Running-Induced Fatigue Changes the Structure of Motor Variability in Novice Runners

Simple Summary Endurance sports, and especially running, are very popular. During running, fatigue inevitably occurs, and especially in novices. Surprisingly, the effects of fatigue have been studied less extensively in novice runners compared with experienced runners. Regardless of the level of expertise, it has been shown that, by analyzing motor variability, valuable insights can be gained regarding the control of important variables. Motor variability in running is understood as step-to-step deviations. One example of an important variable is the body’s center of mass, as it provides a simplified representation of the overall movement. Thus, an analysis of the motor variability of the body’s center of mass might lead to valuable insights. Therefore, the purpose of this study was to investigate the effects of fatigue on the motor variability of the body’s center of mass in novice runners. It was found that, with fatigue, the motor variability increased, and the control of the body’s center of mass decreased. Moreover, there was a correlation between the decrease in control and the degree of fatigue. Further studies should investigate which training methods can mitigate this effect. Abstract Understanding the effects of fatigue is a central issue in the context of endurance sports. Given the popularity of running, there are numerous novices among runners. Therefore, understanding the effects of fatigue in novice runners is an important issue. Various studies have drawn conclusions about the control of certain variables by analyzing motor variability. One variable that plays a crucial role during running is the center of mass (CoM), as it reflects the movement of the whole body in a simplified way. Therefore, the aim of this study was to analyze the effects of fatigue on the motor variability structure that stabilizes the CoM trajectory in novice runners. To do so, the uncontrolled manifold approach was applied to a 3D whole-body model using the CoM as the result variable. It was found that motor variability increased with fatigue (UCMꓕ). However, the UCMRatio did not change. This indicates that the control of the CoM decreased, whereas the stability was not affected. The decreases in control were correlated with the degree of exhaustion, as indicated by the Borg scale (during breaking and flight phase). It can be summarized that running-induced fatigue increases the step-to-step variability in novice runners and affects the control of their CoM.

Simple Summary: Endurance sports, and especially running, are very popular. During running, fatigue inevitably occurs, and especially in novices. Surprisingly, the effects of fatigue have been studied less extensively in novice runners compared with experienced runners. Regardless of the level of expertise, it has been shown that, by analyzing motor variability, valuable insights can be gained regarding the control of important variables. Motor variability in running is understood as step-to-step deviations. One example of an important variable is the body's center of mass, as it provides a simplified representation of the overall movement. Thus, an analysis of the motor variability of the body's center of mass might lead to valuable insights. Therefore, the purpose of this study was to investigate the effects of fatigue on the motor variability of the body's center of mass in novice runners. It was found that, with fatigue, the motor variability increased, and the control of the body's center of mass decreased. Moreover, there was a correlation between the decrease in control and the degree of fatigue. Further studies should investigate which training methods can mitigate this effect.
Abstract: Understanding the effects of fatigue is a central issue in the context of endurance sports. Given the popularity of running, there are numerous novices among runners. Therefore, understanding the effects of fatigue in novice runners is an important issue. Various studies have drawn conclusions about the control of certain variables by analyzing motor variability. One variable that plays a crucial role during running is the center of mass (CoM), as it reflects the movement of the whole body in a simplified way. Therefore, the aim of this study was to analyze the effects of fatigue on the motor variability structure that stabilizes the CoM trajectory in novice runners. To do so, the uncontrolled manifold approach was applied to a 3D whole-body model using the CoM as the result variable. It was found that motor variability increased with fatigue (UCM Biology 2022, 11,942 The RV can be stabilized through covariance among individual deviations. This means that a deviation in one a corresponding change in another EV. In this way, movement executions that does not affect the RV; thus, w movement solutions exist. This portion of motor variabil also referred to as UCM‖. It is considered beneficial be flexible movement executions by providing a multit solutions. If changes in an EV are not compensated for by occur. The motor variability that leads to changes in the portion of variability is potentially "unwanted", as it affe is thus interpreted as an indication of a high degree of con the UCM ꓕ , the chosen RV is assumed to be stabilized and the control hypothesis about the RV is accepted. The (UCMRatio) is used to quantify the degree of stabilization [ The UCM has yielded valuable results in a number o the total force applied by the fingers as the RV showed t UCMꓕ increased under fatigue, and therefore increased t In another study that analyzed the gender-specific ef stability during a pointing task, higher UCM‖ and UCM comparison with men in a fatigued state [33]. The UCM times for locomotion studies (e.g., Qu et al. reported a UCMRatio in the frontal plane during walking in a fatigue [28]), which means that the CoM is less stabilized in recovered state. Möhler et al. showed that, even under expert runners was stabilized over the entire gait cycle [26 were unaffected by fatigue, the UCMꓕ increased in the flig al. [11] showed that, at higher speeds, the motor variabi novices than in experts.
The UCM is a promising approach to gain further in on the variability structure that stabilizes the CoM in nov the aim of the present study was to analyze the influence motor variability structure that stabilizes the CoM traje hypothesized that novice runners would lack strategies to state and would not stabilize their CoM trajectories, w decreased UCMRatio and increased UCMꓕ, respectively parameters would correlate with the level of fatigue.

Introduction
The popularity of running has been expanding worldwide for many years. The number of participants in official running competitions increased from less than 2 million in 2001 to more than 7.9 million by 2018 [1]. The goal of runners, whether they are at the novice, recreational or expert level, is to improve their fatigue resistance [2]. However, the mechanisms and effects of fatigue are not yet fully understood [3]. In addition, research The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM . It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCM Biology 2022, 11,942 The RV can be stabilized through covariance among the EVs, which compens individual deviations. This means that a deviation in one EV can be compensated a corresponding change in another EV. In this way, there is variability acr movement executions that does not affect the RV; thus, with respect to the RV, equ movement solutions exist. This portion of motor variability that does not affect th also referred to as UCM‖. It is considered beneficial because it constitutes a so flexible movement executions by providing a multitude of equivalent mo solutions. If changes in an EV are not compensated for by another EV, changes in occur. The motor variability that leads to changes in the RV is referred to as UCM portion of variability is potentially "unwanted", as it affects the RV. A low level of is thus interpreted as an indication of a high degree of control. If the UCM‖ is great the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous s and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the (UCMRatio) is used to quantify the degree of stabilization [23,25].
The UCM has yielded valuable results in a number of studies. Studies that an the total force applied by the fingers as the RV showed that the values of the UC UCMꓕ increased under fatigue, and therefore increased the variability in the EVs In another study that analyzed the gender-specific effects of fatigue on low stability during a pointing task, higher UCM‖ and UCMꓕ were reported for wo comparison with men in a fatigued state [33]. The UCM approach has been used times for locomotion studies (e.g., Qu et al. reported a significantly lower value UCMRatio in the frontal plane during walking in a fatigued state than in a recovere [28]), which means that the CoM is less stabilized in the fatigued state than recovered state. Möhler et al. showed that, even under fatigue, the CoM trajec expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Mö al. [11] showed that, at higher speeds, the motor variability during running is hi novices than in experts.
The UCM is a promising approach to gain further insight into the effects of on the variability structure that stabilizes the CoM in novices during running. The the aim of the present study was to analyze the influence of fatigue on the stride-to motor variability structure that stabilizes the CoM trajectory in novice runners. hypothesized that novice runners would lack strategies to control their CoM in a fa state and would not stabilize their CoM trajectories, which would be indicate decreased UCMRatio and increased UCMꓕ, respectively. The changes in these parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. The de the participants, experimental protocol and data collection are repeated in the fol subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; stature: 1.8 m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; other sports per week: 110 ± 71 min) participated in the study. Exclusion criteria were regular r exercise more than once a month, any injury or pain in the lower limbs within the months prior to data collection and a BMI higher than 25 kg/m 2 . All participants pr written informed consent to voluntarily participate in this study. The study was ap by the ethics committee of the Karlsruhe Institute of Technology.
. This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCM Biology 2022, 11,942 The RV can be stabilized through covariance among the EVs, which com individual deviations. This means that a deviation in one EV can be compen a corresponding change in another EV. In this way, there is variability movement executions that does not affect the RV; thus, with respect to the RV movement solutions exist. This portion of motor variability that does not affe also referred to as UCM‖.
It is considered beneficial because it constitutes flexible movement executions by providing a multitude of equivalent solutions. If changes in an EV are not compensated for by another EV, chang occur. The motor variability that leads to changes in the RV is referred to as portion of variability is potentially "unwanted", as it affects the RV. A low lev is thus interpreted as an indication of a high degree of control. If the UCM‖ is the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nerv and the control hypothesis about the RV is accepted. The ratio of the UCM‖ t (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies th the total force applied by the fingers as the RV showed that the values of the UCMꓕ increased under fatigue, and therefore increased the variability in the In another study that analyzed the gender-specific effects of fatigue on stability during a pointing task, higher UCM‖ and UCMꓕ were reported fo comparison with men in a fatigued state [33]. The UCM approach has been u times for locomotion studies (e.g., Qu et al. reported a significantly lower v UCMRatio in the frontal plane during walking in a fatigued state than in a rec [28]), which means that the CoM is less stabilized in the fatigued state recovered state. Möhler et al. showed that, even under fatigue, the CoM t expert runners was stabilized over the entire gait cycle [26]. While the UCMRat were unaffected by fatigue, the UCMꓕ increased in the flight phase. In additio al. [11] showed that, at higher speeds, the motor variability during running novices than in experts.
The UCM is a promising approach to gain further insight into the effec on the variability structure that stabilizes the CoM in novices during running the aim of the present study was to analyze the influence of fatigue on the stri motor variability structure that stabilizes the CoM trajectory in novice run hypothesized that novice runners would lack strategies to control their CoM i state and would not stabilize their CoM trajectories, which would be ind decreased UCMRatio and increased UCMꓕ, respectively. The changes in parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. Th the participants, experimental protocol and data collection are repeated in th subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; stature m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; other sp per week: 110 ± 71 min) participated in the study. Exclusion criteria were regu exercise more than once a month, any injury or pain in the lower limbs within months prior to data collection and a BMI higher than 25 kg/m 2 . All participan written informed consent to voluntarily participate in this study. The study wa by the ethics committee of the Karlsruhe Institute of Technology.
is thus interpreted as an indication of a high degree of control. If the UCM is greater than the UCM Biology 2022, 11,942 The RV can be stabilized through covariance among the E individual deviations. This means that a deviation in one EV c a corresponding change in another EV. In this way, there movement executions that does not affect the RV; thus, with re movement solutions exist. This portion of motor variability tha also referred to as UCM‖.
It is considered beneficial because flexible movement executions by providing a multitude solutions. If changes in an EV are not compensated for by anot occur. The motor variability that leads to changes in the RV is portion of variability is potentially "unwanted", as it affects the is thus interpreted as an indication of a high degree of control. I the UCM ꓕ , the chosen RV is assumed to be stabilized by th and the control hypothesis about the RV is accepted. The ratio (UCMRatio) is used to quantify the degree of stabilization [23,25] The UCM has yielded valuable results in a number of stud the total force applied by the fingers as the RV showed that th UCMꓕ increased under fatigue, and therefore increased the va In another study that analyzed the gender-specific effects stability during a pointing task, higher UCM‖ and UCMꓕ we comparison with men in a fatigued state [33]. The UCM appro times for locomotion studies (e.g., Qu et al. reported a signif UCMRatio in the frontal plane during walking in a fatigued stat [28]), which means that the CoM is less stabilized in the f recovered state. Möhler et al. showed that, even under fatig expert runners was stabilized over the entire gait cycle [26]. Wh were unaffected by fatigue, the UCMꓕ increased in the flight ph al. [11] showed that, at higher speeds, the motor variability d novices than in experts.
The UCM is a promising approach to gain further insigh on the variability structure that stabilizes the CoM in novices d the aim of the present study was to analyze the influence of fat motor variability structure that stabilizes the CoM trajectory hypothesized that novice runners would lack strategies to cont state and would not stabilize their CoM trajectories, which decreased UCMRatio and increased UCMꓕ, respectively. The parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published the participants, experimental protocol and data collection are subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4. m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 per week: 110 ± 71 min) participated in the study. Exclusion cri exercise more than once a month, any injury or pain in the low months prior to data collection and a BMI higher than 25 kg/m 2 written informed consent to voluntarily participate in this study by the ethics committee of the Karlsruhe Institute of Technolog , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM to the UCM Biology 2022, 11, 942 3 of The RV can be stabilized through covariance among the EVs, which compensates f individual deviations. This means that a deviation in one EV can be compensated for b a corresponding change in another EV. In this way, there is variability across th movement executions that does not affect the RV; thus, with respect to the RV, equivale movement solutions exist. This portion of motor variability that does not affect the RV also referred to as UCM‖.
It is considered beneficial because it constitutes a source flexible movement executions by providing a multitude of equivalent moveme solutions. If changes in an EV are not compensated for by another EV, changes in the R occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ.
Th portion of variability is potentially "unwanted", as it affects the RV. A low level of UCM is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater tha the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCM (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyze the total force applied by the fingers as the RV showed that the values of the UCM‖ an UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32 In another study that analyzed the gender-specific effects of fatigue on lower-lim stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women comparison with men in a fatigued state [33]. The UCM approach has been used sever times for locomotion studies (e.g., Qu et al. reported a significantly lower value of th UCMRatio in the frontal plane during walking in a fatigued state than in a recovered sta [28]), which means that the CoM is less stabilized in the fatigued state than in th recovered state. Möhler et al. showed that, even under fatigue, the CoM trajectory expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and UCM were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Möhler al. [11] showed that, at higher speeds, the motor variability during running is higher novices than in experts.
The UCM is a promising approach to gain further insight into the effects of fatigu on the variability structure that stabilizes the CoM in novices during running. Therefor the aim of the present study was to analyze the influence of fatigue on the stride-to-strid motor variability structure that stabilizes the CoM trajectory in novice runners. It w hypothesized that novice runners would lack strategies to control their CoM in a fatigue state and would not stabilize their CoM trajectories, which would be indicated by decreased UCMRatio and increased UCMꓕ, respectively. The changes in these UC parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. The details the participants, experimental protocol and data collection are repeated in the followin subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; stature: 1.82 ± 0. m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; other sports activi per week: 110 ± 71 min) participated in the study. Exclusion criteria were regular runnin exercise more than once a month, any injury or pain in the lower limbs within the last s months prior to data collection and a BMI higher than 25 kg/m 2 . All participants provide written informed consent to voluntarily participate in this study. The study was approve by the ethics committee of the Karlsruhe Institute of Technology.
(UCM Ratio ) is used to quantify the degree of stabilization [23,25].
The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM and UCM The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖.
It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ. This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the UCMRatio in the frontal plane during walking in a fatigued state than in a recovered state [28]), which means that the CoM is less stabilized in the fatigued state than in the recovered state. Möhler et al. showed that, even under fatigue, the CoM trajectory of expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and UCM‖ were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Möhler et al. [11] showed that, at higher speeds, the motor variability during running is higher in novices than in experts.
The UCM is a promising approach to gain further insight into the effects of fatigue on the variability structure that stabilizes the CoM in novices during running. Therefore, the aim of the present study was to analyze the influence of fatigue on the stride-to-stride motor variability structure that stabilizes the CoM trajectory in novice runners. It was hypothesized that novice runners would lack strategies to control their CoM in a fatigued state and would not stabilize their CoM trajectories, which would be indicated by a decreased UCMRatio and increased UCMꓕ, respectively. The changes in these UCM parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. The details of the participants, experimental protocol and data collection are repeated in the following subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; stature: 1.82 ± 0.06 m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; other sports activity per week: 110 ± 71 min) participated in the study. Exclusion criteria were regular running exercise more than once a month, any injury or pain in the lower limbs within the last six months prior to data collection and a BMI higher than 25 kg/m 2 . All participants provided written informed consent to voluntarily participate in this study. The study was approved by the ethics committee of the Karlsruhe Institute of Technology.
increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM and UCM Biology 2022, 11,942 The RV can be stabilized through covariance among the EVs, which compensa individual deviations. This means that a deviation in one EV can be compensated a corresponding change in another EV. In this way, there is variability acro movement executions that does not affect the RV; thus, with respect to the RV, equi movement solutions exist. This portion of motor variability that does not affect the also referred to as UCM‖.
It is considered beneficial because it constitutes a sou flexible movement executions by providing a multitude of equivalent mov solutions. If changes in an EV are not compensated for by another EV, changes in t occur. The motor variability that leads to changes in the RV is referred to as UCM portion of variability is potentially "unwanted", as it affects the RV. A low level of is thus interpreted as an indication of a high degree of control. If the UCM‖ is greate the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous s and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that an the total force applied by the fingers as the RV showed that the values of the UCM UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [ In another study that analyzed the gender-specific effects of fatigue on lowe stability during a pointing task, higher UCM‖ and UCMꓕ were reported for wom comparison with men in a fatigued state [33]. The UCM approach has been used s times for locomotion studies (e.g., Qu et al. reported a significantly lower value UCMRatio in the frontal plane during walking in a fatigued state than in a recovered [28]), which means that the CoM is less stabilized in the fatigued state than recovered state. Möhler et al. showed that, even under fatigue, the CoM traject expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Mö al. [11] showed that, at higher speeds, the motor variability during running is hig novices than in experts.
The UCM is a promising approach to gain further insight into the effects of f on the variability structure that stabilizes the CoM in novices during running. The the aim of the present study was to analyze the influence of fatigue on the stride-to motor variability structure that stabilizes the CoM trajectory in novice runners. hypothesized that novice runners would lack strategies to control their CoM in a fa state and would not stabilize their CoM trajectories, which would be indicated decreased UCMRatio and increased UCMꓕ, respectively. The changes in these parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. The det the participants, experimental protocol and data collection are repeated in the foll subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; stature: 1.82 m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; other sports a per week: 110 ± 71 min) participated in the study. Exclusion criteria were regular ru exercise more than once a month, any injury or pain in the lower limbs within the l months prior to data collection and a BMI higher than 25 kg/m 2 . All participants pro written informed consent to voluntarily participate in this study. The study was app by the ethics committee of the Karlsruhe Institute of Technology.
were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the UCM Ratio in the frontal plane during walking in a fatigued state than in a recovered state [28]), which means that the CoM is less stabilized in the fatigued state than in the recovered state. Möhler et al. showed that, even under fatigue, the CoM trajectory of expert runners was stabilized over the entire gait cycle [26]. While the UCM Ratio and UCM were unaffected by fatigue, the UCM The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖.
It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ.
This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the UCMRatio in the frontal plane during walking in a fatigued state than in a recovered state [28]), which means that the CoM is less stabilized in the fatigued state than in the recovered state. Möhler et al. showed that, even under fatigue, the CoM trajectory of expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and UCM‖ were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Möhler et al. [11] showed that, at higher speeds, the motor variability during running is higher in novices than in experts.
The UCM is a promising approach to gain further insight into the effects of fatigue on the variability structure that stabilizes the CoM in novices during running. Therefore, the aim of the present study was to analyze the influence of fatigue on the stride-to-stride motor variability structure that stabilizes the CoM trajectory in novice runners. It was hypothesized that novice runners would lack strategies to control their CoM in a fatigued state and would not stabilize their CoM trajectories, which would be indicated by a decreased UCMRatio and increased UCMꓕ, respectively. The changes in these UCM parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. The details of the participants, experimental protocol and data collection are repeated in the following subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; stature: 1.82 ± 0.06 m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; other sports activity per week: 110 ± 71 min) participated in the study. Exclusion criteria were regular running exercise more than once a month, any injury or pain in the lower limbs within the last six months prior to data collection and a BMI higher than 25 kg/m 2 . All participants provided written informed consent to voluntarily participate in this study. The study was approved by the ethics committee of the Karlsruhe Institute of Technology.
increased in the flight phase. In addition, Möhler et al. [11] showed that, at higher speeds, the motor variability during running is higher in novices than in experts.
The UCM is a promising approach to gain further insight into the effects of fatigue on the variability structure that stabilizes the CoM in novices during running. Therefore, the aim of the present study was to analyze the influence of fatigue on the stride-tostride motor variability structure that stabilizes the CoM trajectory in novice runners. It was hypothesized that novice runners would lack strategies to control their CoM in a fatigued state and would not stabilize their CoM trajectories, which would be indicated by a decreased UCM Ratio and increased UCM Biology 2022, 11, 942 3 The RV can be stabilized through covariance among the EVs, which compensate individual deviations. This means that a deviation in one EV can be compensated fo a corresponding change in another EV. In this way, there is variability across movement executions that does not affect the RV; thus, with respect to the RV, equiv movement solutions exist. This portion of motor variability that does not affect the R also referred to as UCM‖.
It is considered beneficial because it constitutes a sour flexible movement executions by providing a multitude of equivalent move solutions. If changes in an EV are not compensated for by another EV, changes in th occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ. portion of variability is potentially "unwanted", as it affects the RV. A low level of U is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous sys and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the U (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that anal the total force applied by the fingers as the RV showed that the values of the UCM‖ UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31 In another study that analyzed the gender-specific effects of fatigue on lowerstability during a pointing task, higher UCM‖ and UCMꓕ were reported for wom comparison with men in a fatigued state [33]. The UCM approach has been used se times for locomotion studies (e.g., Qu et al. reported a significantly lower value o UCMRatio in the frontal plane during walking in a fatigued state than in a recovered [28]), which means that the CoM is less stabilized in the fatigued state than in recovered state. Möhler et al. showed that, even under fatigue, the CoM trajecto expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and U were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Möh al. [11] showed that, at higher speeds, the motor variability during running is high novices than in experts.
The UCM is a promising approach to gain further insight into the effects of fa on the variability structure that stabilizes the CoM in novices during running. There the aim of the present study was to analyze the influence of fatigue on the stride-to-s motor variability structure that stabilizes the CoM trajectory in novice runners. It hypothesized that novice runners would lack strategies to control their CoM in a fati state and would not stabilize their CoM trajectories, which would be indicated decreased UCMRatio and increased UCMꓕ, respectively. The changes in these U parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. The deta the participants, experimental protocol and data collection are repeated in the follo subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; stature: 1.82 ± m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; other sports act per week: 110 ± 71 min) participated in the study. Exclusion criteria were regular run exercise more than once a month, any injury or pain in the lower limbs within the la , respectively. The changes in these UCM parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. The details of the participants, experimental protocol and data collection are repeated in the following subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; stature: 1.82 ± 0.06 m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; other sports activity per week: 110 ± 71 min) participated in the study. Exclusion criteria were regular running exercise more than once a month, any injury or pain in the lower limbs within the last six months prior to data collection and a BMI higher than 25 kg/m 2 . All participants provided written informed consent to voluntarily participate in this study. The study was approved by the ethics committee of the Karlsruhe Institute of Technology.

Experimental Protocol
The measurements were conducted in the Biomechanics Laboratory of the BioMotion Center at the Institute of Sports and Sports Science of the Karlsruhe Institute of Technology. The participants ran on a treadmill (h/p/cosmos Saturn, Nussdorf-Traunstein, Germany) with a slope of 1% [34]. The participants were first familiarized with the treadmill by walking at a speed of 5 km/h for 6 min [35] and running at a speed of 8 km/h for 6 min [36], followed by 10 s of running at a speed of 13 km/h. After the treadmill familiarization, the subjects had two minutes to recover before running at a fixed speed of 13 km/h until subjective exhaustion. The speed of 13 km/h was chosen by experience as a speed that would challenge novice runners sufficiently to lead to exhaustion after several minutes. Participants were instructed to look straight ahead and to not perform undesired movements. To prevent falls, all participants wore a safety harness during the experiment. The treadmill was stopped immediately when the participants indicated exhaustion. After having indicated exhaustion and after the treadmill stopped, the participants were asked to rate their fatigue on the Borg scale [37].

Data Collection and Processing
First, 22 anthropometric measures were manually taken from each participant, and 42 reflective markers were attached to participants' skin, in accordance with the ALASKA modeling system (Advanced Lagrangian Solver in Kinetic Analysis, INSYS GmbH, Chemnitz, Germany; [38]). During the treadmill protocol, 16 Vicon cameras (Vicon Motion Systems; Oxford Metrics Group, Oxford, UK) recorded the subjects' kinematics, with a recording frequency of 200 Hz.
Data recording started 10 s after the treadmill reached the speed of 13 km/h and lasted until the treadmill stopped. The marker data were preprocessed using Vicon Nexus V2.11.0 software and filtered with a second-order low-pass Butterworth filter, with a cutoff frequency of 15 Hz, using MATLAB R2020b (The MathWorks, Natick, MA, USA). The anthropometric measurements (22 measured manually, 43 determined from the reflective markers, according to the requirements of the ALASKA modeling system) and the marker trajectories were used to calculate joint angles using inverse kinematics with the full-body Dynamicus model (ALASKA, INSYS GmbH, Chemnitz, Germany; [38]).
From the recorded data, the first 35 gait cycles represented the PRE condition (rested state), and the last 35 gait cycles represented the POST condition (fatigued state). Gait cycles were detected based on the sign change of the heel or forefoot marker and the vertical acceleration of the toe marker at initial contact and toe off, respectively [39]. For the subsequent UCM analysis, each gait cycle was time-normalized to 101 data points using cubic spline interpolation.

Uncontrolled Manifold Approach
Variability analyses may be sensitive to the number of gait cycles included [40,41]. Considering the application of UCM purely in terms of the mathematical aspect, more movement repetitions would provide more reliable results [40]. However, in the present case, the fatigue effects might interfere when including too many cycles. Therefore, N cycle = 35 was chosen as the number of gait cycles to be included in the present analysis.
The necessary steps for the UCM approach were first explained by Scholz and Schöner [24]. The calculations used in this study, with specific consideration to the 3D CoM, are outlined in the following section.
To calculate the CoM based on the EVs, a subject-specific anthropometric 3D model consisting of 17 segments and 50 degrees of freedom (47 segmental angles and 3 hip rotations [26]) was used. The whole-body CoM (r CoM ) was calculated as a weighted sum of the body segments (e.g., [46]), as in Equation (1): where N is the number of segments; V i is the volume of the ith segment; and r i is the center-of-gravity vector of the ith segment. This model enabled changes in joint angles (Θ, EV) to be linked to changes in the CoM trajectory (r CoM , RV), where the RV was expressed as a function of the EVs, as in Equation (2). Because the EVs in a UCM model must have the same unit, only joint angles were used, instead of different quantities (e.g., joint angles combined with kinematic variables; see also [47]): The space in which changes in the EVs do not cause changes in the RV corresponds to the null space of the linearized Jacobian. According to the UCM approach [24], it is calculated as in Equation (3): In the next step, deviations from Θ 0 were separated into those parallel to the UCM (stabilizing the RV (σ k, ), Equation (4)): and those orthogonal to the UCM (changing the RV (σ k, Biology 2022, 11,942 The RV can be stabilized through covariance among the EVs, whic individual deviations. This means that a deviation in one EV can be co a corresponding change in another EV. In this way, there is var movement executions that does not affect the RV; thus, with respect to movement solutions exist. This portion of motor variability that does n also referred to as UCM‖. It is considered beneficial because it cons flexible movement executions by providing a multitude of equi solutions. If changes in an EV are not compensated for by another EV, occur. The motor variability that leads to changes in the RV is referred portion of variability is potentially "unwanted", as it affects the RV. A is thus interpreted as an indication of a high degree of control. If the UC the UCM ꓕ , the chosen RV is assumed to be stabilized by the centra and the control hypothesis about the RV is accepted. The ratio of the U (UCMRatio) is used to quantify the degree of stabilization [23,25].
The UCM has yielded valuable results in a number of studies. Stu the total force applied by the fingers as the RV showed that the values UCMꓕ increased under fatigue, and therefore increased the variability In another study that analyzed the gender-specific effects of fatig stability during a pointing task, higher UCM‖ and UCMꓕ were repor comparison with men in a fatigued state [33]. The UCM approach has times for locomotion studies (e.g., Qu et al. reported a significantly l UCMRatio in the frontal plane during walking in a fatigued state than in [28]), which means that the CoM is less stabilized in the fatigued recovered state. Möhler et al. showed that, even under fatigue, the expert runners was stabilized over the entire gait cycle [26]. While the U were unaffected by fatigue, the UCMꓕ increased in the flight phase. In a al. [11] showed that, at higher speeds, the motor variability during ru novices than in experts.
The UCM is a promising approach to gain further insight into th on the variability structure that stabilizes the CoM in novices during r the aim of the present study was to analyze the influence of fatigue on motor variability structure that stabilizes the CoM trajectory in novi hypothesized that novice runners would lack strategies to control their state and would not stabilize their CoM trajectories, which would decreased UCMRatio and increased UCMꓕ, respectively. The chang parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [ the participants, experimental protocol and data collection are repeate subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; ot per week: 110 ± 71 min) participated in the study. Exclusion criteria we exercise more than once a month, any injury or pain in the lower limbs months prior to data collection and a BMI higher than 25 kg/m 2 . All par written informed consent to voluntarily participate in this study. The st by the ethics committee of the Karlsruhe Institute of Technology.
), Equation (5)): The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖.
It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ.
This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the UCMRatio in the frontal plane during walking in a fatigued state than in a recovered state [28]), which means that the CoM is less stabilized in the fatigued state than in the recovered state. Möhler et al. showed that, even under fatigue, the CoM trajectory of expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and UCM‖ were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Möhler et al. [11] showed that, at higher speeds, the motor variability during running is higher in novices than in experts.
The UCM is a promising approach to gain further insight into the effects of fatigue on the variability structure that stabilizes the CoM in novices during running. Therefore, the aim of the present study was to analyze the influence of fatigue on the stride-to-stride motor variability structure that stabilizes the CoM trajectory in novice runners. It was hypothesized that novice runners would lack strategies to control their CoM in a fatigued state and would not stabilize their CoM trajectories, which would be indicated by a decreased UCMRatio and increased UCMꓕ, respectively. The changes in these UCM parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. The details of the participants, experimental protocol and data collection are repeated in the following subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; stature: 1.82 ± 0.06 m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; other sports activity per week: 110 ± 71 min) participated in the study. Exclusion criteria were regular running exercise more than once a month, any injury or pain in the lower limbs within the last six months prior to data collection and a BMI higher than 25 kg/m 2 . All participants provided written informed consent to voluntarily participate in this study. The study was approved by the ethics committee of the Karlsruhe Institute of Technology.
where k = 1, . . . , N cycle and N cycle is the number of included gait cycles (here: N cycle = 35). These calculations were performed for each time point of the normalized gait cycle, which led to a total of 101 UCM calculations. Subsequently, the variabilities parallel and orthogonal to the UCM were calculated as the variance over the 35 gait cycles, as in Equation (6) and Equation (7), respectively: and UCM Biology 2022, 11, 942 3 of 12 The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖.
It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ. This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the UCMRatio in the frontal plane during walking in a fatigued state than in a recovered state [28]), which means that the CoM is less stabilized in the fatigued state than in the recovered state. Möhler et al. showed that, even under fatigue, the CoM trajectory of expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and UCM‖ were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Möhler et al. [11] showed that, at higher speeds, the motor variability during running is higher in novices than in experts.
The UCM is a promising approach to gain further insight into the effects of fatigue on the variability structure that stabilizes the CoM in novices during running. Therefore, the aim of the present study was to analyze the influence of fatigue on the stride-to-stride motor variability structure that stabilizes the CoM trajectory in novice runners. It was hypothesized that novice runners would lack strategies to control their CoM in a fatigued state and would not stabilize their CoM trajectories, which would be indicated by a decreased UCMRatio and increased UCMꓕ, respectively. The changes in these UCM parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. The details of the participants, experimental protocol and data collection are repeated in the following subsections.
The ratio between these two quantities was calculated: The UCM Ratio quantifies the degree of stabilization of the RV (in this study, the CoM). It has a theoretical range of [−1, 1], where a positive value indicates a stable state, and vice versa [42,43].

Statistics
Statistical analyses were performed using MATLAB R2020b (The MathWorks, Natick, MA, USA). First, all data were checked for normal distribution using the Shapiro-Wilk test [48]. The time courses were tested for differences between the PRE and POST for each dependent parameter (UCM , UCM The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖. It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ. This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the and UCM Ratio ) using a dependent t-test with statistical parametric mapping (SPM) from the spm1d toolbox [49]. The significance level was set at p = 0.05. Afterwards, the entire gait cycle was divided into two flight phases (FPs), and the left and right stance phases, according to the detected gait events [39]; both stance phases were divided into a breaking phase (BP) and a propulsion phase (PP) [50]. Right and left flight phases, as well as right and left breaking and propulsion phases, were then averaged to calculate one value for each phase in the following abbreviations: FP, BP and PP. Afterwards, the changes from PRE to POST (∆ Pre-Post ) in the mean values of each UCM parameter were calculated by subtracting the mean value of POST from the mean value of PRE for the three phases of the gait cycles: FP, BP and PP. Spearman correlations between the ∆ Pre-Post and the Borg-scale rating were separately calculated for each UCM parameter and each phase of the gait cycle individually. These calculations were performed as a cross-check to assess whether the UCM parameters can reflect the effects of fatigue captured by the Borg-scale ratings.

Results
The participants maintained the speed of 13 km/h for 6.18 ± 2.45 min. Their exhaustion was confirmed by a Borg-scale rating of 18.7 ± 1.0 (scale ranging from 6 to 20), which corresponds to "very very hard" in terms of the difficulty level.
The UCM analysis showed that both the UCM Biology 2022, 11,942 The RV can be stabilized through covariance among the EVs, which com individual deviations. This means that a deviation in one EV can be compen a corresponding change in another EV. In this way, there is variability movement executions that does not affect the RV; thus, with respect to the RV movement solutions exist. This portion of motor variability that does not affe also referred to as UCM‖. It is considered beneficial because it constitutes flexible movement executions by providing a multitude of equivalent solutions. If changes in an EV are not compensated for by another EV, chang occur. The motor variability that leads to changes in the RV is referred to as portion of variability is potentially "unwanted", as it affects the RV. A low le is thus interpreted as an indication of a high degree of control. If the UCM‖ is the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nerv and the control hypothesis about the RV is accepted. The ratio of the UCM‖ t (UCMRatio) is used to quantify the degree of stabilization [23,25].
The UCM has yielded valuable results in a number of studies. Studies th the total force applied by the fingers as the RV showed that the values of the UCMꓕ increased under fatigue, and therefore increased the variability in the In another study that analyzed the gender-specific effects of fatigue on stability during a pointing task, higher UCM‖ and UCMꓕ were reported fo comparison with men in a fatigued state [33]. The UCM approach has been u times for locomotion studies (e.g., Qu et al. reported a significantly lower UCMRatio in the frontal plane during walking in a fatigued state than in a rec [28]), which means that the CoM is less stabilized in the fatigued state recovered state. Möhler et al. showed that, even under fatigue, the CoM expert runners was stabilized over the entire gait cycle [26]. While the UCMRat were unaffected by fatigue, the UCMꓕ increased in the flight phase. In additio al. [11] showed that, at higher speeds, the motor variability during running novices than in experts.
The UCM is a promising approach to gain further insight into the effec on the variability structure that stabilizes the CoM in novices during running the aim of the present study was to analyze the influence of fatigue on the str motor variability structure that stabilizes the CoM trajectory in novice run hypothesized that novice runners would lack strategies to control their CoM i state and would not stabilize their CoM trajectories, which would be ind decreased UCMRatio and increased UCMꓕ, respectively. The changes in parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. T the participants, experimental protocol and data collection are repeated in th subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; stature m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; other sp per week: 110 ± 71 min) participated in the study. Exclusion criteria were regu exercise more than once a month, any injury or pain in the lower limbs withi months prior to data collection and a BMI higher than 25 kg/m 2 . All participan written informed consent to voluntarily participate in this study. The study w by the ethics committee of the Karlsruhe Institute of Technology. and UCM were higher in the POST state (see Figure 1) throughout the gait cycle. However, the differences concerning the UCM did not reach the level of statistical significance, whereas the differences for the UCM Biology 2022, 11, 942 3 of 12 The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖.
It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ.
This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the UCMRatio in the frontal plane during walking in a fatigued state than in a recovered state [28]), which means that the CoM is less stabilized in the fatigued state than in the recovered state. Möhler et al. showed that, even under fatigue, the CoM trajectory of expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and UCM‖ were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Möhler et al. [11] showed that, at higher speeds, the motor variability during running is higher in novices than in experts.
The UCM is a promising approach to gain further insight into the effects of fatigue on the variability structure that stabilizes the CoM in novices during running. Therefore, the aim of the present study was to analyze the influence of fatigue on the stride-to-stride motor variability structure that stabilizes the CoM trajectory in novice runners. It was hypothesized that novice runners would lack strategies to control their CoM in a fatigued state and would not stabilize their CoM trajectories, which would be indicated by a decreased UCMRatio and increased UCMꓕ, respectively. The changes in these UCM parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. The details of the participants, experimental protocol and data collection are repeated in the following subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; stature: 1.82 ± 0.06 m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; other sports activity per week: 110 ± 71 min) participated in the study. Exclusion criteria were regular running exercise more than once a month, any injury or pain in the lower limbs within the last six months prior to data collection and a BMI higher than 25 kg/m 2 . All participants provided written informed consent to voluntarily participate in this study. The study was approved by the ethics committee of the Karlsruhe Institute of Technology. did (see Figure 2). The UCM Ratio did not show any statistically significant changes throughout the gait cycle between the PRE and POST. In both states, it was greater than zero throughout the entire gait cycle.
was set at p = 0.05. Afterwards, the entire gait cycle was divided into two flig (FPs), and the left and right stance phases, according to the detected gait events stance phases were divided into a breaking phase (BP) and a propulsion phase Right and left flight phases, as well as right and left breaking and propulsion ph then averaged to calculate one value for each phase in the following abbreviatio and PP. Afterwards, the changes from PRE to POST (ΔPre-Post) in the mean valu UCM parameter were calculated by subtracting the mean value of POST from value of PRE for the three phases of the gait cycles: FP, BP and PP. Spearman c between the ΔPre-Post and the Borg-scale rating were separately calculated for parameter and each phase of the gait cycle individually. These calculat performed as a cross-check to assess whether the UCM parameters can reflect of fatigue captured by the Borg-scale ratings.

Results
The participants maintained the speed of 13 km/h for 6.18 ± 2.45 m exhaustion was confirmed by a Borg-scale rating of 18.7 ± 1.0 (scale ranging fro which corresponds to "very very hard" in terms of the difficulty level. The UCM analysis showed that both the UCM ꓕ and UCM‖ were higher in state (see Figure 1) throughout the gait cycle. However, the differences conc UCM‖ did not reach the level of statistical significance, whereas the differen UCM ꓕ did (see Figure 2). The UCMRatio did not show any statistically significa throughout the gait cycle between the PRE and POST. In both states, it was gr zero throughout the entire gait cycle.  The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖.
It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ. This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the UCMRatio in the frontal plane during walking in a fatigued state than in a recovered state [28]), which means that the CoM is less stabilized in the fatigued state than in the recovered state. Möhler et al. showed that, even under fatigue, the CoM trajectory of expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and UCM‖ were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Möhler et al. [11] showed that, at higher speeds, the motor variability during running is higher in novices than in experts.
The UCM is a promising approach to gain further insight into the effects of fatigue on the variability structure that stabilizes the CoM in novices during running. Therefore, the aim of the present study was to analyze the influence of fatigue on the stride-to-stride motor variability structure that stabilizes the CoM trajectory in novice runners. It was hypothesized that novice runners would lack strategies to control their CoM in a fatigued state and would not stabilize their CoM trajectories, which would be indicated by a decreased UCMRatio and increased UCMꓕ, respectively. The changes in these UCM parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20]. The details of the participants, experimental protocol and data collection are repeated in the following subsections. To investigate the relationship between the changes in the UCM parameters from the PRE to the POST (∆ Pre-Post ) and the level of exhaustion, Spearman correlations (ρ) were calculated. Table 1  The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖.

Participants
It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ. This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the UCMRatio in the frontal plane during walking in a fatigued state than in a recovered state [28]), which means that the CoM is less stabilized in the fatigued state than in the recovered state. Möhler et al. showed that, even under fatigue, the CoM trajectory of expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and UCM‖ were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Möhler et al. [11] showed that, at higher speeds, the motor variability during running is higher in novices than in experts.
The UCM is a promising approach to gain further insight into the effects of fatigue on the variability structure that stabilizes the CoM in novices during running. Therefore, the aim of the present study was to analyze the influence of fatigue on the stride-to-stride motor variability structure that stabilizes the CoM trajectory in novice runners. It was hypothesized that novice runners would lack strategies to control their CoM in a fatigued state and would not stabilize their CoM trajectories, which would be indicated by a decreased UCMRatio and increased UCMꓕ, respectively. The changes in these UCM parameters would correlate with the level of fatigue.
for the BP and FP correlated with the Borg scale, with p values of 0.055 and 0.032, respectively. The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖.
It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ. This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu  The RV can be stabilized throu individual deviations. This means a corresponding change in anoth movement executions that does no movement solutions exist. This por also referred to as UCM‖. It is co flexible movement executions by solutions. If changes in an EV are n occur. The motor variability that le portion of variability is potentially is thus interpreted as an indication the UCM ꓕ , the chosen RV is as and the control hypothesis about th (UCMRatio) is used to quantify the d The UCM has yielded valuabl the total force applied by the finge UCMꓕ increased under fatigue, and In another study that analyzed t stability during a pointing task, h comparison with men in a fatigued times for locomotion studies (e.g., UCMRatio in the frontal plane durin [28]), which means that the CoM recovered state. Möhler et al. sho expert runners was stabilized over were unaffected by fatigue, the UC al. [11] showed that, at higher spe novices than in experts.
The UCM is a promising appr on the variability structure that sta the aim of the present study was to motor variability structure that sta hypothesized that novice runners w state and would not stabilize the decreased UCMRatio and increase parameters would correlate with th

Materials and Methods
This study reanalyzes the dat the participants, experimental prot subsections.

Discussion
The aim of our study was to analyze the influence of fatigue on the stride-to-stride motor variability structure that stabilizes the CoM trajectory in novice runners. Therefore, a UCM analysis was performed in which the CoM was chosen as the RV. The results revealed significant increases in the UCM Biology 2022, 11, 942 3 of 12 The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖.
It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ. This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the UCMRatio in the frontal plane during walking in a fatigued state than in a recovered state [28]), which means that the CoM is less stabilized in the fatigued state than in the recovered state. Möhler et al. showed that, even under fatigue, the CoM trajectory of expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and UCM‖ were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Möhler et al. [11] showed that, at higher speeds, the motor variability during running is higher in novices than in experts.
The UCM is a promising approach to gain further insight into the effects of fatigue The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖.
It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ. This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the UCMRatio in the frontal plane during walking in a fatigued state than in a recovered state [28]), which means that the CoM is less stabilized in the fatigued state than in the recovered state. Möhler et al. showed that, even under fatigue, the CoM trajectory of expert runners was stabilized over the entire gait cycle [26]. While the UCMRatio and UCM‖ were unaffected by fatigue, the UCMꓕ increased in the flight phase. In addition, Möhler et al. [11] showed that, at higher speeds, the motor variability during running is higher in novices than in experts.
can be interpreted as decreased control over the CoM [24]. However, this decrease did not affect the stability of the CoM, as revealed by the unaffected UCM Ratio . The fact that the UCM Ratio was constantly above zero shows that the CoM is controlled throughout the whole gait cycle in both the PRE and POST states. The variability in terms of the UCM The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖.
It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ. This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the UCMRatio in the frontal plane during walking in a fatigued state than in a recovered state was increased in the POST state, which is in line with previous studies [3,26]. Furthermore, the increases in the UCM The RV can be stabilized through covariance among the EVs, which compensates for individual deviations. This means that a deviation in one EV can be compensated for by a corresponding change in another EV. In this way, there is variability across the movement executions that does not affect the RV; thus, with respect to the RV, equivalent movement solutions exist. This portion of motor variability that does not affect the RV is also referred to as UCM‖. It is considered beneficial because it constitutes a source of flexible movement executions by providing a multitude of equivalent movement solutions. If changes in an EV are not compensated for by another EV, changes in the RV occur. The motor variability that leads to changes in the RV is referred to as UCMꓕ. This portion of variability is potentially "unwanted", as it affects the RV. A low level of UCMꓕ is thus interpreted as an indication of a high degree of control. If the UCM‖ is greater than the UCM ꓕ , the chosen RV is assumed to be stabilized by the central nervous system, and the control hypothesis about the RV is accepted. The ratio of the UCM‖ to the UCMꓕ (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studies that analyzed the total force applied by the fingers as the RV showed that the values of the UCM‖ and UCMꓕ increased under fatigue, and therefore increased the variability in the EVs [31,32]. In another study that analyzed the gender-specific effects of fatigue on lower-limb stability during a pointing task, higher UCM‖ and UCMꓕ were reported for women in comparison with men in a fatigued state [33]. The UCM approach has been used several times for locomotion studies (e.g., Qu et al. reported a significantly lower value of the correlated with the Borg-scale ratings in the BP and FP (p = 0.055 and p = 0.032, respectively), which indicated the plausibility of using UCM analysis for analyzing the control of the CoM trajectory under fatigue.
In our previous study [20], we showed that novice runners had a lower CoM position around the heel strike and increased RoM of the CoM in the medio-lateral direction in the POST state. Combined with the results from this study, it may be concluded that the decreased control of the CoM might be the reason for the changes in the CoM trajectory. Furthermore, in the present study, increases in the UCM Biology 2022, 11,942 The RV can be stabilized through covariance among the EVs, which individual deviations. This means that a deviation in one EV can be com a corresponding change in another EV. In this way, there is varia movement executions that does not affect the RV; thus, with respect to th movement solutions exist. This portion of motor variability that does no also referred to as UCM‖. It is considered beneficial because it constit flexible movement executions by providing a multitude of equiv solutions. If changes in an EV are not compensated for by another EV, c occur. The motor variability that leads to changes in the RV is referred portion of variability is potentially "unwanted", as it affects the RV. A lo is thus interpreted as an indication of a high degree of control. If the UCM the UCM ꓕ , the chosen RV is assumed to be stabilized by the central and the control hypothesis about the RV is accepted. The ratio of the UC (UCMRatio) is used to quantify the degree of stabilization [23,25].
The UCM has yielded valuable results in a number of studies. Stud the total force applied by the fingers as the RV showed that the values o UCMꓕ increased under fatigue, and therefore increased the variability in In another study that analyzed the gender-specific effects of fatigu stability during a pointing task, higher UCM‖ and UCMꓕ were reporte comparison with men in a fatigued state [33]. The UCM approach has b times for locomotion studies (e.g., Qu et al. reported a significantly lo UCMRatio in the frontal plane during walking in a fatigued state than in [28]), which means that the CoM is less stabilized in the fatigued s recovered state. Möhler et al. showed that, even under fatigue, the C expert runners was stabilized over the entire gait cycle [26]. While the UC were unaffected by fatigue, the UCMꓕ increased in the flight phase. In ad al. [11] showed that, at higher speeds, the motor variability during run novices than in experts.
The UCM is a promising approach to gain further insight into the on the variability structure that stabilizes the CoM in novices during run the aim of the present study was to analyze the influence of fatigue on th motor variability structure that stabilizes the CoM trajectory in novice hypothesized that novice runners would lack strategies to control their C state and would not stabilize their CoM trajectories, which would b decreased UCMRatio and increased UCMꓕ, respectively. The changes parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [2 the participants, experimental protocol and data collection are repeated subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; st m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; oth per week: 110 ± 71 min) participated in the study. Exclusion criteria were exercise more than once a month, any injury or pain in the lower limbs w months prior to data collection and a BMI higher than 25 kg/m 2 . All parti written informed consent to voluntarily participate in this study. The stud by the ethics committee of the Karlsruhe Institute of Technology. after the right and left heel strikes were observed (Figure 1).
Variability is an important and ubiquitous feature of human movement. It has been hypothesized that decreased variability could indicate a tendency for overuse injury [51]. When studying variability, the level of analysis of the motor variability is crucial. For instance, a minimum variability in the level of joint angles could be important to avoid overuse injuries, whereas keeping the step frequency constant may be important for an efficient running style [52]. However, this does not necessarily constitute a contradiction. As assessed within the UCM approach, the covariation of individual parameters can be used to provide flexible movement variants while keeping an RV constant [23]. This can be interpreted as a compensatory mechanism for stabilizing the RV. In the present study, and in our previous ones, we have shown that UCM with a 3D whole-body model is applicable to running movement for stride-to-stride analyses, and that this method provides meaningful results [11,26]. The fact that the changes in the UCM Biology 2022, 11,942 The RV can be stabilized through covariance among the EVs, which c individual deviations. This means that a deviation in one EV can be com a corresponding change in another EV. In this way, there is variab movement executions that does not affect the RV; thus, with respect to the movement solutions exist. This portion of motor variability that does not also referred to as UCM‖.
It is considered beneficial because it constitu flexible movement executions by providing a multitude of equival solutions. If changes in an EV are not compensated for by another EV, ch occur. The motor variability that leads to changes in the RV is referred to portion of variability is potentially "unwanted", as it affects the RV. A low is thus interpreted as an indication of a high degree of control. If the UCM the UCM ꓕ , the chosen RV is assumed to be stabilized by the central n and the control hypothesis about the RV is accepted. The ratio of the UCM (UCMRatio) is used to quantify the degree of stabilization [23,25]. The UCM has yielded valuable results in a number of studies. Studie the total force applied by the fingers as the RV showed that the values of UCMꓕ increased under fatigue, and therefore increased the variability in In another study that analyzed the gender-specific effects of fatigue stability during a pointing task, higher UCM‖ and UCMꓕ were reported comparison with men in a fatigued state [33]. The UCM approach has be times for locomotion studies (e.g., Qu et al. reported a significantly low UCMRatio in the frontal plane during walking in a fatigued state than in a [28]), which means that the CoM is less stabilized in the fatigued sta recovered state. Möhler et al. showed that, even under fatigue, the Co expert runners was stabilized over the entire gait cycle [26]. While the UCM were unaffected by fatigue, the UCMꓕ increased in the flight phase. In add al. [11] showed that, at higher speeds, the motor variability during runn novices than in experts.
The UCM is a promising approach to gain further insight into the e on the variability structure that stabilizes the CoM in novices during runn the aim of the present study was to analyze the influence of fatigue on the motor variability structure that stabilizes the CoM trajectory in novice hypothesized that novice runners would lack strategies to control their Co state and would not stabilize their CoM trajectories, which would be decreased UCMRatio and increased UCMꓕ, respectively. The changes parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published study [20] the participants, experimental protocol and data collection are repeated i subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 years; sta m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 min; othe per week: 110 ± 71 min) participated in the study. Exclusion criteria were r exercise more than once a month, any injury or pain in the lower limbs w months prior to data collection and a BMI higher than 25 kg/m 2 . All partici written informed consent to voluntarily participate in this study. The study by the ethics committee of the Karlsruhe Institute of Technology. observed in this study correlated with the Borg-scale rating of the participants supports the plausibility of applying the UCM approach to study the effects of fatigue on the control of the CoM trajectory. To deepen our understanding of the compensatory mechanisms, one possibility could be to relate the UCM parameters to the movement patterns and their changes, which would help us to find out which EV has the largest effect on the changes in the RV. Mathematically, this information is contained in the Jacobian matrix, as it relates the rate of the changes in all the EVs to changes in the RV. However, there is a separate Jacobian for each time step, and a separate entry for each dimension of the RV. In this entry, there is the contribution of all the EVs. Due to this complexity, the extraction of the desired information is not straightforward.
The fact that the UCM Ratio is above zero is interpreted as an indicator for the control of the RV [42,43], which, in our case, is the CoM. Although it has been suggested that the CoM is a controlled variable, and especially in the postural studies [53], it is hard to verify this experimentally. Scholz et al. [54] showed that, during balance recovery, participants tended to re-establish the position of the CoM rather than those of the joint configurations. Consequently, they suggested that the CoM is the key variable controlled by the central nervous system. In the present study, overall, both the UCM Biology 2022, 11,942 The RV can be stabilized through covariance among the EV individual deviations. This means that a deviation in one EV c a corresponding change in another EV. In this way, there movement executions that does not affect the RV; thus, with res movement solutions exist. This portion of motor variability tha also referred to as UCM‖.
It is considered beneficial because flexible movement executions by providing a multitude solutions. If changes in an EV are not compensated for by anot occur. The motor variability that leads to changes in the RV is portion of variability is potentially "unwanted", as it affects the is thus interpreted as an indication of a high degree of control. I the UCM ꓕ , the chosen RV is assumed to be stabilized by th and the control hypothesis about the RV is accepted. The ratio (UCMRatio) is used to quantify the degree of stabilization [23,25] The UCM has yielded valuable results in a number of stud the total force applied by the fingers as the RV showed that th UCMꓕ increased under fatigue, and therefore increased the var In another study that analyzed the gender-specific effects stability during a pointing task, higher UCM‖ and UCMꓕ wer comparison with men in a fatigued state [33]. The UCM appro times for locomotion studies (e.g., Qu et al. reported a signifi UCMRatio in the frontal plane during walking in a fatigued state [28]), which means that the CoM is less stabilized in the f recovered state. Möhler et al. showed that, even under fatig expert runners was stabilized over the entire gait cycle [26]. Wh were unaffected by fatigue, the UCMꓕ increased in the flight ph al. [11] showed that, at higher speeds, the motor variability du novices than in experts.
The UCM is a promising approach to gain further insight on the variability structure that stabilizes the CoM in novices d the aim of the present study was to analyze the influence of fati motor variability structure that stabilizes the CoM trajectory hypothesized that novice runners would lack strategies to contr state and would not stabilize their CoM trajectories, which decreased UCMRatio and increased UCMꓕ, respectively. The parameters would correlate with the level of fatigue.

Materials and Methods
This study reanalyzes the data of a previously published the participants, experimental protocol and data collection are subsections.

Participants
A total of 14 healthy young novice runners (age: 27.4 ± 4.3 m; body mass: 77.5 ± 10.3 kg; running activity per week: 14 ± 18 and UCM were higher in the fatigued state, indicating more variable joint angles. Because UCM analysis examines variances from trial to trial, the number of repetitions (e.g., gait cycles) taken into consideration plays a crucial role in obtaining reliable results [40]. If fatigue effects are studied by using the UCM approach, then subjects cannot sustain the performance indefinitely. However, there are currently only a few studies that address this issue. The number of repetitions needed varies widely depending on the movement task, method and UCM parameter [40,41,55,56]. The large variations can be explained by the inherent noise in the original signals. The kinematics of a multi-joint movement, such as running, are generally associated with higher noise than the kinematics of a singlejoint movement, such as elbow flexion. Therefore, even though 15-20 cycles have been recommended for analyzing gait data, we chose to include 35 cycles in our analysis [41].
There are some limitations that should be considered. (1) When studying variability, performing a study on a treadmill may be a drawback, since a certain amount of variability is abolished due to the fixed running speed. On the one hand, the runners were not able to adapt the running speed, which is contrary to real-world scenarios. On the other hand, it has been shown that running style depends on running speed [16]. Moreover, if the speed is variable, then it is difficult to distinguish between the motor variability and the effects due to the changes in the speed. [57]. When studying the effects of fatigue, it is thus preferable to keep the running speed constant, even though reducing the running speed may be a possible strategy to maintain the CoM trajectory [7]. (2) Even though the Borg scale is a standardized method, a purely subjective measure may be considered as a limitation. On the other hand, a subjective measure might be better able to capture the state of fatigue resented by the participants, as fatigue has many facets. (3) In this study, the CoM was chosen as an RV. However, it was not necessarily the single right one for such studies. A multitude of parameters regarding efficiency, injury prevention and performance could also be of interest, as these are important during running and should therefore be stabilized or held constant. (4) When calculating the SPM over the whole gait cycle, an error might occur because the ratio between the stance-and swing-phase durations might change with fatigue [21]. However, in our previous study, we found no changes in the spatiotemporal parameters. Therefore, any bias due to temporal shifts in the normalization process can be ignored.

Conclusions
Thousands of people start running every year. Although the motivations are numerous, one that is common to many novice runners is to improve their fitness and, therefore, to push themselves to their limits. The occurrence of fatigue is thus inevitable. Understanding the effects of fatigue on movement patterns and its variability is crucial. By using the UCM approach, the effects of fatigue on the variability structure were analyzed in the present study by looking at the CoM trajectory as a crucial parameter in locomotion. In a fatigued state, the control of the CoM was decreased, but its stability was unaffected. The decrease in control was correlated with the ratings of perceived exertion. In contrast to expert runners, the structure of the motor variability in novices is thus affected by fatigue.
Because the UCM was able to capture the effects due to fatigue on the CoM trajectory, it may be used to evaluate the outcome of running-technique training, as the aim of technique training is ultimately to improve the running style, and the running style can be operationalized by the CoM trajectory [16]. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available in [20].