Higher Hamstrings Strength and Stability Are Related to Lower Kinematics Alteration during Running after Central and Peripheral Fatigue

n be classi ed as peripheral or central depending on the extent of its effects. Muscle strength reduction, associated with the appearance of fatigue during running, produces kinetics and kinematics modi cations which could lead to an increased risk of injury. This study aimed to analyze the effect of peripheral and central fatigue protocols in running kinematics and to investigate the relationship between isokinetic strength and dynamic stability in fatigue related changes. Eighteen male recreational runners participated in the study. The dynamic postural stability index (DPSI) and quadriceps and hamstring isokinetic strength were assessed before the fatigue test. Then, angular kinematics during treadmill running were evaluated in pre- and post-fatigue states (central and peripheral). The results showed that runners with higher hamstring isokinetic strength and better DPSI had lower modi cations after central fatigue of stance time, knee exion, vertical and leg stiffness, and ankle dorsi exion during the absorption and propulsion phases (r > 0.400,p< 0.05). Moreover,

Then, angular kinematics during treadmill running were evaluated in pre- and post-fatigue states (central and peripheral). The results showed that runners with higher hamstring isokinetic strength and better DPSI had lower modi cations after central fatigue of stance time, knee exion, vertical and leg stiffness, and ankle dorsi exion during the absorption and propulsion phases (r > 0.400,p< 0.05). Moreover, small changes in ankle dorsi exion at initial contact after peripheral fatigue are related to a better DPSI and higher hamstring isokinetic strength (r > 0.400,p< 0.05). In summary, high values of hamstring isokinetic concentric strength and dynamic stability are related to lower increases of range of movements during running after central and peripheral fatigue. So, fatigue may affect to a lesser extent the running technique of those runners with higher hamstring strength and stability values. Keywords:running; fatigue; strength; stability; kinematics 1. Introduction Neuromuscular fatigue has been suggested as one of the main causes of injury in running [1] since it is characterized by decreasing muscle strength or power and produces kinetics and kinematics modi cations [1–7]. Neuromuscular fatigue is commonly classi ed as peripheral and central to clarify the origin of these changes. Peripheral fatigue causes alterations at a muscular level and in contractile elements (e.g., alteration at the cross- bridges level, sarcolemma excitability, or excitation–contraction coupling failure) [8,9]. On the other hand, central fatigue is produced by limitations at the neuromuscular junction (e.g., limiting maximal voluntary activation or neural drive to the muscle) [8,9]. Peripheral fatigue during running decreases muscle strength and activity [3], modi es running biomechanics and spatiotemporal parameters [3,10], and increases ground reaction forces [10] and shock absorption [3]. Central fatigue, also alters muscle strength and activity [11–13], promoting changes in movement patterns and spatiotemporal parameters [11,14], increases in ground reaction forces [15] and shock absorption [16], as well as decreases in stiffness characteristics [17], Sensors2022,22, 1990.

Sensors2022,22, 1990 2 of 14 increases in tissue vibration [18] and plantar pressures [19], and decreased postural stability dynamics [20], trunk stability [21,22] or in the performance of cognitive tasks [23]. In addition to fatigue, some of these changes are also affected by the surface [24,25], footwear [26], orthoses [27] or compression garments [16]. Excessive vertical ground reaction forces (vGRF) maintained for a long time during running causes important stress in the musculoskeletal system, and its capacity to sustain these forces can be reduced because of the fatigue [1,3–5]. For this reason, vGRF is con- sidered a relevant outcome parameter for running assessment since high peak values or a loading rate that is higher than the runner capacity increases injury risk [28,29]. In order to minimize vGRF after both peripheral [3,4] and central fatigue [1,5], an increase of range of movements is produced as a compensatory strategy, characterized by a greater knee- exion angle during ground contact. It has been suggested that hamstrings muscles play an essential role during running fatigue and injury risk because an inhibition of the hamstring muscles is produced before the onset of fatigue, which causes a dominance of the quadriceps in the loading response phase and that induces an increased knee exion [3,4]. However, these modi cations can increase the metabolic cost, making the running technique less ef cient [5,6]. So, high levels of muscle strength could prevent, or at least delay, the kinematic changes associated with fatigued running [30]. During forward jump landing tasks, dynamic postural control imitates the initial contact and absorption phases of running, where knee exion–extension strength has a leading role in a safe landing [31]. It has been shown that fatigue also increases the range of movements and decreases vGRF in this type of task [32,33]. Biomechanics changes due to fatigue lead to suboptimal movements that can increase the risk of injury [19]. In addition, injuries occur especially nearly at the end of competitions or training where the fatigue processes are very advanced [34]. Therefore, the high popularity of running and the high injury incidence suggest that identifying and comparing

in this type of task [32,33]. Biomechanics changes due to fatigue lead to suboptimal movements that can increase the risk of injury [19]. In addition, injuries occur especially nearly at the end of competitions or training where the fatigue processes are very advanced [34]. Therefore, the high popularity of running and the high injury incidence suggest that identifying and comparing the biomechanical changes produced by peripheral and central fatigue, depending on factors such as strength or stability, could add a further step in the prevention of running injuries. We hypothesized that after fatigue protocols the running kinematics will change, adopting a less ef cient running pattern. We also expected that central fatigue will affect the running kinematics more than peripheral fatigue protocol, showing greater changes during landing and absorption phases during central fatigue. While we expect to observe compensatory patterns to maintain running ef ciency after peripheral fatigue. Different levels of strength and/or dynamic stability would be hypothesized to affect running kinematics after fatigue, speci cally runners with a higher isokinetic strength and/or higher dynamic stability would reduce the kinematic changes expected after central and/or peripheral fatigue protocols. Describing the relationship between the appearance of fatigue (central and peripheral), the alterations that it produces on the running kinematics, and the levels of strength and stability, would be an advance in the understanding of the internal processes related to the factors of running injury risks. We explored the running kinematics before and after two fatigue protocols (central and peripheral) and related to leg exion–extension the isokinetic strength pro le and dynamic stability. Our objective was to quantify running kinematic changes because of central and peripheral fatigue protocols and to relate the magnitude of the changes to the levels of isokinetic strength and dynamic stability. The key contributions of this paper can be summarized as follows: We investigated the differences in running kinematics after two fatigue protocols to identify the responses associated with fatigue. We have described that central fatigue induce changes in running kinematics to a lesser ef ciency running pattern. We described the relationship between isokinetic strength and dynamic stability

and dynamic stability. The key contributions of this paper can be summarized as follows: We investigated the differences in running kinematics after two fatigue protocols to identify the responses associated with fatigue. We have described that central fatigue induce changes in running kinematics to a lesser ef ciency running pattern. We described the relationship between isokinetic strength and dynamic stability vari- ables as a predictor of prevention effects of the fatigue processes.

Sensors2022,22, 1990 3 of 14 We have described that higher hamstring isokinetic strength and dynamic stability are related to lower kinematic changes in the running pattern. 2. Materials and Methods 2.1. Participants Eighteen male recreational runners (n= 18) participated in the study (age:28.2 8.6 years ; height: 1.77 0.065 m; body mass: 71.7 8.4 kg; estimated maximal oxygen consumption: 62.2 4.7 mL/kg/min; running experience: 7.3 5.3 years). For the selection of the sample, a non-probabilistic sampling by quotas was used, whose quotas were represented by the inclusion and exclusion criteria. As inclusion criteria for the study, it was decided that all the participants were men, recreational runners, aged between 18 and 45 years, who ran regularly [35], at least twice a week in the last year [36], and who did not present lesions at the time of the investigation or in the 6 months prior to it [35]. Participants who did not meet all the criteria set were excluded. The inclusion of the athletes who met the conditions to participate in the study was carried out once the informed consent to participate in the project had been granted. Informed consent was provided to all participants before inclusion in the study, which was approved by the Ethics Committee of the University (registry number: 6775). A sample size calculation was performed based on the ANOVA repeated measures within factors design, using the G-Power 3 software (version 3.1.9.7, Düsseldorf, Germany). This analysis indicated that at least a sample of 16 cyclists was required to detect signi cant differences in the different variables analyzed with a minimum detectable effect size of f = 1.0 (large) ( = 0.05, = 0.05, power = 0.94). 2.2. Experimental Setups Each participant completed three evaluation sessions, separated by 48–72 h each one. In the rst session, the maximal aerobic speed (MAS) was estimated using the 5-min Running Field Test [37]. The second session was carried out as follows: (I) warm-up, (II) isokinetic strength registration, (III) angular kinematics recording before and after peripheral fatigue. Finally, the third session was performed as follows: (I) warm-up, (II) evaluation of dynamic postural

48–72 h each one. In the rst session, the maximal aerobic speed (MAS) was estimated using the 5-min Running Field Test [37]. The second session was carried out as follows: (I) warm-up, (II) isokinetic strength registration, (III) angular kinematics recording before and after peripheral fatigue. Finally, the third session was performed as follows: (I) warm-up, (II) evaluation of dynamic postural stability, (III) angular kinematics recording before and after central fatigue. It should be noted that the second and third laboratory sessions were randomized. All measurements were registered in the dominant limb [38]. The warm-up consisted of running freely for 10 min, which also allowed them to familiarize themselves with the treadmill (Excite ® + Run MD Inclusive, Technogym Trading S.A., Barcelona, Spain) [37] (Figure). Figure 1.Experimental protocol followed in the study.

Sensors2022,22, 1990 4 of 14 2.2.1. Isokinetic Strength Assessment Regarding isokinetic strength registration, peak concentric torque values in quadriceps and hamstring muscles were recorded using an isokinetic dynamometer (Biodex System Pro 3™, Biodex Medical Systems, Inc., New York, NY, USA). From a seated position (80 hip exion), two sets of concentric/concentric knee exion–extension movements were performed, with a range of motion ranging from 0 (full extension) to 90 of knee exion [3,39]. In the rst set, three sub-maximal concentric contractions at 60 /s were performed as a familiarization. In the second set, three maximal concentric contractions at 120 /s were carried out to determine peak concentric torque values. Peak concentric torque values in quadriceps (QTORQ) and hamstrings (HTORQ) were registered, considering the highest value for analysis, and expressed as a percentage of the body weight. The angles at which QTORQ and HTORQ were reached were also recorded (QANG-TORQ and HANG-TORQ, respectively). The hamstrings/quadriceps strength ratio (H/Q ratio) was also calculated. 2.2.2. Dynamic Stability Assessment Dynamic postural stability was registered through an adaptation of the Dynamic Postural Stability Index (DPSI) test [40]. Before the test, each participant performed three valid countermovement jumps to calculate the 50% of their maximum jump height, using the highest jump [40]. Runners were placed 70 cm from the center of a force platform (Kitsler 9286BA, Kistler Group, Winterthur, Switzerland) and they were instructed to double limb jump over an elastic band set at 50% of their maximum jump height, with hands on hips and looking forward, landing on their dominant limb, and stabilizing as quickly as possible. To familiarize themselves with the test, a minimum of three practice attempts were required [41]. After the practice, three attempts were performed to evaluate the mediolateral (MLSI), anteroposterior (APSI), vertical (VSI), and global (DPSI) stability indices [40], recording the ground reaction force (GRF) signals at a frequency of 1000 Hz. The rst three seconds after impact were used for analysis [40]. Isokinetic strength and dynamic postural stability descriptive variables are shown in Table. Table 1.Descriptive parameters of isokinetic strength and Dynamic postural stability. Mean SD VSI * 0.325 0.056

vertical (VSI), and global (DPSI) stability indices [40], recording the ground reaction force (GRF) signals at a frequency of 1000 Hz. The rst three seconds after impact were used for analysis [40]. Isokinetic strength and dynamic postural stability descriptive variables are shown in Table. Table 1.Descriptive parameters of isokinetic strength and Dynamic postural stability. Mean SD VSI * 0.325 0.056 MLSI * 0.114 0.010 APSI * 0.031 0.005 DPSI * 0.346 0.055 QTORQ (%) 245.28 39.60 HTORQ (%) 124.77 31.26 QANG-TORQ ( ) 56.71 5.08 HANG-TORQ ( ) 40.65 11.28 H/Q ratio (%) 50.6 8.2 SD: Standard Deviation, *: Dimensionless, VSI: Vertical Stability Index, MLSI: Mediolateral Stability Index, APSI: Anteroposterior Stability Index, DPSI: Dynamic Postural Stability Index, Q: quadriceps, H: hamstrings, TORQ: Peak Torque, ANG: Peak Torque Angle. 2.2.3. Angular Kinematics Assessment Regarding angular kinematics recording, the measurement protocol was the same in both sessions, modifying only fatigue protocol (peripheral or central fatigue). Angular kinematics were recorded in a 2-min treadmill running period at 3.89 m/s and 0% slope, both before and after fatigue conditions. Retro-re ective markers were placed on the lateral of the greater trochanter, femoral condyle, lateral malleolus and 5th metatarsal head. Four posterior markers were also located on the shoe and lower leg [42] (Figure). An Optitrack V120:Trio infrared motion capture system (NaturalPoint, Inc., Corvallis, OR, USA), running at 120 Hz, was used during the last 30 s of each 2-min period (pre and post-fatigue) to

Sensors2022,22, 1990 5 of 14 track the markers. No recovery time was allowed. Once they nished central/peripheral fatigue protocol, the 2-min running test was performed as quickly as possible to avoid recovery processes. Figure 2.Kinematic markers setup model employed in the study. Forty- ve stride cycles were approximately registered in each condition, and data processing was performed using Motive software (NaturalPoint, Inc., Corvallis, OR, USA). Marker data were ltered with a fourth-order low-pass Butterworth lter with a cut-off fre- quency of 6 Hz. A custom routine performed with the MatLab R2013b program (Mathworks Inc., Natick, MA, USA) was used to calculate running kinematics. The angle convention (Figure) was used to detect the movements of thigh and knee exion–extension, shank oscillation, ankle dorsi exion-plantar exion and rearfoot eversion–inversion. Standing calibration of body segments was considered as zero degrees. Thus, positive values rep- resented hip exion, knee exion, greater shank oscillation, ankle plantar exion and rearfoot inversion, while negative values described hip extension, knee extension, lower shank oscillation, ankle dorsi exion and rearfoot eversion. Root mean square error (RMSE) was calculated to determine the 3D reconstruction accuracy, obtaining a systematic error of 0.005, 0.012 and 0.037 mm for X (mediolateral), Y (anteroposterior), and Z (vertical) axes, respectively. Gait cycles were normalized to 101 data points, and the stance was divided into absorp- tion and generation phases. The absorption phase was represented from the initial contact (IC) to maximum knee exion (MKF) in the midstance, while the generation/propulsion phase was interpreted from MKF to toe-off (TO) [43]. The best method to identify the IC regardless of the foot strike pattern is through the vertical velocity of the pelvis [44]. So, IC was identi ed as the frame of maximum downward velocity of the trochanter. MKF was detected as the peak knee exion located between the two peaks knee extensions produced in IC and TO. TO was identi ed as the second peak knee extension [44]. Finally, maximum oscillation during swing (MO) as peak knee exion located between the two peaks knee extensions produced in the TO and IC. Additionally, spatiotemporal parameters (stride frequency,

trochanter. MKF was detected as the peak knee exion located between the two peaks knee extensions produced in IC and TO. TO was identi ed as the second peak knee extension [44]. Finally, maximum oscillation during swing (MO) as peak knee exion located between the two peaks knee extensions produced in the TO and IC. Additionally, spatiotemporal parameters (stride frequency, stride length, stride time, stance time, and swing time) were calculated. Finally, leg and vertical stiffness were also estimated from the kinematics variables using the spring-mass model [45]. 2.2.4. Fatigue Protocols Regarding fatigue generation, on the one hand, central fatigue was induced by30-min of treadmill running (0% slope) at 85% of MAS [37]. Furthermore, runners had to manifest a perceived effort equal to or greater than 17 or “Very Hard” [2] on the Borg's Scale 6–20 [46]. On the other hand, peripheral fatigue was induced with an isokinetic dynamometer. Contin- uous concentric/concentric knee exion–extension movements at 120 /s were performed, exerting maximal effort through the whole range of motion, without rest. Fatigue protocol nished when the concentric peak torque fell below 50% for 3 consecutive movements in both directions [39].

Sensors2022,22, 1990 6 of 14 2.3. Statistics Data were analyzed with the statistics software SPSS Statistics (SPSS v.26, Chicago, IL, USA). After checking the normality of the variables with the Kolmogorov–Smirnov test, two-way repeated measures ANOVA (normal distribution variables) or Friedman test (non-normal distribution variables) was carried out to compare running kinematics (I) pre vs. post central fatigue and (II) pre vs. post peripheral fatigue. Delta (D) or pre-post fatigue modi cations between peripheral and central fatigue were evaluated by paired samplest-test (normal distribution variables) or Wilcoxon test (non-normal distribution variables). Statistical signi cance was set atp< 0.05. When differences were signi cant, con dence intervals (95% CI) and Cohen's d effect size (ES) were also calculated, where >0.2 is considered small, >0.5 moderate, and >0.8 large [47]. The relationship between research factors and post-fatigue variables was evaluated through Pearson's Correlation Coef cient (r), where magnitude was interpreted as: <0.1, trivial; 0.1–0.3, small; 0.3–0.5, moderate; 0.5–0.7, large; 0.7–0.9, very large; 0.9–1.0, almost perfect; and 1.0, perfect [48]. Moreover, the coef cient of determination (R 2 ) (i.e., the percentage of the variance in the dependent variable that can be explained by variations in independent variables) was calculated elevating r squared and multiplying it by 100 [49]. 3. Results Kinematics modi cation pre vs. post peripheral and central fatigue, and the com- parison of kinematics modi cations after central vs. peripheral fatigue, are shown in Tables, respectively. Regarding the effects of fatigue on spatiotemporal variables, in our study, only stance time and propulsion time was signi cantly higher after central fatigue (p= 0.025 andp= 0.033, respectively) (Table). No differences were observed on stiffness variables. At initial contact, shank inclination (p= 0.034) and ankle- exion (p= 0.035) were increased after central fatigue and peripheral fatigue, respectively. Knee- exion was increased (p= 0.000) after the central fatigue protocol during the maximum knee exion phase. During the take-off phase, knee- exion increased (p= 0.003) after the peripheral fatigue protocol, as well as the shank inclination (p= 0.020), which also increased after the central fatigue protocols (p= 0.002) (Figure). No differences were observed on maximum

This study analyzes the effect of fatigue on running kinematics and strength.