biomechanics of cycling

97 downloads 0 Views 4MB Size Report
application of biomechanical principles to the sport and exercise of cycling. ...... objectives such as leisure, travel to work and to transport small items in ...... changing muscle recruitment in order to give priority for activation of muscles that.
 

BIOMECHANICS OF CYCLING

 

EDITORS RODRIGO R BINI, PhD FELIPE P CARPES, PhD

Publisher: Adis-Springer

2

This book is dedicated to the memory of Prof. Antônio Carlos Stringhini Guimarães who was our first supervisor and a passionate by biomechanics and cycling. Prof. Guimarães gave us the opportunity to study cycling and involved us with his passion on this unique sport. Sadly, Prof. Guimarães died after being hit by a car when cycling his bicycle in a morning of October 2005. We hope this book somehow accomplishes his wills.

3 EDITED BY

Rodrigo Rico Bini, BPhysEd, MSci, PhD Exercise Research Laboratory Universidade Federal do Rio Grande do Sul Rua Felizardo, 750 - Bairro Jardim Botânico Porto Alegre - RS ZIP 90690-200 Brazil Email: [email protected]

Felipe Pivetta Carpes, BPhysEd, MSci, PhD Applied Neuromechanics Research Group Federal University of Pampa Faculty of Health Sciences Campus Uruguaiana, BR 472, km 592, PO box 118 Uruguaiana - RS ZIP 97500-970 Brazil Email: [email protected]

4 FOREWORD

Recent decades have witnessed a remarkable expansion in the knowledge and application of biomechanical principles to the sport and exercise of cycling. There has been wide publication of new experimental findings in the biomechanics of cycling. Cycling coaches and cyclists have been quick to realize the importance of acquiring new biomechanical cycling knowledge that they may put to effective use. New findings often prove to be invaluable for the enhancement and safety of cycling performance. In that regard, biomechanics through their research have sought for ceilings in cycling responses in an attempt to identify those factors that limit or enhance cycling efficiency. Their work, discussed in this text, provides fresh understandings how numerous personal and environmental variables contribute to cycling excellence. One goal of this text is to meet the

growing

need

for

a

comprehensive

presentation

of

contemporary

cycling

biomechanical knowledge which will positively influence the activity of cycling in a global fashion. To achieve this goal the present text bridges the gap between the scientist and the cyclist. In doing so it becomes a valuable resource for cycling coaches and cyclists. It also provides useful cause-effect insight for sports medicine practitioners and physical therapists that must address cycling-specific anatomical stress and injury. The general aim of this text is to afford the reader with an up-to-date foundation for understanding the biomechanics of cycling.

To accomplish this, the focus of text is on promoting an

understanding of the biomechanical principles that govern the effectiveness of cycling. Biomechanical fundamentals within the conceptual framework of cycling performance are reviewed. Throughout the text the authors provide interpretation of research findings and their application to cycling efficiency and safety. An effort has been made to keep the text as concise and clear as possible yet as comprehensive as necessary. Illustrations and examples are commonly used to clarify and further explain biomechanical issues. For clarity and understanding, the most striking graphic presentations and illustrations are employed. Sufficient detail is presented for those who may lack a strong biomechanical background.

To facilitate a better

understanding of some of the biomechanical events encountered during cycling, a certain amount of basic physics is included. straightforward style.

Concepts are presented in a clear and

Using this format, the reader is presented with the necessary

5 knowledge skills and technology to identify the critical aspects necessary to experience cycling success. The text is organized into ten chapters with each chapter beginning with an introduction to the topic of discussion and ending with conclusions and practical applications that places theory and knowledge into practice. Uniqueness of this text is the inclusion of cycling overuse injuries, and their prevention. Acute and chronic adaptations to the peculiarities of cycling are reviewed and discussed. Each chapter of the book contains extensive practical applications, including equipment

and

procedures

to

evaluate

cycling

performance.

Biomechanics

instrumentation used for quantitative analysis are illustrated and described. Unique to this text is Chapter 9 that includes an example research design that compares two types of pedal force feedback during 4-km time trial training on a bicycle ergometer. I firmly believe this text will be a valuable addition to libraries of those who perform cycling research, engage in clinical practice related to exercise science, practice sports medicine, serve as cycling coaches or are professional, amateur or recreational cyclists. It is hoped that the information presented will entice you to seek additional knowledge in this exciting area of the biomechanics of cycling. Irvin E. Faria, Ph.D., FACSM

6 CONTENTS Foreword ................................................................................................................................4  List of contributors ..................................................................................................................9  Preface .................................................................................................................................11  Acknowledgements ..............................................................................................................12  CHAPTER 1: Introduction to biomechanical analysis for performance enhancement and injury prevention ...................................................................................................................13  Rodrigo R. Bini and Felipe P. Carpes ..................................................................................13  1.1 Introduction .............................................................................................................................. 13  1.2 Anthropometry ......................................................................................................................... 14  1.3 Motion analyses ....................................................................................................................... 16  1.4 Force and pressure analysis...................................................................................................... 18  1.5 Muscle activation ..................................................................................................................... 19  1.6 Inverse dynamics and muscle modeling .................................................................................. 22  1.7 Conclusions and practical applications .................................................................................... 24  CHAPTER 2: Measuring pedal forces .................................................................................25  Rodrigo R. Bini and Felipe P. Carpes ..................................................................................25  2.1 Introduction .............................................................................................................................. 25  2.2 Pedal force components ........................................................................................................... 25  2.3 Principles of force measurements in instrumented pedals ....................................................... 26  2.4 Historical changes in methods for force measurements........................................................... 28  2.5 Angular sensors for kinematics assesments ............................................................................. 31  2.6 Sources of errors in pedal force measurements ....................................................................... 32  2.7 Conclusions and practical applications .................................................................................... 33  CHAPTER 3: Muscle activity ...............................................................................................35  Rodrigo R. Bini and Felipe P. Carpes ..................................................................................35  3.1 Introduction .............................................................................................................................. 35  3.2 Methodological aspects and use of EMG in cycling................................................................ 37  3.3 Muscle activation during cycling ............................................................................................. 39  3.3.1 Workload ........................................................................................................................................ 40  3.3.2 Pedaling cadence ............................................................................................................................ 41  3.3.3 Body position on the bicycle .......................................................................................................... 42  3.3.4 Fatigue ............................................................................................................................................ 43  3.3.5 Cycling skill ...................................................................................................................................... 43 

3.4 Conclusions and practical applications .................................................................................... 44  CHAPTER 4: Joint kinematics .............................................................................................46  Felipe P. Carpes, Rodrigo R. Bini and Jose I. P. Quesada .................................................46  4.1 Introduction .............................................................................................................................. 46  4.2 Crank kinematics...................................................................................................................... 49  4.3 Upper body motion during cycling .......................................................................................... 50 

7 4.4 Lower body motion during cycling.......................................................................................... 52  4.5 Conclusions and practical applications .................................................................................... 57  CHAPTER 5: Kinetics and pedalling technique ...................................................................58  Rodrigo R. Bini and Mateus Rossato...................................................................................58  5.1 Introduction .............................................................................................................................. 58  5.2 Pedalling kinetics ..................................................................................................................... 58  5.3 Pedal forces .............................................................................................................................. 60  5.4 Pedalling technique .................................................................................................................. 63  5.5 Factors affecting pedal force effectiness and pedaling technique............................................ 65  5.5.1 Workload level ................................................................................................................................ 66  5.5.2 Pedalling cadence ........................................................................................................................... 67  5.5.3 Body position on the bicycle .......................................................................................................... 67  5.5.4 Fatigue ............................................................................................................................................ 68  5.5.5 Skill and experience ........................................................................................................................ 69 

5.6 Pedaling technique vs. performance ........................................................................................ 69  5.7 Conclusions and practical applications .................................................................................... 70  CHAPTER 6: Non-traumatic injuries in cycling....................................................................71  Rodrigo R. Bini and Thiago A. Di Alencar ...........................................................................71  6.1 Introduction .............................................................................................................................. 71  6.2 Occurence of overuse injuries in cycling ................................................................................. 71  6.3 Overuse injury mechanisms in cycling .................................................................................... 73  6.4 Proposed strategies to reduce overuse injury risk .................................................................... 76  6.5 Conclusions and practical applications .................................................................................... 80  CHAPTER 7: Bicycle types and sizes .................................................................................82  Rodrigo R. Bini, Frederico Dagnese and Julio Kleinpaul ....................................................82  7.1 Introduction .............................................................................................................................. 82  7.2 Types of bicycles ..................................................................................................................... 83  7.2.1 Road bike ........................................................................................................................................ 83  7.2.2 Mountain bike ................................................................................................................................ 84  7.2.3 Hybrid bicycles ................................................................................................................................ 84  7.2.4 Triathlon bicycles ............................................................................................................................ 85 

7.3 Frame size ................................................................................................................................ 86  7.4 Crank length and noncircular chainrings ................................................................................. 89  7.5 Handlebars ............................................................................................................................... 90  7.7 Conclusions and practical applications .................................................................................... 91  CHAPTER 8: Optimizing bicycle configuration and cyclists’ body position to prevent overuse injury using biomechanical approaches .................................................................92  Rodrigo R. Bini, Patria A. Hume, James L. Croft and Andrew E. Kilding ............................92  8.1 Introduction .............................................................................................................................. 92 

8 8.2 Methods.................................................................................................................................... 93  8.3 Findings.................................................................................................................................... 93  8.4 Optimization of configuration of bicycle components and cyclists’ body position ................ 94  8.4.1 Handlebars vertical and horizontal position .................................................................................. 94  8.4.2 Saddle height and horizontal position ............................................................................................ 96  8.4.3 Position of the foot on the pedal ................................................................................................... 96 

8.5 Biomechanical approaches for optimizing bicycle configuration and cyclists’ body position to prevent overuse injury .................................................................................................................... 98  8.5.1 Anthropometrics .......................................................................................................................... 102  8.5.2 Dynamic joint kinematics ............................................................................................................. 103  8.5.3 Pedal forces and joint kinetics ...................................................................................................... 104  8.5.4 Muscle activation ......................................................................................................................... 105 

8.6 Future research ....................................................................................................................... 106  8.7 Conclusions and practical applications .................................................................................. 107  CHAPTER 9: Pedalling technique changes with force feedback training in competitive cyclists: Preliminary study ................................................................................................. 108  Rodrigo R. Bini, Patria A. Hume and James L. Croft ....................................................... 108  9.1 Introduction ............................................................................................................................ 108  9.2 Methods.................................................................................................................................. 110  9.2.1 Participants and allocation to pedal force feedback groups ........................................................ 110  9.2.2 Pre‐training protocol .................................................................................................................... 111  9.2.3 Training sessions ........................................................................................................................... 111  9.3.4 Data analyses ................................................................................................................................ 114  9.3.5 Statistical analyses ........................................................................................................................ 114 

9.4 Results .................................................................................................................................... 114  9.5 Discussion .............................................................................................................................. 120  9.6 Conclusions and practical applications .................................................................................. 122  CHAPTER 10: Technology in cycling ............................................................................... 123  Rodrigo R. Bini and Felipe P. Carpes ............................................................................... 123  10.1 Introduction .......................................................................................................................... 123  10.2 Motion analisys .................................................................................................................... 124  10.3 Force measurements ............................................................................................................. 128  10.4 Muscle function.................................................................................................................... 129  10.5 Aerodynamics ...................................................................................................................... 131  10.6 Conclusions and practical applications ................................................................................ 134  Editors’ conclusion ............................................................................................................ 135  REFERENCES ................................................................................................................. 136

9 LIST OF CONTRIBUTORS

Jose Ignacio Priego Quesada, Lic PhysEd Researcher en GIBD (Research Group in Sports Biomechanics) PhD student at the Instituto de Biomecánica de Valencia (IBV) Universidad de Valencia, Spain Email: [email protected]

Mateus Rossato, MSci Senior Lecturer Faculty of Physical Education and Physiotherapy Federal University of Amazonas Manaus, Brazil Email: [email protected]

Frederico Dagnese, MSci Lecturer at Portal Fitness Centre Santa Maria, Brazil Email: [email protected]

Thiago Ayala Di Alencar, FT Bike Fit Studio Fisio Vitale Clinical Centre Goiania, Brazil Email: [email protected]

Júlio Kleinpaul, MSci Lecturer at São Francisco de Barreiras Faculty Barreiras, Brazil Email: [email protected]   

Patria Anne Hume, PhD Professor at AUT University, New Zealand Director of the Sports Performance Research Institute New Zealand

10 Email: [email protected]

James Croft, PhD Research Fellow at Otago University, New Zealand Email: [email protected]

Andrew Kilding, PhD Associate Professor at AUT University, New Zealand Sports Performance Research Institute New Zealand Email: [email protected]  

11 PREFACE Research became part of our lives in early stages of our undergraduate course when we had the opportunity to be involved with studies on biomechanics of human motion. The nature of the motion at that time involved cycling and questions on how we could improve the configuration of bicycle components in order to enhance performance of cyclists was around. We ended learning more on the biomechanics of cyclists then we could imagine but there were always lots of questions that no study we found on the literature could explain. For that reason we kept going towards a more clear understanding on why and how humans could take their best during bicycle riding. More than ten years after that time, we had the opportunity to provide our perspective on the biomechanics of cycling. We do not expect that this book will become a final word, although some experts believe that research on cycling biomechanics is a “dead issue”. A contrary argument is that day after day cyclists come up with different strategies on their pedaling that we cannot easily explain how it works. We then go back to our laboratories and do research to see how much effect a novel setting could have on varying parameters associated to performance or injury prevention. In the end, I am not expecting to see a final word because creation of new methods is potentially a continuum. We only expect that cycling in future years could elicit a better interaction between cyclists and their bicycles then in the past. We expect that this book should provide a primary source for new students in the field of cycling biomechanics and hopefully an update for those how have been involved for a long time. The aim was to provide an overview on the complex interaction between the bicycle and the human body and we are certain that there is more to sayon cycling biomechanics away from the contents in this book. To accomplish this goal, Dr. Felipe Carpes, a very long term fellow and friend, helped me to take together ideas and content in order to have a book that could be used as a stepping stone for those who are willing to get fascinated with the biomechanics of cycling. We also had important help from fellows who are co-authors in various chapters. Without their support we would not have completed the hard work of finishing this book. At last, we hope this book improves the link between practice and science in cycling, given most of the time there is a long bridge between empirical knowledge of cyclists and coaches and evidence from research in this field. Rodrigo Rico Bini January 2014, Porto Alegre - Brazil

12 ACKNOWLEDGEMENTS We would like to thank reviewers who provided feedback on the very early stage of this book. Their inputs helped us to increase the quality of our book in its various chapters. They were all experts in biomechanics and/or human movement sciences: Ms. Alice Cavalheiro Tamborindeguy Dr. Carlos Bolli Mota Dr. André Cervieri Dr. Cláudia Tarragô Candotti Dr. Cláudia Silveira Lima Mr. Fabricio Duarte Dr. Flávio Antônio de Souza Castro Mr. Frederico Dagnese Dr. Reginaldo Fukushi Mr. Ricardo Peterson da Silveira We would like to thank Prof. Irvin Faria for his very kind opening words in this book and for all support provided in many of our research projects. We would like to thank all fellows who helped us during our undergraduate and graduate studies and all members of the Cycling Research and Study Group (GEPEC – www.gepecbrasil.com) for their support.  

Rodrigo Bini and Felipe Carpes  

13 CHAPTER

1:

INTRODUCTION

TO

BIOMECHANICAL

ANALYSIS

FOR

PERFORMANCE ENHANCEMENT AND INJURY PREVENTION

Rodrigo R. Bini and Felipe P. Carpes

1.1 Introduction Bicycles have become complex due to an increased need for reducing energy cost during pedaling [1]. The changes from the Hobby Horse to modern bikes have led to a large number of components being individually set to accommodate cyclists of varying body dimensions. In parallel, racing profiles has become more complex, and bicycles are now used on roads, tracks and other types of surfaces. Moreover, different disciplines in competitive cycling have expanded to include road, track, mountain bike, BMX and even merging bike to running and swimming in triathlon. To enable optimal performance in these cycling events, bicycles have been substantially changed and they now largely differ among those disciplines, including dimensions and weight. In long tours performed by road cyclists, it is possible to observe cyclists using particular bicycles for different stages, in order to better accommodate for differences in duration and geographies of the race. Along with the solid use in competitive sport, bicycles have been a common device to enhance physical fitness level in gyms and training centers. For fitness monitoring, cycle ergometers have been developed to enable the measurements of mechanical work during exercise testing. For rehabilitation purposes, bicycles are an option to reduce the stress after anterior cruciate knee ligament (ACL) reconstructions [2], for people who have chondromalacia patellae [3] or menisci ruptures [4] rather than walking or running. For transportation, bicycles have been an alternative to engine or animal based vehicles for decades. In Amsterdam, approximately 20% of travels involve bicycles. In Denmark, the use of bicycles are popular since childhood [5]. These reports enforce the option for bicycles as an alternative vehicle for transportation as indicated by the World Commission on Environment and Development. In that regard, the bicycle reduces the number of automobiles in large cities thereby lowering pollution from oil, petrol and gas based engines. In parallel to a greater number of cyclists there is an increased number of traumatic and overuse based injuries [6]. Reports indicate that 85% of cyclists could sustain one or

14 more sites of physical pain [7]. These sites vary from low back, knees to the groin areas, particularly when bicycles are rode for longer than three hours [8]. Assuming that riding bicycles could lead to overuse injuries, prevention could be accomplished by setting bicycle components to the dimensions of the cyclist [9-11]. Along with the common practice of matching bicycle to the body dimensions, the analysis of joint angles, force production and muscle actions during pedaling have been referred to improve the configuration of bicycles [12]. Merging motion analysis and external forces could lead to an estimative of forces acting on the related joints using biomechanical modeling [3]. Patellofemoral compressive forces have been shown to change depending on workload and pedaling cadence [13], which could provide a predictive perspective for overloading the knee joint. Throughout this book, biomechanical methods will be presented with application to the assessment of cyclists in order to provide resources for the prevention of injuries and potential improvements in performance of cyclists with aims ranging from commuting to racing. A brief introduction will be present along this opening chapter.

1.2 Anthropometry The most common anthropometric variables used in human movement analysis are probably body mass and standing height. However, for cycling, these two variables are relatively limited without measurements of segments lengths. Given that the dimensions of the bicycle are varied, it is essential that the set of fixed and moving parts are presented according to the rider's body dimension to maximize comfort [14]. The adjustment of the bicycle to the cyclist's body is one of the key factors that affect the magnitude of the forces produced by muscles, and therefore, the load applied to the joints [15]. In a simple mechanical layout including a rigid segment rotating over a hinge axis, the torque produced depends on the lever arm and the magnitude and direction of the applied force. In cycling, segment lengths act like levers transferring forces from the upper to the lower parts of the body towards the pedals and cranks. Therefore, the question is raised as to how the combination of settings in bicycle components could then affect the lengths of levers for body segments to transfer force to the pedals and cranks. That question could be answered if measurements of muscle forces in vivo were possible, which are not very often conducted due to invasive procedures. To better understand muscle forces, biomechanists have measured forces applied to the pedals and cranks, which will be latter introduced in this chapter.

15 In Figure 1.1, bicycle and body dimensions are presented to highlight potential changes needed in bicycles to accommodate different lengths for segments of the cyclist.

Fig 1.1. (A) Illustration of saddle height and saddle to handlebar height and length. (B) Lower limb measure is commonly used to configure saddle height (1- inseam length, 2- throcanteric length and 3- ischial tuberosity length).

In Figure 1.1-A, we can see that bicycles enable users to change the distance from the saddle to the pedals and the horizontal distance from the saddle to the handlebars. These settings are somewhat related to lower limb dimensions (i.e. throcantherion height shown in Figure 1.1-B). As an example, optimal saddle height should be 96-100% of the trochanterion height [16]. Bicycles are designed with varying components and large differences among models are common. As an example, saddles can be changed from the traditional to a more complex design when the anterior node is removed to reduce pressure in groin areas [17]. Similarly, wheels and frames are designed to improve aerodynamical profile [18,19]. Apart from the potential changes in each component, it is critical to allow minimum changes in a bicycle in order to accommodate a small range of body dimensions. For that purpose, saddle vertical (height) and horizontal positions are possible to adjust even in very basic bicycle models. Assuming that body dimensions should ideally match bicycle configurations, De Vey Mestdagh [10] presented a series of configurations of different components of the bicycle and their link to body dimensions. These have been largely used as an initial “guess” for selecting bicycle sizes and configuration of components. However, studies looking at the

16 relationship between configurations of components taken from body dimensions and comfort showed a weak link [20,14]. For professional cyclists, reduced comfort has been found when cyclists were asked to use predictive “optimal” bicycle settings [20]. Similar results were observed in commuting cyclists with further findings on the weak relationship between self-selected saddle to handlebars length (as seen in Figure 1.1-A) and grip reach [14]. Therefore, it is unclear how valid predictive equations could be for definitions of “optimal” or “ideal” configuration of bicycle components [21]. Among bicycle components that cannot be easily set, the crank length has been on focus for some studies [22,23]. In the field, crank length is generally set based on upper leg length and cycling discipline [24]. Cyclists with longer thigh length tend to use longer crank lengths than smaller subjects. However, varying crank lengths did not result in substantial changes in joint power production during maximal and sub maximal cycling [25]. Therefore, it is unlikely that small changes in crank length could substantially improve cycling performance, and it could explain why a similar crank arm length is very often observed among a range of cyclists. Moreover, few studies showed that changing crank length [23] or chainring shape throughout the crank cycle [26] could improve cycling performance. In contrary, most attempts to change these components had no effect in sub maximal cycling performance [27]. Although anthropometric dimensions of the human body are critical when selecting bicycle size and its components, physiological profile (i.e. muscle force capacity, fiber type distribution) and motor control (i.e. motor unit recruitment and skill) could largely affect cycling performance compared to the configuration of bicycle components. Coyle et al. [28] illustrated the link between better performance and larger population of type I fibers in driving muscles (i.e. vastus lateralis) of professional cyclists. This finding enforces the thought that training is the most important factor in success in cycling. Further studies should examine if larger quadriceps volume could also be part of the determinant effects in cycling performance, as suggested by Coyle et al. [28]. In summary, it is unlikely that “optimal” or “ideal” bicycle configuration could be determined solemnly by the analysis of body dimensions. An additional factor like joint motion could play an important role in this issue.

1.3 Motion analyses Knowledge on the anthropometrics of human bodies is important particularly if they are used to measure human motion. For that purpose, scientists have assessed human

17 motion by modeling the human body and tracking motion of its individual components (i.e. body segments). Most often, human motion is analyzed using cameras that enable the acquisition of image frames at higher frequencies than normal cameras (i.e. greater than 25-30 frames per second). Filming human motion enables the user to assess a given motion qualitatively and objectively [29]. For researchers and clinicians assessing cycling motion, the use of high-speed cameras allows the measurement of joint angles, which is the most common outcome [30-32]. That is because joint angles will dictate muscle lengths (and affect muscle forces) and because cycling could be assessed as a continuous motion. Sagital plane (side view) are the most common approach because less motion is observed in frontal or transverse plane during sub maximal steady state cycling [33]. Body segments are converted in rigid segments with hinge type joints that link segments to the pedal. In Figure 2.1-A, a bidimensional (sagital only) model of the right lower limb is illustrated along with a definition for joint angles measured during stationary cycling motion (Figure 2.1-B).

Fig 2.1. (A) Spatial model of the sagital plane used to assess lower limb cycling motion. (B) Joint and segment angles commonly used to describe kinematics during cycling motion [34,35].

More recently, inertial sensors (i.e. gyros and accelerometers) have been used to reconstruct motion of human segments in order to reduce time spent with preparation of subjects and to enable analysis out of the laboratory [36]. Recent research [37] presented custom made motion analysis systems that eliminates the use of video cameras and reduces the costs and preparation time for evaluation. Unfortunately, at this time, only one commercial system focused on cycling motion analysis (i.e. Retul® - www.retul.com) and uses similar technology but it is restricted to indoors use. Moreover, some studies use

18 global positioning system (GPS) units to track the motion of each cyclist, in the case of a long duration training or racing [38]. The GPS provides information on traveling speed and gradient, which could be used to infer on power production (see STRAVA www.strava.com). Regardless of the measurement technique, motion analysis should provide information on position, velocity and acceleration of each given reference point (e.g. body segment) through time. This information assists on the analysis of human motion because fast motion of segments cannot be totally assessed by the human eye. Indeed, the objective comparison of motion taken at varying conditions requires reliable assessment techniques, which enforces the use measurement techniques from motion analysis systems. For practical application, motion analysis helps to assess body position on the bicycle and fatigue state along with effects from changes in cadence-power combination.

1.4 Force and pressure analysis Motion analysis in general does not take into account the sources for movement, which are only added to the assessment when forces involved in motion are monitored. In cycling, the attachment between the cyclist and the bicycle is defined in a three-point link, involving the handlebars, the saddle and pedals. These links are forces transfer points from the body to the bicycle and vice-versa. For that purpose, force sensors have been attached to bicycle parts like handlebars and seat tube [39] and pedals [40] to capture the reaction forces from these components to body segments. Although during steady state road cycling traveling speed is fairly constant, the forces involved in pedaling action largely fluctuate throughout the crank cycle [41]. The main reason is that muscle force production depends on muscle length, and consequently on joint angles [42]. Given that joint angles change during pedaling action and that levers for each muscle depend on the position of the crank relative to the leg segments, force applied to the pedals fluctuate from this coupling. This limited connection between cranks and legs lead to a limited percentage of the force applied to the pedals that drive the cranks in their movement direction. Therefore, power production is limited by the direction of the force applied to the pedals in relation to the cranks. Research showed that the propulsive forces could be 40-60% of the total force [43-45] and that varying factors could change this range (i.e. body position, workload, cadence, fatigue and others). The majority of studies to date were limited to laboratorial assessments of forces involved in cycling motion due to the electronics involved in recording force signals into

19 computers [46]. In order to measure power output, commercial power meters have been largely used in training and racing [47]. Monitoring training and exercise intensity using power production has been shown to be more reliable than assessing heart rate because this measure is influenced by environmental factors (i.e. humidity and temperature) [48]. More recently, force sensors have been implemented to provide force and power output in order to improve the information related to exercise effort (see MEP® - www.aip-mep.com). This should be added to other commercial systems that will provide measurements during outdoors cycling exercises (see Garmin-Vector® - http://sites.garmin.com/vector/). Bilateral pedal force measurements are illustrated in Figure 3.1, with detail on the total force, effective force (tangential to crank), normal and anterior-posterior components. Fig 3.1. Ensemble forces applied to the pedal taken along a full crank cycle. Illustration of effective (tangential to the crank), total (resultant from normal and anterior-posterior), normal (to pedal surface) and anteriorposterior (to pedal surface). Positive values indicate effective force in favor of crank motion, normal pulling force and anterior force on the pedal surface.

Force measurements could be combined to the kinematics of segments in order to infer on joint contact forces and moments. These measurements provide insights on the individual joint actions and coordination of muscle groups [49].

1.5 Muscle activation Skeletal muscles transfer force to bones once the central nervous system sends action potential to muscle fibers. The combined action potential from various motor units can be assessed using surface or indwelling electromyography measurements [50]. For

20 surface electromyography pairs of sensors are attached to the skin to gather the electrical signals from muscle action-potentials which provides an overall profile of muscle recruitment during various tasks. In cycling, most studies [51-55] opted for surface electrodes because indwelling sensors may result in pain and unnatural cycling motion. A benefit from indwelling sensors is that small muscles can be more properly assessed, which has been largely explored by Chapman and co-workers [31,56-58]. Better insight has been gained by these research on differences in muscle activation between cyclists of different levels of experience. Research has suggested that muscle activation could also be affected by various factors, like workload level, pedaling cadence, body position on the bicycle, fatigue state and others. Better efficiency in driving muscles, which is gained through training, has been suggested when cyclists showed lower activations for a given workload compared to untrained subjects [59]. However, gains from training seem to be detected only for muscles that are strongly linked to power production, which is not the case of calf muscles. These muscles have been suggested to act as force transfer units [60], potentially because they tend to be less sensitive to changes in workload and to differences between cyclists and non-cyclists [61]. In Figure 4.1, activation from six lower limb muscles is illustrated as average and standard-deviation from signals acquired during ten crank cycles [44].

21 Fig 4.1. Ensemble muscular activation (mean + standard deviation of ten crank cycles) of vastus lateralis, gluteus maximus, rectus femoris, gastrocnemius medialis, the long head of biceps femoris and tibialis anterior of a competitive cyclist [44]. Dashed line indicates a threshold for onsetoffset of muscle activation (10% of maximal activation).

Following common practice in assessment of muscle activation in cycling, activation is presented as a function of crank angle to enable readers to assess parts of crank revolution where each muscle has larger activation. A generic threshold (10% of maximal activation) is showed (dashed line) in order to provide insight on the activation timing (whenever muscles cross the threshold) [62]. With that in mind, it is possible to observe that the order of muscle recruitment follows a sequence from the knee extensors to hip extensors, ankle plantar-flexors and knee flexors/ankle dorsi flexors. This sequence has been consistent among studies with particular differences for ankle muscles between cyclists [55]. It will be taken further on the potential in using electromyography to gather insights on potential links between workload and activation [63], co-activations and muscle coordination [64] in a following chapter.

22 1.6 Inverse dynamics and muscle modeling Mechanical and biological modeling of the human body has been used to calculate forces acting on body segments. For injury prevention in cyclists, inverse dynamics has been used to assess joint contact forces and net moments in order to gain insights into potential mechanism for injuries [30,65,66]. To accomplish the measurements of joint loads (i.e. forces and moments) modeling anthropometrical dimensions of human body segments must be linked to the global position of these segments along with forces acting at the end point of the kinetic chain. In cycling, pedal forces has been measured to compute reaction forces at the foot, which can be combined to segment motion taken from video analysis to calculate joint loads [67]. Along with joint forces during varying conditions in cycling, (i.e. changing bicycle configuration and cycling motion) [30,65] pedaling technique has been assessed looking at the contribution from individual joints to total lower limb mechanical work [68,69]. Likewise, net joint moments have been calculated to gather insights into potential action from muscle groups (i.e. increase in knee flexion moments to sustain increasing workload) [70]. The starting step for inverse dynamics is the assessment of forces acting in individual lower limb segments using free-body diagrams (see Figure 5.1).

Fig 5.1. Illustration of the free-body diagram used to compute net moments by inverse dynamics (adapted from Nigg [71]).

23 For preventing overuse injuries, it is critical to anticipate potential overloading in joints because even physiological loads applied to a joint after degeneration could be detrimental to joint cartilage matrix [72]. With that in mind, peak compressive forces and pressure on the joint have been assessed to track if either motion or workload could lead to potential injury risk for cyclists [15]. For performance, minimum joint mechanical moments has been linked to the preferred and optimal pedaling cadences [73]. This link indicates that moments could be sensitive to changes in muscle forces, although moments are limited to the percentage of cocontraction between antagonists at a given joint. Converting muscle forces into crank power is critical for optimizing cycling performance. For that achievement, cyclists rely on the production and transfer of power across the hip, knee and ankle joints, which can be measured by assessing net moments and angular velocities at these joints. Studies indicated that the knee joint is the main driver for power (~60% of the total power) followed by the hip and the ankle, which mostly transfers power from large muscles (i.e. hip and knee extensors) to the cranks [74]. In Figure 6.1, power produced by the lower limb joints is shown along with total power from the lower limb [68].

Fig 6.1. Ensemble power measures at the ankle, knee and hip joints during cycling at 80% of the maximal aerobic power output. Total power measured at the three lower limb joints is illustrated [68].

Changes in percentage of contribution from each lower limb joint to the total joint power could be achieved by increasing workload level [69]. In contrast, consistent contribution from each joint was found when pedaling cadence was changed ±20% from

24 cyclists preferred cadence [68]. Further changes (e.g. training, fatigue, etc) could be assessed in order to track if cyclists sustain joint contributions.

1.7 Conclusions and practical applications Biomechanical methods have evolved during the last 20 years and they have become available at varying research centers. However, few cycling teams and clinics have access to these methods for assessment of athletes, which varies across different regions in the world. Research in cycling biomechanics have also expanded and now covers more countries, which enhances the application of research in daily cycling training. The recent publication of articles in a specific journal on cycling (i.e. Journal of Science and Cycling - http://www.jsc-journal.com) may boost the access of research to a greater number of cyclists and coaches. Future years may provide increasing knowledge on cycling biomechanics, particularly with reducing pricing of motion analysis systems and increased access to commercial instrumented pedals to more research centers and cycling teams.

25 CHAPTER 2: MEASURING PEDAL FORCES

Rodrigo R. Bini and Felipe P. Carpes

2.1 Introduction Bicycle components have changed over the years in order to minimize resistive forces and energy cost for pedaling with purpose of maximizing cycling performance [1]. Along these lines, the assessment of forces exerted by cyclists is important for the analysis of pedaling technique and to anticipate injury risk factors. Cyclists continuously aim to produce maximal possible power output for longer duration, particularly when power delivered to the cranks can be translated into bicycle speed. To ascertain the optimal transfer of forces applied to the pedals to cranks, the measurement of pedal forces and pedal motion is critical for development of interventions with focus on increasing maximal crank torque. An alternative approach is to define a given speed (or power output) and to seek for alternative ways to minimize peak crank torque and pedal forces in order to maximize the use of pedal force application. During the last two decades, various instruments have been developed to provide force measurements [75,76,40]. However, most of these devices were limited to the laboratorial environment. A recent device has been presented in order to enable measurements during track cycling [77]. In line with that, commercial systems developed instrumented pedals to provide wireless measurements during outdoors pedaling exercise (http://www.polar.com/en/about_polar/news/Polar_and_LOOK_launch_together_power_pe dals). Despite of this, commercial wireless systems may be limited to the assessment of power output rather than actual pedal forces. In this chapter, a description of possibilities for measuring forces and the potential benefit of each component of pedal forces will be presented. Further, pedal and crank kinematics will be introduced to highlight their importance for assessment of pedaling kinetics and technique.

2.2 Pedal force components Muscle force is transferred to the pedals throughout bones and tendons but the direction of force application on the pedals depend on the position of the foot in relation to the pedal surface. For the analysis of force directions on the pedal surface, total (or

26 resultant) pedal force is separated into three orthogonal components (normal – Fy, anterior-posterior – Fx and medio-lateral – Fz), as shown in Figure 1.2. Fig 1.2. Illustration of three dimensional components of the force applied to the pedal surface (Fx – anterior-posterior, Fy – vertical and Fz – mediolateral.

Only normal and anterior-posterior pedal force components can be translated into crank torque depending on the position of the pedal in relation to the crank. This helps to explain why most studies focused only on the measurement of these two force components. Another reason could be associated to the increase in complexity from electronic settings when the medio-lateral force is measured along with rotation moments on the three-dimensional axes of the pedal. However, research has shown that increases in lateral force application on the pedal surface [78] and larger internal rotation moment along the Fy axle [79] could be associated to overload on the knee joint soft tissues [66].

2.3 Principles of force measurements in instrumented pedals The most common approach to measure forces applied to the pedals is by measuring the deformation on the pedal structure resulting from force applied to the material. This principle follows Hook’s law, where a linear deformation is expected when force is applied at a constrained bandwidth. Whenever, force application exceeds the

27 elastic properties of the material, permanent deformation is observed and changes in force to deformation relationship is observed. Strain gauges are the most used sensor in pedal force measurements due to their low cost and easy handling compared to other sensors (e.g. piezoelectric crystals). Strain gauges are attached to areas on the material where larger deformations are expected, commonly simulated using finite elements modeling [80]. In Figure 2.2, an example of strain gauge is shown with illustration of components of the sensor. Once strain gauges are attached to the material, deformation from force application on the material is transferred to the strain gauge, which changes the resistance on the foil structures. For a given current level passing to the foil, changes in structure of the foil will affect voltage output due to changes in electrical resistance. Therefore, net changes in voltage (input to output) will be linked to the force applied to the material. In Figure 3.2, an instrumented pedal is shown with strain gauges attached to the pedal spindle.

Fig 2.2. Expanded view of a commercial strain gauge element for strain measures.

28 Fig 3.2. Strain gauges attached to the pedal spindle [81].

2.4 Historical changes in methods for force measurements Faria e Cavanagh [82] reported that in 1889 R.P. Scott was the first to perform measurements of the force applied to the pedals. Guye [83] also described that R.P. Scott used a mechanical device that, similarly to the one presented by Sharp [84], enabled small blades to draw on paper whenever force was applied to the pedal. Assuming a linear relation between the force applied to the size of drawing on the paper, the magnitude of forces could then be quantified. However, at that moment, only the normal force component (Fy) was measured. This limitation was latter addressed by instrumented pedals introduced by Dal Monte et al. [85], that enabled bilateral measurements of normal and anterior-posterior pedal forces during stationary cycling in the laboratory. In the early 80s, the first three dimensional pedals were presented [40], which provided full assessment of loads produced during cycling. At that time, attention was given to the relationship between magnitude and direction of pedal forces [40] and to the potential risk of development of overuse injuries in cycling [86,78,66]. The knee joint was the focus due to larger lateral pedal force during cycling with more medial position of the knees [78] and larger internal rotation moment along the Fy axle for injured cyclists [79]. Another avenue is the assessment of the percentage of the force applied to the pedals that drives the crank. The index of effectiveness is the traditional analysis of the percentage of use of pedal forces that produce propelling power on the bicycle and it can

29 be measured as the ratio between the impulse of the effective force (i.e. force perpendicular to the crank) and the total pedal force (see equation 1.2).

Equation 1.2. Index of effectiveness (IE), computed by the ratio between the impulse of the effective (EF) to the resultant (total, RF) pedal force (RF) [87].

In Figure 4.2, components from resultant pedal (RF) force in sagital plane are shown, as normal (Fy) and anterior-posterior (Fx) along with the effective force (EF), which states for pedal force applied perpendicular to the crank. Fig 4.2. Illustration of pedal force components in sagital plane (Fy) and anterior-posterior (Fx) to the pedal surface along with the effective force (EF), which stands for the percentage of resultant force (FR) that is acting perpendicular to the crank.

Most studies assessing pedal forces were limited to road cycling. In the late 90s, instrumented pedals were customized for mountain bike riders [88]. That involved the adaptation of cleat system to SPD Shimano, rather than the road Shimano or Look type cleat, which are often used in road cycling and triathlon. Along with limitations on the use of most instrumented pedals to laboratorial environment, limited cleat types were observed. In cycling, cleats vary among brands and exchanging cleats involves changing cleat to shoe position and sometimes changing type of shoes. In Figure 5.2-A, we

30 introduce an instrumented pedal force system developed in our university [89] and used in various studies from our research group [90,91,30,92,45]. This system was designed for Look-Delta cleats. In Figure 5.2-B, a second instrumented pedal force system was designed by our research group to provide measurements of forces using SPD-Shimano cleats, commonly observed in mountain bikers [81]. Fig 5.2. Instrumented pedals designed for measurements of normal (Fy) and anteriorposterior (Fx) components of pedal forces. (A) instrumented pedal for LookDelta cleats [89], commonly used by road cyclists and triathletes. (B) Instrumented pedals for SPD-Shimano cleats [81], commonly used by mountain bikers.

Throughout the years, different prototypes were presented in order to provide an alternative to instrumented pedals. The first prototype with a different design was based on the attachment of a cycle ergometer to a 3D force plate [93]. In that project, forces were determined using inverse dynamics of the bicycle and force plate connection. That enabled cyclists to use their own pedals, which improved instrumented pedals design. A second approach was based on the development of an adaptor between the pedals and the cranks [94]. This system provided normal and anterior-posterior force measurements in the adaptor, which were linked to forces applied to the pedals. Again, the main benefit was the possibility for cyclists to use their own cleats and shoes. A great challenge for research on cycling biomechanics is the ecological validity of measurements taken in laboratorial environment. Indeed, differences in pedal forces have been observed when pedaling on a fixed cycle trainer to cycling on a treadmill [46]. To address this issue, Dorel e fellows [77] introduced a partially wireless system that enabled measurements to be taken during track cycling. An analogical to digital converter was carried by the cyclists on a backpack and signals were then wirelessly transferred to a

31 computer sitting in the centre of the track. With this device, coaches and biomechanics could either save data for off-line analysis or could process data with minimal delay between acquisition and decision making. Improvements in automation and data transferring technology largely impacted progress in force measurement systems. Wireless technologies are developing fast and will be the main avenue to enable more ecological assessments of cyclists and triathletes of different disciplines. On the other hand, high cost still limits a wide spread of this technology to varying levels of cycling (from recreational to professional). At the moment, commercial systems are slowly appearing in the market. Instrumented cranks were adapted to provide crank torque measurements outdoors [95]. However, they have been shown to limit the assessment of force exerted by each leg [96], which limits any analysis of asymmetries in cycling.

2.5 Angular sensors for kinematics assessments Exclusive measurement of pedal force components can be incomplete if pedal kinematics is not provided. To decouple perpendicular to resultant pedal force, information on crank position and pedal angle is required because pedal coordinate system does not follow crank or global coordinate system (Figure 6.2). In Figure 6.2, it is possible to observe that pedal inclination (θ) changes Fy force orientation. Therefore, measurements of pedal inclination are critical to decompose Fy and Fx from the pedal to the crank and/or global coordinate systems. Fig 6.2. Illustration of the pedal to global coordinate system force decomposition using the pedal angle ().

32

Two methods are commonly used to assess pedal inclination. The first involves filming cyclists pedaling using either one or more cameras. That sometimes involves long time for tracking markers and off-line synchronization with pedal force data. A second approach uses angular sensor (e.g. encoders or potentiometers) attached to pedal spindle tracking rotation of pedal axle in relation to the crank motion. Calibration procedures permit converting changes in voltages in changes in angles during cycling motion. This approach is beneficial because signals can be directly synchronize with force data and because processing time is largely reduced.

2.6 Sources of errors in pedal force measurements Static calibration of loads is necessary to ascertain on the voltage to load relation of each pedal force system. For that purpose, known loads are applied to the X and Y axis of the pedal surface or pedal spindle (Figure 7.2). Using force data (in Volts) collected during this calibration procedure enables users to correct and update calibration factors of instrumented pedals using angular inclination of linear regression taken from voltage (independent variable) to load (dependent variable).

Fig 7.2. Illustration of a static load calibration procedure using a 10 kg load attached by a rigid cable to the pedal spindle in X and Y pedal coordinate system, ascertain for no pedal inclination.

33

In order to gather accurate force measurements, three main sources are usually referred. The first, as previously described, results from errors in calibration of forces during static load tests. The second is the possible cross-talk between orthogonal force components when force applied is aligned to a given pedal axis. A third source is the drift in pedal force measured throughout time. Most of these can be minimized or mathematically amended. As shown in Figure 7.2 inset, the control of pedal inclination is critical to contain load application within the Fy axis of the pedal. Any changes in load direction will affect voltage readings and will compromise the calibration procedures. In addition to that, crosstalk is commonly observed in single axle force application. This phenomenon can be observed when, as per the example given in Figure 7.2, load is applied at a single axis but changes in voltage are also found in other components (e.g. Fx). In principle, assuming that line of force direction is controlled, Poisson effect, which involves the changes in form of a volume when load is applied, is the key explanation. Following the concept introduced by Siméon Poisson, whenever a material is compressed in one direction, an expansion is observed in other directions. In the example provided in Figure 7.2, compression resulting from load applied at the Fy axis will expand the pedal towards the Fx axis (in minimal dimension) and will be gathered by sensors related to Fx deformation readings. Cross-talk correction has been used for 2D [94] and 3D force components [80], which enables the compensation on calibration factors for unwanted changes in force readings. A third effect is drifting in force readings, which have been only reported by Stapelfeldt and co-workers [94]. In this study, changes were -0.02 N/min, which were considered minimum. Strain gauges manufactures recommend that warming up gauges by 20-30 minutes with power supply and no load application to the material could be a solution to minimize temperature effects on force readings. Taken together, the control of calibration procedures, cross-talk and drifts could largely improve accuracy in force measurements.

2.7 Conclusions and practical applications Assessment of pedal forces can provide cyclists and coaches outcomes from changes in pedaling technique along with the possibility to monitor force exerted by

34 cyclists in various pedaling actions. Pedal force measurements have been preferably conducted in laboratorial environment due to limitations in carrying data analog systems and in transferring data for storage systems. Linking pedal forces to measurements of cycling motion (via kinematics assessment) is key to determine further variables related to biomechanics of lower limbs during cycling (i.e. joint kinetics). At the moment, a few commercial systems offer measurements of force and/or power produced at each pedal, which enables the use of force measurements during cycling motion at the clinical and training environment.

35 CHAPTER 3: MUSCLE ACTIVITY

Rodrigo R. Bini and Felipe P. Carpes

3.1 Introduction One of the key issues in biomechanics is the possibility to measure muscles forces during contraction. To achieve this, sensors must be surgically implanted in tendons (e.g. buckle transducers, fiber optic and others) enabling tendon force records [97]. However, these methods require invasive procedures which are rarely observed in the literature [2,98]. With that in mind, biomechanists moved to a more indirect reflection of muscle force, which is muscle activation. The assessment of muscle activation as a reflection of muscle force has many limitations, but for some muscles in vary particular tasks, muscle force and surface electromyography (EMG) have shown to be related in animal models [99,100]. The physiological link is that, to increase force production, central nervous system generally relies on increasing muscle drive in order to enhance calcium availability for cross-bridges formation. In Figure 1.3, we briefly illustrate this sequence when motor neuron action potential is directed to muscle fibers, leading to recruitment of motor units and force production.

36 Fig 1.3. (A) Illustration of a motor unit, which involves a motor neuron and the various muscle fibers innerved. (B) Activation recorded using surface electromyography from vastus lateralis synchronized to pedal force records (C) taken from 10 crank cycles.

In Figure 3.1 we can observe that Vastus Lateralis muscle is largely activated before peak pedal force is recorded, which suggests a delay in activation to force production. In the case of cycling, there are a series of reasons for the delay coming from physiological (e.g. time taken from action potential to elicit spread of calcium in sarcolema) to mechanical (e.g. transfer of force from knee to pedals via bone-on-bone contact and tendon elongations). The electromechanical delay is then linked to the time taken from peak activation to be observed in surface electromyography [101,102] and to the mechanical link of lower limbs to the pedals, passing from action of knee and ankle joint muscles [103,74]. Assessment of muscle activation in cycling has been mostly done using surface electromyography. This enables the measurement of varying effects in muscle recruitment, linked workload level [104], pedaling cadence [51,53], body position on the bicycle [105], fatigue state [106,52,107], and others. Combined use of surface EMG in laboratorial performance assessments (e.g. graded cycling tests) can enable the definition of metabolic related intensities [108], which are useful in training prescription.

37 Looking at the wide number of applications of surface EMG in cycling, the goal of this chapter is to illustrate some uses of EMG in assessment of cyclists. Attention will be given to existing evidence from studies assessing varying affecting factors in muscle activation. Brief comments will be drawn on methods and procedures using surface EMG because key articles cover this issue more completely [50,109].

3.2 Methodological aspects and use of EMG in cycling The use of surface EMG for the assessment of muscle activation involves the measurement of electrical energy traveling from muscle fibers to the skin, resulting from motor neuron drive. Along with surface measurements taken on the skin level, indwelling methods have been used in cycling to gather activation of small size and deep muscles. Chapman and colleagues presented a series of studies looking at recruitment of lower leg muscles, whose are not always of access using surface EMG [31,110-112,56-58]. A great advantage in using indwelling EMG measurements is that reduced cross-talk among muscles (e.g. recording activation from nearby muscles) and assessment of activation of deep muscles (e.g. tibialis posterior) are possible. On the other hand, difficulties in introducing needles and fine wires at a deeper muscle level may limit the participation of some subjects. Measurements taken using electrodes attached to the skin of cyclists are illustrated in Figure 2.3.

38 Fig 2.3. Illustration of bipolar surface electromyograph y electrodes used during assessment of muscle activation of cyclists. Inset highlights the electrodes attached to the skin of the cyclist.

In the case of surface EMG, two electrodes are attached on the skin at the most prominent belly region of the muscle in order to capture electrical energy that arrives to muscle fibers. A third method that recently became available is the matrix of electrodes for records of muscle activation at multiple sources on the skin level [113]. This technique enables users to record, among other measurements, the motor unit action potential velocity [54]. Macdonald e fellows [54] observed stable motor unit action potential velocity for vastus lateralis and vastus medialis during fatiguing cycling exercises. However, during graded exercise testing, increased motor unit action potential velocity was observed, suggesting larger recruitment of faster motor units. In many cases, EMG hardware is limited to a given number of channels, which contains the number of muscles to be assessed. Therefore, the election of muscles for analysis is critical in order to properly track the main drivers of a given motion. Another issue is the size and shape of muscles, particularly those of small sizes or multipennated could be limit proper attachment of electrodes. Muscles of small size may be more prone of cross-talk while multipennated muscles may suffer from need of attachment of electrodes in each head in order to gather measurements from each muscle component. Another limiting factor is the size of underlying adipose tissue, which reduces the magnitude of signals that travel from muscles to the skin [50]. This is a key factor in

39 assessments conducted in different days because subjects may change their diet and therefore change contents of adipose tissue. Surface EMG has been used to determine metabolic thresholds during graded exercise testing [114]. Validity has been confirmed for the majority of subjects assessed using surface EMG via comparisons to lactate concentrations [114] and via comparisons to ventilation [108]. An example is shown in Figure 3.3. Fig 3.3. Illustration of EMG-VO2 plot to determine the EMG threshold. Muscle activation is presented by the RMS normalized by the activation recorded during maximal isometric voluntary contractions (CVM). Illustration courtesy of Dr. Fernando Diefenthaeler.

The option for the assessment of surface EMG in determining metabolic thresholds can be useful given the lower cost of EMG systems compared to metabolic carts. On the other hand, the analysis of EMG signals and subject preparation could be as complex as those observed in gas exchanges assessments.

3.3 Muscle activation during cycling Surface EMG has been used to monitor responses from muscle activation due to changes in workload [62,104], pedaling cadence [53,115], body position on the bicycle [116,105], fatigue state [52], and cycling skill [31,53]. At steady state cycling under controlled level of workload and pedaling cadence, muscle activation usually follow the pattern showed in Figure 4.3.

40 Fig 4.3. Ensemble activation from Tibialis Anterior, Biceps Femoris, Adductor Longus, Gastrocnemius Medialis, Rectus Femoris, Vastus Lateralis and Gluteus Maximus from a competitive cyclist. Adapted from Bini et al. [44].

From a qualitative perspective, lower limb muscles are activated and de-activated in a given section of crank cycle. For knee extensors, (e.g. vastus lateralis and rectus femoris), larger activation is observed close to the top dead centre (e.g. 0 of crank cycle). This early recruitment has been associated to early knee extension required to move the crank forward. Hip joint extensors (e.g. gluteus maximus) have latter activation, close to the most forward crank position (e.g. 90 of crank cycle) in order to accelerate the crank and drive large forces to the pedals. Plantar flexors (e.g. gastrocnemius medialis) act in transferring force from the legs to the pedals during the later section of crank cycle. The combined action of plantar flexors and knee flexors help on the backward rotation of the crank at the bottom of the stroke (e.g. 180 of crank cycle). For some cyclists, an active pulling backward is observed in order to change force direction on the crank (from downward to backward). Some muscles act during the upward phase (e.g. recovery phase) of crank cycle, in order to help pulling the pedals upward via flexion of hip and knee and ankle dorsi flexion (e.g. tibialis anterior).

3.3.1 Workload Workload is managed by changes in resistance imposed to the cyclists while pedaling. Workload is computed by the integral of power output over time and it is expressed in mechanical work units (i.e. Joules). In cycling, workload level has been shown to largely affect muscle activation. During self-paced cycling exercise, greater

41 workloads were achieved by tuning muscle activation of vastus lateralis and rectus femoris during a 40-km time trial performed indoors [117]. In addition Hug and co-workers [118] observed that hamstrings also increase activation when workload is greater during graded exercise tests. It has been suggested that individual selection for recruitment of hamstrings may dictate the contribution of this muscle group to power delivered to the cranks by tuning co-activation between knee flexors and knee extensors [53]. Therefore, during sub maximal cycling exercise, increases in workload should enhance activation of hip and knee joint extensors (e.g. vastus lateralis), with more varied response for hamstrings and ankle plantar flexors (e.g. soleus and gastrocnemius medialis).

3.3.2 Pedaling cadence During laboratorial trials, cyclists are commonly assessed using controlled workload and pedaling cadence, when changes in cadence are constantly balanced by changes in cycle ergometer resistance in order to keep constant power output. However, in outdoors cycling, pedaling cadence hardly detaches from workload, given cyclists need to control gear ratio to sustain similar exercise effort when resistive forces are changed. Muscle force-velocity relationship also naturally links force capability to muscle shortening velocity and both depend on pedaling cadence. In line with that, pedaling cadence that minimizes muscle activation increases with greater power output [104]. The reason for that is also partially explained by greater contribution from inertial forces to crank torque at higher pedaling cadence, which can reduce muscle force requirements [119]. Muscle activation timing (onset-offset) is also influenced by pedaling cadence, with individual responses from each muscle [62] and a consistent anticipation on muscle onset for faster pedaling cadences [120] was observed. However, no effects were observed in co-activation of knee joint muscles [53]. Therefore, increases in pedaling cadence for a constant workload can lead to lower muscle activation. For training, there could be less fatigue if cyclists opt for using higher pedaling cadences, which is in line with evidence that cyclists opt for cadences close to 90 rpms in order to minimize joint moments [73]. Further data add to that by showing that professional cyclists minimize knee joint extensors activation pedaling at ~100 rpms rather than at lower cadences [121].

42 3.3.3 Body position on the bicycle Upper and lower body position is critical to dictate muscle activation [51,116]. In this issue, changes in upper body lean and saddle height have been referred as the two most important settings in body position of cyclists during pedaling [116,12,105]. Changes in vertical and horizontal saddle position affect lower limb joint angles [30,32], which will change muscle lengths and their potential to produce force during crank cycle. Given that muscle activation is largely influenced by muscle length [99,100], changes in mono and bi-articular muscle lengths from changes in body position on the bicycle should influence muscle activation. Conflicting results have been reported for various lower limb muscles when changes are conducted in saddle height [51,122,123]. Variance in results between studies can be associated to the assessment of non-athletes with few or no cycling experience along with differences in workload and pedaling cadences. One study showed increased activation for triceps surae muscles when cyclists were pedaling at higher saddle heights [51]. Duration of activation (onset-offsets) was observed to remain stable for biceps femoris, vastus lateralis and gastrocnemius medialis when changes in saddle height were performed, although less eccentric contraction was observed for gastrocnemius and vastus was shown [124]. In this particular study, trained cyclists were tested using saddle heights of 96-100% of the throcanterion height, which did not affect cycling economy. Therefore, along with previous observations, changes of ~4 cm in saddle height affect muscle activation without large impact in cycling economy [12]. For changes in horizontal position of the saddle, Ricard and fellows [125] observed that more forward projection of cyclists on the bicycle resulted in less activation of biceps femoris during Wingate tests. Differently, Bini and others [126] found lower activation for gastrocnemius medialis when cyclists were asked to move forward and backward on their given saddle position. At the most backward position, rectus femoris also presented less activation compared to their preferred position. These results suggest that towards a more forward position on the bicycle, cyclists shift force drivers from hamstrings to quadriceps. Indeed, this is a common practice during training, when cyclists tend to move backward on the saddle during uphill cycling and forward during sprinting, in order to share force production among varying muscles. Changing upper body position has a consistent effect in activation of upper and lower body muscles [105]. Triathletes and time trialists use aerobars to reduce frontal projected area and improve aerodynamics. This practice reduces activation o biceps brachii and upper trapezius compared to the Superman position, when cyclists pedal with their arms fully extended lying on a support longer than aerobars used by triathletes [127].

43 Dorel and fellows [116] observed that pedaling at the aerobars enhances activation from gluteus maximus, vastus medialis and vastus lateralis compared to upright cycling. Taken together, studies suggest that cyclists should take care when changing bicycle setup or even during changes in their position for a given setup in order to properly share force requirements among a greater number of muscles to postpone fatigue effects and sustain cycling performance.

3.3.4 Fatigue Different experimental designs have been used to study fatigue in cycling. Among the most common include, constant load time to exhaustion [52,106], incremental load tests [54] and time trial tests [117,128] were used. Regardless of the experimental model, fatigue appears to play an important role on increases in muscle activation [52,106] and reduction in muscle force capability [54]. Dorel and fellows [52] observed large activation for biceps femoris and gluteus maximus at final stages of a constant load time to exhaustion test. This finding highlights an increase in firing rate and motor units recruitment due to changes in fatigue state. In contrary, during sprint cycling, lower activation was observed for biceps femoris, which have been linked to a reduction in coactivation at the knee joint for sustaining knee joint extensor moments [129]. Two studies looked at cycling time trials, when cyclists can manage gear ratio and pedaling cadence to choose their preferred pacing to complete as fast as possible a predefined distance. Duc et al. [128] showed unchanged activation during a 30-min time trial, which differs from larger activation for vastus lateralis found during a 40-km time trial [117]. Studies showed that during time trials, cyclists subconsciously manage muscle activation in order to postpone exhaustion [130]. In summary, changes in fatigue throughout exercise seem to largely increase activation from hip and knee joint muscles, which are responsible for driving power to the cranks. This increase in activation could be linked to higher firing rates and greater number of motor units being recruited. Therefore, subconscious control from central nervous system acts during cycling time trials and all exercise under self-selected intensity in order to postpone exhaustion.

3.3.5 Cycling skill The ability of cyclists has been commonly assessed by measuring force direction during crank cycle [131]. A larger percentage of the force applied to the pedal that drives

44 the crank forward has been related to greater skill in cycling [132], which has been shown to distinguish cyclists to non-cyclists [133]. Trained cyclists present greater ability to activate biceps femoris during pedaling action compared to non-cyclists [134,59], which could be explained by the role of this muscle to change crank forces from downward to backward at the bottom of the stroke. Variability in activation has been linked to consistence in motion and experienced cyclists presented lower variability in activation of lower leg muscles compared to novice cyclists [56,58]. Time frame for gains in consistence in muscle activation is still to be determined. Muscle

activation

of

triathletes

has

been

the

focus

of

some

studies

[126,106,53,112]. The reason for a particular interest in triathletes is that triathletes share training volume in swimming, cycling and running [135], which could lead to different adaptations in their musculoskeletal system. As an example, running and cycling have shown to change muscle-length-force properties [136,137], which could at some point impact triathlon performance. Differences between triathletes and cyclists have been shown for pedal force effectiveness (reduced for triathletes [138]), larger variability in muscle activation for triathletes [57] and greater co-activation in triathletes compared to cyclists [53]. Therefore, for similar cycling task, triathletes present reduced performance compared to cyclists [139] potentially because neuromuscular system of triathletes adapt different to training then cyclists. Further research is warranted to shed light on this hypothesis.

3.4 Conclusions and practical applications Various studies assessed muscle activation during cycling task of cyclists, triathletes and non-athletes pedaling at varying workload levels, pedaling cadences, body positions on the bicycle and fatigue state. These investigations provided a large body of knowledge on how many factors (e.g. workload, cadence, etc.) could impact muscle recruitment in cycling. The majority of studies focused on activation from lower limb muscles, given these are the main drivers for properlying the cranks. On the other hand, upper body muscles have been assessed in order to track fatigue effects on posture related muscles. Overall, increases in workload lead to higher activation of the majority of muscles, which is more varied when pedaling cadence is changed. Fatigue leads to increased recruitment of hip and knee joint muscles in order to add motor units at higher firing rates to sustain performance. Greater forward lean of the upper body increases recruitment of hip and knee joint extensors. Moving the saddle in vertical or horizontal

45 directions have led to conflicting results due to differences in magnitude of changes and experience of subjects (i.e. non-cyclists). Cycling skill has been shown to enhance the use of biceps femoris in cyclists compared to non-cyclists, along with the reduction in variance in activation and co-activations.

46 CHAPTER 4: JOINT KINEMATICS

Felipe P. Carpes, Rodrigo R. Bini and Jose I. P. Quesada

4.1 Introduction Motion analysis involves detecting the position of joints and segments in a global coordinate system, which enables the assessment of translations and rotations. Exclusive analysis of motion does not take into account forces acting on the body and interactions to varying systems (e.g. bicycle components). In biomechanics, the most common approach for motion analysis is by filming subjects performing a given motion and tracking segments and joints throughout various frames [140]. For that purpose, reference markers are attached to the skin at anatomical sites related to joint coordinate systems [33]. Tracking these markers throughout motion is important to assess changes in segment and joint motion during a given task. Video cameras can vary in terms of the number of frames taken for a given time (30-1000 frames per second - fps) and resolution (640x480 – full HD), which result in differences in precision in detecting segments and joint coordinates, also affecting resolution for tracking markers during high speed movements. When two or more cameras are used simultaneously, three-dimensional perspective of motion can be gathered as long as a minimum of two cameras track a given coordinate, by using mathematical algorithms [33]. In cycling, studies were mostly limited to the evaluation of sagital plane motion of lower limb joints [141,35,142] due to the large ranges of motion for hip (42-44°), knee (7378°) and ankle joints (21-25°) [35] compared to motion at the frontal and transverse planes [33]. Another reason for tracking motion exclusively at the sagital plane is the reduced complexity compared to three-dimensional data acquisition and processing and to the good agreement between 2D and 3D sagital plane angles in cycling [33]. Angles are the common output given the constrained offered by bicycle components and the continuum action in pedaling, with a common assumption for bilateral symmetry in joint motion [141,35,142]. Methods related to data acquisition and processing for tracking markers in video frames are well described by Page and others [143], and will not be covered in this chapter. However, care should be taken for acquisition of motion of cyclists potentially due

47 to effects from errors in camera position in relation to the cyclists for 2D assessments (see Figure 1.4). Fig 1.4. Position of a single camera for 2D motion analysis of cyclists in the sagital plane.

For three-dimensional assessments, two or more cameras are needed securing that each marker is visible for a minimum of two cameras in order to determine X-Y-Z coordinates using DLT transformation [144]. Technical documents are provided in a very comprehensive web-page [145] on ways to conduct motion analysis. An example is shown in Figure 2.4 on the three-dimensional assessment of cyclists.

48 Fig 2.4. Position of cameras for 3D motion analysis of cyclists.

Technology evolved and cameras used for motion analysis moved from Video Home Systems (VHS) (Figure 3.4-A) to Ethernet based systems where cameras employing infrared increase processing speed by detecting 2D markers coordinates rather than acquiring full video (Figure 3.4-B). Fig 3.4. Video Home System (VHS) cameras used in older motion analysis (A) and existing Ethernet based cameras with infrared technology to track 2D coordinate markers (B).

In cycling, static poses are used to assess body position on the bicycle [146,141,147] in order to infer on pedaling motion. Further, lower limb action is tracked during cycling at varying workloads [148,35], pedaling cadences [68,115], body positions

49 on the bicycle [126,149,16] and fatigue states [150,151]. The option for varying combinations of workloads and pedaling cadences elicited changes in joint angles [35,68], with effects on muscle lengths and activations [54,62]. To accomplish this assessment, lower limb motion is commonly assessed as average patterns from a given number of crank cycles determined from top to top dead centers (see Figure 4.4).

Fig 4.4. Crank cycle and quarters used to assess motion during pedaling.

4.2 Crank kinematics Crank motion is commonly contained by the angular path of the crank eliciting a combination of linear and angular motion to the lower limbs. Time (i.e. cadence) at which cyclists cover a full crank cycle varies depending on fatigue state, workload level and other factors. In general, road cyclists pedaling at cruising speed opt for a cadence of 90-100 rpm for varying reasons (e.g. minimize muscle activations, reduce pedal force per crank cycle, etc) [73,104], which lead to a time duration of ~0.6 s per crank cycle. In order to gather an ensemble kinematic pattern for joints and segments, it is assumed that cyclists cover each crank cycle at constant angular velocity, which has not been confirmed [152,153]. Indeed, within cycle variation have ranged from 75-115 rpm for cyclists pedaling at an average of 90 rpm [153]. This assumption has been followed in many studies using single sensors to detect crank position [117] or motion analysis systems [70]. Although

50 changes within cycle were observed for pedaling kinetics [153], it is uncertain of this effect on comparisons between cyclists who present different crank angular velocities. Fig 5.4. Crank angular velocity during a full crank revolution highlighting variation of angular velocity throughout the crank cycle [154].

4.3 Upper body motion during cycling Cycling involves mostly motion of the lower limbs. However, recruitment of upper body muscles is used to sustain trunk position. Indeed, excessive recruitment has been linked to low back pain particularly during long duration riding (>30 min) [155]. Methods to measure spine motion vary in terms of the partition of the spine in sections [156,157] or representing the spine as a single rigid segment [146]. Two-dimensional assessment of trunk and low back angles have been suggested in order to track for between-cyclists differences in low back angles for a given upper body lean [158] (see Figure 6.4).

51 Fig 6.4. Illustration of the trunk angle commonly assessed in cyclists [146,159] (A) and a method using the last rib a an anatomical site for partition of spine in upper and lower sections (B) [158].

Motion of the pelvis has been related to saddle pressure and differences between genders were observed with greater anterior pelvic tilt for female cyclists [160]. In this regard, the pelvis have briefly shown to tilt on the frontal plane in a case study, which has been linked to greater risk for low back pain in cyclists [161]. Similarly, cyclists with existing low back pain showed a greater lumbar flexion than age-mached subjects [156]. This finding enforces that position of the pelvis is critical to lumbar angle during pedaling and to low back pain related to bicycle riding [162]. Upper body position has been related to changes in activation of lower limb muscles [163], which were later shown to affect performance in cycling [164]. Greater trunk flexion was found in triathletes [146] due to the need for reducing frontal projected area [159], with detrimental effects in pedal pulling forces [116]. Therefore, given triathletes seek for minimum drag force during the cycling leg of the race, reducing trunk angle is critical to achieve optimal performance [165]. Changes in position of the hands on the handlebars were related to pelvic tilt and trunk angles. In road cyclists, changing from the top flat section of the handlebars to the drops led to greater trunk (141 ±0.87 vs. 156 ±0.73) and pelvic angles (16 ±1.38 vs. 29 ±1,59) [166]. The greater the forward lean, the larger effects were observed in lumbar pelvis, with more flexion when cyclists opted for pedaling at the aerobars [167]. Greater tilt for the anterior section of the saddle was also found to increase lumbar extension,

52 therefore, replicating more closely lumbar angles expected during upright standing [162]. The increase in tilt for the anterior section of the saddle was indeed effective in reducing low back pain potentially because of better force direction from the lumbar vertebras to the sacrum. However, cyclists showed an anterior pelvic tilt [168], which may be linked to a less pronounced tilt for the anterior section of the saddle. Further research should be conducted looking at potential effects in performance after changing saddle tilt.

4.4 Lower body motion during cycling Lower body is the main focus in motion analysis of cyclists due to their large contribution to crank power production via action of lower limb muscles. Cyclic motion performed in stationary bicycles enables researchers to track joint and segments kinematics in order to calculate joint ranges of motion. Studies looking at joint angles were mostly contained to sagital plane motion [34,141,151]. Although differences in bilateral kinematics were found [169], unilateral assessments were conducted in uninjured cyclists potentially to reduce the volume of data. The option for sagital plane analysis of motion of cyclists was also supported by small differences in hip, knee and ankle angles comparing two to three dimensional assessments [33]. This study then lends support to the assumption of a parasagital motion used to track joint angles [170,171], following the layout shown in Figure 7.4.

53 Fig 7.4. Lower limb joint and segment angles used to assess cycling motion [146].

Overall, ranges of motion in cycling were found to vary approximately at 45º for the hip, 75º for the knee and 20º for the ankle [172]. However, differences between cyclists with varying experience [31] and between road and mountain bike cyclists [142] were observed. Changes in saddle height largely affected knee and ankle joints with greater knee flexion and larger plantar flexion found for higher heights [12]. Workload level was also found to affect ankle [35] and knee joint angles [148], which were associated to changes in muscle actions. Increases in pedaling cadence led to smaller ranges of motion for the ankle joint, which were associated to an attempt to sustain ankle stiffness and to improve ankle plantar flexors action as a force transfer link to the cranks [68,115]. Hip, knee and ankle motions along with pedal angles taken at the sagital plane are illustrated in Figure 8.4.

54 Fig 8.4. Hip, knee and ankle motions along with pedal angles taken at the sagital plane [35].

Fatigue was also observed to increase ankle range of motion [150,173] (see Figure 9.4), with a potential requirement for sustaining ankle muscles stiffness. Further, variability in tibial rotation during a 60-min cycling time trial was shown to increase [173], which suggests a compromise in controlling moments of rotation at the tibiofemoral joint when fatigue develops. For the upper body, evidence shows that cyclists increase forward lean when fatigue is increased during pedaling [174,175] potentially as an attempt to increase the contribution from hip joint extensors (via longer muscle-tendon unit length) to crank power [116]. Further research is needed to provide evidence on how much cyclists could benefit from training with kinematics feedback in order to postpone changes in joint motion. Research could provide information on mechanisms for performance optimization via delays in fatigue by changing (or not) joint motions.

55 Fig 9.4. Changes in ankle angle during a 60 min cycling. Adapted from Wiest et al. [176].

Differences between road and mountain bike cyclists were evident for ankle motion during similar exercise intensity and pedaling cadence (see Figure 10.4) [177]. Studies are needed to add to this comparison of motion analysis between cyclists of different disciplines. One potential reason for differences between disciplines is due to differences in bicycle configuration, which may in turn affect joint angles [146], but differences in the pedaling technique may also contribute to the specific kinematics between road and mountain cyclists. Further research is need to provide normative ranges of motion and joint angles, which can then be applied for adjusting bicycle components for cyclists of different disciplines.

56 Fig 10.4. Comparison of ankle angle between road and mountain bike cyclists pedaling at the workload of their ventilatory threshold. Adapted from Carpes at al. [178].

Lately, concern on excessive foot and tibia rotation to the pedal surface has been related to a greater risk for knee joint overuse injuries [79]. Moreover, injured cyclists were found to present a larger medial projection of their knees in relation to the pedal axle in the frontal plane [179]. Assuming a piston like motion for the thigh and shank, medio-lateral motion should not be intended given larger medial motion were found to increase varus loading on the knees [78], which may also impair the production of effective force during pedaling. With that in mind, Carpes et al. [180] observed ~4 cm medial projection for the shank in the frontal plane for uninjured trained cyclists. Moreover, this study showed that dominant limb sustained more consistent motion with larger medio-lateral variations for the non-dominant limb. At the moment, only one study presented a device to track foot rotation during pedaling in order to assess potential risk factors in cyclists [181]. Indeed, it is important to highlight that there is need for determining a range for foot rotation that could minimize knee joint loading. That is important because excessive or too little motion were linked to anterior knee pain [79]. Bicycle components may have an effect in joint motion. Shan [182] observed larger range of motion for the ankle joint when non-cyclists pedaled using a device that reduced crank length at the power phase of crank cycle (from 12 to 3 o’clock). Smaller rotation was

57 observed for the tibia when trained cyclists pedaled in the laboratory using the Rotor IV compared to circular chainrings [183] (Figure 11.4). Fig 11.4. Tibia rotation assessed from cyclists pedaling using circular (Conv) and non-circular (Rotor) chainrings. Adapted from Carpes et al. [184].

The reduced external rotation for the tibia could be beneficial given knee loading could be diminished using the Rotor IV. However, further research is needed to provide pedal and knee load measurements of cyclists using different chanring systems.

4.5 Conclusions and practical applications Motion analysis of cyclists can provide very important information on pedaling skill and technique. Changes in workload, pedaling cadence, body position on the bicycle and fatigue state have a large impact in lower limb joint angles. Few studies assessed 3D motion analysis and asymmetries in pedaling, which suggests that more research is needed in these areas. However, motion analysis should be assessed in combination with muscle activation and pedal forces in order to gain a better picture of coordinative patterns of cyclists. Comparison of motion from cyclists of different disciplines (i.e. road, mountain bike, triathletes, etc) should be conducted in order to provide normative data for bicycle configuration. Outdoors assessment will provide further information during training and/or racing in a more ecological scenario.

58 CHAPTER 5: KINETICS AND PEDALLING TECHNIQUE

Rodrigo R. Bini and Mateus Rossato

5.1 Introduction Improving the interaction between cyclists and their bicycles is a key issue to enhance performance. The reason for that is linked to the optimal use of force applied from cyclists at the pedals, handlebars and saddle in order to improve bicycle speed at the minimum possible energy cost. With that in mind, theoretical research was conducted in order to understand how cyclists apply forces during pedaling [185]. This work was later complemented by experimental data assessed in cyclists during stationary pedaling [40,34,13,52]. Simulation modeling was also employed to gain more general information of potential effects from changes in workload, pedaling cadence and bicycle configuration on joint kinetics [186]. Therefore, computer simulation and experimental research have been conducted to provide evidence on how joint kinetics responds to changes in pedaling action. In some sports, technique is closely linked to performance [187,188], which enforces athletes to perform optimally in order to increase chances for success. Somewhat differently, cycling has been shown to at least in part detach from this perspective [133,189]. With the aim of assessing more closely potential explanations for this particular relation between technique and performance in cycling, this chapter we will give attention to joint kinetics and its potential link to pedaling technique. Discussion on limitation from exclusive analysis of pedal forces as an analog to pedaling technique will be provided.

5.2 Pedaling kinetics To gather a full picture of the biomechanics of cycling, it is important to determine all forces action on the cyclist-bicycle during pedaling. Given cyclists produce force in their muscles and transfer that through the skeletal system to the pedals, pedal forces are critical. They will respond to external forces like drag, weight and rolling, along with forces applied to the saddle and handlebars (see Figure 1.5).

59 Fig 1.5. Illustration of forces acting on the bicycle during outdoor cycling.

During seated cycling, body weight is shared between the saddle, pedals and handlebars, with approximately 60% of cyclist’s body mass directed to the back wheel [190]. Therefore, body weight adds to the gravity and inertial resistance along with the weight of bicycle components. During flat riding at speed greater than 20 km/h (~5.5 m/s), approximately 90% of the drag force is used to overcome air resistance [190]. Whenever level of terrain is changed, body-bicycle weight adds to the overall resistance (or addition) of forces need to move the bicycle forward [191]. Changes throughout the years in design of bicycle components largely reduced contribution from rolling resistance by lowering the cost of transportation [1]. Existing bicycles should then require minimum possible energy cost to the drive transmission translating the majority of forces applied to the pedals into forward driving motion. Looking at forces applied to the saddle and to the handlebars, Stone e Hull [39] showed a coherence between pedal forces and other forces reacting to cyclists’ body. Therefore, even during level cycling, cyclists use upper body muscles to help in power production at the cranks. That power is increased during larger workload level or greater incline during pedaling [39].

60 5.3 Pedal forces Force applied to the pedal is a combination of normal, anterior-posterior and mediolateral components. Given competitive cyclists of various disciplines opt for a cleat to fix the position of the cycling shoes to the pedals, the contribution of the medio-lateral pedal force component to crank properling is minimum. That helps to illustrate the normal and anterior-posterior components along with total force on the sagital plane (Figure 2.5). Along with these components, the percentage of the total force that is perpendicular to the crank (effective force) is also illustrated in Figure 2.5. Fig 2.5. Normal (Fy) and anterior-posterior force (Fx) components of the total (resultant) force (RF) applied to the pedal. Effective (perpendicular - EF) component of the total force applied to the pedal.

Pedal force profile of one cyclist pedaling at 350 W of power output and 90 rpm of pedaling cadence is shown in Figure 3.5.

61 Fig 3.5. Illustration of effective (A), resultant (B), normal (C), anteriorposterior (D) forces applied to the pedals for one cyclists pedaling at 350 W of power output and 90 rpm of pedaling cadence [192].

The percentage of the total force translated into effective force is critical to the magnitude of force that drives the bicycle. If cyclists are unable to direct force perpendicular to the crank, this portion is directed towards the axis of the crank, therefore, reducing power output for a given pedaling cadence and force applied to the pedal. A parallel scenario is when cyclists apply force perpendicular to the crank but at an opposite direction of crank motion. That is commonly observed during the recovery phase of crank cycle (from 180-360° of crank angle). The reason for this second waste in force is due to the inertial effect of lower limbs that create countermovement at the crank, which is, in some cases, opposed by the active pulling action from lower limb flexors [134,133]. In Figure 4.5, an illustration is shown on the theoretically optimal direction of normal and anterior-posterior force components to direct the largest percentage for total force towards crank motion assuming realistic pedal inclination angles.

62 Fig 4.5. Illustration of theoretical optimal direction of normal and anteriorposterior force components to direct the largest percentage for total force towards crank motion assuming realistic pedal inclination angles.

Substantial differences can be found comparing pedal force application for our representative cyclist (Figure 3.5) and the theoretical optimal force direction (Figure 4.5). One of the major differences is found when the crank goes throughout the bottom of the stroke (180°) towards the recovery phase of crank cycle. In this part of crank cycle, cyclists rarely follow the theoretical optimal force direction for some reasons. The most accepted hypothesis is that to active pull the pedals backward and upward at this section a large recruitment is required for knee and hip flexors, which increases pedaling energy cost [189]. Another hypothesis is that multiple changes in force direction could increase voluntary cortical activity thereby challenging cyclists to perform different force directions for both pedals at high pedaling cadences, a common situation when the effectiveness of pedal force is reduced [138,193]. Differences in magnitude and direction of pedal forces can be related then to the ability and skill level of cyclists. Therefore, mechanically more effective pedaling action should be related to larger effectiveness of pedal forces (i.e. large ratio between effective and total pedal force). The analysis of pedaling technique has been then mostly contained by the assessment of force direction during cycling.

63 5.4 Pedaling technique With the goal of summarizing pedaling technique and compare different cyclists at varying conditions (workloads, cadences, etc), researchers presented three main mathematical approaches to compute pedal force effectiveness. The first was based on the computation of the index of effectiveness (IE) based on the ratio between the impulse of the effective force and the impulse of the total pedal force (equation 1.5).



Equation 1.5. Index of effectiveness (IE) computed from the ratio between the impulse of the effective force (EF) and the impulse of the total pedal force (TF) during a full crank cycle [87].

This index allows the calculation of a single value for the whole crank cycle as a measure of an overall effectiveness. Variations of this index have been used by computing the effectiveness at particular sections of the crank cycles (i.e. propulsion and recovery phases) [45]. This last approach provides the analysis of cyclists’ effectiveness at particular sections of crank cycle in order to assess if a given cyclist is optimally direction forces and if there is need to provide feedback at specific parts of the crank motion. In parallel to that, the computation of an instantaneous index of effectiveness [194] is illustrated in Figure 5.5.

64 Fig 5.5. Illustration of the instantaneous index of effectiveness taken from eight trained cyclists pedaling at 80% of their maximal aerobic power output for three pedaling cadences (±20% from their preferred cadence) (unpublished data from Rossato et al. [45]).

Larger effectiveness of pedal forces can be observed during the propulsive phase of crank cycle (0-180° of crank angle) followed by negative effectiveness at the recovery phase (180-360° of crank angle). This pattern indicates that cyclists drive the crank during the propulsive phase of crank cycle but apply force during the recovery phase in opposite direction to the crank motion. Although the index of effectiveness is similarly in opposition at the propulsive and at the recovery phases, the magnitude of forces related to the drive of cranks at the propulsive phase is larger leading to an effective motion of the crank towards bicycle propulsion. However, it is still under debate if cyclists should reduce the negative effectiveness observed at the recovery phase, in order to improve the average effectiveness during the whole crank cycle. The link between pedal force effectiveness and pedaling technique has been shown less obvious [195,70]. Cyclists presented better effectiveness than non-cyclists [133], but differences were not observed comparing competitive to recreational cyclists [194]. Changes in pedaling cadence (±20% from the preferred cadence) did not result in differences in pedal force effectiveness [45]. Likewise, increased fatigue [196] and graded workloads [70] did not largely affect pedal force effectiveness in trained cyclists. One potential explanation for this disconnection between pedaling technique and force effectiveness (measured by the index of effectiveness) is that a single measure was taken during the whole crank cycle. To provide a different view to technique assessment, the index of effectiveness was computed in separate sections of crank cycle (propulsive and recovery) [52]. These authors observed that cyclists presented differences in effectiveness of pedal forces at the propulsive phase of crank cycle when pedaling at high

65 pedaling cadences. A second approach to technique assessment is the analysis of joint moments, which showed that cyclists change peak joint moments at higher workload levels (greater hip and knee flexions) without changes in pedal force effectiveness [70]. These authors suggest that cyclists change pedaling technique at different testing conditions but that does not necessarily translate into changes in pedal force effectiveness. It is likely that changes in joint motion and moments could be more important than changes in pedal force effectiveness with focus on sustaining cycling performance. Muscle activation could also be used to gain insight into muscle coordination during cycling motion [53,56,117]. In this case, cyclists and triathletes have been shown to differ in terms of muscle activation [53,56]. Candotti et al. [138] observed greater pedal force effectiveness for cyclists compared to triathletes at 60 and 75 rpm of pedaling cadences, which was not sustained when these athletes were pedaling at 90 rpm in this study and others [126,149]. Given 90 rpm is usually similar to cyclists self selected pedaling cadence, differences in technique between cyclists and triathletes could be contained to uphill cycling when cadence is reduced [191]. Apart from that, greater co-activation for knee [138] and ankle muscles [57] were shown for triathletes compared to road cyclists, which enforces that these athletes have different coordination during pedaling. These differences have latter been shown to lead to larger mechanical work for the ankle joint and smaller hip angle for the hip joint for triathletes compared to cyclists [126,149], which may add to previously determined differences in performance in favor of road cyclists [139]. Motion analysis adds to the assessment of pedaling technique by showing that cyclists with less experience present larger variability in ankle motion during pedaling than more experienced cyclists [112]. Likewise, preliminary data suggests that road and mountain bike cyclists present different ankle motion when pedaling at similar workload [142]. These differences in motion pattern may provide insights into differences in coordinative pattern gathered from training experience.

5.5 Factors affecting pedal force effectives and pedaling technique Factors affecting pedal force effectiveness and pedaling technique in cycling could be summarized for workload level [197,198], pedaling cadence [138,199], body position on the bicycle [116,200], fatigue state [196,52] and cycling skill/experience [138,194].

66 5.5.1 Workload level Changes in pedal forces during varying workload levels are illustrated in Figure 6.5.

Fig 6.5. Illustration of normal, anteriorposterior, effective and resultant pedal forces applied by competitive cyclists pedaling at three workload levels at 90 rpm (unpublished data from Bini at al. [201]).

When workload is increased, larger peak normal and anterior-posterior pedal forces are found associated to a change in anterior-posterior force direction (greater pedal pulling backward). These results were expanded to joint moments and motion by showing substantial changes due to greater workload levels [70]. Therefore, larger hip and knee flexion moments observed at higher workload levels should be associated to changes in pedal force application to gather a full picture of changes in coordination during cycling. Evidence suggests that towards larger workload, cyclists may increase force contribution from hip and knee flexors in order to enhance pedal force application and to share higher power requirements among a larger number of muscles [70]. Reduced hip joint range of motion is balanced by larger ankle joint range of motion, which may affect muscle force-length relationships [51,115]. These changes do not minimize the higher contribution from hip and knee joint extensors to power production [69,202,203,117].

67 5.5.2 Pedaling cadence Effects from changes in pedaling cadence on pedal force are unclear. Studies showed that cyclists may minimize pedal force application when pedaling close to 90 rpm [138,120,199,204,87]. However, others showed an inverse relationship between pedaling cadence and pedal force effectiveness [138,204,87]. Likewise, pedaling at higher cadences have been shown to reduce muscle activation [104], joint moments [73,59] and minimize co-activations [119,53]. Although effects in pedal forces were observed in other studies [138,120,199,204,87] trained cyclists did not change pedal force effectiveness during full crank revolutions when pedaling 20% faster or slower than their preferred pedaling cadence [45]. Competitive cyclists opt for high pedaling cadences (~100 rpm), which has been shown to minimize activation from knee joint extensors (e.g. Vastus Lateralis) [205] and reduce lower limb joint moments [73]. An inverse relationship was also found between pedaling cadence and ankle joint range of motion [68,115], suggesting an adaptation of muscle force-length-velocity relationship. In summary, increased pedaling cadence lead to reduced pedal force application per pedal stroke along with reduced force effectiveness, less muscle activation and reduced ankle range of motion in order to sustain a given power output.

5.5.3 Body position on the bicycle Body position on the bicycle is affected by changes in configuration of bicycle components or by changes in posture of the cyclist for a given bike setup. These changes will lead to different joint motion [51,32], which influence muscle activation and muscletendon unit length [51] along with different energy cost [32,206]. Only three published studies assessed pedal force at varying saddle heights. Ericson and Nisell [204] evaluated non-athletes while Bini and others [149] assessed trained cyclists and triathletes with no changes in pedal force effectiveness from the manipulation of saddle height. Differently, Diefenthaeler and co-workers [200] observed that 1 cm upwards and downwards shift in saddle height affected pedal force effectiveness by ~5%. However, only three cyclists were assessed in this study, which limits the expansion of their findings to other athletes. Apart from changes in saddle height, the horizontal position of the saddle can be changed in regular bicycles. Bini and others [126] did not find changes in pedal force effectiveness when cyclists and triathletes were assessed at their most forward and backward position on the saddle. In contrary, less activation was found for hamstrings when cyclists pedaled at a more forward saddle position [125]. In addition, larger activation

68 for gastrocnemius medialis and reduced activation for rectus femoris were found when cyclists and triathletes were assessed at their most backward saddle position [126]. Changes in upper body position could also lead to changes in pedal force by their effect in muscle-tendon unit lengths. A greater forward trunk position was found to reduce by ~9% the pulling forces at the recovery phase potentially due to the reduced length for rectus femoris and other hip flexors [116]. This shorter length could shift the hip flexors away from their optimal length for force production. Therefore, changes in body position can largely affect joint motion and muscle activation with no clear effects in pedal force effectiveness. Connick and Li [124] clearly showed that the neuromuscular system re-organizes the recruitment of lower limb muscle when cyclists pedal at different saddle heights in order to sustain energy cost. That leads to little or no effect on pedal forces when saddle height is changed by ±4%. In contrary, a forward projection of the upper body leads to significant reduction in upward forces during the recovery phase by lowering activation and potentially force production from hip flexors.

5.5.4 Fatigue Conflicting results have been found in terms of fatigue effects on pedal forces and pedaling technique. The index of effectiveness has been referred to remain largely stable when changes in fatigue state are observed (e.g. via changes in muscle activation). [196,151]. In contrary, large increases in pushing downward and pulling backward forces on the pedal were observed [52,207], along with less pulling upward force at the cranks [151]. These results suggest that coordination during pedaling can be changed by fatigue state, re-enforced by larger ankle range of motion and greater time at ankle dorsi-flexion [207,151]. This latter observation could be linked to an attempt to increase the length of triceps surae during pedal pushing. Greater activation for hip [52] and knee [106] joint extensors is contrasting to the stability in joint kinetics during fatiguing cycling [150]. Changes in muscle co-activation were refuted during fatiguing sprint cycling [129] and during cycling time trial [64]. In summary, exhaustion during cycling seems to be postponed by increases in activation of main muscle drivers (e.g. hip and knee extensors) along with changes in ankle joint motion to sustain similar force transfer to the cranks and keep power production. Therefore, little effect should be expected in pedal force effectiveness measured by the index of effectiveness.

69 5.5.5 Skill and experience In cycling, pedaling skill has been related to the ability to transfer the majority of the force applied to the pedals into crank motion (i.e. larger index of effectiveness). Accordingly, greater index of effectiveness was found for cyclists compared to nonathletes [133] and for cyclists compared to triathletes [138]. However, these findings were not supported by the comparison between competitive and recreational cyclists [194] or by comparing cyclists to triathletes in a latter study [126,149]. Trained cyclists also presented greater hip flexion than non-athletes without differences for knee of ankle mean angles and range of motion. Recreational cyclists [208,131] and non-athletes [209-213] improved their pedal force effectiveness when training using visual and verbal feedback of pedal forces. Larger hip and knee extension were found along with less dorsi flexion after a single session with pedal force feedback [209]. These changes resulted in lower knee flexion moment and larger ankle plantar flexion moment after training. In summary, cyclists with better skill levels present greater index of effectiveness, less co-activation and larger hip and knee flexor moments. Further research is needed to assess if long term cycling training could affect joint angles.

5.6 Pedaling technique vs. performance It would be intuitive to expect that more skilled cyclists could have better performance. However, a study assessing 14 competitive cyclists showed that cyclists with better indices of effectiveness were not the ones who achieved better performance in cycling time trials [28]. More recent studies observed that when cyclists were asked to improve pedal force effectiveness by changing their pedaling technique, they reduced cycling efficiency for a given workload [133,189]. The potential reason for this reduction is a limitation for cyclists to recruit hip and knee flexors during aerobic exercise. The possible low percentage of type I fibers in these muscles could lead to greater energy cost to increase activation from hip and knee flexors [189]. A different experimental design was used decoupled cranks to elicit a more active crank traveling during the recovery phase. Training with these decoupled cranks (PowerCranks® or SmartCranks®) did elicit larger activation from hamstrings [214] and better indices of effectiveness [215]. However, no improvements in cycling performance were observed for trained cyclists [216,215]. Therefore, following current interventions

70 presented in the literature, it is not likely that better performance could be achieved by training to improve pedal force effectiveness.

5.7 Conclusions and practical applications Many studies directed their attention to kinetics and kinematics of cycling. The measurement of pedal forces has been assessed in order to provide feedback to improve performance of cyclists. Therefore, it is critical to assess the effects of changes in pedaling technique and pedal force application on energy cost in cycling. For achieving this purpose, combined measures of pedal force application, joint motion and muscle activation is critical. Actually there is good understanding on the effects of workload level and pedaling cadence on pedaling technique and pedal force application. On the other hand, further assessment of changes in body position on the bicycle and pedaling technique training is needed to provide optimal practical advices to coaches and athletes. Effects of changes in bike configuration and training to change pedal force direction using existing methods have been shown not to provide improvements in performance.

71 CHAPTER 6: NON-TRAUMATIC INJURIES IN CYCLING

Rodrigo R. Bini and Thiago A. Di Alencar

6.1 Introduction Non-traumatic (overuse) injuries are commonly reported among cyclists [7]. There are indeed multiple injuries in cyclists which may enhance the referred percentage of total injuries [7]. Different from traumatic injuries, there are potential mechanisms related to the development of overuse injuries, which somewhat limits the proposition of effective preventive strategies. During one year, 33 million of Americans used their bicycles, in an average of six times per month for one hour of cycling [6]. Assuming that 85% of cyclists will develop one or more injuries during their lifetime, approximately 23 million cyclists will get injured at some point [7]. Looking at body sites for injuries, the knee (23-50%) [7] and the low back (46%) [217] are among the most common sites of pain and/or injury in cyclists. An expected link between knee pain and increased knee flexion angle was observed in a preliminary study [218]. Latter findings did not indicate that cyclists with non-specific knee pain could present similar joint forces and knee flexion angles from pain-free cyclists [219]. In addition to bicycle configuration, cycling training profile (volume and intensity) has been suggested to affect the likelihood for development of knee injuries [220]. However, to date, no evidence has been provided in this direction. With that in mind, the aim of this chapter is to present evidence on the occurrence of injuries in cyclists. Potential preventive strategies will be briefly presented.

6.2 Occurrence of overuse injuries in cycling The acquisition of data on injuries mostly depends on the use of surveys in a crosssection perspective [221,217] or on the interviews of cyclists after a long tour [222,223]. The first method is limited to a reduced number of cyclists given cyclists are commonly assessed individually and population could be specific in the case of professional cyclists [217]. In the second method, cyclists usually respond to a survey reporting body sites for injuries or pain during a particular race.

72 Regardless of data collection method, Dettori and Norvell [7] observed that up to 85% of cyclist can develop an overuse injury during their lifetime, potentially because multiple injuries can be reported in some cases. Amongst the most common sites of injuries were the knee (21-65%), the upper back (9-66%), hands/wrist (10-70%), buttocks (42-64%) and the low back (30-75%) [7]. The anterior knee pain has been commonly linked to reports of chondromacea patellae, patellar tendinitis, iliotibial band syndrome and hamstrings tendinitis [11]. For the upper back, compression of braquial nerves due to long term recruitment of trapezius has been reported [224]. Compression of the ulnar nerve has been also linked to longer support of weight on the wrists and hands [225,226]. Pain on the buttocks has been associated to the occlusion of cavernosal arterial inflow [227] and a potential risk of sexual functions [228]. The lumbar pain has been linked to potential compression of nerve roots that drive action potentials to and from the spinal cord due to posterior protrusion of the intervetebral disk [157]. In parallel, deactivation of erector spinae muscles have been related to the non-specific low back pain [220], which resulted in a low back flexion pattern [156]. An alternative approach to assess injury mechanism is the use of biomechanical methods to assess kinetics and kinematics of joint movements. The common approach assesses cyclists (or non-cyclists) with and without injuries in order to compute the predicted loads at a given joint (e.g. knee). For changes in saddle height, studies opted for this approach [30,3] given changes in saddle height potentially affect joint motion and knee loads. In Figure 1.6, a biomechanical model of the tibiofemoral and the patellofemoral joints is shown to illustrate the forces acting in these joints during pedaling.

73 Fig 1.6. Illustration of a biomechanical model used to calculate tibiofemoral and patellofemoral compressive forces during pedaling [13].

A limiting factor of this model is that it assumes joint follows a hinge type motion with no translation between bones. The assessment of individual-subjects moment-arms is also difficult and reference values from the literature are used to compute muscle-tendon forces from joint moments. In the case of multi-joint systems (e.g. spine), it is difficult to state a rotation axis and other approaches are preferred (e.g. finite elements analysis). For inverse dynamics models, muscle-tendon forces are calculated from net joint moments, which are commonly affected by effects from varying muscles. At the knee joint, force from quadriceps could be cancelled by hamstrings and gastrocnemius, which could reduce predicted muscle-tendon and joint forces.

6.3 Overuse injury mechanisms in cycling Table 1.6 presents the most referred injury mechanisms among cyclists. Data are presented in order to assess the strength of evidence surrounding the like hood of each mechanism in determining occurrence of injuries. Guidelines from Law and Macdermid [229] were used to rate articles for the impact in 1a (systematic review of randomized controlled trials), 1b (randomized control trial study), 2a (systematic review from controlled case studies), 2b (controlled case studies), 3a (systematic review of case studies), 3b

74 (case studies), 4 (experimental studies with little control on intervenient variables), 5 (opinion from experts).

75 Table 1.6. Most commonly referred mechanisms for the development of overuse injuries in cyclists. Body site

Type of injury

Advocated mechanism

Strength of evidence

Knee

Chondromalacea patellae

Saddle too low or with excessive anterior projection (anterior knee injuries)

4 [11], 5 [224], 5 [230], 2a [41], 4 [231], 2b [3,30],

Tendinitis patelae Iliotibial band friction Hamstrings tendinities Upper back

Cramps Miofacial pain Compression of brachial nerves

Saddle too high or excessively posterior projection (iliottibial and hamstrings injuries) Handlebars too far away from the saddle

5 [224]

Use of aerobars

Hands/wrists

Compression of the ulnar nerve

Long rides

2b [225]

Buttocks

Pudendal nerve compressiion

Too narrow saddle

2a [7]

Reduced oxygen supply through the perineal artery

Longer time of sited riding

Compression of posterior nerve roots on the spine

Handlebars too far away from the saddle

Overload on the anterior portion of the sacrum promontorium

Excessive incline for the anterior portion of the saddle

Low back

5 [224], 2b [162]

76 Following reports from Dettori and Norvell [7], support to existing knowledge on injury mechanism has not been based on evidence from randomized controlled trials. Only a few studies conducted case controlled studies to assess the effectiveness of preventive strategies in reducing pain or injury in cyclists. The major link between biomechanics and injury risk has been provided by studies assessing predicted loads on the joints when changing bicycle configuration. One example is that large reductions in saddle height (i.e. greater than 9% of a reference height) could lead to large compressive forces at the patelofemoral joint [3].

6.4 Proposed strategies to reduce overuse injury risk Due to the limited evidence found in the literature when the efficacy of various strategies to prevent injuries in cycling is analyzed, the majority of existing practice has been only supported by empirical knowledge. The most used strategies are changing configuration of bicycle components in order to seek for an “optimum” body posture in the bicycle. That could lead to lower joint loading. Following again the example of the knee joint loading, changes in vertical (height) and horizontal (foreback) position of the saddle could theoretically change knee forces, as per illustrated in Figure 2.6. Fig 2.6. A - Illustration of the projection of the patella in relation to the pedal spindle when crank is at the 90° (3 o’clock) position [232]. BIllustration of the knee flexion angle () when the crank is at the 180° (6 o’clock) position [12,233].

76

77 Excessive anterior projection of the patella in relation to the pedal spindle (Figure 2.6-A) and/or increased knee flexion angle at the bottom of the stroke (Figure 2.6-B) have been linked to larger patellofemoral compressive forces [9,233]. The main reason for this expected change is based on the inverse relationship between saddle height and knee flexion angle [32]. This change in knee angle could lead to changes in muscle-tendon [234] and potentially to changes in muscle-tendon force. An illustration of effects from increased knee angle on patellofemoral compressive force is shown in Figure 3.6. Fig 3.6. An increase in saddle height from position illustrated in A to B (3 cm) could reduce the knee flexion angle from 47° to 31° and therefore, reduce the compressive patellofemoral force (FP vector). TQ is for quadriceps tendon and TP is for patella tendon.

The example illustrated in Figure 3.6 has been experimentally tested by Bini and Hume [219] when assessing recreational cyclists with and without non-specific knee pain. In this study, changes in saddle height of 2-7% of cyclists preferred height showed that although large increases were found in knee flexion angle (13-20%), patelofemoral compressive force was not largely affected (~4%). The reason for this result is potentially linked to the use of controlled power output and pedaling cadence across different saddle height trials. A second reason could be linked to combined changes in hip and ankle joint angles to accommodate increases (or reductions) in saddle height in order to equally share additional forces among the three lower limb joints. Similar findings were observed when assessing trained cyclists and triathletes pedaling at saddle heights 4-6% lower (and higher) than their 77

78 preferred height. [91]. It is important to state that when larger decreases were conducted in saddle height (i.e. 9%>) increases were observed in patelofemoral compressive forces [3]. That could highlight the example provided in Figure 3.6, which may in turn damage the cartilage matrix if longer training is sustained using too low height of the saddle. Regardless of the height, the saddle can also be changed in the horizontal direction (fore-back). Silberman and fellows [232] suggested that optimal saddle horizontal position should enforce that the patella lies on the pedal spindle, as per shown in Figure 2.6-A. However, triathletes could be at higher risk of developing anterior knee pain and/or injuries given they commonly opt for a more anterior position of their saddles leading to larger anterior projection of their knees (~6 cm compared to ~2 cm for road cyclists) [146]. Two studies provided evidence in contrary to the potential higher risk of using a larger anterior knee projection. The first [223] observed that occurrence of knee joint injuries in triathlon is similar to the observed in road cycling. The second study compared cyclists and triathletes pedaling at preferred and anterior position on their bikes (6 cm forward then their preferred positions) showing that, although large knee flexion was observed, increases in patellofemoral compressive force did not change significantly (4%) [34]. This latter finding is enforced by observations that workload and pedaling cadence are much more effective in changing knee loads than changes in saddle position [13,3]. Apart from studies assessing impact of acute changes in saddle position on knee joint loads, limited evidence has been presented concerning the potential impact of changes in bicycle configuration and their likelihood for injury prevention (see Table 1). Based on that, it is difficult to define optimal strategies to reduce and/or minimize upper back pain. It has been suggested that refraining from using low handlebars or keep the saddle closer the bars could be effective strategies. These strategies should minimize hyperextension for the cervical section of the spine [224]. Research assessing upper back muscle activations using different settings for the handlebars could help addressing this question. The occurrence of numbness in the hands from the compression of the ulnar nerve has been related to the maintenance of the hand in the same position during prolonged pedaling [225,235]. The use of padded gloves has been proposed as a strategy to reduce the excessive weight on the hands [225]. Slane et al. [236] 78

79 observed that using foam gloves reduced pressure at the hypothenar portion of the hand compared to gel gloves and no gloves. For gel gloves, only changing from 5 mm thickness resulted in similar cushioning than observed for foam gloves. Like handlebars, saddle also supports an important portion of body weight of the rider. Due to the reduced contact area between the saddle and perineal area, high pressure in the perineal area and the consequent compression of the vessels that supply the penis and scrotum have been showed [237]. One approach to reduce the pressure in the perineum is the use of a saddle with a cut relief area. However, this type of saddle may not provide satisfactory results when compared to a conventional saddle [238]. Saddles with a cut relief area tend to decrease the pressure on the medial portion of the perineum, where lies the neurovascular bundle for the penis and scrotum [239]. The cut relief saddle could indeed preserve the space between the saddle and the perineum, but the posture of the rider still remains the most important variable for the reduction of compression [239]. Reductions in saddle pressure were observed only when the saddle was compared to a “noose less” saddle [240] (Figure 4.6). However, further studies indicate that the width of the seat is critical in the prevention of genitourinary tract dysfunction, and the distance between the ischial tuberosities could be an important anthropometrical parameter to determine the optimal saddle for each cyclist [241,242]. A secondary strategy could be the reduction of workload [238], by changing gears and managing riding incline in order to reduce saddle pressure.

79

80 Fig 4.6. Schematic diagram of a conventional saddle (A) and a “noose less” saddle with cut relief area (B).

The lumbar region may have a high incidence of discomfort or pain during and/or after the ride. Often pain presents itself directly related to the time spent on the bike [155] or due to the posture used when pedaling [157]. Another trigger to low back pain is strength of trunk muscles (e.g. abdominal and spine erector muscles). For now, only saddle inclination angle has been experimentally shown effective in reducing low back pain. Reduction in saddle anterior portion in relation to the posterior portion has been shown to reduce low back pain in 72% of cyclists [162]. The reason for the effectiveness in changing saddle angle is linked to reduction in lumbar flexion and potentially to lesser load applied to sacrum promontorium.

6.5 Conclusions and practical applications Non-traumatic injuries may affect up to 85% of cyclists and some cyclists may sustain more than one injury simultaneously. The knee, cervical, hands, back and perineum are the most affected with uncertain mechanisms. Characteristics of bicycle components related to body size of the cyclist have been identified as potential determinants for non-traumatic injuries. Few studies have been conducted to define the relationship between these predictors and the occurrence of injuries. Few strategies for prevention and/or minimization of non-traumatic injuries were

80

81 assessed experimentally. Therefore, clinical practice has been mostly based on empirical knowledge rather than experimental evidence.

81

82 CHAPTER 7: BICYCLE TYPES AND SIZES

Rodrigo R. Bini, Frederico Dagnese and Julio Kleinpaul

7.1 Introduction Bicycle components should be adjusted to enable cyclists to properly use the bicycle aiming at improving health status with the minimum possible injury risk. The importance of proper positioning on the bike based on biomechanical principles has already been demonstrated for both injury prevention and to increase performance [243,244,147]. The primary step for selecting a bicycle is the definition for a frame size which should match cyclist’s body dimensions. For a given frame size, saddle height [245], size of handlebars [10] and crank length [24] should be configured. These settings must be adjusted based on anthropometric characteristics of the cyclist and flexibility, because these parameters will determine force application pattern on the pedals, neuromuscular recruitment strategy, energy expenditure, the likelihood of developing injuries, aerodynamic drag and comfort on the bike [10,246]. An improper body position on the bike may, in most cases, be related to lack of information from cyclists on the proper way to adapt their bikes and its components to their body sizes [247]. Although there are methods to adjust cyclists’ body position on the bike [10], cyclists usually opt for adjustments based on subjective sensations [245,10]. A previous study showed that recreational cyclists present larger errors in saddle height setting (1-8 cm) compared to competitive cyclists (1-2 cm) [247]. In addition to this, training volume and intensity should be taken into account to ascertain on the potential factors for non-traumatic injuries in cyclists. Therefore, errors in bicycle configuration could lead to injuries, reductions in pedal force application, increased energy cost. With that in mind the purpose of this chapter is to present the main aspects related to bicycle configuration. We seek to provide information for cyclists to optimally use bicycle for exercising, training and health improvements.

82

83 7.2 Types of bicycles When choosing a bicycle, there is need to define the goals for its use keeping in mind that there as numerous types of bicycles available on the market. The most commonly used in training and racing are the road bikes, mountain bikes and urban (commuting) bikes. Each of these models has unique characteristics in terms of comfort and potential to configure moving parts.

7.2.1 Road bike Road bikes (Figure 1.7) have been an option for almost all ages and population with recreational or competitive purposes. This bike is used for individual or group practice whenever held in paved roads to allow high traveling speeds (>50 km/h). Low handlebars enable a reduced projected area for better aerodynamic profile and low rolling resistance due to narrow tire type. In addition, the reduced mass for road bike (between 7-10 kg) results from the use of light materials for constructions (aluminum or carbon fiber) and to the absence of headlights or luggage space. The tires are narrow and flat, which also ensures low friction coefficient. The number of gears is between 10 and 20.

Fig 1.7. Illustration of a road bicycle.

83

84 7.2.2 Mountain bike Mountain bikes (Figure 2.7) are very versatile as they allow the practice of various forms due to wide tires that adhered to the surface of contact and damping systems on the front-posterior and medial to minimize the impact in order to protect the structure of the bike and the musculoskeletal system. This bike is then suitable for places where the terrain is uneven. The number of gears is between 20 and 30.

Fig 2.7. Illustration of a mountain bicycle.

7.2.3 Hybrid bicycles The recreational bicycle model (Figure 3.7) merge features from road cycling and mountain biking and can weigh between 12 and 20 kg. Generally this type of bicycle has straight handlebars with several gears (18-27). These bikes should provide greater comfortable in riding due to the closer position of handlebars to vertical, leading to reduced upper body lean. Many recreational bicycles have fenders or headlights. These are used by the majority of the population with different objectives such as leisure, travel to work and to transport small items in various types of terrain.

84

85 Fig 3.7. Illustration of a hybrid (recreational) bicycle.

7.2.4 Triathlon bicycles The triathlon bicycle (Figure 4.7) is particularly different from road bicycles given the saddle is usually projected forward by either changing the seat tube angle or by changing the horizontal position of the saddle [146,16]. Another main feature is the use of aerobars to reduce trunk flexion and improve aerodynamics profile. This change in position of the upper body is also observed in time trial bicycles.

85

86 Fig 4.7. Illustration of a triathlon bicycle.

7.3 Frame size Frame size should accommodate the anthropometric characteristics of the cyclist, regardless the aim of the practice. In shops the choice of frame sizes is usually based on the height of the rider. For the road bikes the unit of measure used is in centimeters (cm) while for mountain bikes, frame sizes are usually in inches. The measurement is taken from bottom bracket to the top of the seat tube (tube where the seat is fixed). Table 1.7 is commonly used and is based on the relationship between the cyclist's height and frame size.

 

86

87 Table 1.7. Determination of the optimal frame size according to the cyclist’s height. Information is presented considering the most common units for measures for road and mountain bikes. Height of cyclist

Road bike frame size

Mountain bike frame size

1.50 – 1.60 m

48 cm

15”

1.60 – 1.70 m

50, 52, 54 cm

16-17”

1.70 – 1.80 m

54, 55, 56 cm

18-20”

1.80 – 1.90 m

57, 58 cm

21-22”

1.90 m – >

60, 62 cm

22” >

Few evidence has provided suggestions for setting bicycle frame size based on cyclists standing height [248]. However, this method is not appropriate for determining frame size given the ratio between the height and leg length can vary between cyclists. Other methods can be used to determine frame size, like the height of the crotch (i.e. inseam leg length). Table 2.7 illustrates this relation to the definitions for road and mountain bikes.

 

87

88 Table 2.7. Definition of bicycle frame size using the inseam leg length. Option for different units was based on the most common definition described in the literature [147]. Mountain bike frame

Inseam leg length

Road bike frame size

72 cm

48 cm

14.5“

74 cm

49 cm

15“

76 cm

50 cm

15.5“

78 cm

51 cm

16“

80 cm

53 cm

16.5“

82 cm

54 cm

17“

84 cm

55 cm

17.5“

86 cm

57 cm

18“

88 cm

58 cm

18.5“

90 cm

59 cm

19“

92 cm

61 cm

19.5“

94 cm

62 cm

20“

size

The inseam leg length is measured with the cyclist standing upright barefeet about hip-width apart and accommodating a level between the legs, maintaining contact with the base of the pelvis (the area between the pubis and the basis of ischial tuberosity). The height from the level to the floor is added to the size of the book, illustrated in Figure 5.7.

88

89 Fig 5.7. Lower limb measures commonly used to configure saddle height (1- inseam length, 2- throcanteric length and 3- ischial tuberosity length).

Although definition of frame size can be an important starting option, there may be differences in frame geometry between brands, which limit the use of references provided in tables 1.7 and 2.7. In this line Burke and Pruitt [233] stated that the seat tube angle is the most important component on bicycle frame. In addition, recent suggestions have been made to define optimal bicycle frame based on measures of height and length of frame (i.e. stack and reach, respectively) with focus on triathlon bicycles [249]. More research is needed in order to illustrate differences in body position of cyclists on different bicycle assuming a given frame size. This would be important information because many bicycle makers develop similar frame sizes with varying components (e.g. different crank lengths and handlebars steam lengths.

7.4 Crank length and noncircular chainrings Anthropometrical characteristics of the cyclist can influence the crank length, which may affect the physiological and biomechanical aspects of cycling performance [250]. The most common anthropometric measure to choose the crank 89

90 length is the thigh length. However, saddle configuration (height and fore-back position) largely influences the chosen crank length. Similar to isolated saddle height configuration, short saddle associated to short crank length could lead to excessive knee flexion, which could lead to increase knee loads [9,3]. In contrary, pedaling cadence has been described to increase by using shorter cranks [22]. It was also observed that the peak force applied to the pedal decreases with increasing crank length [22]. That could minimize the load per pedal stroke assuming that to sustain a given power output, pedalling at higher cadences results in less muscular force due to increase contribution from inertial components [119]. Longer crank lengths (195 mm) resulted in larger activation of biceps femoris, tibialis anterior and soleus [251]. Noncircular chainrings have been designed to optimize variables related to cycling performance (e.g. peak crank torque and efficiency), with conflicting results in terms of performance [27]. Changes in joint kinematics without effects on muscle activation were observed when cyclists were assessed using noncircular chainrings [252,184]. In theory, noncircular chainrings have the potential to enhance cycling performance due to smoother fluctuations in crank angular velocity and reductions in internal work [253,254]. However, conflicting results for economy/efficiency explain the unclear effects of noncircular chainrings on cycling performance [255,256,26]. For more details, a literature review has covered this issue [27].

7.5 Handlebars Type of handlebars is varied among makers, but little research has been done to compare potential effects from using each particular handlebars in cycling posture. Most attention has been given to the position of the hands in road bicycle handlebars [257,164,166]. This provides some relation to position used in mountain bike (top of handlebars) but a gap on effect of changes in distance from saddle to the position of the hands remains present in bike fit research. Latter discussion will be provided on the custom design of aerobars for paralympic cyclist (see chapter 10 [258]). In summary, although variability in design of handlebars exist, no research has given attention to the potential effects of these variation in cycling posture and other outcomes (e.g. upper body angles, pressure distribution on the hands, comfort, etc).

90

91 7.7 Conclusions and practical applications Bicycles manufactured for recreational purposes should allow wider ranges of adjustments for the saddle and handlebars in order to improve comfort for cyclists. Regular low cost bicycles sold for commuters are usually limited to a single frame size, which limits the range of body sizes that could be accommodated. Ideally, customized bicycles should be the target given frame sizes can still vary in terms of ranges of changes for the saddle and handlebars. In order to make the fine tuning of the cyclist to the bicycle, it is important to consider the proper body position and to configure bicycle components as described elsewhere in this book.  

 

91

92 CHAPTER 8: OPTIMIZING BICYCLE CONFIGURATION AND CYCLISTS’ BODY POSITION

TO

PREVENT

OVERUSE

INJURY

USING

BIOMECHANICAL

APPROACHES Rodrigo R. Bini, Patria A. Hume, James L. Croft and Andrew E. Kilding

8.1 Introduction Overuse injuries in cycling are common [6] with up to 85% of cyclists sustaining one or more overuse injuries during their lifetime [7]. In one year, 33 million USA residents rode a bicycle an average of six days a month, for an average of >1 hour a day [6], suggesting that approximately 23 million cyclists may develop at least one overuse injury in their lifetime. Overuse injuries of the anterior knee joint (e.g. chondromalacea) are most common, affecting about 50% of injured cyclists [7]. Cycling overuse injuries are not confined to the lower extremity as 46% of elite cyclists reported low back pain [217]. A cyclist is in contact with the bicycle via the handlebars, saddle and the pedals. Consequently, the way a bicycle is configured can change a cyclists’ body riding position. For example, if the saddle height is too low then the larger knee flexion angle (defined as the angle between the tibia and the forward projection line of the femur) close to peak pedal force may increase knee joint forces and lead to overuse injuries [7]. A common strategy to reduce the risk of overuse injuries is to change bicycle configuration which leads to a different body position on the bicycle. Therefore, it is necessary to review existing guidelines for bicycle configuration that aim to help prevent overuse injuries, and to present potential benefits and limitations of biomechanical methods employed in cycling research to help prevent overuse injuries. Sports medicine personnel and coaches should be aware of what has been recommended for prevention of overuse injuries and which methods, if any, are effective. The aim of this review was to evaluate evidence for the effectiveness of bicycle configuration and cyclists’ body position to prevent overuse injuries in cycling using current biomechanical approaches.

92

93

8.2 Methods Peer-reviewed

journals,

books,

theses,

and

conference

proceedings

published since 1960 and up to July 2011 were searched using Medline, Scopus, ISI Web of Knowledge, EBSCO, and Google Scholar data bases. Keywords searched included ‘bicycle’, ‘overuse injury’, ‘body position’, ‘bike fitting’, ‘kinetics’, ‘kinematics’ and ‘electromyography/EMG’. Articles were excluded if they did not have at least an English abstract, or if they focused solely on acute cycling injuries. Sixty references were reviewed.

8.3 Findings The low back and knee were the main body sites of overuse injuries in cyclists [217,7]. Most preventive strategies for overuse injuries were based on optimizing the interaction between body position and bicycle components [6]. No randomized controlled trials investigating risk factors for overuse injury in cycling were found. One prospective case control study reported that 70% of cyclists reduced low back pain when the saddle was tilted with the anterior portion downward by 10-15° [162] (see Table 1). Some cross-sectional studies compared kinematics and kinetics of injured versus uninjured cyclists [259,260], though no study provided evidence of likely thresholds for overuse injury risk for the various biomechanical variables analyzed. Although a relationship between improper bicycle configuration and injuries in cyclists has been suggested [6], evidence for the effectiveness of clinical biomechanical evaluation of body position and bicycle configuration in preventing the likelihood of overuse injuries in cycling was not found. Therefore, this review reports approaches to optimize configuration of bicycle components (handlebars vertical and horizontal position; saddle height and horizontal position; position of foot on the pedal)

and

cyclists’

body

position

using

biomechanics

approaches

of

anthropometrics, dynamic joint kinematics, pedal forces and joint kinetics, and muscle activation measures.

93

94 8.4 Optimization of configuration of bicycle components and cyclists’ body position The height of handlebars, saddle height and horizontal position, and position of the foot on the pedal (or position of the cleats) can all be changed [10] but the distance from the saddle to the handlebars and crank length seem to be the most important for bicycle configuration components. Optimal muscle force production occurs within a certain range of muscle length, and consequently within a range of joint angles [42]. Therefore, it may be expected that optimal joint angles would lead to maximal force production. However, most of the references for optimizing the configuration of bicycle components are based only on the length of body segments [10] without concern for joint angles.

8.4.1 Handlebars vertical and horizontal position The horizontal (distance from the handlebars to the saddle) and vertical (height) positions of the handlebars affect the upper body flexion angle [10,65]. The position of the handlebars has been empirically related to the sum of lengths of the torso and the arm [10]. Usually the horizontal position of the handlebars is fixed unless the headset of the bicycle is changed. Therefore, the effects of the combined horizontal and vertical position of the handlebars will dictate the angles of the trunk and pelvis (see Figure 1.8).

94

95 Fig 1.8. Flexion angle of the trunk (θT) and pelvis (θP) in the sagital plane. Insert A: Three positions of the hands on the handlebars. Insert B: Forward-backward configuration of the shoe on the pedal.

To allow comparisons between studies, trunk and pelvis angles have been defined as presented in Figure 1. Trunk and pelvis angles were smaller with the hands on the top of the handlebars (141 ±0.9 for the trunk and 16 ±1.4 for the pelvis) than when the hands were on the brake hoods (156 ±0.7 for the trunk and 29 ±1.6 for the pelvis) [166]. This finding highlights that if the hands are further away in relation to the body, trunk and pelvic angles will be increased. In further support, increased angles between the lumbar vertebral disks when comparing hands on the top of the handlebars (~2.5) with the hands on the drops of the handlebars (~3) by radiologic analysis [167] of three professional cyclists positioned statically on their bicycles. This result indicates that a forward inclined upper body position of cyclists (compared to upright standing) will affect lumbar disk displacement with small effects from changes in position of the hands on the handlebars. High rates of low back pain (up to 60%) [157] have been reported for cyclists, which may result from forward trunk flexion and pronounced lumbar kyphosis on the bicycle compared to the standing upright position [167]. Burnett et al. [157] reported no significant difference between the pelvic angle for cyclists with (15 ±5.1) and without (23 ±5.8) low back pain during fatiguing cycling with the hands on the drops of the handlebars. It is possible however that large between-subject variability in

95

96 trunk and pelvic angles may hide any relationship between lower back pain and upper body position in cycling [157]. Furthermore, both saddle shape [166] and inclination angle [162] have been found to affect the angle of the pelvis which may make it difficult to isolate the relationship between trunk angle and low back pain. The interaction effects between the horizontal and vertical positions of the handlebars, the angles of the trunk and pelvis and any relationships with overuse injuries are not yet understood.

8.4.2 Saddle height and horizontal position The saddle can be adjusted in the vertical and horizontal directions without changing its inclination angle in the sagital plane. The position of the saddle can be measured relative to the bicycle components (e.g. saddle position relative to the bottom bracket) or relative to the cyclist (e.g. saddle height relative to the throchanteric length). When using the length of body segments to configure the saddle height, the distance from the greater throchanter to the floor, the distance from the pubis to the floor and the distance from the ischial tuberosity to the floor have all been used [12]. However, there is concern that using length-based configuration methods may not result in similar joint kinematics [147]. Therefore, the range of the knee flexion angle (e.g. 25-30 when the crank is at the 6 o’clock crank position) has been recommended as a more reliable method of configuring saddle height [147] to achieve consistent lower limb kinematics. For the horizontal position of the saddle, it has been recommended that the vertical projection of the knee should intersect the pedal axis [6] based on the assumption that knee positions forward of the pedal axis may result in higher compressive forces on the patellofemoral joint [6]. Triathletes often use a more forward position of the saddle compared to road cyclists [125] which suggests that they should have a higher risk of knee injuries than road cyclists but prevalence of knee injuries in triathletes (14% to 63%) [223] and cyclists (21% to 65%) are similar [7].

8.4.3 Position of the foot on the pedal The position of the foot on the pedal can be changed in several ways. When cyclists use a cleat between the shoe and the pedal the position of the foot-shoe in

96

97 relation to the pedal can be changed in the anterior-posterior direction and rotated about the longitudinal axis (see Figures 1.8 and 2.8, respectively). Fig 2.8. Medial projection of the knee (continuous line) in relation to the pedal axis (dashed line) in the frontal plane. Insert: Medio-lateral and rotational configuration of the shoe in relation to the vertical axis of the pedal.

The most common recommendation for optimal foot positioning is that the ball of the foot should lie over the pedal axis (see Figure 1) [6]. However, when the horizontal position of the foot on the pedal is varied (forward or backward, see Figure 1), no significant effects on knee joint forces [3], muscle activity [261], pedal forces [204] or oxygen uptake [262] are observed. Therefore, there is no current evidence to support that the horizontal position of the foot on the pedal can be optimized to reduce the risk of overuse injuries or improve performance. Most cleats allow the shoe to rotate around the vertical axis (toes inward or outward) to a certain extent. It has been suggested that cleats that do not enable such rotation movements may increase forces on the knee joint and increase the likelihood of injuries [79]. It is likely that some rotational movement between the cleat and the shoe with respect to the vertical axis may improve force transfer from the 97

98 legs to the pedals. However, the optimal range of motion around the vertical axis is currently unknown. The position of the foot-shoe along the longitudinal axis of the pedal (mediolateral displacement of the foot) has been related to the knee angle in the frontal plane [263]. It is usually recommended that the knee should remain directly over the pedal axis in the frontal plane during the pedaling cycle (see Figure 2) [264]. However, the knee moves medially between the 12 o’clock and 6 o’clock crank positions (propulsive phase) and externally between the 6 o’clock and 12 o’clock crank positions (recovery phase)[179]. The amount of medial knee movement that is tolerated is unknown but cyclists with anterior knee pain have been shown to have 320% greater medial displacement than cyclists without anterior knee pain [179] indicating that excessive medial movement may be associated with knee pain. It is therefore important to consider the position of the foot-shoe relative to the pedal and its effect on the position of the knee (e.g. such as when using orthotics) to reduce knee injury risk and improve cycling economy.

8.5 Biomechanical approaches for optimizing bicycle configuration and cyclists’ body position to prevent overuse injury The most frequently suggested strategy to reduce injuries in cyclists is to optimize body position [6] by adjusting the bicycle configuration to fit the cyclist’s anthropometry [10]. Additional approaches include altering pedaling technique using dynamical video analysis and force measurement, muscle activity and simulation models. Table 1.8 summarizes studies that have used biomechanical approaches to provide analyses of joint kinetics and kinematics and muscle activity in the scope of injury prevention and rehabilitation for overuse injuries in cyclists. Table 1.8. Summary of studies using a biomechanical approach to provide analyses of joint kinematics, kinetics or muscle activity within the scope of injury prevention and rehabilitation for overuse injuries in cyclists. Biomechanical method; Study

Participants; Injury in

Main results and notes

focus.

98

99 Gregor

Review/case study; 1

Increased moment about the vertical

and

cyclist with anterior knee

axis of the tibia for the symptomatic

Wheeler

pain, 27 uninjured cyclists.

cyclist compared to the uninjured

[79]

No gender or age given.

cyclists. Magnitude of difference not reported.

McLeod

Joint kinematics, muscle

Proposed cyclists should activate

and

activity; 6 male uninjured

hamstrings, hip flexors and tibialis

Blackburn

non-cyclists aged 21 to 30

anterior to prevent excessive

[265]

years. Injuries related to

compressive forces on the patella

tibiofemoral (shear and

from quadriceps activation.

compressive) and patellofemoral forces. Semwal

Joint kinetics and

Software graphical interface for

and

kinematics; 4 elite

analyses of joint moments and

Parker

uninjured cyclists with

kinematics. Focus on provision of

[266]

asymmetry on bilateral

summarized feedback for cyclists

knee movement. No

and coaches. Descriptive analysis of

gender or age given. Knee

possible increased risk of injuries

injuries related to

due to greater rotational moment

rotational moment on the

about the vertical axis of the tibia.

vertical axis of the tibia. Farrell et

Joint kinematics; 6 male

Analysis of knee flexion angle to

al. [231]

and 4 female uninjured

provide possible overload of iliotibial

non-cyclists aged 31 ±5.5

band. Diverse causes for iliotibial

years. Iliotibial band.

band impingement (anatomical, kinematics, forces). More time in the impingement zone (knee flexion < 30) in cycling because of more repetitive action compared to running.

Bailey et

Joint kinematics; 24 active

A more medial position of the knee

al. [179]

male cyclists (14 uninjured

(~2.5 - 320%) in the injured group

and 10 injured) aged 28

compared to the uninjured group. 99

100 ±8.4 years. Patellar tendinitis and/or anterior knee pain. Burnett et

Joint kinematics, muscle

Low back pain group presented

al. [157]

activity; 9 cyclists without

~50% greater lumbar flexion

pain (four male and five

compared to pain-free cyclists with

female aged 37 ±7.9

~50% increased rectus abdominals

years) and 9

activity.

cyclists/triathletes with non-specific chronic low back pain (4 male and 5 female aged 42 ±9.7 years). Joseph et

Joint kinematics; 1 male

Descriptive increased hip movement

al. [161]

27 year old elite cyclist

in the frontal plane related to back

with non-specific chronic

pain.

low back pain. Salai et al.

Joint kinematics; 40

70% reduced low back pain after 10-

[162]

uninjured recreational

15 reduced saddle angle (lower

cyclists with non-specific

forward portion compared to

chronic low back pain. No

backward portion of the saddle). A

gender or age given.

prospective case control study.

Tamborind Joint kinetics and

No differences in peak knee force

-eguy and

kinematics; 9 male

when reducing or increasing saddle

Bini [30]

uninjured non-cyclists

height by 3 cm compared to the

aged 22 and 36 years.

throchanteric height when cycling at

Evaluated potential for

low workload and pedalling cadence.

injuries related to

Knee flexion angle at the 6 o’clock

tibiofemoral compressive

crank position was increased by 41%

and shear, and

at the low saddle position compared

patellofemoral

to the reference and by 18%

compressive load.

compared to the high saddle position. Optimising knee angle may 100

101 reduce overload on the knee joint for low workload and pedalling cadence. Ericson et. Joint kinetics; 6 male

Descriptive profile of medio-lateral

al. [78]

uninjured non-cyclists

moments on the knee. Higher medial

aged 20 and 31 years.

knee forces (~120%) when knees

Evaluated potential for

close to the bicycle frame compared

injuries related to medio-

to the preferred position of the knees.

lateral moments on the knees. Ericson

Joint kinetics; 6 male

Description of knee moment profile.

and Nisell

uninjured non-cyclists

Peak knee moment and compressive

[267]

aged 20 and 31 years.

force between 60 and 100 knee

Evaluated potential for

angle. Peak anterior shear force

injuries related to

close to 65 knee flexion angle.

compressive and shear

Compressive force increased with

forces on the tibiofemoral

higher workload (~100%) or lower

joint.

saddle height (~64%). No effects of pedalling cadence on tibiofemoral forces.

Ericson

Joint kinetics; 6 male

Peak patellofemoral force close to

and Nisell

uninjured non-cyclists

80 knee flexion angle.

[3]

aged 20 and 31 years.

Patellofemoral compressive force

Evaluated potential for

increased at higher workload (~75%)

injuries related to

and lower saddle height (~25%). No

compressive forces on the

effects of pedalling cadence on

patellofemoral joint.

Patellofemoral compressive force.

Ruby et al. Joint kinetics; 11 male

Descriptive profile of the three-

[66]

uninjured cyclists aged 23

dimensional forces and moments on

to 45 years. Evaluated

the knee joint. Important to analyze

potential for injuries

the knee based on three-dimensional

related to rotational

kinetics. Hypothesised that increased

moments on the vertical

rotational moment on the vertical axis

axis and medio-lateral

of the tibia may lead to greater injury 101

102 moments on the knees.

risk.

Srinivasan

Muscle activity; 14 male

Higher fatigue level based on muscle

and

cyclists (7 with non-

activity analysis for the right

Balasubra

specific chronic low back

trapezius medialis and erector spinae

-manian

pain, 7 without low back

for the low back pain group after 15

[155]

pain) aged 25.4 ±1.9

minutes of outdoor cycling covering 6

years.

km.

Studies were mostly conducted using uninjured cyclists and/or non-cyclists. Cyclists with and without patellar tendonitis and/or anterior knee pain [179], or nonspecific low back pain have been compared [157] and characteristics of cyclists with low back pain before and after changing saddle angle [162] have been reported. However, no study has yet reported likely thresholds for biomechanical variables in predicting injury.

8.5.1 Anthropometrics Several studies have highlighted the relationship between improper bicycle configuration and injuries in cyclists [6]. However, there are various complications in applying optimal configurations. Appropriate joint angles and distances from body segments to specific components of the bicycle that have been suggested are not usually based on scientific research [147]. In addition, various measures are interrelated and there is disagreement regarding which should be measured [147]. Typical static analysis [10] is inaccurate because it fails to account for the variability of joint dynamics observed while cycling, which is greater for less experienced cyclists [31]. Farrell et al. [231] described that after setting the saddle height for a knee flexion angle of 30º (0 is full extension) with the crank at the 6 o’clock crank position prior to exercise, this angle changed during cycling to a range between 3045º in ten non-cyclists. Defining optimal parameters based on knee and ankle angles measured dynamically should be better than using angles derived from static postures. The relationship between anthropometric characteristics of the cyclist (e.g. inseam leg length) and the configuration of bicycle components (e.g. saddle height) 102

103 has been related to comfort [14]. However, the use of anthropometric methods based on segment length results in different knee flexion angles for different cyclists [147]. Reduced accuracy of length-based methods for configuration of saddle height may increase the between-subject variability for joint kinematics, as highlighted in the example of Farrell et al. [231]. Poor relationships between the horizontal configuration of the saddle and the upper-leg length have been presented [14]. The optimum saddle height for young cyclists with similar height (e.g. 145-149 cm) was calculated using length-based methods [248] and found to have high variability (17%). Therefore, predicting the relative position of bicycle components from cyclists’ anthropometric dimensions may not be as accurate as has been advocated [10]. No published study has explored the possible relationship between injury and bicycle configuration by relating anthropometric characteristics and the history of injuries to the joint angles based on static or dynamic body positions on the bicycle. Limitations exist by assuming that the static position is representative of the joint position during riding and this must be taken in account.

8.5.2 Dynamic joint kinematics The rapid development in technology has provided significant benefits for movement

analysis [29]. For example, three-dimensional body kinematics

measurements are now available without the need to attach skin markers [268] and the result is more realistic tracking of movement. Unfortunately, these systems are expensive and unaffordable for cycling coaches and bicycle shops, where the configuration of bicycle components for an individual usually occurs. Fortunately, technological advancement and the reduced price of high-speed video cameras have been helpful for coaches using basic image analysis to provide feedback to cyclists. Joint kinematics are sensitive to changes in body position on the bicycle [51] and to the presence of injury [179]. Upper [161] and lower body injuries [179] have been related to altered kinematics which may lead to overload on soft tissues. The major resistive force in cycling, aerodynamic drag, can be reduced by increasing trunk angle to a more flexed position which decreases the frontal projected area [159]. However, for cyclists that suffer from low back pain, increasing upper body flexion may overload the posterior annulus of intervertebral discs [157] and may cause asymmetries in frontal plane kinematics of the low back [161]. However, no 103

104 study has reported trunk and pelvic flexion angles that could minimize overload on intervertebral discs. The relationship between kinematics and injuries (using cross sectional studies comparing injured versus non-injured cyclists) has been described for the knee joint [179] but not for the hip and the ankle joints. High variability in shank rotation about the vertical axis has been reported for international level uninjured cyclists [266] and therefore may increase the risk of knee injuries. Increased frontal plane movement of the knees [179] and higher rotation moment about the vertical axis [79] have been observed in cyclists with anterior knee pain compared to uninjured cyclists. High variability of joint angles has been associated with poor movement skill in cyclists [31] and to the likelihood of injury occurrence in other sports [269]. For the ankle joint, kinematics were found to be substantially different among cyclists of similar performance levels [198] and between cyclists of different disciplines [142]. However, it is unclear if variability in joint kinematics can affect injury risk in cyclists.

8.5.3 Pedal forces and joint kinetics Studies have quantified pedal forces during controlled laboratory trials [52]. To increase ecological validity, Stapelfeldt et al. [94] developed an adaptor to be used between the pedal and the crank that allowed cyclists to use their own pedal. With further development, portable systems that can be used in the field [77] may provide more ecological measurements of cycling mechanics. Unfortunately, these systems are not yet widely available for use by most coaches and cyclists, and usually only biomechanics laboratories provide this equipment. However, whilst wireless systems will no doubt become more available in the future, there is still a need to determine their reliability and accuracy. The quantification of bilateral pedal forces may be useful in evaluating asymmetries of anterior cruciate ligament injured cyclists [259,260]. The combined measurement of joint kinematics and pedal forces can provide an estimate of joint kinetics, using an inverse dynamics model [30], which has been used to describe changes in joint kinetics in injured cyclists [260]. A higher rotational moment about the vertical axis of the tibia has been reported for one cyclist with anterior knee pain [79], while significantly lower knee extensor moments have been reported for anterior cruciate ligament injured non-athletes during pedaling [260]. 104

105 The uninjured leg is often used as a control when looking at the effects of injury on joint kinetics. However, asymmetries in hip [270] and knee joint kinetics [266] have been reported in trained uninjured cyclists which make comparisons difficult to interpret. It is currently unclear what threshold of asymmetry can be defined as “functional” for uninjured cyclists for the purpose of injury prevention. Cycling is commonly analyzed as a pure sagital plane movement, which neglects medio-lateral forces that affect non-driving structures of the joints [66]. Complete three-dimensional analysis of joint kinetics has rarely been reported [66]. To give coaches an understanding of the various forces acting upon a cyclist, Semwal and Parker [266] developed an animated computer software that provides selectable analyses of three-dimensional pedal forces and joint kinematics. Coach friendly computer software like this may enhance the use of joint kinetics as a tool for injury prevention. One example is the identification of rotational moments at the knees that are related to overload on the menisci and cruciate ligaments [79]. Once assessed using three-dimensional joint kinetics analysis, rotational knee moments can be minimized by using floating rather than fixed cleats at the shoe-pedal interface [41]. Inverse dynamics analysis in the sagital plane has been used to calculate the moments and resultant forces on the joints and such measures may be useful in preventing joint overload [267]. Tibiofemoral compressive and shear force components in the sagital plane have been compared using different bicycle configurations (e.g. saddle height), however, conflicting results were found [30,271,267] potentially due to differences in cycling experience (i.e. cyclists versus non-athletes) or magnitude of changes in saddle height. Therefore, further research should determine the effects of bicycle configuration on sagital and threedimensional forces of the lower limb joints.

8.5.4 Muscle activation The profile of neural drive using surface electromyography has been reported in studies comparing cyclists of varying experience [53,57], comparing effects of different bicycle configurations [123], and in injured cyclists [260]. Concern has been raised when using EMG over multiple days due to problems in repositioning electrodes [50]. However, a study [272] that normalized electromyographic signals using a maximal power output cycling trial reported no significant differences in 105

106 muscle activation when comparing sub-maximal workload levels using different between-electrodes distances. McLeod and Blackburn [265] conducted a qualitative analysis of muscle activation during cycling and suggested intervention based on muscle recruitment patterns (e.g. increasing activation of flexor group on the recovery phase) for rehabilitation during a single cycling bout. Experimental studies are yet to be conducted looking at the effectiveness of feedback from muscle activation while cycling motion. A model for saddle height optimization has been proposed based on lowering the activation of vastus lateralis, vastus medialis, biceps femoris and gastrocnemius lateralis [273]. However, the case study approach and limitations in methods and analysis preclude any conclusions at this stage. A better definition of workload and cadence, the use of more comprehensive muscle activation amplitude analysis, assessment of more cyclists and the use of a reference condition to normalize the amplitude of muscle activation may improve the reliability of this technique for optimizing bicycle configuration. The use of electromyography for biofeedback for patients after stroke [274] and functional electrical stimulation for spinal cord injured patients [275] may be useful as a methodology for modifying activation patterns while cycling. No study was found using biofeedback or functional electrical stimulation to change muscle activation patterns in cyclists. Simulation models of musculoskeletal mechanics incorporating muscle activation characteristics have shown ~33% lower patellofemoral compressive force during forward compared to reverse pedaling, which has been used in rehabilitation procedures and has the potential to prevent soft tissue overuse injuries [15]. Further research using simulation models is needed to assess changes in bicycle configuration and its effect on joint forces.

8.6 Future research The biomechanical approach for preventing overuse injuries in cyclists is still predominately based on the optimization of body position on the bicycle. For this approach, static analysis of the cyclist on the bicycle is the most used procedure but this neglects variance in force and kinematics of the joints. The references for 106

107 “optimum” or “appropriate” configuration of bicycle components have mostly been based on empirical knowledge rather than experimental results. To date, very few studies have shown that increased variability of force [79] and joint kinematics [179] leads to more injuries in cyclists. Pedal force and joint kinetics analyses are still limited to biomechanics laboratories because of the complexity and high cost of purchasing these systems. The use of electromyography based biofeedback has not been reported for cycling, though successful results for other injured populations (e.g. stroke patients) suggests this could be a useful avenue for further research. Sports equipment has been optimized based on the knowledge of muscle physiology characteristics [276], but few studies have used mathematical models to optimize the configuration of bicycle components [186]. Unknown reliability of biomechanics techniques has compromised bicycle configuration effectiveness as a clinical tool to prevent injuries in cycling. The effectiveness of various strategies and biomechanical techniques to prevent injuries in cycling therefore need to be tested by prospective controlled trial studies using appropriate epidemiological analysis to provide risk ratios for various bicycle configurations, body positions and types of overuse injuries.

8.7 Conclusions and practical applications Optimizing body position on the bicycle theoretically has the potential to help prevent overuse injury in cycling, however biomechanical studies to date have not shown clear evidence that they can reduce overuse injury via optimization of bicycle configuration. If attempting to optimize bicycle configuration and cyclists’ body position to prevent overuse injury then it is suggested that dynamic assessment of joint kinetics and/or kinematics is undertaken. Studies involving bicycle set-up using a variety of the biomechanics methodologies available plus prospective injury and training data collection are required.

107

108 CHAPTER 9: PEDALLING TECHNIQUE CHANGES WITH FORCE FEEDBACK TRAINING IN COMPETITIVE CYCLISTS: PRELIMINARY STUDY

Rodrigo R. Bini, Patria A. Hume and James L. Croft

9.1 Introduction Natural lower limb mechanics in cycling are characterized by coordinated flexionextension of the hip, knee and ankle joints resulting in propulsive torque when the pedal is at the propulsive phase of crank revolution (from 12 o’clock to 6 o’clock) and resistive torque during the recovery phase of crank revolution (from 6 o’clock to 12 o’clock) [60]. Pedal force effectiveness has been computed as the percentage of the total force applied to the pedal that results in propulsive torque [277]. To maximize propulsive torque, cyclists are instructed to apply force in different directions during crank revolution. Figure 1.9 shows an example of “ideal” force application on the pedal to optimize pedal force effectiveness.

Fig 1.9. Illustration of “ideal” normal and anteriorposterior force components application on the pedal to optimize force effectiveness.

108

109

The success of pedal force feedback has commonly been evaluated using a reduction in resistive force at the 3rd and 4th quarters of pedal revolution [131] and the difference between an “ideal” force and the measured force applied on the pedal [209]. Training can improve pedal force effectiveness and change the “natural” lower limb mechanics [209,278,279,210,131] by reducing the resistive crank torque and the total force applied to the pedal at the same workload. Greater activation of knee joint flexors and ankle dorsiflexors have been observed when pedal force effectiveness is improved [133]. Changes in muscle activation due to greater pedal force effectiveness have also been found to affect efficiency [133,189] and possibly performance in cycling. Efficiency has been related to performance in endurance cycling [280] and theoretically reducing resistive force on the pedal may lead to higher efficiency. However, improving pedal force effectiveness has resulted in higher [195,197] and lower efficiency in cycling [133,189]. Studies that reported lower efficiency when cyclists aimed to improve force effectiveness during pedalling suggested that greater activation of hamstrings [133] may result in additional energy expenditure compared to a “natural” pedalling action. Therefore, it is uncertain if cyclists should focus on improving pedal force effectiveness to reduce resistive force or if they should focus on producing high pushing force on the pedal when the crank is close to the 3 o’clock position due to the greatest moment arm. The latter would require greater force production from quadriceps muscle groups, which may be more efficient in trained cyclists [28]. Studies using pedal force feedback to improve pedal force effectiveness have been conducted using low exercise intensity (up to 80% of maximal oxygen uptake) and pedaling cadence (80% of maximal oxygen uptake) and cadence (>80 rpm) using visual feedback of pedal force effectiveness. 109

110 To ascertain if using pedal force feedback would enhance force effectiveness and also result in better performance in cycling, we compared two types of pedal force feedback during training on 4-km cycling time trial performance. Two types of feedback were used: overall pedal force effectiveness during pedal revolution or peak normal force applied to the pedal. Our preliminary target was to assess the feasibility of using pedal force feedback during training at high intensity exercise (i.e. 4-km time trial) to improve pedal force effectiveness and performance in cycling.

9.2 Methods 9.2.1 Participants and allocation to pedal force feedback groups Four male and two female athletes with competitive experience in cycling and in triathlon, ranked as “club riders” according to Ansley and Cangley [281] were invited to participate in the study (mean ±SD: 28 ±7 years, 176 ±15 cm, 64.8 ±15 kg, 354 ±94 W maximal aerobic power output, and 60 ±6 ml.kg-1.min-1 of VO2Max) and signed an informed consent form in agreement with the research ethics committee of the institution where the study was conducted. Three cyclists were assigned to the “force effectiveness” group (FEG) that received visual feedback during training sessions of average pedal force effectiveness and a diagram indicating pedal force application in each quadrant. Three cyclists were assigned to the “peak force” group (PFG) (n = 3) that only received information on the average peak normal pedal force. Cyclists in each group were matched by similar performance (e.g. maximal aerobic power output and VO2Max) and gender (e.g. one female to FEG and one female to PFG). The cyclists knew neither the nature of the other group’s feedback nor the aims of the study. They were informed only that time trial training was going to be provided with use of pedal force feedback to help improve their performance and were asked to keep their regular cycling training. The choice for providing feedback of peak normal force applied on the pedal was to offer feedback that would not lead to substantial changes in pedal force effectiveness. The reason is that peak normal force occurs around 90° of crank revolution and would therefore be orientated perpendicular to the crank, influencing both effective and total forces proportionally [28].

110

111

9.2.2 Pre-training protocol At the start of the first session body mass and height were measured according to ISAK protocols [282]. Cyclists completed the Waterloo inventory to allow the determination of lower limb dominance [283]. Cyclists’ bicycle saddle height and horizontal position were measured to set up the stationary cycle ergometer (Velotron, Racemate, Inc). Cyclists performed an incremental cycling test to exhaustion with workload starting at 100 W for the first three minutes and increasing in steps of 25 W each minute [284]. During the study, pedaling cadence was visually controlled by the cyclists at 90 ±2 rpm using the Velotron Coaching software 2008 (Velotron, Racemate, Inc). Gas exchanges were continuously sampled from a mixing chamber where samples were drawn into the oxygen and carbon dioxide analyzers for continuous measurement using a metabolic cart (TrueOne 2400, Parvo Medics, Salt Lake City, UT, USA). Prior to the test, the oxygen and carbon dioxide analyzers were calibrated according to the manufacturer’s recommendations. Maximal aerobic power output and maximal oxygen uptake were defined as the highest power output measured during the test and as the highest oxygen uptake value computed over 15 s of data, respectively. After 10 minutes of rest, cyclists were familiarized with the 4km time trial where they self-selected gear-ratio and pedaling cadence.

9.2.3 Training sessions In the second session, the athletes performed two bouts of a 4-km time trial separated by 10 minutes of active rest on the bicycle. The 4-km time trial was used to elicit a maximal aerobic exercise effort under racing condition (i.e. change in pedaling cadence and gear ratio). During all time trials in the second session and the following seven sessions, normal and anterior-posterior forces were measured using a pair of strain gauge instrumented pedals [138], with pedal-to-crank angle measured using angular potentiometers. A reed switch attached to the bicycle frame detected the time that the crank passed the bicycle frame and was used to compute pedaling cadence. All data were acquired at 600 Hz by an analogue to digital converter (PCI-MIO-16XE-50, National Instruments, USA) using a custom Matlab (Mathworks Inc, MA) data acquisition script. Data were acquired for 10 s every 500 111

112 m of the time trial during the first two sessions. Feedback was provided for 5 s for the right pedal and then 5 s for the left pedal every 500 m in the first training session. Bilateral pedal force effectiveness, which is based on the ratio between the angular impulse of the effective (tangential) force on the crank and the resultant (total) force applied to the pedal surface, was computed along with bilateral average peak normal pedal force for every bout of feedback provided for five pedal revolutions. During training sessions, feedback of forces on the right and left pedals was provided (see Figure 2.9) at distance points showed in Table 1.9 with a delay up to 15 s due to acquisition and processing of data using a custom made Matlab script. Frequency of feedback was reduced throughout the training period to minimize the participants dependence on the feedback and to improve the kinaesthetic sensory pathways responsible for fine-tuning pedal force application [278].

Table 1.9. Occurrence of force feedback during the 4-km time trial for the eight training sessions. Distance 1st and 2nd

3rd and 4th

5th and 6th

7th and 8th

points

sessions

sessions

sessions

sessions

500 m

Yes

Yes

No

No

1.0 km

Yes

No

Yes

No

1.5 km

Yes

Yes

Yes

No

2.0 km

Yes

Yes

No

Yes

2.5 km

Yes

Yes

Yes

No

3.0 km

Yes

No

No

Yes

3.5 km

Yes

Yes

Yes

No

3.8 km

Yes

Yes

No

No

112

113 Fig 1.9. Illustration of “ideal” normal and anteriorposterior force components application on the pedal to optimize force effectiveness.

 

Fig 2.9.

Example of the feedback screen shown to cyclists of the normal and

anterior-posterior forces applied on the right pedal, effectiveness of pedal forces (%) and peak normal force (N). White arrows were presented when pedal force application resulted in propulsive torque, and black arrows were shown when pedal force resulted in resistive torque on the crank. Force diagram and force effectiveness value were shown only to the FEG and the peak force value was shown only to the PFG.

113

114

9.3.4 Data analyses Off-line force data (peak normal force, resultant force and effectiveness of pedal forces) of both pedals were computed for five revolutions. Pedaling cadence was computed for each revolution from the time difference between each pulse of the reed switch signal. Force and pedaling cadence from each of the five revolutions were averaged for each of the eight distance points during the 4-km time trial. Force and pedaling cadence data from each revolution were then averaged for the first and second 4-km time trials of each session, resulting in eight values of force and pedaling cadence for each revolution. Average power output and 4-km time recorded by the Velotron Coaching software 2008 (Velotron, Racemate, Inc) for both of the 4km time trials within each session were averaged.

9.3.5 Statistical analyses Means and standard deviations were computed for each group and normalized by the results of the first training session for graphical presentation. To compare group responses to training, pedal force data were averaged for the first two training sessions (1st and 2nd) and for the last two training sessions (7th and 8th) for the right and left pedals and analyzed using effect sizes. Cohen‘s effect sizes (ES) were computed for the analysis of magnitude of the differences between means and were rated as trivial (1.0) [285]. We chose large effect sizes for discussion of results to ascertain non-overlap between mean scores greater than 55% [286].

9.4 Results Lower limb symmetry assessed by the Waterloo inventory indicated only one cyclist from the FEG as left leg dominant (>60% of left leg preference in the inventory) with all other cyclists stating right leg dominance. There was a trend for increases in right and left peak normal force, right and left resultant force and right pedal force effectiveness over the training sessions (see Figure 3.9). 114

115

Fig 3.9. Means and standard deviations for force effectiveness group (FEG) and peak force group (PFG) for right and left peak normal force (A), resultant force (B) and effectiveness of pedal force (C) normalised by the results of the first training session. N = 3 for each group.

115

116 There were large increases in right peak normal force for the FEG and PFG, followed by large changes in right resultant force for PFG and increases in right force effectiveness for both groups (see Table 2.9).

116

117 Table 2.9. Means and standard deviations and session change scores as percentages, and effect sizes and for peak normal force (NF), resultant force (RF) and force effectiveness (FE) for cyclists of FEG and PRG. Peak normal force and resultant force presented in Newtons and force effectiveness presented as % of linear impulse of resultant force. Abbreviations used are for forward (Fwd) and backward (back) positions on the saddle and effect sizes of trivial (T), small (S), moderate (M) and large (L). Large differences are highlighted in bold italics. Force effectiveness group (FEG)

Peak force group (PFG)

FEG vs. PFG

1&2

7&8

Change (%; ES)

1&2

7&8

Change (%; ES)

Differences (%; ES) between

sessions

sessions

between 1&2 and

sessions

sessions

between 1&2 and 7&8

1&2 and 7&8 session changes

sessions

for FEG versus PFG

7&8 sessions NF

338

352

4%;

349

367

5%;

1%;

Right

±77

±58

1.2, L

±123

±129

1.8, L

0.4, S

Left

308

320

4%;

320

324

1%;

3%;

±62

±51

0.8, M

±106

±56

0.3, S

0.6, M

RF

152

156

2%;

159

161

2%;