 -.04 -0.13 (-0.30 to 0.05) .08 0.60* (0.02 to 1.11)
BMI 2 -.02 -0.02 (-0.07 to 0.03) .10 0.19* (0.03 to 0.36)
Pubertal Status 2 -.02 -0.04 (-0.10 to 0.02) .08 0.07 (-0.07 to 0.19)
Intellectual Development
2 .04 0.06 (-0.01 to 0.13) -.12 -0.19* (-0.36 to -0.02)
Sport Competence
2 .01 0.00 (-0.01 to 0.02) -.01 -0.00 (-0.03 to 0.03)
Fear 1 .01 0.01 (-0.01 to 0.02) .09 0.04 (-0.00 to 0.08)
Frustration 1 .05 0.02 (-0.00 to 0.03) .04 0.02 (-0.03 to 0.06)
Depressive Symptoms
1 .01 0.00 (-0.02 to 0.02) .18*** 0.02** (0.02 to 0.11)
Δ Fear 1-3 -.06a -0.02 (-0.04 to 0.00) -.11 -0.06 (-0.13 to 0.02)
Δ Frustration 1-3 -.01 -0.00 (-0.03 to 0.02) -.04 -0.03 (-0.09 to 0.04)
Δ Depressive Symptoms
1-3 .01 0.00 (-0.02 to 0.02) -.08 -0.03 (-0.09 to 0.04)
a = p was 0.052. T1 = baseline wave, T3 = third measurement wave, ∆ = change between T1 (age 11) and T3 (Age 16), BMI = body mass index, CI = Bias Corrected and Accelerated confidence intervals, rp = partial correlations between relative-age and outcome adjusted for real age at time of testing. Regression estimates were bootstrapped (k = 10000, bias corrected intervals), and indicate change in outcome per month in relative-age, after adjustment for age at testing. Note that for change variables we also adjusted for change in age between T1 and T3. Details on all measures and procedures can be found in the method section. All correlations between all variables are given in table 2.***p<0.001, **p<0.01, *p<0.05, two-tailed.
125RELATIVE-AGE EFFECTS IN ADOLESCENTS
Table A3. Relative-Age Effects on School Progress (Normative Development is reference).
Adj. for Real Agea Adj. for Real Age & SESb
Binary Outcome OR CI (95%) OR CI (95%)
Repeated Grade 0.83*** (0.80 to 0.86) 0.82*** (0.79 to 0.86)
Skipped Grade 1.47*** (1.30 to 1.67) 1.66*** (1.44 to 1.92)
Special Education 0.96 (0.91 to 1.01) 0.81*** (0.75 to 0.87)
N = 2230 (50.8% women). a The odds are based on bootstrapping (k = 10000). b Odds are not based on bootstrapping. CI = Bias Corrected and Accelerated confidence intervals, OR = Ratio of the probability that an event will happen to all possible cases for that event. ***p<0.001, **p<0.01, *p<0.05, two-tailed.
CHAPTER VI126
Table A4. Relative-Age as the predictor of the Odds of Being Popular or Rejected (Normative Development is reference) in Adolescents with a Normative School Progress and Adolescents who Repeated a Grade, adjusted for Real Age at testing.
School Progress
Social Status
rp OR CI (95%) p B CI (95%)
Normative Popular -.00 1.00 (0.94 to 1.06) 0.90 -0.00 (-0.07 to 0.06)
Rejected .09*** 1.12 (1.05 to 1.19) 0.00 0.11 (0.04 to 0.18)
Repeated a Grade
Popular .04 1.11 (0.83 to 1.49) 0.55 0.11 (-0.53 to 0.66)
Rejected .02 1.04 (0.85 to 1.28) 0.67 0.04 (-0.20 to 0.24)
N = 2230 (50.8% women). In total 137 (8.1%) adolescents were rated as popular and 132 (7.9%) as rejected. Regression estimates are based on bootstrapping (k = 10000). CI = Bias Corrected and Accelerated confidence intervals, OR = Ratio of the probability that an event will happen to all possible cases for that event, rp = partial correlations between relative-age and outcome, adjusted for real age at time of testing. The values for the presented correlations and regression remained almost identical after additional adjustment for family SES. ***p<0.001, **p<0.01, *p<0.05, two-tailed.
127RELATIVE-AGE EFFECTS IN ADOLESCENTS
Table A5. Social Status stratified over Relative-Age Position in Quartiles and School Progress.
Normative School Progress Repeated a Grade
Rejected n Popular n Rejected n Popular n
Relatively Young
9.3% 37 7.5% 30 2.7% 5 1.6% 3
Second 6.1% 27 9.5% 42 3.3% 3 1.1% 1
Third 6.5% 29 8.4% 37 4.9% 3 1.6% 1
Relatively Old
9.8% 39 7.0% 28 4.9% 2 2.4% 1
Total 7.9% 132 8.1% 137 3.4% 13 1.6% 6
n = number of subjects.
Table A6. Quartiles of Socioeconomic Status (SES) stratified over School Progress.
Normative Repeated a Grade Skipped a Grade Special Education
% n % n % n % n
Low SES 65.3% 357 22.7% 124 0.9% 5 11.2% 61
Second Q 72.0% 394 21.4% 117 0.7% 4 5.9% 32
Third Q 77.1% 422 15.2% 83 3.8% 21 3.8% 21
High SES 87.6% 479 8.6% 47 3.1% 17 0.7% 4
n = number of subjects, Q = Quartile.
CHAPTER VI128
A7. Power Calculations
We calculated the power of all specific linear regression analyses (Selya, Rose, Dierker, He-
deker, & Mermelstein, 2012). In the adolescents with a normative school progress reasonably
precise estimations (given 80% power) were guaranteed beyond Cohen’s f 2 = 0.01 (~R2 = 0.01,
d = ~0.18, 2 predictors, α = 0.05, and N≥1100 for all outcomes). In the group of adolescents
who had repeated a grade we could estimate effects beyond Cohen’s f 2 = 0.03 (~R2 = 0.03,
d = ~0.34, given 80% power, 2 predictors, α = 0.05, N≥320 for all outcomes). However, we con-
cluded we lacked the power to test for relative-age effects in the group that skipped a grade
(N = 48, Cohen’s f 2 = 0.24, R2 = 0.19, d = ~0.98). These post-hoc power calculations indicate
that we obtained reasonably precise estimations in our study, and we are therefore confi-
dent that substantial relative-age effects are indeed absent.
REFERENCES
Selya, A. S., Rose, J. S., Dierker, L. C., Hedeker, D., & Mermelstein, R. J. (2012). A practical guide to calculating Cohen’s f2, a measure of local effect size, from PROC MIXED. Frontiers in Psychology, 3, 111. doi:10.3389/fpsyg.2012.00111
The Dynamic Relationship between Physical Activity and Mood in the Daily Life of Depressed and Non-Depressed Individuals: a Within-Subject Time-Series Analysis
“I get my exercise running to the funerals of my friends who exercise.”
Barry Gray, in New York Magazine, 1980
“Getting fit is all about mind over matter. I don’t mind, so it doesn’t matter.”
Adam Hargreaves, in Mr Lazy's Guide to Fitness, 2000
Stavrakakis N, Booij SH, Roest AM, de Jonge P, Oldehinkel AJ & Bos EH
Under preparation
CHAPTER
CHAPTER VII130
ABSTRACT
Purpose: The association between physical activity and mood found in longitudinal observa-
tional studies is generally small to moderate. It is unknown how this association generalizes
to individuals. The aim of the present study was to investigate inter-individual differences in
the bidirectional dynamic relationship between physical activity and mood, in depressed and
non-depressed individuals, using time-series analysis.
Methods: Twenty pair-matched depressed and non-depressed participants (30% males,
mean age = 36.6, SD = 8.9) wore accelerometers and completed electronic questionnaires
three times a day for 30 days. Physical activity was operationalized as the total energy ex-
penditure (EE) per day segment (i.e., 6 hours). The multivariate time series (T = 90) of every
individual were analyzed using vector autoregressive modeling (VAR), with the aim to assess
direct as well as lagged (i.e., over 1 day) associations between EE and positive and negative
affect.
Results: We observed large inter-individual differences in the strength, direction and tem-
poral aspects of the relationship between physical activity and positive and negative affect.
An exception was the direct (but not the lagged) influence of physical activity on positive
affect, which was positive in nearly all individuals.
Conclusion: This study showed that the association between physical activity and affect
varied considerably across individuals. Thus, while at the group level the effect of physical
activity on mood may be small, in some individuals the effect may be clinically relevant.
131PHYSICAL ACTIVITY AND MOOD IN DAILY LIFE
INTRODUCTION
Physical activity or exercise is considered to be an effective treatment for depression (Dunn,
Trivedi, Kampert, Clark, & Chambliss, 2005; Ströhle et al., 2007; Ströhle, 2009). Although this
proposition is appealing, since physical activity can be implemented easily and with rela-
tive low costs, the empirical evidence of the effectiveness of physical activity is not robust
(Chalder et al., 2012; Cooney et al., 2013; Krogh, Nordentoft, Sterne, & Lawlor, 2010). Simi-
larly, observational studies have shown an inverse association between physical activity
and depressive symptoms (Dunn, Trivedi, & O’Neal, 2001; Farmer et al., 1988; Salmon, 2001;
Strawbridge, Deleger, Roberts, & Kaplan, 2002), with small to moderate effect sizes in dif-
ferent populations (Biddle & Asare, 2011; Johnson & Taliaferro, 2011). Although large cohort
studies with a few measurements are useful to detect general patterns and traits in the
population, they are not suitable for studying dynamic relationships between variables that
fluctuate in daily life (Hamaker, 2012; Hamaker, Dolan, & Molenaar, 2005), which is the case
for physical activity and depressive symptoms.
Some studies have adopted a more ecologically valid approach to investigate the associa-
tion between depressive symptoms and physical activity. Ecological momentary assess-
ment (EMA) studies address the dynamic nature of factors by investigating them repeat-
edly over time in participants’ daily life. Since slight changes in positive and negative affect
might accumulate over time to result in depressed mood, it is important to investigate how
physical activity influences daily affect levels (Rosmalen, Wenting, Roest, de Jonge, & Bos,
2012). A recent literature overview of EMA studies investigating the relationship between
physical activity and momentary affective states in daily life (Kanning, Ebner-Priemer, &
Schlicht, 2013) showed a generally consistent positive association between physical activ-
ity and positive affect across studies, which became stronger when only high-quality stud-
ies were included. In contrast, the influence of physical activity on negative affect, although
observed in some studies (see for example: Dunton, Liao, Intille, Wolch, & Pentz, 2011 and
LePage & Crowther, 2010), was non-existent in the majority of studies reviewed (see: Kan-
ning et al., 2013).
The above EMA studies all used multilevel analysis to estimate dynamic associations. Al-
though multilevel analysis allows for inter-individual differences when estimating the rela-
tionship between physical activity and mood, it is in principle a between-subject (group) ap-
proach; individual regression terms are averaged across the group and the same regression
model is imposed on all subjects rather than modeling the dynamic relationship for each
subject individually. Only under strict conditions can findings at the inter-individual level be
generalized to the intra-individual level (see: Hamaker, 2012), and these conditions rarely
apply to the study of dynamic psychological processes (Hamaker et al., 2005; Molenaar &
CHAPTER VII132
Campbell, 2009). In addition, a within-subject time-series approach may be more suitable for
investigating bidirectional relationships than multilevel analysis, because these models are
built at the individual level and all variables in the model can be used as both predictors and
outcomes (Brandt & Williams, 2007). The added value of such an approach was illustrated
by a recent time-series study investigating the dynamic relationship between physical activ-
ity and depressive symptoms in four post-myocardial infarction patients (Rosmalen et al.,
2012). Inter-individual differences in the magnitude as well as the direction of the association
were shown, in spite of the relatively homogeneous study sample.
The present study aimed to further elucidate inter-individual differences in the bidirec-
tional association between physical activity and mood. Adopting an intensive time-series
approach, we studied the dynamic relationship between directly measured physical activ-
ity (i.e., accelerometers) and affect in the daily life of pair-matched depressed and non-de-
pressed individuals, and individual differences therein.
METHODS
Participants
The sample was part of the ‘Mood and movement in daily life’ (MOOVD) study, which was
set up to investigate the dynamic relationship between physical activity and mood in daily
life, and the role of several biomarkers therein. Participants (age 20-50) were intensively
monitored in their natural environments for 30 days, by means of electronic diaries, saliva
sampling, and continuous actigraphy. The multiple repeated measurements per individual
(T = 90) allowed assessing within-person temporal relationships between variables at the
individual level. For the current study, 10 depressed and 10 non-depressed participants, who
were pair-matched on gender, smoking status, age, and Βody Μass Ιndex (BMI), were in-
cluded (the first 20 participants that finished the study).
Depressed individuals were recruited from the psychiatry department of the University Med-
ical Center Groningen (UMCG) and the Center for Integrative Psychiatry (CIP) in Groningen,
the Netherlands. Non-depressed participants were recruited from the general population by
means of posters and advertisements in local newspapers. Participants were screened for
depressive symptoms with the Beck Depression Inventory version 2 (BDI-II; Beck, Erbaugh,
Ward, Mock, & Mendelsohn, 1961) and for several other health conditions with a general
health questionnaire. Individuals with a BDI-II score of >14 (i.e., depressed group) and indi-
viduals with a BDI-II score of <9 (i.e., non-depressed group) were invited for the inclusion
interview. The inclusion interview covered the Composite International Diagnostic Interview
(CIDI; World Health Organization, 1990), several questionnaires, and a briefing in which the
133PHYSICAL ACTIVITY AND MOOD IN DAILY LIFE
use of equipment and the electronic diary was explained. Depressed individuals were includ-
ed if they met DSM-IV criteria for a major depressive disorder (current episode or in remission
for no longer than 8 weeks). Non-depressed individuals were included if they were free of
mood disorders at the moment of inclusion. Individuals who reported chronic somatic illness
or medication use that may directly influence functioning of the Hypothalamic Pituitary Ad-
renal (HPA) axis or the Autonomic Nervous System (ANS) were excluded. Other reasons for
exclusion were: pregnancy, significant hearing or visual impairments, and having a current
or recent (within the last two years) psychotic or bipolar disorder as assessed with the CIDI
interview. Participants received a fee for participation and a personal report of their daily
mood and activity patterns. The MOOVD study design was approved by the Medical Ethical
Committee and all participants gave written informed consent.
Ambulatory Sampling
Participants completed questionnaires on an electronic diary, the PsyMate (PsyMate BV,
Maastricht, the Netherlands) (Myin-Germeys, Birchwood, & Kwapil, 2011) for a total of 32
days, with the first two days used to get familiar with the device. The PsyMate was pro-
grammed to generate beeps at three predetermined moments a day with equidistant in-
tervals: in the morning (mean≈10:00 AM), six hours later in the afternoon (mean≈4:00 PM),
and six hours later in the evening (mean≈10:00 PM). This fixed interval design was chosen
to allow the application of time-series analysis. To capture most of participants’ daily life
without intruding with their sleep habits, beeps were planned at the end of the morning,
afternoon, and evening, with the exact time points depending on the individual participant’s
sleep-wake schedule. This schedule was assessed by the Munich Chronotype Questionnaire
(MCTQ; Roenneberg et al., 2007) and an extra question regarding the time they would go to
bed if they were to go to bed ‘on time’. After every alarm beep, participants were asked to
fill out the electronic diary. They were instructed to do so immediately after the beep, but a
time window of one hour was used for situations in which this was not possible. Through-
out the study period, participants wore the ActiCal® (Respironics, Bend, OR, USA), an om-
nidirectional water-resistant accelerometer, on the wrist of the non-writing arm. They were
instructed never to remove the ActiCal except when entering a sauna (high temperatures).
The ActiCal recorded data over 1-min sampling intervals. For more details about the ActiCal,
see Heil (2006).
Measures
The electronic diary questionnaire contained 60 items on mood, cognition, and daily activi-
ties. Positive and negative affect scores were computed from mood items adopted from
Bylsma, Taylor-Clift, and Rottenberg (2011), rated on a 7-point Likert scale. For positive
CHAPTER VII134
affect, seven positive mood items were averaged, namely feeling talkative, enthusiastic,
confident, cheerful, energetic, satisfied, and happy (range 1-7). For negative affect, seven
negative mood ratings were averaged, namely feeling tense, anxious, distracted, restless,
irritated, depressed, and guilty (range 1-7).
Energy expenditure (EE) data from the ActiCal was used as a measure for physical activity.
For each participant, the amount of EE (kcal/min) across all sampling minutes over one day
segment (i.e., 360 minutes per segment) was summed. This resulted in total EE per day seg-
ment (kcal/day segment).
Person characteristics including age, gender, educational level, smoking status, coffee and
alcohol intake, and the average amount of exercise per week (min) were obtained from ques-
tionnaires filled out at the baseline assessment. BMI was calculated from the weight and
height of each individual.
Statistical Analysis
Missing diary data were imputed by means of Expectation-Maximization imputation in IBM
SPSS Statistics 20. For every participant, all variables used in the time-series model as well
as auxiliary variables that were significantly associated with these variables were used as
predictor variables in the imputation. On average, participants had 8 missing values on the
diary variables (8.8%). Of the 20 participants, three participants had missing actigraphy data
at the last 9 or 10 days of their study, because of technical problems. These relatively large
periods without activity records were discarded and not imputed, leaving 60-63 valid data
points for these three participants, which is still considered sufficient for the application of
time-series analysis (Lütkepohl, 2007; Rosmalen et al., 2012).
The multiple time series of every individual were analyzed using vector autoregressive (VAR)
modeling. A VAR model is a multivariate autoregressive model that consists of a set of re-
gression equations for a system of two or more variables, in this case positive affect, nega-
tive affect, and EE (Brandt & Williams, 2007). All three variables in the system were treated
as endogenous, which means that they could be both determinant and outcome. Each of
the three endogenous variables was regressed on its own p lagged values and the p lagged
values of the other variable. The errors of the VAR model should be serially uncorrelated but
can be contemporaneously correlated. VAR analysis is especially suitable for investigating
the dynamic relationships between two or more variables: inferences can be made about the
temporal order of effects, which can involve bidirectional effects and feedback loops (Brandt &
Williams, 2007). Moreover, by analyzing all variables in one system, the most important ef-
fects will be identified. For more details about VAR modeling and a non-technical introduc-
tion, see Brandt & Williams (2007) and Rosmalen et al. (2012), respectively.
135PHYSICAL ACTIVITY AND MOOD IN DAILY LIFE
The number of lags included in the models was a priori set to three, which is equivalent to a
period of one day. A fixed number of lags was chosen to ease comparison of results for dif-
ferent participants. Considering the study’s exploratory nature, a relatively high number of
lags was chosen, allowing for the investigation of relatively long-term lagged effects (i.e.,
over 3 time intervals = 1 day). To account for structural lower morning values for total EE,
compared to afternoon and evening values – participants spent part of the morning lying in
bed – as well as diurnal rhythms in mood, dummy variables for morning and evening were in-
cluded (afternoon served as a reference category). Variables for time and the square of time
were included into the model if this was necessary to render the series stationary. Total EE
was divided by 100 to accommodate the difference in scaling between EE and positive and
negative affect. Maximum likelihood estimation with a degrees-of-freedom adjustment ad-
vocated for small samples was used for estimating the VAR coefficients (Lütkepohl, 2007).
VAR model assumptions, namely stability of the model, independence, homoscedasticity
and normality of residuals, were assessed using diagnostic checks (Lütkepohl, 2007). If one
of the tests indicated a violation of the model assumptions, models were adjusted, re-esti-
mated, and re-evaluated, in an iterative model-building process, until all assumptions were
met. Outlier modeling was applied only if this was required to meet model assumptions,
using the criterion of residual SD>3 (stepwise lowering to SD>2, if necessary; Field, 2009).
Log transformation of the variables was applied if the residuals remained skewed. If scat-
terplots of skewed variables with other variables suggested curvilinear relationships, non-
skewed variables were left untransformed. In other cases, all endogenous variables were
log-transformed.
To assess the direct effect of EE on positive and negative affect, we examined the contempo-
raneous correlations between the variables, which can be retrieved from the residuals of the
final models. These correlations represent the simultaneous correlations between the vari-
ables, i.e., between the scores at the same assessment point. To assess the overall lagged
effects of EE on positive and negative affect, and vice versa, the effect size coefficients of
lags 1 to 3 were averaged into one net effect size coefficient, and standardized so that they
could be compared across participants. To test the significance of the overall lagged effects
the Granger causality Wald test was used. This is a test for the directionality of the influence
between two time series (Granger, 1969; Lütkepohl, 2007). In view of the exploratory nature
of the study, a significance level of 0.05 was used for all tests. All analyses were done in
STATA 11 using the suite of VAR commands (StataCorp., 2009).
CHAPTER VII136
RESULTS
Descriptive Statistics
The demographic and clinical characteristics of the depressed and non-depressed partici-
pants are presented in Table 1.
Analysis
Short-Term Effect of Physical Activity on Positive and Negative Affect
As depicted in Figure 1a, the majority of depressed (red bars) and non-depressed (blue bars)
participants showed a positive contemporaneous association between EE and positive af-
fect. As the EE scores covered the period directly before the mood rating, this can be inter-
preted as a direct effect of EE on positive affect. For five individuals (three non-depressed
and two depressed) these positive associations were significant (p<0.05). The size of the
correlations was small to moderate, according to the suggestions provided by Cohen (1992).
Figure 1b shows the direct association between EE and negative affect for depressed (red
bars) and non-depressed (blue bars) participants. Eight of them showed a positive (non-sig-
nificant) contemporaneous association between EE and negative affect, while nine partici-
pants showed a negative association, without any apparent differences between the groups
(depressed and non-depressed). For two individuals (one depressed and one non-depressed)
the negative association between EE and negative affect was significant (p<0.05). The size of
these correlations was again small to moderate.
137PHYSICAL ACTIVITY AND MOOD IN DAILY LIFE
Table 1. Demographic and Clinical Characteristics for the Depressed and Non-Depressed Group.
Depressed Group (n = 10) Non-Depressed Group (n = 10)
Mean SD Min Max Mean SD Min Max
Demographics
Age 36.4 10.3 22 49 36.7 7.9 24 46
Gender (% male) 30.0 — — — 30.0 — — —
Body Mass Index (BMI) 23.8 4.9 17.4 34.9 22.2 2.3 19.5 26.9
Educational Level (0-3) 2.6 0.5 2 3 2.6 0.5 2 3
Lifestyle
Coffee Intake (per day) 2.2 2.5 0 8 1.8 1.4 0 4
Alcohol Intake (per month) 10.8 14.5 0 40 8.4 14.4 0 48
Self-Reported Exercise Frequency (per week)
1.7 1.5 0 4 2 1.5 0 4
Self-Reported Duration of Exercise (min. per week)
83 89 0 240 133 109 0 240
Average EE over one day segment
233 77 126 385 258 69 120 369
Mood and Depressive Symptoms
Pre-BDI-II 32.5 10 20 51 2.8 3.4 0 10
Post-BDI-II 25.9 15.8 2 47 4.4 4.6 0 13
Average PA (1-7) 3.1 1.3 1.1 4.5 4.6 1.3 1.7 6.1
Average NA (1-7) 3.6 1.2 1.4 5.3 1.6 0.6 1.0 2.8
BDI-II = Beck Depression Inventory, EE = energy expenditure, PA = positive affect, NA = negative affect, SD = standard deviation.
CHAPTER VII138
Figure 1a and 1b. Direct Association between Physical Activity and Positive (1a) and Negative (1b) Affect for the Depressed (red bars) and Non-Depressed (blue bars) Individuals.
EE = energy expenditure. * significant at the 0.05 level. Values on y-column depict correlation coefficients (r).
5 4 16 1 6 12 9 17
815
7
113 13 19 2
1018
Participant Number
1b. EE - Negative Affect
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4
r
12 85 19 2 16 1 9 17
Participant Number
1a. EE - Positive Affect
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4
r
6143182013117154
139PHYSICAL ACTIVITY AND MOOD IN DAILY LIFE
Lagged Αssociation between Physical Activity and Positive, and Negative Affect
The lagged associations between previous values of EE (i.e., the joint effects of lag 1, lag 2,
and lag 3) and current values of positive affect varied greatly between individuals (Figure
2a), without any apparent differences between the groups (depressed versus non-de-
pressed). Seven participants showed a positive lagged association between EE and posi-
tive affect, but only for two (one depressed and one non-depressed individual) this effect
was significant. For the remaining 12 participants a negative lagged association of EE on
subsequent positive affect was observed, and for four individuals (all non-depressed) this
association was significant. Mixed effects (i.e., both positive and negative associations on
different lags) were present in two of them (participant 15 and 16). Figure 2b shows the
results for EE and negative affect. The majority of participants (eleven compared to seven
participants) showed a negative lagged association between EE and subsequent negative
affect, which was significant in three participants (one depressed and two non-depressed).
Overall, the effect sizes of the lagged associations between EE and positive and negative
affect were small.
Lagged Association between Positive and Negative Affect and Physical Activity
With regard to the lagged association between positive affect and subsequent EE (Figure
3a), the majority of the participants (n = 15) showed a positive association between positive
affect and subsequent EE. For two participants (both depressed) this association was sig-
nificant. Four participants (two depressed and two non-depressed) showed a non-significant
negative association. The majority of participants also showed a positive lagged association
between negative affect and subsequent EE (eleven compared to six participants) (see Fig-
ure 3b). For only one (depressed) participant this association reached statistical significance
(p<0.05). Overall, the effect sizes of the lagged associations between positive and negative
affect and EE were small.
CHAPTER VII140
EE = energy expenditure; * significant at the 0.05 level; ~significant association with mixed sign of coefficients at different lags. Values on y-column depict standardized coefficients (beta).
16 9 12 7 19 31
11 15 134 8
6 2 10 518 17
Participant Number
2b. EE (lagged) - Negative Affect
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
Bet
aFigure 2a and 2b. Lagged Associations between Physical Activity and Positive (2a) and Negative (2b) Affect in the Depressed (red bars) and Non-Depressed (blue bars) Individuals.
14 17 11 9 6 133
1912
2 157 8
5 16 18 20 1 4
Participant Number
2a. EE (lagged) - Positive Affect
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
Bet
a
141PHYSICAL ACTIVITY AND MOOD IN DAILY LIFE
Figure 3a and 3b. Lagged Associations between Positive (3a) and Negative (3b) Affect and Physical Activity in the Depressed (red bars) and Non-Depressed (blue bars) Individuals.
11 12 18 4 16 21714156191398
317 5 20
Participant Number
3a. Positive Affect (lagged) - EE
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
Bet
a
EE = energy expenditure. * significant at the 0.05 level. Values on y-column depict standardized coefficients (beta).
3b. Negative Affect (lagged) - EE
5 3 19 12 4 2161876110
13 1517 9 8 11
Participant Number
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
Bet
a
CHAPTER VII142
DISCUSSION
To our knowledge, this is the first study to use a within-subject time-series approach to
investigate the dynamic relationship between physical activity and affective states in de-
pressed and non-depressed individuals. The associations found were not consistent across
individuals, ranging from negative to positive, bidirectional or not, and with varying effect
sizes. An exception was the direct association between physical activity and positive affect,
which was positive in nearly all individuals, albeit with varying effect sizes.
A positive relationship between physical activity and positive affect has also been found in
the majority of group-based EMA studies (for example see: Giacobbi, Hausenblas, & Frye,
2005 and Wichers et al., 2012). However, the delayed association between physical activity
and positive affect was not consistently found in the current study; in some individuals it
was positive, while in others it was negative. Furthermore, the strength of the relationship
was small to negligible for most individuals. Thus, the results suggest that physical activity
enhances positive affect directly while only in some individuals there is also a delayed effect.
The increase in positive affect following physical activity did not seem to differ between de-
pressed and non-depressed individuals, although any potential differences between the two
groups were not evaluated statistically due to the small sample size. A recent EMA study
comparing patients with a depressive disorder and non-depressed controls (Mata et al.,
2012) showed similar improvements in positive affect after physical activity in both groups.
Hence, physical activity seems to (directly) improve positive affect in individuals regardless
of their depression status.
While the direct association between physical activity and positive affect was quite consis-
tent across individuals, this was not the case for negative affect. Both positive and nega-
tive associations between physical activity and negative affect were observed, though we
found the strongest effect sizes for a negative association. A similar pattern emerged for the
lagged association between physical activity and subsequent negative affect. So, in some
individuals there seems to be a negative effect of physical activity on negative affect, but in
most individuals this relationship is weak or non-existent. When aggregating data of indi-
viduals with opposing effects of physical activity on negative affect, the group-level effect
will be small and inconsistent across different study samples. Hence, individual differences
may explain inconsistent results in EMA studies regarding the relationship between physical
activity and negative affect (Kanning et al., 2013).
Observational (between-subject) studies investigating the direction of the relationship
between physical activity and depressive symptoms have found a reciprocal relationship
between the two variables under investigation (Jerstad, Boutelle, Ness, & Stice, 2010;
143PHYSICAL ACTIVITY AND MOOD IN DAILY LIFE
Lindwall, Larsman, & Hagger, 2011; Stavrakakis, de Jonge, Ormel, & Oldehinkel, 2012). We
extended these findings by showing that the reciprocal relationship differed between indi-
viduals. In some individuals increases in positive mood were followed by increases in their
physical activity levels, while in others it was the other way around. Interestingly, in only a
few individuals there appeared to be a noticeable bidirectional relationship. If these data
were aggregated, we might have concluded that there was a bidirectional relationship at the
group level, while in most individuals the relationship was unidirectional, albeit in different
directions.
The effect sizes in the present study were generally small, but considering the non-experi-
mental daily-life design and the short intervals between the measurements this could have
been expected (Wichers et al., 2012). Nonetheless, small effects can accumulate over time
and become larger, because of the way these changes are disseminated through the sys-
tem and mutually enhanced through possible feedback loops. This can especially be true
if physical activity becomes a daily habit, for instance through regular sport participation,
which could result in systematic long-term improvements in positive mood. Future research
using intensive time-series designs should explore this putative accumulating effect in more
detail, in order to elucidate whether the short-term benefits of physical activity on mood can
accumulate into larger benefits over time, and if so, in which individuals.
Clinical Relevance
Two findings from the current study might be of interest to clinicians. First, our results sug-
gest that physical activity benefits positive affect in the short-term. A recent meta-analysis
of randomized controlled studies showed evidence of a short-term benefit of physical activ-
ity on depression (Krogh et al., 2010). Our results support and extend this finding by showing
that these short-term benefits might be acting through an elevation of positive mood rather
than through a decrease in negative mood for the majority of individuals. The mechanisms
behind this short-term benefit of physical activity on positive affect are currently not under-
stood. Some hypotheses have been proposed implicating the endorphin system (Dishman &
O’Connor, 2009; Thoren, Floras, Hoffmann, & Seals, 1990) and increases in self-efficacy
(Bodin & Martinsen, 2004), but future research is needed to explore these in more depth.
Second, in some individuals high levels of positive or negative affect were associated with
increases in physical activity in the current study. Feeling good might increase the level of
physical activity engagement, but apparently also feeling bad might have such an effect
in some individuals. This concurs with the suggestion that physical activity might be per-
ceived as an efficient strategy to repair negative mood, i.e., ‘walking off’ negative affectivity
(Clark & Isen, 1982). Our study cannot address this hypothesis properly, but suggests inter-
CHAPTER VII144
individual differences in the way affect influences subsequent activity levels that need to
be taken into account.
Strengths & Limitations
This study has several notable strengths. First, the time-series design allows the investiga-
tion of the associations at the individual level, thanks to the multiple repeated measure-
ments over time, and also allows for the exploration of temporal patterns and the direction
of the relationship. Second, although obtaining multiple repeated measurements can be
troublesome, time-consuming, and demanding, especially for participants with depression,
the compliance of the participants did not seem to deteriorate significantly in this study.
There was no indication that the participants were unable or found it difficult to comply with
the study protocol, and any systematic missing values were due to technical problems with
the accelerometers. Even in these cases the number of obtained measurements per indi-
vidual (>60) was still adequate to perform time-series analysis.
The results from this study should be interpreted with some caution due to several limita-
tions. First, the generalizability of within-subject studies to the population at large is lim-
ited, and this is the trade-off that comes with the increased specificity of within-subjects
time-series studies. Therefore, a combination of within- and between-subject approaches
is warranted in order to improve our understanding of the complex relationship between
physical activity and daily mood. Second, the models applied in this study assume that the
relationship between physical activity and affect is linear. It is possible that the relationship
between the two is curvilinear and a specific threshold (i.e., frequency, duration or inten-
sity) of physical activity is needed to benefit individuals with regards to their affect, while
exceeding this threshold might result in negative consequences. Another potential limita-
tion is that the strength of an association depends on the degree of variability in the data.
Therefore, it is possible that for some individuals the variability between the measures was
not large enough to detect an association. Finally, we considered the temporal ordering of
the contemporaneous associations to be from physical activity to affect, given the fact that
physical activity covered each time period as a whole and affect was assessed at the end of
each time period. However, it is possible that reverse effects (affect to physical activity) were
acting here as well, which might have introduced bias in our interpretation. However, due to
the design of the study we could not assess this eventuality further.
145PHYSICAL ACTIVITY AND MOOD IN DAILY LIFE
CONCLUSION
We explored inter-individual differences in the magnitude and direction of the relationship
between physical activity and positive and negative affect using a within-subjects time-se-
ries design. Our findings provided evidence that the strength and direction of the relationship
between physical activity and affect differs across individuals. Future research investigating
this dynamic relationship could try to identify subgroups of individuals that might benefit
most from physical activity. These findings could be informative for both research and clini-
cal care intended to improve the mood of depressed individuals.
CHAPTER VII146
REFERENCES
Beck, A. T., Erbaugh, J., Ward, C. H., Mock, J., & Mendelsohn, M. (1961). An inventory for measuring depres-sion. Archives of General Psychiatry, 4(6), 561-71.
Biddle, S. J. H., & Asare, M. (2011). Physical activity and mental health in children and adolescents: A review of reviews. British Journal of Sports Medicine, 45(11), 886-895.
Bodin, T., & Martinsen, E. W. (2004). Mood and self-efficacy during acute exercise in clinical depression. A randomized, controlled study. Journal of Sport & Exercise Psychology, 26(4), 623-633.
Brandt, P. T., & Williams, J. T. (2007). Multiple time series models. Sage, Inc.
Bylsma, L. M., Taylor-Clift, A., & Rottenberg, J. (2011). Emotional reactivity to daily events in major and minor depression. Journal of Abnormal Psychology, 120(1), 155-167.
Chalder, M., Wiles, N. J., Campbell, J., Hollinghurst, S. P., Searle, A., Haase, A. M., et al. (2012). A pragmatic randomised controlled trial to evaluate the cost-effectiveness of a physical activity intervention as a treatment for depression: The treating depression with physical activity (TREAD) trial. Health Technol-ogy Assessment, 16(10), 1-+.
Clark, M. S., & Isen, A. M. (1982). Toward understanding the relationship between feeling states and social behavior. In A. H. Hastorf, & A. M. Isen (Eds.), Cognitive social psychology (pp. 73-108). New York: Elsevier.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155-159.
Cooney, G. M., Dwan, K., Greig, C. A., Lawlor, D. A., Rimer, J., Waugh, F. R., et al. (2013). Exercise for depres-sion. Cochrane Database of Systematic Reviews, (9), CD004366.
Dishman, R. K., & O’Connor, P. J. (2009). Lessons in exercise neurobiology: The case of endorphins. Mental Health and Physical Activity, 1(2), 4-9.
Dunn, A. L., Trivedi, M. H., Kampert, J. B., Clark, C. G., & Chambliss, H. O. (2005). Exercise treatment for de-pression: Efficacy and dose response. American Journal of Preventive Medicine, 28(1), 1-8.
Dunn, A. L., Trivedi, M. H., & O’Neal, H. A. (2001). Physical activity dose-response effects on outcomes of de-pression and anxiety. Medicine and Science in Sports and Exercise, 33(6 Suppl), S587-97; discussion 609-10.
Dunton, G. F., Liao, Y., Intille, S., Wolch, J., & Pentz, M. A. (2011). Physical and social contextual influences on children’s leisure-time physical activity: An ecological momentary assessment study. Journal of Physi-cal Activity & Health, 8, S103-S108.
Farmer, M. E., Locke, B. Z., Moscicki, E. K., Dannenberg, A. L., Larson, D. B., & Radloff, L. S. (1988). Physical activity and depressive symptoms: The NHANES I epidemiologic follow-up study. American Journal of Epidemiology, 128(6), 1340-1351.
Field, A. (2009). Discovering statistics using SPSS (3rd ed.). SAGE publications Ltd.
Giacobbi, P. R., Hausenblas, H. A., & Frye, N. (2005). A naturalistic assessment of the relationship between personality, daily life events, leisure-time exercise, and mood. Psychology of Sport and Exercise, 6(1), 67-81.
Granger, C. W. J. (1969). Investigating causal relations by econometric models and cross-spectral methods. Econometrica, 37(3), 424-38.
Hamaker, E. L. (2012). Why researchers should think “within-person”. In M. Mehl, & T. Conner (Eds.), Hand-book of research methods for studying daily life (pp. 43-61). New York: The Guildford Press.
Hamaker, E. L., Dolan, C. V., & Molenaar, P. C. M. (2005). Statistical modeling of the individual: Rationale and application of multivariate stationary time series analysis. Multivariate Behavioral Research, 40(2), 207-233.
Heil, D. P. (2006). Predicting activity energy expenditure using the actical (R) activity monitor. Research Quarterly for Exercise and Sport, 77(1), 64-80.
Jerstad, S. J., Boutelle, K. N., Ness, K. K., & Stice, E. (2010). Prospective reciprocal relations between physical activity and depression in female adolescents. Journal of Consulting and Clinical Psychology, 78(2), 268-272.
147PHYSICAL ACTIVITY AND MOOD IN DAILY LIFE
Johnson, K. E., & Taliaferro, L. A. (2011). Relationships between physical activity and depressive symptoms among middle and older adolescents: A review of the research literature. Journal for Specialists in Pedi-atric Nursing: JSPN, 16(4), 235-251.
Kanning, M. K., Ebner-Priemer, U. W., & Schlicht, W. M. (2013). How to investigate within-subject associa-tions between physical activity and momentary affective states in everyday life: A position statement based on a literature overview. Frontiers in Psychology, 4, 187.
Krogh, J., Nordentoft, M., Sterne, J. A., & Lawlor, D. A. (2010). The effect of exercise in clinically depressed adults: Systematic review and meta-analysis of randomized controlled trials. The Journal of Clinical Psychiatry, 72(4), 529-538.
LePage, M. L., & Crowther, J. H. (2010). The effects of exercise on body satisfaction and affect. Body Image, 7(2), 124-130.
Lindwall, M., Larsman, P., & Hagger, M. S. (2011). The reciprocal relationship between physical activity and depression in older European adults: A prospective cross-lagged panel design using SHARE data. Health Psychology, 30(4), 453-462.
Lütkepohl, H. (2007). New introduction to multiple time series analysis. Springer.
Mata, J., Thompson, R. J., Jaeggi, S. M., Buschkuehl, M., Jonides, J., & Gotlib, I. H. (2012). Walk on the bright side: Physical activity and affect in major depressive disorder. Journal of Abnormal Psychology, 121(2), 297-308.
Molenaar, P. C. M., & Campbell, C. G. (2009). The new person-specific paradigm in psychology. Current Direc-tions in Psychological Science, 18(2), 112-117.
Myin-Germeys, I., Birchwood, M., & Kwapil, T. (2011). From environment to therapy in psychosis: A real-world momentary assessment approach. Schizophrenia Bulletin, 37(2), 244-247.
Roenneberg, T., Kuehnle, T., Juda, M., Kantermann, T., Allebrandt, K., Gordijn, M., et al. (2007). Epidemiology of the human circadian clock. Sleep Medicine Reviews, 11(6), 429-438.
Rosmalen, J. G. M., Wenting, A. M., Roest, A. M., de Jonge, P., & Bos, E. H. (2012). Revealing causal hetero-geneity using time series analysis of ambulatory assessments: Application to the association between depression and physical activity after myocardial infarction. Psychosomatic Medicine, 74(4), 377-386.
Salmon, P. (2001). Effects of physical exercise on anxiety, depression, and sensitivity to stress: A unifying theory. Clinical Psychology Review, 21(1), 33-61.
Stavrakakis, N., de Jonge, P., Ormel, J., & Oldehinkel, A. J. (2012). Bidirectional prospective associations between physical activity and depressive symptoms. The TRAILS study. Journal of Adolescent Health, 50(5), 503-508.
Strawbridge, W. J., Deleger, S., Roberts, R. E., & Kaplan, G. A. (2002). Physical activity reduces the risk of subsequent depression for older adults. American Journal of Epidemiology, 156(4), 328-334.
Ströhle, A., Hofler, M., Pfister, H., Muller, A. G., Hoyer, J., Wittchen, H. U., et al. (2007). Physical activity and prevalence and incidence of mental disorders in adolescents and young adults. Psychological Medicine, 37(11), 1657-1666.
Ströhle, A. (2009). Physical activity, exercise, depression and anxiety disorders. Journal of Neural Trans-mission, 116(6), 777-784.
Thoren, P., Floras, J. S., Hoffmann, P., & Seals, D. R. (1990). Endorphins and exercise - physiological-mecha-nisms and clinical implications. Medicine and Science in Sports and Exercise, 22(4), 417-428.
Wichers, M., Peeters, F., Rutten, B. P. F., Jacobs, N., Derom, C., Thiery, E., et al. (2012). A time-lagged mo-mentary assessment study on daily life physical activity and affect. Health Psychology, 31(2), 135-144.
World Health Organization (1990). Composite international diagnostic interview (CIDI), ver. 1.0. Geneva: WHO.
Summary and General Discussion
“The great tragedy of science – the slaying of a beautiful hypothesis by an ugly fact”
Thomas Henry Huxley, in Collected Essays, Vol. 8, 1870
CHAPTER
151DISCUSSION
Summary and General Discussion
Overview of the Research Objectives and Key Results
Physical activity (PA) has played a very important role in human evolution. Humans living in
modern societies are relatively inactive compared to our early ancestors and modern hunter-
gatherers, which clearly has led to health problems. However, the association between being
physically active and developing depressive symptoms is not that clear. The overall aim of
this thesis was to investigate the nature of the relationship between PA and depressive
symptoms and to provide a better understanding of this relationship.
The first part of this thesis (chapters II and III) investigated the strength and direction of
the relationship between PA and depressive symptoms in adolescents. The data used came
from the TRacking Adolescents’ Individual Lives Survey (TRAILS). More specifically, the re-
lationship between PA and depressive symptoms was investigated prospectively (over three
measurement waves spanning approximately 6 years) using Structural Equation Modeling
(SEM) in chapter II. The results showed a weak bidirectional relationship between the two
factors. This bidirectional relationship was observed for cognitive but not for somatic symp-
toms of depression. Adolescents who were more physically active were less likely to experi-
ence depressive symptoms over time, and vice versa, adolescents with more depressive
symptoms were less likely to be physically active over time.
The study described in chapter III focused on the protective hypothesis of PA, by investigat-
ing its potential to prevent the onset of depression. Using a sample of adolescents, followed
up for approximately 5 years, this study found no evidence that PA protected against the on-
set of a first major depressive episode in early adulthood in either boys or girls. Adolescents
who were physically active at an early age were, regardless of the frequency, duration and
intensity of the activity, equally likely to develop a first episode of depression later in life as
their more sedentary counterparts.
The overall aim of the second part of this thesis was to explore possible genetic (chapter IV)
and psychosocial (chapter V and VI) modifiers of the relationship between PA and depressive
symptoms. The samples used were derived from TRAILS as well.
In chapter IV, the role of so-called plasticity genes in the prospective reciprocal relation-
ship between PA and depressive symptoms was investigated. Specific alleles of these genes
have been suggested to confer a susceptibility to environmental influences. This susceptibil-
ity has been hypothesized to increase with increasing numbers of plasticity alleles (Belsky
& Beaver, 2011). Therefore, a cumulative plasticity index was created, in order to investigate
its putative modifying effects on the relationship between PA and depressive symptoms.
The individual polymorphisms were investigated as well. Overall, there was no evidence that
CHAPTER VIII152
either the cumulative plasticity gene index or the individual polymorphisms modified the
bidirectional prospective association between PA and depressive symptoms. Adolescents
with genotypes assumed to reflect high plasticity did not benefit more from PA with regards
to their depressive symptoms and did not exhibit more depressive symptoms after being
physically inactive than individuals without these susceptibility alleles.
The study in chapter V investigated the effects of sport participation from an evolutionary
perspective. We explored whether participating in competitive sports was inversely related to
baseline depressive symptoms and changes in depressive symptoms over time. In addition, we
examined moderating effects of peer-perceived sport competence and gender. Overall, sport
participation was inversely associated with current depressive symptoms, but not with symp-
tom changes over time. The association between sport participation and current depressive
symptoms was largely explained by sport competence. Finally, there was no evidence that
sport competence and gender moderated the relationship between sport participation and de-
pressive symptoms: boys who participated in sports and were considered to be highly athleti-
cally competent did not exhibit fewer depressive symptoms than girls and less competent boys.
The study in chapter VI explored the effects of classroom positioning according to birth date
in multiple domains of psychological development and other outcomes, including baseline
depressive symptoms, depressive symptom changes during adolescence, and teacher-rated
sport competence. In a large sample of adolescents followed for approximately 6 years, sub-
stantial relative-age effects were observed in relation to school progress: relatively younger
adolescents were four times more likely to repeat a grade, while relatively older adolescents
were up to 20 times more likely to skip a grade. In the subgroup that repeated a grade,
the relatively young adolescents reported fewer depressive symptoms than their relatively
older peers. However, the relative-age effects did not influence the evaluation of sport com-
petence in either the normative group or the subgroup that repeated a grade.
The final part of this thesis (chapter VII)30 examined temporal patterns in the association
between PA and affect, using a within-subject approach (time-series analysis), with the aim
to provide a better understanding of the strength and direction of the relationship between
PA and affect within individuals31. A sample of depressed and non-depressed adults from the
Mood and Movement in Daily Life (MOOVD) study was used. The results showed inter-indi-
vidual differences in the strength and the sign of the relationships between PA and affect. An
exception was formed by the correlations between concurrent PA and positive affect, which
were positive in nearly all individuals. This suggests a consistent pattern of enhancement of
positive affect during and immediately after PA, but no consistent improvement of negative
affect. The longer-term (delayed) effects of PA on affect varied greatly across individuals,
regardless of their depression status.
153DISCUSSION
The Direction and Strength of the Relationship between PA and Depressive Symptoms
In this section, the direction of the relationship between PA and depressive symptoms will
be discussed, followed by some considerations on the strength of the association, and how
the findings from the studies of this thesis relate to previous research.
Direction
Chapter II describes a weak bidirectional relationship between PA and depressive symptoms
in adolescents. Thus, physically active adolescents were less depressed than sedentary
adolescents and adolescents with many depressive symptoms were more sedentary than
those with few depressive symptoms, both cross-sectionally and over time. This finding is
consistent with the results from previous studies in adolescent girls (Jerstad, Boutelle, Ness,
& Stice, 2010) and older adults (Lindwall et al., 2011)32. Birkeland et al. (2009) did not find
any evidence of a bidirectional relationship in a longitudinal study of adolescents between
the ages of 13 and 23. Although they observed a relationship between PA and depression at
age 13, changes in PA did not predict changes in depressive symptoms and vice versa. The
reasons for this discrepancy are not quite clear33. Nonetheless, the findings from chapter II
support both the protective and the inhibition hypotheses previously proposed. Bidirectional
relationships between PA and positive and negative affect were also found in the MOOVD
study described in chapter VII, but with substantial individual differences. Some individuals
exercised more when they felt better, which resulted in feeling even better, while for others
this spiral was reversed or not the case at all. Individual differences in the direction of the
relationship between PA and depressive symptoms are of great interest to both researchers
and clinicians and should be examined in more detail, since most observational studies ag-
gregate results across individuals and ignore differences among them systematically.
A number of mechanisms have been proposed to explain the putative benefits of PA on de-
pressive symptoms (protective hypothesis). The predominant models include biological and
psychosocial factors that are altered through PA and in turn influence the trajectory of de-
pressive symptoms. It is important to point out that the proposed biological and psychoso-
cial mechanisms are not in any way mutually exclusive.
The main biological mechanisms involve neuromodulators such as serotonin (5-HT; Dishman
et al., 2006a), dopamine (DA; Bliss & Ailion, 1971), adrenaline and noradrenaline (NA; Dish-
man et al., 2006a; Helmich et al., 2010), endorphins (Boecker et al., 2008), and neurotrophins
such as the Brain Derived Neurotrophic Factor (BDNF; Ploughman, 2008). Levels of these
biomarkers are altered during and after exercise, which in turn triggers a cascade of changes
in brain functioning that might explain the alleviation of depressive symptoms (Dishman
et al., 2006a; Helmich et al., 2010; Struder & Weicker, 2001; Toups et al., 2011; van Praag,
CHAPTER VIII154
Christie, Sejnowski, & Gage, 1999; van Praag, Kempermann, & Gage, 1999; van Praag, 2008).
For example, the monoamine hypothesis of depression postulates that imbalances in 5-HT
and DA34 underlie features and symptoms of depression (Lapin & Oxenkrug, 1969; van Praag
& Korf, 1970). Although this hypothesis has been questioned recently (Delgado, 2000; Leo &
Lacasse, 2008; Luscher, Shen, & Sahir, 2011 but also see Cowen, 2008), evidence from imag-
ing studies supports a dysfunction in 5-HT1A receptors in depressed individuals (Drevets
et al., 2007), and suggests that this dysfunction is not merely an artifact of antidepres-
sant medication (Hirvonen et al., 2008). Likewise, decreased DA35 activity has been linked
to motoric and cognitive symptoms of depression (Nestler & Carlezon, 2006; Willner, 1995),
but also to motivational problems and anhedonia in patients with major depressive disorder
(MDD; Dunlop & Nemeroff, 2007). Because of interactions between the DA and 5-HT sys-
tems (Dremencov et al., 2004; Sasaki-Adams & Kelley, 2001), dysfunctions in one of these
systems can initiate secondary changes in others, such as the GABA or adrenaline systems
(Luscher et al., 2011; Thase, 2009).
The psychosocial mechanisms that have been proposed to explain the association between
PA and depressive symptoms are based on self-determination and self-efficacy theories
(Biddle & Mutrie, 2008; Deci & Ryan, 1985; Deci & Flaste, 1995; Sallis & Owen, 1999; Salmon,
2001). The rationale is that by engaging in PA, individuals become more confident in their
abilities, and feel more in control. The increase in self-esteem induces an improvement in
mood (Dishman et al., 2006b). There is also evidence that activities in which individuals form
social networks provide a feeling of belonging and empowerment, which might result in alle-
viation of depressive symptoms as well (McGale, McArdle, & Gaffney, 2011). Finally, a recent
review (Thompson-Coon et al., 2011) suggests that particularly outdoor activities benefit
mental health. However, this area of research is relatively underdeveloped and although the
initial results seem promising, future research is needed to explain the particular impact of
outdoor activities on individuals’ mood and depressive symptoms.
The inhibition hypothesis posits that depressed mood, at least to some degree, demoti-
vates the individual from being physically active (Birkeland et al., 2009). Even sub-thresh-
old depressive symptoms, i.e., symptoms that do not reach the threshold for clinical di-
agnosis, have been associated with psychosocial and functional impairments (Georgiades,
Lewinsohn, Monroe, & Seeley, 2006; Lewinsohn, Solomon, Seeley, & Zeiss, 2000). These
symptoms may trigger a negative feedback loop in which it is difficult for individuals to be
energetic and motivated enough to exercise36 (Goodwin, 2003). The majority of studies to
date have focused on the benefits of PA on depressive symptoms, while only very few have
focused on the inhibition hypothesis (Roshanaei-Moghaddam, Katon, & Russo, 2009). The
findings of chapter II and VII indicate that there is a need to acknowledge the inhibition hy-
pothesis, at least for certain individuals.
155DISCUSSION
Strength of the Αssociation
Overall, whenever a statistically significant association between PA and depressive symp-
toms was observed in the studies described in this thesis, the effects were weak. PA ex-
plained approximately 1 to 2% of the variance in depressive symptoms, and sport participa-
tion, regardless of the competitive element of each sport, slightly more, approximately 3
to 5%. Besides methodological limitations, which will be discussed later, there are several
possible reasons why the effects were weak.
First, the weak effects might indicate that the relationship between PA and depressive
symptoms is not likely to be causal. Although a plethora of hypotheses have been proposed
explaining the relationship between PA and depressive symptoms, a genetic study in a large
sample of monozygotic and dizygotic twins by de Moor et al. (2008) has raised doubts on the
causality of the relationship. They observed cross-sectional and prospective inverse asso-
ciations between leisure-time PA and anxiety and depressive symptoms, with small effect
sizes. However, monozygotic twins that exercised a lot did not report fewer anxiety and de-
pressive symptoms than their more sedentary co-twins (de Moor et al., 2008). Moreover, in
the prospective analyses, changes in PA were not related to changes in anxiety and depres-
sive symptoms, which contradicts one of the predictions made by the causal hypothesis37.
This indicates that the association between PA and depressive symptoms might be caused
by the effects of common genes acting independently, i.e., influencing the propensity to ex-
ercise while at the same time also protecting against symptoms of depression, but without
a causal link between the two.
If a long-term causal relationship exists, it may be expected that PA will exhibit clear ben-
efits in preventing the onset of depression. This has been shown in a recent review of pro-
spective studies in adults, adolescents and children, where PA seemed to prevent the devel-
opment of depression (Mammen & Faulkner, 2013). However, the findings from chapter III do
not indicate such benefits. This is contrary to the findings of a previous study (Jacka et al.,
2011), based on a large Australian study in over 2000 adults (median age 56 years), in which
PA during childhood and depression in later life were retrospectively measured. In this study,
low levels of self-reported childhood PA increased the odds of reporting depression in adult-
hood by 35%. Even though these findings might be affected by retrospective bias, it is still
possible that PA prevents depression onset in middle to late adulthood where the average
onset of MDD is estimated to be between the ages of 20 and 3038 (Kessler & Wang, 2009).
A second reason for the weak associations between PA and depressive symptoms could be
that the relationship between PA and depressive symptoms is modified by other factors.
For example, puberty, school transitions and other important life changes, biological fac-
tors, social factors, sleep and diet, and personality traits may all affect both one’s mood and
CHAPTER VIII156
engagement in PA, and perhaps also moderate the relationship between the two. Chapters
IV, V and VI investigated some of these factors in more detail and their impact is discussed
in the next section.
Another possible reason for the weak associations is that PA may have a differential relation-
ship with different clusters of depressive symptoms. Chapter II showed that PA was inversely
associated with the cognitive symptom but not with the somatic symptom subscale of de-
pression39. This difference might have been due to the low reliability of the somatic symptom
subscale of depressive symptoms40. Possibly, individual somatic symptoms are differentially
related to PA, for instance, PA could be inversely related to sleep problems and positively
related to energy levels. Currently, there is a lack of studies examining this possibility.
Finally, the weak effects might reflect that PA is beneficial in reducing depressive symptoms
for some individuals but not for all41. In large-scale epidemiological studies like TRAILS,
information is aggregated from groups of individuals (between-subject design) and group av-
erages of the association between PA and depressive symptoms are presented42. Even large
effects at the group level, however, may not apply to fluctuations within individuals, since
a between-subject correlation is not the same as a within-subject correlation (Hamaker,
Dolan, & Molenaar, 2005; Molenaar & Campbell, 2009; Rosmalen, Wenting, Roest, de Jonge,
& Bos, 2012). This between-subject variability needs to be taken into account when studying
the relationship between PA and depressive symptoms, in order to avoid unjustified infer-
ences of group-level findings with regard to temporal patterns within individuals (de Jonge &
Bos, 2013; Hamaker, 2012; Molenaar & Campbell, 2009). To learn how PA and affect influence
each other over time within persons, studies using time-series analysis43 with a large number
of repeated assessments per person offer a great research tool. The results of the study in
chapter VII suggest that PA had a consistent positive short-term influence on positive affect
for the majority of participants, but not on negative affect. This finding is in line with the
majority of ecological momentary assessment (EMA) studies, which have generally shown
that the relationship between PA and positive affect is more consistent than the relationship
between PA and negative affect (Kanning et al., 2013). However, even though the majority of
individuals showed a consistent positive relationship between PA and positive affect, only
for a few individuals (5 out of 19) was this relationship significant and of at least moderate
strength. This indicates that, even though for some individuals the relationship was moder-
ate, a group aggregation approach would most likely dilute the effects and be weak over all
participants. This was even more pronounced in the relationship between PA and negative
affect, where the associations were very heterogeneous among individuals, that is, some
participants showed a negative association, a few others a positive one44. Therefore, the
results from the study in chapter VII indicate that the weak effects found in prior chapters are
probably due to individual heterogeneity in the relationship between PA and affect. Future
157DISCUSSION
studies should use a combination of between- and within-subject approaches, in order to
identify commonalities between individuals and elucidate why some individuals benefit or
are harmed by PA in relation to their mood.
Genetic and Psychosocial Modifiers
The second part of this thesis (chapters IV, V and VI) explored possible genetic and social modi-
fiers of the PA-depressive symptom relationship. Both will be discussed in the following part.
Genetic Modifiers
Belsky et al. (2009) identified a number of genes that are thought to be involved to suscep-
tibility to environmental influences, and have been labeled plasticity genes. The plasticity
genes hypothesis extends the classic diathesis-stress model45, by stating that individuals
with a specific genotype will not only be at increased risk when exposed to negative life
events and experiences, but also benefit more from positive ones, such as social support46
(Belsky & Pluess, 2009; Belsky et al., 2009; Ellis, Boyce, Belsky, Bakermans-Kranenburg, &
van Ijzendoorn, 2011). These putative plasticity genes, which were investigated in chapter
IV, influence mainly the serotonergic and dopaminergic system, and have been shown to act
in a ‘For better and for worse’ manner consistently in the literature (Belsky & Pluess, 2009;
Belsky et al., 2009). In the study described in chapter IV, there was no evidence to suggest
that either the cumulative plasticity index or the individual polymorphisms modified the bidi-
rectional relationship between PA and depressive symptoms prospectively.
This is contrary to findings from three recent studies. Two studies on the serotonin trans-
porter polymorphism (5-HTTLPR; Rethorst, Landers, Nagoshi, & Ross, 2010 and Rethorst,
Landers, Nagoshi, & Ross, 2011) concluded that carriers of the short allele of 5-HTTLPR had
a larger reduction in depressive symptoms following exercise than carriers of the long allele.
Another study, which concerned the BDNF val/met polymorphism, showed that carriers of
the met allele that exercised more exhibited fewer depressive symptoms than carriers of the
val allele (Mata, Thompson, & Gotlib, 2010). Both were cross-sectional studies conducted
in adults with relatively small sample sizes, and investigated the role of these genes in
the unidirectional effect of PA on depressive symptoms. In contrast, the study in chapter IV
was prospective and investigated the moderating effects of these genes on the bidirectional
relationships between PA and depressive symptoms in a large population sample of adoles-
cents. Therefore, differences in the design and the population sample might account for the
discrepant findings.
Mixed findings concerning the 5-HTTLPR gene and especially the gene-environment inter-
action of the short allele on depression are ubiquitous in the literature (Caspi et al., 2003;
CHAPTER VIII158
Karg, Burmeister, Shedden, & Sen, 2011; Munafo, Durrant, Lewis, & Flint, 2009; Risch et al.,
2009). Therefore, the influence of the short-allele of the 5-HTTLPR as a moderator of the
relationship between PA and depressive symptoms remains inconclusive. The discrepant
findings concerning the BDNF gene are more surprising, considering the large body of re-
search showing that BDNF levels are altered by PA (Toups et al., 2011; van Praag et al., 1999;
van Praag, 2008) and changes in BDNF levels are implicated with depression47 (for a review
see Hashimoto, 2010). It is possible that this discrepant finding is due to methodological
limitations of our study: since the number of participants homozygous to the met allele was
small, we merged this group with the individuals carrying the met/val alleles, which may
have introduced bias in the results.
The BDNF and other plasticity genes have numerous polymorphisms, of which only a small
selection were investigated in our study. Investigation of additional polymorphisms may help
to identify genetically based subgroups of individuals that will benefit from PA with regard to
their depressive symptoms48. However, considering the difficulties and inconsistencies with
candidate genes replications in MDD (Flint & Kendler, 2014), among other disorders, this is
speculative. Depressive disorders are multifactorial and due to the inability to identify clear
candidate genes for these ailments, we cannot exclude the possibility that multiple genes
and their interactions influence their onset and trajectory. This polygenic effect of multiple
loci of genes interacting has been recently investigated in several psychopathologies, in-
cluding MDD, and yielded some promising results (Peyrot et al., 2014; Smoller et al., 2013).
Future research is needed to elucidate the complex interactions between genes and their
influence on behavior, but there are reasons for optimism in this line of research.
Psychosocial Modifiers
Two possible psychosocial modifiers of the relationship between PA and depressive symp-
toms have been explored, namely sport competence as perceived by peer evaluations and
classroom positioning according to the month of birth.
PA , Spor t Par ticipation a nd Spor t Co m petence
The competitive element of sport participation has been argued to have an evolutionary ori-
gin (Lombardo, 2012) for several reasons. First, sports resemble primitive hunting and war-
fare of our ancestors. Second, sports are universal and can be found in almost every human
group throughout history. Finally, athletic competitions play a very influential role in society.
Lombardo argued that the competitive element of sports influences hierarchy formations
in males. This hierarchy formation will in turn affect the chances of males being selected by
females as sexual partners (Lombardo, 2012).
159DISCUSSION
In turn, this hierarchy formation has been linked to depression49. Price et al. (1994) suggest-
ed in their Social Competition Hypothesis that depression might be an adaptive strategy,
similar to hibernation in some animals50. Hence, individuals who are likely to lose in ritual
agonistic competitions may develop depression in order to be protected from further dam-
age and loss of their position within a social group51. In other words, the Social Competition
Hypothesis predicts that lower status individuals are more likely to suffer from depression.
The study in chapter V investigated whether the Social Competition Hypothesis extends to
sport participation and depressive symptoms. Overall, there was a positive association be-
tween adolescents who participate in competitive sports and social status, supporting the
hypothesis that sport participation acts as a hierarchy formation in adolescents52. This is
consistent with other studies showing that sport participation, especially in boys, is impor-
tant for adolescent hierarchy formation (Gill, 1992). Moreover, sport competence was also
related to concurrent depressive symptoms, as was suggested by Price et al. (1994), i.e., rel-
atively incompetent individuals exhibited more depressive symptoms53. Sport competence
explained the observed relationship between sport participation and depressive symptoms,
which indicates that sport competence may have a larger influence on depressive symp-
toms than participation in sports. There was no evidence that sport participation and sport
competence were related to changes in depressive symptoms. Furthermore, boys who were
considered to be highly sport competent did not benefit more from sport participation than
girls or less competent boys.
PA has had an important role in the evolution of humans, thus any beneficial effects of PA
on depressive symptoms could be put into a larger evolutionary theoretical framework. Our
study supported some predictions based on these evolutionary perspectives, but not the
whole postulated relationship between PA, sport participation and depressive symptoms. It
is important to investigate the relationship between PA or sport participation and depres-
sive symptoms from an evolutionary perspective, because through this larger framework
important insights can be achieved. Evolutionary perspectives offer the possibility for new
hypotheses to be tested, which can lead to a better understanding of the relationship be-
tween PA and mental health.
Classroo m Position in g a nd Relative-Age Ef fects
Children are allocated into classrooms according to their date of birth54. This might result
in classrooms encompassing children of up to a year older being in the same grade with
relatively younger children. Previous research has shown relative unfavorable outcomes
in certain domains for the relatively young compared to the relatively old children (see
for example: Bedard & Dhuey, 2006 and Morrow et al., 2012). These favorable or unfavor-
able outcomes might be a result of the evaluation of adults or their peers. Relative older
CHAPTER VIII160
adolescents will be slightly taller, stronger and physically more mature, which would result
in a higher likelihood to be selected by coaches and excel in sports (Gladwell, 2008)55. If being
perceived or being actually competent in sports influences social status (as was shown, to
some extent, in chapter V) then the relative older adolescents might report fewer depressive
symptoms through this attainment of higher status (as shown in chapter V as well). Thus,
the relative-age effects might moderate the relationship between PA and depressive symp-
toms. Although not directly explored here, the possibility that relative-age effects influence
hierarchy formations and so result in better opportunities for the relative older individuals
appears to be a promising line for future research. The findings in chapter VI indicated that
there are no relative-age effects (favorable or unfavorable) in adolescents with a norma-
tive school progress. However, within adolescents who repeated a grade, the relative young
ones reported fewer depressive symptoms than their relatively older peers. These relative-
age effects did not translate into differences in the evaluation of sport competence by teach-
ers, in either group. The possibility that these relative-age effects influence peer perceptions
of sport competence was not considered. Chapter V showed a difference in the perceptions
of sport competence between teachers and peers. Hence, it would be interesting to explore
how peer perceptions might be affected by relative-age differences.
In sum, most of the moderation analyses performed in the studies in this thesis have not
yielded clear factors that might help identify individuals that can benefit or be harmed by PA,
regarding depressive symptomatology. The study in chapter VII, however, showed individual
differences in the relationship between PA and affect, which still suggests that there might
be common factors between individuals that influence whether PA can be beneficial for men-
tal health. Future research using a combination of within and between subject designs might
prove imperative in identifying these common factors. The elucidation of the reasons why
some individuals benefit from PA while others do not, might prove helpful in the develop-
ment of individual-targeted PA interventions that will be effective in treating depression.
Strengths & Limitations
The longitudinal design and the large sample sizes used in most chapters is an important
strength of this thesis. In most cases, sample sizes of more than 1000 adolescents were used,
which provided adequate power to detect relatively small effect sizes. Additionally, we used two
or more measurement waves, covering a relatively long period of time. This facilitates the exclu-
sion of time-invariant unobserved individual differences and informs of the temporal order of
events, thus making the exploration of the direction of the effects under investigation possible.
There are also limitations that need to be acknowledged. In most studies in this thesis,
self-report measures of PA and depressive symptoms were used, with different degrees of
161DISCUSSION
specificity. Self-reports typically depend on the memory of participants, who estimate their
depressive symptoms or PA based on an average week. The affective scale of the Youth Self-
Report, which assesses depressive symptoms corresponding to DSM-IV criteria (Achenbach
& Rescorla, 2001), is an empirically based and well-validated questionnaire with an accept-
able test-retest reliability (Achenbach, Dumenci, & Rescorla, 2003). However, the use of
self-reports to estimate PA has been criticized by researchers (see for example: Johnson &
Taliaferro, 2011), since these might be prone to over or under-estimation of the levels of PA
engagement. Although acceptable test-retest reliability was shown in independent samples
(Booth, Okely, Chey, & Bauman, 2001; Haskell, 2012)56, recent reviews comparing self-re-
ports to more direct measures of PA, such as accelerometers, revealed only a moderate to
weak correlation between the two (Adamo, Prince, Tricco, Connor-Gorber, & Tremblay, 2009;
Prince et al., 2008). For example, Adamo et al. (2009) observed that self-reports underes-
timated light or moderate PA, while vigorous PA was consistently over-estimated, when
compared to more direct measures in children and adolescents. Prince et al. (2008) found
that studies using self-reports generally overestimated the levels of PA participation when
compared to direct measures, especially for women. Another limitation is that, throughout
this thesis, the relationship between PA and depression was modeled using linear regression
techniques. It is possible that PA is beneficial up to a certain point (curvilinear relationship),
while exceeding this threshold results in adverse mental health consequences, as studies on
the negative effects of overtraining on mood have shown (Bresciani et al., 2011; Armstrong &
VanHeest, 2002). In most chapters of this thesis, the effects of duration and intensity of PA
were not considered. It is possible that only individuals who engage at a specific frequency,
duration and intensity of PA experience fewer depressive symptoms (Dunn, Trivedi, Kam-
pert, Clark, & Chambliss, 2005; Goldfield et al., 2011; Sanders, Field, Diego, & Kaplan, 2000).
A study in Chinese adolescents (Tao et al., 2007), for example, showed that only low to mod-
erate PA intensity but not high intensity was related to depressive symptoms. Therefore,
it is possible that the different intensities of PA might have a differential association with
depression. However, in the study in chapter III we examined the dose-response effect of PA
and found no evidence that different frequencies, durations and intensities of activities pro-
tected against an onset of depression. To conclude, it is still possible that the different dura-
tions and intensities of activities differentially relate to depressive symptoms, but based on
the results in this thesis, this seems unlikely.
Clinical Relevance
Throughout this thesis, the effect sizes observed were small. Still, weak effects observed
in population studies (epidemiology) do not necessarily imply that the use of PA or exercise
in a clinical trial cannot alleviate such symptoms. Findings from population studies cannot
CHAPTER VIII162
be used to justify exercise as a treatment without relying on randomized controlled trials
(RCTs), regardless of the strength of the relationship observed in these studies. Even small
effect sizes found in population studies can translate to substantial clinical benefits in cer-
tain individuals. To illustrate this point, Rosenthal showed that prescribing aspirin resulted
in a substantial decrease in heart attacks (85 out of approximately 11000 patients), despite
the estimated effect being weak (r = -.03; Rosenthal, 1990)57.
Concerning the relationship between PA and depression, initially promising results (in both
observational and intervention studies) have been prematurely echoed and exaggerated58,
mainly because of the elegant and intuitively appealing hypothesis that PA might be used
as a treatment for depression. Some support for this hypothesis was observed in the study
in chapter VII59. The results suggested that, although higher levels of PA resulted in lower
negative affect in a few subjects60, there were also individuals in whom high levels of nega-
tive affect were associated with more activity. This may indicate that for some individuals PA
may be an efficient strategy to alleviate negative mood. However, a recent high quality RCT
(Chalder et al., 2012) and meta-analyses of RCTs (Cooney et al., 2013; Mead et al., 2009) have
not found sufficient support for this hypothesis at the group level61.
Finally, the results in chapter VII indicate that PA has a direct effect on elevating positive
mood for the majority of individuals, regardless of whether they are depressed or not, but
the delayed effect of PA on positive mood seems to be much less consistent. The reasons
for a short-term effect of PA on positive affect, and for why this does not carry over to longer
time periods, are currently not understood. Cogent and elegant hypotheses implicating bio-
logical factors such as the endorphin system62 (Dishman & O’Connor, 2009; Thoren, Floras,
Hoffmann, & Seals, 1990) have been proposed, but future research is needed to investigate
these mechanisms in further detail. By identifying the reasons why these benefits do not
carry over the longer term, clinicians might be able to manipulate factors and identify spe-
cific individuals that might be helped by PA in terms of their depressive symptoms. Until a
better understanding is achieved, however, any conclusions about this relationship should
be tentative and conservative until further studies demonstrate substantial clinical benefits
that are specific to PA.
Concluding Remarks
The relationship between PA and depressive symptoms has been investigated extensively
over the last few years. Initial studies bore the promise that PA might reduce depressive
symptoms, which would make it a very elegant and potentially cost-effective treatment
for depression. Major organizations such as the WHO have implemented exercise referral
schemes for the treatment of depression. However, some authors have questioned the cau-
163DISCUSSION
sality of the relationship between PA and depressive symptoms (Birkeland et al., 2009; de
Geus & de Moor, 2008; de Moor et al., 2008), while others have argued that the protective
effects of exercise might have been overestimated (Daley & Jolly, 2012).
This thesis illustrates that the relationship between PA and depressive symptoms is bi-
directional, not robust and more complex than what was previously envisioned. This sug-
gests that PA might not be directly related to mental health but instead represents a small
fraction in a large dynamic network of factors interacting with each other. Additionally, this
thesis demonstrated large inter-individual differences in the relationship between PA, affect
and depressive symptoms, where some individuals benefited from PA, while others did not.
It was also evident that for some individuals depressed mood hindered them from being
physically active, while for others this negative mood motivated them to ‘walk it off’.
Future research should focus on delineating other factors that influence the relationship
between PA and depressive symptoms, in order to accurately identify individuals that might
benefit from PA. This accurate identification of individuals will hopefully help to devise per-
sonal exercise interventions for treating depression.
Notes30 This study has many differences from all other studies included in this thesis and therefore has been in-
cluded in a separate section. However, because this study explored the strength and the direction of the relationship between PA and affect in individuals, the findings will be discussed in parts of the discussion that are most relevant.
31 Affect is thought to be the most elementary form of a conscious feeling and is defined as “a neurophysi-ological state that is consciously accessible as a simple, non-reflective feeling that is an integral blend of hedonic (pleasure-displeasure) and arousal (sleepy-activated) values” (see: Russell, 2003, table 1). A prolonged core affect leads to what is termed mood. Affect is divided into positive affect and negative affect. Long lasting changes on affect and especially increases in negative affect can lead to a depressed mood (Clark & Watson, 1991), and consequently a prolonged depressed mood can lead to further exacer-bation of depressive symptoms and subsequently to the development of depression. This study explored the influence of PA on positive and negative affect in depressed and non-depressed individuals because of the close connection between affect and depressed mood.
32 Lindwall et al. (2011) established that the direction of the relationship was stronger from PA to depres-sive symptoms than the other way around, but the model fit suggested that the relationship was bidi-rectional.
33 Numerous reasons could be behind this discrepancy, but most of these are speculative. For instance, it could be that Birkerland et al.'s results reflected a false negative. Perhaps the protective and inhibi-tion hypotheses are true only in adulthood and not in adolescence. It is also possible that the results of Birkerland et al. were due to a narrow definition of PA. Birkerland et al. focused on leisure-time PA, which, according to some authors, should be considered as exercise and not PA (Caspersen, Powell, & Christen-son, 1985). Finally, the relationship between PA and depressive symptoms might not be a causal one (this point will be elaborated further in later parts of this thesis).
CHAPTER VIII164
34 NA has also been implicated in depression (Thase, 2009).35 It should be noted that most antidepressants do not seem to directly impact DA levels. However, MAOIs,
buproprion and sertraline have been shown to have some effect on dopaminergic neuromodulation in animal models of depression (Thase, 2009; Willner, 1995; Willner, 1997). Moreover, in animals these anti-depressants countermand dysfunctions in DA systems, especially in conditions of chronic stress (Cuadra, Zurita, Gioino, & Molina, 2001; Nestler & Carlezon, 2006). Therefore, DA and its possible effect on depres-sion is of great interest for future research.
36 Similarly, a positive mood may result in higher energy levels and motivation, which can influence engage-ment in PA.
37 The causal hypothesis on the relationship between PA and depressive symptoms, with the protective and inhibition hypotheses as two specifications, predicts that increases in frequency and intensity of PA over time would reduce depressive symptoms over time, and conversely, that decreases in PA levels over time would increase depressive symptoms over time, through a causal chain.
38 With a later peak between the ages of 30 and 40 (see: http://www.health.am/psy/major-depressive-disorder/).
39 Cognitive symptoms subscale includes: lack of interest; feelings of worthlessness; loss of pleasure; cry-ing; self-harm; suicidal ideation; and feelings of guilt and sadness. The somatic symptoms subscale in-cludes: lack of appetite; overtiredness; trouble with sleeping and lack of energy.
40 This finding is counter-intuitive, because it would be expected that PA, if anything, improves sleep qual-ity, increases appetite and (unless overtraining) increases energy levels (Biddle & Mutrie, 2008). Fur-thermore, a recent study has shown that PA reduced somatic symptoms of depression in patients with chronic heart failure (Redwine et al., 2012).
41 Note here that the modifying analyses in this thesis explored this idea further, without conclusive find-ings. The study in chapter VII, however, showed individual differences in the relationship between PA and affect more clearly and these findings will be discussed in this section.
42 Between-subject designs are also called nomothetic approaches, while within-subject designs are also called idiographic approaches. Because the terms ‘nomothetic’ and ‘idiographic’ are not used in the origi-nally intended way (Lamiell, 1998), the terms between-subject and within-subject designs are used in-stead, in order to avoid any further confusion.
43 Which is a within-subject approach measuring the same individual over multiple measurement points. Although this approach cannot compare effects between individuals, it can elucidate the pattern of the relationship in individuals over time.
44 It should be noted here, that these results could be interpreted differently. It could be argued that the relationship between PA and negative affect was rather homogeneous, since the majority of participants did not show an effect of PA on negative affect or vice versa. However, the fact that some individu-als showed positive while others showed negative associations between PA and negative affect, even though a minority, should make it clear that there is some heterogeneity in the association between these two factors.
45 Individuals with specific genotypes (diathesis) might be vulnerable to developing mental health problems after important negative life events (stressors) have occurred.
46 This is referred to as ‘For better and for worse’.47 BDNF is targeted by a variety of antidepressants (Hashimoto, 2010). In MDD and Bipolar Disorder (BPD)
studies, antidepressant medication increased the expression of BDNF mRNA in limbic structures in re-sponse to chronic treatment but not acute treatment (see for example: Hashimoto, 2010). Finally, the time that it takes to increase BDNF expression coincides with the time typically required for antidepres-sants to work in depressed patients (Hashimoto, 2010).
165DISCUSSION
48 Or conversely increase vulnerability to depression that in turn can affect PA engagement levels.49 This hypothesis is consistent with gender differences observed in depression and has been dealt with in
previous publications from the authors (Price, 1988). Briefly, they suggest that females still compete for ranking, but female competition is more inconspicuous and less common than male competition.
50 Notably, not all evolutionary perspectives converge on the idea that depression might be adaptive (Al-len & Badcock, 2006; Nettle, 2004). Some theories propose that personality traits, such as neuroticism, might be adaptive and therefore survived evolution. Individuals, who are on the far side of this affective reactivity distribution will suffer detrimental effects, i.e., develop depression (Nettle, 2004).
51 Throughout the early human evolutionary history, this loss would have likely resulted in physical harm, lower chances of reproduction or death.
52 And as such, supported the prediction made by the evolutionary perspective of sport participation (Lom-bardo, 2012), i.e., more athletically competent individuals will attain higher status.
53 Therefore also supporting one of the predictions of the Social Competition Hypothesis (Price et al., 1994) that lower status individuals will suffer more from depressive symptoms than higher status individuals.
54 There is considerable variation to this rule in different cultures (Bedard & Dhuey, 2006).55 Especially since in childhood and adolescence even small differences in age (a year or less) result in large
differences in physique and abilities.56 Even a single item is considered to provide an accurate representation of an individual’s engagement in
PA (Milton, Clemes, & Bull, 2012).57 This might seem as a rather small proportion of people being protected, but this clinical trial was termi-
nated prematurely, because the evidence that aspirin protected from heart-attacks was so overwhelm-ing that it was deemed unethical to continue giving placebo medication to the control group (see Rosen-thal, 1990).
58 There are intervention studies that show a benefit of PA on depressive symptoms (see for example: Blu-menthal et al., 1999 and Brown, Pearson, Braithwaite, Brown, & Biddle, 2013) but have been criticized for low quality methodological designs including: small samples, short follow-ups, poor randomization, no ad-equate control group and reliance on non-clinical volunteers (Daley & Jolly, 2012). Adding to these problems the possibility of publication bias, it is imperative to design intervention studies of high methodological quality in order to conclusively demonstrate that PA can be used as a treatment for depression.
59 Although not an intervention study.60 And therefore supporting (to an extent) the bidirectional relationship observed in prospective (observa-
tional) studies.61 Also see a review on EMA studies by Kanning et al. (2013).62 Briefly, this theory suggests that PA increases endorphin levels, resulting in what colloquially is termed
‘Runner’s high’. These changes have been shown to last only for a very brief period of time.
CHAPTER VIII166
REFERENCES
Achenbach, T. M., & Rescorla, L. A. (2001). Manual for the ASEBA school-age forms and profiles. Burlington, VT: University of Vermont, Research center for children, youth and families.
Achenbach, T. M., Dumenci, L., & Rescorla, L. A. (2003). DSM-oriented and empirically based approaches to constructing scales from the same item pools. Journal of Clinical Child and Adolescent Psychology, 32(3), 328-340.
Adamo, K. B., Prince, S. A., Tricco, A. C., Connor-Gorber, S., & Tremblay, M. (2009). A comparison of indirect versus direct measures for assessing physical activity in the pediatric population: A systematic review. International Journal of Pediatric Obesity, 4(1), 2-27.
Allen, N. B., & Badcock, P. B. T. (2006). Darwinian models of depression: A review of evolutionary accounts of mood and mood disorders. Progress in Neuropsychopharmacology & Biological Psychiatry, 30(5), 815-826.
Armstrong, L. E., & VanHeest, J. L. (2002). The unknown mechanism of the overtraining syndrome - clues from depression and psychoneuroimmunology. Sports Medicine, 32(3), 185-209.
Bedard, K., & Dhuey, E. (2006). The persistence of early childhood maturity: International evidence of long-run age effects. Quarterly Journal of Economics, 121(4), 1437-1472.
Belsky, J., & Beaver, K. M. (2011). Cumulative-genetic plasticity, parenting and adolescent self-regulation. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 52(5), 619-626.
Belsky, J., Jonassaint, C., Pluess, M., Stanton, M., Brummett, B., & Williams, R. (2009). Vulnerability genes or plasticity genes? Molecular Psychiatry, 14(8), 746-754.
Belsky, J., & Pluess, M. (2009). Beyond diathesis stress: Differential susceptibility to environmental influ-ences. Psychological Bulletin, 135(6), 885-908.
Biddle, S. J. H., & Mutrie, N. (2008). Psychology of physical activity: Determinants, well-being and interven-tions (2nd ed.). New York, USA: Routledge.
Birkeland, M. S., Torsheim, T., & Wold, B. (2009). A longitudinal study of the relationship between leisure-time physical activity and depressed mood among adolescents. Psychology of Sport and Exercise, 10(1), 25-34.
Bliss, E. L., & Ailion, J. (1971). Relationship of stress and activity to brain dopamine and homovanillic acid. Life Science Part 1 Physiology & Pharmacology, 10(20), 1161-9.
Blumenthal, J. A., Babyak, M. A., Moore, K. A., Craighead, W. E., Herman, S., Khatri, P., et al. (1999). Effects of exercise training on older patients with major depression. Archives of Internal Medicine, 159(19), 2349-2356.
Boecker, H., Sprenger, T., Spilker, M. E., Henriksen, G., Koppenhoefer, M., Wagner, K. J., et al. (2008). The run-ner’s high: Opioidergic mechanisms in the human brain. Cerebral Cortex, 18(11), 2523-2531.
Booth, M. L., Okely, A. D., Chey, T., & Bauman, A. (2001). The reliability and validity of the physical activ-ity questions in the WHO health behaviour in schoolchildren (HBSC) survey: A population study. British Journal of Sports Medicine, 35(4), 263-267.
Bresciani, G., Cuevas, M. J., Molinero, O., Almar, M., Suay, F., Salvador, A., et al. (2011). Signs of overload after an intensified training. International Journal of Sports Medicine, 32(5), 338-343.
Brown, H. E., Pearson, N., Braithwaite, R. E., Brown, W. J., & Biddle, S. J. H. (2013). Physical activity inter-ventions and depression in children and adolescents: A systematic review and meta-analysis. Sports Medicine, 43(3), 195-206.
Caspersen, C. J., Powell, K. E., & Christenson, G. M. (1985). Physical-activity, exercise, and physical-fitness: Definitions and distinctions for health-related research. Public Health Reports, 100(2), 126-131.
Caspi, A., Sugden, K., Moffitt, T. E., Taylor, A., Craig, I. W., Harrington, H., et al. (2003). Influence of life stress on depression: Moderation by a polymorphism in the 5-HTT gene. Science, 301(5631), 386-389.
167DISCUSSION
Chalder, M., Wiles, N. J., Campbell, J., Hollinghurst, S. P., Searle, A., Haase, A. M., et al. (2012). A pragmatic randomised controlled trial to evaluate the cost-effectiveness of a physical activity intervention as a treatment for depression: The treating depression with physical activity (TREAD) trial. Health Technol-ogy Assessment, 16(10), 1-+.
Clark, L. A., & Watson, D. (1991). Tripartite model of anxiety and depression - psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 100(3), 316-336.
Cooney, G. M., Dwan, K., Greig, C. A., Lawlor, D. A., Rimer, J., Waugh, F. R., et al. (2013). Exercise for depres-sion. Cochrane Database of Systematic Reviews, (9), CD004366.
Cowen, P. J. (2008). Serotonin and depression: Pathophysiological mechanism or marketing myth? Trends in Pharmacological Sciences, 29(9), 433-436.
Cuadra, G., Zurita, A., Gioino, G., & Molina, V. (2001). Influence of different antidepressant drugs on the effect of chronic variable stress on restraint-induced dopamine release in frontal cortex. Neuropsychopharma-cology, 25(3), 384-394.
Daley, A., & Jolly, K. (2012). Exercise to treat depression. British Medical Journal (Clinical Research Ed.), 344, e3181.
de Geus, E. J. C. & de Moor, M. H. M. (2008). A genetic perspective on the association between exercise and mental health. Mental Health and Physical Activity, 1, 53-61.
de Jonge, P., & Bos, E. H. (2013). Tobacco smoking and depression: Time to move on to a new research para-digm in medicine? BMC Medicine, 11, 138.
de Moor, M. H. M., Boomsma, D. I., Stubbe, J. H., Willemsen, G., & de Geus, E. J. C. (2008). Testing causality in the association between regular exercise and symptoms of anxiety and depression. Archives of General Psychiatry, 65(8), 897-905.
Deci, E. L., & Flaste, R. (1995). Why we do what we do: Understanding self-motivation. New York: Penguin.
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior. New York, USA: Plenum Press.
Delgado, P. L. (2000). Depression: The case for a monoamine deficiency. Journal of Clinical Psychiatry, 61, 7-11.
Dishman, R. K., Berthoud, H. R., Booth, F. W., Cotman, C. W., Edgerton, V. R., Fleshner, M. R., et al. (2006a). Neurobiology of exercise. Obesity, 14(3), 345-356.
Dishman, R. K., Hales, D. P., Pfeiffer, K. A., Felton, G., Saunders, R., Ward, D. S., et al. (2006b). Physical self-concept and self-esteem mediate cross-sectional relations of physical activity and sport participation with depression symptoms among adolescent girls. Health Psychology, 25(3), 396-407.
Dishman, R. K., & O’Connor, P. J. (2009). Lessons in exercise neurobiology: The case of endorphins. Mental Health and Physical Activity, 1(2), 4-9.
Dremencov, E., Gispan-Herman, I., Rosenstein, M., Mendelman, A., Overstreet, D. H., Zohar, J., et al. (2004). The serotonin-dopamine interaction is critical for fast-onset action of antidepressant treatment: In vivo studies in an animal model of depression. Progress in Neuropsychopharmacology & Biological Psychia-try, 28(1), 141-147.
Drevets, W. C., Thase, M. E., Moses-Kolko, E. L., Price, J. S., Frank, E., Kupfer, D. J., et al. (2007). Serotonin-1A receptor imaging in recurrent depression: Replication and literature review. Nuclear Medicine and Biol-ogy, 34(7), 865-877.
Dunlop, B. W., & Nemeroff, C. B. (2007). The role of dopamine in the pathophysiology of depression. Archives of General Psychiatry, 64(3), 327-337.
Dunn, A. L., Trivedi, M. H., Kampert, J. B., Clark, C. G., & Chambliss, H. O. (2005). Exercise treatment for de-pression: Efficacy and dose response. American Journal of Preventive Medicine, 28(1), 1-8.
CHAPTER VIII168
Ellis, B. J., Boyce, W. T., Belsky, J., Bakermans-Kranenburg, M. J., & van Ijzendoorn, M. H. (2011). Differential susceptibility to the environment: An evolutionary–neurodevelopmental theory. Development and Psy-chopathology, 23(1), 7-28.
Flint, J., & Kendler, K. S. (2014). The genetics of major depression. Neuron, 81(3), 484-503. doi: 10.1016/ j.Neuron.2014.01.027.Epub 2014 Feb 5
Georgiades, K., Lewinsohn, P. M., Monroe, S. M., & Seeley, J. R. (2006). Major depressive disorder in adoles-cence: The role of subthreshold symptoms. Journal of the American Academy of Child and Adolescent Psychiatry, 45(8), 936-944.
Gill, D. L. (1992). Gender and sport behaviour. Advances in sport psychology (pp. 143-160). Champaign, IL: Human Kinetics.
Gladwell, M. (2008). Outliers: The story of success. USA: Little, Brown and Company.
Goldfield, G. S., Henderson, K., Buchholz, A., Obeid, N., Nguyen, H., & Flament, M. F. (2011). Physical activity and psychological adjustment in adolescents. Journal of Physical Activity & Health, 8(2), 157-163.
Goodwin, R. D. (2003). Association between physical activity and mental disorders among adults in the United States. Preventive Medicine, 36(6), 698-703.
Hamaker, E. L. (2012). Why researchers should think “within-person”. In M. Mehl, & T. Conner (Eds.), Hand-book of research methods for studying daily life (pp. 43-61). New York: The Guildford Press.
Hamaker, E. L., Dolan, C. V., & Molenaar, P. C. M. (2005). Statistical modeling of the individual: Rationale and application of multivariate stationary time series analysis. Multivariate Behavioral Research, 40(2), 207-233.
Hashimoto, K. (2010). Brain-derived neurotrophic factor as a biomarker for mood disorders: An historical overview and future directions. Psychiatry and Clinical Neurosciences, 64(4), 341-357.
Haskell, W. L. (2012). Physical activity by self-report: A brief history and future issues. Journal of Physical Activity & Health, 9 Suppl 1, S5-10.
Helmich, I., Latini, A., Sigwalt, A., Carta, M. G., Machado, S., Velasques, B., et al. (2010). Neurobiological alterations induced by exercise and their impact on depressive disorders. Clinical Practice and Epidemi-ology in Mental Health: CP & EMH, 6, 115-25.
Hirvonen, J., Karlsson, H., Kajander, J., Lepola, A., Markkula, J., Rasi-Hakala, H., et al. (2008). Decreased brain serotonin 5-HT1A receptor availability in medication-naive patients with major depressive disorder: An in-vivo imaging study using PET and [carbonyl-C-11]WAY-100635. International Journal of Neuropsy-chopharmacology, 11(4), 465-476.
Jacka, F. N., Pasco, J. A., Williams, L. J., Leslie, E. R., Dodd, S., Nicholson, G. C., et al. (2011). Lower levels of physical activity in childhood associated with adult depression. Journal of Science and Medicine in Sport / Sports Medicine Australia, 14(3), 222-226.
Jerstad, S. J., Boutelle, K. N., Ness, K. K., & Stice, E. (2010). Prospective reciprocal relations between physi-cal activity and depression in female adolescents. Journal of Consulting and Clinical Psychology, 78(2), 268-272.
Johnson, K. E., & Taliaferro, L. A. (2011). Relationships between physical activity and depressive symptoms among middle and older adolescents: A review of the research literature. Journal for Specialists in Pedi-atric Nursing: JSPN, 16(4), 235-251.
Kanning, M. K., Ebner-Priemer, U. W., & Schlicht, W. M. (2013). How to investigate within-subject associa-tions between physical activity and momentary affective states in everyday life: A position statement based on a literature overview. Frontiers in Psychology, 4, 187.
Karg, K., Burmeister, M., Shedden, K., & Sen, S. (2011). The serotonin transporter promoter variant (5-HT-TLPR), stress, and depression meta-analysis revisited: Evidence of genetic moderation. Archives of Gen-eral Psychiatry, 68(5), 444-454.
169DISCUSSION
Kessler, R. C., & Wang, P. S. (2009). Epidemiology of depression. In I. H. Gotlib, & C. L. Hammen (Eds.), Hand-book of depression (2nd ed., pp. 5-22). New York: The Guildford Press.
Lamiell, J. T. (1998). ‘Nomothetic’ and ‘idiographic’ - contrasting Windelband’s understanding with contem-porary usage. Theory & Psychology, 8(1), 23-38.
Lapin, I. P., & Oxenkrug, G. F. (1969). Intensification of the central serotoninergic processes as a possible determinant of the thymoleptic effect. Lancet, 1(7586), 132-136.
Leo, J., & Lacasse, J. R. (2008). The media and the chemical imbalance theory of depression. Society, 45(1), 35-45.
Lewinsohn, P. M., Solomon, A., Seeley, J. R., & Zeiss, A. (2000). Clinical implications of “subthreshold” de-pressive symptoms. Journal of Abnormal Psychology, 109(2), 345-351.
Lindwall, M., Larsman, P., & Hagger, M. S. (2011). The reciprocal relationship between physical activity and depression in older European adults: A prospective cross-lagged panel design using SHARE data. Health Psychology, 30(4), 453-462.
Lombardo, M. P. (2012). On the evolution of sport. Evolutionary Psychology, 10(1), 1-28.
Luscher, B., Shen, Q., & Sahir, N. (2011). The GABAergic deficit hypothesis of major depressive disorder. Mo-lecular Psychiatry, 16(4), 383-406.
Mammen, G., & Faulkner, G. (2013). Physical activity and the prevention of depression: A systematic re-view of prospective studies. American Journal of Preventive Medicine, 45 (5), 649-57. doi:10.1016/j.amepre.2013.08.001
Mata, J., Thompson, R. J., & Gotlib, I. H. (2010). BDNF genotype moderates the relation between physical activity and depressive symptoms. Health Psychology, 29(2), 130-133.
McGale, N., McArdle, S., & Gaffney, P. (2011). Exploring the effectiveness of an integrated exercise/CBT inter-vention for young men’s mental health. British Journal of Health Psychology, 16, 457-471.
Mead, G. E., Morley, W., Campbell, P., Greig, C. A., McMurdo, M., & Lawlor, D. A. (2009). Exercise for depres-sion. Cochrane Database of Systematic Reviews (Online), (3), CD004366.
Milton, K., Clemes, S., & Bull, F. (2012). Can a single question provide an accurate measure of physical activ-ity? British Journal of Sports Medicine, 47(1), 44-8.
Molenaar, P. C. M., & Campbell, C. G. (2009). The new person-specific paradigm in psychology. Current Direc-tions in Psychological Science, 18(2), 112-117.
Morrow, R. L., Garland, E. J., Wright, J. M., Maclure, M., Taylor, S., & Dormuth, C. R. (2012). Influence of relative age on diagnosis and treatment of attention-deficit/hyperactivity disorder in children. Canadian, Medical Association Journal, 184(7), 755-62.
Munafo, M. R., Durrant, C., Lewis, G., & Flint, J. (2009). Gene X environment interactions at the serotonin transporter locus. Biological Psychiatry, 65(3), 211-219.
Nestler, E. J., & Carlezon, W. A. (2006). The mesolimbic dopamine reward circuit in depression. Biological Psychiatry, 59(12), 1151-1159.
Nettle, D. (2004). Evolutionary origins of depression: A review and reformulation. Journal of Affective Dis-orders, 81(2), 91-102.
Peyrot, W. J., Milaneschi, Y., Abdellaoui, A., Sullivan, P. F., Hottenga, J. J., Boomsma, D. I., et al. (2014). Effect of polygenic risk scores on depression in childhood trauma. British Journal of Psychiatry, 205(2), 113-9.
Ploughman, M. (2008). Exercise is brain food: The effects of physical activity on cognitive function. Develop-mental Neurorehabilitation, 11(3), 236-240.
Price, J. S. (1988). Alternative channels for negotiating asymmetry in social relationships. In M. R. A. Chance (Ed.), Social fabrics of the mind (pp. 157-195). Erlbaum.
CHAPTER VIII170
Price, J. S., Sloman, L., Gardner, R., Gilbert, P., & Rohde, P. (1994). The social competition hypothesis of de-pression. British Journal of Psychiatry, 164, 309-315.
Prince, S. A., Adamo, K. B., Hamel, M. E., Hardt, J., Gorber, S. C., & Tremblay, M. (2008). A comparison of direct versus self-report measures for assessing physical activity in adults: A systematic review. International Journal of Behavioral Nutrition and Physical Activity, 5, 56.
Redwine, L. S., Tsuang, M., Rusiewicz, A., Pandzic, I., Cammarata, S., Rutledge, T., et al. (2012). A pilot study exploring the effects of a 12-week T’ai chi intervention on somatic symptoms of depression in patients with heart failure. Journal of Alternative and Complementary Medicine, 18(8), 744-748.
Rethorst, C. D., Landers, D. M., Nagoshi, C. T., & Ross, J. T. (2010). Efficacy of exercise in reducing depressive symptoms across 5-HTTLPR genotypes. Medicine and Science in Sports and Exercise, 42(11), 2141-2147.
Rethorst, C. D., Landers, D. M., Nagoshi, C. T., & Ross, J. T. (2011). The association of 5-HTTLPR genotype and depressive symptoms is moderated by physical activity. Journal of Psychiatric Research, 45(2), 185-189.
Risch, N., Herrell, R., Lehner, T., Liang, K. Y., Eaves, L., Hoh, J., et al. (2009). Interaction between the sero-tonin transporter gene (5-HTTLPR), stressful life events, and risk of depression: A meta-analysis. JAMA: The Journal of the American Medical Association, 301(23), 2462-2471.
Rosenthal, R. (1990). How are we doing in soft psychology. American Psychologist, 45(6), 775-777.
Roshanaei-Moghaddam, B., Katon, W. J., & Russo, J. (2009). The longitudinal effects of depression on physi-cal activity. General Hospital Psychiatry, 31(4), 306-315.
Rosmalen, J. G. M., Wenting, A. M., Roest, A. M., de Jonge, P., & Bos, E. H. (2012). Revealing causal hetero-geneity using time series analysis of ambulatory assessments: Application to the association between depression and physical activity after myocardial infarction. Psychosomatic Medicine, 74(4), 377-386.
Russell, J. A. (2003). Core affect and the psychological construction of emotion. Psychological Review, 110(1), 145-172.
Sallis, J. F., & Owen, N. (1999). Physical activity and behavioral medicine. Thousand Oaks: SAGE publi-cations, Inc.
Salmon, P. (2001). Effects of physical exercise on anxiety, depression, and sensitivity to stress: A unifying theory. Clinical Psychology Review, 21(1), 33-61.
Sanders, C. E., Field, T. M., Diego, M., & Kaplan, M. (2000). Moderate involvement in sports is related to lower depression levels among adolescents. Adolescence, 35(140), 793-797.
Sasaki-Adams, D. M., & Kelley, A. E. (2001). Serotonin-dopamine interactions in the control of conditioned reinforcement and motor behavior. Neuropsychopharmacology, 25(3), 440-452.
Smoller, J. W., Craddock, N., Kendler, K., Lee, P. H., Neale, B. M., Nurnberger, J. I., et al. (2013). Identifica-tion of risk loci with shared effects on five major psychiatric disorders: A genome-wide analysis. Lancet, 381(9875), 1371-1379.
Struder, H. K., & Weicker, H. (2001). Physiology and pathophysiology of the serotonergic system and its im-plications on mental and physical performance. Part II. International Journal of Sports Medicine, 22(7), 482-497.
Tao, F. B., Xu, M. L., Kim, S. D., Sun, Y., Su, P. Y., & Huang, K. (2007). Physical activity might not be the pro-tective factor for health risk behaviours and psychopathological symptoms in adolescents. Journal of Paediatrics and Child Health, 43(11), 762-767.
Thase, M. E. (2009). Neurobiological aspects of depression. In I. H. Gotlib, & C. L. Hammen (Eds.), Handbook of depression (2nd ed., pp. 187-217). New York: The Guildford Press.
Thompson-Coon, J., Boddy, K., Stein, K., Whear, R., et al. (2011). Does participating in physical activity in outdoor natural environments have a greater effect on physical and mental wellbeing than physical ac-tivity indoors? A systematic review. Environmental Science & Technology, 45(5), 1761-72.
171DISCUSSION
Thoren, P., Floras, J. S., Hoffmann, P., & Seals, D. R. (1990). Endorphins and exercise - physiological-mecha-nisms and clinical implications. Medicine and Science in Sports and Exercise, 22(4), 417-428.
Toups, M. S. P., Greer, T. L., Kurian, B. T., Grannemann, B. D., Carmody, T. J., Huebinger, R., et al. (2011). Effects of serum brain derived neurotrophic factor on exercise augmentation treatment of depression. Journal of Psychiatric Research, 45(10), 1301-1306.
van Praag, H. M., Christie, B. R., Sejnowski, T. J., & Gage, F. H. (1999). Running enhances neurogenesis, learn-ing, and long-term potentiation in mice. Proceedings of the National Academy of Sciences of the United States of America, 96(23), 13427-13431.
van Praag, H. M., Kempermann, G., & Gage, F. H. (1999). Running increases cell proliferation and neurogen-esis in the adult mouse dentate gyrus. Nature Neuroscience, 2(3), 266-270.
van Praag, H. M., & Korf, J. (1970). L-tryptophan in depression. Lancet, 2(7673), 612.
van Praag, H. M. (2008). Neurogenesis and exercise: Past and future directions. Neuromolecular Medicine, 10(2), 128-140.
Willner, P. (1995). Dopaminergic mechanisms in depression and mania. In F. E. Bloom, & D. J. Kupfer (Eds.), Psychopharmacology: The fourth generation of progress (pp. 614-620). New York: Raven Press.
Willner, P. (1997). The mesolimbic dopamine system as a target for rapid antidepressant action. Interna-tional Clinical Psychopharmacology, 12, S7-S14.
Summary in English
175SUMMARY IN ENGLISH
Depression is a global public health problem, mainly because of the relatively high lifetime
prevalence and the substantial disability of individuals afflicted with this disorder. In the gen-
eral population, an estimated one out of 20 people will suffer from major depression at any
given point in time. The causes of the disorder include a complex interaction of psychological,
biological and behavioral factors, which makes depression difficult to treat. The effective-
ness of current interventions to treat depression, such as cognitive behavioral therapy and
antidepressant medication, are intensely debated. In the last few decades, research has
focused on the potential benefits of physical activity in treating depression. Early studies
suggested that physical activity might be effective in alleviating depressive symptoms, thus
providing new grounds for optimism and potentially offering a valuable intervention for the
treatment of depression. However, the evidential base for this optimism is not very strong,
especially concerning depressive symptoms in adolescents.
The aim of this dissertation was to explore in depth the relationship between physical
activity and depressive symptoms in adolescents and adults. A better understanding of
the direction and strength of this relationship and identification of possible moderators
may contribute to an effective prevention and treatment of depression. The research de-
scribed in this thesis was part of the studies Tracking Adolescents’ Individual Lives Survey
(TRAILS) and Mood and Movement in Daily Life (MOOVD). TRAILS has followed a large
cohort of Dutch adolescents from early adolescence up until adulthood. MOOVD examines
the temporal relationship between physical activity and mood in daily life in depressed and
non-depressed adults.
In chapter II, the direction of the relationship between physical activity and depressive
symptoms was explored and discussed. Previous research has primarily focused on the ef-
fects of physical activity on depressive symptoms, and largely ignored the potential effects
of depressive symptoms on physical activity. We observed a bidirectional prospective re-
lationship. Adolescents who were active exhibited fewer depressive symptoms over time
than more sedentary adolescents, and the opposite was also true, that is, adolescents with
many depressive symptoms became more sedentary over time than adolescents with fewer
depressive symptoms. Cognitive symptoms of depression such as loss of interest and de-
pressed mood were more likely to be associated with physical activity than somatic symp-
toms such as loss of appetite, sleep and energy.
The study described in chapter III investigated the protective hypothesis of physical ac-
tivity, which states that early engagement in physical activity might prevent the onset of
depression. This was explored regarding the period from adolescence to early adulthood,
which spanned approximately 6 years. Effects of specific characteristics of physical activity,
i.e., its nature, duration, frequency and intensity, were examined in the analysis. Physical
SUMMARY IN ENGLISH176
activity did not protect against a first episode of depression in this study, so the hypoth-
esis that early physical activity might prevent the onset of depression was not supported
empirically.
Chapter IV concerns the potential role of so-called plasticity genes on the (bidirectional)
relationship between physical activity and depressive symptoms. The term plasticity genes
refers to polymorphisms that have been reported to influence the actions of some neuro-
modulating and neurotrophin systems. These systems, among which are the serotonergic
and dopaminergic systems, can be altered by physical activity and depression. Overall, ado-
lescents with these polymorphisms assumed to reflect high plasticity did not differ from
other adolescents with regard to any of the associations between physical activity and de-
pressive symptoms. In other words, we found no evidence that the proposed genes directly
influence the relationship between physical activity and depressive symptoms.
The study in chapter V explored whether adolescents participating in competitive sports
reported fewer concurrent depressive symptoms and developed fewer depressive symp-
toms over time. In addition, we examined the influence of peer-perceived sport competence
and gender on the relationship between competitive sport participation and depressive
symptoms. Overall, both boys and girls who took part in competitive sports reported fewer
concurrent depressive symptoms, but did not differ from others with regard to symptom
changes over time. The association between sport participation and concurrent depressive
symptoms was largely explained by peer-perceived sport competence. This indicates that
athletically competent adolescents are less likely to exhibit depressive symptoms, regard-
less of their competitive sport participation.
Chapter VI concerns the effects of classroom positioning according to birth dates on multiple
outcomes, including depressive symptoms, depressive symptom changes during adoles-
cence, and teacher-rated sport competence. Substantial relative-age effects were observed
in relation to school progress: relatively young adolescents were four times more likely to
repeat a grade than the relatively old adolescents, while relatively old adolescents were up
to 20 times more likely to skip a grade. In the subgroup of adolescents who had repeated a
grade, the younger ones exhibited fewer depressive symptoms than relatively older peers
who had repeated a grade. Finally, the relative-age effects did not influence the evaluation
of sport competence in either the normative group or the subgroup that repeated a grade.
The final empirical chapter of the thesis (chapter VII) examined the relationship between
physical activity and affect in depressed and non-depressed adults, using a within-subject
approach. During a month, physical activity and affect (positive and negative) were mea-
sured three times a day by means of accelerometers and diaries. The results showed indi-
vidual differences in both the direction and the strength of the relationships. Cross-sectional
177SUMMARY IN ENGLISH
correlations between physical activity and positive affect were positive in nearly all individu-
als, though not always significant, which suggests that positive affect was enhanced during
and immediately after physical activity. No consistent improvement was observed in nega-
tive affect. The long-term effects of physical activity on positive and negative affect varied
greatly across individuals, regardless of their depression status. While for a few individuals
physical activity improved (positive) affect in the long term, for the majority of individuals
the relationship did not exist or was even reversed.
In conclusion, this thesis confirms the existence of a relationship between physical activity
and depressive symptoms, but it is rather weak, and shows substantial individual differ-
ences. Some individuals tend to feel better after having been physically active, while others
feel worse or show no relationship between mood and activity. The reasons behind these
differences are still unknown, and we do not know yet which individuals might benefit from
an active lifestyle. This knowledge is needed to design effective personalized interven-
tions for the treatment of depression. Up to this point, however, the evidence that physical
activity can help treat depressive symptoms is relatively weak, and additional research is
required before physical activity can be implemented as an effective strategy for the treat-
ment of depression.
Summary in Dutch
181SUMMARY IN DUTCH
Depressie is een wereldwijd gezondheidsprobleem. Dit komt voornamelijk doordat relatief
veel mensen een depressie doormaken in hun leven, maar ook omdat depressie gepaard
kan gaan met fors functieverlies. Naar schatting 1 op de 20 mensen krijgt gedurende de le-
vensloop een depressieve stoornis. De oorzaken van depressieve stoornissen omvatten een
complexe interactie van psychologische, biologische, en gedragsmatige factoren, hetgeen
depressie lastig te behandelen maakt. De effectiviteit van de beschikbare interventies voor
depressie, zoals cognitieve gedragstherapie en antidepressieve medicatie, zijn onderwerp
van hevig debat. Wetenschappers hebben zich de afgelopen decennia gericht op de poten-
tiële voordelen van lichamelijke beweging als alternatieve behandeling van depressie. Tot
dusver hebben enkele studies laten zien dat lichamelijke beweging effectief kan zijn in het
verminderen van depressieve symptomen en dus een waardevolle aanvulling kan zijn. Er is
echter nog weinig bewijs voor dit optimisme, in het bijzonder met betrekking tot depressieve
symptomen in de adolescentie.
Het doel van dit proefschrift is om de associatie tussen lichamelijke activiteit en depressieve
symptomen uit te diepen. Een beter begrip van de richting en sterkte van de associatie tus-
sen lichamelijke (in)activiteit en depressieve symptomen en identificatie van mogelijke mo-
deratoren kan bijdragen aan het voorkomen en behandelen van depressie. Het proefschrift
beschrijft onderzoek dat is uitgevoerd in de studies TRAILS (‘Tracking Adolescents’ Indivi-
dual Lives Survey’) en MOOVD (Mood and Movement in Daily Life). TRAILS volgt een groot
cohort van Nederlandse adolescenten van de vroege adolescentie tot in de volwassenheid.
MOOVD onderzoekt de temporele relatie tussen lichamelijke activiteit en stemming in het
dagelijks leven in depressieve en niet-depressieve volwassenen.
In hoofdstuk II wordt de richting van de verbanden tussen lichamelijke activiteit en depres-
sieve symptomen verkend en bediscussieerd. Eerder onderzoek was voornamelijk gericht
op de effecten van lichamelijke activiteit op depressieve symptomen, terwijl de potentiële
effecten van depressieve symptomen op lichamelijke activiteit grotendeels werden gene-
geerd. In onze studie werd een wederkerige relatie geobserveerd. Adolescenten die actief
waren ervoeren minder depressieve symptomen op het volgende meetmoment dan ado-
lescenten die minder bewogen, maar het tegenovergestelde was ook waar: adolescenten
met veel depressieve symptomen bewogen vervolgens minder dan adolescenten met minder
depressieve symptomen. Cognitieve symptomen van depressie, zoals verlies van interesse
en depressieve stemming, hingen sterker samen met lichamelijke activiteit dan lichamelijke
symptomen zoals verlies van eetlust, slaap en energie.
De studie beschreven in hoofdstuk III onderzocht de hypothese dat lichamelijke activiteit be-
schermt tegen het ontstaan van een depressie. Dit werd in kaart gebracht voor een ongeveer
6 jaar durende periode die liep van de adolescentie tot de vroege volwassenheid. Effecten
SUMMARY IN DUTCH182
van de aard, duur, frequentie en intensiteit van beweging werden nagegaan in de analyse.
Lichamelijke activiteit voorkwam een eerste episode van depressie niet in deze studie; de
hypothese dat vroege lichamelijke activiteit beschermt tegen het ontstaan van een depres-
sie werd dus niet ondersteund.
Hoofdstuk IV betreft de potentiële rol van de zogenaamde plasticiteitsgenen in de (weder-
kerige) relatie tussen lichamelijke activiteit en depressieve symptomen. De term plastici-
teitsgenen verwijst naar genvarianten waarvan bekend is dat ze de activiteit van bepaalde
neuromodulatoren en neurotropische systemen beïnvloeden. Deze systemen, waaronder
het serotonerge en dopaminerge systeem, kunnen worden beïnvloed door fysieke activi-
teit en depressie. Adolescenten met genvarianten waarvan verondersteld wordt dat ze een
hoge mate van plasticiteit weerspiegelen lieten geen sterkere (of zwakkere) verbanden
tussen depressieve symptomen en lichamelijke inactiviteit zien dan adolescenten zonder
deze varianten. Met andere woorden, er werd geen bewijs gevonden dat de voorgestelde
genen een directe invloed hadden op de associatie tussen lichamelijke activiteit en depres-
sieve symptomen.
De studie in hoofdstuk V onderzocht of adolescenten die deelnamen aan competitieve spor-
ten minder depressieve symptomen hadden en minder depressieve symptomen ontwikkel-
den in de loop der jaren. Daarnaast testten we de invloed van geslacht en sportcompetentie
volgens klasgenoten op de associatie tussen deelname aan competitieve sporten en de-
pressieve symptomen. Zowel jongens als meisjes die deelnamen aan competitieve sporten
rapporteerden minder depressieve symptomen, maar lieten geen verschil zien met betrek-
king tot toekomstige veranderingen in symptoomniveaus. De associatie tussen deelname
aan sport en depressieve symptomen werd grotendeels verklaard door sportcompetentie
zoals waargenomen door klasgenoten. Dit suggereert dat atletisch competente adolescen-
ten minder depressieve symptomen rapporteren, ongeacht hun deelname aan competitieve
sporten.
Hoofdstuk VI betreft de effecten van de relatieve leeftijd van adolescenten in hun schoolklas
op verschillende uitkomsten, waaronder depressieve symptomen, veranderingen in depres-
sieve symptomen tijdens de adolescentie en door de docent waargenomen sportcompeten-
tie. Substantiële relatieve leeftijdseffecten werden geobserveerd in relatie tot de kans om
zitten te blijven of een jaar over te slaan: adolescenten die relatief jong waren in hun klas
hadden een vier keer zo grote kans dat ze een jaar moesten overdoen dan hun relatief oudere
klasgenoten; terwijl de relatief oude adolescenten een 20 keer zo grote kans hadden dat ze
een klas mochten overslaan. In de subgroep van adolescenten die een jaar moest overdoen
hadden de relatief jonge adolescenten minder depressieve symptomen dan relatief oudere
leeftijdsgenoten die ook een jaar over moesten doen. De relatieve leeftijdseffecten hadden
183SUMMARY IN DUTCH
geen invloed op de waargenomen sport competentie in zowel de normatieve groep als de
subgroep die een jaar bleef zitten.
In het laatste empirische hoofdstuk van het proefschrift, hoofdstuk VII, werd de relatie tus-
sen lichamelijke activiteit en stemming in depressieve en niet-depressieve volwassenen
onderzocht door middel van een zogenaamde binnen-persoon benadering. Gedurende een
maand werden lichamelijke activiteit en affect (positief en negatief) driemaal daags gemeten
met een accelerometer en vragenlijsten. We vonden individuele verschillen in zowel de rich-
ting als de sterkte van de verbanden. Bijna alle individuen rapporteerden meer positief affect
tijdens of onmiddellijk na lichamelijke activiteit (positief cross-sectioneel verband), maar de
associatie was niet altijd significant. Er werd echter geen consistente verbetering in nega-
tief affect geobserveerd. De lange-termijn effecten van lichamelijke activiteit op positief en
negatief affect varieerde sterk tussen mensen, onafhankelijk van hun depressieve status.
Voor sommige personen voorspelde lichamelijke activiteit een verbetering van de stemming
over de tijd, maar voor de meerderheid van de deelnemers bestond dit verband niet of was
de relatie zelfs omgekeerd.
Samengevat bevestigt dit proefschrift het bestaan van een relatie tussen lichamelijke ac-
tiviteit en depressieve symptomen, maar het verband is zwak en vertoont forse individuele
verschillen. Sommige mensen voelen zich beter na lichamelijke inspanning, terwijl er voor
andere personen geen relatie bestaat tussen stemming en activiteit en weer anderen zich
juist slechter voelen na lichamelijke inspanning. De redenen voor deze verschillen zijn tot
dusver onbekend; en we weten nog niet welke mensen baat kunnen hebben van een actieve
leefstijl. Deze kennis is nodig om effectieve gepersonaliseerde interventies voor de behan-
deling van depressie te ontwikkelen. Tot dusver is het bewijs dat lichamelijke activiteit kan
helpen om depressieve symptomen te behandelen relatief zwak. Kortom, meer onderzoek is
nodig voordat lichamelijke activiteit kan worden geïmplementeerd als een effectieve strate-
gie voor de behandeling van depressie.
Acknowledgements
187ACKNOWLEDGEMENTS
The writing of this book has been a great adventure. I have always argued that the message
is more important than the messenger, however I have to thank all of the people that have
helped with the delivery of this message. I have to express my gratitude to Peter de Jonge
for giving me the opportunity to undertake this project. Without his keen eye for talent and
without him taking a risk with me, this book would have not materialized. I also want to
thank him for the support and guidance throughout these years. A special thank you to
Tineke Oldehinkel. Every new scientist and author should have the experience of working
with Tineke. Tineke’s insights, honesty, skeptical mind and amazing ability with details have
been a revelation for me. Tineke, thank you for your support and for being the best educator
I had over those years. I would also like to thank Annelieke Roest, who although came late
in the project, she has been a great guidance through patchy moments. Thank you Annelieke
for being an empathic individual who encouraged me and provided me with new perspectives
that I was not previously aware of.
I am very thankful to Liesbeth, Margo, Gerry and Martha (the lovely secretaries), for mak-
ing me feel like home in a foreign land. Without your cheerful personalities, incredible help
and guidance, life at the ICPE would have been very difficult for me and rather boring. Thank
you ladies.
I would like to thank Hans (Jans!) Ormel for all the critical comments and sharp insights. I
know Jans that you can make life very difficult for your collaborators, but life and science
would not be as much fun if they were easy. A big thank to Elske Bos for some truly great
statistical guidance and philosophical discussions. Keep dancing Elske! Dennis Raven, thank
you for being patient with me but also for your great guidance with the TRAILS dataset. I am
sure you will miss me appearing at your door with all those sticky notes.
I would like to thank Esther Nederhof, Richard Oude Voshaar, René Veenstra and Frank Ver-
hulst for their input throughout my project. I have enjoyed working with you, even if brief
in some cases, and hopefully have learned a few important skills from our collaborations.
Thanks to all volunteers who have participated in this project. A special thank you also to
Catharina Hartman, who although was not directly involved in this project, she always had
her door open for me with whatever I needed. Working with you and Xiaochen has taught me
a lot about myself, and what is needed to be a good teacher.
Many friends and relatives have been supportive during this project. Thank you Ma and Pa
(pappou!) for keeping me in line when facing tough times. Pappou: eagles should not be
feeding on flies, and you might be right. A special thank goes to my sister. Andriana, I am
extremely grateful for your time and input on this project. Jez, Kosta, Jan and Harm Jan (my
dear friends) thank you for your support, intriguing discussions and funny moments. Thanks
for being there for me whenever I needed you guys.
ACKNOWLEDGEMENTS188
Thanks to my roomies for five years. Sanne, Nynke and Bertus, three great young minds that
have helped me, supported me and inspired me to finish. Funny that I managed to finish first,
hey? Thank you guys for putting up with my absurdity, fidgety behavior and noisiness. Sanne
and Bertus it has been a great pleasure for me to collaborate with you on two papers of this
thesis. Nynke thank you for helping me with the student seminars and fruitful discussions
when I was stuck (damn statistics!). Above all, I thank you guys for being amazing friends.
Also thanks to the lovely girls next door, Arunima, Elise, Anna-Roos, Charlotte and Petra
for the sometimes absurd and mind opening discussions. Arunima thank you for giving me
feedback on early versions of the discussion of this thesis. Anna-Roos you have become a
great friend over the years, even though our initial moments were shaky (that crappy bike!).
I feel immense gratitude that you provided me with a house to stay for four months and
introduced me to some great friends of yours.
Finally, many thanks to the wonderful people that I have failed to mention here, but who
nonetheless played an important part in my life in Groningen.
Nikolaos
About the Author
191ABOUT THE AUTHOR
On the 23rd of June 1981 Nikolaos Stavrakakis was born in Athens, Greece. He finished sec-
ondary education at the International Baccalaureate Moraitis school in Athens in 1999. In
the same year he started his study in psychology at Reading University in the UK. After he
received his bachelor’s degree, Nikolaos undertook a master’s degree in Neuroscience in Im-
perial College in London in 2004. There he worked as an internship at the Cyclotron building
of Hammersmith hospital in London investigating emotional regulation and the serotonin
transporter gene. From the beginning of 2010 until august 2014 Nikolaos has worked on the
research presented in this dissertation. Currently, Nikolaos is living in Athens.
List of Publications
195LIST OF PUBLICATIONS
Stavrakakis N, de Jonge P, Ormel J, Oldehinkel AJ. “Bidirectional prospective associations
between physical activity and depressive symptoms. The TRAILS Study.” Journal of Adoles-
cent Health 2012; 50(5):503-8.
Stavrakakis N, Oldehinkel AJ, Nederhof E, Oude Voshaar RC, Verhulst FC, Ormel J, de Jonge
P. “Plasticity genes do not modify associations between physical activity and depressive
symptoms.” Health Psychology 2013; 32(7):785-92.
Stavrakakis N, Roest AM, Verhulst F, Ormel J, de Jonge P, Oldehinkel AJ. “Physical activ-
ity and onset of depression in adolescents: Α prospective study in the general population
cohort TRAILS.” Journal of Psychiatric Research 2013; 47(10):1304-8.
Rhodes RA, Murthy NV, Dresnet MA, Selvaraj S, Stavrakakis N, Babar S, Cowen PJ, Grasby
PM. “Human 5-HT transporter availability predicts amygdala reactivity in vivo.” Journal of
Neuroscience 2007; 27(34):9233-7.
Submitted or under preparation
Jeronimus BF, Stavrakakis N, Veenstra R, Oldehinkel AJ. “Relative Age Effects in Adoles-
cents: The TRAILS study” (under review).
Stavrakakis N, Roest AM, Verhulst FC, Veenstra R, Ormel J, de Jonge P, Oldehinkel AJ.
“Competitive sport participation, sport competence and depressive symptoms in adolescent
boys and girls” (under preparation).
Stavrakakis N, Booij SH, Roest AM, de Jonge P, Oldehinkel AJ, Bos EH. “The dynamic rela-
tionship between physical activity and mood in the daily life of depressed and non-depressed
individuals: a within-subject time-series analysis” (under preparation).
Xiaochen L, Stavrakakis N, Penninx BW, Bosker FJ, Nolen WA, Boomsma DI, de Geus EJC,
Smit JH, Snieder H, Nolte IM, Hartman CA. “Does refining the phenotype improve replication
rates? A candidate gene study on Major Depressive Disorder and Chronic Major Depressive
Disorder” (under preparation).
Recent TRAILS Dissertations
199RECENT TRAILS DISSERTATIONS
Some of the work described in this thesis is part of the Tracking Adolescents’ Individual
Lives Survey (TRAILS) and was performed at the Interdisciplinary Centre Psychopathology
and Emotion regulation (ICPE) of the University Medical Centre Groningen (UMCG), Univer-
sity of Groningen, the Netherlands. More information concerning TRAILS and the research
within can be found in the dedicated internetsite: www.trails.nl
Recent TRAILS theses
Mathyssek, CM (2014). The development of anxiety symptoms in adolescents. Supervisors:
Prof. dr. F.C. Verhulst, dr. F.V.A. van Oort.
Prince van Leeuwen, AL (2013). Blunt vulnerabilities. Identifying risks for initiation and con-
tinued use of cannabis in a Dutch adolescent population. Supervisors: Prof. dr. A.C. Huizink,
Prof. dr. F.C. Verhulst, dr. H.E. Creemers.
Verboom, CE (2012). Depression and role functioning. Their relation during adolescence and
adulthood. Supervisors: Prof. dr. J. Ormel, Prof. dr. W.A. Nolen, Prof. dr. B.W.J.H. Pennix, dr.
J.J. Sijtsema.
Marsman, H (2013). HPA-axis, genes and environmental factors in relation to externalizing
behaviors. Supervisors: Prof. dr. J.K. Buitelaar.
Griffith-Lendering, MFH (2013). Cannabis use, cognitive functioning and behaviour prob-
lems. Supervisors: Prof. dr. H. Swaab, dr. S.C.J. Huijbregts.
Vink, NM (2013). The role of stress in the etiology of asthma. Supervisors: Prof. dr. H.M.
Boezen, Prof. dr. J.G.M. Rosmalen, Prof. dr. D.S. Postma.
Laceulle, OM (2013). Programming effects of adversity on adolescent adaptive capacity. Su-
pervisors: Prof. dr. J. Ormel, Prof. dr. M.A.G. van Aken, dr. E. Nederhof.