Asymmetrical spatiotemporal adjustment between the limbs after self-mobilization exercise program

Abstract

[Purpose] To examine the immediate effect of self-mobilization exercise program on spatiotemporal, kinematic, and kinetic parameters in gait. [Participants and Methods] Thirty-two healthy adults were randomly allocated to the Exercise group that performed three warm-up positions and seven small movements or the Control group that lay supine. Before and immediately after the interventions, we assessed participants’ gait at their self-selected comfortable speeds to capture spatiotemporal parameters, kinematics, and ground reaction force (GRF). [Results] In the Exercise group, the double-support and right stance phases of gait remained unchanged, whereas the left stance phase revealed significant changes after the intervention. Furthermore, the Exercise group demonstrated significant decreases in peak hip flexion and increases in peak hip extension bilaterally, while maintaining a more upright trunk position. Neither group showed significant changes in GRFs or joint angular velocities. [Conclusion] The exercise program altered the dynamic control of the trunk-hip motion in gait. The kinematic changes without concomitant changes in force and joint angular velocity potentially reflect an adjustment of the movement control strategy.

INTRODUCTION

Activities in daily living, such as dressing, grooming, and walking, require continuous control of the body’s center of gravity (CoG) to prevent loss of balance. Balance maintenance requires the continuous adjustment of the CoG within the boundaries of the base of support. While this balance control occurs during both static and dynamic activities, the inherent instability and the continuously varying base of support during dynamic movements necessitate a sophisticated coordination system to maintain the CoG. For example, trunk equilibrium and flexion and extension movements can control the CoG over the base of support and contribute to balance maintenance during dynamic activities in young healthy^1,2,3) and patient populations⁴⁾. To control the CoG, appropriate motor commands must be sent primarily to the extensor muscles of the trunk and hip while integrating sensory feedback from the somatosensory, vestibular, and visual systems^5,6,7). Thus, this multisensory integration plays a crucial role in sensing body position and maintaining postural stability during physical activities.

Postural adjustments continuously influence proprioceptive feedback from muscles, joints, and skin receptors to the central nervous system for stability control^{8, 9)}. Common postural deviations, including forward head, rounded shoulder, and lumbar hyperlordosis, can increase postural sway¹⁰⁾ and elevate muscle activation in the cervical and thoracic erector spinae muscles^11,12,13). Similar to the adjustments in a static balance, these neuromuscular adaptations can occur during dynamic balance, such as gait. Specifically, increased trunk flexion during gait is associated with substantial changes in joint kinematics and kinetics at the hips, knees, and ankles, resulting in greater activity in trunk extensors to counteract the anteriorly shifted CoG^14,15,16). These findings suggest that abnormal posture can impose additional muscle demands and alter strategies for stability control in both static and dynamic activities. However, it remains unclear if a correction of postural deviations would cause adjustments in gait kinematics and kinetics.

An exercise program using a cylinder-shaped tube was developed to mobilize the shoulder, hip, sacroiliac, and intervertebral joints. The program consists of three preparatory positionings and seven motions (Supplementary Material Fig. 1 for details). These positionings and motions in the supine position on the tube enhance the stretching of the anterior chest structures and facilitate sacrum nutation, while aligning the thoracic spine curvature with the tube. This self-mobilization exercise program has been shown to induce significant postural improvements^{17,18,19,20,21)}. For example, a 4-week intervention resulted in radiographic evidence of a decrease in the anterior curvature between the lumbar vertebrae and sacrum, indicating decreased lumbar lordosis¹⁷⁾. Similarly, the protrusion of the head and the angle of the anterior pelvic tilt of healthy young adults decreased immediately following the exercise program¹⁸⁾. Additionally, the expandability of the thoracic cavity of healthy elderly adults increased after performing the exercise program twice a day for one week¹⁹⁾.

In addition to the postural changes, the exercise program also reduced postural sway, as indicated by a decrease in the area, distance, excursion, and fluctuation of the plantar center of pressure during static standing²²⁾. However, it is unknown to what extent the self-mobilization exercise program would influence the dynamic control of the body in gait. Postural changes in gait are directly associated with ground reaction force (GRFs), which are used to estimate forces transmitted from the foot to the proximal joints²³⁾, as well as with joint kinematics and kinetics at the hips, knees, and ankles¹²⁾. While the self-mobilization exercise program is known to induce immediate postural changes during static conditions, the research gap lies in the lack of empirical evidence demonstrating the immediate transfer of these induced postural and kinematic adjustments to a dynamic, functional task like gait. Specifically, the effect of these acute postural changes on gait mechanics and external forces remains unquantified.

The purpose of this study was to investigate the immediate effects of a self-mobilization exercise program on gait mechanics. Specifically, we examined a) spatiotemporal parameters, b) kinematics including joint angles and angular velocities, and c) GRFs during gait. The self-mobilization exercise program primarily intervenes in spinal alignment and hip joint angle^{17, 18)}, which is expected to cause changes in the coordination of the foot placement and the limb joint motion during gait^{8, 9)}. Greater hip extension during gait is associated with greater muscle contraction and GRFs¹⁵⁾. Therefore, we hypothesized that the exercise program would improve the pelvic and spinal position, and these structural changes would translate into the following gait pattern: a) altered spatiotemporal parameters; b) greater hip and trunk extension; and c) increases in GRFs resulting from the modified gait kinematics.

PARTICIPANTS AND METHODS

Thirty-two young healthy adults voluntarily participated in this study without financial compensation. Eligible participants had no previous history of major orthopedic injuries, neurological trauma, psychiatric conditions, excessive curvature of the spine, or pregnancy. Eligibility for participation was determined based on self-reported health conditions documented during the screening process. Participants were randomly assigned into one of two groups: the Exercise Group, which performed exercises on a Styrofoam cylinder-shaped tube (98-cm length, 15-cm diameter, durometer hardness: 24°, StretchPole, LPN Corporation, Nagoya, Japan; Exercise group), or the Control Group, which lay supine on a flat surface (Table 1). All participants gave their written informed consent, and the study protocols were approved by the Office of the Institutional Review Board (Approval Number: 20914).

Table 1. Demographic and anthropometric characteristics of participants by group.

	Exercise (n=16)	Control (n=16)
Age (years)	22.4 ± 5.4	22.8 ± 6.3
Sex (male/female)	9/7	9/7
Height (cm)	170.3 ± 10.3	168.6 ± 11.6
Weight (kg)	71.3 ± 17.9	73.7 ± 21.3
Body mass index (kg/m2)	24.3 ± 4.2	25.5 ± 5.1

Exercise group: performed the exercise program on a cylinder-shaped tube; Control group: lay supine on a flat surface. Data are given as mean ± SD except for sex.

This study employed a randomized controlled, pre-test/post-test experimental design to examine the immediate effects of a self-mobilization exercise program on gait mechanics. We used a convenience sampling technique to recruit participants from the university and surrounding communities. This technique was deemed appropriate given the study’s focus on healthy young adults and the need for a controlled laboratory environment.

The exercise program has been previously detailed elsewhere^{18, 19, 24)} and is illustrated in the Supplementary Material Fig. 1. Briefly, participants in the Exercise group performed three preparatory positionings and seven motions with small magnitude while lying supine with the knee and hip flexed on a Styrofoam cylinder-shaped tube. The three positionings involved a static supine position on the tube, held for 30 seconds each without repetition. These positions were characterized by slight shoulder abduction, hip external rotation, and contralateral shoulder abduction and slight hip extension. The seven motions were performed sequentially for approximately 90 seconds each. These motions, characterized by slow movements of the upper and lower extremities (e.g., shoulder abduction/adduction, scapular protraction/retraction, and hip external/internal rotation), were designed to mobilize the shoulder and hip joints. Additionally, the positioning and rolling action on the tube can also promote sacroiliac and thoracolumbar glides. The order of exercises was fixed and consistent across all the participants. The duration of the exercise program was 15 minutes. A single experimenter instructed each participant on how to perform the exercise program. Participants in the Control group rested in a supine position on a flat surface for 15 minutes with their legs and hips flexed, thus mimicking the positions used in the exercise program. This control position was selected to eliminate any slight gliding forces that might otherwise be inadvertently applied to the shoulder, hip, sacroiliac, and intervertebral joints simply by lying supine on the cylinder tube, even in the absence of explicit movements. All participants were required to refrain from strenuous exercise for one week before the study. Before and immediately after their respective interventions, participants walked along an 11.8 m walking path 8 times at their self-selected, comfortable speed. Before data collection and the intervention, a practice period of 3–5 trials was given to become familiar with the environment and establish a comfortable speed. All testing procedures were completed in a single-day session for each participant.

Three-dimensional kinematic data were collected using a 10-camera 3D motion capture system (100 Hz, VICON Ltd., Oxford, UK). GRFs were recorded using three force plates (1,000 Hz, AMTI. Inc., Watertown, MA, USA) embedded flush within a custom-built walkway. To minimize anticipation (e.g., targeting effects), the force plates were camouflaged with the same floor color and material as the surrounding walkway. Gait events, including initial contact and toe-off, were detected using the vertical GRFs and target pattern recognition based on the heel markers’ trajectory. Specifically, we used a 50N vertical force threshold to precisely define initial contact. This ensures synchronization between our kinetic and kinematic data. Reflective spherical markers were placed on bony landmarks to create a 6-degree-of-freedom model for the trunk, the Charnwood Dynamic (CODA) pelvis model, and lower extremities, as previously reported^25,26,27). When the X-Y-Z axes of the pelvis and trunk are aligned with each segment, the position is referred to as absolute upright. These markers remained in place throughout the experiment.

Visual 3D (HAS-Motion, Kingston, Ontario, Canada) was used to create rigid segmental models for the lower limbs, pelvis, and trunk²⁶⁾ and to compute spatiotemporal, kinematic, and kinetic variables. Marker trajectories and GRF were low-pass filtered at 6 Hz and 20 Hz, respectively, using a Butterworth bidirectional filter. The Euler rotational sequence (i.e., sagittal-frontal-transverse) was used to calculate joint angles (Fig. 1A). Dependent variables included: joint angles of the trunk, hips, knees, and ankles; joint angular velocities of hips, knees, and ankles; and spatiotemporal parameters, including gait speed and the durations of the double limb support phase and stance phases. All dependent variables were averaged across 8 trials and normalized to the gait cycle (0−100%), defined from initial contact to subsequent ipsilateral initial contact. Within the gait cycle, we computed the local maximum and minimum values of joint angle, joint angular velocity, and GRFs for analysis (Fig. 1B). To mitigate potential bias, a single investigator processed and analyzed the data after de-identifying the group allocation of the participants.

Fig. 1.

Trunk rotations and ground reaction forces (GRFs).

A: Trunk rotations are shown at the instant of initial contact. Due to the use of a single coordinate system, the trunk rotations and lateral flexion occur in opposite directions. When analyzing each gait cycle, the trunk’s movement toward the supporting limb is considered positive to reflect the range of motion. For instance, positive values indicate ipsilateral motion: lateral flexion to the left and counter-clockwise rotation during left initial contact (left), and lateral flexion to the right and clockwise rotation during right initial contact (right). B: Representative GRFs, normalized to the % of body weight, are shown in medial-lateral, antero-posterior, and vertical directions (from top to bottom) in a 100% gait cycle. F_x Max and F_x Min: maximum and minimum values of GRFs in the lateral and medial directions, respectively. F_y Max and F_y Min: maximum and minimum values of GRFs in the anterior and posterior directions, respectively. F_z Max 1^st and F_z Max 2^nd: the first and second peaks of GRFs at the loading response and terminal stance, respectively, in the vertical direction.

Statistical analyses were performed using RStudio (RStudio, Boston, MA, USA). We used independent t-tests to determine differences in participants’ demographic and anthropometric characteristics between the two groups. A χ² test was used to assess frequency differences in gender between groups.

Additionally, gait speed is expected to exhibit variability across repeated trials within a participant, even when walking at a self-selected pace. The covariate of gait speed allows us to statistically isolate the specific effects of the intervention on the coordination of limbs, independent of fluctuations in gait speed. To control the effects of gait speed, we used analysis of covariances (ANCOVAs) with speed as the covariate to examine changes in the spatiotemporal parameters before and after the intervention within the groups. The relations between pre-test and post-test measures within groups were assessed, where appropriate, with linear regression analysis.

We tested violations of assumptions for the ANCOVA using Shapiro–Wilk test and Q-Q plots for the normality and Levene’s test for the homogeneity of variances (p>0.05). After assumption verification, we used ANCOVAs to examine the effect of the intervention between the groups on joint angles, joint velocities, and GRFs with pre-test measures as the covariate. Post hoc tests, with a Bonferroni adjustment, examined differences between pre-test and post-test measures within the groups. The alpha levels are set at 0.05.

RESULTS

The demographic and anthropometric characteristics of the participants (Table 1) showed no significant difference in age, height, weight, body mass index, or frequency of gender between the two groups (p>0.05). ANCOVA comparing the double-leg support phase and the left and right stance phases within groups between pre-test and post-test measures showed the main effects of speed (p<0.05, Table 2). There were no main effects of time on spatiotemporal parameters for either group (p>0.05), but a significant time ×speed interaction was found in the % stance phase of the left leg for the Exercise group (F_[1,28]=50.16, p=0.019, η²=0.04, Table 2). Significant correlations were found between gait speed and the following spatiotemporal parameters in both groups at both pre-test and post-test: double leg support phase, % stance phase left, and % stance phase right (p<0.05, Fig. 2). The only exception was the post-test stance phase left in the Exercise group. Linear regression analysis in the Exercise group revealed that pre-test measures of % stance phase left decreased as a function of gait speed (F_[1,14]=13.15, p=0.003, r²=0.48, cross and dashed line, Fig. 2). In contrast, post-test measures did not show a significant correlation between the % stance phase left and gait speed (F_[1,14]=1.65, p=0.220, r²=0.11, triangle and solid line, Fig. 2).

Table 2. Analysis of covariance (ANCOVA) comparing the spatiotemporal parameters within groups between pretest and posttest measures.

		Speed main effect			Time main effect			Speed × Time interaction
		F[1,28]	η2	p	F[1,28]	η2	p	F[1,28]	η2	p
Exercise group
	% Double leg support phase	11.45	0.889	0.002	0.238	0.018	0.629	0.188	0.015	0.668
	% stance phase left	975.5	0.995	0.001	3.98	0.001	0.131	50.16	0.040	0.019
	% stance phase right	9.675	0.252	0.004	0.616	0.016	0.439	0.008	0.001	0.929
Control group
	% Double leg support phase	9.26	0.247	0.005	0.192	0.005	0.664	0.034	0.009	0.855
	% stance phase left	5.556	0.159	0.026	1.020	0.029	0.321	0.402	0.011	0.531
	% stance phase right	9.259	0.800	0.005	1.176	0.102	0.287	0.134	0.010	0.717

Fig. 2.

Relations between gait spatiotemporal parameters and gait speed.

Columns represent the Control group (left) and Exercise group (right). Rows show correlations with gait speed for: % double-support phase (top), % right-leg stance phase (middle), and % left-leg stance phase (bottom). Pre-test measures are indicated by cross markers (×) and dashed lines; post-test measures are indicated by triangle markers (△) and solid lines. Each plot includes the slope (β), coefficient of determination (r²), and p-value. n.s.: not significant.

ANCOVA comparing the mean peak joint angular velocities between the Exercise and Control groups revealed that, after adjustment for pre-test measures, there was no significant difference between the groups in peak joint angular velocity across all measured joints (all p>0.05, Supplementary Tables 1, 2).

ANCOVA comparing the mean peak joint angles showed that, after adjustment for pre-test measures, there was a statistically significant difference between the groups in the post-test mean peak joint angles of the left hip (Flexion; F_{[1, 29]}=5.60, p=0.03, η²=0.162, Extension; F_{[1, 29]}=10.34, p=0.003, η²=0.263), the right hip (Flexion; F_{[1, 29]}=24.44, p=0.001, η²=0.457, Extension; F_{[1, 29]}=11.04, p=0.002, η²=0.276), and the trunk (Extension; F_{[1, 29]}=5.83, p=0.022, η²=0.167, Flexion; F_{[1, 29]}=9.80, p=0.004, η²=0.253, Supplementary Table 3). Within the exercise group, post-test measures revealed a significant reduction in the adjusted means of left and right hip flexion, hip extension, and trunk flexion angles compared to pre-test measures (p<0.05, Fig. 3, Tables 3 and 4). The adjusted mean of the trunk extension angle was significantly greater following the exercise intervention (p<0.05, Fig. 3, Table 4).

Fig. 3.

Adjusted means of joint angles between pretest and posttest measures.

Left column: adjusted means of the left hip angle flexion (top) and extension (bottom) between pretest and posttest measures for each group. Middle column: adjusted means of the right hip angle flexion (top) and extension (bottom) between pretest and posttest measures for each group. Right column: adjusted means of the trunk angle flexion (top) and extension (bottom) between pretest and posttest measures for each group. Cross markers (×) and dashed lines denote data of the Control group, whereas triangle markers (△) and solid lines denote data of the Exercise group. Data are mean values averaged across all subjects after adjustment for pretest measures. The vertical lines denote the standard error of the mean. Asterisks denote a significant change (p<0.05) between pretest and posttest measures within each group.

Table 3. Mean peak joint angle for exercise and control groups.

		Exercise group				Control group
		Left		Right		Left		Right
		Pre	Post	Pre	Post	Pre	Post	Pre	Post
Ankle
	Dorsiflexion	77.7 (1.7)	78.0 (1.6)	76.5 (1.6)	76.7 (1.6)	79.7 (1.2)	79.3 (1.1)	78.9 (1.1)	79.2 (1.1)
	Plantarflexion	−45.9 (2.9)	−45.7 (2.9)	−44.2 (2.0)	−43.9 (2.0)	−46.7 (1.9)	−45.7 (1.5)	−46.1 (1.9)	−45.7 (1.9)
	Abduction	5.78 (1.4)	5.77 (1.3)	12.7 (0.9)	12.9 (0.8)	7.05 (1.1)	7.26 (1.2)	10.5 (1.5)	10.7 (1.3)
	Adduction	−9.62 (1.5)	−10.0 (1.2)	−4.50 (0.9)	−4.75 (0.8)	−9.11 (1.2)	−9.35 (1.2)	−6.05 (1.5)	−6.17 (1.4)
	Inversion	11.5 (0.8)	11.2 (1.0)	14.8 (1.3)	14.7 (1.4)	11.7 (1.0)	11.1 (0.8)	15.1 (1.5)	14.3 (1.1)
	Eversion	−2.16 (2.3)	−2.53 (2.0)	2.98 (1.3)	2.31 (1.2)	−1.15 (1.0)	−1.33 (0.9)	2.73 (1.1)	1.58 (1.1)
Knee
	Extension	2.63 (1.2)	2.31 (1.2)	3.34 (0.9)	3.61 (0.9)	2.98 (0.8)	2.61 (0.9)	2.89 (1.2)	2.59 (1.2)
	Flexion	−63.8 (1.4)	−64.0 (1.5)	−63.3 (1.3)	−63.5 (1.4)	−64.3 (0.8)	−64.9 (0.8)	−64.3 (0.9)	−64.7 (0.9)
	Valgus	1.29 (0.8)	1.22 (0.8)	5.33 (0.7)	5.37 (0.8)	4.38 (0.9)	4.33 (0.9)	6.02 (1.0)	5.96 (1.0)
	Varus	−8.92 (1.2)	−9.71 (1.3)	−4.61 (0.6)	−4.53 (0.5)	−7.41 (1.0)	−7.45 (0.9)	−4.61 (0.8)	−4.72 (0.8)
	Internal rotation	12.8 (1.4)	13.1 (1.4)	13.1 (1.3)	13.6 (1.3)	13.0 (1.7)	12.9 (1.5)	12.7 (1.4)	12.3 (1.5)
	External rotation	−1.92 (0.9)	−1.63 (1.0)	−0.35 (1.1)	0.09 (1.0)	−1.97 (1.4)	−1.68 (1.2)	0.42 (1.1)	0.76 (1.0)
Hip
	Flexion	33.6 (2.0)*	31.9 (2.3)*	33.7 (1.9)*	30.2 (2.1)*	31.1 (1.7)	32.1 (1.4)	30.6 (1.5)	31.7 (1.5)
	Extension	−6.82 (1.7)*	−9.59 (2.1)*	−7.11 (1.5)*	−10.1 (1.9)*	−9.52 (2.0)	−8.92 (1.8)	−9.28 (1.9)	−9.04 (1.8)
	Abduction	7.23 (0.7)	7.02 (0.7)	−9.28 (0.8)	−9.04 (0.6)	8.71 (0.8)	8.93 (0.8)	6.61 (1.0)	6.72 (0.9)
	Adduction	−6.40 (0.8)	−6.95 (0.8)	−6.27 (0.6)	−6.04 (0.6)	−5.39 (0.7)	−5.50 (0.7)	−7.04 (0.8)	−7.21 (0.7)
	Internal rotation	13.0 (1.4)	13.9 (1.6)	7.11 (1.7)	7.15 (1.6)	12.7 (1.4)	12.7 (1.4)	7.82 (1.7)	8.11 (1.7)
	External rotation	1.13 (1.7)	1.41 (1.7)	−5.58 (1.8)	−6.00 (1.7)	−0.45 (1.8)	−0.52 (1.8)	−5.39 (2.1)	−5.42 (2.1)

Data are reported in degrees as means (SD). Asterisk: significant difference between pretest and posttest (p<0.05).

Table 4. Mean peak trunk angle for exercise and control groups.

	Exercise group		Control group
	Pre	Post	Pre	Post
Extension	−3.08 (1.4)*	0.20 (2.0)*	−1.64 (2.0)	−1.69 (1.7)
Flexion	7.58 (1.6)*	3.51 (1.9)*	5.37 (1.9)	5.13 (1.6)
Ipsilateral Flexion at LIC	5.72 (0.7)	6.31 (0.8)	5.99 (0.8)	6.45 (0.8)
Ipsilateral Flexion at RIC	6.29 (0.6)	5.82 (0.7)	6.50 (0.6)	6.50 (0.7)
Ipsilateral Rotation at LIC	7.24 (0.7)	7.30 (0.6)	7.73 (0.7)	7.69 (0.6)
Ipsilateral Rotation at RIC	6.88 (0.8)	7.31 (0.7)	5.83 (0.8)	6.06 (0.7)

Data are reported in degrees as means (SD). Asterisk: significant difference between pretest and posttest (p<0.05). LIC: left initial contact; RIC: right initial contact.

ANCOVA comparing the mean peak GRFs between the Exercise and Control groups revealed that, after adjustment for pre-test measures of GRFs, there was no significant difference between the groups in all peak GRFs of both limbs (all p>0.05, Table 5, Supplementary Table 4).

Table 5. Mean peak ground reaction force.

		Exercise group				Control group
		Left		Right		Left		Right
		Pre	Post	Pre	Post	Pre	Post	Pre	Post
Fx (Medio-Lateral)
	Max (Lateral)	2.48 (0.3)	2.38 (0.4)	2.92 (0.3)	2.94 (0.3)	2.86 (0.4)	2.84 (0.3)	3.18 (0.5)	3.70 (0.5)
	Min (Medial)	7.74 (0.5)	7.84 (0.4)	7.15 (0.4)	7.37 (0.5)	7.36 (0.3)	7.41 (0.4)	7.26 (0.3)	7.25 (0.4)
Fy (Anterior-Posterior)
	Max (Acceleration)	21.0 (1.0)	21.6 (1.1)	20.5 (1.1)	20.5 (1.2)	20.9 (0.9)	21.4 (0.9)	21.1 (1.1)	21.9 (1.0)
	Min (Deceleration)	18.4 (0.8)	18.8 (1.0)	18.6 (1.0)	19.2 (1.0)	18.8 (1.2)	19.3 (1.2)	19.8 (1.2)	20.4 (1.1)
Fz (Vertical)
	Max 1st	112 (2.5)	113 (2.8)	112 (2.5)	113 (2.8)	112 (2.0)	114 (2.0)	113 (2.0)	114 (2.0)
	Max 2nd	110 (1.2)	111 (1.3)	110 (1.6)	110 (1.5)	109 (1.3)	110 (1.2)	111 (1.7)	111 (1.6)

Data are reported in % of body weight as means (standard error of the mean).

DISCUSSION

This study examined the immediate effects of the self-mobilization exercise program on gait spatiotemporal parameters, kinematics, and GRFs in healthy young adults. The main findings are, after the self-mobilization exercise program, a) the double-support and stance phases remained unchanged after the exercise program, except for the stance phase of the left leg, b) the peak hip flexion decreased, the peak hip extension increased, resulting in a more upright trunk position, and c) GRFs were not significantly changed. We interpret these findings as a strategic adjustment of movement patterns to maintain constant reactive external forces, potentially indicating a more biomechanically advantageous coordination pattern.

The coordination of lower limb joint motions is fundamentally linked to muscle activation and the body’s CoG²⁸⁾, and the control of the CoG trajectory is a critical determinant of mechanical work during gait^{29, 30)}. Biomechanically, the more upright posture theoretically reduces the demand on the trunk and hip extensor muscles^{15, 31)}. By potentially offloading muscular work onto the skeletal structure, the intervention may optimize body coordination during functional tasks. The ability to adjust kinematics while maintaining joint velocity and GRFs highlights the successful integration of immediate postural changes into a dynamic gait pattern.

The kinematic adjustments observed exclusively in the hip and trunk in the sagittal plane indicate that the intervention’s effects are direction-dependent, i.e., anteroposterior direction. This finding is consistent with previous reports of improved spinopelvic alignment and decreased postural sway following the exercise intervention^{18, 22)}. The positioning and movements during the exercise program explain this directional specificity. While lying supine on the pole, the sacrum, thoracic spinal processes, and head were supported, while the hip, shoulder, and spinopelvic joints remained unsupported. We speculate that this position, combined with the small-magnitude motions, facilitates an accessory posterior joint mobilization. This promotes the stretching of the anterior chest structures and facilitates sacrum nutation¹⁹⁾.

The exercise program altered the relationship between the stance phase and gait speed in the left limb only. While healthy adults typically decrease stance and double-limb support phases as the gait speed increases³²⁾, the left limb in the exercise group did not show this pattern following the intervention (Fig. 2). This interlimb asymmetrical adjustment may represent a functional adjustment in response to the improved range of motion and environmental constraints^{33, 34)}. Specifically, the right limb may have been preferentially used for propulsive force, while the left limb maintained a relatively longer stance phase to ensure stability³⁵⁾. Therefore, the asymmetrical interlimb adjustment following the exercise program may represent a sound coordination strategy among healthy adults intended to maintain the reciprocal pattern of movement in gait.

The interlimb asymmetrical adjustment may also be attributed to different neural control mechanisms between limbs. During unilateral limb tasks, movements of dominant and non-dominant limbs involve asymmetrical hemisphere activations, suggesting a reliance on feedforward and feedback mechanisms, respectively³⁶⁾. Although the laterality of limb control has been extensively studied using the upper extremities^{36, 37)}, reactive changes in response to resistance to the non-dominant leg during gait have also been documented²⁸⁾. Thus, the spatiotemporal adjustments of the left leg found in the present study may reflect a feedback-driven adjustment to the exercise program.

The present study demonstrated the altered interlimb spatiotemporal asymmetry in healthy young adults following the exercise intervention. Quantifying gait asymmetry is essential for assessing therapeutic outcomes, as these patterns reflect functional limb differences, underlying deficiencies due to pathologies, and the overall efficacy of an intervention³⁸⁾. These adjustments in gait kinematics suggest a potential clinical application for populations characterized by excessive trunk and hip flexion, such as individuals with age-associated kyphosis or lower back pain. Given that the program successfully induced a more upright alignment in our healthy cohort, it may be useful for mitigating pathological posture and improving gait efficiency.

However, several limitations should be acknowledged. First, we did not assess limb dominance as an inter-limb comparison, based on dominance was not the primary objective. Thus, we cannot confirm if the observed interlimb asymmetrical adjustments were due to neural or functional differences between the dominant and non-dominant limbs. Limb dominance may play a role in how the central nervous system prioritizes stability versus propulsion during gait coordination³⁵⁾. However, since our exercise intervention consisted strictly of symmetrical movements and positions, the emergence of asymmetrical adjustments remains a noteworthy finding that warrants further investigation in future studies specifically powered to detect dominance-related effects. Second, the sample population consisted of healthy young adults, which may limit the generalizability of these results to clinical populations or elderly individuals with more pronounced gait pathologies or structural asymmetries. Third, we examined only the immediate effects of the self-mobilization program. Consequently, the long-term retention of these kinematic changes and the effective implementation required for sustained improvement remain unknown.

In conclusion, the self-mobilization exercise program altered the relationship between the stance phase and gait speed of the left limb. This interlimb spatiotemporal asymmetry can be interpreted as a functional control strategy used by healthy young adults to maintain postural stability during the dynamic adjustment in gait. Furthermore, the peak joint angles of the hip and trunk decreased after the exercise program, resulting in a more upright trunk position during gait. This kinematic change potentially reduces associated muscle demand by shifting the body’s CoG. The self-mobilization exercise program induced sagittal-plane postural realignment associated with altered trunk-hip kinematics during walking.

Funding

This work has been partially supported by Office of the Vice President for Research and Research Allocation Committee at the University of New Mexico.

Conflict of interest

The authors declare no conflicts of interest.

Funding Statement

This work has been partially supported by Office of the Vice President for Research and Research Allocation Committee at the University of New Mexico.