The Magnitude of Temporal–Spatial Gait Asymmetry Is Related to the Proficiency of Dynamic Balance Control in Children with Hemiplegic Cerebral Palsy: An Analytical Inquiry

Abstract

Children with hemiplegic cerebral palsy (hemi-CP) frequently experience deficits in dynamic balance, a crucial factor influencing gait function. This imbalance can manifest as temporal–spatial gait asymmetry, where movement patterns differ between the affected and less affected sides. This study investigated how temporal–spatial gait asymmetries and dynamic balance are associated in children with hemi-CP. Eighty-five children with hemi-CP (age: 13.27 ± 1.72 years) were included. The temporal (AI_Temporal) and spatial (AI_Spatial) gait asymmetry indices were, respectively, computed with reference to the swing time and step length of affected and less affected sides, which were collected through a 3D gait analysis. Measures of dynamic balance included the directional dynamic limit-of-stability (D-LOS_directional) assessed across multiple directions (forward, rearward, affected, and less affected) and the overall dynamic limit-of-stability (D-LOS_overall) during static stance, in addition to the heel-to-heel base of support (BOS_H-to-H) during walking, the dynamic gait index (DynGI), and the Timed Up and Down Stair (TUDS) test.The D-LOS_overall correlated negatively with the temporal (r = −0.437, p < 0.001) and spatial (r = −0.279, p = 0.009) asymmetries. The D-LOS_directional (forward, rearward, affected, and less affected) correlated negatively with temporal asymmetry (r ranged from −0.219 to −0.411, all p < 0.05), but only the D-LOS_directional rearward (r = −0.325, p = 0.002) and less affected (r = −0.216, p = 0.046) correlated with spatial asymmetry. The BOS_H-to-H correlated positively with both temporal (r = 0.694, p < 0.001) and spatial (r = 0.503, p < 0.001) asymmetries. The variation in D-LOS_overall and BOS_H-to-H accounted for 19.1% and 48.2%, respectively, of the variations in the temporal asymmetry and 7.8% and 25.3% of the variations in the spatial asymmetry. The findings of this study suggest that dynamic balance control is related to the magnitude of temporal–spatial gait asymmetries in children with hemi-CP. This evidence lays the groundwork for further research into the mechanism linking gait asymmetry and dynamic balance, potentially leading to a deeper understanding of these impairments, while also highlighting the need for longitudinal studies with the inclusion of a broader population to enhance the generalizability of the findings.

1. Introduction

Cerebral palsy (CP) is a group of permanent disorders related to the development of movement and posture, which results in activity limitations. These disorders arise from non-progressive disturbances in the developing brain during the fetal or early infancy stages. These motor disorders frequently coexist with sensory, perceptual, cognitive, and behavioral disturbances in sensation, perception, cognition, as well as behavior, in addition to epilepsy and secondary musculoskeletal issues [1]. The estimated prevalence rate of CP in high-income countries is 2.1 per 1000 live newborns [2], while the exact rate in middle and low-income countries remains uncertain, but appears to be higher, with greater physical disability [3]. Hemiplegic CP (hemi-CP) is a frequent type of CP in which motor deficits are predominant on one side [4]. This type accounts for 33–39% of all CP occurrences [5].

Children with hemi-CP frequently present with distinctive gait deviations that hinder their mobility and daily functioning. Winter’s classification categorizes these deviations based on sagittal plane kinematics, each reflecting unique functional challenges. Type I is marked as a “drop foot” during the swing phase, while Type II features a true equinus due to spasticity of the gastrocnemius muscles. Type III presents a stiff knee gait resulting from co-contraction of the hamstrings and quadriceps, and Type IV shows more pronounced proximal involvement, resembling patterns seen in spastic diplegia [6]. A thorough understanding of the interplay between these gait deviations and other clinical impairments—such as muscle weakness, spasticity, restricted range of motion, loss of selective motor control, and impaired balance control—is crucial. This knowledge can guide clinicians in developing targeted management strategies that effectively address the multifaceted challenges these children encounter.

Efficient and safe locomotion is the overriding objective of gait training in children with hemi-CP [7,8]. Enhancing gait function ensures higher levels and diversity of activity/participation in their daily life [9,10]. Children with hemi-CP, even after rehabilitation, continue to have spasticity, weakness, and abnormal patterns of muscle activation, all of which have been shown to restrict children’s mobility and impede them from fully participating in daily activities [11]. In addition, children with hemi-CP experience a pronounced asymmetry of body weight distribution between lower limbs during quiet standing. They become more reliant on the less affected limb compensating for the weakness of the affected side [12]. The propensity to keep the body weight shifted on the less affected limb is also seen during ambulation [13,14,15,16]. As a consequence, children may evolve step-length and limb-phasing asymmetries [17]. The ensuing asymmetries are specifically associated with slow, inefficient walking and poor balance control [15,18].

Earlier research has sought to pinpoint key factors influencing gait asymmetry in children with hemi-CP [19,20]. Several impairments, including spasticity, muscle weakness/stiffness, and impaired proprioception, have been marked among the major determinants of gait asymmetries [19]. The latest research suggested that in addition to kinematic deviations of the affected lower limb [20], reduced balance capability during quiet standing is associated with asymmetrical gait patterns [16,18]. Despite the fact that quiet standing balance and gait are two distinct levels of motor functions, they have integrated control mechanisms (i.e., they share some organizational principles, such as the control of the center of mass and the frame of reference for their respective kinematic coordination). Additionally, standing balance and gait have a number of associations and interdependencies at different CNS levels—that is, the afferent feedback serves as a mediator in the relationship between balance and gait regulation and many neural pathways operate to shape the pattern of CNS responses [21].

The foregoing analyses were limited to studying the relationship between static balance and gait asymmetries, where instead of using data from a computerized instrument, the balance was determined by the degree of weight distribution towards the less affected side or a clinical balance scale [16,22]. This, unlike dynamic balance data that are obtained via computerized equipment, does not involve the dynamic limits of postural control—a direction-specific shifting of the body weight (i.e., forward/rearward or the affected/less affected side direction) during standing and/or dynamic balance control in functional contexts such as walking and stair climbing. The earlier findings, therefore, may fall short of fully explaining the link between gait asymmetries and dynamic balance.

A nuanced understanding of the relationship between multifarious balance metrics and gait function is imperative to establish a credible rationale for targeted rehabilitation strategies that enhance functional mobility [23], ultimately helping physical rehabilitation practitioners make informed treatment decisions. Accordingly, this study sought to determine whether an association exists between temporo-spatial gait asymmetries and dynamic balance control—both during standing and during walking—in a convenience sample of children with hemi-CP. The initial hypothesis was that gait asymmetries would be associated with reduced dynamic balance control.

2. Materials and Methods

2.1. Study Protocol and Ethics

This was an analytical inquiry undertaken between December 2019 and August 2021 at Biomechanics Laboratories (Gait/Motion Analysis Lab and Balance Assessment Lab) of the Department of Physical Therapy at Prince Sattam Bin Abdulaziz University, Al-Kharj, KSA. Participants and their parents/legal representatives were informed about the study’s objective and procedures before joining the study and signing a written consent form. The study protocol was approved by the Physical Therapy Research Ethics Committee at the university (RHPT/0019/0040). Procedures conformed with the ethical guidelines of the Declaration of Helsinki, which was released in 1975.

2.2. Participants

Participants were recruited via the Neurology/Physical Therapy departments of four local hospitals in Riyadh/Al-Kharj, KSA. Attending clinicians who normally see children with CP in the outpatient clinics of these hospitals were given a thorough explanation of the study so that they could make appropriate referrals. Participants were further screened for eligibility by the principal investigator, who had more than 20 years of experience in pediatric physical therapy. Inclusion criteria were: (1) a pediatric neurologist-verified hemi-CP diagnosis [24]; (2) age between 8–15 years (since walking behavior in this age group tends to be biomechanically more consistent with the adult patterns) [19,25]; (3) motor function level I or II per the gross motor function classification system (GMFCS) [26], representing children with mild-to-moderate motor impairments who are capable of independent ambulation, which allows for investigation of gait asymmetry and dynamic balance in more active and mobile group; (4) mild spasticity, that is, a spasticity grade of 1 or 1+ on the modified Ashworth scale (MAS) [27]—assessed through hip adductors, knee extensors and flexors, and ankle plantar flexors—to minimize variability and potential confounding factors associated with the more severe spasticity; and (5) adequate mental capacity—this was verified if the children’s records contained pertinent evidence proving normal intellectual function, if they were capable of perceiving, comprehending, and following instructions effectively during screening, and if they were enrolled in regular traditional school classes where they could focus, interact, and work collaboratively with others. Exclusion criteria were neurolytic blocking agents in the past six months, corrective musculoskeletal surgery through the past year, severe contractures, leg-length discrepancy, and visual/auditory deficits.

Power Analysis and Sample Size Determination

The sample size was determined depending on a 95% confidence interval (95%CI: 0.275–0.646) for a Pearson correlation coefficient representing the relationship between dynamic balance and temporal asymmetry index. These were obtained by analyzing data from the first 12 observations. A sample size of 68 children was required to produce a two-sided 95%CI with a width equal to 0.371 when the estimate of Pearson’s product–moment correlation was (r = 0.482). However, this study collected data from 85 children, expecting that ~20% of participants might be lost at random. The power analysis was performed using the PASS software, v16.0.12 (NCSS, Kaysville, UT, USA).

2.3. Measurements

Each child participated in two measurement sessions held on two consecutive days. The first session involved 3D gait analysis. The second session included a laboratory-based assessment of dynamic balance, which utilized the Biodex Balance System to determine both the directional and overall dynamic limits of postural stability. This was followed by an evaluation of clinical metrics related to dynamic balance, specifically the timed up and down stairs test and dynamic gait index. A 15- to 20-min rest period was incorporated between these measurements.

2.3.1. Computation of Gait Asymmetry

The 3D VICON MX motion-capture system (Oxford Metrics Ltd., Oxford, UK) was used to collect gait data at a 1200 Hz sampling frequency. Twelve near-infrared cameras (4-megapixel resolution, maximum speed of 370 frames/second) were employed to track 17 light-reflective markers during walking in a 3D coordinate system within a 1.5 mm spatial displacement error. The system was calibrated, and the projection area was tuned as part of the measurement preparation. Markers were affixed to the skin overlaying particular bone landmarks, as previously outlined [20,28]. Participants were instructed to walk barefoot five times along a 10-m walking route at a comfortable pace. Data were processed using the VICON Polygon software (v4.3.2). To minimize the acceleration and deceleration effects, three usable gait cycles were picked up from the middle of the walking route. The average temporo-spatial gait parameters were computed. The spatial asymmetry index (AI_Spatial) was computed for the step length of the affected (StepL_AFF) and less affected (StepL_Less-AFF) sides as follows: AI_Spatial = abs [(StepL_AFF − StepL_Less-AFF)/[(StepL_AFF + StepL_Less-AFF)]. The temporal asymmetry index (AI_Temporal) was computed for the swing times of the affected (SwingT_AFF) and less affected (SwingT_Less-AFF) sides as follows: AI_Temporal = abs [(SwingT_AFF − SwingT_Less-AFF)/[(SwingT_AFF + SwingT_Less-AFF)] [22]. Higher values mean more asymmetrical patterns, and a value of “zero” denotes symmetrical patterns [19,20].

2.3.2. Quantification of Dynamic Balance

The dynamic limit of stability (D-LOS) during static stance was quantified using the Biodex balance system (Biodex Medical Systems, Shirley, NY, USA). The D-LOS designates the maximum excursion of the center of gravity (COG) that children were deliberately able to cover in different directions through the base of support (BOS) without stepping or losing balance. In the D-LOS test, the transfer of the COG to intercept eight consecutive targets arising randomly at intervals of 45° around the center of pressure was assessed in terms of timing and accuracy. These targets, according to the manufacturer, emerge at 50% of the maximum possible COG excursion, which depends on each child’s height.

Participants stood on the balance plate in a natural stance, barefoot, with their feet hip-width apart and their arms at their sides. Then, they were directed to lean their bodies to the furthest extent to move the COG (represented by a cursor on a display panel) toward a target, and then to return to the center before the next target was displayed. Children were taught to perform the test as accurately and quickly as was feasible while maintaining a straight body and employing their ankles as key axes of movement. Throughout the test, children were observed by the assessor to ensure that the motions were dominant at their ankles and there were no wide segmental motions at the trunk, hip, or knee. If the test revealed noticeably increased COG trajectories (the biomechanical expression of large segmental motion), it was repeated to make sure of the correct execution of the test without segmental motions. Once the eight targets had been accomplished, the test was terminated. Three trials were allowed, and the average score was recorded. For the purpose of this study, the D-LOS_directional (the average directional control score for the forward, rearward, affected, and less affected directions only) and the D-LOS_overall were used for statistical computations. Higher D-LOS scores indicate better dynamic balance control. Below are the algorithms for computing the D-LOS_directional and D-LOS_overall scores [14,29,30].

$${\mathrm{D}\text{-}\mathrm{LOS}}_{\mathrm{directional}}\left(\%\right)={\displaystyle \frac{\mathrm{Rectilinear} \mathrm{distance} \mathrm{to} \mathrm{target}}{\mathrm{Entire} \mathrm{distance} \mathrm{covered}}} \times 100$$

$${\mathrm{D}\text{-}\mathrm{LOS}}_{\mathrm{overall}} (\%)={\sum }_{\mathrm{i}=1}^{\mathrm{i}=4}({\mathrm{D}-\mathrm{LOS}}_{\mathrm{directional}} \mathrm{score}) ÷ 4 (\mathrm{Mean} \mathrm{of} \mathrm{four} \mathrm{targets})$$

In addition, the heel-to-heel BOS (BOS_H-to-H; the perpendicular distance between the heel center of one footfall at the affected side and the progression line formed by two footfalls on the less affected side) [30,31], as one of the variables measured through gait analysis, was employed to assess the ability to maintain dynamic balance during walking. A larger BOS_H-to-H value indicates a lower dynamic balance competence.

2.3.3. Clinical Metrics of Dynamic Balance

The dynamic gait index (DynGI) was used to evaluate the dynamic balance control and risk of falling while walking. It is a valid, performance-based, and simple-to-use instrument for assessing the children’s potential to modify their balance, not only during habitual steady-state walking but also for walking during more challenging tasks (i.e., in the presence of external demands), in children from 8 to 15 years old [32]. The DynGI is composed of eight items that are assessed on a four-point ordinal scale (0–3, with 0 representing severe impairment and 3 indicating normal performance). The DynGI items include the following: leveled-surface walking, walking pace change, horizontal and vertical head turns, stepping over and around obstacles, walking and pivot turns, and stepping up/down stairs. The maximum DynGI score is 24, which indicates that the child is able to walk safely [32].

Another clinical metric for assessing dynamic balance capabilities was the Timed Up and Down Stairs (TUDS) test. The TUDS test has been demonstrated to be valid and reliable in children with typical development and children with CP whose motor function is classified as level I or II on the GMFCS [33]. The test was conducted on a flight of stairs with 14 steps (each measuring 20 cm in height). For testing, children stood up 30 cm away from the bottom step. They were then instructed to climb up the stairs as quickly as they felt safe, turn around a mark on the top of the stairs, and go all the way downstairs until both feet hit the starting point. Children were free to use any stair-climbing strategy they wanted (examples of these are, skipping steps, running up the stairs, using a step-to or foot-over-foot sequence, or others), but were advised to face the moving direction (that is; to face up and down as they traverse the steps, rather than to the side). They were given a signal “ready, 1, 2, 3, and go”. The tests were conducted while wearing shoes, but not lower limb orthotics. The time (second) between the “go” signal and the landing was measured through a stopwatch. Shorter times suggest a higher performance level.

2.4. Data Analysis

The NCSS Statistical Program for Windows, version 11.0.13 (NCSS Statistical Software©, Kaysville, UT, USA) was used for all statistical analyses, and GraphPad PRISM 9 (GraphPad Software Inc.; San Diego, CA, USA) was used for graph generation. The evidence against the null hypothesis was indicated by a p-value of 0.05 or lower. Descriptive statistics (mean ± StDev and min/max observations) were computed to provide a comprehensive overview of the key characteristics. The Pearson correlation coefficient was used to characterize the degree/direction of the relationship between dynamic balance and temporo-spatial gait asymmetries. Linear regression analysis was performed to determine how dynamic balance affected gait asymmetries.

3. Results

Eighty-five children completed the required measurements, and their data were included in the analysis. Demographic and clinical characteristics of the participating children are outlined in

Table 1. Characteristics (demographic, anthropometric, and clinical) of the participating children.

Variables	Values	Min–Max
Age, years	13.27 ± 1.72	8–15
Gender (boys/girls), n (%)	54 (63.5)/31 (36.5)	NA
Height, m	1.39 ± 0.10	1.19–1.61
Weight, Kg	41.87 ± 6.53	30–61
BMI. Kg/m2	21.41 ± 1.34	18.97–24.13
Side affected (RT/LT), n (%)	39 (45.9)/46 (54.1)	NA
Spasticity per MAS (1/1+), n (%)	48 (56.5)/37 (43.5)	NA
GMFCS level (I/II), n (%)	57 (67.1)/28 (32.9)	NA
Insult/MalDev site (C/SC/C+SC), n (%)	19 (22.3)/56 (65.9)/10 (11.8)	NA
Insult type (WMI/GMI/B-MalDev), n (%)	53 (62.4)/23 (27.1)/9 (10.6)	NA
AFO use (yes/no), n (%)	29 (34.1)/56 (65.9)	NA

Age, height, weight, and BMI are shown as mean ± standard deviation. Other variables are listed as frequency (percentage). Abbreviations: RT/LT: right/left side, MAS: Modified Ashworth Scale, GMFCS: gross motor function classification system, MalDev: maldevelopment, C: cortical, SC: subcortical, WMI: white matter insult, GMI: grey matter insult, B-MalDev: brain maldevelopment, AFO: ankle–foot orthosis, NA: not applicable.

, whereas the descriptive statistics related to gait and dynamic balance measurements are summarized in

Table 2. The temporo-spatial gait parameters, asymmetry indices, and dynamic balance variables in the participating children.

Variable		Mean ± StDev	95% LCL, UCL for Mean	Min–Max	p-Value
Temporo-Spatial Gait Parameters
Walking speed, cm/s		92.13 ± 15.25	88.84–95.42	48.11–120.42	NA
StepL, cm	Affected	0.62 ± 0.05	0.61–0.63	0.48–0.71	˂0.001 **
	Less affected	0.49 ± 0.04	0.48–0.50	0.39–0.57
SwingT, s	Affected	0.49 ± 0.06	0.47–0.50	0.34–0.64	˂0.001 **
	Less affected	0.38 ± 0.05	0.37–0.39	0.29–0.51
Asym. indices	AISpatial	0.11 ± 0.04	0.10–0.12	0.03–0.19	NA
	AITemporal	0.12 ± 0.05	0.11–0.13	0.01–0.23	NA
Dynamic balance measures
D-LOSdirectional	Forward	43.88 ± 5.46	42.70–45.10	27–58	NA
	Rearward	42.51 ± 6.21	41.17–43.85	28–54	NA
	Affected	46.76 ± 8.04	45.03–48.50	27–65	NA
	Less affected	54.68 ± 6.32	53.32–56.05	41–68	NA
D-LOSoverall		46.96 ± 4.25	46.04–47.87	34–59	NA
BOSH-to-H, cm		17.31 ± 4.69	16.24–18.38	6.51–26.50	NA
DynGI		18.67 ± 2.12	18.21–19.13	14–22	NA
TUDS, s		17.12 ± 4.55	16.14–18.10	8.5–27.4	NA

Abbreviations: StepL: step length, SwingT: swing time, Asym: asymmetry, AISpatial: spatial asymmetry index, AITemporal: temporal asymmetry index, D-LOS: dynamic limit of stability, BOSH-to-H: heel-to-heel base of support, DynGI: dynamic gait index, TUDS: timed up and down stair test, NA: not applicable. StDev: standard deviation, LCL: lower confidence limit, UCL: upper confidence limit, p-value indicates the difference between affected and non-affected sides, ** significant at p ˂ 0.01.

.

The analysis uncovered a significant negative correlation between the AI_Spatial and some measures of the D-LOS_directional, specifically the rearward [r = −0.325; p = 0.002] and less affected direction [r = −0.216; p = 0.046] and the D-LOS_overall [r = −0.279; p = 0.009]. Furthermore, there was a significant positive correlation between AI_Spatial and BOS_H-to-H [r = 0.503; p < 0.001] and TUDS duration [r = 0.265; p = 0.014] and a negative correlation between AI_Spatial and DynGI [r = −0.280; p = 0.009] (

Table 3. Associations between dynamic balance metrics and temporo-spatial gait asymmetry indices.

	AISpatial		AITemporal
	r (95%CI)	p-Value	r (95%CI)	p-Value
D-LOSdirectional
Forward	–0.204 (0.009, −0.399)	0.059	–0.219 (–0.007, −0.411)	0.043 *
Rearward	–0.325 (–0.119, −0.501)	0.002 **	–0.411 (–0.216, −0.572)	<0.001 **
Affected side	–0.028 (0.185, −0.239)	0.797	–0.272 (–0.062, −0.456)	0.012 *
Less affected side	–0.216 (–0.119, −0.501)	0.046 *	–0.243 (–0.031, −0.432)	0.025 *
D-LOSoverall	–0.279 (–0.003, −0.409)	0.009 **	–0.437 (–0.245, −0.592)	<0.001 **
BOSH-to-H	0.503 (0.323, 0.645)	<0.001 **	0.694 (0.563, 0.791)	<0.001 **
DynGI	–0.280 (–0.071, −0.463)	0.009 **	–0.297 (–0.089, −0.478)	0.006 **
TUDS	0.265 (0.055, 0.451)	0.014 *	0.214 (0.001, 0.406)	0.049 *

Abbreviations: AISpatial: spatial asymmetry index, AITemporal: temporal asymmetry index, D-LOS: dynamic limit of stability, BOSH-to-H: heel-to-heel base of support, DynGI: dynamic gait index, TUDS: timed up and down stair test, r: Pearson correlation coefficient, CI: confidence intervals. * Significant at p < 0.05, ** significant at p < 0.01.

).

A significant negative correlation was observed between AI_Temporal and all measures of the D-LOS_directional [forward: r = −0.219; p = 0.043, rearward: r = −0.411; p < 0.001, affected: r = −0.272; p = 0.012, and less affected direction: r = −0.243; p = 0.025], as well as the D-LOS_overall [r = −0.437; p < 0.001]. There was also a significant positive correlation between the AI_Temporal and the BOS_H-to-H [r = 0.694; p < 0.001] and TUDS duration [r = 0.214; p = 0.049], as well as a negative correlation between AI_Temporal and DynGI [r = −0.297; p = 0.006] (Table 3).

The estimated change in AI_Spatial per unit change in D-LOS_overall was –0.0024. The D-LOS_overall accounted for ~7.8% of variations in the AI_Spatial (R² = 0.0776). The value of AI_Spatial when the D-LOS_overall was zero was 0.2250. The straight-line equation relating AI_Spatial and D-LOS_overall was estimated as: AI_Spatial = (0.2250) + (−0.0024) D-LOS_overall (Figure 1a). The analysis also revealed that the estimated change in AI_Spatial per unit change in BOS_H-to-H was 0.0038. The proportion of variation in AI_Spatial that can be explained by the variation in BOS_H-to-H was ~25.3% (R² = 0.2533). The value of AI_Spatial when the BOS_H-to-H was zero was 0.0448. The straight-line equation relating AI_Spatial and D-LOS_overall was estimated as: AI_Spatial = (0.0448) + (0.0038) BOS_H-to-H (Figure 1b).

The estimated change in AI_Temporal per unit change in D-LOS_overall was –0.0050. The D-LOS_overall accounted for ~19.1% of variations in the AI_Temporal (R² = 0.1909). The y-intercept, the value of AI_Temporal when the D-LOS_overall is zero, was 0.3582. The regression-line equation relating AI_Temporal and D-LOS_overall was estimated as: AI_Temporal = (0.3582) + (–0.0050) D-LOS_overall (Figure 2a). The estimated change in AI_Temporal per unit change in BOS_H-to-H was 0.0069. The proportion of variation in AI_Temporal that could be explained by the variation in BOS_H-to-H was ~48.2% (R² = 0.4821). The value of AI_Temporal when the BOS_H-to-H was zero was 0.0036. The regression-line equation relating AI_Temporal and BOS_H-to-H was estimated as: AI_Temporal = (0.0036) + 0.0069) BOS_H-to-H (Figure 2b).

4. Discussion

The fundamental objective of the current work was to determine if an association existed between the proficiency of dynamic balance control (on standing and during walking) and the asymmetry patterns of gait (spatial and temporal) in children with hemi-CP. The most evident finding to emerge from this analytical study is that the dynamic balance control in a static stance or during walking is significantly related to the magnitude of temporo-spatial asymmetries of gait. More specifically, higher D-LOS scores (i.e., shifting body weight in different directions more efficiently) during static stance, narrower BOS_H-to-H distance during walking, lower DynGI score, and shorter TUDS duration are associated with the more symmetrical patterns. These findings substantiate the initial hypothesis that gait asymmetries are intricately related to diminished dynamic balance control in children with hemi-CP.

It is generally recognized that children with hemi-CP have variable degrees of postural control and balance impairments, which are key elements of locomotor issues [34]. However, no research exists exploring the relationship between dynamic balance and temporospatial gait asymmetries. Children with hemi-CP tend preferentially to support more body weight on the less affected lower limb while standing and during walking [12]. This differential loading resulted in decreased balance control of the affected limb and was intrinsically linked to the deteriorated temporo-spatial asymmetries of gait in such a patient population [16,17].

The current findings revealed that temporal gait asymmetry decreases among children who have a greater capacity for adjusting their COG within a fixed BOS in both forward/rearward and affected/less affected side directions without losing their balance. Children with hemi-CP mostly have trouble shifting their weight toward the affected lower limb, resulting in more instability during the stance phase, and leading eventually to a shorter swing phase on the less affected side [12,16,18]. Nevertheless, the capacity to transfer body weight in antero-posterior and lateral directions involves the affected side’s direction and permits further stance stability during walking; therefore, the less affected lower limb’s swing phase is lengthened and temporal asymmetry is diminished. These findings are further corroborated by the evidence from a preceding analysis which pointed out that an augmented weight-bearing capability of the affected limb promotes the enhancement of the swing time symmetry [35].

It has also been observed, surprisingly, that spatial gait asymmetry is lower in children who are able to shift their COG than it is in children who are unable to shift their COG toward specific directions (rearward or less affected sides’ direction) rather than others. Seemingly, different factors like ankle plantar flexors’ spasticity, plantar and dorsiflexors’ swing trajectory contribution, and kinematic deviations of the lower limbs have greater associations with the spatial asymmetry of gait (although not directly measured in this study) and play even more important roles in determining the spatial asymmetry patterns than the ability to maintain dynamic balance (just as body weight shifting) [19,20]. The plantar flexor spasticity brings on further instability during the stance phase and inefficient weight support on the affected limb. Therefore, children with hemi-CP are likely to reduce the StepL_Less-AFF in comparison with the StepL_AFF [19]. Insufficient strength of the ankle joint muscles could be linked to the shortened StepL_AFF. The ankle dorsiflexors prevent foot dragging along the ground and inhibit premature foot contact [36], therefore extending the StepL directly, whereas the plantar flexors act during the pre-swing phase to lengthen the StepL by creating the ground reaction force [36,37]. Thus, factors that directly influence the StepL, like the spasticity of the plantar flexors and the strength of the dorsi and plantar flexors, have stronger associations with the spatial asymmetry of gait than dynamic balance control during static standing [19].

The current study determined that the temporo-spatial gait asymmetries decreased in children who exhibited narrower BOS_H-to-H distributions, greater DynGI scores, and shorter TUDS durations. That is, children with more proficient dynamic balance functions were capable of walking within a narrow BOS and navigating stairs more quickly, implying that they were able to effectively maintain their balance during either walking or performance of other locomotor activities by shifting their COG in all forward/rearward and affected/less affected directions. These findings are in accordance with prior studies on adults with post-stroke hemiplegia, which found that patients who experience more temporo-spatial gait asymmetries walk with a wider step, which is linked to decreased balance competencies [30,38].

4.1. Study Merits and Literature Contribution

This study constitutes a groundbreaking investigation into the complex interplay between temporo-spatial gait asymmetries and dynamic balance control specifically within the context of the pediatric population with hemi-CP. To the author’s knowledge, it represents the first comprehensive exploration of this critical relationship, thereby offering novel insights and contributing significantly to the existing body of literature. While previous studies have examined these variables independently [13,34,39,40,41], the findings demonstrated herein underscore the interrelatedness of the gait mechanics and balance, highlighting that impaired dynamic balance can significantly impact the pattern of gait symmetry. Additionally, this study utilized a relatively large sample and a high power (95%), crucial elements that allow for the detection of smaller but meaningful relationships and significantly enhance the reliability and validity of the findings. Moreover, the current findings advocate for the implementation of integrated assessment approaches in clinical practice, where both gait asymmetries and dynamic balance deficits are evaluated concurrently. This holistic perspective is essential for informing targeted therapeutic interventions that address both aspects simultaneously.

4.2. Study Limitations

Despite the significance of the findings demonstrated herein, certain limitations may apply to their generalizability. The paucity of comparable results is one source of uncertainty. Therefore, further research into the relationship between gait symmetry and dynamic balance control in children with hemi-CP would contribute to developing a higher level of precision in this issue. The study included children whose brain insult/mal-development site was heterogeneous (i.e., cortical, subcortical, or a combination of both), and temporo-spatial parameters of gait might differ from one another. Therefore, studies with larger samples stratified by the brain insult/mal-development site are warranted to affirm or disprove the current findings. The study analyzed data from children within the 8–15 years age bracket. So, the question of whether children with hemi-CP in younger or older age groups would have similar outcomes remains unanswered, and this could be a starting point for further analyses in the future. A note of caution is also due here since the sample was restricted to children with mild spasticity (i.e., level 1 or 1+ on the MAS), which may make the findings less generalizable to all children with hemi-CP, especially as some of those children could have more spasticity in at least some of the joints (hamstrings, rectus, or gastrocnemius) in the lower extremities. Therefore, additional work is required to confirm the viability of the current findings for children with higher spasticity. The current findings should also be interpreted with caution given the fact that some children (about 1/3) were assessed for gait and dynamic balance barefoot, although they occasionally wore an ankle–foot orthosis. In future investigations, it might be possible to collect/analyze data with and without the use of orthosis, thereby drawing a more definitive conclusion. Since no reference values exist for the dynamic balance measures employed in the present study, it was not possible to indicate what was expected from children aged 8–15 and functioning at GMFCS level I or II; this could, therefore, be a line of inquiry for upcoming studies. Finally, the proprioceptive function was not considered when children were selected to join the study. Gait asymmetries have been observed to worsen with increasing degrees of position-sense error [13]. Thus, forthcoming research should consider this factor, thereby providing more definitive evidence.

4.3. Clinical and Research Implications

The study’s findings on the relationship between gait asymmetry and dynamic balance can guide clinicians and physical rehabilitation practitioners in developing targeted therapeutic interventions. An in-depth understanding of specific gait patterns that correlate with balance deficits would help therapists create individualized rehabilitation programs that address both aspects simultaneously, enhancing overall mobility outcomes. Further, the insights from this study can lead to the establishment of standardized assessment protocols that evaluate gait asymmetry and dynamic balance, allowing for more accurate monitoring of patient progress and timely adjustments to the treatment plans. Furthermore, identifying specific gait asymmetries that predispose children to balance issues also enables the implementation of preventive strategies, ensuring early intervention before more significant mobility challenges arise.

This study also provides a foundation for future research to explore the underlying mechanisms linking gait asymmetry and dynamic balance in greater depth. Elucidation of these mechanisms could lead to the design of targeted interventions that address identified deficits. Prospective cohort studies could facilitate the monitoring of gait asymmetry and dynamic balance, contributing to a better understanding of the long-term trajectory of these impairments in children with hemiplegic cerebral palsy. Longitudinal studies are also warranted on how changes in gait asymmetry and dynamic balance evolve with various interventions, offering insights into their effectiveness and long-term impact. Additionally, expanding the research to include broader populations with varying severity levels of hemiplegic cerebral palsy would enhance the generalizability of the findings, informing more inclusive guidelines.

5. Conclusions

This study highlights a relationship between the magnitude of temporal–spatial gait asymmetry and dynamic balance control in children with hemi-CP. Generally, children exhibiting greater dynamic balance control tend to demonstrate more symmetrical patterns. Specifically, the D-LOS_directional data suggest that increasing difficulties in shifting weight toward the affected limb are associated with increased gait asymmetries. However, the observed correlations were weak to moderate, suggesting that while these relationships were notable, they warrant further investigations to fully understand the complexities involved, particularly how dynamic balance control impacts gait symmetry in this population. Future investigations are, therefore, recommended to delve into the causal pathway linking gait asymmetry and dynamic balance, as this could deepen our understanding of their interplay in children with hemi-CP.