Differences between Sexes and Speed Levels in Pelvic 3D Kinematic Patterns during Running Using an Inertial Measurement Unit (IMU)

o assess the 3D kinematic pattern of the pelvis during running and establish differences between sexes using the IMU sensor for spatiotemporal outcomes, vertical acceleration symmetry index, and ranges of motion of the pelvis in the sagittal, coronal, and transverse planes of movement. The kinematic range in males was 5.92 –6.50 , according to tilt. The range of obliquity was between 7.84 and 9.27 and between 9.69 and 13.60 , according to pelvic rotation. In females, the results were 6.26 –7.36 , 7.81 –9.64 , and 13.2 –16.13 , respectively. Stride length increased proportionally to speed in males and females. The reliability of the inertial sensor according to tilt and gait symmetry showed good results, and the reliability levels were excellent for cadence parameters, stride length, stride time, obliquity, and pelvic rotation. The amplitude of pelvic tilt did not change at

, 7.81 –9.64 , and 13.2 –16.13 , respectively. Stride length increased proportionally to speed in males and females. The reliability of the inertial sensor according to tilt and gait symmetry showed good results, and the reliability levels were excellent for cadence parameters, stride length, stride time, obliquity, and pelvic rotation. The amplitude of pelvic tilt did not change at different speed levels between sexes. The range of pelvic obliquity increased in females at a medium speed level, and the pelvic rotation range increased during running, according to speed and sex. The inertial sensor has been proven to be a reliable tool for kinematic analysis during running. Keywords:biomechanics; kinematics; pelvis; running; wearables; exercise 1. Introduction Running is one of the most popular and accessible activities for the population [1], and its popularity has grown exponentially in the last 50 years [2]. Consequently, there has been an increase in rates of injury, especially in beginners lacking experience, with up to 30% of new runners affected every year [3]. Moreover, the repetitive nature of running makes it an activity with a high injury risk [4], which ranges from 3.2% to 84.9% [5,6], with a median prevalence of 44.6% 18.4% [7]. Out of the many running injuries, 70–80% of them are caused by overuse [7], with Aquilian tendinopathy, plantar fasciitis and patellofemoral, iliotibial band, and the medial tibialis stress syndromes being the most prevalent [7]. Running injuries' etiology is multifactorial, yet it is not possible to determine the exact cause for every injury because movement during running requires a precise inter-segmental coordination [8]. Among studied risk factors, previous injuries [9], high body mass index (BMI) [10], sex, age, experience [11], training alterations [3,9,12], biomechanical issues [13], and fatigue [14,15] are the most prevalent. It has been observed that injured runners change their movement pattern to prevent further damage [14]. Despite this, overall results turn out to be inconsistent due to the large number of potential outcomes that could direct runners to injury. Int. J. Environ. Res. Public Health2023,20, 3631.

It has been observed that injured runners change their movement pattern to prevent further damage [14]. Despite this, overall results turn out to be inconsistent due to the large number of potential outcomes that could direct runners to injury. Int. J. Environ. Res. Public Health2023,20, 3631.

Int. J. Environ. Res. Public Health2023,20, 3631 2 of 12 According to a biomechanical analysis carried out during running, the pelvis plays a stabilizing role and transfers energy between the lower extremity and the rest of the body [16], creating stress in the back [17] and distal structures of the lower limb [18] if there is an alteration in the coordination of pelvic and vertebral movements. The biomechanical analysis of running is an important way to assess movement in individuals with injuries, in- cluding simple spatiotemporal parameters and complex three-dimensional movements [19]. To achieve this, it is necessary to establish the role of the biomechanical coordination of the pelvis during running. Three-dimensional (3D) optoelectronic systems are considered the Gold Standard in the analysis of movement, surpassing clinical observation [20,21]. Despite this, due to its relatively high cost and the time and space required to develop an analysis using these systems [22], in addition to the dif culty of analyzing certain planes in 3D, the techno- logical progress has enabled the development of more affordable, accessible, and feasible devices [23]. An example of this is the inertial measurement unit (IMU), a portable, valid, and reliable device [24] that facilitates assessment of the orientation of the segments and the articular angles [25,26]. In addition, this device shows multiple correlation coef cients above 0.95 with respect to the Gold Standard when comparing different running speed levels [27], presenting high correlations for the angles of tilt, obliquity, and rotation [28,29]. Research studies, such as the one conducted by Novacheck [30], determine refer- ence ranges for the pelvis, which are considered the normative standards for running patterns in the sagittal, coronal, and transverse planes of movement, through analysis with optoelectronic devices. However, in order to account for progress in technology, footwear [31], cultural variations [32], and an increase in physical activity levels in seden- tarypopulations [33] , among other factors, these ranges must be revised, particularly given the kinetic and kinematic biomechanical differences between males and females during running [34]. Analyses should also be established according to different speed levels to determine how the biomechanics

account for progress in technology, footwear [31], cultural variations [32], and an increase in physical activity levels in seden- tarypopulations [33] , among other factors, these ranges must be revised, particularly given the kinetic and kinematic biomechanical differences between males and females during running [34]. Analyses should also be established according to different speed levels to determine how the biomechanics behave in every type of runner. Knowing the kinematic ranges of the pelvis at different speed levels during running and according to sex could help form an understanding of the potential role its alterations might have in functional or structural injuries, by determining how the pelvis interacts during each phase of the running cycle. The main objective of this study is to determine the 3D kinematic pattern of the pelvis during running and to establish differences between sexes using the IMU sensor for the spatiotemporal outcomes, gait symmetry index, and amplitude of motion ranges of the pelvis in the sagittal, coronal, and transverse planes of movement. A secondary objective is to determine the reliability of the IMU sensor for these variables. 2. Materials and Methods 2.1. Participant Charateristics A total of 101 participants were included in the study to determine normative values: 51 males and 50 females. Ages ranged from 18 to 53 years, with a mean of 31.3 years. Mean weight was 65.7 Kgs and mean height was 170 cm (Table). Out of the initial 107 participants, 6 were excluded due to several reasons (medical criterian= 1 and foot blistersn= 5). The analysis of the reliability of the IMU sensor included 29 participants, with a mean age of 31.2 years, a mean weight of 66 Kgs, and a mean height of 172 cm (Table). Participants in this study were healthy subjects with no current injuries who had at least 1 year's running experience, and who had trained for at least a total of 90 min, dis- tributed across weekly training sessions. Participants were excluded if they were older than 65 years, had suffered an injury in the lower limb in the last year, had undergone a surgical intervention, or

healthy subjects with no current injuries who had at least 1 year's running experience, and who had trained for at least a total of 90 min, dis- tributed across weekly training sessions. Participants were excluded if they were older than 65 years, had suffered an injury in the lower limb in the last year, had undergone a surgical intervention, or had neurological problems that might have altered the biomechanics of the standard running cycle pattern.

Int. J. Environ. Res. Public Health2023,20, 3631 3 of 12 Table 1. Demographic data of participants to establish normative values and sensor reliability according to sex. Values are presented with mean and standard deviation (SD). Normative Values Sensor Reliability Outcome Males Females Males Females n 51 50 14 15 Age (years) 32.49 8.61 30.16 8.94 32.36 9.09 30 8.08 Weight (Kgs) 74.02 6.69 57.25 6.11 74.65 6.69 58 6.09 Height (cm) 176 5.70 165.70 5.79 178 4.94 167 7.02 Recruitment took place through circulation in the electronic channels of the triathlon clubs of the Valencian Community and running teams. This study was approved by the Ethics Committee of the University CEU Cardenal Herrera, in Valencia (CEI 14/018), and was conducted according to the basic principles of the Declaration of Helsinki. Every participant was briefed regarding the nature of the study and was asked to give written consent to participate. 2.2. Procedure In the rst phase of the study, standardized values and differences between sexes were established involving the same outcomes. The study then determined the reliability of the IMU sensor in the biomechanical analysis of running. Several spatiotemporal outcomes were assessed, including cadence, running cycle, stride length, and vertical acceleration symmetry index. Anterior–posterior tilt pelvic ranges, obliquity, and pelvic rotation amplitudes were also assessed in females and males. The amplitude of the pelvis 3D movements and spatiotemporal outcomes were as- sessed using an inertial sensor BTS G-Sensor (BTS Bioengineering, Garbagnate Milanese, Italy) with an ergonomic belt at the height of S1 (Figure) to capture different kinematic and spatiotemporal outcomes. This IMU comprised a 16-axis triaxial accelerometer with multiple sensitivities ( 2, 4, 6, 8, and 16 g) with a frequency of 4 Hz to 1000 Hz, a triaxial gyroscope with multiple sensitivities ( 250, 500, 1000, 2000 o/s), with a fre- quency oscillating between 4 Hz to 8000 Hz, and a triaxial 13-bit magnetometer ( 1200 uT), with a frequency exceeding 100 Hz.Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 3 of 12 Table 1. Demographic data of participants to establish normative values and sensor reliability

multiple sensitivities ( 250, 500, 1000, 2000 o/s), with a fre- quency oscillating between 4 Hz to 8000 Hz, and a triaxial 13-bit magnetometer ( 1200 uT), with a frequency exceeding 100 Hz.Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 3 of 12 Table 1. Demographic data of participants to establish normative values and sensor reliability ac- cording to sex. Values are presented with mean and standard deviation (SD). Normative Values Sensor Reliability Outcome Males Females Males Females n 51 50 14 15 Age (years) 32.49 ± 8.61 30.16 ± 8.94 32.36 ± 9.09 30 ± 8.08 Weight (Kgs) 74.02 ± 6.69 57.25 ± 6.11 74.65 ± 6.69 58 ± 6.09 Height (cm) 176 ± 5.70 165.70 ± 5.79 178 ± 4.94 167 ± 7.02 Participants in this study were healthy subjects with no current injuries who had at least 1 year’s running experience, and who had trained for at least a total of 90 min, dis- tributed across weekly training sessions. Participants were excluded if they were older than 65 years, had suffered an injury in the lower limb in the last year, had undergone a surgical intervention, or had neurological problems that might have altered the biome- chanics of the standard running cycle pattern. Recruitment took place through circulation in the electronic channels of the triathlon clubs of the Valencian Community and running teams. This study was approved by the Ethics Committee of the University CEU Cardenal Herrera, in Valencia (CEI 14/018), and was conducted according to the basic principles of the Declaration of Helsinki. Every par- ticipant was briefed regarding the nature of the study and was asked to give written con- sent to participate. 2.2. Procedure In the first phase of the study, standardized values and differences between sexes were established involving the same outcomes. The study then determined the reliability of the IMU sensor in the biomechanical analysis of running. Several spatiotemporal outcomes were assessed, including cadence, running cycle, stride length, and vertical acceleration symmetry index. Anterior–posterior tilt pelvic ranges, obliquity, and pelvic rotation amplitudes were also assessed in females and

study, standardized values and differences between sexes were established involving the same outcomes. The study then determined the reliability of the IMU sensor in the biomechanical analysis of running. Several spatiotemporal outcomes were assessed, including cadence, running cycle, stride length, and vertical acceleration symmetry index. Anterior–posterior tilt pelvic ranges, obliquity, and pelvic rotation amplitudes were also assessed in females and males. The amplitude of the pelvis 3D movements and spatiotemporal outcomes were as- sessed using an inertial sensor BTS G-Sensor (BTS Bioengineering, Garbagnate Milanese, Italy) with an ergonomic belt at the height of S1 (Figure 1) to capture different kinematic and spatiotemporal outcomes. This IMU comprised a 16-axis triaxial accelerometer with multiple sensitivities (±2, ±4, ±6, ±8, and ±16 g) with a frequency of 4 Hz to 1000 Hz, a triaxial gyroscope with multiple sensitivities (±250, ±500, ±1000, ±2000 o/s), with a fre- quency oscillating between 4 Hz to 8000 Hz, and a triaxial 13-bit magnetometer (±1200 uT), with a frequency exceeding 100 Hz. Figure 1.Placement of the IMU in S1.

Int. J. Environ. Res. Public Health2023,20, 3631 4 of 12 In this study, we used a treadmill (BH Fitness Columbia Pro 130 cm 40 cm) to establish standardized conditions under which the kinematic outcomes of running would be more reproducible. We set the incline to 1 and allowed each participant to select the speed [34,35] at which they regularly trained (self-selected speed). The participants performed in their regular training shoes and were allowed a 5 min warm-up period to adjust to the treadmill. According to protocols used in previous running biomechanics studies [36,37], the initial speed was progressively increased over 2 min and was then maintained for 3 min while the data were collected. Spatiotemporal outcomes including cadence, stride length, and running cycle were registered. Vertical acceleration symmetry index, kinematic ranges of tilt, obliquity, and pelvic rotation, as well as the participant's age, weight, height, running experience, and weekly training volumes were also registered. To determine the reliability of the sensor, participants performed 2 tests of 5 min each with a 30 min margin between them, under the same circumstances. Environmental conditions were 22–25 and relative humidity levels ranged from 40 to 55%. The participants who were assessed to establish normative values for pelvic kinematics during running using the inertial sensor were strati ed into three groups: “Slow speed” was considered for values between 9.98 km/h and 8.75 km/h, “medium speed” was considered for values between 9.98 and 11.70 km/h and between 8.75 km/h and 10.11 km/h, and “fast speed” was considered for values above 11.71 km/h and 10.11 km/h for males and females, respectively. 2.3. Statistical Analysis To describe the demographical data of the sample, descriptive statistics were indepen- dently calculated for each sex. Normality distribution was assessed for the independent outcomes using the Kolmogorov–Smirnov test and homogeneity of variance was assessed using Levene's test. Regarding the reliability analysis, the Shapiro–Wilk test was used to determine the normal distribution of the sample. Maximal and minimal values were identi ed at every speed level for both males and females and these were divided into 2 percentiles: 33 and 66. Participants

the independent outcomes using the Kolmogorov–Smirnov test and homogeneity of variance was assessed using Levene's test. Regarding the reliability analysis, the Shapiro–Wilk test was used to determine the normal distribution of the sample. Maximal and minimal values were identi ed at every speed level for both males and females and these were divided into 2 percentiles: 33 and 66. Participants were allocated to one of 3 groups between these percentiles: slow, medium, and fast speed. To assess the reliability of the IMU sensor, the Intraclass Correlation Coef cient (ICC) was calculated. Inferential statistics were performed according to sex and speed. A 2-way factorial ANOVA between subjects was used to determine the kinematic outcomes, the factors being sex and speed. The level of signi cance was established atp< 0.05 with a con dence interval of 95%. Statistical calculations were performed with the SPSS software in its 18.0 version. Epidat 4.2 software was used to obtain the sample size of the 2 tests. In both cases, an alpha error of 0.05 was accepted with a power of 85%. To establish normative values, calculations were performed for two independent means, with a standard deviation of 1.68 [38] and a minimal detectable change of 1.1 units, requiring a minimum sample of 45 subjectsin each group. To check the reliability of the sensor, calculations were performed for two dependent means, with a minimum sample of 25 subjects in total. 3. Results 3.1. Normative Values of the Pelvic Kinematics and Spatiotemporal Outcomes Regarding kinematic ranges, the mean amplitude of pelvic tilt for males and females oscillated between 5.92 and 7.36 (Figure) without statistically signi cant differences, but with a tendency to increase as speed increased in females, whilst in males a bigger range was observed at medium speed (Table).

Int. J. Environ. Res. Public Health2023,20, 3631 5 of 12Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 5 of 12 Figure 2. Pelvic tilt trace with mean amplitude in males and females. Negative degrees represent a retroversion of the pelvis. Positive degrees represent anteversion of the pelvis. No statistical differ- ences found in remaining speed levels or between sexes. Table 2. Spatiotemporal outcomes according to sex and speed levels. Males (SD) Females (SD) Mean Differences (min–max) Slow Speed Symmetry index (%) 99.04 (0.69) 99.37 (0.34) −0.33 (−0.80–0.14) Cadence (p/m) 170.1 (9.7) 171.9 (15.5) −1.76 (−9.56–6.04) Stride time (s) 0.70 (0.04) 0.69 (0.06) 0.01 (−0.03 a −0.04) Stride length (m) 1.81 (0.12) 1.30 (0.14) 0.22 (0.08–0.35) Pelvic tilt (°) 6.27 (1.70) 6.26 (2.40) −0.01 (−1.37–1.77) Pelvic rotation (°) 10.64 (2.47) ** 13.21 (2.98) ** −2.57 (−5.09 a −0.05) Pelvic obliquity (°) 9.27 (2.53) 8.11 (1.34) 1.16 (−1.18–2.51) Medium Speed Symmetry index (%) 99.43 (0.44) 98.89 (0.73) 0.55 * (0.07–1.03) Cadence (p/m) 173.0 (10.4) 172.2 (10) 7.81 (−7.02–8.59) Stride time (s) 0.67 (0.05) 0.69 (0.04) −0.01 (−0.04–0.02) Stride length (m) 2.02 (0.23) 1.84 (0.10) 0.18 (0.06–0.31) Pelvic tilt (°) 5.92 (2.46) 6.57 (2.58) 0.65 (−1.45–1.65) Pelvic rotation (°) 9.96 (3.76) ** 15.74 (3.99) ** −5.78 (−8.26 a −3.30) Pelvic obliquity (°) 7.84 (2.16) 9.64 (1.77) −1.80 *(−3.12 a −0.47) Fast Speed Symmetry index (%) 98.62 (0.82) 98.99 (0.73) −0.37 (−0.81–0.08) Cadence (p/m) 171.3 (9.39) 173.5 (10) −2.117 (−9.47–5.24) Stride time (s) 0.70 (0.04) 0.69 (0.24) 0.01 (−0.02–0.04) Stride length (m) 2.47 (0.27) 2.14 (0.14) 0.33 (0.21–0.45) Pelvic tilt (°) 6.50 (2.00) 7.36 (2.22) 0.86 (−2.38–0.58) Pelvic rotation (°) 13.60 (3.68) ** 16.13 (3.42) ** −2.53 (−4.91 a −0.16) Pelvic obliquity (°) 8.75 (1.44) 7.81 (1.99) 0.95 (−2.23–0.34) * Significant difference p < 0.025 between groups (Males–Females) ** Significant differences p < 0.05 within groups (sex–speed levels). Mean pelvic obliquity oscillated between 7.84° and 9.64° (Figure 3). The obliquity in females at medium speed was significant with respect to males (p < 0.05). No statistical differences were found in the remaining speed level. Mean pelvic rotation ranged from 9.96°

The study analyzes pelvic kinematics in runners of different sexes and speeds.