Search

It is often observed that there is a high individual variability in the response to exercise-induced muscle damage (EIMD), even when tested in a homogeneous cohort accounting for age, sex, ethnicity and physical activity. The response to EIMD is very complex as several tissues, including skeletal muscle fibres, the extra-cellular matrix (ECM), and tendon, play a potential role in the damage response. Therefore, the overall aim of this PhD thesis was to investigate the physiological and genetic factors underpinning the response to muscle damaging exercise. For that, the following objectives were (i) to comprehensively assess the physiological mechanisms and recovery pattern of neuromuscular fatigue of the hamstring muscle group following an intermittent sprint (IS) intervention; (ii) to investigate inter-individual differences in skeletal muscle repair/recovery after an artificial wounding (scratch) assay using of primary human skeletal muscle cells in vitro; (iii) to ascertain whether multiple genetic variations, which are linked to varying tissues, forming a polygenic profile could distinguish between high and low responders following muscle damage in vivo and in vitro; and (iv) to assess whether a genetic profile is linked with the response to both EIMD and chronic resistance exercise. The methodological and analytical approaches utilised in this thesis identified a number of important, novel and impactful findings. Following IS, the impaired hamstring muscle function and delayed recovery is probably caused primarily by damage to the contractile tissue, and participants with a greater force generating capacity (larger physiological cross-sectional area) of the biceps femoris long head were less susceptible to hamstring strength loss immediately after IS, providing evidence that the structure of the muscle protects it against peripheral fatigue/damage. The in vitro study showed that skeletal muscles with an increased number of stem cells of the connective tissue (fibroblasts) might have a better capacity to reorganise the complex ECM, which results in a faster muscle strength recovery after muscle damaging exercise. However, a larger number of active muscle stem cells (myoblasts) seems to be important for the latter stage of muscle regeneration. Individuals possessing a non-preferential genetic profile demonstrated increased rate of muscle damage biomarkers than individuals with a preferential genetic profile. Lastly, we calculated a second polygenic profile which was linked with both the EIMD and the chronic resistance exercise response. These polygenic profiles may be used to anticipate an individual's response/adaptation to EIMD and to chronic resistance exercise, thus enabling resistance exercise to be prescribed on a personalised level to improve muscle health and function.

The area of recovery following soccer match play has become an increasingly popular area of focus in recent years. There is limited research around the activities and monitoring processes undertaken within elite soccer clubs. To the authors knowledge, there is only one previous research article that has attempted to understand the activities undertaken during this period within elite soccer (Nedelec et al., 2013). The aim of this thesis was to understand the activities undertaken by elite soccer players during the recovery period following match play and to investigate the suitability of the various approaches to monitor physiological status during the period. The aim of study 1 (Chapter 3) was to provide an overview of the activities undertaken by soccer players (n=53) during periods of the training week dedicated to recovery following match play. The findings of this study show that massage therapy accounted for the longest amount of time on matchday +1 (MD+1) (43min) and pitch based active recovery on matchday +2 (MD+2) (44min). Pool based active recovery was the most consistently completed method on MD+1 (64%) and pitch based active recovery on MD+2 (80%). These data may not only provide useful information on the recovery processes included in the training plans of elite soccer players but also give potential insights into the strategies that may be used to effectively monitor this recovery between games. The aim of study 2 (Chapter 4) was to determine if heart rate variability (HRV) obtained through a purpose built smartphone application measured on the first day (MD+1) and second day (MD+2) following match play can detect changes in the physical loads undertaken by elite soccer players (n=21). No significant relationships between any of the physical loading variables assessed in this study and smartphone derived HRV MD+1 or HRV MD+2 (natural logarithm of the square root of the mean squared differences of successive R-R intervals [Ln rMSSD]) were found. This suggests that measurements of HRV through the 'ithlete' method may not be suitable for use in monitoring physiological status following soccer match play. The aim of study 3 (Chapter 5) was to develop a reliable exercise protocol that could be used for calculation of heart rate recovery (HRR) and percentage of maximum heart rate reached (%MAXHRr) with elite soccer players (n=15) while adhering to the specific characteristics of an exercise test that would facilitate completion with an elite soccer team. The data presented in this study suggests that the most suitable exercise protocol with regards to the reliability of the heart rate (HR) response and adherence to the physical loads used during the recovery period with elite soccer players was 6x80m straight-line runs at 5m/s with an exercise: rest ratio of 16:14s. This protocol presented coefficient of variation (CV), standard error of measurement (SEM) and physical loading data that were superior to data associated with other similarly styled protocols considered throughout this study. The aim of study 4 (Chapter 6) was to assess the face validity of the HR response to the developed standardised exercise protocol (SEP) under controlled conditions following a high physical load simulated soccer training session (HSSTS) and a low physical load simulated soccer training session (LSSTS) (n=10). The findings of this study suggest that the HR response to the SEP was able to detect manipulated changes in physiological status as a result of a high physical load and a low physical load simulated soccer training session. The %MAXHR during the SEP was found to be significantly higher following the HSSTS (92%) in comparison to the LSSTS (90%). HRR was found to be lower following the HSSTS (17%) when compared to the LSSTS (22%). The aim of study 5 (chapter 7) was to assess the effectiveness of using HRR and %MAXHRr measured through the SEP on the first day (MD+1), second day (MD+2), and third day (MD+3) following match play to detect changes in physical loads experienced by players (n=16) during the preceding soccer game. The findings of this study show the SEP to have limited capabilities in the detection of changes in HRR and %MAXHR on MD+1, MD+2, or MD+3. The data presented may suggest that the SEP is most suited for implementation on MD+2 or MD+3. It seems that HRR is the more sensitive of the markers investigated. In summary, the studies undertaken as part of this thesis provided an insight into the structure of the activities employed during the recovery period following elite soccer match play and the potential opportunities to monitor physiological status during this process. Initially this involved providing an understanding of the types of activities that were completed by players during this period. This was followed by an assessment of one of the current methods of monitoring physiological status (HRV analysis in chapter 4) during this time frame. This provided the basis for a novel method to be developed. The latter studies then focussed on developing a reliable and valid method of monitoring that could be easily implemented into the schedules during the recovery period within an elite soccer club which were outlined at the outset.

Football is an intermittent (Bangsbo, 1994; Shephard, 1999; Reilly, 2005) and non-linear team sport. Coaches and practitioners must, therefore, prepare players to complete relatively large volumes of multidirectional activity that are short in duration due to frequent changes. One key aspect of ensuring the appropriateness of any training strategy is to develop and implement an effective training load monitoring strategy. Global Positioning Systems (GPS) incorporated into Micro-Electro-Mechanical Systems (MEMS) devices appear to be the technology that has been most widely adopted to determine activity (Akenhead & Nassis, 2016). The commonly used GPS technologies are regularly accompanied with tri-axial accelerometer within the MEMS hardware. It may be hypothesised that the progression of multiplanar MEMS accelerometer technologies may allow the frequent change of directions and velocities to be more accurately measured and, therefore, evaluate elite football training more effectively. There is, however, currently a lack of applied research, which has attempted to establish the utility of MEMS accelerometers to appropriately capture the movement requirements associated with elite football training. The overall aim of the research contained within the present thesis was, therefore, to investigate the relevance of indicators of external load for the evaluation of the movement requirements in elite football.

Limited research data is available outlining the resistance training characteristics of elite football players. The aim of the first study (Chapter 3) was to compare approaches to calculating resistance training volume during 4 weeks of pre-season training in 23 English Premier League footballers. Volume was calculated using four different methods of quantification; Repetition volume (RV), Set Volume (SV), Volume Load (VL) and Maximum Dynamic Strength Volume Load (MDSVL). Overall there was a significant difference between resistance training volumes calculated by the different methods used to monitor resistance training load (P < 0.001). More specifically, significant differences were observed between RV and SV methods (P < 0.001), RV and MDSVL (P = 0.001), SV and VL (P = 0.010), SV and MDSVL (P = 0.033) and VL and MDSVL (P = 0.002). Only RV and VL methods were similar in the information they provided on training load (P = 0.411). While the lack of a gold standard measure of volume makes it is unclear which, if any, method represents the most accurate measure of volume the discrepancies between methodological approaches highlight that these different approaches are not directly transferable as strategies to monitor resistance training. The understanding of the differences between each method may therefore enable appropriate, situation specific, approaches to be designed and implemented for both practical and research purposes. The aim of the second study (Chapter 4) was to analyse the resistance training loads completed by an elite professional football team across a competitive season. Resistance training data was collected from 31 elite football players competing in the English Premier League over a 46 week period in the 2012-2013 season. A total of 1685 individual training observations were collected during the pre-season and in-season competition phases, with a median of 42 training sessions per player (range = 9 – 124). Training load data was separated into 7 blocks of 6 weeks for analysis. These periods included pre-season (6 weeks duration) and in-season (40 weeks duration) phases. Set volume was selected as a measure of total volume. Data was analysed using 3 separate linear mixed modelling analysis using the statistical software package R (Version 3.0.1). Weekly resistance training frequency (mean±SD) ranged from 1±1 to 2±1 sessions per week during the pre and in season phases. Significant differences in session frequency were seen between weeks 1-6 and weeks 7-12 (pre-season) (P ˂ 0.05), weeks 7-12 and weeks 13-18 (P ˂ 0.05), and weeks 7-12 and weeks 37-42 (P ˂ 0.05). Mean weekly training volume ranged from 18±16 to 30±24 sets.wk-1. The total training volume demonstrates a clear minimum during weeks 7-12. Significant differences in total training volume were also observed between weeks 1-6 and weeks 7-12 (pre-season) (P ˂ 0.01), weeks 7-12 and weeks 13-18 (P ˂ 0.05), and weeks 7-12 and weeks 19-24 (P ˂ 0.05). There was no significant difference in training intensity between weeks 1-6 (pre-season) and weeks 7-12. Training intensity during weeks 1-6 however was significantly lower than during weeks 13-18 (P ˂ 0.05), 19-24 (P ˂ 0.01), 25-30 (P ˂ 0.01), 31-36 (P ˂ 0.05), and 37-42 (P ˂ 0.01). Training intensity during weeks 7-12 was also significantly lower than during weeks 13-18 (P ˂ 0.01), 19-24 (P ˂ 0.05), 25-30 (P ˂ 0.05), 31-36 (P ˂ 0.05), and 37-42 (P ˂ 0.001). The findings would suggest that resistance training loading is limited during different periods of the season. This is predominantly as a consequence of low training frequency, potentially due to a high prevalence of competitive fixtures. The aim of the third study (Chapter 5) was to attempt to quantify the impact of resistance training completed by players, through evaluating the change in the lower body power outputs of an elite professional football team across a competitive season. Resistance training data was collected from 22 elite football players competing in the English Premier League over a 38 week period. A total of 246 individual power output observations were collected during the in-season competition phase. Power output of the lower body was assessed using a pneumatic resistance leg press machine with software and digital display (Keiser Sports Health Equipment Inc., Fresno, Ca). Data was analysed by means of linear mixed modelling analysis using the statistical software package R (Version 3.0.1). Power outputs ranged from 2200W to 4078W with a mean value of 3022±374W. Linear mixed effects show a significant effect of week on power output across the season (coefficient= 7.76W, p=0.0132). Specifically, when accounting for within player effects, power output increased 7.76W per week during the season. Individual weekly power coefficients ranged from +39.9W to -18.13W per week, thus indicating that the trend for increased power output across the season is not uniform for all the players. These data may suggest that lower body power performance is maintained or minimally enhanced over the course of a full competitive season in elite football players. Combined with the training load data previously examined in this thesis it can be concluded that whilst one resistance training session per week may be sufficient to avoid in season de-training or minimally improve power performance in elite football players, a frequency of two sessions per week may be necessary to obtain significant performance enhancements. In our fourth study (Chapter 6) we provide two case studies that outline and evaluate a structured approach to increasing resistance training loading with the primary goal of developing strength and power during the competitive season in elite football players. The purpose of our initial case was to examine a resistance training programme to enhance strength and power performance, alongside body composition during a period of rehabilitation from injury. The study intervention commenced following two weeks of recovery following the “Laterjet” surgical procedure. Initial assessments were performed for body composition via dual-energy x-ray absorptiometry (DXA) (QDR Series Discovery A, Hologic Inc., Bedford, MA) and lower body power output via using a pneumatic resistance leg press machine with software and digital display (Keiser Sports Health Equipment Inc., Fresno, Ca). Assessments were repeated 8 weeks post-surgery, i.e. following 6 weeks of resistance training. The six-week intervention consisted of three strength training sessions per week for the initial 3 weeks, followed by 2 sessions per week for the subsequent 3 weeks. Training volume (number of sets) equalled a total of 20 sets total per session. Total increase in body mass over the intervention period equated to 5.4kg, of which 4.2 kg increase in lean mass and a 1.3 kg increase in fat mass. Peak power output increased by 21%. Power to weight ratio also increased by 4.4 %. These data illustrate that it is possible to increase physical performance when rapid short-term increase in resistance training load is completed. The purpose of our second case was to examine a resistance training programme to enhance both strength and power performance parameters during a full competitive season. The player plays as a goalkeeper, regularly playing for his club 1st team. Prior to the onset of this case study this player did not present with any current injuries. This season long intervention consisted of two phases of training. Phase 1 was 16 weeks in duration and represented the beginning to the mid-point of the season. During this phase the goal was to gradually and safely increase resistance training loading. Phase 2 was 20 weeks in duration and represented the mid-point to the end of the season. This phase represented a period of consistent high loading following the initial systematic increase in these variables. Assessment data was collected at the beginning, mid-point and end of the 2013-14 season. The player was first assessed for body composition via DXA (QDR Series Discovery A, Hologic Inc., Bedford, MA). Secondly, lower body power output was assessed using a pneumatic resistance leg press machine with software and digital display (Keiser Sports Health Equipment Inc., Fresno, Ca). Finally, the player’s upper body strength was assessed via 6 repetition maximum assessments of the dumbell bench press and prone row. The player completed a mean weekly volume of 41±24 sets per week and a mean frequency of 2±1 sessions per week for the initial phase of the study. The player completed a greater mean weekly volume in the later phase of the season compared to the initial training period (65±28 set per week vs. 41±24 sets per week in the initial phase of the season). A greater mean session frequency was also associated with the second training phase (3±1 vs. 2±1 session per week). There was a total decrease in body mass over the initial intervention period of 4kg, of which 2.7kg decrease in fat mass and a further 0.9 kg decrease in lean mass. Over the second phase of the intervention there was a total increase in body mass of 1.2kg, of which 2.4kg increase in lean mass and 1.2kg decrease in fat mass. During the initial phase of training peak power output increased by 25%, whilst power to weight ratio increased by 30%. During the later phase peak power output increased by a further 9% whilst the power to weight ratio increased by a further 10%. Upper body pressing (Dumbell Bench press) and upper body pulling (Dumbell Prone pull) strength was also increased by 14% and 21% respectively during the initial phase and a further 19% and 24% respectively during the later phase of the season.

The physical demands of soccer match play have been extensively studied. Muscle glycogen is the major energy source required to meet these demands and strategies to maximise this provide clear performance benefits to match play. Such information has allowed sports nutritionists to develop specific guidelines to optimise physical performance and recovery. However, the physical demands of soccer training have only recently started to be examined. For this reason, Study 1 quantified training load in English Premier League soccer players (n=12) during one, two and three game weekly micro-cycles of the 2013-2014 season. Study 1 identified soccer training being significantly less than match play and identified that soccer training displayed evidence of training periodisation. Having identified typical training load during the weekly micro-cycle, it was recognised that soccer match play comprises a large portion of the weekly physical load. Accordingly, Study 2 quantified differences in season long physical load (inclusive of both training and match play) between players who were classified as starters (n=8, started ≥60% of games), fringe players (n=7, started 30-60% of games) and non-starters (n=4, started < 30% of games). Study 2 identified that unlike total seasonal volume of training (i.e. total distance and duration), seasonal high-intensity loading patterns are dependent on players’ match starting status thereby having potential implications for training programme design and prescription of player-specific nutritional guidelines. Additionally, daily energy expenditures (EE) and energy intakes (EI) of elite players are also not currently known. Therefore, studies 3, 4 and 5 quantified EE and EI in English Premier League soccer players consisting of outfield positions (n=6), a professional GK (n=1) and a player undergoing a rehabilitation period from an ACL reconstruction (n=1), respectively. Studies 3 and 4 were conducted over a 7-day period of the 2015-2016 season, consisting of two match days (MD) and five training days (TD). Study 5 consisted of six training days and one day off. Studies 3 and 4 identified CHO periodisation strategies employed by English Premier League Players such that CHO intake was greater on MD than TD. Additionally, players readily achieve current guidelines for daily protein and fat intakes, although energy and macronutrient intakes are skewed on TD. Study 4 also identified that the GK exceeded average daily EE with EI although he failed to meet current recommendations for meals on MD. In study 5 the player was operating in an energy deficit and he was able to decrease his total body mass in the initial 1-6 weeks post injury, which was attributable to largely fat loss. In summary, the work undertaken in this thesis has quantified the typical physical loading patterns of professional soccer players according to fixture schedule, starting status and in special populations. Additionally, the quantification of EI and EE (using DLW) also provides the first report of EE in elite soccer players from the English Premier League. When taken together, these data therefore provide a theoretical framework for soccer-specific nutritional guidelines especially in relation to the concept of nutritional (specifically, carbohydrate) periodisation. Further studies are now required to quantify the specific energy and CHO cost of habitual training sessions completed by elite soccer players as well as examining the manipulation of CHO availability on soccer-specific training adaptations.

American football is one of the most popular sports in the United States. However, unlike other football codes, little is known about its physical demands. Aside from a limited amount of research conducted on college players during training and matches, no research exists on players at the elite level, in the National Football League (NFL). Therefore, the primary aim of this thesis was to evaluate the physical demands of training in the NFL. This aim was achieved by establishing a systematic approach to training evaluation using three main phases of study: (1) Evaluation of monitoring strategies in American football; (2) Description of American football training demands with an emphasis on periodization; and, (3) Examination of the consequences of training with an emphasis on injury risk. The first study of this thesis (Chapter 3) showed that three commercially available inertial sensors were able to differentiate between fundamental American football actions (e.g., sprinting, change of direction, and collisions) during movement tasks in a controlled setting and may be useful for quantifying the physical demands of training. During training sessions, Session Rating of Perceived Exertion exhibited a variety of individual responses making sRPE challenging to use when exclusively evaluating the physical demands of training (Chapter 4). Therefore, more objective measures (e.g., GPS and inertial sensors) were evaluated during training (Chapter 5) and indicate that commonly used measures of distance and velocity may not adequately describe the physical demands for some position groups. As such, inertial sensors offer more flexibility to classify a broad range of activities within the sport. A number of inertial sensor metrics are available to the practitioner in commercially used systems. Chapter 6 utilized a principal components analysis to reduce eleven variables to 3 principal components, explaining 79% of the variance within the data. These findings suggest that a small number of variables (e.g., Player Load and IMA) may be adequate when describing the training demands of the sport. Given the reduction in measures to report, Chapter 7 used Player Load and IMA to describe the periodization strategies across a season and within the training week employed by the coaches of one NFL team. Training load was observed to decreases across the season with no clear periodization structure. Conversely, within the weekly micro-cycle, coaches appear to employ some pattern of periodization whereby training load is seen to systematically decrease as the game nears. The final phase of this thesis (Chapter 8) investigated the consequences of American football training by exploring the relationship between training load measures (Player Load, IMA, and Impacts) and non-contact soft tissue injury. Several logistic regression models were compared using Bayesian Information Criterion (BIC). The best model suggested that sessions with greater volume (PLTotal) and intensity (ImpactsHigh) were associated with non-contact soft tissue injury in American football players and may have implications for practitioners when designing training programs within the sport. Collectively, this thesis has the potential to not only offer practitioners within American football a way forward in terms of evaluating training demands but also may be influential to the broader scope of sports science given some of the novel statistical approaches taken to understanding training load monitoring.

Elite youth soccer player’s performance depends on a multitude of factors (Stølen et al. 2005). Muscular force production capability is of likely importance due to the frequency of powerful actions that occur within match play (Faude et al. 2012). Little is known though of the training undertaken at elite youth soccer academies to increase muscular force production capabilities particularly across maturation groups. There is also little evidence around the nature, or genetic influence on this variable in these populations. Greater knowledge on these areas would be beneficial to aid an understanding of performance and to plan practical interventions. Eight weeks of training for players in the under 9 (U9) through to under 21 (U21) age groups (Chapter 3) were recorded to investigate the duration of each training type completed. The total training duration increased from U9 to U14 before reducing at U15 and then remaining unchanged through to, and including, the U21 age group. Soccer training accounted for 97 ± 4 % of session time in the U9 to U14 groups and 74 ± 3 % in the U15 to U21 groups. The remainder of training was made up of work that was not soccer based. The data in this case study suggest that training time is focussed on the technical/tactical development throughout the academy, particularly in the younger age groups. Study 2, part A (Chapter 4) provided baseline isometric maximal voluntary force (MVF) data for players and an maturation-matched non-elite control group. MVF was slightly higher in the elite compared to control cohorts during an isometric mid-thigh pull (IMTP, 118.29 ± 13.47 N compared to 109.69 ± 17.00 N). Such data may indicate that ability to produce force, specifically isometric MVF, may not be a crucial determinant of performance in elite youth soccer based on this sample. The purpose of study 2, part B (Chapter 4) was to establish the effect of 8 weeks typical training on elite youth soccer players’ IMTP MVF. This was also compared to a non-elite control cohort matched for maturation status and not undertaking training. Isometric MVF did not change in either group over the 8 week period (P = 0.386). These data suggest that this elite youth soccer training simply maintained current strength levels and was not sufficient to elicit increases in isometric MVF. Study 3 (Chapter 5) examined variations in four separate genes, all identified as potentially having an influence on muscular force production capabilities: PPARA rs4253778, NOS3 rs2070744, COLIA1 rs2249492 and VDR rs2228570. Allele and genotype frequency was determined along with the influence of those single nucleotide polymorphisms (SNPs) on isometric MVF. Only NOS3 was different in genotype distribution between cohorts with TT genotype showing 45.0 % and 30.4 % frequency in elite and control cohorts respectively (P < 0.001). Furthermore, no difference was seen between cohorts for isometric MVF data when comparing influence of any genotypes of any gene. These data provide novel information around genotype frequency in this population and would suggest that the gene variations examined here might not play a role in force production in these populations. Overall, these findings suggest that muscular force production may be important for elite youth soccer performance, but that current training practices to improve this characteristic should be revised. This seems especially relevant given the lack of genetic association with force production in the gene variations examined here, i.e. elite youth soccer players do not appear to be selected based on a genetic predisposition for greater isometric force production capabilities.

Contemporary research into training load in team sports primarily focusses on the physiological load demands, whereas the biomechanical load still remains largely unexplored. While the former refers to the work-energy relationship when the players move around the pitch, the latter refers to the external forces the players are exposed to from their movements around the pitch. Monitoring of the biomechanical load helps practitioners estimate the stresses on an athlete’s musculoskeletal structures as a consequence of the external forces acting on their body. Monitoring of the biomechanical load is currently restricted to laboratory settings, but the recent introduction of GPS devices with integrated accelerometers in team sports may enable practitioners to monitor whole-body biomechanical load during training sessions and match-play. The aim of this thesis was therefore to explore if body-worn accelerometry can be used for whole-body biomechanical load monitoring in team sports. The first study of this thesis showed that although a linear relationship exists between body-worn accelerometry (e.g. from GPS integrated accelerometers) and whole-body accelerations, the linear relationship based on Newton’s second law of motion is weak regardless of accelerometer location (trunk, pelvis or tibia). Body-worn accelerometry only measures the acceleration of the segment it is attached to and is therefore inadequate to measure the complex multi-segment dynamics of the whole body during team sports movements. The second study of this thesis did however offer a potential solution to that problem, and it was demonstrated that the complex multi-segment dynamics of the body and the associated ground reaction forces (GRF), a surrogate for whole-body biomechanical load, can be estimated with a mass-spring-damper model (MSD-model). Nonetheless, the MSD-model’s accuracy to estimate GRF slightly decreases for sharp changes of direction at high intensities, because the absorption of energy and generation of energy are decoupled. Finally, a novel approach to estimate GRF from the combination of trunk-mounted accelerometry and a MSD-model was introduced in this thesis. This approach showed that trunk accelerometry data has the potential to generate the eight model parameters required to estimate GRF from a MSD-model, though further work is required in particular towards improving the model’s ability to estimate GRF across a wide range of activities. The novel approach introduced in this thesis has the potential to give practitioners in team sports the opportunity to monitor whole-body biomechanical load due to player-ground interaction in the field, a necessity if they wish to predict the consequent musculoskeletal structural adaptations of training sessions and match-play.

Hamstring muscle injuries constitute a major concern in football and a major challenge for physiotherapists working in this sport, being an injury with long absence from playing and training. Although clinical strategies to rehabilitate these injuries and clinical injury risk assessments have been explored over the years, a broader comprehension of how variables regarding running performance may in some form relate to hamstring injury risk has been missing for clinical professionals. Together with this, its incidence has been increasing despite many preventive efforts, which reveals a necessity for developing risk assessment methods to better inform preventive strategies. The key involvement of hamstring muscles during accelerations and decelerations during football running actions justifies research into acceleration related observations. Therefore, the aims of the current programme of research were to develop new laboratorial and load monitoring strategies related to acceleration actions, by exploring biomechanical factors from a physiotherapist perspective. Additionally, implementing assessments and exposing some key limitations of these assessments in professional clubs is also described throughout the experimental studies of this thesis (chapters 3 and 5). For the purposes of this thesis, force development variables were analysed in chapter 3, during maximal accelerations on a non-motorised treadmill, and comparisons between professional players with and without previous injuries were performed. Results from this study revealed no differences between groups during both maximal acceleration and steady state of a maximal sprint effort. A second approach regarding risk analysis and acceleration variables considered the mechanical load based on trunk-mounted accelerometry used in outfield training, as detailed in chapter 5 of the present thesis. In this study mechanical load expressed by PlayerLoadTM, an accelerometer-derived variable aimed to express the rate of change in acceleration, was collected for the training sessions during three weeks previous to a hamstring injury event, in English Premier League clubs, using matched healthy controls. Although the results did not show significant differences between experimental and control group, this exploratory method may constitute a promising method to assess hamstring injury risk. Reliability and validity of the acceleration related variables were addressed first for each of the two experimental studies detailed in chapter 3 and 5. For this purpose, a pilot study on reliability of force collection using a non-motorised treadmill was performed to test the experimental protocol with results showing good overall reliability. For the PlayerLoadTM, a laboratorial study detailed in chapter 4 using a laboratorial overground soccer simulation protocol was adopted and convergent validity with subjects´ anthropometrics together with reliability analysis of four isolated football actions (jogging, side cut, stride and sprint) was performed. Results of this study revealed no association between PlayerLoadTM and the subjects height or body mass and also an overall good reliability for the four actions analysed. In summary, the research presented in this thesis helped better understand the current value and limitations of screening and monitoring acceleration related variables in the context of hamstring (re-)injury prevention in professional football, introducing to the clinical field a different perspective of addressing hamstring behaviour during acceleration actions, and its hypothetical relation with hamstring injury.

Sleep is a recurring state that is considered a critical process in the optimal attainment of daily functions and recovery in athletes. However individuals from elite sports, such as soccer, may be exposed to a number of situations that may impact sleep within the competitive season (such as inconsistent schedules and travel), which may result in sub-optimal sleeping patterns. However, research documenting the sleep of soccer players is at present limited. Therefore it would seem important to investigate how soccer players sleep to further the understanding of how sleep may be impacted. On this basis, the aim of the current thesis was to examine the typical sleeping patterns of youth soccer players and the factors effecting sleep. This was completed through a series of investigations conducted during the competitive youth soccer season. The aim of the first study (Chapter 3) was to validate a commercially available wireless sleep-monitoring device (WS). This was done in an attempt to provide a viable methodology to measure sleep within the habitual environment of soccer players. Nine randomly selected male participants were monitored over 3 nights and comparisons were made between the WS and other established field measures of sleep (Wristwatch actigraphy, sleep diary and Firstbeat bodyguard heart rate system). The relationships between the WS and the other sleep devices, indicated strong to very strong correlations (r > 0.6) and no significant differences between a range of outputs; total sleep time (Actigraphy assumed sleep time [0.97] & Sleep Diary [0.87] p > 0.05), sleep onset latency (Actigraphy [0.69] p > 0.05) and number of awakenings (Sleep Diary [0.69], p > 0.05). There were also low bias and narrow limits of agreement (LOA) within the comparison of mean differences with the WS for assumed sleep time (2 ± 17 min 95% LOA: -30 to 34 min [Actigraphy]), sleep onset latency (7 ± 17 min, 95% LOA -28 to 40 min [Actigraphy]), and number of awakenings (0.05 ± 1, 95% LOA -3 to 3 [Sleep Diary]). These results suggested that the WS is a viable device for the detection of these selected sleep variables. Chapter 4 looked to provide a comparison of sleep measures between a sample of youth soccer players (N=8) and non-athletes (N=8). Both groups were monitored over a period of 6 days within the habitual setting using the WS. The findings showed the soccer player group attained greater amounts of sleep quantity in comparison to the non-athlete group (504 ± 22 vs. 433 ± 46 min [+71 min] total sleep time, ES: 2.0, Large, p < 0.01). This may have been as a result of a later time of final awakening (08:54 ± 00:14 vs. 07:34 ± 00:46 [+77 min], ES: 1.7, Large, p < 0.01). Such an observation suggested that the soccer players were afforded greater time in bed as a result of the imposed soccer schedule. The soccer players also displayed a statistically greater time spent in wake (13(13) vs. 3(5) min [+10 min], PS: 0.86 ES: 1.5, Large, p < 0.05) on average each night. This data suggested that the sleep of the youth soccer players might be less efficient (as a consequence of greater levels of disturbance), despite desirable quantities of sleep being attained than non-athlete controls. Chapter 5 provided a daily comparison of sleep measures conducted over a 14-day assessment period. It is apparent that youth soccer players attained more sleep quantity in the nights preceding the match day (M-2: 488 ± 53 min [ES: 0.91, Moderate; p = 0.06] & M-1: 486 ± 64 min [ES: 0.84, Moderate; p = 0.02] respectively) in comparison to the night of the day after the match day (M+1: 422 ± 61 min). Such a finding suggested that youth soccer players adopt behaviours that reduce their sleep quantity on the designated recovery day (M+1) by >60 min, which may impact the recovery processes associated to this day. Relationships between sleep parameters and training and match load indicated a 100 au rise in RPELOAD (RPE * Duration) increased the time spent in wake (42 s [90% CI: 0 to 84 s]; ES: 0.36, Small; p = 0.098). It was also observed that an increase of 1000 m total distance increased the time spent in wake (40 s [90% CI: 5 to 75 s]; ES: 0.33, small; p = 0.06) A 100 m rise in high-speed running distance increased the number of awakenings observed (0.14 [90% CI: 0.03 to 0.25]; ES: 0.28, p =0.04) and the time spent in wake on average each night (1.5 min [90% CI: 0.78 to 2.3 min]; ES: 0.57, Small; p = 0.04). A similar outcome was observed in Chapter 6 were a 100 m rise in average high-speed running distance across three different 14-day training periods during the youth soccer season showed a 5 min increase in the time spent in wake on average (ES: 0.88, moderate; p = 0.04). Such outcomes provided a potential link between increases in training intensity (i.e. high-speed running distance) and sleep disturbances within youth soccer players. Increases in high-speed running distance also related to an increase of 24 min (90% CI: 12 to 36 min) on average for total sleep time (ES: 1.3, large; p < 0.01). Similarly increased high intensity heart rate (>85% max HR) was shown to effect total sleep time +20 min (90% CI: 6 to 32 min; ES: 0.87, moderate; p = 0.035). This may suggest that increases in intensity also may impact the amount of sleep quantity within youth soccer players. At present the mechanism for this response largely remains unknown. Within Chapter 7, a practical sleep hygiene strategy (10 min showering at ~40 °C, 20 min before time of lights out) was investigated. A group of ten youth soccer players were evaluated under normal sleeping conditions (control) and a shower intervention period, each consisting of three days within a randomized cross over trial design. Sleep information was collected using the WS. In addition to skin temperature, which was evaluated using iButton skin thermistors. The iButtons were used to establish both distal and proximal skin temperatures. This data was also used to create the distal to proximal gradient (average of distal measures – average of proximal measures = DPG). The data demonstrated that the shower intervention elevated distal skin temperature by (+1.1 °C [95% CI: 0.1 to 2.1 °C]; ES: 0.44, Small; p = 0.04) on average during a 10-minute period prior to lights out in comparison to the control condition. This elevation was also present during the first 30 minutes following lights out (1.0 °C [95% CI: 0.4 to 1.6 °C]; ES: 0.65, Moderate; p < 0.01), which was also accompanied by an increased DPG between conditions (0.7 °C [95% CI: 0.3 to 1.2 °C]; ES: 0.45, Small; p < 0.01). Additionally it was observed that on average the sleep onset latency of the youth soccer players was lower in the shower intervention condition (-7min [95% CI: -13 to -2 min]; ES -0.55, Moderate; p = 0.007). However no other sleep variable was affected as a result of the intervention. These findings demonstrate that a warm shower performed before lights out may offer a practical strategy to alter the thermoregulatory properties of distal skin temperature, which may advance sleep onset latency within youth soccer players.

The purpose of the thesis was to provide a detailed analysis of the physical demands of competitive international female soccer match-play and the physical characteristics of elite players. To date, the majority of research has focussed on sub-elite players with a lack of information available on international level competitors. The aim of the first study (Chapter 4) was to analyse match physical performance using a computerised tracking system (Prozone Sports Ltd., Leeds, England). A total of 167 individual match observations from 122 players competing in competitive international matches during the 2011-2012 and 2012-2013 seasons were completed. Total distance and total high-speed running distances (>14.4 km.h-1) were influenced by outfield playing position, with central midfielders completing the highest (10985 ± 706 m and 2882 ± 500 m) and central defenders (9489 ± 562 m and 1901 ± 268 m) the lowest distances, respectively. Greater total very high-speed running distances (>19.8 km.h-1) were completed when a team was without (399 ± 143 m) compared to with (313 ± 210 m) possession of the ball. The majority of sprints (>25.1 km.h-1) were over short distances with 95 % being less than 10 m. This study provides novel findings regarding the physical demands of different playing positions in competitive international female match-play and important insights for physical coaches preparing elite female players for competition. The aim of the second study (Chapter 5) was to determine the incidence and nature of repeated sprint and high-speed activity in match-play. Repeated sprint activity (a minimum of two efforts (>25.1 km.h-1) with 20 s or less recovery between efforts) was found to be rare during international female match-play with 1.1 ± 1.1 bouts per match. Repeated high-speed activity (a minimum of two efforts (>19.8 km.h-1) was influenced by playing position; with attacking-based players completing more bouts (37-40 bouts per match) than defensive players (22-33 bouts per match). Repeated sprint and high-speed bouts frequently comprised two efforts per bout, with a maximum of three and six efforts respectively. Collectively, this study provides physical coaches with useful data for replicating the demands of repeated high-speed activity and an understanding of the positional demands in order to aid the specificity of training. The aim of the third study (Chapter 6) was to attempt to apply a suitable approach for determining speed zones and to evaluate the application of specific zones to influence data outcome. Maximum match-play running speed in elite females was measured using Global Positioning System technology (STATSports, Viper, Ireland) in 230 individual match observations of 67 outfield players, during 19 international matches from 2011-2015. Female-specific speed zones and activity classifications were scaled appropriately to maximum match-play running speed. The resultant female-specific speed zones were on average 12.5 % lower than the standardised male zones, which if applied to the data in Chapter 4 would result in a small increases in total high-speed running (25 % to 28 %) and total very high-speed running (8 % to 9 %) relative to total distance. The calculated female-specific sprinting threshold (>22.0 km.h-1) corresponds to 82 % of the average maximum female match-play running speed presently observed and consequently might be more representative than the standardised male sprinting threshold (>25.1 km.h-1). However, as it was not possible to validate activity classifications in the current study it is suggested that the standardised thresholds should continue to be used to permit between playing position and gender comparisons, however, the activity classifications (e.g. walking, jogging, sprinting etc.) should be removed and replaced with the actual velocities. The aim of the fourth study (Chapter 7) was to examine the reliability of both anthropometric and performance measures in elite female soccer players. The data suggest that both junior and senior elite female players are able to adequately reproduce a variety of anthropometric (coefficient of variation = 0.1-1.3 %) and performance (coefficient of variation = 0.6-7.7 %) related tests and that reliable measures can be obtained using the present protocols and one familiarisation session. The sample size estimations (n<20) provided important insights for the participant recruitment in Chapter 8 and also suggest that the anthropometric and performance assessments are suitable for the longitudinal tracking of the fitness status of elite female players. The aim of the fifth study (Chapter 8) was to examine the physical characteristics of elite players, which were assessed in 471 national team players from 2011-2015. Anthropometric and performance variables improved with age; with large differences observed between U15s and seniors for body mass (53.9 ± 7.8 v 62.5 ± 5.8 kg), 30 m linear speed (4.78 ± 0.22 v 4.52 ± 0.17 s), countermovement jump (28.3 ± 4.0 cm v 33.4 ± 4.0 cm) and Yo Yo Intermittent Recovery Test Level 1 (1101 ± 369 m v 1583 ± 416 m). Similarities were observed for anthropometric and performance variables between the younger (U15 and U17) and older (U19 and senior) age groups. Goalkeepers generally exhibited inferior anthropometric and performance capabilities to outfield players. Faster linear speed times over short distances observed were in attackers (1.047 s v 1.061-1.077 s), greater repeated speed performance in wide midfielders and attackers (4.89-4.91 s v 4.92-4.99 s) and improved intermittent endurance performance in wide defenders (1483 m v 1260-1336 m) compared to other outfield playing positions. The normative physical characteristics presented, provide unique data for professionals involved in player recruitment and talent identification, whilst the positional differences in performance characteristics, coupled with an in-depth understanding of the demands of match-play can be applied to ensure training specificity. Collectively, the present data provides the most in-depth description of the physical demands and physical characteristics of elite female soccer players to date. The data describing the demands of match-play provides valuable insights for physical coaches preparing elite female players for competition, whilst the normative physical characteristic data provides important information to professionals involved in player recruitment and talent identification and those responsible for physical development.

The transition from academy to the professional level is one of the pivotal key stages in the development of elite professional soccer players and our understanding of this process is incomplete (Mills et al. 2012). A comprehensive detailed assessment of elite youth players match performance is limited and a clear gap in the literature exists which has consequences for the development programs of academy players (Harley et al. 2010). Global positioning systems (GPS) have been developed for use within a sports context to measure distance and velocity, however the methods employed for determining accurate capture conditions has not been examined and rely solely on historical criteria established before its use in dynamic sports such as soccer. The studies contained in this thesis are designed to further examine how an increase understanding of the match demands of youth soccer through the use of GPS technology to allow the development of highly tailored training programs to maximise the development of youth players. In summary, the work undertaken from the studies in this thesis provides novel information in relation to the measuring of match performance when using GPS and the application of this information to the programming of training programmes with elite youth soccer players. It has also examined and validated the use of a simple field test to highlight bilateral lower limb imbalances and possible training programs to improve the bilateral ratio and to maximise the development of youth players.