Variability and coordination patterns of walking with different speeds in active and non-active children with Down syndrome: A cross-sectional case-control study

Abstract

Purpose: Children with Down syndrome (DS) have multiple difficulties in gait pattern. So, the effect of the activity level and speed on the gait coordination and variability was investigated.

Methods: In this case-control observational study, 24 participants in three groups of active and non-active children with DS, and the control group without intellectual disability were asked to walk on a treadmill with two speeds of 0.8 and 1.2 m/s. Continuous Relative Phase (CRP) and variability of CRP in thigh-leg and leg-foot coupling were assessed.

Results: CRP and variability of CRP in the leg-foot coupling in the control group were significantly higher than active and non-active groups with DS. Speed led to increase the CRP of leg-foot in the active group with DS and increase the variability of this coupling in non-active group with DS.

Conclusion: In this study, the activity level provided the compatibility with speed changes of walking in CRP of leg-foot in children with DS.

Down syndrome (DS) is one of the most common genetic causes of intellectual disability (ID; De Graaf et al. 2017). In the children with DS, gait pattern is variable because of the laxity of ligaments, weakness in the neuromuscular structure and balance (Galli et al. 2008). There is a difference between the gait pattern of individuals with DS compared to the typical children without ID, such as increased variability of the step length and width, reduced duration of stride, increased movement variability of the center of pressure (COP; Agiovlasitis et al. 2009, Millar et al. 1993), more variable medio lateral movement of COP and lesser range of motion in the thigh, knee and wrist (Wuet al. 2014). These factors may result in less coordinated gait patterns (Wu et al. 2007) and as a result increase the energy cost of walking (Agiovlasitis et al. 2009, Mendonca et al. 2011).

Coordination is defined as a functional relationship between the components of a system including muscles, joints and nerves. The system functions as a coordinated structure that controls a large number of degrees of freedom when performing various tasks (Bernstein 1967). Gait pattern is one of the most common locomotor skills that require successful interaction with environment (Salami et al. 2014). In the ecological view, physical condition of an individual to perform a task in the environment in which the task is performed determines the constraints and the way the task is performed (Stergiou et al. 2013). Considering the anatomical condition of limbs and joints relative to each other, self-organization of movement patterns may explain the effective motor control in the human body in different physical conditions (McMorris 2004). Therefore, any change in control parameters such as the speed may change the movement stability from out of phase to in-phase or vice versa (Krasovsky et al.2010). Considering that children with DS have multiple difficulties in posture and balance control, (Latash 1992), an improvement in walking coordination helps maintaining the balance and preventing falls (Galli et al. 2008). Literature review reveals that the activity level; (Chow et al. 2008, Agiovlasitis et al. 2009, Agiovlasitis et al. 2015, Chen et al. 2005) and speed change (Behbehani et al. 1990, Giphart et al. 2007) have an effect on the coordination and variability of coordination when walking.

Based on the dynamical systems view, any change in the control parameters such as speed may result in a change in stability condition (Krasovsky et al. 2010). In fact, increasing the variability to a critical point forces the system to change its motor behavior to a more stable (McMorris 2014). According to literatures in patients who suffer from diseases such as Parkinson (Krasovsky et al.2010), myocardial infarction (Balasubramanian et al. 2009), patella pain (Heiderscheit et al. 2002) or DS (Smith et al. 2007, Smith et al. 2011), the stride-to-stride variability is higher compared to the healthy individuals. Another factor that may have any effect on gait pattern and variability is the level activity. Research shows that walking training under the auditory rhythmic conditions(del Olmo et al.2005, Roerdink et al. 2007), weight-bearing treadmill exercise (Daly et al. 2011) and sensory attention focused exercise (Sage and Almeida 2009) improve the coordination and timing of gait pattern in individuals without ID. In active individuals with ID including those with cerebral palsy, treadmill exercise (Johnston et al. 2011) and intense weight-bearing treadmill training program (Provost et al. 2007) had improved the coordination and spatial and temporal parameters. The results of literature review including active and non-active patients with DS shows that aerobic exercise has a positive effect on the leg muscle strength and improves the kinematic parameters and walking speed (Galli et al. 2010, Fox et al. 2019). In a study conducted by Smith et al. (2007), it was found that active individuals with DS trained on a treadmill needed less control strategy to increase their stability and as a result stiffness and step width was decreased. In addition, a 10-week aerobic exercise program decreased the stride width and stiffness, impulses and variability in the subjects with DS (Millar et al.. 1993). The results of a broad survey of multiple peer-reviewed databases including PubMed, Google scholar and so on, showed no existing research that specifically examining the coordination and variability in the children with DS compared to the control children. In addition, the activity level and speed of walking effects on gait coordination and variability of children with DS are unknown.

Walking is a complex oscillatory skill that requires many components acting coordinately with each other. According to the dynamical system view, reaching a control parameter such as speed to a critical level may enforce the system to shift to a more stable pattern (Stergiou et al. 2013). Considering the neurophysiologic problems that children with DS encounter, the question is whether the change in walking speed can lead to a change in speed to a critical level speed and change of walking coordination pattern changes, and whether there is a difference between the active and non-active children with DS? Continuous relative phase (CRP) is the phase relationship between two oscillating segments in the whole cycle and its variability is a kinematics indicator to investigate the variability and the stability of gait coordination patterns (Seay et al. 2006, Bingham et al. 1999). Therefore, the hypothesis of this research was CRP and its variability as coordination indices were different between active and non-active children with DS, in the lower extremity couplings in different speeds of the gait.

Materials and methods

Participants

A nonrandomized availability sampling method was employed to select children with DS (10–15 years-old) with mild ID (intelligent quote (IQ) range of 50–75). This range of intelligence was chosen to ensure that participants were able to understand the instructions. They had a medical record in the school and a psychiatrist had already confirmed DS for each of them. Although we had the IQ scores of all subjects from their medical record in the school, but we tested the IQ with Raven Progressive Matrix Test for all participants.

The sample size for the study was calculated by using G*Power software with statistical power of 0.86 and effect size set to 0.05 (Ulrich et al. 2001), 12 subjects for each group. Unfortunately, this sample size reduced from 36 to 24 subjects because of the unwillingness of a few subjects to walk on the treadmill or not continuing the study. But this sample size was acceptable based on the previous studies (Ulrich et al. 2001). Children with DS (n = 17) were divided into active (n = 9, boys = 4, age = 12 ± 2 y, height = 146 ± 6 cm, weight = 38 ± 5 kg) and non-active (n = 8, boys = 5, age = 12 ± 1 y, height = 150 ± 5 cm, weight = 42 ± 4 kg), plus children without ID were included in a control group (n = 7, boys = 4, age = 12 ± 2 y, height = 137 ± 6 cm, weight = 40 ± 7 kg). The demographic data are shown in Table 1.

Table 1.

Distribution of Tanner’s stages.

Tanner stage	Active DS		Non-active DS		Control
	Boy (n = 4)	Girl (n = 5)	Boy (n = 5)	Girl (n = 3)	Boy (n = 4)	Girl (n = 3)
1	0	0	0	0	0	0
2	0	0	1	0	0	0
3	1	1	1	1	1	1
4	3	2	2	1	2	1
5	0	2	1	1	1	1

The personal information form and questionnaire about the frequency and duration of the physical activity of children (Baecke et al. 1982) in the last six months was completed by their parents. The child was considered as active if he/she had aerobic training in three sessions per week for six months; those who had only one session or no exercise were considered as non-active. The control group included non-active children without ID enrolled in the elementary and secondary schools and voluntarily participated in the project. The weight was measured with light clothes and barefooted by an electronic scale (model 1609 N; Tanita Corp., Tokyo, Japan), and the height was measured using a measuring tape placed on the wall. The participants stood erected with their back and heels against the wall while the examiner placed a ruler on the top of head to measure the height. The weight and height were measured to the nearest precision of 0.1 kg and 0.1 cm, respectively. Also sexual maturation status in boys and girls was determined using self-reported gender-specific Tanner’s scale (Chen et al. 2005). Using this scale, pictures of five developmental stages of pubic hair growth in two genders, female breast and male genitalia development were shown to parents and were asked to describe the appropriate developmental stage of their child. The children were categorized to early-onset, intermediate and late-onset puberty. Distribution of Tanner stages in participants is shown in Table 2. The control group was matched with DS groups in age, height, weight and sexual maturity. Inclusion criteria were having no joint and skeletal disability, no history of fracture in the leg and having normal or corrected vision that was determined based on the medical records available in school. Exclusion criteria were participants unwillingness to walking on the treadmill, lack of cooperation in conducting the study or incomprehension the researcher's instructions.

Table 2.

Mean and standard deviation of variables in the three groups and two speeds.^a

	Non-active with DS				Active with DS				Control		Significance
	Speed 2c	Speed 1b		Significance	Speed 2c	Speed 1b		Significance	Speed 2c	Speed 1b
CRP thigh-leg (°)	59.17 ± 15.04	59.47 ± 9.54		0.11	60.08 ± 16.61	58.68 ± 14.81		0.34	56.74 ± 14.96	60.51 ± 14.02	0.32
Std. thigh-leg (°)	22.78 ± 11.02	22.44 ± 12.51		0.07	18.36 ± 10.89	22.15 ± 12.28		0.12	19.70 ± 18.03	16.70 ± 9.29	0.51
CRP leg-foot (°)	53.22 ± 23.72	36.39 ± 15.20		0.13	43.20 ± 16.89	40.92 ± 17.47		0.16	36.83 ± 19.25	28.49 ± 4.81	0.01*
Std. leg-foot (°)	26.50 ± 17.00	16.96 ± 11.03		0.04*	23.39 ± 14.24	14.47 ± 6.89		0.21	11.62 ± 11.72	6.77 ± 2.05	0.01*

^aResults of Mann Whitney Test.

^bSpeed 1 = 0.8 m/s

^cSpeed 2 = 1.2 m/s.

*Difference is significant. p < 0.025.

CRP: Continuous Relative Phase.

Std: standard deviation.

Instruments

Five reflective markers (16 mm) were attached to the center of the right leg joints at the fifth metatarsal, heel, lateral malleolus, lateral epicondyle of the distal femur and the greater trochanter using the doubled-sided and hypoallergenic adhesive (Shapiro et al. 1981). Anthropometrics measurements were taken for every child to determine the joint centers and markers' positions. preparing the subjects for data collection done by the one examiner and testing and giving the instructions for all subjects done by another examiner. An eight-camera motion capture system (Qualisys: Qualisysinc, Gothenburg, Sweden) was used to collect the data at 100 frames/sec. A Treadmill (VECTO model KL1318) was used to control the speed at 0.8 and 1.2 m/s, respectively (Agiovlasitis et al. 2009) and located in the center of laboratory in appropriate space between cameras. The treadmill was not levelled. The test duration included 40 s in different stages; the first one lasted for 20 s to stabilize the speed, the second one took 10 s to acclimatize the participant with the speed and the final stage lasted for 10 s to analyze the data. During this end stage, five strides were selected for analysis. Each stride included heel contact of a leg until heel contact of the same leg.

Procedure

The testing procedure was explained to both children and parents, and then they completed the written informed consent and the demographic questionnaire in the lab. All the procedures performed in the project conformed to the ethical standards of the institutional and/or national research committee based on the 1964 Helsinki declaration. Before the start of the test, sufficient time was given to the participants for learning how to walk on the treadmill with markers for 10 min. Verbal and food rewards were used as encouragement as well as sweets as a reward to keep the participants motivated until the end of the test. Because most individuals with DS have balance problems compared to individuals without ID (Latash 1992), all subjects were asked to hold the treadmill handle as lightly as possible in order to decrease its effect on CRP and variability of CRP.

In the beginning, the participants walked on the treadmill at speed of 0.8 m/s; then walking speed increased to 1.2 m/s. Every child performed six trials, three trials per stage. The participants were given 3–5 min of rest interval for every trial.

Data reduction

Kinematics data were filtered using a low pass filter and cut off rate of 12 Hz. Position and angular velocity of thigh; knee and foot were calculated across the cycle in the sagittal plane. The first step in calculating CRP is to construct the phase plane of two segments from the position-velocity (or angular position-angular velocity) time series. A crucial step in the calculating CRP involves normalizing the angular position and angular velocity profiles using below methods (Van Emmerik et al. 2014):

$${\theta }^{\mathrm{\prime }}\left(i\right)=2\times [{\theta }_i-\min \left({\theta }_i\right)]/\max \left({\theta }_i\right)-\min {(\theta }_i{)}_{}-\mathrm{ }1$$ (1)

where ${\theta }^{\mathrm{\prime }}$ is normalized angular position, ${\theta }_i$ the original angular position and $i$ is a data point in the cycle.

$${{\mathrm{ }\omega }^{\mathrm{\prime }}}_i=w_i/\max \left[\max \left(w_i\right)-\max \left({-w}_i\right)\right]$$ (2)

where ${\omega }^{\mathrm{\prime }}$ is normalized angular velocity, w is original angular velocity and $i$ is a data point in the cycle.

Finally, the phase angle is obtained for each segment using Equation (3):

$$\mathrm{Ø}\left(i\right)=\hspace{0.17em}{\text{tan}}^{-1}{(\omega }^{\mathrm{\prime }}\left(i\right)/{\theta }^{\mathrm{\prime }}\left(i\right))$$ (3)

where $\mathrm{Ø}$ is the phase angle, ${\omega }^{\mathrm{\prime }}$ is the normalized angular velocity, ${\theta }^{\mathrm{\prime }}$ is the normalized angular position and $i$ is a data point in the cycle.

The CRP angle of the two segmental couplings is then calculated as:

$$\text{CRP}\left(i\right)\text{=}{\mathrm{Ø}}_A\left(i\right)-{\mathrm{Ø}}_B\left(i\right)$$ (4)

where${\text{ Ø}}_A$ is the phase angle of proximal segment and ${\mathrm{Ø}}_B$ is the phase angle of distal segment.

To calculate the phase discontinuity in the final CRP measure, the absolute value of phase angles was determined that included the range of 0 and 180°. When the phase angle was equal to 0°, the movements of the two oscillators were completely in phase. When the phase angle was equal to 180°, the movements of the two oscillators were completely out of phase. Whenever the phase angle was between 0 and 180°, the movements of the two oscillators could be in phase (closer to 0°) or out of phase (closer to 180°; Van Emmerik et al. 2014). Figure 1 shows the process by which the CRP was calculated.

Figure 1.

The process by which the CRP was calculated. (A) The normalized phase plane for each segment. (B) The phase angle derived from each phase plane.(C) difference between two phase angles as the continuous relative phase or CRP.

Data analysis

In the right leg, five cycles were used to analyze the CRP; at the first, data points were interpolated to 100 data points in each cycle. The first five data points with each other averaged the second five data points with each other and so on. Their standard deviations were calculated in the same way. Output included 100 data points of mean and 100 data points of standard deviations. Finally, the average and standard deviation of the final 100 data points was taken and used for further analysis.

Statistical analysis was performed using SPSS software (SPSS, Chicago, IL, version 18.0). Because the presumption of using the parametric tests such as normality was not assumed, non-parametric tests were used to statistical analysis. Normality was not achieved even using the Box-cox conversion and Logarithmic transformations. Kruskal-Walis test was used to analyze the effect of group (with three levels) and Wilcoxon test was used to analyze the effect of speed (with two levels) on the coordination and variability during walking. U-Mann-Whitney test was used for pair wise comparisons between the groups. In all measurement, the significance level was set at alpha = 0.025.

Results

Mean and standard deviation of CRP and the variability of CRP in the thigh-leg and leg-foot couplings are presented in the Table 3. Based on this table, the control and DS active group performed similar adjustments with increasing the speed. As the speed increased, CRP of the leg-foot coupling in the control group increased from 28 to 36° and in the DS active group changed from 36 to 53°. This change in the non-active DS group was small and ranged from 40.9 to 43°. Figure 2 demonstrates CRP of the leg-foot in exemplar participant in three groups.

Table 3.

Results of Kruskal-Wallis test.^a

	Z	Chi-Square	Significance
CRP thigh-leg (°)	−0.12	0.14	0.90
Std. thigh-leg (°)	−1.91	3.97	0.06
CRP leg-foot (°)	−2.69	9.75	0.00
Std. leg-foot (°)	−4.00	20.72	0.00

^aGrouping variable: Group.

CRP: Continuous Relative Phase.

Std: standard deviation.

Figure 2.

CRP of leg-foot in exemplar participant in three groups: (A) participant from the control group, (B) participant from an active group with DS and (C) participant from the non-active group with DS.

The results of Kruskal-Wallis test showed that there was a significant difference in groups for the CRP of the leg-foot coupling${\mathrm{ }(\mathrm{\chi }}_2^{(2)}=9.757\text{ and }p=0.008)\mathrm{ }$ and the variability of the leg-foot coupling${\mathrm{ }(\mathrm{\chi }}_2^{(2)}=20.722\text{ and }p=0.000).$ Results of the Mann-Whitney test showed that in control group, CRP of the leg-foot coupling was significantly lower than both the active (p = 0.007) and non-active (p = 0.004) groups with DS. Also the variability of CRP in the leg-foot coupling in the control group was significantly very lower than both the active (p < 0.001) and non-active (p < 0.001) groups with DS. Groups did not show any significant difference in both CRP and variability of the thigh-leg coupling. Results of the Wilcoxon test showed that increasing the speed induced to the significant increase in the CRP of the leg-foot coupling in active group with DS (p = 0.012). Also, the speed had a significant effect on the increasing the variability of the leg-foot coupling in non-active groups with DS (p = 0.017).

Discussion

The aim of the current study was to examine the coordination patterns and variability of the coordination during walking on the treadmill with two different speeds in active and non-active DS groups and a control group without ID. The results of this study showed that there was a difference between the CRP and variability of CRP in the gait pattern of active and non-active DS groups compared to the control group. This differences were so large. According to the findings, CRP and the variability of CRP of leg-foot coupling in the control group were lower than both DS groups; however, the speed resulted in an increase in the CRP of the leg-foot coupling in the active DS group. In addition, it induced increasing the variability of CRP in the leg-foot coupling in the non-active group with DS.

It seems that two potential factors may have effects on the CRP and variability of CRP, physical and cognitive factors. Agiovlasitis et al. (2015) claims that the higher metabolic energy costs in the gait patterns of the DS subjects compared to the individuals without ID is the result of the altered gait pattern in DS cases. Despite the fact that the variables in the study of Agiovlasitis et al. (2015) included the range of the body center of mass , mediolateral position, anterioposterior velocity and the variability of step length and width, the inefficiency of energy may be triggered by the change in coordination pattern of walking (Schmidt et al. 2013). Based on the report by Jordan et al. (2007) , individuals without ID compared to the DS subjects have minimized level of metabolic energy cost for walking using passive and mechanical features of the leg and show more stable and less variable gait patterns. In the current study, higher variability of the coordination patterns in the leg-foot coupling in both DS groups is consistent with the finding of other studies (Smith et al. 2007, Chang et al. 2009). Wu et al. (2007) and Agiovlasitis et al.(2015) also confirmed these results about difference between gait patterns in the DS individuals and control group without ID. It seems that the differences in some of the anatomical and functional properties such as flat feet deformity, hallux valgus (Galli et al. 2008), increased variability of the length and width of step (Millar et al. 1993) and increased the movement variability of the COP (Agiovlasitis et al. 2009) may explain the higher variability of the coordination of the gait pattern in two groups with DS compared to the control group without ID.

Based on the ecological view, it seems that along the anatomical and functional problems, cognitive factors such as decreased motivation and lack of the fast adaptation with unfamiliar environment in individuals with DS (Latash 1992, Vali Noghondar et al. 2019b) may cause on the CRP and variability of CRP in the gait. Central nervous system regulation of movement efficiency places limitations on movement, with an emphasis on economy rather than locomotive/ambulatory efficacy and aesthetics, (Latash et al. 1996, Latash 2007, Aruin et al. 1997, Vali Noghondar et al. 2019a) if different motor strategies are presented to the central nervous system and implemented during the person tests and repeat by exercise. As a result, the central nervous system may induce differential changes to priorities in the next subsequent motor situations and likely use these novel but preferred new strategies for motor control (Latash 2007, Latash et al.1996).

In the present study, speed did not have a significant effect on the CRP and variability of CRP in two couplings of the control group. This finding is in agreement with the findings reported by Seay et al. (2006) that showed speed had not a significant effect on the coordination patterns and variability of the coordination patterns in the gait of adults without ID. Seay et al. (2006) examined the intra-limb coordination patterns of the leg in transition from walking to running. They reported that the variability of CRP in the thigh-leg and leg-foot couplings in different speeds of gait did not change. In the current study, although changing the control parameter of speed from 0.8 to 1.2 m/s in the control group required a higher amount of energy cost, but it was not a challenging task to change the coordination patterns or the variability. Also no difference was observed may have been because all participants were holding the handrails on the treadmill, which may have affected walking gait. Nonetheless, Giphart et al. (2007) studied the effect of speed on coordination of the gait in two conditions of virtual reality and real, and reported that increasing the speed has effect on increasing the mean relative phase between ipsilateral arm and leg movements in the real condition. So the effect of speed should be studied in the future researches. Also these results were compatible with Behbehani et al. (1990).

Coordination of the gait patterns in the thigh-leg coupling in both groups with DS and control group was similar. It seems that this is due to the high range of motion and instability of the ankle joint (distal joint) compared to the knee joint (proximal joint). The same condition was present for the distal segments because of the instability and more range of the motion. In addition, viewing the angles of CRP in the leg-foot coupling shows that these angles are closer to 0° rather than 180°; therefore, the movement of the two oscillators to each other was in phase.

It seems that the active group with DS attempted to adjust with increasing the speed and this caused a change in the coordination pattern of the gait in the leg-foot coupling. Based on the dynamical systems view, whenever the variability reaches to a critical level, the system may shift to a more stable pattern with less variability (Stergiou et al. 2013). Increasing the joint stiffness in the individuals with DS is a compensation mechanism that represents the muscular weakness (Galli et al. 2008). In this sense, it could be hypothesized that active group with DS decreased the joint stiffness and increased the mobility of the ankle joint compatible with the speed changes; however, non-active group with DS was not able to make the appropriate adaptations to increasing the speed. As a result, no significant increase was observed in CRP of the leg-foot coupling in this group (2.2°). The non-active group with DS showed increased variability of the coordination pattern in the leg-foot coupling with speed increasing. This increase in the non-active group with DS did not induce to change in coordination pattern in leg-foot coupling. In summary, it seems that increasing the speed of gait was sufficient to modify the gait pattern in the DS active group and increasing the variability in non-active group with DS.

There were a few limitations to working with the subjects with DS in the current study. The first limitation was the low sample size. Another limitation was about the aerobic program. Because this research was a retrospective study, we did not supervise during the execution of the aerobic program. Also because of the balance and gait problems of subjects with DS in walking on the treadmill, all the subjects were asked to hold the treadmill handle; so likely this could have affected on the CRP and variability of CRP. Although there was no considerable difference in the participants’ height, it was not possible to adjust the treadmill handle based on the height of the participants individually.

Conclusion

To sum up, the coordination pattern and the variability of the coordination pattern in the control group versus the active and non-active groups with DS in the leg-foot segmental coupling were different. Active DS group attempted induced adaptations to make compatibility with the speed increase by changing the coordination pattern of the leg-foot segmental coupling more through ankle flexibility. The non-active DS group was not able to adapt to the task requirements but had to instead increase the variability. The activity level in the children with DS can is could be compatible with the task needs (like the speed increasing) by adjusting the coordination pattern and variability. In DS patients with problems in balance, posture and gait, these results have important clinical implications for professionals who guide them. For future research, using the equipment and devices such as braces or supramalleolar orthoses combining the training can be explored in people with DS.

Disclosure statement

The authors declare no conflicts of interest.