Performance Level Affects Full Body Kinematics and Spatiotemporal Parameters in Trail Running—A Field Study

running is an emerging discipline with few studies performed in ecological conditions. The aim of this work was to investigate if and how biomechanics differ between more pro cient (MP) and less pro cient (LP) trail runners. Twenty participants (10 F) were recruited for a 9.1 km trail running time trial wearing inertial sensors. The MP athletes group was composed of the fastest ve men and the fastest ve women. Group differences in spatiotemporal parameters and leg stiffness were tested with the Mann–Whitney U-test. Group differences in joint angles were tested with statistic parametric mapping. The nish time was 51.1 6.3 min for the MP athletes and 60.0 5.5 min for the LP athletes (p< 0.05). Uphill sections: The MP athletes expressed a tendency to higher speed that was not signi cant (p> 0.05), achieved by combining higher step frequency and higher step length. They showed a tendency to shorter contact time, lower duty factor and longer ight time that was not signi cant (p> 0.05) as well as signi cantly lower knee exion during the stance phase (p< 0.05). Downhill sections: The MP athletes achieved signi cantly higher speed (p< 0.05) through higher step length only. They showed signi cantly higher knee and hip exion during the swing

shorter contact time, lower duty factor and longer ight time that was not signi cant (p> 0.05) as well as signi cantly lower knee exion during the stance phase (p< 0.05). Downhill sections: The MP athletes achieved signi cantly higher speed (p< 0.05) through higher step length only. They showed signi cantly higher knee and hip exion during the swing phase as well as higher trunk rotation and shoulder exion during the stance phase (p< 0.05). No differences were found with respect to leg stiffness in the uphill or downhill sections (p> 0.05). In the uphill sections, the results suggest lower energy absorption and more favorable net mechanical work at the knee joint for the MP athletes. In the downhill sections, the results suggest that the more ef cient motion of the swing leg in the MP athletes could increase momentum in the forward direction and full body center of mass' velocity at toe off, thus optimizing the propulsion phase. Keywords:trail running; biomechanics; incline running; ecological study 1. Introduction Trail running is a strongly emerging endurance running discipline [1]. It has been previously de ned as any running that takes place in an open country on unpaved surfaces (i.e., off road) with <20–25% paved surface [2]. Mountains represent the ideal scenario with mild, steep or technical uphill (UH) and downhill (DH) traits, where athletes must cope with both physical and mental fatigue. Trail running competitions cover many distances from ve up to several hundred kilometers. In the last two decades, the number of studies focusing on the biomechanics and physiology of graded running (i.e., running on non- level conditions, on negative/positive inclines) grew in parallel with race attendance [3–6]. Several lab-based studies reported differences in spatiotemporal parameters, kinematics, kinetics and muscle activation when comparing level and graded running [7]. Nonetheless, a major shortcoming of lab settings is the absence of interaction between sport performers and the external environment, which is critical to any outdoor discipline. In fact, in a lab-based setting, several factors critical for trail running may be hardly reproduced, such as the irregularity of the ground and different

muscle activation when comparing level and graded running [7]. Nonetheless, a major shortcoming of lab settings is the absence of interaction between sport performers and the external environment, which is critical to any outdoor discipline. In fact, in a lab-based setting, several factors critical for trail running may be hardly reproduced, such as the irregularity of the ground and different terrains (gravel, grass, etc.), to name a few. Therefore, eld-based investigations may convey valuable information to coaches Sports2023,11, 188.

Sports2023,11, 188 2 of 11 and practitioners. Novel wearable technology has enabled the research of trail running outside the laboratory. For example, Defer et al. [8] used inertial sensors to report that footwear with a smaller heel-to-toe drop may elicit changes in the foot-strike pattern and improve performance in short DH runs; Giandolini et al. [9] used portable EMG sensors to investigate neuromuscular fatigue during ecological DH running, reporting both central and peripheral fatigue-induced dysfunctions. Björklund et al. [10] utilized pressure insoles to investigate the kinetics in a trail running eld test, reporting a higher force impulse in UH sections and higher peak force in DH sections. Born et al. [11] reported trail running intensity to be accurately re ected by the tissue saturation index, measured with a compact near-infrared spectroscopy device during a trial run in outdoor conditions. Always in an ecological framework, Townshend et al. [12] found running speed to depend more on stride length than on stride frequency. However, in comparison to other endurance disciplines, little is known about if and how biomechanics differ between trail runners of different performance levels. To the best of our knowledge, only one study investigated such differences in running performance for level and UH running on a treadmill [13]. The authors reported more pro cient (MP) trail runners to express lower (better) cost of running, lower step frequency, lower leg stiffness and lower vertical stiffness compared to less pro cient counterparts during level running at the same speed. In the same study, no differences were observed in uphill running biomechanics, whereas there still exists no evidence about possible different biomechanical behavior of trail running athletes of different performance levels in downhill running. During competitions, trail runners of different performance levels were reported to differ in both absolute and relative speed (as a fraction of average race speed) with more pro cient athletes running UH sections at lower relative speed and DH sections at higher relative speed compared to less pro cient counterparts [14]. Therefore, it is possible that movement pattern and related mechanical parameters differ between trail runners of different performance

reported to differ in both absolute and relative speed (as a fraction of average race speed) with more pro cient athletes running UH sections at lower relative speed and DH sections at higher relative speed compared to less pro cient counterparts [14]. Therefore, it is possible that movement pattern and related mechanical parameters differ between trail runners of different performance levels when running at self-selected speed in ecological conditions. New insights in this regard combined with previous ndings may represent an important step towards the development of speci c training protocols aiming to (I) prevent injuries and (II) optimize performance in a growing endurance discipline. Therefore, the purpose of this study was to investigate if and how spatiotemporal parameters, full-body kinematics and leg stiffness differ between performance levels in a eld trail running session. It was hypothesized that MP athletes would express different kinematics and spatiotemporal parameters (STP) as well as lower leg stiffness compared to less pro cient (LP) counterparts. 2. Materials and Methods 2.1. Participants and Experimental Design Twenty subjects (10 M, 10 F) were recruited from local trail running associations (age [years]: 32.8 8.3 M, 33.4 8.1 F; body height [cm]: 177.2 6.0 M, 166.3 6.9 F; body mass [kg]: 71.9 5.8 M, 61.6 6.9 F; experience in trail running [years]: 3.3 1.5 M, 4.1 1.2 F). This study was approved by the ethical committee of the local university (GZ-10/2022). The inclusion criteria required that participants were between 18 and 50 years old, had at least 1 year experience in trail running, had a minimum training frequency of 2 times per week, ran a minimum weekly training volume of 30 km, had sustained no injuries in the previous 3 months prior to the study. After providing their informed consent, anthropometric data, including body height, body mass and length of relevant body segments, were recorded. On a separate day, the participants completed a 9.1 km trail running time trial consisting of 7 laps of the same 1.3 km route (Figure). Each lap presented an ascent of 60 m, resulting in a 420 m gain across the entire trial.

informed consent, anthropometric data, including body height, body mass and length of relevant body segments, were recorded. On a separate day, the participants completed a 9.1 km trail running time trial consisting of 7 laps of the same 1.3 km route (Figure). Each lap presented an ascent of 60 m, resulting in a 420 m gain across the entire trial. Before the test, the participants were accompanied during a complete lap of the running route to familiarize themselves with the test environment followed by a warm-up at a self-selected intensity, consisting of short UH and DH runs and static

Sports2023,11, 188 3 of 11 stretching exercises. During the test, time and positional data, STP, full-body kinematics and leg stiffness were recorded. The performance level was determined according to the nish time: the fastest 5 men and the fastest 5 women were assigned to the more pro cient MP athletes group; the slowest 5 men and the slowest 5 women were assigned to the LP athletes group. Figure 1. Elevation pro le of trail running test lap; a full test consisted of 7 repetitions. Each lap presents one UH and one DH section, where data were retained for further analysis. 2.2. Materials During the trail running test, the participants wore a GPS watch (Garmin Forerunner 935), a foot sensor to record STP (Stryd Summit Powermeter; Stryd, Inc., Boulder, CO, USA, sampling frequency 1 Hz) and a full-body motion-capture system (Xsens Link, Xsens Technologies BV, Enschede, The Netherlands). The latter consisted of 15 inertial measurement units (IMUs, model MTx, size 36 24.5 10 mm, mass 10 g, sampling frequency 240 Hz); IMU sensors were located on the head, shoulders (2 ), arms (2 ), forearms (2 ), thighs (2 ), legs (2 ), feet (2 ), sternum and pelvis. 2.3. Data Analysis The data from speci c UH and DH sections were retained for further analysis. Those sections (Figure, slope ca 12%, length = 80 m UH and 120 m DH) were selected so as not to present abrupt changes in steepness and ground morphology. Step frequency (SF), step length (SL), contact time (CT) and ight time (FT) were measured with the Stryd foot sensor. STP were made adimensional in order to control for anthropometric differences between the participants [15]. In particular, SF was scaled with p g/l0 ; SL was scaled withl0; and CT and FT were scaled with p l0/g , wherel0is the body height of each subject andgis the gravity acceleration. Duty factor (DF) is the percentage amount of time spent in CT during a stride, where stride indicates the time between two consecutive ipsilateral foot strikes. DF was calculated from the adimensional CT and FT as DF

scaled withl0; and CT and FT were scaled with p l0/g , wherel0is the body height of each subject andgis the gravity acceleration. Duty factor (DF) is the percentage amount of time spent in CT during a stride, where stride indicates the time between two consecutive ipsilateral foot strikes. DF was calculated from the adimensional CT and FT as DF = 100 CT/(2(CT + FT)). It was assumed that gait phases have similar duration for both legs with contralateral foot strike occurring at 50% of gait cycle. The joint angles of the ankle, knee, hip, trunk and shoulder were processed with the software MVN Analyze © (Xsens Link, Xsens Technologies BV, Enschede, The Netherlands) in HD mode and No-Level scenario, thus yielding joint angle trajectories as recommended for biomechanical applications [16] and validated in previous works [17]. In particular, the No-Level scenario is well suited when investigating joint angles speci cally rather than other quantities (e.g., center of mass trajectory). Also, in HD mode, the software processes data over a larger time window to obtain an optimal (and more consistent) estimate of the position and orientation of each body segment, ultimately resulting in a precise estimation of joint angle trajectories [16]. Leg stiffness was measured with the Stryd foot sensor. Validation studies reported that such a portable device provides values comparable to those obtained with gold-standard methods [18]. All subsequent analyses were performed in a Python environment. Gait events were identi ed with a previously validated algorithm [19]. Gait cycle data were resampled to 100 data points. For

Sports2023,11, 188 4 of 11 the laps 2–6 (the rst and last were discarded, thus resulting in 5 retained laps) in each UH and DH section, the rst 20 strides per leg for each subject were analyzed. 2.4. Statistics Separately for the UH and DH sections and for each participant, the data from all strides were averaged. The graphical representations for STP and leg stiffness are provided as boxplots, while joint angles are represented as time series against gait cycle percentage. Group differences in STP and leg stiffness were assessed via the Mann–Whitney U-test, alpha = 0.05. Group differences in joint angle kinematics were assessed with statistical parametric mapping [20], t-test with alpha = 0.05. 3. Results One female participant was not able to complete the trail running test, thus resulting in a sample size of ten for MP athletes (5 M, 5 F) and nine for LP athletes (5 M, 4 F). The general characteristics of the MP and LP athletes are presented in Table. Body mass was signi cantly higher for the LP athletes (p< 0.05), while height, age and experience in trail running did not differ (p> 0.05). Table 1.General characteristic of MP and LP athletes. * =p< 0.05, ns = not signi cant. MP Athletes LP Athletes Signi cance Height [cm] 170.7 9.1 173.6 7.7 ns Body mass [kg] 63.3 6.7 71.1 7.9 * Age [years] 32.9 8.0 33.3 8.4 ns Experience [years] 4.0 1.4 3.3 1.3 ns The nish times for the trail running test as well as the split times for the speci c UH and DH sections where the present data were collected are presented in Table. The nish time for the trail running test and split times in the DH sections were signi cantly lower for the MP athletes (p< 0.05). In the UH sections, the difference between the groups was not signi cant (p> 0.05). Table 2. Time comparison between performance levels for the trail running test as well as for the speci c sections of the trail running test where the present data were collected. * =p< 0.05, ** =p<

signi cantly lower for the MP athletes (p< 0.05). In the UH sections, the difference between the groups was not signi cant (p> 0.05). Table 2. Time comparison between performance levels for the trail running test as well as for the speci c sections of the trail running test where the present data were collected. * =p< 0.05, ** =p< 0.01, ns = not signi cant. MP Athletes LP Athletes Signi cance Finish time [min] 51.1 6.3 60.0 5.5 * DH sections [s] 36.0 5.0 42.0 3.3 ** UH sections [s] 29.8 4.4 33.3 3.8 ns 3.1. Spatiotemporal Parameters No signi cant differences (p> 0.05) were found with respect to SF in the UH or in the DH sections (Figurea). The MP athletes showed signi cantly longer SL in the DH sections only (p< 0.05, Figureb). No signi cant differences were found in CT and FT (p> 0.05) despite that the data distribution of the MP athletes showed generally lower values for CT in the UH sections and higher values for FT in both the UH and DH sections (Figurec,d). No signi cant differences (p> 0.05) were found with respect to DF in the UH or DH sections despite that the data distribution of the MP athletes showed generally lower values, especially in the UH sections (Figuree). The overall average DF was 39% in the UH sections and 33% in the DH sections. A schematic representation of the gait phases during a complete gait cycle is provided in Figure.

Sports2023,11, 188 5 of 11 Figure 2. Spatiotemporal parameters by terrain and performance level. MP = more pro cient athletes, LP = less pro cient athletes, ns = no signi cant difference between groups, * =p< 0.05. Sub gures: (a) step frequency; (b) step length; (c) contact time; (d) ight time; (e) duty factor. Figure 3. Duration of gait phases with respect to gait cycle in uphill (UH) and downhill (DH) conditions. Values are averaged across all participants to provide a general overview, as differences in DF between the MP and LP athletes were not signi cant and the values differ by only ca 2–3% between performance levels. 3.2. Kinematics Signi cant differences were found at the shoulder, trunk, hip and knee in the DH sections in both the stance and swing phases (p< 0.05). Also, signi cant differences were found at the knee joint in the UH sections in the early stance and late swing phases(p< 0.05). No differences were found for the ankle joint (p> 0.05). Uphill: The MP athletes expressed signi cantly higher knee exion (p< 0.05) during the initial stance phase and very late swing phase (Figurea). DownhillHip and knee:

Sports2023,11, 188 6 of 11 Signi cantly higher peak exion (p< 0.05) was observed during the swing phase for the MP athletes (Figureb,c). Shoulder:The MP athletes showed signi cantly higher exion (p< 0.05)during the mid-late stance phase compared to the LP athletes (Figured). Trunk: Signi cantly larger ROM (due to larger peak values) was expressed by the MP athletes (p< 0.05) compared to that of the LP athletes (Figuree). Figure 4. Kinematic comparison between performance levels. (a–c): lower body; (d,e): upper body. MP = more pro cient athletes, LP = less pro cient athletes, SPM = statistical parametric mapping. 3.3. Leg Stiffness No signi cant differences (p> 0.05) for leg stiffness were found between the athletes of different performance levels, neither in the UH sections nor in the DH sections (Figure). Figure 5. Leg stiffness separated by performance level in UH and DH sections. MP = more pro cient athletes, LP = less pro cient athletes, ns = not signi cant difference between groups. 4. Discussion The present results con rmed the hypothesis that full-body kinematics would differ between athletes of different performance levels. The hypothesis that STP would differ between athletes of different performance levels was con rmed in DH only. Finally, the hypothesis that leg stiffness would be lower for MP athletes was not supported by our results. In the present study, body mass was signi cantly lower in the MP athletes. This is in line with a previous investigation that compared trail runners of different performance levels in a laboratory setting with elite athletes showing lower body mass compared to the LP counterparts [13]. The present results show how kinematics and STP differ between performance levels in outdoor trail running. To the best of our knowledge, this is the rst study to report on full-body kinematics across a whole trail running eld test. The MP athletes presented a systematic tendency to longer SL with a signi cant difference in the DH sections, while no signi cant differences were found for SF between performance levels (Figurea,b). Townshend et al. [ 12] found that running speed was mainly regulated

This study investigates biomechanical differences in trail running based on performance level.