Ci a ion: Rico-González, M.;
Puche-O uño, D.; Clemen e, F.M.;
Aquino, R.; Pino-O ega, J. The Mos
Demanding Exe cise in Di e en
T aining Tasks in P o essional Female
Fu sal: A Mid-Season S udy h ough
P incipal Componen Analysis.
Heal hca e 2022,10, 838. h ps://
doi.o g/10.3390/heal hca e10050838
Academic Edi o : Ma ia Chia a
Gallo a
Recei ed: 24 Ma ch 2022
Accep ed: 28 Ap il 2022
Published: 2 May 2022
Publishe ’s No e: MDPI s ays neu al
wi h ega d o ju isdic ional claims in
published maps and ins i u ional a il-
ia ions.
Copy igh : © 2022 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license (h ps://
c ea i ecommons.o g/licenses/by/
4.0/).
heal hca e
A icle
The Mos Demanding Exe cise in Di e en T aining Tasks in
P o essional Female Fu sal: A Mid-Season S udy h ough
P incipal Componen Analysis
Ma kel Rico-González 1, Daniel Puche-O uño 2, Filipe Manuel Clemen e 3,4,5 , Rod igo Aquino 6,*
and JoséPino-O ega 2,7
1Depa men o Didac ics o Musical, Plas ic and Co po al Exp ession, Uni e si y o he Basque Coun y,
UPV-EHU, 48940 Leioa, Spain; [email p o ec ed]
2Depa men o Physical Ac i i y and Spo , Facul y o Spo Science, Uni e si y o Mu cia,
30100 Mu cia, Spain; [email p o ec ed] (D.P.-O.); [email p o ec ed] (J.P.-O.)
3
Escola Supe io Despo o e Laze , Ins i u o Poli écnico de Viana do Cas elo, Rua Escola Indus ial e Come cial
de Nun’Ál a es, 4900-347 Viana do Cas elo, Po ugal; [email p o ec ed]
4Ins i u o de Telecomunicações, Delegação da Co ilhã, 1049-001 Lisboa, Po ugal
5Resea ch Cen e in Spo s Pe o mance, Rec ea ion, Inno a ion and Technology (SPRINT),
4960-320 Melgaço, Po ugal
6LabSpo , Pos -G adua e P og am in Physical Educa ion, Cen e o Physical Educa ion and Spo s,
Fede al Uni e si y o Espí i o San o, Vi ó ia 29075-910, B azil
7BIOVETMED & SPORTSCI Resea ch G oup, Uni e si y o Mu cia, 30100 Mu cia, Spain
*Co espondence: [email p o ec ed]
Abs ac :
The con ex ual ac o s ela ed o aining asks can play an impo an ole in how a playe
pe o ms and, subsequen ly, in how a playe ains o ace a compe i ion. To da e, he e has been
no s udy ha has in es iga ed he mos demanding exe cise in di e en aining asks in emale
u sal. The e o e, his s udy aimed o de e mine he mos demanding e o s du ing di e en aining
asks in a coho s udy conduc ed in p o essional biological women u sal playe s using p inci-
pal componen analysis (PCA). A o al o 14 eli e women u sal playe s (age = 24.34
±
4.51 yea s;
heigh = 1.65
±
0.60 m; body mass = 63.20
±
5.65 kg) pa icipa ed in his s udy. Se en y ain-
ing sessions o an eli e p o essional women’s eam we e egis e ed o e i e mon hs (p e-season
and in-season). Di e en ypes o exe cises we e g ouped in o six clus e s: p e en i e exe cises;
analy ical si ua ions; exe cises in midcou ; exe cises in
3/4
o he cou ; exe cises in ull cou ; su-
pe io i ies/in e io i ies. Each exe cise clus e was composed o 5–7 p incipal componen s (PCs),
conside ing om 1 o 5 main a iables o ming each, explaining om 65 o 75% o he physical o al
a iance. A o al o 13–19 sub- a iables explained he playe s’ e o s in each aining ask g oup.
The i s PCs o explain he o al a iance o aining load we e as ollows: p e en i e exe cises (accel-
e a ions; ~31%); analy ical si ua ions (impac s; ~23%); exe cises in midcou (high-in ensi y e o s;
~28%); exe cises in 3/4o he cou (~27%) and supe io i ies/in e io i ies (~26%) (ae obic/anae obic
componen s); exe cises in ull cou (anae obic e o s; ~24%). The PCs ex ac ed om each exe cise
clus e p o ide e idence ha may assis esea che s and coaches du ing aining load moni o ing.
The desc ip i e alues o he aining load suppo a scien i ic base o assis coaches in he planning
o aining schedules.
Keywo ds: indoo oo ball; da a mining; da a educ ion echniques; women; load moni o ing
1. In oduc ion
Fu sal (also known as indoo oo ball) is an in e mi en exe cise, in which epea ed
e o s occu in e spaced by pe iods o sho eco e y [
1
,
2
]. High-demanding e o s a e
associa ed wi h u sal, conside ing ha , in his game, a lo o quick ansi ions o possession
o he ball occu wi h implica ions o physical demands [
3
]. In he speci ic case o women’s
Heal hca e 2022,10, 838. h ps://doi.o g/10.3390/heal hca e10050838 h ps://www.mdpi.com/jou nal/heal hca e
Heal hca e 2022,10, 838 2 o 13
socce , o al dis ance can a y be ween 1659 and 1547 m pe hal , while 41–42% o he
dis ance is co e ed be ween 6.1 and 12 km/h, 6–7% be ween 15.5 and 18.3 km/h and
4–5% abo e 18 km/h [
4
]. The maximal sp in speed in women’s u sal is be ween 26 and
31 km/h, depending on he hal [
4
]. The e o e, eli e women u sal playe s usually possess
high le els o maximal oxygen up ake (~58.7 mL/kg/min) [
4
] and sp in ing abili ies, since
hey can achie e imes o 1.6s, 2.98s and 4.16s in 10, 20 and 30 m sp in es s [5].
The e o e, unde s anding he aining con ex is impo an o cha ac e izing egula
s imuli on u sal playe s and iden i ying how hey a e managed o achie ing speci ic
goals [
6
]. Al hough he e is a possible non-linea e ec be ween load and indi idual
esponses o s imuli in u sal [
7
], i is expec ed ha speci ic de e minan s o in ensi y can ex-
plain some pe o mance and biological a ia ions [
8
,
9
]. As an example, a s udy conduc ed
on women socce playe s examined he changes in sali a y immunoglobin A, co isol and
uppe espi a o y ac in ec ion and es ed he ela ionship wi h aining in ensi y [
10
].
In his s udy [
10
], i was ound ha a highe aining in ensi y had signi ican ela ion-
ships wi h an inc ease in uppe espi a o y ac in ec ion, hus sugges ing he impac o
he aining p ocess on acu e physiological and immunological esponses. Simila ly, a
s udy conduc ed in women’s socce e ealed ha mode a e-in ensi y aining enhanced
cy okine pa e ns [11].
Na u ally, he con ex ual ac o s associa ed wi h he u sal compe i i e schedule [
12
],
momen o he season [
13
] and ma ch demands will modula e he aining p ocess. Thus,
i is expec ed ha di e en aining d ills occu ing in aining scena ios may esul in
di e en ou comes, and he de e minan s o exe cise can be di e en based on he u sal
d ill ypology [
6
,
14
,
15
]. The a ia ion in hose de e minan s should be unde s ood consid-
e ing he dose– esponse e ec . Despi e he impo ance o conside ing he dose– esponse
e ec o speci ic in e en ion p og ams on changes occu ing in i ness a iables in u sal
playe s [
16
,
17
], i is also impo an o unde s and he cha ac e is ics o he aining in ensi y
dis ibu ion in he mic ocycle [
18
] and he indi idual e ec s o speci ic aining exe cises
on he playe s [
19
]. Al hough he e has been some esea ch desc ibing he a ia ions in
in ensi y wi hin and be ween mic ocycles [
18
,
19
], i did no conside he speci ic impac o
ypes o exe cises and he con ibu ion o ypes o exe cises o he in ensi y imposed on
he playe s. Al hough no ound in women’s u sal, his cha ac e iza ion o he in ensi y
imposed by di e en ypes o exe cises was desc ibed in socce , sugges ing ha he highes
olume o aining and he g ea es locomo o demands (physical demands) imposed on
he playe s we e based on game-based d ills [20].
Desc ip i e analysis is impo an o unde s and how he in ensi y is dis ibu ed ac oss
he di e en aining d ill ypes. Howe e , ela ionships be ween d ills and locomo o
demands may imp o e he capaci y o unde s and he aining p ocess’s de e minan
ou comes. Conside ing he g ea numbe o a iables ypically ex ac ed by mic oelec-
omechanical sys ems, i is impo an o ind a educ ion echnique ha allows iden i ying
he mos impo an a iables o be analyzed. One o he app oaches o do ha is applying
p incipal componen analysis (PCA) [
21
] ha , in his con ex , may play a ole as a da a
educ ion echnique o highligh he mos demanding e o s. In male u sal playe s, PCA
allowed iden i ying ha using 3–4 p incipal componen s was he mos app op ia e ap-
p oach o de e mine playe s’ pe o mance in ma ches [
22
]. In he case o women’s u sal,
PCA allowed iden i ying 22 ex e nal and in e nal in ensi y measu es ha may be used o
in ensi y moni o ing du ing o icial ma ches [
23
]. Despi e hese indings ega ding PCA
in o icial ma ches, he e is no ela ed s udy (as a as we know) ha has applied PCA in
aining d ills and sessions o women u sal playe s. This may help o unde s and which
measu es a e impo an o de ec o moni o ing aining in ensi y in women’s u sal ain-
ing sessions. The e o e, he pu pose o his s udy was o de e mine he mos demanding
e o s du ing di e en aining asks in a coho s udy conduc ed in p o essional emale
u sal playe s conside ing a ange o local posi ioning sys em (LPS) (e.g., dis ances co e ed
a di e en eloci ies, accele a ions and decele a ions) and hea a e-based a iables.
Heal hca e 2022,10, 838 3 o 13
2. Ma e ials and Me hods
2.1. Design
The da a we e egis e ed om daily playe moni o ing in which he a hle es’ mo ion
was ou inely measu ed h oughou he season. Measu es we e ga he ed by a eal- ime
mo ion acking sys em which includes a local posi ioning sys em (LPS) de ice, based
on UWB echnology, and an ine ial measu emen uni (IMU; WIMU PROTM, RealT ack
Sys ems, Alme ia, Spain). In addi ion, one GARMIN HRM-PRO ches s ap (Ga min L d.,
Ola he, KS, USA) was loca ed below he ches . Da a we e sen o he WIMU PRO de ice,
which s o es all da a se s. WIMI PRO de ices we e daily a ached o he playe s’ uppe
back in a pocke a ached o a igh - i ing ga men , placed be ween he scapulae a he
T2–T4 le els o a oid unwan ed mo emen s. The igh - i ing ga men was he same o
each playe in each game. The same ou ine was implemen ed each day, and bo h p o ocol
and ask con ol we e daily checked by he same eam s a .
In his coho s udy, all asks conside ed h oughou he season du ing 70 aining
sessions ( om Sep embe o Feb ua y) we e clus e ed in o di e en g oups depending
on hei cha ac e is ics (see Table 1). Du ing hese aining sessions, an LPS and mic o-
elec omechanical sys em (MEMS) we e used o ex ac in o ma ion om mo e han 250
a iables, which may be summa ized as accele a ions/decele a ions, sho - ime e o s,
dis ance co e ed a di e en in ensi ies and physiological a iables such as hea a e.
Th ough a mul i a ia e educ ion echnique (PCA), he mo e ele an a iables we e ex-
ac ed, highligh ing he mos demanding e o s by p o essional emale u sal playe s in
each g oup o asks.
Table 1. T aining asks g ouped in o exe cise clus e s.
Clus e Cha ac e is ics Task Desc ip ion
1 P e en i e exe cises (indi idual)
- Plyome ic
- S eng h ci cui s
- Taba a
- P op iocep ion
- Eccen ic
- Reac i e eloci y
- Veloci y/accele a i e
2 Analy ical si ua ions
- Analy ical ac ical mo emen s wi h and wi hou opponen s
- Co ne s (wi hou opposi ion)
- Th ow-ins
- Sho on goal om ouls
- Sho on goal om he penal y spo
3 Exe cises in midcou (20 ×20 m) - Games (wi h and wi hou wildca d) om 3 s. 3 o 6 s. 6
- Co ne s a egy (exe cise wi h opposi ion s a ing om a co ne )
4 Exe cises in 3/4o he cou (20 ×28 m) - Games (wi h and wi hou wildca d) om 3 s. 3 o 6 s. 6
- Coun e a ack exe cises
5 Exe cises in ull cou (20 ×40 m) - Games (wi h and wi hou wildca d) om 3 s. 3 o 6 s. 6
- Coun e a ack exe cises
6 Supe io i ies/in e io i ies - Si ua ions wi h nume ical supe io i ies/in e io i ies
2.2. Pa icipan s
Da a we e collec ed om 14 p o essional biological women u sal playe s (means
±
s an-
da d de ia ions: age = 24.34
±
4.51; heigh = 1.65
±
0.60 m; body mass = 63.20
±
5.65 kg)
om a Spanish 1s di ision op-le el eam du ing he 2020–2021 p e-season and in-season
Heal hca e 2022,10, 838 4 o 13
pe iods ( om Sep embe o Feb ua y). This sample was conside ed due o he high num-
be o indi idual obse a ions and he eal-wo ld scien i ic p ac ice in high-le el se ings.
Playe s we e allowed o pa icipa e in a aining session i hey we e heal hy and hey did
no ha e inju ies o did no pa icipa e in a pos -inju y eadap a ion p og am. Due o he
c i ical si ua ion as a esul o se e e acu e espi a o y synd ome (SARS-CoV-2), playe s
we e excluded om he analysis when hey we e con ined. All playe s we e no i ied o
he aim o he s udy and p ocedu es in acco dance wi h he Decla a ion o Helsinki. The
s udy was app o ed by he e hical commi ee o he Uni e si y o Mu cia (p o ocol code
3180/2020).
2.3. Da a Collec ion
To conduc a s ic desc ip ion o he use o echnology, a ecen ly published su ey
was ollowed [
24
]. Fo he use o an ul a-wideband (UWB)-based LPS, 21 poin s ou o 23
we e explained, while o he use o a MEMS, 17 poin s ou o 20 we e de ailed. The es o
he i ems canno be explained because he in o ma ion was no a ailable o he au ho s.
2.4. Measu es
The UWB echnology ope a es on a much wide equency band han o he adi ional
adio communica ion echnologies (a leas 0.5 GHz), and a p e ious s udy did no e-
po any p oblems in UWB-based acking sys em accu acy in mul ipa h condi ions (i.e.,
28 de ices u ned on) (Bas ida Cas illo e al., 2018). This sys em holds bo h FIFA In e na-
ional Ma ch S anda d and Quali y ce i ica es [
25
]. Each de ice consis s o an in e nal
mic op ocesso , 2 GB lash memo y and a high-speed USB in e ace, o eco d, s o e and
upload da a.
2.4.1. Ul a-Wideband
F om he UWB dis ance co e ed a di e en in ensi ies, eloci y and load indica o s
we e egis e ed (see Table 2). The da a we e eco ded in a aining space a om me allic
ma e ials. The UWB sys em was composed o a e e ence sys em and acked de ices
ca ied by he playe s. The an ennae a e ansmi e s and ecei e s o adio equency
signals. The an ennae (mainly he mas e an enna) compu e ize he posi ion o he de ices
ha a e in he play a ea, while he de ice ecei es ha calcula ion using ime di e ence o
a i al (TDOA). The eigh an ennae we e ins alled i e hou s be o e he ma ch, o ming an
oc agon o be e signal emission (4.5 m om he pe ime e line o an ennae loca ed in
he co ne s, and 5.5 m om he pe ime e line o he an ennae loca ed in he middle o he
cou and behind he goals) and ecep ion a a heigh o 3 m, and held by a ipod. Once
ins alled, hey we e swi ched on one by one, wi h he mas e an enna u ned on las . F om
ha momen , i was necessa y o espec a 5-min p o ocol o a oid echnology lock [
24
]. To
allow da a ime synch oniza ion, he mas e an enna managed he ime using a common
clock which allows da a eco ding a he same ime. When all de ices we e swi ched on in
he cen e o he e e ence sys em, a p ocess o au oma ic ecogni ion be ween an ennae
and de ices was ca ied ou o 1 min. In his s udy, he aw da a we e eco ded a an 18 Hz
sampling equency because low equencies ha e been shown o ha e a lowe quali y o
measu emen , and 18 Hz wi h UWB has no shown less accu acy because o noise p oblems.
The condi ions we e main ained wi h low empe a u es, humidi y g adien s and slow ai
ci cula ion o allow easie posi ioning. This UWB sys em has demons a ed alid and
eliable measu es du ing con inuous si ua ions [26].
Heal hca e 2022,10, 838 5 o 13
Table 2. Ex e nal and in e nal aining load a iables.
Abb e ia ion Sub-Va iables
1Dis ance co e ed a di e en
in ensi ies
Dis (m·min−1)Dis ance co e ed
Expl dis Dis ance co e ed a explosi e in ensi y
HSR Abs (m·min−1)High-speed unning
HIBD (m·min−1) High-in ensi y b eak dis ance wi h DEC > 2 m·s−2
2Hea a e- ela ed a iables
HR % (50–60) Time spen om 50 o 60 % o maximum hea a e
HR % (70–80) Time spen om 70 o 80 % o maximum hea a e
HR % (80–90) Time spen om 80 o 90 % o maximum hea a e
HR % (90–95) Time spen om 90 o 95 % o maximum hea a e
HR % (>95) Time spen up o 95 % o maximum hea a e
MAX HR (bpm) Maximum hea a e achie ed
Rel HR % % o hea a e sco ed pe minu e
3Veloci ies
Vel Abs (0–6) (m·min−1) Absolu e dis ance co e ed om 0 o 6 m·min−1
Vel Abs (18–21) (m·min−1) Absolu e dis ance co e ed om 18 o 21 m·min−1
Vel Abs (21–24) (m·min−1) Absolu e dis ance co e ed om 21 o 24 m·min−1
Vel Max Maximum eloci y achie ed by a playe
4Accele a ions
Acc/min Numbe o accele a ions pe minu e
Dis Acc Dis ance accele a ing
MAX Acc (m·s2)Maximum accele a ion
Acc Abs (0–1)/min Absolu e accele a ions lowe han 1 pe minu e
Acc Abs (1–2)/min Absolu e accele a ions om 1 o 2 pe minu e
Acc Abs (3–4)/min Absolu e accele a ions om 3 o 4 pe minu e
Acc Abs (4–5)/min Absolu e accele a ions om 4 o 5 pe minu e
Acc Abs (5–6)/min Absolu e accele a ions om 5 o 6 pe minu e
Acc Abs (6–10)/min Absolu e accele a ions om 6 o 10 pe minu e
5Decele a ions
Dec Abs (−1, 0)/min Absolu e decele a ions lowe han 1 pe minu e
Dec Abs (−2, −1)/min Absolu e decele a ions om 1 o 2 pe minu e
Dec Abs (−4, −3)/min Absolu e decele a ions om 3 o 4 pe minu e
Dec Abs (−5, −4)/min Absolu e decele a ions om 4 o 5 pe minu e
Dec Abs (−6, −5)/min Absolu e decele a ions om 5 o 6 pe minu e
Dec Abs (−10, −6)/min Absolu e decele a ions om 6 o 10 pe minu e
6Impac s
Impac s (0–3) G Impac s a in ensi y lowe han 3 G
Impac s (3–5) G Absolu e decele a ions om 3 o 5 G
Impac s (0–3)/min F om 0 o 3 impac s pe minu e
Impac s (3–5)/min F om 3 o 5 impac s pe minu e
Impac s (5–8)/min F om 5 o 8 impac s pe minu e
7 Landings Landing (5–8)/min F om 5 o 8 landings pe minu e
8Load indica o s
Playe Load /min Playe load ex ac ed om accele ome e ’s 3 axes
Powe Me abolic Ene gy consumed kg·s
HML (10–25.5) (m)
Dis ance co e ed om 10 o 25.5 m a 25.5 W/kg,
which co esponds o 5.5 m/s2o signi ican
accele a ion/decele a ion e o s
HML (25.5–35) (m)
Dis ance co e ed om 25.5 o 35 m a 25.5 W/kg,
which co esponds o 5.5 m/s2o signi ican
accele a ion/decele a ion e o s
DSL/min Dynamic s eng h load. To al impac s wi h high
in ensi y o 2 G.
HBD (m·min−1)
No e: G = 9.8 m/s2.
2.4.2. Ine ial Measu emen Uni s
F om he IMU, accele a ion, decele a ion, impac and landing- ela ed a iables we e
egis e ed (see Table 2). The alidi y o he ine ial measu emen uni (IMU; WIMU PROTM,
Real T ack Sys ems, Alme ia, Spain) was assessed in a p e ious s udy [
27
]. The esul s
o he alidi y in his s udy we e sa is ac o y and had a bias o 0.0006
±
0.0018 s. The
Heal hca e 2022,10, 838 6 o 13
calcula ion o he eloci y was conduc ed h ough di e en ial Dopple and he accele a ion
was calcula ed om eloci y. Finally, he minimum e o du a ion and minimum speed o
a oiding un ealis ic da a men ioned we e de ined by he manu ac u e o a oid ou lie s.
2.5. Pe o med T aining Task Clus e s and Va iables
Posi ional da a eco ded du ing aining asks (Table 1) ex ac ed om PCA a e sum-
ma ized in Table 2.
The aining ask clus e ing ollowed a echnical classi ica ion p oposed by eam
s a , which ocused on he a ionale o he classi ica ion. Fu he , he ask clus e s we e
es ablished depending on he e o s in each o hem. In clus e numbe one, wa m-
up exe cises we e classi ied, which playe s pe o med a he beginning o each aining
session. In clus e numbe wo, he ha bo exe cises we e es ablished (i.e., low physical
s ess), whe e eam s a hal ed he exe cise o explain ac ical in o ma ion. Clus e s h ee,
ou and i e ocused he classi ica ion’s a ionale on he numbe o playe s in ol ed
(indi idual aining asks, pu e opposi ion, opposi ion and collabo a ion and opposi ion
and collabo a ion wi h nume ical unbalances). Addi ionally, asks wi h opposi ion and
collabo a ion ( om 3 s. 3 o 6 s. 6) we e classi ied in o 3 clus e s depending on space
(i.e., midcou ,
3/4
o he ull cou and ull cou ). These clus e s included he mos
ac ical and echnical aining en ichmen exe cises o he u sal playe s, close o he eal
si ua ions. Finally, clus e numbe six e e s o all hese exe cises, usually wi h a high-s ess
si ua ion, whe e playe s wi h supe io i y/in e io i y need o make decisions wi h high
a igue and isk.
The aining exe cises non-cha ac e ized by ask cons ain s we e excluded om he
analysis. In addi ion, abb e ia ions and de ini ions om he a iables ex ac ed om
egis e ed aining asks and ha o med PCs a e desc ibed in Table 2. See Figu e 1 o
de ails ega ding he expe imen al p o ocol.
Heal hca e 2022, 10, x FOR PEER REVIEW 7 o 15
Figu e 1. The expe imen al p o ocol and me hod.
2.6. S a is ical Analysis
To desc ibe he p o ocol o PCA, he su ey pu posed by Rojas-Val e de e al. [28]
was ollowed. The p o ocol was explained s ep by s ep in Oli a-Lozano e al. [29], which
con ains he ollowing s eps: ield da a collec ion, so wa e ou come (inclusion o all a i-
ables measu ed), co ela ion ma ix explo a ion (all a iables co ela ed be ween each
o he ), assump ion con i ma ion ( a iables we e cen e ed and scaled (Z-Sco e), sui abili y
con i med by KMO and Ba le sphe ici y alues), PCA (eigen alues g ea e han 1 we e
included o ex ac ion), loading g oupings in each PC (loadings > 0.7 we e conside ed o
PCA g ouping and inclusion) and inal PCA ou come (each PCA was p esen ed and
g ouped by loading) [29]. F om mo e han 250 a iables eco ded, 12 (clus e 1), 32 (clus e
2), 32 (clus e 3), 28 (clus e 4), 27 (clus e 5) and 31 (clus e 6) a iables we e explo ed
using a co ela ion ma ix in o de o selec he mos ep esen a i e a iables, and hose
wi h co ela ions < 0.7 be ween a iables we e conside ed o ex ac ion [30]. A e ex-
clusion o a iables wi h a iance = 0, 22-21 a iables we e scaled and cen e ed using Z-
Sco es. PCA sui abili y was con i med h ough he Kaise –Meye –Olkin alue (KMO =
0.64–0.78) and Ba le sphe ici y es signi icance (p < 0.05) [31]. Eigen alues > 1 we e con-
side ed o he ex ac ion o each p incipal componen , and a a imax o hogonal o a-
ion me hod was used o iden i y high co ela ions be ween componen s o o e di e en
in o ma ion. PC loadings > 0.6 we e conside ed o ex ac ion, and when a c oss-loading
was ound be ween PCs, only he highes ac o loading was e ained [31].
3. Resul s
F om 250 a iables ex ac ed om he acking sys em and MEMS, 8 main a iables
and load indica o s we e ex ac ed, which con ain a o al o 42 sub- a iables: (1) dis ance
co e ed a di e en in ensi ies (Dis . (m·min
−1
), Expl dis and HSR Abs (m/min)); (2) hea
a e- ela ed a iables (maximum HR, ela i e HR and ime spen a hea a es o di e en
in ensi ies); (3) eloci y a di e en in ensi ies ( om 0 o 6, om 18 o 21, om 21 o 24
and maximum eloci y); (4) a iables ela ed o accele a ions (Acc/min, dis ance accele -
a ing, maximum accele a ion and absolu e accele a ions a di e en in ensi ies ( i e
Figu e 1. The expe imen al p o ocol and me hod.
2.6. S a is ical Analysis
To desc ibe he p o ocol o PCA, he su ey pu posed by Rojas-Val e de e al. [
28
]
was ollowed. The p o ocol was explained s ep by s ep in Oli a-Lozano e al. [
29
], which
con ains he ollowing s eps: ield da a collec ion, so wa e ou come (inclusion o all
Heal hca e 2022,10, 838 7 o 13
a iables measu ed), co ela ion ma ix explo a ion (all a iables co ela ed be ween each
o he ), assump ion con i ma ion ( a iables we e cen e ed and scaled (Z-Sco e), sui abili y
con i med by KMO and Ba le sphe ici y alues), PCA (eigen alues g ea e han 1 we e
included o ex ac ion), loading g oupings in each PC (loadings > 0.7 we e conside ed
o PCA g ouping and inclusion) and inal PCA ou come (each PCA was p esen ed and
g ouped by loading) [
29
]. F om mo e han 250 a iables eco ded, 12 (clus e 1), 32
(clus e 2), 32 (clus e 3), 28 (clus e 4), 27 (clus e 5) and 31 (clus e 6) a iables we e
explo ed using a co ela ion ma ix in o de o selec he mos ep esen a i e a iables,
and hose wi h co ela ions < 0.7 be ween a iables we e conside ed o ex ac ion [
30
].
A e exclusion o a iables wi h a iance = 0, 22–21 a iables we e scaled and cen e ed
using Z-Sco es. PCA sui abili y was con i med h ough he Kaise –Meye –Olkin alue
(KMO = 0.64–0.78) and Ba le sphe ici y es signi icance (p< 0.05) [
31
]. Eigen alues > 1
we e conside ed o he ex ac ion o each p incipal componen , and a a imax o hogonal
o a ion me hod was used o iden i y high co ela ions be ween componen s o o e
di e en in o ma ion. PC loadings > 0.6 we e conside ed o ex ac ion, and when a
c oss-loading was ound be ween PCs, only he highes ac o loading was e ained [31].
3. Resul s
F om 250 a iables ex ac ed om he acking sys em and MEMS, 8 main a iables
and load indica o s we e ex ac ed, which con ain a o al o 42 sub- a iables: (1) dis ance
co e ed a di e en in ensi ies (Dis . (m
·
min
−1
), Expl dis and HSR Abs (m/min)); (2) hea
a e- ela ed a iables (maximum HR, ela i e HR and ime spen a hea a es o di e en
in ensi ies); (3) eloci y a di e en in ensi ies ( om 0 o 6, om 18 o 21, om 21 o 24 and
maximum eloci y); (4) a iables ela ed o accele a ions (Acc/min, dis ance accele a ing,
maximum accele a ion and absolu e accele a ions a di e en in ensi ies ( i e le els)); (5)
decele a ing- ela ed a iables (six le els o decele a ions a di e en in ensi ies); (6) impac s
(measu ed in bo h G (9.8 m
·
s
2
) and numbe pe minu e ( ou le els)); (7) a a iable ela ed
o landing ( om 5 o 8 landings pe minu e); and (8) six load indica o s (playe load, powe
me abolic, HML ( wo le els), DSL, HBD and HIBD). Howe e , he con ibu ions o hese
sub- a iables o explaining playe s’ ex e nal and in e nal load esponses di e ed be ween
ypes o exe cises (Table 3). Desc ip i e s a is ics (e.g., mean, median, s anda d de ia ion,
pe cen ile) o each a iable ex ac ed om PCA o each exe cise clus e a e desc ibed in
Supplemen a y File S1.
Table 3.
Rep esen a ion o p incipal componen analysis om ex e nal and in e nal load esponses in
di e en aining asks pe o med by eli e p o essional emale u sal playe s du ing mid-season.
PC 1234567
Clus e 1 (p e en i e exe cises)
Eigen alue 11.5 9.8 7.0 5.7 5.2
% Va iance 31.4 42.9 52.8 59.9 65.6 70.7
HR % (50–60) 0.798
HR % (70–80) 0.771
HR % (90–95) 0.792
MAX Acc (m/s2)0.766
Acc Abs (2–3)/min 0.780
Acc Abs (4–5)/min 0.798
Acc Abs (5–6)/min 0.759
Dec Abs (−6, −5)/min 0.716
Impac s (0–3) G 0.850
Impac s (0–3) min
Impac s (3–5)/min 0.874
Impac s (5–8)/min 0.917
Impac s (8–100)/min 0.950
Landing (5–8)/min 0.729
Heal hca e 2022,10, 838 8 o 13
Table 3. Con .
PC 1234567
Clus e 2 (analy ical si ua ions)
Eigen alue 13.8 10.5 8.2 6.9 5.3
% Va iance 23.4 37.3 47.7 55.9 62.8 68.1
HR % (50–60) 0.736
HR % (80–90) 0.817
DSL/min 0.766
Vel Abs (0–6)
(m·min−1)0.705
Acc/min 0.903
Acc Abs (0–1)/min 0.903
Acc Abs (1–2)/min 0.737
Acc Abs (2–3)/min 0.766
Acc Abs (4–5)/min 0.799
Dec Abs (−5, −4)/min 0.801
Impac s (0–3) G 0.926
Impac s (3–5) G 0.734
Impac s (0–3) min 00.752
Impac s (5–8)/min 0.855
Clus e 3 (exe cises in midcou )
Eigen alue 14.9 9.4 7.3 6.4 4.9 4.2
% Va iance 28.4 43.3 52.8 60.1 66.5 71.4 75.6
Expl dis (m) 0.867
HR % (50–60) 0.736
HR % (80–90) 0.720
HR % (90–95) 0.802
HR % (>95) 0.880
HIBD (m·min−1)0.859
Vel Abs (18–21)
(m·min−1)0.803
Acc/min 0.900
Dis Acc .726
MAX Acc (m/s2)0.730
Acc Abs (3–4)/min 0.708
Acc Abs (5–6)/min 0.738
Acc Abs (6–10)/min 0.840
Dec Abs (−1, 0)/min
Dec Abs (−2, −1)/min 0.919
Impac s (0–3) G 0.956
Impac s (3–5) G 0.949
Impac s (0–3) min 0.775
Impac s (5–8)/min 0.837
Clus e 4 (exe cises in 3/4o he cou )
Eigen alue 14.0 10.8 7.1 5.3 4.6 4.0
% Va iance 26.9 40.8 51.6 58.7 64.0 68.6 72.6
Expl dis (m) 0.725
HSR Abs (m·min−1)0.807
HR % (50–60) 0.713
HR % (70–80) 0.842
HR % (80–90) 0.764
Acc/min 0.885
Heal hca e 2022,10, 838 9 o 13
Table 3. Con .
PC 1234567
Acc Abs (2–3)/min 0.700
Acc Abs (3–4)/min 0.882
Acc Abs (4–5)/min 0.760
Acc Abs (6–10)/min 0.852
Dec Abs (−1, 0)/min 0.854
Impac s (3–5) G 0.802
Impac s (5–8)/min 0.928
Playe Load /min 0.816
Powe Me abolic (kg
·
s)
0.779
Clus e 5 (exe cises in ull cou )
Eigen alue 16.3 10.9 7.7 6.2
% Va iance 24.5 40.8 51.8 59.5 65.8
Expl dis (m) 0.920
HML (10–25.5) (m) 0.911
HML (25.5–35) (m 0.861
HBD (m/min) 0.854
Vel Abs (18–21)
(m·min−1)0.748
Acc/min 0.839
Acc Abs (5–6)/min 0.714
Acc Abs (6–10)/min 0.854
Dec Abs (−2, −1)/min 0.909
Dec Abs (−6, −5)/min 0.783
Dec Abs (−10,
−6)/min 0.704
Impac s (0–3) G 0.880
Impac s (5–8)/min 0.944
Clus e 6 (supe io i ies/in e io i ies)
Eigen alue 14.0 10.7 8.9 5.7 5.1 4.2
% Va iance 26.4 40.4 51.1 60.0 65.7 70.9 75.1
Expl dis (m) 0.700
Dis (m·min−1)0.778
HR % (50–60) 0.745
HR % (70–80) 0.765
MAX HR (bpm) 0.795
Rel HR % 0.885
Vel Abs (18–21)
(m·min−1)0.867
Acc/min 0.781
Dis Acc 0.727
MAX Acc (m·s2)0.727
Acc Abs (0–1)/min 0.852
Acc Abs (2–3)/min 0.799
Acc Abs (3–4)/min 0.864
Acc Abs (4–5)/min 0.763
Dec Abs (−4, −3)/min 0.752
Dec Abs (−5, −4)/min 0.743
Impac s (0–3) min 0.787
The p e en i e exe cise clus e (exe cise clus e 1) was composed o 6 PCs o med by
14 a iables, which explained a o al o 70% o playe s’ beha io du ing hese exe cises. In
his case, he i s PC was composed o i e sub- a iables ela ed o accele a ions: MAX
Acc (m
·
s
2
), Acc Abs (2–3)/min, Acc Abs (4–5)/min, Acc Abs (5–6)/min and Dec Abs (
−
6,
−
5)/min. In addi ion, clus e 2 was composed o 6 PCs o med by 14 a iables, which
explained a o al o 68% o playe s’ physical beha io du ing hese exe cises. In his case,
