Resea ch A icle
T ajec o ies o Physical Ac i i y P edic
he Onse o Dep essi e Symp oms bu No Thei P og ession:
A P ospec i e Coho S udy
Kaisa Kase a,1Tom Rosens öm,1Taina Hin sa,1Lau a Pulkki-Råback,1,2
Tuija Tammelin,3Ja i Lipsanen,1Xiaolin Yang,3Mi ka Hin sanen,1,4 Ch is ian Hakulinen,1
Ka ja Pahkala,5,6 Mi ja Hi ensalo,7Nina Hu i-Kähönen,8
Olli T. Rai aka i,5,9 and Liisa Kel ikangas-Jä inen1
1Uni o Pe sonali y, Wo k and Heal h Psychology, Ins i u e o Beha io al Sciences, P.O. Box 9, Uni e si y o Helsinki,
00014 Helsinki, Finland
2Helsinki Collegium o Ad anced S udies, Fabianinka u 24, P.O. Box 4, Uni e si y o Helsinki, 00014 Helsinki, Finland
3LIKES, Resea ch Cen e o Spo and Heal h Sciences, Rau pohjanka u 8, 40700 Jy ¨
askyl¨
a, Finland
4Uni o Psychology,Uni e si yo Oulu,P.O.Box8000,90014Oulu,Finland
5Resea ch Cen e o Applied and P e en i e Ca dio ascula Medicine, Kiinamyllynka u 10, Uni e si y o Tu ku, 20520 Tu ku, Finland
6Paa o Nu mi Cen e, Spo s and Exe cise Medicine Uni , Depa men o Physical Ac i i y and Heal h, Kiinamyllynka u 10,
Uni e si y o Tu ku, 20520 Tu ku, Finland
7Depa men o Spo Sciences,P.O.Box35(L),Uni e si yo Jy
¨
askyl¨
a, 40014 Jy ¨
askyl¨
a, Finland
8Depa men o Pedia ics,P.O.Box2000,Uni e si yo Tampe eandTampe eUni e si yHospi al,33521Tampe e,Finland
9Depa men o Clinical Physiology and Nuclea Medicine, Kiinamyllynka u 4-8, Tu ku Uni e si y Hospi al, 20520 Tu ku, Finland
Co espondence should be add essed o Liisa Kel ikangas-J¨
a inen; Liisa.Kel ikangas-Ja [email p o ec ed]
Recei ed 1 Ma ch 2016; Re ised 12 Augus 2016; Accep ed 31 Augus 2016
Academic Edi o : Ian L. Swaine
Copy igh © 2016 Kaisa Kase a e al. This is an open access a icle dis ibu ed unde he C ea i e Commons A ibu ion License,
which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed.
This p ospec i e, communi y-based s udy examined ajec o ies o physical ac i i y om childhood o adul hood and whe he
hese ajec o ies con ibu ed o dep essi e symp oms in adul hood o a g ea e deg ee han adul hood physical ac i i y. Pa icipan s
(𝑛 = 3596) we e om he ongoing Ca dio ascula Risk in Young Finns S udy which s a ed in 1980. Dep essi e symp oms we e
measu ed wi h Beck Dep ession In en o y (BDI-II) in 2012, and physical ac i i y was assessed om 1980 o 2011 wi h sel - epo s.
Analyses we e adjus ed o age, sex, childhood nega i e emo ionali y, socioeconomic ac o s, p e ious dep essi e symp oms, social
suppo , body mass index, and smoking s a us (1980–2007). Highly, mode a ely, and ligh ly physically ac i e ajec o y g oups we e
iden i ied. Highly ac i e pa icipan s epo ed lowe le els o dep essi e symp oms compa ed o ligh ly ac i e ones (𝑝 < 0.001)
and compa ed o mode a ely ac i e ones (𝑝 = 0.001). Mode a ely ac i e pa icipan s had less symp oms han ligh ly ac i e ones
(𝑝 < 0.001). High le els o adul hood physical ac i i y associa ed wi h lowe le els o dep essi e symp oms (𝑝 < 0.001). The indings
did no wi hs and adjus men o p e ious dep essi e symp oms (𝑝 > 0.05). Li elong physical ac i i y ajec o ies o adul hood
physical ac i i y was no associa ed wi h he p og ession o dep essi e symp oms in adul hood. Thus, physical ac i i y his o y does
no con ibu e o he p og ession o he dep essi e symp oms o a g ea e deg ee han adul hood physical ac i i y.
1. In oduc ion
Dep essi e diso de s a ec 340 million people (app oxi-
ma ely 9%) globally ega dless o gende , cul u e, o e hnici y
[1]. F om 1990 ill p esen ime, dep ession has been a ed
one o he leading causes o yea s li ed wi h disabili y (YLDs)
[2]. Dep ession associa es wi h a wide ange o subsequen
heal h ou comes. Fo ins ance, dep ession p edic s u u e
co ona y e en s and ca diac dea hs in heal hy indi iduals
and in hose wi h es ablished ca dio ascula disease (CHD)
Hindawi Publishing Co po a ion
Jou nal o Spo s Medicine
Volume 2016, A icle ID 8947375, 9 pages
h p://dx.doi.o g/10.1155/2016/8947375
2Jou nal o Spo s Medicine
[3, 4]. Dep ession has also been shown o be associa ed
wi h se e al o he physical diseases, including cance and
os eopo osis [4], as well as wi h se e e psychia ic diso de s
such as schizoph enia [5]. Fu he mo e, dep ession ela es o
low educa ional a ainmen and educed wo k p oduc i i y
[6]. Iden i ying ac o s ha con ibu e o dep ession and
de eloping ools ha can be applied in ea ly p e en ion o
dep ession ha e been p ominen ly highligh ed [7].
E en hough some con lic ing indings conce ning he
associa ion be ween physical ac i i y and dep ession exis [8,
9], p io e idence p o ides suppo o he alue o physical
ac i i y in educing dep essi e symp oms in bo h heal hy
and clinical popula ions [10]. Li e a u e has also indica ed
ha e en low doses o physical ac i i y may p o ec agains
dep ession [10]. The ela ion be ween physical ac i i y and
dep ession may also be bidi ec ional, as physical ac i i y may
alle ia e dep essi e symp oms, bu hese symp oms may also
dec ease he likelihood o ini ia ing physical ac i i y [8, 11].
Physical ac i i y has also been ela ed o neu obiological
unc ioning, which may p eemp dep essi e mood [12]. In
addi ion, physical ac i i y has been shown o be associa ed
wi h he de elopmen o indi idual quali ies such as sel -
con idence, emo ional sel - egula ion [13], and capaci y o
social bonding [14], which, in u n, may be p o ec i e agains
dep ession [13, 14].
Se e al s udies ha e indica ed ha egula physical ac i -
i y associa es wi h educed isk o dep ession in all age g oups
om ea ly childhood o la e adul hood [8, 10, 15]. The e is also
e idence ha childhood physical ac i i y may lead o main e-
nance o li e ime physical ac i i y pa e ns conduci e o well-
being (e.g., [16]). These pa e ns, howe e , ha e a endency
o decline owa ds la e adul hood [17]. Many p e ious ind-
ings ega ding he associa ion be ween physical ac i i y and
dep ession de i e om c oss-sec ional designs [10]. P e ious
longi udinal o in e en ion designs co e ela i ely sho
ollow-up phases o physical ac i i y [10, 18]. P ospec i e
s udies examining he associa ion be ween li elong physical
ac i i y and dep essi e symp oms a e e y a e [8].
I has been s a ed ha mo e sophis ica ed me hods
a eneededwhens udying hede elopmen ande iology
o dep ession[19].I hasalsobeenshown ha people’s
heal h beha io s may ha e a ious de elopmen al endencies
ins ead o a one g ow h ajec o y [20, 21], and some o hese
beha io al p o iles may be mo e de imen al o heal h han
o he s. Fo ins ance, i is heo e ically possible ha physical
ac i i y’s decline owa ds la e li e [17] may s a ea lie and/o
bes eepe amongspeci icg oups hanino he s,and hus
he di e en physical ac i i y p o iles may associa e wi h dis-
inc heal h ou comes. The need o examina ions assessing
he de elopmen and decline o physical ac i i y has been
acknowledged [17, 22]. I is also essen ial o s udy whe he
long- e m physical ac i i y con ibu es o heal h o e and
abo e he concu en physical ac i i y. Fu he mo e, iden-
i ying people, whose beha io al endencies associa e wi h
dec eased heal h, migh be use ul o p o essionals who aim
o adjus li es yle in e en ions o a ge ed g oups. I is also
impo an o gain in o ma ion o he beha io al de e minan s
and con ibu o s o good heal h [23].
G ow h mix u e modeling p o ides an app op ia e
amewo k o s udying de elopmen al p ocesses [20, 24].
The key ad an age o such modeling is ha i allows o
he es ima ion o in e indi idual a iabili y in in aindi idual
pa e ns o change o e ime [20, 24]. Cu en app oaches
o g ow h modeling ha e also been ega ded as lexible
o ins ance in e ms o including pa ially missing da a,
unequally spaced measu emen poin s, nonno mally dis-
ibu ed o disc e ely scaled epea ed measu es, and linea ly
o nonlinea ly shaped ajec o ies [24]. These issues, o some
o hese, ypically eme ge in de elopmen al esea ch. I has
been deno ed ha hese issues p esen challenges o mo e
adi ional me hods [24].
Ex ensi e adjus men o po en ial con ounde s has been
lacking in many p e ious s udies ega ding physical ac i i y
andheal hou comes[16,18].De elopmen o dep ession
o en depends on he in e play o mul iple psychological and
li es yle ela ed ac o s [7] ha ideally should be conside ed
when examining associa ions be ween heal h beha io s and
dep ession [25–30]. Child’s ea ly emo ional expe iences (e.g.,
nega i e emo ionali y) may ha e e ec s on he de elopmen
o dep ession la e in li e [25]. Childhood amily’s socioe-
conomic s a us has been shown o be associa ed wi h he
de elopmen o heal h beha io s and well-being [26]. In
adul hood, occu ence o dep essi e episodes has been shown
o p edic subsequen dep ession [27]. I has been p o en ha
dep ession is ela ed o age [28], and women end o epo
highe le els o dep ession han men [28]. Adul hood socioe-
conomic posi ion, expe iences o social suppo , body mass
index, and smoking s a us a e also associa ed wi h dep ession
in adul hood [26, 29, 30].
We examined he po en ial he e ogenei y in physical
ac i i y ajec o ies in ela ion o dep essi e symp oms o e a
30-yea pe iod om childhood o adul hood in a popula ion-
based sample wi h six coho s and including eigh s udy
wa es. To ou knowledge, his is he i s s udy assessing
physical ac i i y and dep essi e symp oms in such design.
The analyses we e adjus ed o se e al isk ac o s o dep es-
sion [25–30]. Physical ac i i y measu emen s we e pe o med
du ing pa icipan s’ childhood and adul hood ( om he
age o 9 o 49), and dep essi e symp oms we e assessed in
pa icipan s’ adul hood (pa icipan s aged om 35 o 50).
To gain a comp ehensi e pic u e o he associa ion be ween
physical ac i i y and dep essi e symp oms, he associa ion
was s udied c oss-sec ionally, wi h espec o change and
wi h espec o long- e m ajec o ies. The objec i es o his
s udy we e (1) o explo e he po en ially dis inc ajec o ies o
physical ac i i y om childhood o adul hood, (2) o examine
whe he physical ac i i y was associa ed wi h c oss-sec ional,
sho - e m, and long- e m changes in dep essi e symp oms
in adul hood, and (3) o examine whe he he physical ac i i y
ajec o ies con ibu ed o dep essi e symp oms in adul hood
o a g ea e deg ee han adul hood physical ac i i y.
2. Me hods
2.1. S udy Design and Pa icipan s. The s udy pa icipan s
we e om he ongoing p ospec i e Ca dio ascula Risk in
Jou nal o Spo s Medicine 3
Table 1: Desc ip i e s a is ics o he o iginal sample (𝑛 = 1724–3596)∗.
Va iables Measu emen yea 𝑛Mean ±SD/% Range
Co a ia es
Age 1980 3596 10.44 ±4.99 3–18
Childhood nega i e emo ionali y 1980 3177 1.06 ±0.11 1–2
Pa en al educa ion†1980 3540 1.90 ±0.77 1–3
Pa en al income 1980 3453 4.80 ±1.94 1–8
Symp oms o dep ession 1992 2330 2.14 ±0.60 1–4.57
Symp oms o dep ession 1997 2099 2.15 ±0.67 1–4.57
Symp oms o dep ession 2001 2097 2.07 ±0.67 1–4.62
Symp oms o dep ession 2007 2056 2.06 ±0.68 1–4.67
Pa icipan s’ educa ion‡2007 2022 2.11 ±0.90 1–3
Pa icipan s’ income 2007 2146 3.50 ±1.56 1–8
Social suppo 2007 2055 4.15 ±0.80 1.08–5.00
Body mass index 2007 2170 26.00 ±4.75 16.56–58.82
Smoking s a us 2007 2224 3.81 ±1.53 1–5
Physical ac i i y indices§
Physical ac i i y 1980 2224 9.05 ±1.83 5–14
Physical ac i i y 1983 2116 9.03 ±1.88 5–14
Physical ac i i y 1986 2320 8.90 ±2.01 5–14
Physical ac i i y 1989 2619 8.63 ±2.10 5–14
Physical ac i i y 1992 2192 9.08 ±1.92 5–14
Physical ac i i y 2001 2442 8.86 ±1.96 5–15
Physical ac i i y 2007 2166 8.81 ±1.81 5–15
Physical ac i i y 2011 1910 9.02 ±1.88 5–15
Physical ac i i y ajec o y g oups§
Ligh ly physically ac i e 1980–2011 371 10.4%
Mode a ely physically ac i e 1980–2011 3046 85.5%
Highly physically ac i e 1980–2011 147 4.1%
Dependen a iable Symp oms o dep ession (BDI-II) 2012 1724 5.04 ±6.60 0–58
∗The o iginal sample size was 3596, and 1764 (49.1%) o he pa icipan s we e males and 1832 (50.9%) we e emales.
†The equencies o pa en s’ educa ional le els we e as ollows: low, 𝑛 = 1228 (34.7%), a e age, 𝑛 = 1428 (40.3%), high, and 𝑛 = 884 (25.0%).
‡The equencies o pa icipan s’ educa ional le els we e as ollows: low, 𝑛 = 713 (35.3%), a e age, 𝑛 = 376 (18.6%), high, and 𝑛 = 933 (46.1%).
§Physical ac i i y indices ≤7indica elow,>7 o10<mode a e, and ≥10 high le els o physical ac i i y. Fac o s sco es, which we e p edic ed om physical
ac i i y indices (1980–2011) (see Supplemen a y Table 3), we e used in LCGA.
Young Finns S udy ha began in 1980 [31].The o iginal sample
consis ed o 3596 child en and adolescen s (83.2% o hose
in i ed, 1832 emales and 1764 males) om six bi h coho s
(aged3,6,9,12,15,and18).Toob aina ep esen a i esample,
Finland was di ided in o i e a eas based on he loca ions o
uni e si ies wi h medical schools (Helsinki, Kuopio, Oulu,
Tampe e, and Tu ku), and he pa icipan s we e andomly
selec ed based on hei social secu i y numbe s om nea by
u ban and u al a eas. In o med consen was eques ed om
each pa icipan (o om he pa en s o small child en), and
he s udy was app o ed by he local e hics commi ees. The
s udy was conduc ed acco ding o Decla a ion o Helsinki
and Ame ican Psychological Associa ion’s e hical p inciples.
The sample was ollowed in 8 wa es, 1983, 1986, 1989, 1992,
1997, 2001, 2007-2008, and 2012, in which medical, psycho-
logical, and physical ac i i y s udies we e conduc ed. Physical
ac i i y om childhood o middle adul hood was assessed in
1980, 1983, 1986, 1989, 1992, 2001, 2007, and 2011, esponse
a e anging om 53.1% o 72.8% (𝑛 = 1910–2619)o he
o iginal s udy pa icipan s (Table 1). Dep essi e symp oms
we e measu ed in 2012 and 47.9% (𝑛 = 1724) o he o iginal
s udy subjec s pa icipa ed in he examina ion (Table 1).
P e ious s udies o sample a i ion ha e shown ha
he e has no been sys ema ic selec ion bias ega ding s udy
pa icipan s’ medical p o iles o physical ac i i y [31, 32],
bu some selec i e a i ion wi h espec o pe sonali y and
dep essi e symp oms exis s [33, 34]. Pa icipan s who we e
less sel -di ec ed and less ag eeable and had highe le els o
neu o icismaswellasdep essi esymp omshaddiscon inued
he s udy mo e o en han o he s [33, 34].
2.2. Measu es
2.2.1. Physical Ac i i y. Physical ac i i y was assessed wi h
sel -adminis e ed ques ionnai es. In o ma ion conce ning 3-
and 6-yea -old child en’s (bo n in 1974 and 1977) physical
ac i i y le els is missing om his s udy, because hey we e
no able osel - epo hei physicalac i i yle elsin1980.
Pa icipan s, om whom physical ac i i y was assessed, we e
aged om 9 o 18 in 1980 and om 34 o 49 in 2011.
F om 1980 o 1989, physical ac i i y ques ionnai es
consis ed o i e ques ions ocusing on he in ensi y and
equency o pa icipan s’ leisu e ime physical ac i i y,
pa icipa ion in spo s-club aining, pa icipa ion in spo s
4Jou nal o Spo s Medicine
compe i ions, and pa icipan s’ usual way o spending leisu e
ime [35]. F om 1992 ill p esen ime, ques ionnai es con-
sis ed o i e ques ions as well, assessing he in ensi y and
equency o leisu e ime physical ac i i y, hou s spen on
physical ac i i y pe week, a e age du a ion o a physical
ac i i y session, and pa icipa ion in o ganized physical
ac i i y [32]. F om 1980 o 1992, he answe s o he ques ions
we e coded in o 3 ca ego ies ( anging om 1 o 3), excluding
he i ems ha assessed pa icipa ion in spo s compe i ions
(1980–1989) and pa icipa ion in o ganized spo s (1992)
whichhada esponse ange om1 o2.F om2001 o2011,
all esponses o he ques ions we e coded in o 3 ca ego ies
( esponse scale anging om 1 o 3). A sum sco e (physical
ac i i y index) o ques ion esponses was c ea ed o each
pa icipan each yea (Table 1), highe sco es e lec ing highe
physical ac i i y le el. The index has been ound eliable and
alid [32].
2.2.2. Dep essi e Symp oms. Pa icipan s’ dep essi e symp-
oms we e assessed wi h Beck Dep ession In en o y II (BDI-
II)[36].Thesesymp omswe emeasu edin2012when he
pa icipan s we e aged om 35 o 50. BDI-II consis s o 21
symp oms wi h a se e i y ange om 0 (no symp oms) o
3 (se e e le el o dep essi e symp oms). A sum sco e o all
i ems was compu ed o each pa icipan (Table 1), and no
missing i ems we e allowed. The eliabili y es ima e (C on-
bach’s 𝛼) o he dep essi e symp om sco es was >0.90. BDI-II
has demons a ed o be a alid ins umen [36–38], and i has
been ega ded as an acknowledged s anda d in he measu e-
men o dep essi e mood [36–39]. I is applicable in clinical
and nonclinical con ex s, including in gene al popula ions
[36–39]. BDI-II co ela es highly wi h i s ea lie e sions,
including modi ied BDI [36, 38], which has also been con-
side ed as a alid measu e o assessing dep essi e symp oms
in gene al popula ions [34]. Fu he mo e, BDI-II co ela es
well wi h o he widely used scales o dep ession [38]. The
ins umen has been designed and also demons a ed o be a
use ul sc eening ool o po en ial dep essed cases [36, 38].
2.2.3. Co a ia es. Childhood, adul hood, and gene al (age,
sex, and body mass index) co a ia es we e con olled o
in his s udy (Table 1) [25–30]. Pa icipan s’ nega i e emo-
ionali y [25] was epo ed by he p ima y ca e ake ia six
ques ions e lec ing pa icipan s’ beha io in childhood (e.g.,
“The child hi s/kicks o he child en “acciden ally””), on a
scale om 1 ( ue) o 2 (no ue), and a e age o he i ems was
calcula ed o each pa icipan . As some o he pa icipan s
we e adolescen s in 1980, hei ca e ake s esponded o
his ques ion e ospec i ely. Pa icipan s’ pa en s’ socioeco-
nomic s a us was assessed ia wo indices, educa ional and
income le els [26]. Pa en s’ educa ional le el was de e mined
ia educa ional in o ma ion collec ed om pa icipan s’
mo he s’ and a he s’ (1 = below 9 yea s/comp ehensi e
school, 2 = 9 o 12 yea s/seconda y school, and 3 = o e
12 yea s/academic educa ion). I pa en s’ educa ional in o -
ma ion di e ed, we based pa en al educa ional s a us on
he in o ma ion collec ed om he pa en wi h he highe
educa ional le el. I educa ional in o ma ion was a ailable o
only one pa en , amily’s educa ional s a us was de e mined
using his/he educa ional in o ma ion. Family’s income le el
was a ed in an 8-poin scale [1 =<15 000 ma ks (∼2523 eu os)
and 8 =>100 000 ma ks (∼16819 eu os)].
Symp oms o pa icipan s’ adul hood dep ession we e
de e mined in 1992, 1997, 2001, and 2007 ia a modi ied
e sion o Beck Dep ession In en o y, e e ed o as a
modi ied BDI [34, 40]. I ems o he measu e we e a ed
in a 5-poin scale, and a e age o he i ems was compu ed
each yea o each pa icipan . Pa icipan s’ socioeconomic
s a us (2007) was de e mined ia wo indices; educa ion was
assessed iaa3-ca ego yscale(1=comp ehensi eschool,
2 = seconda y school, and 3 = academic le el) and income
le el ia an 8-poin scale (1 =<10 000 eu os and 8 =>70 000
eu os). Addi ionally, pa icipan s’ expe iences o social
suppo , body mass index, and smoking s a us measu ed
in 2007 we e con olled o [29, 30]. Social suppo was
de e mined ia a 12-ques ion in en o y [41] using a 5-poin
scale, and a mean sco e o he i ems was calcula ed o each
pa icipan . Pa icipan s’ smoking s a us was examined ia a
5-ca ego y scale (1 = smokes a ciga e e pe day o mo e, 2 =
smokes once in a week, 3 = smokes less han once in a week,
4 = has qui ed smoking, and 5 = has ne e smoked).
2.3. S a is ical Analyses. Physical ac i i y ques ionnai es,
which we e designed o child en and adolescen s (1980–
1989) and adul s (1992–2011) di e ed sligh ly in hei con en .
To assu e ha he indings o he p esen s udy we e based on
changes in physical ac i i y and no due o a measu emen
a i ac , a con i ma o y ac o model was used o examine
whe he he physical ac i i y indices consis ing o i e indi-
ca o a iables had measu emen and s uc u al in a iance
o e ime [42, 43]. Weigh ed leas squa es means and a iance
adjus ed (WLSMV) es ima ion was used o all analyses [43].
The goodness o i o scala in a iance was assessed wi h
compa a i e i index (CFI), Tucke -Lewis index (TLI), and
oo -mean squa e e o o app oxima ion index (RMSEA).
Fac o sco es de i ed om his examina ion we e used in
subsequen analyses.
Wi hin he g ow h mix u e modeling amewo k, La en
Class G ow h Analysis (LCGA) was used o explo e he
ajec o ies o physical ac i i y om childhood o adul hood.
LCGA cap u es in o ma ion abou de elopmen al p ocesses
a in e - and in aindi idual le els, de ec ing subpopula ions
wi h dis inc g ow h ajec o ies [20]. The de e mina ion o
henumbe o subg oups o physicalac i i ywasbasedon
Akaike’s In o ma ion C i e ion (AIC) [44]. In addi ion, he
de e mina ion o he g oups was based on he classi ica ion
quali y es ima ions and p ac ical conside a ions [20, 45].
Wi hin he LCGA model, he a e age empo al ajec o ies
in he physical ac i i y g oups we e modeled by eg ession
equa ions, in which bo h he linea and quad a ic e ms we e
es ed o he independen a iable ( ime).
The associa ions be ween physical ac i i y ac o sco es
(assessed om age 9 o 49) and dep essi e symp oms (pa -
icipan s aged om 35 o 50) we e i s examined c oss-
sec ionally and longi udinally wi h linea eg ession analyses.
Due o he po en ial mul iple es ing p oblem, Bon e oni-
co ec ed 𝑝 alues (𝑝 < 0.003) we e used in de e min-
ing he signi ican associa ions. The ea e , he associa ions
Jou nal o Spo s Medicine 5
be ween he physical ac i i y ajec o y g oups and dep essi e
symp oms measu ed in 2012 we e examined wi h analyses
o a iance, and pos hoc es s we e also pe o med (Bon-
e oni’s me hod). Fu he mo e, we examined he longi u-
dinal associa ions be ween physical ac i i y le els assessed
in pa icipan s’ adul hood (2007, including pa icipan s aged
om 30 o 45) and dep essi e symp oms (2012) using a
linea eg ession. Due o he numbe o missing alues, he
a iance analyses and he eg ession analyses in which he
adul hood physical ac i i y (2007) was used as a p edic o
we e pe o med in ano he da ase which was impu ed using
he expec a ion-maximiza ion (EM) algo i hm [46]. Analy-
ses we e pe o med in s a is ical so wa e p og ams Mplus
( e sion 7.1 and e sion 7.2), IBM SPSS ( e sion 21), and S a a
( e sion 13). 𝑝 alues o <0.05 we e conside ed signi ican .
3. Resul s
Desc ip i es o he o iginal sample (𝑛 = 1724–3596)a e
shown in Table 1. Supplemen a y Tables 1and 2(see Supple-
men a y Ma e ial a ailable online a h p://dx.doi.o g/10.1155/
2016/8947375) p o ide desc ip i es o he sample in he
comple e (𝑛 = 648)andimpu ed(𝑛 = 3564–3596)da a,
espec i ely. Al hough he scala in a iance model o phys-
ical ac i i y did no demons a e s ong ac o ial in a iance
o e ime, he i o pa ial scala in a iance model was
adequa e (CFI = 0.90, TLI = 0.90, and RMSEA = 0.047), gi en
pa ial in a iance o he h eshold pa ame e s. Fo RMSEA,
alues <0.05 indica e a e y close model i , and CFI and TLI
alues close o 0.90 deno e an adequa e i [47]. Since he
pa ial scala in a iance model was conside ed accep able,
ac o sco es we e p edic ed o each subjec o be used in
subsequen analyses (Supplemen a y Table 3).
The p e equisi es o g ow h mix u e modeling (e.g., [24])
we e me sa is ac o ily. LCGA sugges ed ha a h ee- ac o
(g oup) solu ion was he bes i ing model o he da a
basedonAICindices(AIC=48915.44,48915.39,48901.43,
48902.85,and48907.08 o 1,2,3,4,and5g oups, esp.;a
lowe AIC implies a be e model). The classi ica ion quali y
o he h ee- ac o modelwasalsoadequa ebasedon hea e -
age p obabili y es ima es (g oup 1 = 0.74, g oup 2 = 0.86, and
g oup 3 = 0.71). Values >0.70 delinea e ha he g oup consis s
o indi iduals wi h simila pa e ns o change [45]. Linea
pa ame e es ima es we e ound o he physical ac i i y
g oups (Supplemen a y Table 4). Figu e 1 shows he cen al
endencies o he h ee g oups, ligh ly (𝑛 = 371), mode a ely
(𝑛 = 3046), and highly (𝑛 = 147)physicallyac i eg oups
byage.Basedon hees ima edma ginalmeans,pa icipan s’
physical ac i i y le els emained ela i ely unchanged om
childhood o adul hood in each g oup, al hough hese le els
appea ed o diminish minimally owa ds middle adul hood
in all pa icipan s (Figu e 1).
The eg ession analyses pe o med in he o iginal sample
(𝑛 = 255–1467) indica ed ha low le els o physical
ac i i y ac o sco es we e, in mos examined ages, associa ed
wi h highe le els o dep essi e symp oms in pa icipan s’
adul hood (𝑝 < 0.05), al hough some o he associa ions
a enua ed when Bon e oni-co ec ed 𝑝 alues (𝑝 < 0.003)
we e applied (Table 2).
−1.00
−0.80
−0.60
−0.40
−0.20
0.00
0.20
0.40
0.60
0.80
1.00
Highly physically ac i e g oup (obse ed means)
Highly physically ac i e g oup (es ima ed means)
Mode a ely physically ac i e g oup (obse ed means)
Mode a ely physically ac i e g oup (es ima ed means)
Ligh ly physically ac i e g oup (obse ed means)
Ligh ly physically ac i e g oup (es ima ed means)
Pa icipan s’ age (yea s)
494645434240393736343330272421181512
9
Physical ac i i y ( ac o sco es)
Figu e 1: Means o he highly physically ac i e (𝑛=147),mode a ely
physically ac i e (𝑛=3046),andligh lyphysicallyac i e(𝑛=371)
ajec o y g oups om childhood o middle adul hood.
The analyses o a iance in he o iginal sample (𝑛=
1722) indica ed ha he physical ac i i y ajec o y g oups
p edic ed symp oms o dep ession [𝐹(2,1719 = 8.12, 𝑝<
0.001,adjus ed𝑅2=0.01)].Pos hoc es sshowed ha
highly physically ac i e g oup had lowe le els o dep essi e
symp oms han ligh ly ac i e g oup (mean di e ence = −3.26;
𝑝 < 0.001,95%CI:−5.25 o −1.26). Highly physically ac i e
pa icipan s had lowe le els o dep ession in compa ison o
mode a ely ac i e ones (mean di e ence = −1.92, 𝑝 = 0.02,
95%CI: −3.66 o −0.18). Also mode a ely ac i e pa icipan s
had lowe le els o dep essi e symp oms han ligh ly ac i e
pa icipan s (mean di e ence = −1.33, 𝑝 = 0.02,95%CI:−2.48
o −0.19). Based on he analyses o a iance a e adjus ing
o he co a ia es [25–30], he associa ions a enua ed o non-
signi icance [𝐹(2, 631) = 2.13, 𝑝 = 0.12,adjus ed𝑅2=0.25].
In pa icula he p e ious symp oms o dep ession measu ed
in1992,1997,2001,and2007a enua ed heassocia ions( o
de ails, see Supplemen a y Table 5).
We also examined he unadjus ed associa ions in he
da a wi h ull in o ma ion on all s udy a iables (𝑛 = 648),
in which case he associa ion be ween he physical ac i i y
ajec o y g oups and dep essi e symp oms became only
ma ginally signi ican [𝐹(2,645) = 2.76, 𝑝 = 0.06,adjus ed
𝑅2= 0.01]. As he sample a i ion migh ha e a ec ed
his associa ion (i.e., by educing he s a is ical powe ), he
analyses we e also pe o med in a da ase which was impu ed
using EM-algo i hm (𝑛 = 3564–3596). This me hod was
applied as he Li le’s MCAR es [48] con i med ha he da a
we e no missing comple ely a andom (𝜒2= 3903.02, 𝑑𝑓 =
2923, 𝑝 < 0.001). In he impu ed da ase , physical ac i i y

6Jou nal o Spo s Medicine
Table 2: Physical ac i i y ac o sco es (assessed a pa icipan s’ ages
om 9 o 49) as p edic o s o symp oms o dep ession (pa icipan s
aged om 35 o 50) (𝑛 = 255–1467).
Pa icipan s’ age 𝑏SE Β𝑝
∗95% CI†
9−1.06 0.51 −0.07 0.040 −2.06 o −0.05
12 −1.73 0.39 −0.13 <0.001 −2.49 o −0.98
15 −1.72 0.30 −0.15 <0.001 −2.30 o −1.14
18 −1.21 0.29 −0.11 <0.001 −1.78 o −0.63
21 −1.33 0.34 −0.11 <0.001 −1.99 o −0.66
24 −1.53 0.30 −0.15 <0.001 −2.12 o −0.95
27 −1.52 0.35 −0.15 <0.001 −2.21 o −0.84
30 −0.86 0.35 −0.08 0.015 −1.56 o −0.17
33 −1.16 0.47 −0.10 0.014 −2.08 o −0.24
34 −0.51 0.51 −0.06 0.324 −1.52 o 0.50
36 −0.68 0.49 −0.06 0.161 −1.64 o 0.27
37 −0.89 0.60 −0.09 0.140 −2.07 o 0.29
39 −1.59 0.44 −0.15 <0.001 −2.46 o −0.73
40 −0.34 0.72 −0.03 0.643 −1.76 o 1.09
42 −1.96 0.60 −0.18 0.001 −3.14 o −0.79
43 −1.65 0.61 −0.16 0.007 −2.85 o −0.45
45 −1.31 0.57 −0.13 0.022 −2.44 o −0.19
46 −2.22 0.58 −0.21 <0.001 −3.35 o −1.09
49 −1.34 0.54 −0.14 0.014 −2.41 o −0.27
∗Bon e oni-co ec ed 𝑝 alues (𝛼 = 0.05/19, 𝑝 < 0.003)we eusedin
de e mining signi ican associa ions.
†CI: con idence in e al.
g oups p edic ed he symp oms o dep ession [𝐹(2,3561) =
16.74, 𝑝 < 0.001,adjus ed𝑅2=0.01].Pos hoc es sindica ed
ha highly physically ac i e g oup had lowe le els o dep es-
si e symp oms han ligh ly ac i e g oup (mean di e ence =
−2.74; 𝑝 < 0.001,95%CI:−3.91 o −1.56) (Figu e 2). Highly
physically ac i e pa icipan s had lowe le els o dep essi e
symp oms in compa ison o mode a ely ac i e ones (mean
di e ence = −1.59, 𝑝 = 0.001,95%CI:−2.60 o −0.57)
(Figu e 2). Also mode a ely ac i e pa icipan s had lowe
le els o dep essi e symp oms han ligh ly ac i e pa icipan s
(mean di e ence = −1.15, 𝑝 < 0.001,95%CI:−1.81 o −0.49)
(Figu e 2). When he co a ia es [25–30] we e adjus ed o , he
esul s became nonsigni ican [𝐹(2,3547) = 0.53, 𝑝 = 0.59,
adjus ed 𝑅2= 0.47]. In pa icula he p e ious symp oms
o dep ession measu ed in 1997 and 2001 a enua ed he
associa ions ( o de ails, see Supplemen a y Table 5).
The ea e , we examined he longi udinal associa ions
be ween adul hood physical ac i i y (assessed in 2007, pa -
icipan s’ aged om 30 o 45) and symp oms o dep ession
(assessed in 2012). In he o iginal sample, adul hood physical
ac i i y was associa ed wi h dep essi e symp oms (𝑏=−1.10,
𝑝 < 0.001,95%CI:−1.58 o −0.61, adjus ed 𝑅2=0.01).When
he co a ia es [25–30] we e adjus ed o , he associa ion
a enua ed o nonsigni icance (𝑏=−0.17, 𝑝 = 0.62,95%CI:
−0.85 o 0.51, adjus ed 𝑅2=0.24).Inpa icula hep e ious
symp oms o dep ession assessed in 2001 and 2007 a enua ed
he associa ion ( o de ails, see Supplemen a y Table 6).
The ea e , he analyses we e pe o med in a sample wi h
ull in o ma ion on all s udy a iables (𝑛 = 648), in which
case he adul hood physical ac i i y was no associa ed wi h
Highly physically ac i e g oup
Mode a ely physically ac i e g oup
Ligh ly physically ac i e g oup
Unadjus ed model∗
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
Symp oms o dep ession (means)
Adjus ed model∗†
Figu e 2: Symp oms o dep ession (2012) in he physical ac i i y
ajec o y g oups in unadjus ed∗andadjus edmodels
∗† (𝑛=3564).
∗S anda d e o s a e ep esen ed in he igu e by he e o ba s
a ached o each column. †Pa icipan s’ age, sex, childhood nega i e
emo ionali y, pa en al educa ion, pa en al income (1980), p e i-
ous symp oms o dep ession (1992–2007), pa icipan s’ educa ion,
income, social suppo , body mass index, and smoking s a us (2007)
we e adjus ed o in he model.
dep essi e symp oms (𝑏=−0.55, 𝑝 = 0.16,95%CI:−1.31 o
0.21, adjus ed 𝑅2= 0.002). Due o he sample a i ion, he
associa ion was s udied also in he impu ed da a (𝑛 = 3596),
and he esul s showed ha he adul hood physical ac i i y
was associa ed wi h dec eased le els o dep essi e symp oms
(𝑏=−1.07, 𝑝 < 0.001,95%CI:−1.36 o −0.78, adjus ed 𝑅2
= 0.01). When he co a ia es [25–30] we e adjus ed o , he
associa iona enua ed ononsigni icance(𝑏=0.06,𝑝 = 0.57,
95%CI: −0.16 o 0.28, adjus ed 𝑅2=0.47).Inpa icula he
symp oms o dep ession assessed in 2001 and 2007 a enua ed
he associa ion ( o de ails, see Supplemen a y Table 6).
4. Discussion
This s udy examined whe he dis inc ajec o ies o li elong
physical ac i i y exis ed in he da a and whe he physical
ac i i y was ela ed o dep essi e symp oms in adul hood.
We also s udied whe he he li elong physical ac i i y ajec-
o ies con ibu ed o he ou come o a g ea e deg ee han
adul hood physical ac i i y. This inspec ion was impo an ,
because he in o ma ion whe he li elong ajec o ies con-
ibu e o dep essi e symp oms o e and abo e he concu -
en (adul hood) physical ac i i y is lacking. LCGA e ealed
h ee dis inc g oups, he ligh ly, mode a ely, and highly
physically ac i e g oups. Physical ac i i y le els emained
ela i ely simila om childhood o adul hood in each
Jou nal o Spo s Medicine 7
g oup, al hough he g oups’ physical ac i i y le els dec eased
sligh ly owa ds middle adul hood. These esul s a e in line
wi h p e ious s udies demons a ing he decline o physical
ac i i y wi h age [17].
High physical ac i i y associa ed wi h lowe le el o
dep essi e symp oms in adul hood, which is in acco d wi h
p e ious esea ch [10, 18]. The mechanisms behind he
obse ed associa ion can be bo h physiological and psycho-
logical. Physical ac i i y may alle ia e dep essi e symp oms
h ough neu obiological al e a ions [12]. Physical ac i i y
may also ela e o enhanced sel -con idence, emo ion eg-
ula ion skills, and social capaci ies ha a e associa ed wi h
posi i e mood [13, 14].
The associa ion be ween high physical ac i i y and lowe
le els o dep essi e symp oms in adul hood, howe e , disap-
pea ed when co a ia es (age, sex, childhood nega i e emo-
ionali y, pa en s’ socioeconomic s a us, pa icipan s’ p e i-
ous symp oms o dep ession, pa icipan s’ socioeconomic
s a us, social suppo , body mass index, and smoking s a us)
we e aken in o accoun . The explana o y powe o he
ully adjus ed models was subs an ially highe in compa -
ison o heunadjus edones(46%highe in heimpu ed
da a). These same esul s we e ound when li elong physical
ac i i y ajec o ies and adul hood physical ac i i y (assessed
in 2007) we e used as p edic o s o dep essi e symp oms
in adul hood. Speci ically, physical ac i i y did no p edic
lowe le els o dep essi e symp oms in adul hood when he
expe iences o p e ious dep ession we e aken in o accoun .
The bidi ec ional na u e o he associa ion be ween physical
ac i i y and dep ession has been documen ed [8, 11]. Hence,
i is heo e ically possible ha he childhood symp oms o
dep ession p eceding ou i s measu emen ha e dec eased
he likelihood o ini ia ing physical ac i i y [8, 11] o ha o
main aining adequa e le el o physical ac i i y [11].
Summa izing, he s udy indica ed ha li elong physical
ac i i y ajec o ies o adul hood physical ac i i y le els we e
no associa ed wi h he p og ession o dep essi e symp oms
in adul hood. The s udy sugges s ha li elong physical ac i i y
his o y does no con ibu e o he p og ession o he dep es-
si e symp oms o a g ea e deg ee han adul hood physical
ac i i y.
4.1. Limi a ions and S eng hs. Ou esul s need o be in e -
p e ed in ligh o he ollowing limi a ions. Pa icipan s
didno p o idein o ma ion ega dingeach a iableac oss
he measu emen yea s, which diminished he comple e-
sample size conside ably. Howe e , we a emp ed o con ol
he po en ial a i ion bias ia an impu a ion me hod. The
measu emen s conce ning he pa icipan s’ physical ac i i y
ocused only on he sel - epo ed leisu e ime physical ac i -
i y, no o al physical ac i i y o ene gy expendi u e. Also he
dep essi e symp oms we e assessed wi h sel -adminis e ed
ques ionnai es, and hus he possibili y o subjec i e bias can-
no be excluded. Howe e , his is common in epidemiological
s udies by necessi y. To ou knowledge, compa ably long
ollow-up s udies using diagnos ic in e iews do no exis .
Fu he mo e, ea ly childhood dep ession was no assessed in
he s udy, bu his de iciency was e ie ed o an ex en by
con olling o he childhood nega i e emo ionali y [25].
The s eng hs o he s udy we e he p ospec i e,
popula ion-based s udy design, ela i ely la ge sample size,
use o LCGA, and u iliza ion o an ex ensi e se o co a ia es.
InLCGA, hemainad an ageis ha i allowsa esea che o
model de elopmen al p ocesses a in e - and in aindi idual
le els[20].Suchmodelinghasbeenshown obeespecially
use ul ega ding he de elopmen o heal h beha io s
[20, 21], and he need o u he s udies has been ecognized
[22].Howe e ,in hecaseo hiss udy, he adi ionalc oss-
sec ional and longi udinal s udies yielded o e y simila
esul s. This was no e iden ap io i, and he e o e ou s udy
adds e idence ega ding he e iology o dep ession [19] by
showing ha physical ac i i y ajec o ies appea no o play
a special ole o e and abo e he adul hood physical ac i i y.
Fu he mo e, dep essi e symp oms we e s udied wi h a
well- alida ed ins umen , BDI-II [36–39].
We we e able o s udy he whole a iance o pa icipan s’
dep essi e symp oms ins ead o ca ego ical diagnoses. This
is impo an when a emp ing o ind p e en i e ools o
dep ession, because unc ional impai men is much mo e
s ongly associa ed wi h symp om se e i y han wi h diag-
nos ic symp om coun [49]. Also he people who ha e
expe ienced symp oms o dep ession a e a isk o ge ing a
diagnosis [27].
4.2. Conclusions. This s udy iden i ied h ee dis inc physical
ac i i y g oups, he ligh ly, mode a ely, and highly physically
ac i e ones. Each g oup’s physical ac i i y le els emained
ela i ely unchanged om childhood o adul hood, al hough
he le els ended o diminish sligh ly owa ds la e adul hood.
Physical ac i i y was associa ed wi h dep essi e symp oms
in adul hood. Highly physically ac i e pa icipan s om
childhood o adul hood had lowe le els o dep essi e symp-
oms in adul hood compa ed o ligh ly physically ac i e
ones. Fu he mo e, pa icipan s’ adul hood physical ac i i y
assessed in 2007 was associa ed wi h dec eased le els o
dep essi e symp oms in adul hood. The associa ions be ween
physical ac i i y and dep essi e symp oms disappea ed when
he p eexis ing symp oms o dep ession we e con olled o ,
indica ing ha physical ac i i ies did no associa e wi h he
p og ession o dep essi e symp oms. Thus, he s udy sugges s
ha li elong physical ac i i y his o y does no con ibu e
o he p og ession o he dep essi e symp oms o a g ea e
deg ee han adul hood physical ac i i y. Ob aining in o ma-
ion o men al heal h his o y migh bene i clinicians and
o he p o essionals in e alua ing he ole o physical ac i i y
in well-being.
Disclosu e
The sponso s had no ole in s udy design, collec ion, analysis,
in e p e a ion o he da a, w i ing he epo , and decision o
submi he epo o publica ion.
Compe ing In e es s
The au ho s decla e ha he e is no con lic o in e es s
ega ding he publica ion o his pape .
8Jou nal o Spo s Medicine
Acknowledgmen s
The Young Finns S udy has ecei ed inancial suppo om
he Academy o Finland, Social Insu ance Ins i u ion o
Finland, Kuopio, Tampe e and Tu ku Uni e si y Hospi-
al Medical Funds, Juho Vainio Founda ion, Paa o Nu mi
Founda ion, Finnish Founda ion o Ca dio ascula Resea ch
and Finnish Cul u al Founda ion, Sig id Juselius Founda-
ion, Tampe e Tube culosis Founda ion, and Emil Aal onen
Founda ion. This s udy was also suppo ed by Academy
o Finland [G an nos. 258711, 265869, and 258578 (Liisa
Kel ikangas-J¨
a inen, Lau a Pulkki-R˚
aback, Kaisa Kase a,
Mi ka Hin sanen, and Tom Rosens ¨
om)], Signe and Ane
Gyllenbe g Founda ion (Liisa Kel ikangas-J¨
a inen, Mi ka
Hin sanen, and Lau a Pulkki-R˚
aback), Emil Aal onen Foun-
da ion (Mi ka Hin sanen), Alli Paasiki i Founda ion (Mi ka
Hin sanen), Juho Vainio Founda ion (Lau a Pulkki-R˚
aback),
Finnish Minis y o Educa ion and Cul u e (Tuija Tammelin,
Xiaolin Yang, and Kaisa Kase a), Finnish Cul u al Founda-
ion (Ch is ian Hakulinen), and U heiluopis os¨
a¨
a i¨
o(Kaisa
Kase a). I ina Lisinen and Mau i Niiniaho a e g a e ully
acknowledged o da a managemen .
Re e ences
[1] P. C. Dinas, Y. Kou edakis, and A. D. Flou is, “E ec s o exe cise
and physical ac i i y on dep ession,” I ish Jou nal o Medical
Science, ol.180,no.2,pp.319–325,2011.
[2]T.Vos,A.D.Flaxman,M.Nagha ie al.,“Yea sli edwi h
disabili y (YLDs) o 1160 sequelae o 289 diseases and inju ies
1990–2010: a sys ema ic analysis o he Global Bu den o
Disease S udy 2010,” The Lance , ol.380,no.9859,pp.2163–
2196, 2010.
[3] R.M.Ca neyandK.E.F eedland,“Dep ession,mo ali y,and
medical mo bidi y in pa ien s wi h co ona y hea disease,”
Biological Psychia y, ol.54,no.3,pp.241–247,2003.
[4] D. M. Cla ke and K. C. Cu ie, “Dep ession, anxie y and hei
ela ionship wi h ch onic diseases: a e iew o he epidemiology,
isk and ea men e idence,” Medical Jou nal o Aus alia, ol.
190, no. 7, pp. 54–60, 2009.
[5] K.W.Sax,S.M.S akowski,P.E.KeckJ .,V.H.Upadhyaya,S.
A. Wes , and S. L. Mcel oy, “Rela ionships among nega i e, pos-
i i e, and dep essi e symp oms in schizoph enia and psycho ic
dep ession,” B i ish Jou nal o Psychia y, ol.168,no.1,pp.68–
71, 1996.
[6] J. M. Donohue and H. A. Pincus, “Reducing he socie al bu den
o dep ession: a e iew o economic cos s, quali y o ca e and
e ec s o ea men ,” Pha macoEconomics, ol.25,no.1,pp.7–
24, 2007.
[7] R.F.Mu
˜
noz, P. Cuijpe s, F. Smi , A. Z. Ba e a, and Y. Leykin,
“P e en ion o majo dep ession,” Annual Re iew o Clinical
Psychology, ol. 6, pp. 181–212, 2010.
[8]S.M.Pe ei a,M.Geo oy,andC.Powe ,“In es iga iono
bidi ec ional associa ions be ween dep essi e symp oms and
physical ac i i y om age 11 o 50 yea s in he 1958 B i ish Bi h
Coho ,” The Lance , ol.382,supplemen 80,2013.
[9] D. K i z-Sil e s ein, E. Ba e -Conno , and C. Co beau,
“C oss-sec ional and p ospec i e s udy o exe cise and
dep essed mood in he elde ly: The Rancho Be na do s udy,”
Ame ican Jou nal o Epidemiology, ol.153,no.6,pp.596–603,
2001.
[10] M. Teychenne, K. Ball, and J. Salmon, “Physical ac i i y and like-
lihood o dep ession in adul s: a e iew,” P e en i e Medicine,
ol. 46, no. 5, pp. 397–411, 2008.
[11] M. Aze edo Da Sil a, A. Singh-Manoux, E. J. B unne e
al., “Bidi ec ional associa ion be ween physical ac i i y and
symp oms o anxie y and dep ession: he Whi ehall II s udy,”
Eu opean Jou nal o Epidemiology, ol.27,no.7,pp.537–546,
2012.
[12] I. Helmich, A. La ini, A. Sigwal e al., “Neu obiological al e -
a ions induced by exe cise and hei impac on dep essi e
diso de s,” ClinicalP ac iceandEpidemiologyinMen alHeal h,
ol. 6, pp. 115–125, 2010.
[13] L. C us , “A e iew and concep ual e-examina ion o men al
oughness: implica ions o u u e esea che s,” Pe sonali y and
Indi idual Di e ences, ol.45,no.7,pp.576–583,2008.
[14]E.McAuley,B.Blissme ,D.X.Ma quez,G.J.Je ome,A.F.
K ame , and J. Ka ula, “Social ela ions, physical ac i i y, and
well-being in olde adul s,” P e en i e Medicine, ol.31,no.5,
pp.608–617,2000.
[15] J.M.RuuskanenandI.Ruoppila,“Physicalac i i yandpsycho-
logicalwell-beingamongpeopleaged65 o84yea s,”Age and
Ageing, ol.24,no.4,pp.292–296,1995.
[16] R. Malina, “T acking o physical ac i i y ac oss he li espan,”
P esiden ’s Council on Physical Fi ness and Spo s Resea ch Diges
, ol.3,no.14,pp.1–8,2001.
[17] R. Telama and X. Yang, “Decline o physical ac i i y om you h
o young adul hood in Finland,” Medicine and Science in Spo s
and Exe cise, ol.32,no.9,pp.1617–1622,2000.
[18] X. Yang, M. Hi ensalo, M. Hin sanen e al., “Longi udinal
associa ions be ween changes in physical ac i i y and dep essi e
symp oms in adul hood: he Young Finns s udy,” In e na ional
Jou nalo Beha io alMedicine, ol.21,no.6,pp.908–917,2014.
[19] P. De Jonge and A. M. Roes , “Dep ession and ca dio ascula
disease: he end o simple models,” B i ish Jou nal o Psychia y,
ol. 201, no. 5, pp. 337–338, 2012.
[20] B. O. Mu h´
en and L. K. Mu h´
en, “In eg a ing pe son-cen e ed
and a iable-cen e ed analyses: g ow h mix u e modeling wi h
la en ajec o y classes,” Alcoholism: Clinical and Expe imen al
Resea ch, ol.24,no.6,pp.882–891,2000.
[21] N. R. Riggs, C.-P. Chou, C. Li, and M. A. Pen z, “Adolescen
o eme ging adul hood smoking ajec o ies: when do smoking
ajec o ies di e ge, and do hey p edic ea ly adul hood nico-
ine dependence?” Nico ine & Tobacco Resea ch, ol.9,no.11,
pp. 1147–1154, 2007.
[22] S. C. Duncan, T. E. Duncan, L. A. S ycke , and N. R. Chaume-
on, “A coho -sequen ial la en g ow h model o physical
ac i i y om ages 12 o 17 yea s,” Annals o Beha io al Medicine,
ol. 33, no. 1, pp. 80–89, 2007.
[23] N. M. Peel, R. J. McClu e, and H. P. Ba le , “Beha io al
de e minan s o heal hy aging,” Ame ican Jou nal o P e en i e
Medicine, ol. 28, no. 3, pp. 298–304, 2005.
[24]P.J.Cu an,K.Obeida ,andD.Losa do,“Twel e equen ly
asked ques ions abou g ow h cu e modeling,” Jou nal o
Cogni ion and De elopmen , ol. 11, no. 2, pp. 121–136, 2010.
[25] K. A. Dodge, “Social-cogni i e mechanisms in he de elopmen
o conduc diso de and dep ession,” Annual Re iew o Psychol-
ogy, ol. 44, no. 1, pp. 559–584, 1993.
[26] B.Galoba des,M.Shaw,D.A.Lawlo ,J.W.Lynch,andG.D.
Smi h, “Indica o s o socioeconomic posi ion (pa 1),” Jou nal
o Epidemiology and Communi y Heal h, ol.60,no.1,pp.7–12,
2006.
Jou nal o Spo s Medicine 9
[27] J. Ka s en, C. A. Ha man, J. H. Smi e al., “Psychia ic his o y
and sub h eshold symp oms as p edic o s o he occu ence o
dep essi e o anxie y diso de wi hin 2 yea s,” B i ish Jou nal o
Psychia y, ol. 198, no. 3, pp. 206–212, 2011.
[28] J. Mi owsky, “Age and he gende gap in dep ession,” Jou nal o
Heal h and Social Beha io , ol.37,no.4,pp.362–380,1996.
[29] A. Allg¨
owe ,J.Wa dle,andA.S ep oe,“Dep essi esymp oms,
social suppo , and pe sonal heal h beha io s in young men and
women,” Heal h Psychology, ol.20,no.3,pp.223–227,2001.
[30] F. S. Luppino, L. M. de Wi , P. F. Bou y e al., “O e weigh ,
obesi y, and dep ession: a sys ema ic e iew and me a-analysis
o longi udinal s udies,” A chi es o Gene al Psychia y, ol.67,
no. 3, pp. 220–229, 2010.
[31] O. T. Rai aka i, M. Juonala, T. R¨
onnemaa e al., “Coho p o ile:
he ca dio ascula isk in young Finns s udy,” In e na ional
Jou nal o Epidemiology, ol. 37, no. 6, pp. 1220–1226, 2008.
[32] X. Yang, R. Telama, M. Hi ensalo e al., “Leade ship com-
ponen o ype A beha io p edic s physical ac i i y in ea ly
midli e,” In e na ional Jou nal o Beha io al Medicine, ol.19,
no. 1, pp. 48–55, 2012.
[33] T. Rosens ¨
om, M. Jokela, C. R. Cloninge e al., “Associa ions
be ween dimensional pe sonali y measu es and p eclinical
a he oscle osis: he ca dio ascula isk in Young Finns s udy,”
Jou nal o Psychosoma ic Resea ch, ol.72,no.5,pp.336–343,
2012.
[34] T. Rosens ¨
om, M. Jokela, S. Pu onen e al., “Pai wise measu es
o causal di ec ion in he epidemiology o sleep p oblems and
dep ession,” PLoS ONE, ol.7,no.11,A icleIDe50841,2012.
[35] R. Telama, J. Viika i, and I. V¨
alim¨
aki, “A he oscle osis p e-
cu so s in Finnish child en and adolescen s. X. Leisu e- ime
physical ac i i y,” Ac a Paedia ica Scandina ica. Supplemen ,
ol.74,no.318,pp.169–180,1985.
[36] A. Beck, R. S ee , and G. B own, Manual o he Beck Dep ession
In en o y-II, Psychological Co po a ion, San An onio, Tex,
USA, 1996.
[37]E.A.S o ch,J.W.Robe i,andD.A.Ro h,“Fac o s uc u e,
concu en alidi y, and in e nal consis ency o he beck dep es-
sion in en o y—second edi ion in a sample o college s uden s,”
Dep ession and Anxie y, ol.19,no.3,pp.187–189,2004.
[38] Y.-P. Wang and C. Go ens ein, “Psychome ic p ope ies o he
Beck Dep ession In en o y-II: a comp ehensi e e iew,” Re is a
B asilei a de Psiquia ia, ol.35,no.4,pp.416–431,2013.
[39] A. Beck, R. S ee , and G. B own, BDI-II—Beck Dep ession
In en o y-II: Finnish S anda diza ion, Psychological Co po a-
ion and Psykologien Kus annus Oy, Helsinki, Finland, 2004.
[40] A.T.Beck,C.H.Wa d,M.Mendelson,J.Mock,andJ.E baugh,
“An in en o y o measu ing dep ession,” A chi es o Gene al
Psychia y, ol.4,pp.561–571,1961.
[41] G. Zime , N. Dahlem, S. Zime e al., “The mul idimensional
scale o pe cei ed social suppo ,” Jou nal o Pe sonali y Assess-
men , ol. 52, no. 1, pp. 30–41, 1988.
[42] R. J. Vandenbe g and C. E. Lance, “A e iew and syn hesis o he
measu emen in a iance li e a u e: sugges ions, p ac ices, and
ecommenda ions o o ganiza ional esea ch,” O ganiza ional
Resea ch Me hods, ol.3,no.1,pp.4–70,2000.
[43] L. Mu h´
en and B. Mu h´
en, Mplus Use ’s Guide,Mu h
´
en &
Mu h´
en, Los Angeles, Cali , USA, 6 h edi ion, 1998–2010.
[44] S. I. V ieze, “Model selec ion and psychological heo y: a
discussion o he di e ences be ween he Akaike in o ma ion
c i e ion (AIC) and he Bayesian in o ma ion c i e ion (BIC),”
Psychological Me hods, ol.17,no.2,pp.228–243,2012.
[45] H.And u ,N.Ca a o,A.Thompsone al.,“La en classg ow h
modelling: a u o ial,” Tu o ials in Quan i a i e Me hods o
Psychology, ol.5,no.1,pp.11–24,2009.
[46] A. P. Demps e , N. M. Lai d, and D. B. Rubin, “Maximum
likelihood om incomple e da a ia he EM algo i hm,” Jou nal
o he Royal S a is ical Socie y, ol.39,no.1,pp.1–38,1977.
[47] G. W. Cheung and R. B. Rens old, “E alua ing goodness-
o - i indexes o es ing measu emen in a iance,” S uc u al
Equa ion Modeling. A Mul idisciplina y Jou nal, ol.9,no.2,pp.
233–255, 2002.
[48] R. J. Li le, “A es o missing comple ely a andom o
mul i a ia e da a wi h missing alues,” Jou nal o he Ame ican
S a is ical Associa ion, ol. 83, no. 404, pp. 1198–1202, 1988.
[49] J. Ka s en, C. A. Ha man, J. O mel, W. A. Nolen, and B. W. J. H.
Penninx, “Sub h eshold dep ession based on unc ional impai -
men be e de ined by symp om se e i y han by numbe o
DSM-IV symp oms,” Jou nal o A ec i e Diso de s, ol.123,no.
1–3, pp. 230–237, 2010.