Electromyography in the Four Competitive Swimming Strokes: a Systematic Review ⇑ Jonas Martens A, , Pedro Figueiredo B,C, Daniel Daly A

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Lirias

Journal of Electromyography and Kinesiology 25 (2015) 273–291

Contents lists available at ScienceDirect

Journal of Electromyography and Kinesiology

journal homepage: www.elsevier.com/locate/jelekin

Review Electromyography in the four competitive swimming strokes: A systematic review ⇑ Jonas Martens a, , Pedro Figueiredo b,c, Daniel Daly a

a KU Leuven, Faculty of Kinesiology and Rehabilitation Sciences, Department of Kinesiology, Leuven, Belgium b School of Physical Education, Federal University of Rio Grande do Sul, Brazil c Centre of Research, Education, Innovation and Intervention in Sport, Faculty of Sport, University of Porto, Portugal

article info abstract

Article history: The aim of this paper is to give an overview on 50 years of research in electromyography in the four Received 12 November 2014 competitive swimming strokes (crawl, breaststroke, butterfly, and backstroke). A systematic search of Received in revised form 9 December 2014 the existing literature was conducted using the combined keywords ‘‘swimming’’ and ‘‘EMG’’ on studies Accepted 9 December 2014 published before August 2013, in the electronic databases PubMed, ISI Web of Knowledge, SPORT discus, Academic Search Elite, Embase, CINAHL and Cochrane Library. The quality of each publication was assessed by two independent reviewers using a custom made checklist. Frequency of topics, muscles Keywords: studied, swimming activities, populations, types of equipment and data treatment were determined from Electromyography Swimming all selected papers and, when possible, results were compared and contrasted. In the first 20 years of EMG Systematic review studies in swimming, most papers were published as congress proceedings. The methodological quality Quality assessment was low. Crawl stroke was most often studied. There was no standardized manner of defining swimming phases, normalizing the data or of presenting the results. Furthermore, the variability around the mean muscle activation patterns is large which makes it difficult to define a single pattern applicable to all swimmers in any activity examined. Ó 2014 Elsevier Ltd. All rights reserved.

Contents

1. Introduction ...... 274 2. Methods ...... 274 3. Results...... 279 3.1. Literature search and retrieval ...... 279 3.2. Research question 1: What is the methodological quality of publications on EMG in swimming?...... 279 3.3. Research question 2: What were the purposes and most frequent topics examined? ...... 280 3.4. Research question 3: Which muscles have been investigated? ...... 281 3.5. Research question 4: Which swimming activities and populations have been studied? ...... 281 3.6. Research question 5: What types of equipment were used and how was the data treated?...... 283 4. Discussion...... 283 4.1. Research question 1: What is the methodological quality of publications on EMG in swimming?...... 283 4.2. Research question 2: What were the main purposes and most frequent topics examined? ...... 283 4.2.1. Description of muscle activity of one group or performance level of swimmers ...... 283 4.2.1.1. Crawl stroke...... 284 4.2.1.2. Backstroke ...... 284 4.2.1.3. Butterfly ...... 284 4.2.1.4. Breaststroke ...... 285 4.2.2. Comparison of muscle activation patterns between different groups of participants ...... 285

⇑ Corresponding author at: Faculty of Kinesiology and Rehabilitation Sciences, Department of Kinesiology, KU Leuven, Tervuursevest 101 bus 1500, 3001 Leuven, Belgium. Tel.: +32 16 32 91 04. E-mail address: [email protected] (J. Martens).

4.2.2.1. Crawl ...... 285 4.2.2.2. Breaststroke ...... 285 4.2.3. EMG pattern in variations of a swimming stroke...... 285 4.2.3.1. Crawl ...... 286 4.2.3.2. Backstroke ...... 286 4.2.3.3. Butterfly ...... 286 4.2.3.4. Breaststroke ...... 287 4.2.4. Presentation or comparison of (new) EMG equipment ...... 287 4.2.5. Biofeedback with EMG ...... 287 4.3. Research questions 3 and 5: Which muscles have been investigated and what types of equipment were used and how was data treated? 287 4.4. Research question 4: Which swimming activities and populations have been studied? ...... 289 5. Conclusions...... 289 Conflict of interest ...... 289 Acknowledgements ...... 289 References ...... 289

1. Introduction 3. Which muscles have been investigated? 4. Which swimming activities and populations have been studied? Between the publication of Cureton’s Mechanics and kinesiology 5. What types of equipment were used and how was data treated? of swimming in 1930 (Cureton, 1930) and 1964, numerous attempts have been made to describe anatomical function and 2. Methods muscle involvement in human swimming. Muscle involvement is, however, only one element. The muscle use pattern within a A systematic search of the existing literature was conducted complex rhythmical swimming movement is a more important using the combined keywords ‘‘swimming’’ and ‘‘EMG’’ on studies consideration. Such information cannot be obtained by functional published before August 2013, in the electronic databases PubMed, anatomical deduction (Clarys, 1983a). Therefore, more than ISI Web of Knowledge, SPORT discus, Academic Search Elite, 50 years ago, researchers began to examine the potential to study Embase, CINAHL and Cochrane Library. Furthermore, proceedings muscle activation patterns using electromyography (EMG). from the following congresses were systematically scanned for The primary means of communication on swimming research in relevant papers: the quadrennial congress of Biomechanics and general and in particular EMG has been via abstracts and congress Medicine in Swimming (starting in 1971 following the first con- proceedings, in particular the quadrennial symposium on Biome- gress of 1970), the biannual congress of Biomechanics (starting chanics and Medicine in Swimming. This unfortunately has reper- in 1968, following the first congress in 1967) and the annual cussions for the quality and completeness of some publications as congress of the International Society of Biomechanics in Sports well as their availability in online databases. Nevertheless, the pio- (starting in 1983). In addition to the electronic databases and con- neering work presented over the years in this area deserves men- gress proceedings, the reference lists of all articles found were tion as do the 15 narrative reviews or overviews that have been examined to finalize the search. The inclusion criteria were: (1) written on this topic. They represented the state of the art at the studies containing EMG data on at least one of the four competitive time and usually give a narrative summary of articles, sometimes swimming strokes, (2) swimmers of all performance levels, sexes focused on one theme, without quality assessment of study design and ages and (3) articles written in English. The exclusion criteria and presentation of information. Most are not complete, nor claim were: (1) EMG studies on animals, (2) studies on swimming to be. Seven reviews and overviews focused purely on swimming (Clarys, 1983a, 1988, 1996; Birrer, 1986; Pink and Tibone, 2000; Rouard, 2010; Clarys and Rouard, 2011). In eight others, swimming Table 1 10 Items of the quality assessment checklist. is mentioned as one of a number of sports reviewed (Moynes et al., 1986; Clarys, 1987, 2000, 2006; Clarys et al., 1988a; Bradley and 1. Question/objective sufficiently described? 0, 1, or 2 Tibone, 1991; Glousman, 1993; Clarys and Cabri, 1993). While 2. Study design evident and appropriate? 0, 1, or 2 3. Connection to a theoretical framework/wider body of 0, 1, or 2 reviews on particular aspects of swimming, such as fatigue knowledge? (Rouard, 2010), or injuries (Pink and Tibone, 2000) were made, a 4. Subject characteristics sufficiently described? complete systematic review on the muscle activity in the four com- 4.1. Age (average and standard deviation) 0, 1, or 2 petitive strokes (crawl, breaststroke, butterfly and backstroke) has 4.2. Gender 0 or 2 never been published. The 50th anniversary of the first study on 4.3. Swimming/activity level of the participants 0, 1, or 2 5. Data collection methods clearly described and EMG in swimming is the ideal pretext for this systematic review systematic? which aims to give a comprehensive overview of 50 years of elec- 5.1. Research (swimming) protocol 0, 1, or 2 tromyographic research in swimming of the four competitive 5.2. Type of EMG system 0, 1, or 2 swimming strokes, to assess the quality of each publication and 5.3. Studied muscles 0, 1, or 2 5.4. Uni or bilaterality 0, 1, 2 or to summarize and analyze the topics, population and muscles NA studied and equipment used. In this way the gaps in research still 6. Data analysis clearly described and systematic? existing in this field can be clearly identified. 6.1. EMG filters 0 or 2 To guide this review of EMG in swimming the following 6.2. Normalization method 0, 1, or 2 research questions were examined: 6.3. Data processing protocol 0, 1, or 2 7. Some estimate of variance is reported for the main results 0, 1, 2 or NA 1. What is the methodological quality of publications on EMG in 8. Conclusions supported by the results 0, 1, or 2 swimming? 9. Number of participants sufficient to draw conclusions 0, 2, 4 or 6 2. What were the main purposes and most frequent topics 10. Statistical analysis is described and appropriate 0, 1, 2 or NA examined? J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291 275 movements replicated on dry land, (3) EMG studies on swimmers and (6.3) data processing protocol (e.g. integration, rectification, with injuries (or a history of injuries) and swimmers with an etc.) is described. Items that were not selected from the existing impairment, (4) studies only containing EMG data on competitive quality assessment checklist of Kmet et al. (2004) were: (a) Method swimming starts and turns, (5) studies on fatigue in swimming of subject/comparison group selection or source of information/ using EMG, (6) EMG studies on breathing muscles in underwater input variables described and appropriate, (b) If interventional divers, (7) master and PhD thesis’s. and random allocation was possible, was it described?, (c) If inter- Titles and abstracts were reviewed by two persons and then all ventional and blinding of investigators was possible, was it papers meeting the criteria were subject to a quality assessment reported?, (d) If interventional or blinding of subjects was possible, procedure. A quality assessment checklist of 10 items was created was it reported?, (e) Outcome and (if applicable) exposure mea- based on an existing list for assessing the quality of qualitative and sure(s) well defined and robust to measurement/misclassification quantitative studies (Kmet et al., 2004) as shown in Table 1. Items bias? Means of assessment reported?, (f) Controlled for confound- included from the existing list were items 1, 2, 3, 4, 5, 6, 7, and 8. ing?, (g) Context of study clear?, (h) Sampling strategy described, Item 4 was for the purpose of this review split in 3 sub items: (4.1) relevant and justified?, (i) Use of verification procedure(s) to estab- age (average and standard deviation), (4.2) gender and (4.3) swim- lish credibility?, (j) Reflexivity of the account? And (k) Results ming or activity level of the participants is reported. Item (5) was reported in sufficient detail. Two more items (9 and 10 in Table 1) split into (5.1) research (swimming) protocol, (5.2) type of EMG were added to make this checklist tailor made for EMG studies in system, (5.3) studied muscles and (5.4) uni or bi laterality of the swimming. On all the above mentioned items, with the exception measurements are reported. Item (6) was split into (6.1) EMG fil- of items (4.2) and (6.1) which are yes or no items, articles could ters used are described, (6.2) normalization method is described score a 0, 1 or 2 or NA when non applicable. A score of 0 was given

PubMed ISI Web of Knowledge SPORT discus Academic Search Elite Embase CINAHL Cochrane Library 269 215 74 73 38 4 2

Total 675

Duplicates 211

Title assessment 464

Animal studies 301

Abstract assessment 163

Not relevant to the review topic 65

Reviews and overviews 15

Full text reading 83

Reference checking Not relevant to the review topic Grey literature (congress proceedings) 72 36

Included in review 47

2 duplicate and 1 triple published studies 4

Total included in review 43

Fig. 1. Flowchart of the systematic literature search. Table 2 276 Overview of all selected publications ranked by quality (column 2). Yrs = years, PB = personal best (time), WR = world record, D = deltoideus, DA = deltoideus anterior, DP = deltoideus posterior, SA = serratus anterior, T = trapezius, RM = rectus femoris, SSC = subscapularis, SSP = supraspinatus, ISP = infraspinatus, Tmaj = teres major, Tmin = teres minor, LD = latissimus dorsi, PM = pectoralis major, BB = biceps brachii, TB = triceps brachii, FCU = flexor carpi ulnaris, ECU = extensor carpi ulnaris, BR = brachioradialis, G = gastrocnemius, RA = rectus abdominis, GM = gluteus maximus, ED = extensor digitorum, SCM = sternocleidomastoideus, OE = obliquus externus, ST = semitendinosus, FDS = flexor digitorum superficialis, TA = tibialis anterior, VM = vastus medialis, BF = biceps femoris, ECRB = extensor carpi radialis brevis, S = sacrospinalis, ES = erector spinae, RM = rhomboid major, IP = iliopsoas, RF = rectus femoris.

Author (year) Score (%) Purpose of the study Participants Swimming EMG type Uni/ Muscles studied strokes bilaterallity Ruwe et al. (1994) 95 Compare muscle activity in 12 muscles in 25 Normal shoulders (M age 39 yrs) and Breast Fine wire telemetric Uni DA, D medialis, DP, SA, T breaststroke between swimmers without 14 painful shoulders (M age 31 yrs) in 34 upper part, RM, SSC, SSP, and with shoulder pain collegiate and master level competitive ISP, Tmin, LD, PM swimmers Lauer et al. (2013) 92 Apply the coactivation index to the prime 10 International level male crawl Crawl SEMG on-line Uni BB, TB movers of the elbow in front crawl and swimmers (M age 20.8 ± 2.3 yrs and PB examine how this differed per stroke 200 crawl 109.2 ± 2.3 s)

phases 273–291 (2015) 25 Kinesiology and Electromyography of Journal / al. et Martens J. Caty et al. (2007) 89 Evaluate stabilization of wrist joint during 7 Male international swimmers (M age Crawl SEMG on-line Uni FCU, ECU insweep and outsweep of tethered 22.6 ± 2.7 yrs and PB 100 crawl freestyle swimming 52.08 ± 4.89 s) Pink et al. (1991) 87 Describe muscle activity patterns in 12 20 (15 male and 5 female) collegiate and Crawl Fine wire telemetric Uni DA, D medialis, DP, SA, T muscles during freestyle swimming linked master level competitive swimmers (M upper part, RM, SSC, SSP, to the arm phases age 39 yrs) ISP, Tmin, LD, PM Pink et al. (1992) 87 Describe the muscle activity in 12 muscles 14 (9 male, 5 female) collegiate and Back Fine wire telemetric Uni DA, D medialis, DP, SA, T of the shoulder in competitive swimmers master level competitive swimmers, (M upper part, RM, SSC, SSP, doing backstroke age 31 yrs) who were asymptomatic for ISP, Tmin, LD, PM shoulder pain Rouard and Clarys 84 Determine behavior of antagonist muscles 9 Male good 100 m freestyle French Crawl SEMG telemetric Uni BB, TB, FCU, BR, LD, DA (1995) = Clarys of the upper limb in swimming swimmers (M age 17.3 ± 2.59 yrs and PB and Rouard (4 Â 100 m maximal effort) 100 crawl 58.62 ± 2.5 s) (1995) = Clarys and Rouard (1996) Rouard and Billat 82 Investigate the kinematics of freestyle 30 Males, 20 females (M age Crawl SEMG telemetric Uni (?) DA, D medialis, BB, TB, BR, (1990) combined with EMG to determine 18.1 ± 2.7 yrs) in 3 categories: regional, FCU differences in performance level and sex national and international level Figueiredo et al. 79 Examine differences in EMG of 3 parts of 1 Trained male swimmer (24 yrs) with Crawl SEMG on-line Uni DP, D middle, DA (2007) deltoid in 2 different patterns of recovery 15 years of competitive swimming (lateral and bend arm) and differences experience between inspiratory and non-inspiratory cycles 9 Male swimmers (M age 17.3 ± 2.6 yrs Crawl SEMG telemetric Uni (?) DA, D medialis, BB, TB, BR, Clarys and Rouard 79 Describedown, in muscular and outsweep motor in control frontcrawl during and and PB 100 crawl 58.62 ± 2.5 s) FCU (1996) = Clarys study the behavior of shoulder muscles and Rouard under maximal normalized loading (1995) = Rouard conditions Clarysand Clarys and Rouard (1995) 76 Investigate the sculling and sweeping 9 Male swimmers (M age 17 ± 2.6 yrs and Crawl SEMG telemetric Uni (?) BB, TB, FCU, BR, LD, DA (1995) = Rouard dimension of the propulsive arm PB 100 crawl 58.62 ± 2.5 s) and Clarys movement in crawl using EMG (1995) = Clarys and Rouard (1996) Cabri et al. (1988) 74 Compare EMG and forces between 13 Top male swimmers Crawl SEMG on-line Uni TB long head, PM, FCU, LD, different stroke rates in tethered RF, G lateral head swimming Monteil and Rouard 74 Investigate the activity of 5 muscles 7 Skilled sprinters and long distance Crawl SEMG telemetric Uni PM, BR, BB, TB caput (1992) during crawl swimming with 3 different swimmers (4 men M age 16.7 ± 0.4 yrs and medialis, FCU sizes of paddles 3 women M age 15.6 ± 0.4 yrs) Rouard et al. (1992) 74 Investigate evolution of in kinematic and 30 Males, 20 females (M age Crawl SEMG telemetric Uni (?) DA, D medialis, BB, TB, BR, EMG parameters with increasing speed in 18.1 ± 2.7 yrs) FCU freestyle Pink et al. (1993) 74 Description of the muscle firing patterns 20 (3 female, 17 male) collegiate and Butterfly Fine wire telemetric Uni DA, D medialis, DP, SA, T in 12 muscles of the normal shoulder master level competitive swimmers ( M upper part, RM, SSC, SSP, during butterfly age 39 yrs) ISP, Tmin, LD, PM Rouard et al. (1993) 74 Examine relationship between hand 14 Male freestyle swimmers, PB of 78.7– Crawl SEMG telemetric Uni LD, TB, BR, BB, FCU, DA forces and EMG of 6 upper limb and trunk 98.4% of WR on 100 crawl muscles according to level of performance Monteil and Rouard 74 Investigate muscular activity and 6 Skilled swimmers (3 men, M age Crawl SEMG telemetric Uni PM pars clavicular, TB caput (1994) kinematic parameters between paddle 16.67 ± 0.58 yrs and 3 women, M age medialis, BB, BR, FCU swimming and free swimming at same 15.67 ± 0.58 yrs) all competing at French intensity National Championships Strass (1991) 71 Examine explosive strength ability and 41 (28 male 13 female) for lab study all of Crawl, 3 maximal SEMG telemetric Uni PM, D pars clavicularis, the neuromuscular activity of selected national level ( M age 20.63 ± 5.53 yrs) 17 Isometric Tmaj, LD, TB caput laterale, arm shoulder muscles of sprinters in male sprinters for swim study ( M age contractions in 3 BB technical positions relevant to swimming 23.61 ± 4.96 yrs) different positions with a special test apparatus in the lab on dry land

(with EMG). In addition evaluate if there 273–291 (2015) 25 Kinesiology and Electromyography of Journal / al. et Martens J. are relationships between characteristics found on dry land and kinematic parameters in swimming (without EMG) Monteil and Rouard 71 Describe differences in EMG between 7 Skilled sprinters and long distance Crawl SEMG telemetric Uni (?) PM, BR, BB, TB caput (1992) swimming with paddles at slow, medium swimmers (4 men M age 16.7 ± 0.4 yrs and medialis, FCU and fast speed 3 women M age 15.6 ± 0.4 yrs) Piette and Clarys 68 Describe and compare EMG of several 12 Male swimmers ( M age 20 yrs), 6 Crawl SEMG telemetric Uni (?) BB, LD, PM pars clavicularis (1979) muscles in crawl in top and average level competitive (PB 100 m crawl < 60 s) and 6 and sternocostalis, RA above swimmers PE students (PB 100 m crawl around 80 s) and below umbilicus, GM pars superior and inferior Bollens et al. (1988) 68 Investigate differences in EMG between 13 Male swimmers (PB 100 m crawl of Crawl SEMG on-line Uni TB caput longum, PM pars normal crawl swimming and fully 59 s or less ( M time 56.9 s)) sternocostalis, LD, RF, G tethered swimming at 3 different stroke caput laterale frequencies Monteil et al. (1996) 68 Investigate activity of rotator cuff muscles 9 Male swimmers between 17 and 30 yrs Crawl Fine wire on-line Uni (?) SSP, ISP, Tmin, SSC, PM pars in during 400 yards freestyle test in a ( M age 23.8 ± 4.5 yrs) clavicular, LD flume Clarys et al. 62 Describe and compare EMG of all 60 Participants, 30 competitive and 30 Crawl SEMG telemetric Uni and bilateral BB, TB caput longum and (1983)= Clarys superficial muscles in crawl in good and swimmers with good technical skills laterale, DA, DL, DP, ED, T (1985) average level swimmers (bilateral), SCM, OE (bilateral), G caput laterale, RF, ST, FCU, PM pars clavicularis and sternocostalis, LD, RA pars superior and inferior, GM pars superior and inferior, TA Clarys 61 Describe and compare EMG of all 60 Participants, 30 competitive and 30 Crawl SEMG telemetric Uni and bilateral BB, TB caput longum and (1985) = Clarys superficial muscles in crawl in good and swimmers with good technical skills laterale, DA, DL, DP, ED, T et al. (1983) average level swimmers (bilateral), SCM, OE (bilateral), G caput laterale, RF, ST, FCU, PM pars clavicularis and sternocostalis, LD, RA pars superior and inferior, GM pars superior and inferior, TA Clarys et al. 61 Investigate differences in EMG between 17 Well trained swimmers Crawl SEMG on-line Uni PM, TB, FDS, LD (1988)= Clarys normal crawl swimming at very slow, and Cabri (1988) submaximal (80%) and full 100% sprint speed and MAD swimming

(continued on next page) 277 Table 2 (continued) 278

Author (year) Score (%) Purpose of the study Participants Swimming EMG type Uni/ Muscles studied strokes bilaterallity Nuber et al. (1986) 59 Describe muscle activity of shoulder 10 Male and 1 female, 2 former Crawl Breast, Fine wire on-line Uni (?) BB, SSC, LD, PM clavicular muscles in 3 strokes with fine wire competitive swimmers, 2 amateurs, 7 Butterfly, Crawl on head, SSP, ISP, SA, D ‘‘middle electrodes and compare to imitations of master students swimming (range 19– dry land, Butterfly part’’ these movements on dry land. 35 yrs) on dry land Ikai et al. (1964) 55 Arrange equipment for recording EMG in 9 Male Olympic participants and 5 Crawl, Breast, SEMG on-line Uni FCU, ECU, BB, TB, D, PM, swimming students from University swimming club Butterfly, Back Tmaj, T, LD, RA, GM, RF, BF, TA, G Compare EMG of high level versus lower level swimmers Barthels and Adrian 53 Identify degree to which selected muscles 4 Intercollegiate male swimmers, PB Butterfly, Butterfly SEMG on-line Uni RA, ‘‘Lumbar ES’’, RF, BF, TA, (1971) functioned during 4 different dolphin kick 100 yard butterfly between 50 and 60 s kicking G medialis conditions .Mrese l ora fEetoygah n ieilg 5(05 273–291 (2015) 25 Kinesiology and Electromyography of Journal / al. et Martens J. Bollens and Clarys 53 Compare EMG of normal crawl swimming 8 Male participants (M time of 58.1 s on Crawl SEMG telemetric Uni BB caput longum, TB, BR, ED (1985) to swimming with 6 different designs of 100 m crawl) handpaddles Clarys et al. (1985) 53 Compare wired and telemetric EMG in air 11 Competent swimmers Crawl Elbow SEMG telemetric and Uni BB and in water during contractions and flexion on dry land on-line swimming Yoshizawa et al., 47 Describe EMG and kinematics in 21 Participants ranked in 3 classes: Breast SEMG on-line Uni TA, G, VM, RF, BF, GM, RA, S, 1976 breaststroke Olympians, members of University BB long head, TB lateral swimming club, average adults head DA, DP, PM ‘‘sterno portion’’, LD Oka et al. (1979) 47 Analyze the evolution of muscle activity 31 Infants and children ranging from Crawl flutter kick SEMG on-line Uni (?) TA, G, VM, RF, BF, GM during the learning process of flutter kick 10 month old to 12 year old and 8 skilled adults Daniel and Klauck 47 Comparison of force, speed parameters 1 Male competitive swimmer of 27 years Breast in water ? Uni PM, D, TB, Tmaj (1999) and EMG for muscle coordination old and on dry land between free swimming, tethered swimming, and swimming movements on bench trainer Maes et al. (1975) 45 Compare EMG in 3 swimming strokes 13 Male all-around swimmers and water Crawl, Breast, SEMG telemetric Uni (?) BB, FCU, RA (above and between normal swimmers and water polo players (M age 21 yrs) Backstroke below umbilicus), BF, TA polo players Delhez et al. (1971) 42 Describe EMG of diaphragm in different 2 Swimming champions, 3 physical Crawl, Breast, Esophageal wired NA Diaphragm swimming styles education graduates and 2 physicians Back, Diving surface electrodes with trained4 Olympic in underwater swimmers diving Breastmemory SEMG on-line Uni TA, G, VM, RF, BF, GM, RA, S, Yoshizawa et al. 42 Describe EMG in continuous and glide BB long head, TB lateral (1978) stroke breaststroke head, DA, DP, PM ‘‘sterno portion’’, LD, FCR, ECRB Clarys et al. (1973) 39 Describe similarities and differences of 10 Competent swimmers and water-polo Crawl, Water-polo SEMG telemetric Uni (?) BB, TB, FCU, BR competition and water-polo front crawl players of national division level crawl Lewillie (1976) 39 Evaluate variability of EMG during 10 Swimmers of different levels of ability Crawl SEMG telemetric Bilateral TB swimming (results presented of 1 swimmer) Tokuyama et al. 39 Examine characteristics of swimming in 37 Participants, from 10 month old infant Crawl, Breast, Back SEMG on-line Uni TA, G, VM, RF, BF, GM, RA, S, (1976) infants and children by EMG to unskilled and skilled adults BB long head, TB lateral head, DA, DP, PM ‘‘sterno portion’’, LD Yoshizawa et al. 39 Determination of the effects of EMG- 7 Top level Japanese swimmers Breast, Butterfly, SEMG telemetric Uni D ‘‘pars spinata’’, PM ‘‘pars (1983) biofeedback on training participating in World Championships and Back abdominalis’’, LD, FCR Olympics (2 backstrokers, 2 butterfly swimmers, 3 breaststrokes) Lewillie (1971) 35 Normalization of EMG to maximal 1 Swimmer Crawl SEMG telemetric Uni TB isometric contraction Lewillie (1974) 34 Superimpose EMG on kinematics of elbow Unknown number ‘‘who were not of Crawl, Breast, SEMG telemetric Uni (?) BB, TB joint international caliber’’ Butterfly, Back J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291 279

when the answer on the question was ‘‘no’’ and a score of 2 when the answer was ‘‘yes’’. A score of 1 was given when the information was partially reported, or when it could be derived from the results, tables or images. To give more weight to studies with a larger sample size, item (9), was scored as follows: a 0 was given when less than five participants participated in the study. An article scored 2 when between five and nine participants participated and a four when there were between 10 and 20 participants. When there were more than 20 participants, an article scored 6 on this ‘‘Extensors of fore-arm’’, GM, hamstrings, ‘‘Quadri- ceps femoris’’, ‘‘Six strong muscles of the calf’’, IP item. The adapted checklist was validated by consensus of four experts in the field of at least three of the following four fields: electromyography, swimming, biomechanics and systematic reviewing. The first and second author independently scored all articles and when disagreement occurred, the final score was decided in consultation with the last author. The summation of the scores on the different items was then put on a scale of 0–100 (%) to cancel out the influence of items that were non applicable for some articles.

3. Results SEMG telemetric Uni (?) TB Quadriceps SEMG telemetric Uni (?) TB, RF 3.1. Literature search and retrieval

Fig. 1 shows the results of the systematic literature search. A total of 675 titles were retrieved. Duplicates were eliminated using reference managing software (Endnote X7, Thomson Reuters, USA), leaving 464 references. A search on the keywords ‘‘animal’’, ‘‘fish’’, ‘‘rat’’, and ‘‘mouse’’ eliminated 301 animal stud- Butterfly, Back Butterfly, Back ies. After reading the remaining 163 abstracts, 65 articles did not meet the inclusion criteria. Furthermore 15 reviews or overviews were eliminated as they did not present new information. The full texts of the remaining 83 publications were read as well as 36 additional papers found after reference checking of the retrieved articles and in congress proceedings. Seventy-two of these were further eliminated for not meeting the inclusion criteria. Of the remaining 47 papers, two were duplicate publications, and one study was published three times. In these cases, a study was first presented at a congress and later published as peer reviewed article. The article scoring highest on the quality assessment scale was retained in the final total of studies in this review. An overview of all selected publications is given in ?? Crawl Crawl, Breast, SEMG on-line Uni (?) FDS 3 Female intercollegiate swimmers14 Breast, Race Back swimmers (10 men, 4 women) Crawl SEMG on-line Uni ? DA, DP, TB, Wrist flexors Uni (?) LD, Pectoralis, T, TB, Table 2. The publications are ranked from highest to lowest (column 2) expressed as a percentage of maximum score on the quality assessment checklist. Duplicate and triple publications are included, but marked in column 1 in italic. Studies of EMG in swimming are descriptive or case studies, describing muscular activity in swimming, and/or adaptation and development of equipment for use in the challenging environment of water. Only one intervention study was found (Oka et al., 1979).

3.2. Research question 1: What is the methodological quality of publications on EMG in swimming? normal swimming of hand and foot inbackground. relation to a fixed respect to kinematics of swimming swimming In Fig. 2, the results of the quality assessment are given per Olympiad. The early years were dominated by publications with a lower score. The mean quality of the 20 articles published before 1984 is 42% (SD 12.6%). Of these, only one article (Ikai 18 Define possibilities of telemetry A freestyle swimmer Crawl SEMG telemetric Uni (?) TB, ‘‘Femoral quadriceps’’ 21 Describe EMG in 4 strokes linked to path 24 Compare EMG on the MAD system to 32 Describe EMG in 4 strokes at 3 speeds 1 Swimmer of national class Crawl, Breast, 34 Examine muscular forces of29 the hand with Define motor pattern in freestyle et al., 1964) was peer reviewed. The others were published as congress proceedings. Proceedings books in the early years were reviewed and corrections were made, but nothing was excluded

Clarys unless it was very poor English. More recently, interest in swim- = ming stroke research with EMG has decreased, and furthermore the vast majority of publications (18/27) since 2000 has been on (1988) et al. (1988) Nemessuri (1971) Lewillie (1968) Clarys and Cabri Lewillie (1971) Lewillie (1973) Vaday and Dupuis et al. (1979) fatigue and are therefore excluded here as this was not the topic of this review. 280 J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291

10 system that was developed in the 1980s, the swimmer pushes off with the hands on an underwater rail with pads. With this tool, the force applied by the swimmer during each stroke could be recorded (Hollander et al., 1986). Three studies focussed on variations in style of a certain swimming stroke. These studies com- 5 80-100% pared water-polo crawl and competition front crawl (Clarys 60-80% et al., 1973), a gliding and a continuous breaststroke (Yoshizawa 40-60% et al., 1978) and two types of arm recovery in crawl (Figueiredo

Number of of publicaons Number et al., 2007). Finally, a number of studies investigated the extent 0 0-40% to which ‘‘alternative swimming training methods’’ such as tethered swimming, or dry land swimming result in differences in EMG patterns. The question was: are these alternative methods

1964 -1968 1969 -1972 1973 -1976 1977 -1980 1981 -1984 1985 -1988 1989 -1992 1993 -1996 1997 -2000 2001 -2004 2005 -2008 2009 -2012 sufficiently specific to benefit swimmers. Two studies compared

2013 - present differences in ‘‘free’’ swimming, implying swimming in the water, Fig. 2. Number of publications per Olympiad and per quality assessment category and tethered swimming (Bollens et al., 1988; Daniel and Klauck, expressed as a percentage of a perfect score (100%). 1999). In tethered swimming, the swimmer is attached to the pool wall with a rope or a rubber band usually ﬁxed at the waist, swimming but remaining stationary. In free swimming, the swimmer 3.3. Research question 2: What were the purposes and most frequent moves forward. Two studies did the same for dry land swimming topics examined? and ‘‘free’’ swimming, in freestyle (Strass, 1991) and breaststroke (Daniel and Klauck, 1999). Fig. 3 presents an overview of the general purposes and topics A second group of studies (n = 10) compared muscle activation examined in the included studies. For the second group of studies patterns between different groups of participants. Seven studies (n = 21), speciﬁcations are presented indicating the kinds of focussed on differences between top level swimmers (often Olym- variations of a certain swimming stroke investigated. In all other pic level) and lower level swimmers (Ikai et al. (1964), Delhez et al. studies, the purpose was to examine the EMG pattern of one group (1971), Yoshizawa et al. (1976), Piette and Clarys (1979), Clarys of participants (n = 12), compare the muscle pattern between dif- et al. (1983) duplicate to Clarys (1985), Rouard and Billat (1990), ferent levels of swimmers (n = 10), present (new) EMG equipment Rouard et al. (1993)). In one of these the differences between sexes (n = 3) or use EMG in teaching with audio feedback (n = 1). was also investigated (Rouard and Billat, 1990). One study Maes Five studies covered two purposes and were therefore included et al. (1975) claimed in the title, abstract and conclusions to eval- in this overview twice. 45% of the studies (n = 21) examined the uate swimmers with an impairment. Nevertheless the methods differences in muscle activation patterns in different variations of and the results of the study clearly described a comparison of the a certain swimming stroke. Two studies compared EMG between crawl, breast and backstroke of waterpolo players to competitive different stroke rates in freestyle (Cabri et al., 1988; Bollens swimmers. Because of this discrepancy, this study will not be et al., 1988). Eight showed the changes occurring when swimming further discussed. Finally, two studies focussed on the muscle speed was altered (within the same group of participants) patterns in children (Tokuyama et al., 1976; Oka et al., 1979). (Barthels and Adrian (1971), Delhez et al. (1971), Lewillie (1971b, Twelve separate studies aimed to describe the muscle activa- 1973), Dupuis et al. (1979), Clarys et al. (1988b) duplicate to Clarys tion pattern of one group or one performance level of swimmers and Cabri (1988), Rouard et al. (1992), Monteil and Rouard (Lewillie (1971a), Vaday and Nemessuri (1971), Lewillie (1974), (1992a)). In four studies, differences in muscle activity were Lewillie (1976), Pink et al. (1991), Pink et al. (1992), Pink et al. studied when swum with and without diverse types or different (1993), Clarys and Rouard (1995) duplicate to Rouard and Clarys sizes of swimming equipment, such as hand paddles (Bollens and (1995) duplicate to Clarys and Rouard (1996), Monteil et al. Clarys, 1985; Monteil and Rouard, 1992b, 1994), or when swim- (1996), Caty et al. (2007), Lauer et al. (2013)). One study described ming on the MAD system (Clarys et al. (1988b) duplicate to the pattern of one group of swimmers with the objective to make Clarys and Cabri (1988)). In the MAD (measurement of active drag) comparison to swimmers with a shoulder injury (Ruwe et al.,

20 MAD

dry land

frequency 15 tethered

paddle 10 style

speed Number of publicaons 5

0 Descripon EMG paern Descripon EMG paern Comparison EMG paern Presentaon of (new) EMG Audio feedback to teach variaons diﬀerent levels of equipment swimmers

Fig. 3. Overview of the main purposes of all publications on EMG in the 4 competitive swimming strokes. MAD = measurement of active drag. J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291 281

1994). In this review, only the section of the study with the non- 3.4. Research question 3: Which muscles have been investigated? injured participants is considered. Three studies presented or compared (new) EMG equipment Table 3 shows the frequency of muscles studied in EMG studies and used swimmer(s) to demonstrate the equipment, more specif- of swimming strokes. Duplicate studies are counted only once sub- ically a telemetric EMG system (Lewillie, 1968), a comparison sequently 43 articles remain. Fig. 4 shows the relative frequency of between wired and telemetric EMG recordings (Clarys et al., the muscles studied. Only one study analyzed bilaterally (Lewillie, 1985), and a ﬁne wire EMG system to record muscle activity of 1976) and in one study (Clarys, 1985), all muscles were tested the deep rotator cuff muscles (Nuber et al., 1986). A ﬁnal study pre- unilaterally, with the exception of the trapezius and obliquus sented EMG-biofeedback (Yoshizawa et al., 1983) to teach a certain externus. In all other studies the muscles of interest were tested style of swimming. unilaterally. Although 39% of the studies did not mention this explicitly, this information could be derived from the results or discussion section. The single study analyzing the diaphragm was excluded from this uni-bilateral discussion.

Table 3 frequency of studied muscles in EMG studies on competitive swimming strokes. 3.5. Research question 4: Which swimming activities and populations Muscle Study frequency (n) have been studied? Pectoralis major 11 pectoralis major pars clavicularis 5 The total sum of activities studied in the publications reviewed pectoralis major pars sternocostalis 6 is greater than the number of unique publications itself (43). In 12 pectoralis major ‘‘pars abdominalis’’ 1 studies, more than one swimming activity was studied. All multi- ‘‘Pectoralis’’ 1 ple-stroke studies date from before 1987. After 1987, researchers Triceps brachii 18 Triceps brachii caput laterale 5 focused on one swimming stroke, but went deeper in describing Triceps brachii caput medialis 3 and explaining their findings. In the vast majority (48%) of studies, Triceps brachii caput longum 3 freestyle (also called crawl) was at least one of the swimming Biceps brachii 17 activities investigated. Breaststroke was the second most popular Biceps brachii caput longum 4 swimming stroke (20%), followed by backstroke (14%) and Flexor carpi ulnaris 13 Flexor carpi radialis 2 butterfly (12%) (see Fig. 5). ‘‘wrist flexors’’ 1 Brachioradialis 9 Extensor carpi ulnaris 2 Extensor digitorum 2 Extensor carpi radialis brevis 1 ‘‘extensors of forearm’’ 1 Flexor digitorum superficialis 1 Latissimus dorsi 20 Deltoideus 2 Deltoideus pars anterior 14 Deltoideus pars posterior 10 Deltoideus pars medialis/acromialis 8 Deltoideus pars lateralis 1 Deltoideus ‘‘pars clavicularis’’ 1 Deltoideus ‘‘pars spinata’’ 1 Sternocleidomastoideus 1 Supscapularis 6 Supraspinatus 6 Infraspinatus 6 Trapezius 3 Trapezius pars superior 4 Serratus anterior 5 Teres major 3 Teres minor 5 Fig. 4. Relative frequency of muscles studied in EMG research in swimming in the 4 Rhomboideus (major) 4 competitive strokes. Rectus abdominis 5 Rectus abdominis above umbilicus 3 Rectus abdominis below umbilicus 3 Obliquus externus 1 1% 1% 1% 1% (Lumbar) Erector spinae/Sacrospinalis 4 Diaphragm 1 crawl Quadriceps 3 12% Vastus medialis 4 breaststroke Rectus femoris 10 Gluteus maximus 6 backstroke Gluteus maximus pars superior 2 Gluteus maximus pars inferior 2 48% buerfly Hamstrings 1 14% Biceps femoris 7 waterpolo crawl Semitendinosus 1 Iliopsoas 1 freestyle kick Tibialis anterior 8 Gastrocnemius 5 20% buerfly kick Gastrocnemius pars lateralis 3 Gastrocnemius pars medialis 1 underwater diving ‘‘six strong muscles of the calf’’ 1 Fig. 5. Swimming activities examined in EMG research in swimming. Table 4 282 Overview of muscles studied per general purpose (first column) and per swimming stroke (column 2); the number in column 3 represents the number of studies publish around a certain topic; the numbers in the following columns represent the number of times a certain muscle was studies in a certain topic; Abbreviations are given in Table 2.

Purpose Stroke Studied muscles Arm Leg Shoulder Trunk Internal TB BB FCU FCR FDS BR wrist ECU ED ECRB forearm VM RF Quadriceps GM BF ST IP TA G PM D SSC SSP ISP T SCM T-maj T-min SA RM LD RA OE ES/S Diaphragm flexors extensors Variation in stroke rate crawl 2 2 1 2 2 2 2 Variation in swimming crawl 6 5 2 2 1 2 1 2 1 1 1 speed butterfly 2 1 2 1 1 1 1 1 butterfly 1111111 kick breaststroke 3 2 1 1 1 1 .Mrese l ora fEetoygah n ieilg 5(05 273–291 (2015) 25 Kinesiology and Electromyography of Journal / al. et Martens J. backstroke 3 2 1 1 1 1 underwater 1 1 diving Variation in equipment: crawl 3 3 3 2 3 1 2 handpaddles Variation in equipment: crawl 1 1 1 11 MAD Variation in style crawl 2 1 1 1 1 1 breaststroke 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Variation in environment: crawl 1 1 1 1 1 1 free vs tethered swimming breaststroke 1 1 11 1 Variation in environment: crawl 1 1 1 11 1 1 water vs dry land swimming breaststroke 1 1 11 1 Differences between crawl 6 4 5 4 2 1 1 2 3 1 1 2 2 3 4 2 1 1 4 3 1 1 different levels of swimmers breaststroke 3 2 2 1 1 1 2 2 2 2 2 2 2 1 1 2 2 1 1 butterfly 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 backstroke 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 underwater 1 1 diving Differences between sexes crawl 1 1 1 1 1 1 Differences between crawl 1 1 1 1 1 1 1 1 1 1 1 1 1 1 children and adult swimmers breaststroke 1 1 1 1 1 1 1 1 1 1 1 1 1 1 backstroke 1 1 1 1 1 1 1 1 1 1 1 1 1 1 crawl kick 1 1 1 1 1 1 1 Pattern of 1 group (level) of crawl 9 6 3 2 1 1 1 2 1 1 2 3 2 2 2 2 2 1 1 4 swimmers breaststroke 3 2 1 1 1 1 1 1 1 1 1 1 1 1 backstroke 3 2 1 1 1 1 1 1 1 1 1 1 1 1 butterfly 3 2 1 1 1 1 1 1 1 1 1 1 1 1 Compare or present (new) crawl 3 1 2 1 1 1 1 1 1 1 1 EMG equipment breaststroke 1 1 11111 1 1 butterfly 1 1 11111 1 1 Biofeedback breaststroke 1 1 11 1 butterfly 1 1 11 1 backstroke 1 1 11 1 J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291 283

Table 4 summarizes the purposes of the EMG studies combined is not indicated if it is a raw, a rectified, an integrated, or a root with the muscles studied and swimming strokes. mean squared signal normalized. In the total of 43 studies considered in this review, one study provided no information on the swimmer or swimmers participat- 4. Discussion ing (Lewillie, 1971a). In all other studies, information on the population studied can be categorized in (a) the swimming level, (b) This review presents an overview of all published articles on the personal best times of the swimmers and (c) the age of the EMG in competitive swimming strokes and assessed the quality swimmers. of the articles, and where possible compared results of studies that confirm or contradict one another. The manner in which EMG data 3.6. Research question 5: What types of equipment were used and how was reported, and the limited information presented in the was the data treated? methods section, on the participant characteristics and swimming protocol especially in the first 20 years of EMG research make com- As can be seen in Fig. 6, the majority of studies (79%) have been parison or merger of data from similar studies impossible in most executed using surface electromyography (SEMG). Only 14% used cases. fine wire or needle electrodes. For the final 7% it is unknown what equipment was used, or an esophageal wired electrode was used. 4.1. Research question 1: What is the methodological quality Another distinction can be made between online data acquisition of publications on EMG in swimming? (34.8%) and telemetric data acquisition (55.8%). The latest wireless EMG technology has not yet appeared in publications on muscle The quality of publications of the first 20 years in EMG research activity in the competitive swimming strokes. in swimming was low (cf. Fig. 2). Since the 1980s, the number of In 47% of the studies, a maximal voluntary contraction was used peer reviewed articles rose and quality with it. Essential improve- as a way of normalizing the EMG data (Fig. 7, left panel). In this ments were due to a more detailed description of the participants, method, the participant was asked to perform one or several max- swimming protocol and information on the processing of electrical imal contractions against (manual) resistance. The highest value signal (items 4, 5 and 6 of Table 1). (in volts) found during this maximal contraction was considered 100%. All values found during swimming were then expressed as 4.2. Research question 2: What were the main purposes and most a % of the maximal voluntary contraction. In case of normalization frequent topics examined? to the dynamic maximum, the highest value found during the swimming trail was considered 100%. In 13 studies, no normaliza- To organize the discussion of research question 2, the five main tion was executed. All these studies score lower than 60% on the purposes and most frequent topic categories from Fig. 3 are quality assessment scale. separated. Half of the studies used integration of the electrical signal in processing data (Fig. 7, right panel). Of the 11 studies unclear on 4.2.1. Description of muscle activity of one group or performance level the data processing protocol, nine mention normalization, but it of swimmers The very first three articles examining the muscle activity of one group or level of swimmers were of low methodological qual- 30 ity, but nevertheless delivered pioneers work in the field. Lewillie 25 (1971a) analyzed typical propulsive muscles (quadriceps and triceps) in the four competitive swimming strokes attempting to 20 combine muscle activity with underwater kinematics by using 15 light traces to illustrate the path of the hand in relation to a fixed 10 background. This article was however unclear regarding the swim- 5 ming protocol, participant characteristics and conclusions and is 0 Number of of publicaons Number therefore the second lowest scoring publications in this review. A follow up study (Lewillie, 1974) again analyzed the muscle work of biceps and triceps in all four swimming strokes and concluded that EMG is ‘‘very characteristic of the stroke and much less so of the individual swimmer’’, and that ‘‘no simple relationship seems to link the speed of flexion–extension of the elbow joint and the EMG activity level’’ of biceps and triceps. In a third low scoring publication, Vaday and Nemessuri (1971) attempted to combine Fig. 6. Number of publications per type of EMG equipment. propulsive force measurements with EMG. The publication gives

23% MVC 21% integrated 47% 49% unknown no normalisaon raw

30% dynamic maximum 25% recﬁed

Fig. 7. Normalization methods (left panel) and EMG data treatment (right panel) used in EMG studies on competitive swimming strokes. 284 J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291 a description of muscles active during the various phases of front described the activation pattern of four muscles of the elbow joint crawl swimming and concluded that ‘‘in the traction period the (biceps, brachioradialis, flexor carpi ulnaris and triceps) and found tractor muscles and in case of pulsion the pulsor muscles show that during the early pull-through the highest activity in those main power’’. While it is impossible to determine the ‘‘normal’’ muscles was found, with biceps and brachioradialis as prime mov- muscle activation pattern in the four swimming strokes from these ers and with a clear antagonistic activity in biceps and triceps low quality publications, a series of four high scoring articles (Pink ‘‘most likely to maintain joint stability but also to supplement et al., 1991, 1992, 1993; Ruwe et al., 1994) do create a reference the prime movers’ action by equalizing the pressure distribution base for the muscle activation pattern of swimming respectively in the joints’’. Caty et al. (2007) confirmed these findings regarding crawl, backstroke and butterfly and breaststroke. Information on the flexor carpi ulnaris and its high activity during the early pull- the muscle activity in crawl is furthermore expanded in the work through, together with a strong extensor carpi ulnaris antagonistic of Lewillie (1976), Rouard and Clarys (1995), Monteil et al. activity to fully stabilize the wrist in this arm phase where the (1996), Caty et al. (2007) and Lauer et al. (2013). For clarity, the magnitude of the water resistance is high. In addition during this swimming phases of the arms of all publications are described in phase ‘‘the hand reaches the maximum depth of its trajectory with terms used by Pink et al. (1991). The arm movement starts with an upper limb close to an extension’’. Furthermore, the authors hand entry, followed by the early pull-through, the late pull- state that a lower stabilization of the wrist during the late pull- through, the exit, the early recovery and the late recovery. through could be associated with the decrease of water resistance. Lauer et al. (2013) continued on the study of Rouard and Clarys 4.2.1.1. Crawl stroke. Pink et al. (1991) described the muscle activ- (1995) and examined the coactivation patterns of biceps and tri- ity of 12 shoulder muscles in a group of competitive swimmers ceps brachii in depth. When referring to the earlier study, the with a rather high mean age (39 yrs). It was concluded that in authors state that although ‘‘the level of muscle activity observed crawl ‘‘the patterns of muscular activity at hand entry and exit were lower, EMG patterns were similar’’. Furthermore in the most were similar’’ for the prime movers in these two swimming phases recent study, the highest coactivation was found during the early as well as the recovery, being upper trapezius, rhomboids, supra- pull-through ‘‘to strongly stabilize the elbow joint to overcome spinatus and the middle and anterior deltoid. The muscles that water drag’’. A second high coactivation of both muscles was found were highly active during the underwater swimming phases were before the hand enters the water ‘‘to brake the end of the (recov- the pectoralis major and latissimus dorsi and can therefore be con- ery) movement and control the hand velocity just before its entry sidered as main propulsive muscles in the crawl. Pectoralis major into the water’’. To evaluate the variability of myoelectric signals, showed its peak activity of 71% of MVC (maximal voluntary con- Lewillie (1976) tested ‘‘ten swimmers of different levels of ability’’ traction) in the early pull-through phase and latissimus dorsi (peak over several trials, but presented the results of only one swimmer. of 75% of MVC) in the late pull-through phase when swimming at It was concluded that ‘‘the repeatability of swimming movements moderate pace. Teres minor followed closely the activation pattern by highly skilled swimmers appears to be exceptionally high, as of pectoralis major to stabilize the humerus. Regarding the remain- measured by both duration of and quantified electromyography. ing rotator cuff muscles ‘‘each had a unique role’’ in stabilizing A limited number of stroke cycles may therefore be accepted as scapula and humerus during all swimming phases and all showed valuable information’’. This conclusion should be adopted with more constant activity during the entire stroke cycle. great care, since the results of only one swimmer were presented, Some criticisms could be made on this article. The SDs of the and no detailed information on the swimming protocol is given. muscle activation patterns are high. Potentially within this partic- Furthermore, only one muscle was analyzed (triceps brachii). ipant group more than one typical activation pattern was present. Furthermore, the article describes the S-shaped underwater swim- 4.2.1.2. Backstroke. Pink et al. (1992) studied the same shoulder ming path of the hand which was the common swimming tech- muscles with the same equipment and protocol in backstroke as nique in crawl before 2000 but has since changed certainly in the was done for crawl in Pink et al. (1991). Similar to crawl, the del- sprint events to a more direct backward paddling movement. This toid muscles, supraspinatus and ‘‘scapular muscles’’, being rhom- change in technique might have affected the activation patterns for boids, upper trapezius and serratus anterior were mainly active example of pectoralis major since there is no inward sculling during hand entry and exit. These ‘‘scapular muscles formed a force movement of the hand any longer and consequently this publica- couple to position the scapula in a way that minimized impinge- tion might not describe the modern crawl swimming technique. ment of the rotator cuff muscles and maximized congruency of Finally, the study was executed using fine wire (needle) electrodes the glenoid and humeral head’’. ‘‘Latissimus dorsi, subscapularis inserted in the muscles, which might raise questions on the free- and teres minor were active primarily during the propulsive sub- dom of movement of the swimmers and therefore the representa- phase of the stroke’’, where ‘‘teres minor and subscapularis func- tiveness of their swimming trails with the EMG equipment. tioned together by forming a force couple to control humeral Following up on the study of Pink et al. (Rouard and Clarys, rotation’’. Pectoralis major played a less important role in the back- 1995) investigated the activity of latissimus dorsi and anterior del- stroke compared to crawl because in crawl the S-shape movement toid as well but came to different conclusions. In their much more of the hand is more pronounced then in backstroke. Similar prob- homogenous group of nine young swimmers (M age 17.3 yrs) of lems were identified to the previous article on crawl of variability, presumably higher performance level, latissimus dorsi did not freedom of movement and the mean age of the swimmers. show such a pronounced peak activity as in Pink et al. (1991). The anterior deltoid showed the opposite pattern as in Pink et al. (1991) with the lowest activity during late pull-through opposed 4.2.1.3. Butterfly. Pink et al. (1993) described the activity of the 12 to peak activity in this phase in Pink et al. (1991). As in both stud- shoulder muscles in butterfly very similarly to crawl with a clear ies, variability is high, comparing the mean activity should be trea- coactivation of all parts of deltoid, supraspinatus and infraspinatus, ted carefully. In Monteil et al. (1996) a high correlation between trapezius and rhomboids at hand entry and hand exit and recovery. muscular activities was found between teres minor and latissimus Pectoralis major and latissimus dorsi muscles acted again as main dorsi but not with pectoralis major as was found by Pink et al. propulsor muscles with teres minor and subscapularis closely (1991). Pectoralis major activity showed high correlation with linked to the activation of both these muscles to control the hum- supraspinatus but variability between individuals was again eral head and in the case of subscapularis to aid the latissimus reported to be high. The study of Rouard and Clarys (1995) also dorsi in the humeral internal rotation. J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291 285

4.2.1.4. Breaststroke. In a final study within this research topic, the for young non-swimmers than breaststroke. In the learning pro- activity of the 12 muscles was described in breaststroke (Ruwe cess, the change of body position in the water is most prominent et al., 1994). During the propulsive phase in breaststroke, pectoral- in crawl, going from a vertical to a more horizontal position. is major and latissimus dorsi showed their highest activity with teres minor counteracting the internal rotation of the humerus 4.2.2.1. Crawl. From a study that compared six competitive swim- caused by the pectoralis major and subscapularis preventing the mers with a PB on 100 m crawl in less than 60 s, and six PE students humeral head to lever out of the glenoid by latissimus dorsi. Dur- with a PB of around 80 s, it was concluded that in all muscles, the ing recovery the three heads of the deltoid muscles were clearly mean activation is higher in the more skilled swimmers. The largest more active to lift the arm, supraspinatus depress the humeral differences were found in the latissimus dorsi (4Â higher), rectus head, and upper trapezius and serratus anterior to maintain scap- abdominis (2Â higher) and gluteus maximus pars inferior (4Â ular stability. higher). The study also demonstrated a large inter individual variation among the competitive swimmers in all muscles investigated 4.2.2. Comparison of muscle activation patterns between different (Piette and Clarys, 1979). In the largest EMG swimming study found groups of participants for this review, also known as the Brussels EMG project, Clarys et al. Before dividing the discussion of this second research topic into (1983) reported comparisons between 30 competitive swimmers, the four competitive swimming strokes, a series of rather low scor- and ‘‘30 swimmers with good technical skills’’ for a total of 25 mus- ing studies will be discussed. The results of these studies combine cles. The main propulsive force in crawl swimming is produced by some swimming strokes or have totally different goals such as the arm and shoulder muscles, predominantly the latissimus dorsi, learn to swim or description of the function of the diaphragm. pectoralis major, triceps and deltoid, but this study also confirmed The first study in the history of EMG in swimming was published the idea reported previously by Piette and Clarys, that the trunk in 1964 (Ikai et al., 1964) and compared the EMG patterns of nine muscles and the gluteus maximus play a very important role in a male Olympic swimmers to five members of the university swim- good swimming technique and that differences between both par- ming club. It was concluded that the muscle activation pattern of ticipant groups were noticeable in the trunk–pelvis and leg muscle top swimmers was more effective and in crawl, breaststroke and ‘‘work intensity’’ which was higher in the better swimmers. A sec- butterfly, the extensors of arm and trunk contracted more vigor- ond large study was performed by Rouard and Billat (1990) where ously than the flexors to move the body forward. Backstroke was 30 male and 20 female swimmers were divided into three catego- an exception, where the opposite pattern was found. There was ries: international, national and regional level. When comparing no normalization of the EMG signals in this study which makes it the muscle activities relative to the dynamic maximum of the dif- hard to compare amplitudes between participants or participant ferent performance levels, ‘‘the best swimmers were characterized groups. The EMG diagrams given in this study were also not linked by lower muscular activity and a higher selective recruitment’’, to movement phases. In a unique study analyzing the differences in indicating once again greater efficiency. This study is the only found activity of the diaphragm between ‘‘experienced swimmers’’ and where differences between men and women were investigated. ‘‘champions’’, Delhez et al. (1971) found that ‘‘the respiratory rate There were no differences in patterns but women ‘‘were less effi- of the experienced swimmers is 30% higher to that of the champion cient in the sense that they had a lower speed for a same level of for the breast stroke and 60% superior for the crawl’’ again indicat- muscular activities’’. In 1993, Rouard et al. (1993) confirmed previ- ing more efficient use of muscles in higher level swimmers. ous findings when comparing 14 swimmers of different abilities Tokuyama et al. (1976) examined the characteristics of swim- (ranging from a PB/WR ratio on 100 m crawl from 78.7% to 98.4%) ming in infants and children, and compared them to unskilled concluding that the best swimmers tended to present higher force and skilled adults. It was concluded that crawl and backstroke pro- values for lower muscular recruitments implying a more efficient duced the same characteristics, namely that the ‘‘main discharges pattern and a higher potential for forces in elite swimmers. Com- of the hip and knee extensors and the coordinated movements of bining these findings, it could be concluded that better swimmers the upper and lower limbs in the infants and children in the initial show higher activity in stabilizing muscles around the trunk and period of learning to swim were about the same as the pattern of pelvis, are more selective and are better in timing the activation the skilled adults’’, with some clear differences found in the EMG peaks of their propulsive muscles and are able to reach a higher por- related to the early pull-through phase. ‘‘In breaststroke, the child’s tion of their potential. patterns were quite different from the skilled adult patterns’’. This could indicate that the crawl movement lies closer to the natural 4.2.2.2. Breaststroke. Yoshizawa et al. (1976) followed up on Ikai crossed coordination movement of walking on land and therefore et al. (1964) by giving a more detailed comparison between Olym- demands less time to learn, whereas breaststroke is a more com- pic breaststroke swimmers (65–68 s on 100 m breaststroke), uni- plex movement that needs more practice and experience. How- versity team swimmers and ‘‘average adults’’. They found clear ever, no statistical analysis was performed in this comparative differences in activation patterns e.g. a longer activation of the tib- study, and EMG was only analyzed in its raw form. In another Jap- ialis anterior in top level swimmers allowing a longer dorsiflexion anese study (Oka et al., 1979), the flutter kick was compared of the foot, better timing of the use of the gastrocnemius which between children and skilled adults and three children were tested prevented early plantar flexion of the foot seen in poor swimmers, for three consecutive years to examine their learning process. In resulting in a less effective kick and a high level of activity in the children who could not swim, an excessive flexion of the lower rectus abdominis in the better swimmers, stabilizing the pelvis. limbs was seen, much like a pedaling movement with the upper In the upper body, top level swimmers showed a higher and earlier body nearly in a vertical position. As swimmers became more activity in the biceps in the pull-phase, indicating an earlier elbow experienced, the upper body was nearly parallel to the surface, flexion than the lower level swimmers. They could therefore cor- but still no whip-like action of the lower limbs or sufficient move- rectly orientate the propulsive surface of the upper limb earlier ments of the thighs of the lower limbs was observed. In the last in the stroke. A weak point of this article is the analysis of raw stage of training, muscle recruitment patterns approached those EMG signals and the absence of normalization. of skilled adults. Again, no statistical analysis and normalization of EMG signals were done, as in the study of Tokuyama et al. 4.2.3. EMG pattern in variations of a swimming stroke (1976). Combining these two Japanese studies, it could be con- Early studies, often with small sample size, comparing muscle cluded that total coordination of the crawl stroke is easier to learn patterns and intensity of use agree that when speed increases, 286 J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291 muscle intensity increases as well. Later studies with larger sample (as in the 100% stroke frequency), fast-twitch muscle fibers make a sizes come to different conclusions however. greater contribution’’ and ‘‘the action potential amplitudes of fast- twitch units and their firing frequency have been shown to be 4.2.3.1. Crawl. In a single case study the amplitude of the triceps greater than those of the slow-twitch units’’ (Cabri et al., 1988). brachii reached 65% of its isometric maximum at low speed, Bollens et al. (1988) agree to a certain extent by stating that the 135% at training speed and 180% at sprint speed. In sprint, the ‘‘quality of the arm muscle EMG pattern tended to increase and relaxation time of the triceps brachii becomes almost negligible leg muscle to decrease’’ when decreasing the stroke frequency. as the forward extension of the arm during recovery becomes more For the explanation of how the authors assessed the quality of active (Lewillie, 1971b). These % of maximum isometric contrac- the EMG pattern, they refer to a paper presented at an Olympic tion were later corrected by the same author in Lewillie (1973) Scientific Congress that was unavailable. to 108%, 114% and 128% for triceps brachii. The activity of the rec- When considering variations in style of a certain swimming tus femoris was not high (47–56% of isometric maximum) in swim- stroke, it was concluded that ‘‘no important differences’’ were ming at low speed as the leg kick in crawl is only used to stabilize found in four arm muscles when comparing waterpolo crawl to the body. At sprint speed however, activity of this muscle increased regular crawl stroke (Clarys et al., 1973). Figueiredo et al. (2007) to 90% (Lewillie, 1973). Clarys et al. (1988b) reported an increase in analyzed differences in the three parts of the deltoid muscle when normalized intensity of four arm and shoulder muscles when recovering with laterally, and with a bend arm in crawl, both in swimming speed in crawl increased. However, when increasing inspiratory and non-inspiratory cycles. They concluded that ‘‘little the swimming speed in crawl with hand paddles, Monteil and differences were found between the two patterns, and in terms of Rouard (1992a), concluded that (except for the biceps brachii), muscular activity, there is not a better pattern of arm recovery’’. the greatest activity recorded was not obtained during the fastest When considering the specificity of ‘‘alternative training meth- pace, but during the medium pace. The authors fail to explain ods’’, it was concluded that fully tethered swimming (where the the reason for this divergence from previous work. The same group swimmer held in place by a cable fixed to a dynamometer on the however, also examined the influence of speed in 50 participants pool deck) produces similar arm muscle EMG patterns to free crawl and again concluded that there was no linear relationship between swimming, especially when swimming at a stroke frequency corre- speed and muscular activity of the upper limb. They stated that sponding to sprint velocity. In the leg musculature it is ‘‘difficult to muscular recruitment was greater at lower and higher speeds, conclude that they show similar EMG patterns in fully tethered and moderate for the intermediate speeds and that ‘‘the greater swimming compared with normal free swimming’’ (Bollens et al., activities of the legs, associated with a higher position of the body 1988). on the water, could explain the lower muscular activities of the When assessing the specificity of five dry land strength exer- upper limb for intermediate speeds’’ Rouard et al. (1992). cises for crawl, (Olbrecht and Clarys, 1983) were conclusive in stat- Three studies investigated the effect of crawl swimming with ing that ‘‘there is little electromyographic similarity between hand paddles. When comparing six hand paddle designs, Bollens swimming movements on dry land and the front crawl movement and Clarys (1985) concluded that paddles barely reproduce identi- under normal conditions’’ (in the water). The results of the water cal EMG diagrams to normal swimming, with the most identical tests were however not presented in their study, but only referred hand paddle scoring only 34.8% on the IDANCO classifying system. to. They concluded that ‘‘the idea of specific dryland training can- This system allows researchers to qualitatively assess whether an not be supported’’ by their results. On the contrary, Strass (1991) EMG curve is IDentical, Analogue or Conform another curve. The found that good sprinters were better able to activate the arm system is based on time amplitude similarities and differences shoulder muscles compared to swimmers with poorer sprint per- (Clarys and Rouard, 2011). Furthermore, the mean electrical inten- formance and that the testing apparatus on dry land used in this sity of the contraction of the muscles was found to be lower for all study could be made swim specific by imitating the positions of muscles and hand paddles tested but no explanation was given. the main phases of the arm stroke. Monteil and Rouard (1992b) reported that for crawl stroke there was no linear increase in muscular activity with increasing paddle 4.2.3.2. Backstroke. When speed increased in backstroke, dia- size compared to swimming without paddles, and with three dif- phragm contractions became more intense and longer (Delhez ferent paddle sizes. ‘‘Biceps brachii had the lowest activity, but et al., 1971). Backstroke showed a very similar pattern to crawl its demand seemed to be greater during the pull phase (angle in triceps and rectus femoris with increasing speed (Lewillie, arm–trunk 45–90°) when swimming with paddles compared to 1973). Even though the participants in Dupuis et al. (1979) swam without’’. In a following study (Monteil and Rouard, 1994), the ‘‘at moderate and sprint pace’’ in backstroke, the article failed to same authors confirmed their previous conclusions and stated that report differences in muscle patterns. muscular activity did not seem to be affected (increased or decreased or changed) by the use of paddles when swimming at 4.2.3.3. Butterfly. In butterfly, the increases in hip and knee ampli- equivalent intensity (80% of max speed on 100 m crawl with and tudes and angular velocities when speed increased ‘‘were reflected without paddles), although large individual differences were in corresponding increases in muscular activity levels’’ in leg mus- found. cles such as rectus femoris, biceps femoris, tibialis anterior and Another variation in crawl swimming was investigated by gastrocnemius and lower trunk muscles rectus abdominis and Clarys et al. (1988b) as they compared free swimming to swim- erector spinae between (Barthels and Adrian, 1971). These authors ming against mechanical resistance on the MAD system where however, reported a high variability in hip and knee amplitudes the swimmer pushes off with the hands on an underwater rail with among their four skilled butterfly swimmers. Butterfly showed pads. It was concluded that there was a very high degree of mus- the highest maximum intensities as % to maximal isometric con- cular pattern and intensity similarity in both conditions when tractions for triceps and rectus femoris muscles of all competitive using the IDANCO system and descriptive statistics (mean and swimming strokes at sprint speed (100% compared to 90% for standard deviation) to assess the variability. crawl, and 96% for breaststroke and backstroke in rectus femoris, When stroke frequency decreased, lower muscle intensities and 160% compared to 142% for backstroke, 128% for crawl and were found and the swimmer became more efficient in his force 120 for breaststroke in triceps brachii). The rise in intensity was production ‘‘probably because more time is taken to perform and present in both muscles when speed increased, but was more finish a full stroke’’. Furthermore, in ‘‘the fast contraction velocities pronounced in triceps brachii (Lewillie, 1973). J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291 287

4.2.3.4. Breaststroke. In breaststroke, no changes were found in the these shortcomings, the authors claimed that supraspinatus, infra- intensity of contraction of the rectus femoris when speed spinatus, deltoid and serratus anterior were predominately recov- increased. Triceps however doubled in intensity when sprinting. ery phase muscles in all three strokes. Latissimus dorsi and ‘‘The improvement in performance is due at normal speed to an pectoralis major were mainly used in the pull-through phase. effective increase of the duration of the leg thrust, when at sprint speed it is due to an increase of the arm pull’’ (Lewillie, 1973). 4.2.5. Biofeedback with EMG Differences in muscle activation patterns were found when In a unique study using EMG as a tool for feedback, Yoshizawa comparing continuous breaststroke to glide-stroke breaststroke et al. (1983) helped swimmers mimic the pattern of Olympic (Yoshizawa et al., 1978). In the glide stroke, the swimmer takes winners in backstroke, butterfly and breaststroke. They converted time to glide after the arm recovery phase. In continuous style the electrical discharges during the arm pull of four arm and shoul- breaststroke, the arms begin spreading immediately after the der muscles into audible signals, which were then given to seven recovery phase to generate propulsion. A difference was found in top level Japanese swimmers through waterproof earphones. Par- the kick phase in the activation sequencing of the gastrocnemius, ticipants were asked to bring the audible pattern of their activation tibialis anterior, biceps femoris and rectus femoris mainly caused in accordance with that of the Olympic winners. It was found that by a whip action of the feet toward the surface in the glide stroke. ‘‘only a short period of training of about 30 min was necessary to When considering the upper extremities, differences in discharge obtain sustained discharge patterns for all participants during the patterns were found in triceps brachii and the deltoid muscles arm pull period’’. The authors furthermore report that the change linked to the different recovery style. Daniel and Klauck (1999) in muscle patterns ‘‘did not lead to improvement in the swimming came to a similar conclusion in breaststroke as Bollens et al. records during or just after the training. However, 1–3 months (1988) did for crawl stating that ‘‘the muscular coordination is sim- later their swimming speeds improved’’. ilar between free swimming and during fully tethered swimming’’ in breaststroke. These authors were less clear in their assessment 4.3. Research questions 3 and 5: Which muscles have been investigated of the similarity of breaststroke-like movements performed on a and what types of equipment were used and how was data treated? swim bench. Even though the participants in Dupuis et al. (1979) swam ‘‘at moderate and sprint pace’’ in breaststroke, the article The choice of researchers to study certain muscles more than did not to report differences in muscle patterns. others is driven by two main motives. First and foremost, the function of certain muscles in propelling the body through the 4.2.4. Presentation or comparison of (new) EMG equipment water was already theorized before 1964 and was ever since well Three studies used a swimmer or swimmers swimming a com- documented in the EMG studies aiming to present the muscle pat- petitive swimming stroke to demonstrate a new EMG recording tern in one group of swimmers, in the studies that compare partic- technique or compare different techniques. In 1968, Lewillie ipant groups and in the studies that focused on different variations (1968) presented a two channel telemetric system to record mus- of a certain stroke. Furthermore, recently a book on the anatomy of cle activation in one crawl swimmer. In this system, a transmitter swimming was published (McLeod, 2010) describing the role of the was placed in a water-tight bag fixed to the back of the swimmer. musculoskeletal system in the four competitive swim strokes. The surface electrodes were fixed to the body with adhesive tape and were connected to the transmitter with an antenna sending the EMG signal the receiver on the pool deck up to 20 m away. Table 5 Only two muscles (triceps and quadriceps) were studied, and the Overview of population characteristics in EMG studies in competitive swimming results of the first published study on EMG in swimming with strokes. wired electrodes by Ikai et al. (1964) were confirmed. The quality Swimming level (%) of this study is the lowest in this review because it focused mainly Olympic (international) level 8.5 on the presentation of the equipment more so than the link to bio- National level 13.4 mechanics of the swimmer. In 1985 Clarys et al. (1985) compared a Competitive swimmers 24.2 University team level 2.9 wired with a telemetric system both in air and in water, and using Master level competitive swimmers 14.3 simple isometric contractions, as well as a crawl swimming proto- Students physical education 3.5 col on 11 swimmers. It was concluded that ‘‘both the telemetered Unskilled to skilled swimmers (non competitive) 19.5 and the conventional EMG registration systems give a similar Physicians trained in underwater diving 0.3 Unknown 13.3 pattern. Nevertheless there was a tendency, despite thorough pre- cautions (different tapes and varnish on the electrodes) for water Personal best times (PB) (%) Mean PB 100 crawl 91.85% of WR in 2007 (SD 4.89 s) 1.1 to decrease the detectable electrical output of human muscle’’. Mean PB 100 crawl 85.09% of WR in 1988 2.0 The authors fail to explain this phenomenon but it could be Mean PB 100 crawl 84.25% in 1985 1.2 assumed that there is a change in conductivity of the skin once Mean PB 100 crawl 82.24% of WR in 1995 (SD 2.5 s) 1.4 one enters the water. No links were made in this study between PB better than 82.4% of WR on 100 crawl in 1979 0.9 EMG and the crawl technique. Nuber et al. (1986) presented the (PR/WR) on 100 crawl 78.7–98.4% of WR in 1993 2.1 Mean PB 100 crawl 61.8% of WR in 1979 0.9 first swimming study using fine wire electrodes, although 7 years Mean PB 200 crawl 93.4% of WR in 2013 (SD 2.3 s) 1.5 earlier Okamoto and Wolf (1979) presented a study using these PB range of 50–60 s on 100 yard butterfly (in 1971) 0.6 type of electrodes in underwater walking. The purpose of the study Unknown 88.3 was to analyze deep rotator cuff muscles in crawl, breaststroke and Age (%) butterfly, which are virtually impossible to assess using surface M age 39 14.3 electrodes due to cross talk. The authors however encountered M age 20–27 yrs 17.2 M age 18–20 yrs 15.2 numerous difficulties using this technique (e.g. to find participants M age 15–18 yrs 4.5 willing to swim with needles in their body, and technical difficul- From 10 month old to adults 5.6 ties such as the leads that became inoperable and an extensive From 10 month old to 12 year old 4.7 pickup of noise). The same results were found as in later published From 19 to 35 years old 1.7 work of this group (Pink et al., 1991), but in a non-quantified way. Adults 1.2 Unknown 35.7 Raw EMG data without normalization was presented. Despite 288 J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291

Popular muscles that have a large function in propulsion are To compare muscle activity patterns and amplitudes between triceps brachii, latissimus dorsi, quadriceps and hamstring different participants, or different conditions or variations in one muscles, and in breaststroke pectoralis major can be added. In participant, the electrical signals have to be normalized. The the recovery phase, deltoid plays in important role, and for breast- method of normalization remains a debate in EMG research on stroke the hamstrings as well. Muscles that function principally as swimming. The majority of the studies in this review use the stabilizing muscles to counteract the forces generated on the body EMG of a maximal effort tested on dry land (MVC) as the normal- by the propelling muscles or to position the limbs as propelling izing factor. Clarys and Rouard (2011) state that ‘‘the use of the surfaces and that have been studied extensively are biceps brachii, MVC reference is perfect in all static, e.g. isometric applications’’. trapezius and smaller upper back muscles, the rotator cuff muscles, For all dynamic activities such as swimming, the use of an isomet- the muscles of the forearm, the lower trunk and hip muscles and ric reference is debatable (Clarys, 2000). In several swimming tibialis anterior and gastrocnemius. studies, the researchers report that swimmers exceed the 100% The second reason to choose certain muscles is the possibility of maximum derived from the MVC, with percentages up to 160% studying a muscle with surface EMG. The difficulties of working (Lewillie, 1973; Clarys, 1983a). Furthermore, Masumoto and with fine wire electrodes in swimming were highlighted by Mercer (2008) pointed out that there are differences between Nuber et al. (1986), although researches of the same group in MVC when tested on dry land to in water. To conclude, most California were able to study the deep rotator cuff muscles using researchers do not report what exactly they consider to be the this equipment in the early 1990s. The use of surface EMG with maximum. Is it the maximum value over a series of MVC trails, online data acquisition (wired) has led to choosing muscles that or the average of the maximum values of those trails? The alterna- were accessible and did not hinder the swimmer too much, accept- tive to MVC mostly used in swimming research is a method where ing the possibility of cross talk. The telemetric system, and very normalization is done to the highest peak activity in the dynamic recently the development of wireless EMG enabled researchers to condition, in this case swimming. This method counters the disad- study virtually all superficial muscles with less or no hinder for vantages of the MVC method, but is less internationally accepted the swimmer. and does not permit the comparison of amplitude within one

Deltoid anterior 80 60 40 Pink et al., 1991 20 % of MVC 0 Rouard & Clarys, 1995 Early pull Late pull Early Late through through recovery recovery

Lassimus dorsi 80 60 40 Pink et al., 1991 20 % of MVC 0 Rouard & Clarys, 1995 Early pull Late pull Early Late through through recovery recovery

Biceps brachii 80 60 40 Lauer et al., 2013 20 % of MVC 0 Rouard & Clarys, 1995 Early pull Late pull Early Late through through recovery recovery

Triceps brachii 80 60 40 Lauer et al., 2013 20 % of MVC 0 Rouard & Clarys, 1995 Early pull Late pull Early Late through through recovery recovery

Flexor carpi ulnaris 80 60 40 Caty et al., 2007 20 % of MVC 0 Rouard & Clarys, 1995 Early pull Late pull Early Late through through recovery recovery

Fig. 8. Mean activation patterns in crawl as a % of maximal voluntary contraction. J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291 289 participant for example between left and right since each muscle is differentiates a beginner from an expert on the muscular level, his own reference. An extensive overview of different normaliza- and what steps need to be taken in a learning curve. The use of tion methods is given by Burden (2010). biofeedback, as proposed by Yoshizawa et al. (1983) could help accelerate the learning process. 4.4. Research question 4: Which swimming activities and populations 3. The recent development of wireless surface EMG might create have been studied? opportunities to measure more muscles at the same time, less accessible muscles and with greater ease of use for researchers It is not surprising that crawl stroke is the activity most studied and comfort for the swimmer. using electromyography. Pereira et al. (2007) state this ‘‘might be 4. Front crawl is by far the stroke most investigated. The electro- associated to the fact of crawl being the fastest stroke, and the myographic study of the butterfly has room for exploration. one with the higher number of events in official competitions’’. 5. The swimming techniques used are seldom presented. Further- Crawl is also the stroke chosen by coaches to increase mileage in more these change over time. Example of this is the S-shaped training because it has a lower impact on the shoulder joint, com- curve the path of the hand made under water in crawl before pared to backstroke and butterfly, and on knee and hip joints com- 2000. There is little information on muscle patterns of modern pared to breaststroke. A final reason could be that the 100 m day high leveled swimmers in all four competitive swimming freestyle is widely regarded as the most prestigious event on strokes in the publications selected for this review as in most Olympic Games or international championships. It receives the recent studies on EMG in swimming, the focus is not on describ- highest media attention, and has the most competitors taking part. ing the muscle pattern, or differences in patterns, but in These four reasons attract researchers to this swimming stroke. describing fatigue using EMG. When considering the populations studied in EMG research in 6. The influence of the use of paddles on shoulder muscles has swimming, Table 5 shows that there is unfortunately a great deal not been investigated. This is surprising, since it is generally of unknown information. From the data available, it appears that accepted that the use of hand paddles causes more stress in the majority (53%) of the participants of whom we have informa- the shoulder joint and when shoulder injury is imminent, coa- tion were competitive to high leveled swimmers. 24% could be ches will eliminate pull sessions with paddles in training. While considered average leveled swimmers, and the remaining 22% studies on dry land show that shoulder injuries sometimes orig- low leveled swimmers or children. 72.5% of all participants of inate from lower trunk and lower limb instabilities (Kammer whom we know the age are older than 18. This is no surprise given et al., 1999), this hypothesis has never been tested in the fact that ethical committees are easier to pass when working swimmers. with adults and that most researchers were looking for a description of the pattern of a good execution of a certain stroke, and In an attempt to combine results of different studies, it was not of the pattern of beginners. found that EMG studies in swimming have no standardized manner of defining swimming phases, normalizing the data or of presenting the results. Furthermore, the variability around the 5. Conclusions mean activation patterns is large which makes it difficult to define a single muscle activation pattern applicable to all swimmers. To To summarize the findings of the top quality publications (scor- build up the knowledge in muscle activity patterns in swimming, ing above 80% on the quality assessment) for crawl which is by far a standardized way of presenting the data is needed. the swimming stroke most examined, Fig. 8 presents a generalized overview of the mean activation patterns of the muscles where at least two publications give numeric data. The graphs start and end Conflict of interest with the hand entry. Since the variability around the mean is large in each study, it is difficult to define one definitive muscle activa- The authors declare that there are no conflict of interest. tion pattern that applies to all swimmers. This kind of comparison is impossible for the remaining competitive strokes since compara- Acknowledgements ble data is unavailable. Although the study of EMG in swimming has a history of This project was financed by KU Leuven, Faculty of Kinesiology 50 years, this systematic review reveals some gaps in the literature. and Rehabilitation Sciences. The authors are thankful to the Department of Kinesiology, as well as the Federal University of 1. None of the studies performed on swimming focused on the Rio Grande do Sul. influence of the breathing action on skeletal muscles. This was only mentioned in passing by Figueiredo et al. (2007) but the purpose of their study was to compare different recovery References styles. A study on influence of breathing could be very interest- ing especially in freestyle and butterfly, were stroke cycles with Barthels KM, Adrian MJ. Variability in the dolphin kick under four conditions. In: and stroke cycles without breathing are alternated. It is Lewillie L, Clarys JP, editors. First international symposium on biomechanics in swimming, waterpolo and diving. Brussels, Belgium: Université Libre de assumed that better swimmers can maintain speed during their Bruxelles; 1971. p. 105–18. breathing actions, while beginners have a tendency to change Birrer P. The shoulder, EMG and the swimming stroke. J Swim Res 1986;2:20–3. their stroke rhythm while breathing. No muscular activity evi- Bollens E, Annemans L, Vaes W, Clarys JP. Peripheral EMG comparison between fully tethered and free front crawl swimming. In: Ungerechts BE, Wilke K, Reischle K, dence exists to support this assumption. Furthermore crawl editors. Swimming science V: proceedings of the 5th international symposium swimmers often have a preferred breathing side but little is of biomechanics and medicine in swimming. Bielefeld, Germany: Human known about the cause and the effect. Kinetics Publishers; 1988. p. 173–81. Bollens E, Clarys JP. Peripheral EMG control of handpaddle influence on swimming 2. The changes in muscle activation pattern during a learn- movements. In: Perren SM, Schneider E, editors. Biomechanics: current to-swim program are under evaluated. A longitudinal study interdisciplinary research, selected proceedings of the fourth meeting of the measuring EMG of swimmers with very limited experience with European Society of Biomechanics. Davos, Switzerland: Martinus Nijhoff Publishers; 1985. p. 699–704. a certain stroke at the beginning of a series of lessons, and at the Bradley JP, Tibone JE. Electromyographic analysis of muscle action about the end of the series of lessons could provide more insight in what shoulder. Clin Sports Med 1991;10:789–805. 290 J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291

Burden A. How should we normalize electromyograms obtained from healthy Bedingfield EW, editors. Swimming III, proceedings of the third international participants? What we have learned from over 25 years of research. J symposium of biomechanics in swimming. University of Alberta, Edmonton, Electromyogr Kinesiol: Off J Int Soc Electrophysiol Kinesiol 2010;20:1023–35. Canada: University Park Press; 1979. p. 110–7. Cabri JMH, Annemans L, Clarys JP, Bollens E, Publie J. The relation of stroke Figueiredo P, Fernandes RJ, Sousa A, Vilar S, Pereira S, Gonçalves P, et al. frequency, force, and EMG in front crawl tethered swimming. In: Ungerechts BE, Electromyographic analysis of the deltoideus muscle in two types of arm Wilke K, Reischle K, editors. Swimming science V: proceedings of the 5th recovery in front crawl. In: Arrelano R, Sánchez JA, Navarro F, Morales E, López G, international symposium of biomechanics and medicine in editors. Swimming science I. University of Granada; Granada, 2007. p. 157–62. swimming. Bielefeld, Germany: Human Kinetics Publishers; 1988. p. 183–9. Glousman R. Electromyographic analysis and its role in the athletic shoulder. Clin Caty V, Aujouannet Y, Hintzy F, Bonifazi M, Clarys JP, Rouard AH. Wrist stabilisation Orthop Relat Res 1993:27–34. and forearm muscle coactivation during freestyle swimming. J Electromyogr Hollander AP, De Groot G, van Ingen Schenau GJ, Toussaint HM, De Best H, Peeters Kinesiol 2007;17:285–91. W, et al. Measurement of active drag during crawl arm stroke swimming. J Clarys JP. A review of EMG in swimming: explanation of facts and/or feedback Sports Sci 1986;4:21–30. information. In: Hollander AP et al., editors. Biomechanics and medicine in Ikai M, Ishii K, Miyashita M. An electromyographic study of swimming. Res J Phys swimming: proceedings of the fourth international symposium of biomechanics Educ 1964;7:47–54. in swimming and the fifth international congress on swimming medicine, June Kammer CS, Young CC, Niedfieldt MW. Swimming injuries and illnesses. Phys 21–25, Amsterdam, 1982. Champaign, IL, United States: Human Kinetics Sportsmed 1999;27:51–60. Publishers; 1983a. p. 123–35. Kmet LM, Lee RC, Cook LS. Standard quality assessment criteria for evaluating Clarys JP. Hydrodynamics and electromyography: ergonomics aspects in aquatics. primary research papers from a variety of fields. Alberta Heritage Foundation Appl Ergon 1985;16:11–24. for Medical Research; 2004. Clarys JP. Application of EMG for the evaluation of performance in different sports. Lauer J, Figueiredo P, Vilas-Boas JP, Fernandes RJ, Rouard AH. Phase-dependence of In: Marconnet P, Komi PV, editors. Muscular function in exercise and elbow muscle coactivation in front crawl swimming. J Electromyogr Kinesiol training. Basel, Switzerland: Karger; 1987. p. 200–23. 2013;23:820–5. Clarys JP. The Brussels swimming EMG project. In: Ungerechts BE, Wilke K, Reischle Lewillie L. Telemetrical analysis of the electromyogram. In: Wartenweiler J, Jokl E, K, editors. Swimming science V: 5th international symposium of biomechanics Hebbelinck M, editors. Biomechanics I. Zurich, Switzerland: S. Karger, Basel; and medicine in swimming. Bielefeld, Germany: Human Kinetics Publishers, 1968. p. 147–9. Inc.; 1988. p. 157–72. Lewillie L. Graphic and electromyographic analysis of various styles of swimming. Clarys JP. The historical perspective of swimming science. In: Troup JP, Hollander In: Vredenbergt J, Wartenweiler J, editors. 2nd International seminar on AP, Strasse D, Trappe SW, Cappaert JM, Trappe TA, editors. Biomechanics and biomechanics. Eindhoven, The Netherlands: Karger, Basel; 1971a. p. 253–7. medicine in swimming VII. Taylor & Francis; 1996. p. xi–xxxiii. Lewillie L. Quantitative comparison of the electromyogram of the swimmer. In: Clarys JP. Electromyography in sports and occupational settings: an update of its Lewillie L, Clarys JP, editors. First international symposium on biomechanics in limits and possibilities. Ergonomics 2000;43:1750–62. swimming, waterpolo and diving. Brussels, Belgium: Université Libre de Clarys JP. Memoriam Prof. Dr. Léon Lewillie. In: Biomechanics and medicine in Bruxelles; 1971b. p. 155–9. swimming X: proceedings of the Xth international symposium on biomechanics Lewillie L. Muscular activity in swimming. In: Cerquiglini S, Venerando A, editors. and medicine in swimming, Porto, Portugal. In: Vilas-Boas JP, Alves F, Marques Biomechanics III, third international seminar on biomechanics. Rome: Karger; A, editors. Portuguese Journal of Sport Sciences. p. 9–12. 1973. p. 440–5. Clarys JP, Cabri J. EMG of repetitive and rhythmic (sports) movements. In: Lewillie L. Telemetry of electromyographic and electrogoniometric signals in Myklebust JB, Harris GF, editors. Proceedings of a special symposium on swimming. In: Nelson RC, Morehouse CA, editors. Biomechanics IV. 4th maturing technologies and emerging horizons in biomedical engineering. New International seminar on biomechanics. University Park, Orleans, United States; 1988. p. 58–9. Pennsylvania: University Park Press; 1974. p. 203–7. Clarys JP, Cabri J. Electromyography and the study of sports movements: a review. J Lewillie L. Variability of myoelectric signals during swimming. In: Komi PV, editor. Sports Sci 1993;11:379–448. Biomechanics V-B: proceedings of the fifth international congress of Clarys JP, Cabri J, De Witte B, Toussaint HM, de Groot G, Huying P, et al. biomechanics. Jyvaskyla, Finland: University Park Press; 1976. p. 230–4. Electromyography applied to sport ergonomics. Ergonomics Maes L, Clarys JP, Brouwer PJ. Electromyography for the evaluation of handicapped 1988a;31:1605–20. swimmers. In: Lewillie L, Clarys JP, editors. Swimming II: proceedings of the Clarys JP, Jiskoot J, Lewillie L. A kinematographical, electromyographical, and second international symposium on biomechanics in swimming. Brussels, resistance study of water-polo and competition front crawl. In: Cerquiglini S, Belgium: University Park Press; 1975. p. 268–75. Venerando A, editors. Biomechanics III, third international seminar on Masumoto K, Mercer JA. Biomechanics of human locomotion in water: an biomechanics. Rome: Karger; 1973. p. 446–52. electromyographic analysis. Exerc Sport Sci Rev 2008;36:160–9. Clarys JP, Massez C, Van Den Broeck M, Piette G, Robeaux R. Total telemetric surface McLeod I. Swimming anatomy. Human Kinetics; 2010. EMG of the front crawl. In: Matsui H, Kobayashi K, editors. Biomechanics VIII-B: Monteil KM, Rouard AH. Biomechanical aspects of paddle swimming at different proceedings of the eighth international congress of biomechanics. Nagoya, speeds. In: Maclaren D, Reiley T, Lees A, editors. Biomechanics and medicine in Japan: Human Kinetics Publishers; 1983. p. 951–8. swimming: swimming science VI: proceedings of the sixth international Clarys JP, Robeaux R, Delbeke G. Telemetered versus conventional EMG in air and symposium on biomechanics and medicine in swimming. Liverpool, UK: E & water. In: Winter DA, editor. Biomechanics IX-B: proceedings of the 9th FN Spon; 1992a. p. 63–7. international congress of biomechanics. United States: Human Kinetics Monteil KM, Rouard AH. Influence of the size of the paddles in front crawl stroke. In: Publishers; 1985. p. 286–90. Maclaren D, Reiley T, Lees A, editors. Biomechanics and medicine in swimming: Clarys JP, Rouard AH. The sculling-sweeping perspective versus the pull–push swimming science VI: proceedings of the sixth international symposium on perspective in free style swimming. In: Atkinson G, Reilly T, editors. Sport, biomechanics and medicine in swimming. Liverpool, UK: E & FN Spon; 1992b. p. leisure and ergonomics: proceedings of the third international conference on 99–104. sport, leisure and ergonomics. Burton, UK: E & FN Spon; 1995. p. 146–51. Monteil KM, Rouard AH. Free swimming versus paddles swimming in front crawl. J Clarys JP, Rouard AH. The frontcrawl downsweep: shoulder protection and/or Hum Movement Stud 1994;27:89–99. performance inhibition. J Sports Med Phys Fitness 1996;36:121–6. Monteil KM, Rouard AH, Dufour AB, Cappaert JM, Troup JP. Swimmers’ shoulder: Clarys JP, Rouard AH. The swimming muscle: history, methodology, and EMG of the rotators during a flume test. In: Troup JP, Hollander AP, Strasse D, applications of electromyography in swimming. In: Seifert L, Chollet D, Trappe SW, Cappaert JM, Trappe TA, editors. Biomechanics and medicine in Mujika I, editors. World book of swimming: from science to swimming VII. Liverpool, UK: Taylor & Francis; 1996. p. 83–9. performance. New York: Nova Science Publishers, Inc.; 2011. p. 43–68. Moynes DR, Perry J, Antonelli DJ, Jobe FW. Electromyography and motion analysis of Clarys JP, Toussaint HM, Bollens E, Vaes W, Huijing PA, de Groot G, et al. Muscular the upper extremity in sports. Phys Ther 1986;66:1905–11. specificity and intensity in swimming against a mechanical resistance – surface Nuber GW, Jobe FW, Perry J, Moynes DR, Antonelli D. Fine wire electromyography EMG in MAD and free swimming. In: Ungerechts BE, Wilke K, Reischle K, analysis of muscles of the shoulder during swimming. Am J Sports Med editors. Swimming science V: proceedings of the 5th international symposium 1986;14:7–11. of biomechanics and medicine in swimming. Bielefeld, Germany: Human Oka H, Okamoto K, Yoshizawa M, Tokuyama H, Kumamoto M. Electromyographic Kinetics Publishers; 1988b. p. 191–9. and cinematographic study of the flutter kick in infants and children. In: Cureton TK. Mechanics and kinesiology of swimming. Res Quart Am Phys Educ Terauds J, Bedingfield EW, editors. Swimming III, proceedings of the third Assoc 1930;1:87–121. international symposium of biomechanics in swimming. University of Alberta, Daniel K, Klauck J. Mechanical and electromyographical parameters in breaststroke Edmonton, Canada: University Park Press; 1979. p. 167–72. pull under different moving conditions. In: Keskinen KL, Komi PV, Hollander AP, Okamoto T, Wolf SL. Underwater recording of electromyographic activity using editors. Biomechanics and medicine in swimming VIII: proceedings of the VIII fine-wire electrodes. In: Terauds J, Bedingfield EW, editors. Swimming III, international symposium on biomechanics and medicine in proceedings of the third international symposium of biomechanics in swimming. Jyväskylä, Finland: Gummerus Printing; 1999. p. 33–9. swimming. University of Alberta, Edmonton, Canada: University Park Press; Delhez L, Pirnay F, Deroanne R, Petit JM. Analysis of ventilatory and cardiac activities 1979. p. 160–6. by autonomous memorizing during swimming and underwater diving. In: Olbrecht J, Clarys JP. EMG of specific strength training exercises for the front crawl. In: Lewillie L, Clarys JP, editors. First international symposium on biomechanics in Hollander AP, Huijing PA, de Groot G, editors. Biomechanics and medicine in swimming, waterpolo and diving. Brussels, Belgium: Université Libre de swimming: proceedings of the fourth international symposium of biomechanics Bruxelles; 1971. p. 145–54. in swimming and the fifth international congress on swimming Dupuis R, Adrian MJ, Yoneda Y, Jack M. Forces acting on the hand during swimming medicine. Amsterdam, The Netherlands: Human Kinetics Publishers; 1983. p. and their relationships to muscular, spatial, and temporal factors. In: Terauds J, 136–41. J. Martens et al. / Journal of Electromyography and Kinesiology 25 (2015) 273–291 291

Pereira SM, Vilar S, Gonçalves P, Fernandes S, Fernandes RJ, Roesler H, et al. Yoshizawa M, Tokuyama H, Okamoto K, Kumamoto M. Electromyographic study of Electromyographic analysis of the flip turn technique. In: Menzel HJ, Chagas the breaststroke. In: Komi PV, editor. Biomechanics V-B: proceedings of the fifth MH, editors. Proceedings of the 25th international symposium on biomechanics international congress of biomechanics. Jyvaskyla, Finland: University Park in sports. Ouro Preto, Brazil; 2007. p. 55–8. Press; 1976. p. 222–9. Piette G, Clarys JP. Telemetric EMG of the front crawl movement. In: Terauds J, Bedingfield EW, editors. Swimming III, proceedings of the third international symposium of biomechanics in swimming. University of Alberta, Edmonton, Canada: University Park Press; 1979. p. 153–9. Jonas Martens, is Master in Physical Education; Pink M, Jobe FW, Perry J, Kerrigan J, Browne A, Scovazzo ML. The normal shoulder Master in Biomedical and Clinical Engineering Tech- during the backstroke: an EMG and cinematographic analysis of 12 muscles. niques; finalizing a Ph.D. at KU Leuven (Belgium) in Clin J Sport Med 1992;2:6–12. the field of electromyography and swimming; Pink M, Jobe FW, Perry J, Kerrigan J, Browne A, Scovazzo ML. The normal shoulder swimming finalist at the Sydney and Athens Paral- during the butterfly swim stroke. An electromyographic and cinematographic ympic Games and former World Record holder. analysis of twelve muscles. Clin Orthop Relat Res 1993:48–59. Pink M, Perry J, Browne A, Scovazzo ML, Kerrigan J. The normal shoulder during freestyle swimming. An electromyographic and cinematographic analysis of twelve muscles. Am J Sports Med 1991;19:569–76. Pink MM, Tibone JE. The painful shoulder in the swimming athlete. Orthop Clin North Am 2000;31:247–61. Rouard AH. Muscle fatigue in swimming. In: Kjendlie P-L, Stallman RK, Cabri J, editors. XIth international symposium for biomechanics & medicine in swimming. Oslo: Norwegian School of Sport Science; 2010. p. 33–5. Rouard AH, Billat RP. Influences of sex and level of performance on freestyle stroke: an electromyography and kinematic study. Int J Sports Med 1990;11:150–5. Rouard AH, Clarys JP. Cocontraction in the elbow and shoulder muscles during rapid Pedro Figueiredo Postdoctoral fellow at the School cyclic movements in an aquatic environment. J Electromyogr Kinesiol of Physical Education, Federal University of Rio 1995;5:177–83. Grande do Sul, Brazil. Member of the Centre of Rouard AH, Quezel G, Billat RP. Effects of speed on EMG and kinematic parameters Research, Education, Innovation and Intervention in in freestyle. In: Maclaren D, Reiley T, Lees A, editors. Biomechanics and Sport, University of Porto, Portugal. His research medicine in swimming: swimming science VI: proceedings of the sixth interests are related to the bioenergetics and bio- international symposium on biomechanics and medicine in mechanics of exercise and are focused on the study of swimming. Liverpool, UK: E & FN Spon; 1992. p. 93–7. human locomotion. Rouard AH, Schleihauf RE, Troup JP. Relationships between hand forces and EMG in freestyle swimming. Abstracts of the XIVth congress of the international society of biomechanics. Paris, France; 1993. p. 1150–1. Ruwe PA, Pink M, Jobe FW, Perry J, Scovazzo ML. The normal and the painful shoulders during the breaststroke. Electromyographic and cinematographic analysis of twelve muscles. Am J Sports Med 1994;22:789–96. Strass D. Force–time and electromyographic characteristics of arm shoulder muscles in explosive type force production in sprint swimmers. J Swim Res 1991;7:19–27. Tokuyama H, Okamoto K, Kumamoto M. Electromyographic study of swimming in Daniel Daly is Professor of Kinesiology and Rehabil- infants and children. In: Komi PV, editor. Biomechanics V-B: proceedings of the itation Science at the KU Leuven, Belgium. He was fifth international congress of biomechanics. Jyvaskyla, Finland: University Park born in Cambridge, Mass (USA) and was educated at Press; 1976. p. 215–21. Springfield College, the Pennsylvania State University Vaday M, Nemessuri M. Motor pattern of free-style swimming. In: Lewillie L, Clarys and the KULeuven, Belgium. His main research topic JP, editors. First international symposium on biomechanics in swimming, has been on Biomechanics of Swimming and more waterpolo and diving. Brussels, Belgium: Université Libre de Bruxelles; 1971. p. 167–73. recently in the area of ‘‘Physical activity in water for Yoshizawa M, Okamoto K, Kumamoto M, Tokuyama H, Oka H. Electromyographic persons with chronic disabilities: Competition, Rec- study of two styles in the breaststroke as performed by top swimmers. In: reation, Therapy’’. Asmussen E, Jørgensen K, editors. Biomechanics VI-B. Proceedings of the 6th international congress of biomechanics. Copenhagen: University Park Press; 1978. p. 126–31 (1977). Yoshizawa M, Okamoto T, Kumamoto M. Effects of EMG-biofeedback training on swimming. In: Matsui H, Kobayashi K, editors. Biomechanics VIII-B: proceedings of the eighth international congress of biomechanics. Nagoya, Japan: Human Kinetics Publishers; 1983. p. 828–32.