BJMB
Brazilian Journal of Motor Behavior
Research Article
!
Dias, Freitas, de
Freitas
2025
VOL.19
N.1
1 of 10
Synergic control of vertical body oscillation during stance phase of treadmill running
MATEUS S. DIAS1 | SANDRA M. S. F. FREITAS2 | PAULO B. DE FREITAS1
1 Laboratório de Análise do Movimento (LAM), Instituto de Ciências da Atividade Física e Esporte (ICAFE), Programa de Pós-Graduação Interdisciplinar em Ciências
da Saúde, Universidade Cruzeiro do Sul, São Paulo, SP, Brazil
2 Laboratório de Análise do Movimento (LAM), Programa de Mestrado e Doutorado em Fisioterapia, Universidade Cidade de São Paulo (UNICID), São Paulo, SP,
Brazil
Correspondence to:!Paulo B. de Freitas
Instituto de Ciências da Atividade Física e Esporte (ICAFE), Universidade Cruzeiro do Sul
Rua Galvão Bueno, 868, Liberdade, São Paulo, SP, Brazil
Telephone: +55 11 3385 3103
email: defreitaspb@gmail.com | paulo.defreitas@cruzeirodosul.edu.br
https://doi.org/10.20338/bjmb.v19i1.444
HIGHLIGHTS
• The Uncontrolled Manifold framework was used to
study motor coordination in running.
• The vertical trajectory of the center of mass (COMV) is
stabilized in treadmill running stance phase.
• A multi-joint synergy (sagittal: hip, knee, and ankle;
frontal: pelvis) is organized to stabilize COMV.
• Running speed does not affect the multi-joint synergy
stabilizing COMV.
ABBREVIATIONS
COMV Vertical trajectory of the center of mass
CNS Central nervous system
GRFv Vertical ground reaction force
J Jacobian matrix
SPM Statistical parametric mapping
UCM Uncontrolled Manifold
VUCM "Good" variance
VORT "Bad" variance
ΔV Synergy index
ΔVZ Modified Fisher z-transformation
PUBLICATION DATA
Received 31 10 2025
Accepted 15 04 2025
Published 09 05 2025
BACKGROUND: Stabilization of vertical body oscillation (i.e., vertical trajectory of the center
of mass – COMV) is critical for running efficiency and performance and may be influenced by
running speed.
AIM: To investigate the presence of a multi-joint synergy that stabilizes COMV during treadmill
running and evaluate the effect of running speed on this synergy using the Uncontrolled
Manifold (UCM) framework.
METHODS: Twenty-eight experienced runners (22–51 years old) ran on an instrumented
treadmill at 2.5, 3.5, and 4.5 m/s. Ankle, knee, and hip angles (sagittal plane) and pelvis
obliquity were used as elemental variables (DOF=4) to calculate the synergy index (ΔVZ), the
normalized difference between variance components in the joint space that did not affect
(VUCM) and those that did affect (VORT) vertical body oscillation (i.e., COMV) during the stance
phase (1-100%). Statistical parametric mapping (SPM) analysis was used to identify the
presence of synergy and the influence of speed.
RESULTS: Across all speeds, a synergy (ΔVZ>0) was present, except between 13% and
17% at the slowest speed. ΔVZ was not affected by speed. While VORT remained unchanged
across speeds, VUCM was higher at the fastest speed between 7% and 43% of the stance
phase.
INTERPRETATION: The findings indicate a robust multi-joint synergy that stabilizes vertical
body oscillation during the stance phase of running. Although running speed did not disrupt
this synergy, higher speeds were associated with increased “good” joint variability, suggesting
enhanced flexibility of the motor control system without compromising stability of vertical body
oscillation.
KEYWORDS: Center of mass | Coordination | Motor abundance | Uncontrolled manifold
hypothesis | Synergy
INTRODUCTION
Running is a form of locomotion that enables individuals to move rapidly through space, though it demands higher energy
expenditure compared to walking 1,2. It relies on the cyclical and coordinated motion of multiple body segments 3. According to the
principle of motor abundance 4,5, the central nervous system (CNS) regulates and coordinates relevant motor elements (e.g., fingers,
segments/joints, muscles) to stabilize one or more critical performance variables (e.g., total force, foot or center of mass trajectory, center
of pressure) across a range of motor tasks (e.g., finger pressing task 6,7, walking/running 8–10, upright standing 11,12). The ability of the CNS
to stabilize important performance variables is crucial for motor performance, particularly when facing changes in external forces (e.g.,
ground reaction forces, external loads) or internal states (e.g., sensory feedback, joint stiffness, fatigue) 4. A failure to achieve such
stabilization can lead to inefficient movements, a higher risk of injury, reduced task success, and increased energetic cost, especially in
tasks requiring precision, balance, or rapid adaptation to perturbations.
Vertical body oscillation, represented by the vertical trajectory of the center of mass (COMV), is associated with running energy
BJMB!!!!!!!!!
Brazilian(Journal(of(Motor(Behavior(
(
Dias, Freitas, de
Freitas
2025
VOL.19
N.1
2 of 10
Research Article
expenditure (i.e., running economy) and overall performance 13–16. Therefore, it may represent an important performance variable that the
CNS aims to stabilize. To investigate whether the CNS stabilizes COMV during treadmill running, we employed the framework of
Uncontrolled Manifold (UCM) hypothesis 17–19.
The UCM hypothesis proposes that the CNS organizes multi-element (e.g., multi-joint) synergies to stabilize one or more
relevant performance variables (e.g., COMV position) during the execution of a specific task (e.g., running). In this context, a synergy is
defined as a neural organization that coordinates the actions of a series of elements relevant to the performance of a specific task,
covarying these elements with the goal of stabilizing performance variables or measures 17–19. To test the existence of such synergies,
the variability across repeated trials (or cycles in continuous motor tasks like running) is estimated by the variance of all elemental
variables directly involved in the task (e.g., joint angles, fingers, muscles) and is linked to the performance variable itself. Specifically,
UCM analysis decomposes the total variance observed in the elemental variables into two distinct components. The first component,
known as VUCM or "good" variance, reflects combinations of elements that preserve the stability of the performance variable. The other
component, VORT or "bad" variance, represents combinations among elements that result in performance errors or deviations from
average performance across trials. If VUCM is greater than VORT, it can be concluded that there is synergy controlling that performance
variable. This results in a positive synergy index (ΔV). If ΔV is negative, meaning that VORT is greater than VUCM, it can be concluded that
the CNS is not actively involved in controlling that specific performance variable 17–19.
Previous studies by Möhler and collaborators 20–22 employed the UCM framework to investigate whether a multi-joint or multi-
segmental synergy stabilizes the 2D (sagittal) and 3D COM trajectories during running. They also explored how running experience,
speed, and fatigue affect this synergy and found evidence of robust multi-segmental synergies stabilizing the 2D and 3D COM trajectory
throughout the running cycle. However, they did not observe any significant impact of experience, speed, and fatigue on the strength of
these synergies. According to Möhler and colleagues 22, the limited sensitivity of UCM outcomes to detect potential effects of experience,
speed, and fatigue may reflect the CNS's consistent and precise control over the joints contributing to the regulation of the hypothetical
COM trajectory. However, Möhler and collaborators 20–22 did not investigate the COM stabilization along each individual axis of motion,
focusing instead on the overall 3D trajectory. While informative, this approach may obscure the potential for axis-specific stabilization, as
it is theoretically plausible that the CNS prioritizes COM stabilization in certain directions over others.
As previously mentioned, vertical body oscillation is closely linked to running efficiency and performance 13–16. Yet, whether and
how the CNS stabilizes COMV through multi-joint coordination remains underexplored. Moreover, prior studies using the UCM approach
have predominantly relied on complex 3D biomechanical models to estimate COM trajectory 20–24, which require detailed segmental
anthropometrics and multiple assumptions to relate changes in the performance variable to the elemental variables (e.g., segmental
angles). While these models provide adequate representations of 3D COM trajectory, they also introduce methodological challenges,
such as increased sensitivity to marker placement errors, soft tissue movement artefact, and model assumptions 25,26. In contrast, the
approach adopted in this study employs a simpler linear model based on multiple regression analysis 9,27–29. This method establishes a
direct statistical relationship between the elemental variables (e.g., joint angles) and the performance variable (COMV), without requiring
detailed anthropometric modeling or assumptions about segment masses and inertias. As such, it offers a more accessible and robust
framework for UCM analysis, facilitating broader application in experimental and clinical settings.
In the present study, we investigated the stabilization of COMV trajectory in the stance phase during treadmill running and
examined how this stabilization is influenced by running speed. Based upon the findings of Möhler and colleagues, who demonstrated
that the COM trajectory is stabilized by a multi-segmental synergy during running 20–22, we hypothesized that the CNS would organize a
multi-joint synergy to stabilize COMV during the stance phase. Furthermore, considering that increased movement speed has been
shown to disrupt motor coordination, reflected by greater 'bad' variance (VORT) and weaker synergies 36, we hypothesized that the
strength of this synergy would be reduced at higher running speeds due to an increase in VORT.
METHODS
Participants
Kinematic and kinetic data were collected from 28 long-distance runners (27 males), aged 22 to 51 years (age, mean ±
standard deviation: 34.75 ± 6.7 years). The data source is an openly accessible database (DOI: 10.6084/m9.figshare.4543435).
Experienced runners, who had maintained a training schedule of at least three running sessions per week for an average of 8.4 (±7)
years, were included in the study. Participants typically trained 3 to 5 times per week, covering over 20 km. The group's average running
pace was 4.1 ± 0.4 min/km. This group of runners consisted of 3 elite and 25 competitive runners, who competed in races ranging from 5
km to marathons. Among the participants, 20 were right-foot dominant, and 8 were left-foot dominant. Detailed individual data are
available at https://figshare.com/ndownloader/files/39452935.
On average, runners weighed 69.6 kg (± 7.7 kg) and were 175.9 cm (± 6.8 cm) tall. This study received approval from the
Research Ethics Committee at the Federal University of ABC in Brazil (CAAE: 53063315.7.0000.5594). As reported by the authors, each
participant read and signed an informed consent form before taking part in the study 30.
BJMB!!!!!!!!!
Brazilian(Journal(of(Motor(Behavior(
(
Dias, Freitas, de
Freitas
2025
VOL.19
N.1
3 of 10
Research Article
Study Protocol
The participants initially walked on a dual-belt, instrumented treadmill with embedded force plates (FIT; Bertec, Columbus, OH,
USA) at 1.2 m/s for one minute to become familiar with the treadmill. They then stood on the left treadmill belt while the speed increased
to 2.5 m/s. The participant ran at this speed for 3 minutes and 30 seconds, with kinematic and kinetic data recorded during the final 30
seconds. The treadmill speed was subsequently raised to 3.5 m/s, and the participant ran for another 3 minutes and 30 seconds, with
data recorded in the last 30 seconds. This process was repeated at 4.5 m/s. According to the original authors, runners required three
minutes to adapt to each treadmill speed before data recording.
Twelve Raptor4 cameras (Motion Analysis Corporation, Santa Rosa, CA, USA) were used to record 48 technical and
anatomical markers, some in clusters on the thigh and shank segments. After a 1-second standing calibration trial, most anatomical
markers were removed. Kinematic data were recorded at 150 Hz, while the instrumented treadmill provided kinetic data recorded at 300
Hz. Each 30-second recording allowed for analysis of approximately 38, 40, and 43 cycles at 2.5, 3.5, and 4.5 m/s, respectively.
Data processing
We used a customized LabView routine (National Instruments, Austin, TX, USA) for data analysis. ASCII files with markers' 3D
positions over time and force components from the left belt's embedded force plate were employed to compute the outcomes. We filtered
all markers’ data and the vertical force component using a 4th-order, zero-lag, Butterworth filter with a 10 Hz cutoff frequency. We
estimated the vertical center of mass position (COMV) from the virtual marker created at the midpoint of the markers placed on the right
and left posterosuperior iliac spine 25,31. The markers were also used to compute one segmental (i.e., pelvis obliquity) and three joint (i.e.,
hip, knee, and ankle in the sagittal plane) angles, following the method described by Fukuchi et al. 30. The pelvis obliquity angle was
expressed relative to the horizontal global coordinate plane and the joint angles expressed the distal segment movement relative to the
proximal segment. Foot strike and toe-off events were identified based on the vertical ground reaction force (GRFv), with foot strike
occurring when GRFv exceeded 20 N and toe off when it fell below this threshold 30. To distinguish between left and right foot strikes, we
analyzed the anterior-posterior position of the 5th metatarsal joint markers for both feet. The foot whose marker was positioned further
forward at the moment of foot strike was the one in contact with the treadmill. Data analyses and UCM outcomes calculations were
performed only for the dominant leg based on a previous study 32 and on a preliminary analysis of this dataset that showed no discernible
differences between stance legs for any of the outcomes of interest.
UCM analysis of variance
The UCM analysis was performed to investigate how variations in four angles (i.e., ankle, knee, and hip angles in the sagittal
plane, and pelvis obliquity) across running cycles influence COMV during the stance phase of the running cycle. Initially, a multiple
regression analysis was performed to model the relationship between the mean-free COMV (dependent variable) and the mean-free
values of the four angles (independent variables) throughout the stance phase 9,27–29. From the resulting regression coefficients, a
Jacobian matrix (J) was derived for each participant.
The mean-free joint angles were then time-normalized across the stance phase from 1-100%. At each time point, the mean-
free joint angle vectors were projected into the null space (UCM) and the orthogonal space of J, using matrix decomposition. The total
variance of the projected joint configuration was partitioned at each instant into two components: (1) VUCM, the variance within the null
space of J that does not affect COMV, and (2) VORT, the variance in the orthogonal space that affects COMV.
With VUCM and VORT calculated for each instant, ΔV was computed at each time point as ΔV=(VUCM/3-VORT/1)/(VTOT/4), where
VTOT represents the total variance. Normalization by the dimensionality of the corresponding space was applied to VUCM, VORT, and VTOT.
Since ΔV does not follow a normal distribution, a modified Fisher z transformation was used for normalization,
[ΔVZ=(0.5×ln((4+ΔV)/(1.333-ΔV))) – 0.549], where ln represents the natural log and the constants represent the absolute boundaries of
ΔV (−4 to 1.333) and the z-transformed value when ΔV = 0 (i.e., 0.549). Positive values of ΔVZ indicate the presence of multi-joint
synergy stabilizing COMV during the stance phase, with larger values reflecting stronger synergies.
Statistical Analysis
Statistical parametric mapping (SPM) analyses 33 were employed to test the hypothesis of the study. Three one-sample t-test
SPM analyses, one for each running speed, were performed to test whether ΔVZ was significantly greater than zero, which would
indicate the presence of a multi-joint synergy stabilizing COMV. In addition, F-test SPM analyses were conducted to examine the effect of
running speed on ΔVZ, VUCM, and VORT. When significant effects were found, post-hoc SPM t-tests were performed. Because VUCM and
VORT were not normally distributed, they were log-transformed prior to the SPM analyses. The significance level was set at 0.05 for all
tests.
RESULTS
The COMV position, normalized by the participants’ stature, ranged from 0.45 to 0.63 (Figure 1) during the stance phase,
BJMB!!!!!!!!!
Brazilian(Journal(of(Motor(Behavior(
(
Dias, Freitas, de
Freitas
2025
VOL.19
N.1
4 of 10
Research Article
reaching its minimum height at 41.2% (±2.8), 42.6% (±2.3), and 43.3% (±2.7) of the stance phase for 2.5, 3.5, and 4.5 m/s, respectively.
The SPM F-test revealed a significant effect of speed on COMV across a substantial portion of the stance phase [1-29%
(SPM{F}, F(2,54) = 521, p < 0.001, critical threshold = 3.91) and 50-100% (SPM{F}, F(2,54) = 810, p < 0.001, critical threshold = 3.91)] but
not in the moments near the COMV change of direction (30-49%, SPM{F}, F(2,54) = 3.7, p > 0.05, critical threshold = 3.91).
Figure 1. Across-subject averaged COMV (normalized by stature) trajectories during the running stance phase (1-100%) in the 2.5 m/s (green line), 3.5 m/s (blue line), and
4.5 m/s (red line) speed conditions. Error bars represent standard error for each point in the stance phase (1-100%).
The multiple linear regression analyses used to determine J revealed high coefficients of determination (R2 ≥ 0.9 for all cases,
median R2 ≥ 0.98, across speed conditions). This indicates that the model incorporating the four angles effectively predicted COMV at all
speeds. Additionally, all four angles significantly contributed to the model (p < 0.05).
The one-sample SPM{t} analyses revealed that ΔVZ was consistently greater than zero throughout the stance phase in all three
speed conditions (Figure 2A), except for a brief interval between 13-17% in the 2.5 m/s condition (Figure 2B-D). Additionally, the SPM F-
test indicated a significant main effect of speed on ΔVZ only by the end of the stance phase (98-100%, SPM{F}, F(2,54) = 5.95, p = 0.049,
critical threshold = 5.18, Fig. 3A). However, post-hoc pairwise comparisons (SPM t-test) revealed no significant differences between
speed conditions for ΔVZ (all p > 0.05, Figure 3B-D).
Figure 2. Across-subject averaged synergy index (ΔVZ) time-series at speeds of 2.5 m/s (green line), 3.5 m/s (blue line), and 4.5 m/s (red line) with error bars representing
standard error for each point in the stance phase (A, left-hand side). The plots on the right-hand side (B-D) show one-sample SPM {t} trajectories during the running stance
phase (1-100%) for each tested speed: (B) 2.5 m/s, (C) 3.5 m/s, and (D) 4.5 m/s. Horizontal lines in B, C, and D represent t-critical values. Gray areas in B, C, and D
indicate intervals with p < 0.05.
BJMB!!!!!!!!!
Brazilian(Journal(of(Motor(Behavior(
(
Dias, Freitas, de
Freitas
2025
VOL.19
N.1
5 of 10
Research Article
Figure 3. SPM {F} (A) and post-hot SPM t-tests (B-D) values during the running stance phase for the synergy index (ΔVZ). B: 2.5 m/s vs. 3.5 m/s comparison; C: 2.5 m/s
vs. 4.5 m/s comparison; and D: 3.5 m/s vs. 4.5 m/s comparison. Horizontal lines represent the F-critical (A) and t-critical (B-D) values.
The SPM analysis revealed a significant effect of speed on VUCM (log-transformed, Figure 4A) during an interval between 7%
and 43% of the stance phase (Figure 4B, SPM{F}, F(2,54) = 14.46, p = 0.002, critical threshold = 5.29). Specifically, VUCM was larger at 4.5
m/s compared to 2.5 between 10% and 41% of the stance phase (SPM{t}, t27 = 5.02, p < 0.001, critical threshold = 3.42, Figure 4D) and
larger at 4.5 m/s compared to 3.5 m/s between 36% and 43% (SPM{t}, t27 = 3.57, p = 0.013, critical threshold = 3.4, Figure 4E) running
speed. In contrast, for VORT (log-transformed, Figure 5A), the SPM analysis did not reveal a significant effect of speed (SPM{F}, F(2,54) =
4.07, p > 0.05, critical threshold = 5.03, Figure 5B).
Figure 4. Across-subject averaged VUCM time-series at speeds of 2.5 m/s (green line), 3.5 m/s (blue line), and 4.5 m/s (red line) (A, left-hand side). (B) SPM one-way
ANOVA results [SPM {F}] testing the effect of speed on VUCM. SPM t-test results for the following comparisons: (C) 2.5 m/s and 3.5 m/s, (D) 2.5 m/s and 4.5 m/s, and (E) 3.5
m/s and 4.5 m/s condition. Error bars in (A) represent standard error. The horizontal line in (B) is the SPM F-critical and the horizontal lines in C, D, and E are SPM t-critical
values. Horizontal gray areas represent intervals where p<0.05 for the SMP F-test.
BJMB!!!!!!!!!
Brazilian(Journal(of(Motor(Behavior(
(
Dias, Freitas, de
Freitas
2025
VOL.19
N.1
6 of 10
Research Article
Figure 5. Across-subject averaged VORT time-series at speeds of 2.5 m/s (green line), 3.5 m/s (blue line), and 4.5 m/s (red line) (A, left-hand side). (B) SMP one-way
ANOVA results [SPM {F}] testing the effect of speed on VUCM. SMP t-test results for the following comparisons: (C) 2.5 m/s and 3.5 m/s, (D) 2.5 m/s and 4.5 m/s, and (E) 3.5
m/s and 4.5 m/s condition. Error bars in (A) represent standard error. The horizontal line in (B) is the SPM F-critical and the horizontal lines in C, D, and E are SPM t-critical
values.
DISCUSSION
In this study, we tested two hypotheses: (1) that a multi-joint synergy would stabilize vertical body oscillation during the stance
phase of treadmill running, and (2) that this synergy would weaken with increasing running speed. Our findings confirmed the first
hypothesis but refuted the second. We observed the presence of a multi-joint synergy (ΔVZ) responsible for stabilizing the COMV (i.e.,
body vertical oscillation) throughout the stance phase. Additionally, we found that vertical body oscillation was affected by running speed
at the onset and end of the stance phase but remained stable at its minimum point—when the trajectory reverses from descending to
ascending. Moreover, running speed influenced the 'good' variance component, VUCM, primarily in the early stance phase, from just after
touchdown until the COMV changed direction. Specifically, VUCM was highest in the fastest speed condition within this interval. In contrast,
the bad variance (VORT) was not impacted by running speed.
Previous studies by Mohler and colleagues 20–22 showed that the planar and tridimensional trajectory of the COM is stabilized
during running. However, they did not analyze the stabilization of each axis independently, leaving open the question of how much of this
stabilization is specifically related to vertical body oscillation. Most recently, Liew and collaborators 34, using the same dataset as the
present study, showed that the COMV stability can be achieved through covariation between two parameters from a simple spring-mass
model: leg angle (i.e., the angle between the leg vector and the horizontal surface) and leg length. They applied the UCM concept but
rather than using elemental variables typically considered to be directly controlled by the CNS (e.g., joint or segmental angles), they
treated the covariation between two performance variables (leg angle and leg length), which were presumed to be already stabilized by
the CNS, to maintain the stability of a third performance variable: the vertical (or the horizontal) COM trajectory. Their results suggested
the existence of a covariation pattern between leg angle and leg length that stabilized vertical, but not horizontal, COM trajectories during
running stance, particularly at the beginning and end of this phase.
Our study is the first to demonstrate that COMV is stabilized by a multi-joint synergy based on elemental variables that are more
mechanistically tied to the neural control of movement. Unlike Liew and collaborators 34, who used leg angle and leg length (i.e.,
parameters that are the result of coordinated actions of multiple joints), we utilized individual joint angles in the sagittal and frontal planes
as elemental variables. These angles directly and independently contribute to the vertical displacement of the COM: greater flexion leads
to a lower COMV while greater extension results in a higher COMV. Importantly, these elemental variables are under direct and
independent control by the CNS, as each joint can be actively modulated by specific muscle groups with distinct neural control 35. This
provides a more mechanistically grounded model to analyze motor synergies compared to composite variables such as leg length or
angle, which are emergent outcomes rather than directly controlled elements. Thus, our findings provide stronger evidence that the CNS
forms synergies at the joint level to stabilize vertical body oscillation—a variable known to be critical for running economy 13–16 and
performance.
The notion that vertical body oscillation is tightly controlled by the CNS is further supported by the finding that COMV
BJMB!!!!!!!!!
Brazilian(Journal(of(Motor(Behavior(
(
Dias, Freitas, de
Freitas
2025
VOL.19
N.1
7 of 10
Research Article
displacement at its minimum point is unaffected by running speed. Additionally, the stability of the synergy index (ΔVZ) and the lack of
change in VORT across speeds suggest that the CNS consistently prioritizes control of COMV regardless of locomotor demand. In contrast
to general expectations that increasing speed disrupts motor coordination, reflected by higher 'bad' variance (VORT) and weaker synergies
36, our results indicate that the CNS maintains tight control over COMV, likely due to its biomechanical and metabolic relevance.
Fukuchi and collaborators, using the same dataset, reported substantial changes in several biomechanical variables across
speeds, including stride length and cadence, joint angles, ground reaction forces, torques, and power 30. Additionally, our data show that
vertical body oscillation (i.e., COMV trajectory) was influenced by running speed, with higher COMV values observed at slower speeds
during both the initial and latter portions of the stance phase (see Fig. 1). However, despite these biomechanical differences, increased
running speed did not impair COMV stability. The first indication that vertical body oscillation is tightly controlled by the CNS, as
suggested by Möhler and colleagues 22, is the lack of effect of speed in the COMV displacement during the phase when the COMV
changes direction from descending to ascending. The second indication is that both the bad variance (i.e., VORT) and the index of multi-
joint synergy stabilizing the COMV are not affected by speed at any time during the stance phase.
In general, the movement speed influences control, with faster speeds often leading to decreased task performance, increased
'bad' variance among elemental variables, and reduced synergy strength 36. However, our findings do not support this assumption. Mohler
and collaborators 21 tested the stability of the 3D COM trajectory using the UCM approach in novice and experienced runners. The
participants ran at two different speeds, 10 and 15 km/h, and they observed no differences in the synergy index between groups at both
speeds. Although they did not test the effect of running speed on the synergy index, examining their reported values (Table II in Möhler
and colleagues 21) suggests that the effect of speed is negligible or non-existent. Conversely, our results are partially consistent with
those of Liew and collaborators 34 who, using the same dataset and investigating the stabilization of the COMV through the covariation of
leg angle and leg length, found that the index of motor abundance (IMA), which is similar to our synergy index, was affected by running
speed in certain parts of the stance phase (onset, near midstance, and at the end). However, those researchers did not provide a
plausible explanation for the observed effect of speed. It is important to note that although the performance variable was the same (i.e.,
COMV), the elemental variables differed between our study and that of Liew and collaborators 34.
Interestingly, while VORT and ΔVZ were not influenced by running speed, we did find an effect of speed on VUCM. Specifically,
VUCM was greater in the fastest speed condition (4.5 m/s) compared to the slowest (2.5 m/s) between 7% and 43% of the stance phase,
which corresponds to the braking phase, with 43% mark aligning with the point at which COMV transitions from descending to ascending.
This period is dedicated to absorbing impact forces and controlling the descent of the body through eccentric action of the hip and knee
extensors and ankle dorsiflexors 37,38. VUCM represents the trial-by-trial multi-joint variability that does not interfere with task performance
(i.e., COMV) and is linked to the CNS’s flexibility and adaptability in finding proper joint motion combinations to respond promptly to
perturbations 39–41. Thus, higher VUCM at faster speeds likely represents increased flexibility in joint coordination, allowing the CNS to
prepare for and respond to perturbations within a shorter time frame, primarily during the running braking phase. Notably, regardless of
running speed both VUCM and VORT are close to their minimum values near the moment of COMV reversal, suggesting that the CNS
transiently reduces multi-joint variability to avoid further downward COM displacement and potential collapse.
The importance of UCM analysis to improving running performance lies in its ability to identify how the CNS organizes joint
coordination to stabilize salient performance variables such as COMV. Understanding the structure and strength of motor synergies can
guide coaches and clinicians in designing training protocols that enhance flexibility without compromising stability. For instance, a runner
exhibiting low VUCM in early stance may lack the adaptive capacity to respond to perturbations at higher speeds, under fatigue, or when
running on uneven terrain. Conversely, excessive VORT may indicate poor joint coordination and ineffective movement patterns. By
targeting specific joints or phases of stance where synergy is weak, interventions can be developed to improve running control and
economy. Moreover, UCM-based indices may serve as biomarkers to assess motor competence, track training progress, or monitor
recovery in rehabilitation contexts.
This study has limitations that should be acknowledged. Some may argue that treadmill running differs from overground
running and could influence coordination patterns. However, studies focusing on joint angles and COM trajectory suggest that the
differences between treadmill and overground running are minimal 42,43. Despite these minimal differences, future studies should test the
generalizability of these results to overground running. Another potential limitation is that the vertical trajectory of the COM was not based
on the actual COM position obtained with a full-body biomechanical model but on a virtual marker positioned between the posterior
superior iliac spine markers. However, earlier studies have shown that the vertical trajectory of this virtual marker closely matches the
vertical displacement of the COM calculated using a full-body biomechanical model 25,31. Additionally, as mentioned by Liew and
collaborators 34, since this dataset predominantly includes male participants, it is important to exercise caution when extrapolating our
findings to female participants. Finally, although studies indicate consistent biomechanics in well-trained runners up to the age of 60 44,45,
the relatively broad age range of our sample (22–51 years) may introduce some age-related variability. Therefore, the age range of the
sample should be considered a potential limitation of our study.
Despite these considerations, this study represents a pioneering investigation into how the joints of the lower limbs and pelvis
coordinate to stabilize vertical body oscillation during the running stance phase. Further research would be valuable in sports and
rehabilitation science to examine how acute factors (e.g., central and peripheral fatigue) and chronic conditions (e.g., patellofemoral pain
syndrome, low back pain) may impact the multi-joint synergy involved in stabilizing salient performance variables during running, and how
BJMB!!!!!!!!!
Brazilian(Journal(of(Motor(Behavior(
(
Dias, Freitas, de
Freitas
2025
VOL.19
N.1
8 of 10
Research Article
interventions might enhance this coordination.
CONCLUSION
In conclusion, the findings indicated that the CNS organizes a multi-joint synergy to stabilize vertical body oscillation (i.e.,
COMV) throughout most of the running stance phase. This stabilization plays a crucial role in running performance and economy.
Excessive downward oscillation of the body would require greater eccentric muscle torque to prevent collapse, while excessive upward
movement at the beginning and end of the stance phase could suggest that the generated muscle torque is not being efficiently directed
toward forward propulsion, which is the primary goal of running. The absence of a speed effect on COMV further suggests that vertical
body oscillation is tightly regulated by the CNS, regardless of external conditions. However, the greater good variance (VUCM) in the
fastest condition at the first half of the stance phase, when the COM is descending and approaching its lowest position indicates that the
CNS is better prepared to deal with extreme conditions by increasing the system´s flexibility.
REFERENCES
1. Cavagna GA, Kaneko M. Mechanical work and efficiency in level walking and running. J Physiol. 1977;268(2):467-481. doi:
10.1113/jphysiol.1977.sp011866
2. Margaria R, Cerretelli P, Aghemo P, Sassi G. Energy cost of running. J Appl Physiol. 1963;18(2):367-370. doi:10.1152/jappl.1963.18.2.367
3. Hicheur H, Terekhov AV, Berthoz A. Intersegmental Coordination During Human Locomotion: Does Planar Covariation of Elevation Angles Reflect
Central Constraints? J Neurophysol. 2006;96(3):1406-1419. doi:10.1152/jn.00289.2006
4. Latash ML. The bliss (not the problem) of motor abundance (not redundancy). Exp Brain Res. 2012;217(1):1-5. doi:10.1007/s00221-012-3000-4
5. Latash ML, Scholz JP, Schoner G. Toward a new theory of motor synergies. Motor Control. 2007;11(3):276-308. doi: 10.1123/mcj.11.3.276
6. de Freitas PB, Freitas SMSF, Lewis MM, Huang X, Latash ML. Stability of steady hand force production explored across spaces and methods of
analysis. Exp Brain Res. 2018;236(6):1545-1562. doi:10.1007/s00221-018-5238-y
7. Park J, Sun Y, Zatsiorsky VM, Latash ML. Age-related changes in optimality and motor variability: an example of multifinger redundant tasks. Exp
Brain Res. 2011;212(1):1-18. doi:10.1007/s00221-011-2692-1
8. Yamagata M, Tateuchi H, Shimizu I, Ichihashi N. The effects of fall history on kinematic synergy during walking. J. Biomech. 2019;82:204-210.
doi:10.1016/j.jbiomech.2018.10.032
9. de Freitas PB, Freitas SMSF, Dias MS. Synergic control of the minimum toe clearance in young and older adults during foot swing on treadmill
walking in different speeds. Gait Posture. 2024;111:150-155. doi:10.1016/j.gaitpost.2024.04.025
10. Dias MS, Freitas SMSF, de Freitas, PB. Multi-joint synergy in foot height stabilization across different running speeds: An Uncontrolled Manifold
analysis. Res Q Exerc Sport. 2025. doi:10.1080/02701367.2025.2480143
11. Krishnamoorthy V, Latash ML, Scholz JP, Zatsiorsky VM. Muscle synergies during shifts of the center of pressure by standing persons. Exp Brain
Res. 2003;152(3):281-292. doi: 10.1007/s00221-003-1574-6.
12. Freitas SMSF, de Freitas PB, Falaki A, et al. Synergic control of action in levodopa naïve Parkinson’s disease patients: II. Multi-muscle synergies
stabilizing vertical posture. Exp Brain Res. 2020;238(12):2931-2945. doi:10.1007/s00221-020-05947-z
13. Folland JP, Allen SJ, Black MI, Handsaker JC, Forrester SE. Running technique is an important component of running economy and performance.
Med Sci Sports Exerc. 2017;49(7):1412. doi: 10.1249/MSS.0000000000001245.
14. Tartaruga MP, Brisswalter J, Peyré-Tartaruga LA, et al. The Relationship Between Running Economy and Biomechanical Variables in Distance
Runners. Res Q Exerc Sport. 2012; 83(3): 367-375. doi: 10.1080/02701367.2012.10599870.
15. Leite OHC, do Prado DML, Rabelo NDDA, et al. Two sides of the same runner! The association between biomechanical and physiological markers
of endurance performance in distance runners. Gait Posture. 2024;113:252-257. doi:10.1016/j.gaitpost.2024.06.027
16. Moore IS. Is There an Economical Running Technique? A Review of Modifiable Biomechanical Factors Affecting Running Economy. Sports Med.
2016;46(6):793-807. doi:10.1007/s40279-016-0474-4
17. Latash ML, Scholz JP, Schoner G. Motor control strategies revealed in the structure of motor variability. Exerc Sport Sci Rev. 2002;30(1):26-31. doi:
10.1097/00003677-200201000-00006
18. Scholz JP, Schöner G. The uncontrolled manifold concept: identifying control variables for a functional task. Exp Brain Res. 1999;126(3):289-306.
doi:10.1007/s002210050738
19. Scholz JP, Schöner G. Use of the uncontrolled manifold (UCM) approach to understand motor variability, motor equivalence, and self-motion. Adv
Exp Med Biol. 2014;826:91-100. doi:10.1007/978-1-4939-1338-1_7
20. Möhler F, Fadillioglu C, Scheffler L, Müller H, Stein T. Running-Induced Fatigue Changes the Structure of Motor Variability in Novice Runners.
Biology. 2022;11(6):942. doi:10.3390/biology11060942
21. Möhler F, Marahrens S, Ringhof S, Mikut R, Stein T. Variability of running coordination in experts and novices: A 3D uncontrolled manifold analysis.
Eur J Sport Sci. 2020;20(9):1187-1196. doi:10.1080/17461391.2019.1709561
BJMB!!!!!!!!!
Brazilian(Journal(of(Motor(Behavior(
(
Dias, Freitas, de
Freitas
2025
VOL.19
N.1
9 of 10
Research Article
22. Möhler F, Ringhof S, Debertin D, Stein T. Influence of fatigue on running coordination: A UCM analysis with a geometric 2D model and a subject-
specific anthropometric 3D model. Hum Mov Sci. 2019;66:133-141. doi:10.1016/j.humov.2019.03.016
23. Papi E, Rowe PJ, Pomeroy VM. Analysis of gait within the uncontrolled manifold hypothesis: Stabilisation of the centre of mass during gait. J
Biomech. 2015;48(2):324-331. doi:10.1016/j.jbiomech.2014.11.024
24. Yamagata M, Tateuchi H, Shimizu I, Saeki J, Ichihashi N. The relation between kinematic synergy to stabilize the center of mass during walking and
future fall risks: a 1-year longitudinal study. BMC Geriatr. 2021;21(1):1-10. doi:10.1186/s12877-021-02192-z
25. Napier C, Jiang X, MacLean CL, Menon C, Hunt MA. The use of a single sacral marker method to approximate the centre of mass trajectory during
treadmill running. J Biomech. 2020;108:109886. doi:10.1016/j.jbiomech.2020.109886
26. Pavei G, Seminati E, Cazzola D, Minetti AE. On the Estimation Accuracy of the 3D Body Center of Mass Trajectory during Human Locomotion:
Inverse vs. Forward Dynamics. Front Physiol. 2017;8. doi:10.3389/fphys.2017.00129
27. Freitas SMSF, Scholz JP. A comparison of methods for identifying the Jacobian for uncontrolled manifold variance analysis. J Biomech.
2010;43(4):775-777. doi:10.1016/j.jbiomech.2009.10.033
28. Freitas SMSF, Scholz JP, Latash ML. Analyses of joint variance related to voluntary whole-body movements performed in standing. J Neurosci
Methods. 2010;188(1):89-96. doi:10.1016/j.jneumeth.2010.01.023
29. Tuitert I, Valk TA, Otten E, Golenia L, Bongers RM. Comparing Different Methods to Create a Linear Model for Uncontrolled Manifold Analysis.
Motor Control. 2019;23(2):189-204. doi:10.1123/mc.2017-0061
30. Fukuchi RK, Fukuchi CA, Duarte M. A public dataset of running biomechanics and the effects of running speed on lower extremity kinematics and
kinetics. PeerJ. 2017;5:e3298. doi:10.7717/peerj.3298
31. Gullstrand L, Halvorsen K, Tinmark F, Eriksson M, Nilsson J. Measurements of vertical displacement in running, a methodological comparison. Gait
Posture. 2009;30(1):71-75. doi:10.1016/j.gaitpost.2009.03.001
32. Brown AM, Zifchock RA, Hillstrom HJ. The effects of limb dominance and fatigue on running biomechanics. Gait Posture. 2014;39(3):915-919.
doi:10.1016/j.gaitpost.2013.12.007
33. Pataky TC, Robinson MA, Vanrenterghem J. Vector field statistical analysis of kinematic and force trajectories. J Biomech. 2013;46(14):2394-2401.
doi:10.1016/j.jbiomech.2013.07.031
34. Liew BXW, Rügamer D, Birn-Jeffery AV. Neuromechanical stabilisation of the centre of mass during running. Gait Posture. 2024;108:189-194.
doi:10.1016/j.gaitpost.2023.12.005
35. Scholz JP, Schöner G. Use of the uncontrolled manifold (UCM) approach to understand motor variability, motor equivalence, and self-motion. Adv
Exp Med Biol. 2014;826:91-100. doi:10.1007/978-1-4939-1338-1_7
36. Latash ML. One more time about motor (and non-motor) synergies. Exp Brain Res. 2021;239(10):2951-2967. doi:10.1007/s00221-021-06188-4
37. Novacheck TF. The biomechanics of running. Gait Posture. 1998;7(1):77-95. doi:10.1016/S0966-6362(97)00038-6
38. Gazendam MGJ, Hof AL. Averaged EMG profiles in jogging and running at different speeds. Gait Posture. 2007;25(4):604-614.
doi:10.1016/j.gaitpost.2006.06.013
39. Mattos DJS, Latash ML, Park E, Kuhl J, Scholz JP. Unpredictable elbow joint perturbation during reaching results in multijoint motor equivalence. J
Neurophysiol. 2011;106(3):1424-1436. doi:10.1152/jn.00163.2011
40. Scholz JP, Schöner G, Latash ML. Identifying the control structure of multijoint coordination during pistol shooting. Exp Brain Res. 2000;135(3):382-
404. doi:10.1007/s002210000540
41. Zhang W, Scholz JP, Zatsiorsky VM, Latash ML. What do synergies do? Effects of secondary constraints on multidigit synergies in accurate force-
production tasks. J Neurophysiol. 2008;99(2):500-513. doi:10.1152/jn.01029.2007
42. Fellin RE, Manal K, Davis IS. Comparison of lower extremity kinematic curves during overground and treadmill running. J Appl Biomech.
2010;26(4):407-414. doi:10.1123/jab.26.4.407
43. Van Hooren B, Fuller JT, Buckley JD, et al. Is motorized treadmill running biomechanically comparable to overground running? A systematic review
and meta-analysis of cross-over studies. Sports Med. 2020;50(4):785-813. doi:10.1007/s40279-019-01237-z
44. Willy RW, Paquette MR. The Physiology and biomechanics of the master runner. Sports Med Arthrosc Rev. 2019;27(1):15-21.
doi:10.1097/JSA.0000000000000212
45. Dos Anjos Souza VR, Seffrin A, da Cunha RA, et al. Running economy in long distance runners is positively affected by running experience and
negatively by aging. Physiol Behav. 2023;258:114032. doi:10.1016/j.physbeh.2022.114032
ACKNOWLEDGEMENTS
The authors thank Dr. Claudiane Fukuchi, Dr. Reginaldo Fukuchi, and Dr. Marcos Duarte from Federal University of ABC,
Brazil, for their impressive work recording and processing the data and for making this dataset open to the public.
BJMB!!!!!!!!!
Brazilian(Journal(of(Motor(Behavior(
(
Dias, Freitas, de
Freitas
2025
VOL.19
N.1
10 of 10
Research Article
Citation: Dias MS, Freitas SMSF, de Freitas PB. (2025). Synergic control of the vertical body trajectory during treadmill running. Brazilian Journal of Motor Behavior,
19(1):e444.
Editor-in-chief: Dr Fabio Augusto Barbieri - São Paulo State University (UNESP), Bauru, SP, Brazil. !
Associate editors: Dr José Angelo Barela - São Paulo State University (UNESP), Rio Claro, SP, Brazil; Dr Natalia Madalena Rinaldi - Federal University of Espírito Santo
(UFES), Vitória, ES, Brazil; Dr Renato de Moraes – University of São Paulo (USP), Ribeirão Preto, SP, Brazil.!
Copyright:© 2025 Dias, Freitas and de Freitas and BJMB. This is an open-access article distributed under the terms of the Creative Commons Attribution-Non
Commercial-No Derivatives 4.0 International License which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source
are credited.
Funding: This work was funded by the Sao Paulo Research Foundation (FAPESP, grants #2021/10105-0, #2020/ 11317-9).
Competing interests: The authors have declared that no competing interests exist.
DOI:!https://doi.org/10.20338/bjmb.v19i1.444