Neuromuscular mechanisms and anthropometric modifications in the initial stages of independent gait

Gait & Posture 24 (2006) 375–381 www.elsevier.com/locate/gaitpost

Neuromuscular mechanisms and anthropometric modifications in the initial stages of independent gait P.S.C. Chagas a,*, M.C. Mancini b, S.T. Fonseca b, T.B.C. Soares 1, V.P.D. Gomes 2, R.F. Sampaio b a

b

Department of Physical Therapy, School of Medicine, Federal University of Juiz de Fora (UFJF/MG), Brazil Departments of Physical and Occupational Therapy, Federal University of Minas Gerais State (UFMG), Belo Horizonte, Brazil Received 2 August 2005; received in revised form 21 November 2005; accepted 27 November 2005

Abstract The gait acquisition is an important milestone of motor development. Structural modifications observed during this period add complexity to the process, and the child needs to use appropriate neuromuscular strategies to walk independently. The objective of this study was to document the longitudinal modifications in neuromuscular mechanisms and in functional anthropometric measures during independent gait acquisition in typically developing children. Twelve children were followed for 2 months after gait acquisition, with its initial period documented by the standardized test Alberta Infant Motor Scale. Quantification of the EMG signals of six muscles in the right lower limb allowed the calculation of the co-contraction indexes (CCI) considering pairs of antagonistic muscles, representing the hip, knee and ankle joints. The CCIs were summed up to yield a total index. Anthropometric measures were transformed into gravitational torque (mLg) values for stance and swing phases of gait. Statistical analyses included repeated measures ANOVA models with one factor (week post-acquisition) for the dependent variables mLg and normalized CCI (CCI/mLg). A significant increase was observed in the mLg during the period evaluated both in stance and swing phases ( p = 0.0001). In addition, there was a decrease in the value of the normalized hip and total CCIs in both phases of gait ( p < 0.05). The results revealed changes in the neuromuscular mechanisms used by typical children to deal with the demands involved in the process of gait acquisition. # 2005 Elsevier B.V. All rights reserved. Keywords: Gait; Co-contraction; Child development; Motor acquisition

1. Introduction Gait acquisition is a complex process [1], involving different musculoskeletal structures and functions that are self-organized to fulfill a specific goal, in a particular environment. Typical children walk independently around 12 months post-term, starting at the moment the child is capable of taking five independent steps, without support [2,3]. The process of gait acquisition is mostly concentrated in the first 2

* Corresponding author. Rua Tom Fagundes 80, Apto. 402, Bairro Cascatinha, CEP 36033-300, Juiz de Fora, MG, Brazil. Tel.: +55 32 3241 5319; fax: +55 3229 3843. E-mail address: [email protected] (P.S.C. Chagas). 1 Physical therapist 2 Occupational therapist 0966-6362/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2005.11.005

months, which is the time interval for children to develop the abilities necessary to perform a stable locomotion [4]. The acquisition of gait has been characterized as the result of the dynamic interaction of several subsystems in a particular context [4]. This view proposes a non-linear rhythm of development, with differences among individuals [5], underscoring the variability in motor behavior. Thus, new motor acquisitions are the result of the emergence of different strategies adapted to the individual characteristics of the children, task demands and context conditions [5,6]. Gait development involves dealing with modifications of the body, which occur in an accelerated rhythm in the first 18 months of life [7,8]. During the first year of life, body growth favors the emergence of locomotion by making the legs longer than the trunk, decreasing the disproportion of the cephalic dimension in relation to the rest of the child’s body

376

P.S.C. Chagas et al. / Gait & Posture 24 (2006) 375–381

[8,9]. The relationship between children’s anthropometric characteristics and the development of gait shows that changes in child’s weight and height are associated with changes in properties of the musculoskeletal system, including extensibility, stiffness and strength [7,8]. During gait, the child has to move his body mass against gravity and coordinate the movement of the limbs and trunk, while maintaining the stability of all body segments [5,7]. The changes in the musculoskeletal properties associated to anthropometric modifications impose demands to the child in a period in which the pattern of gait is not yet fully established. Therefore, the emergent pattern of gait is specific to the physical capabilities present during the process of development, which has to be adapted to the demands of this complex functional task [1,10]. A strategy commonly used in the process of acquisition of new motor abilities is the restriction of the degrees of freedom [1,10], which minimizes the complexity of the task. Co-contraction of the muscles around the joints that compose a limb is a neuromuscular mechanism frequently reported to fulfill this goal [1,10]. This phenomenon can be observed as a stiffening of the lower limbs, which restricts the movement of the different joints, diminishing the dimensionality of the system [4,10]. In this sense, modifications in the activation of the muscles of the lower limbs may accompany the process of unfreezing the degrees of freedom in the locomotion task. The objective of this study was to document longitudinal changes observed in muscular co-contraction and in gravitational torque values associated with the process of acquisition of gait in typically development children. The hypotheses (H) of this study were: H1. There will be an increase in the gravitational torque (mLg) in both phases of gait (i.e. stance and swing), during the period comprised between the beginning of acquisition and 2 months after the acquisition of gait in children with typical development. H2. There will be a decrease in muscular co-contraction normalized by the gravitational torque (ICC/mLg) in both phases of gait (stance and swing), during the period comprised between the beginning of acquisition and 2 months after the acquisition of gait in typically developing children.

2. Methods 2.1. Participants A convenience sample of 12 children with typical motor development was selected, including children of both sexes, with minimal chronological age of 10 months. The inclusion criteria for participation in this study were: at term birth, absence of complications in the pre, peri and/or post-term periods, birth weight greater than 2500 g and adequate

motor development at 10 months of age (characterized by a score 25% in the Alberta Infant Motor Scale: AIMS [11]). The children could not be making regular use of any kind of medication or present any sensorial disorders. At 10 months of age, the parents were informed that their children could not use a baby walker during the period pre-acquisition until 2 months after acquisition of independent gait, in order to avoid any interference from this equipment in the dynamics of their emergent gait. Prior to children’s participation, their caregivers were informed about the study’s objectives and procedures, and signed an informed consent. The Ethics Review Committee from the Federal University of Minas Gerais (UFMG) approved this project in May 28, 2003. 2.2. Methods During independent gait, an electromyographer (EMG) MP150WSW (Biopac System1 – E.U.A.), connected to a microcomputer was used to collect the electric activity of six muscles of the hip, knee and ankle joints during gait. Six active surface electrodes were attached with antiallergenic adhesive tapes to the mid third of selected muscles on the right lower limb (RLL) of the children [12]. The selected muscles were: gluteus maximus, rectus femoris, vastus lateralis, biceps femoris, lateral portion of the gastrocnemius and tibialis anterior. A ground electrode was positioned on the tibial tuberosity. These measures were performed in seven assessments, beginning at the week in which the child took five independent steps without support, repeated weekly in the first month post-acquisition and every 15 days in the second month post-acquisition. The evaluations of the independent gait were performed on a level ground, regular surface, where the children were stimulated to walk three times without support for a distance of approximately 4 m. A digital video camera Sony1 TRV740 was positioned 2 m away from the area reserved for gait, in order to register the independent gait of the children in the sagittal plane. Video clips were made during the period of data collection. A light circuit with a small LED was custom-made and positioned between the video camera and the EMG amplifier, in a way that the LED could be visualized in the clips. During data collection, this LED was activated simultaneously with the beginning of the EMG recordings, allowing synchronization between the video clips and the signals obtained by the electromyographer. The caregivers accompanying the child remained at the end of the gait area with ageappropriated toys, and stimulate the child to walk independently in his/her direction. At the end of the EMG evaluation, the following anthropometric measures were obtained from each child, according to the criteria defined by Schneider and Zernicke and Jensen [9,16]: body mass (kg), height (m), length of the thigh, leg and foot (m) and segment perimetry (m).

P.S.C. Chagas et al. / Gait & Posture 24 (2006) 375–381

2.3. Data reduction 2.3.1. Definition of gait cycle The video data of each child were transferred to a Macintosh microcomputer via the software IMOVIE1 at 30 frames per second. The data were used to allow the delimitation of the gait cycles. The gait cycle was defined as the period from the initial contact of the right foot until the moment the right leg performed the next contact with the ground [13]. The gait cycle was divided into stance and swing phases. The withdrawal of the right foot from the floor identified the beginning of the swing phase, which continued until the posterior initial contact of this same foot, which identified the beginning of the stance phase [13]. Three gait cycles were chosen in every longitudinal evaluation of each child. Only cycles with regular and stable steps were used for further analyses. The remaining information was discarded. 2.3.2. Transformation of anthropometric data into gravitational torque (mLg) Measures of perimetry and length of the segments in the lower limb, as well as children’s weight and height were transformed into gravitational torque values for each phase of the gait according to models of hybrid pendulums. In the swing phase, the lower limbs act like a simple pendulum consisting of three connected segments (i.e., thigh, leg and foot), which oscillate about the hip joint [14,15]. Considering the stance phase, the mass of the whole body (i.e., one feet, two legs, two thighs and a composite including head, trunk and arms) oscillates about the ankle joint, which can be illustrated by an inverted pendulum [14,15]. The masses of these segments, under the influence of gravity, generate a torque denominated gravitational torque. In order to calculate the gravitational torque, it was necessary to quantify the equivalent masses and lengths of the simple and inverted pendulums. The equation for the gravitational torque is: gravitational torque ¼ m  L  g where, m, mass of the pendulum; L, equivalent length of the pendulum; g, gravity. The mass of the pendulum equivalent (m) is defined as the sum of the masses of each segment that compose it, estimated from the body mass, using an anthropometric table [9,14,15]. The length of the pendulum equivalent (L) corresponds to the distance from its axis of rotation to the point of the pendulum in which the total mass (m) could be physically represented [14]. Estimates of the total mass, center of mass and radius of gyration of each segment, calculated from total body mass and segment lengths with anthropometric data [9,16], were used to calculate the simple pendulum equivalent (lower extremity around the hip) and the inverted pendulum equivalent (whole body around the ankle) lengths using the parallel axis theorem [6,15].

377

2.3.3. Electromyography After the definition of gait cycles, the software Acqknowledge1 (Biopac System) was used for electromyographic data processing. This procedure included filtering and rectification of the signals. Data were collected at a frequency of 1000 Hz, and high pass and low pass filtered with cut-off frequencies of 10 and 200 Hz, respectively. In order to normalize the electromyographic data, the filtered and rectified signals obtained from each muscle were divided by the peak electromyographic value of each child during the selected gait cycle, in every longitudinal evaluation [18]. Muscular co-contraction was calculated as the overlapping area of the normalized EMG signals of the pairs of antagonistic muscles selected at hip, knee and ankle and collected during the gait cycle. This area of overlap represented the intensity of simultaneous muscle activation (co-contraction level) of each pair of muscles. This method was described by Unnithan et al. [19] and it demonstrated a test–retest ICC of 0.957 [17]. The muscle groups were paired for the co-contraction quantification as following: (a) gluteus maximus and rectus femoris; (b) vastus lateralis and biceps femoris; (c) tibialis anterior and lateral portion of the gastrocnemius. The area of overlap was identified by custom-made software (Matlab1) and allowed the calculation of a mean co-contraction value, or co-contraction index (CCI), for each pair of muscle during the stance and swing phases of gait separately [17]. Therefore, this procedure resulted in a CCI for the hip (a), knee (b) and ankle (c) and the sum of these indexes yielded the total CCI of the lower limb in both stance and swing phases of each gait cycle. A study performed by Obusek et al. [20] demonstrated that variations in the mLg, generated by the addition of weights to the lower limbs, resulted in modifications of movement strategies through increases in joint stiffness. In addition, there is evidence that this stiffness is influenced by modifications in muscular co-contraction [14,15]. Consequently, modifications in the mLg may influence the variations in the CCIs. In order to avoid the possible effect of the mLg on the joint stiffness and on muscular cocontraction, the mean values of the CCIs of each joint during both phases of gait, in every longitudinal evaluation, were divided by the gravitational torque (mLg), creating the variable CCI normalized by gravitational torque values (CCI/mLg). The CCI/mLg of each joint was summed to yield the total normalized CCI. This normalization procedure allowed more objective analyses of the strategies used by the child in the process of gait acquisition [20]. 2.4. Statistical analyses Repeated measures analyses of variance (ANOVA) with one factor (week post-acquisition) were used to evaluate changes in the mean normalized CCIs (CCIs/mLg) of the hip, knee and ankle joints and in the total CCI/mLg, in each phase of gait (stance and swing). These analyses were also

378

P.S.C. Chagas et al. / Gait & Posture 24 (2006) 375–381

Fig. 1. Normalized Co-contraction indexes (CCI) [black traced line] and gravitational torque (mLg) [black line] describing (A) total and joint CCI from the (B) hip, (C) knee and (D) ankle for stance and swing phases, during the longitudinal assessments of all children.

0.0560 (0.0150) 0.0002 * 0.0550 (0.0190) 0.0001 * 0.0670 (0.0220) 0.0480 * 0.0580 (0.0210) 0.0005 * 0.0620 (0.0150) 0.0047 * 0.0630 (0.0230) 0.0090 * 0.0780 (0.024)a 0.0026 (0.0009) 0.0258 * 0.0025 (0.0006) 0.0010* a

*

p < 0.05: Significant level after post hoc inferential analyses. Standard deviations for each week.

0.0028 (0.0006) 0.0430* 0.0028 (0.0006) 0.0632 0.0026 (0.0005) 0.0054* 0.0031 (0.0007)a (C) Total CCI/mLg p-Value

0.0029 (0.0010) 0.3003

0.019 (0.0079) 0.0001 * 0.022 (0.0128) 0.0010 * 0.028 (0.0128) 0.1072 0.023 (0.0082) 0.0020 * 0.025 (0.0071) 0.0104 * 0.0010 (0.0001) 0.0001* 0.0010 (0.0002) 0.0003* 0.0011 (0.0002) 0.0043 * 0.0010 (0.0002) 0.0002* 0.0011 (0.0003) 0.0027* 0.0014 (0.0004)a (B) Hip CCI/mLg p-Value

p-Value

0.0010 (0.0002) 0.0003 *

0.034 (0.012)a

0.026 (0.0111) 0.0222 *

2.95 (0.46) 0.0001 * 2.87 (0.48) 0.0001 * 2.85 (0.43) 0.0001 * 2.79 (0.40) 0.0001 * 2.67 (0.33) 0.0002 * 2.58 (0.33) 0.0488 * 2.48 (0.40) a 60.96 (11.3) 0.0001 * 60.27 (11.0) 0.0003* 58.94 (8.7) 0.0245* 59.29 (10.0) 0.0090* (A) mLg

56.90 (9.9)a

57.08 (8.8) 0.8175

59.54 (10.4) 0.0042 *

Week 8 Week 6 Week 4 Week 3 Week 2 Week 1 Swing phase

Week 0 Week 8 Week 6 Week 4

The results of the repeated measures ANOVA demonstrated significant differences in the mean mLgs, throughout the seven longitudinal evaluations, both in the stance (F = 5.789; p = 0.0001) and swing (F = 24.883; p = 0.0001) phases of gait. Post hoc comparisons revealed that in the stance phase, differences in the mean mLgs identified by the bivariate comparisons (i.e., between weeks) were significant, especially between the initial weeks after acquisition (i.e., weeks 0 and 1) and the remaining weeks of the study. However, in the swing phase of gait, these bivariate differences were observed when the measures obtained in the first month (i.e., first 5 weeks) were compared with the remaining weeks post-gait acquisition analyzed. The direction of the differences demonstrated a progressive increase in the mean mLgs. Results from the ANOVA tests performed with the variables CCIs/mLg revealed a significant decrease in the indexes of the hip joint, both in the stance (F = 4.329; p = 0.0010) and swing (F = 3.858; p = 0.0023) phases of gait, and also in the total indexes observed in the stance (F = 2.605; p = 0.0251) and swing (F = 4.271; p = 0.0011) phases. Pre-planned contrast analyses revealed that in the stance phase, the magnitudes of the hip CCI/mLg and of the total CCI/mLg were significantly greater at the week of gait acquisition (week 0) compared to the remaining weeks of the study period. In the swing phase, the magnitudes of the CCIs/mLg of the same joints were also significantly greater at the week 0 when compared to the other weeks, and at the week 4 when compared to weeks 6 and 8. These results are illustrated in Table 1. No other difference in CCI/mLg was observed in the study.

Week 3

3.2. Inferential results

Week 2

All 12 children, 6 males and 6 females, concluded the study. The beginning of independent gait occurred in average at 12 months of age (S.D.: 2 months; minimum: 10 months and maximum: 14 months). At the moment of gait acquisition, every child obtained a score of 54 points in the Alberta Infant Motor Scale (AIMS) [11]. Fig. 1 shows the temporal evolution of the mean CCIs/mLg and mean mLgs of the 12 participant children.

379

Week 1

3.1. Descriptive results

Week 0

3. Results

Stance phase

used to evaluate changes in the mean gravitational torque (mLg) of the children quantified in the stance and swing phases of gait in each longitudinal evaluation. Pre-planned contrasts (i.e., post hoc) were used to locate the differences between each pair of weeks. The results concerning the dependent variables (CCIs/mLg and mLg) were analyzed separately in the stance and swing phases. The level of significance (a) was established at 0.05.

Table 1 Evolution of the mean (A) mLg on stance and swing phases, (B) hip CCI/mLg on stance and swing phases and (C) total CCI/mLg on stance and swing phases, during the longitudinal assessments of all children

P.S.C. Chagas et al. / Gait & Posture 24 (2006) 375–381

380

P.S.C. Chagas et al. / Gait & Posture 24 (2006) 375–381

4. Discussion The results of this study illustrate the modifications in the neuromuscular strategies used by the children in order to explore their capabilities in the presence of constant modifications of their anthropometric properties, during the emergence of independent gait. The transformation of the anthropometric measures into gravitational torque (mLg) values allowed the study of the impact of the child’s structural modifications in the process of gait acquisition. Such modifications of the body create a different demand (i.e., changing gravitational torques) from that inherent to the process of acquisition of a new motor competence. This effect may impose perturbations to the overall process of gait acquisition. In this study, the mLg was a good indicator of the anthropometric variation observed in the children during the period of acquisition and development of gait. The anthropometric modifications of the children in the present study seem to have a greater dynamic effect in the swing than in the stance phase of gait, as reflected by the changes in gravitational torque for the two phases. Such result may be attributed to the fact that this variable reflects anthropometric characteristics of different segments in each phase of gait [14]. During the swing phase, the variable mLg is related to the torque generated by the masses and lengths of the segments of the right lower limb of the children under the influence of gravity. In the stance phase, the mLg relates to the torque generated by the masses and lengths of not only the right lower limb, but also of the left lower limb, trunk, arms and head. These results suggest that the anthropometric changes represented by the gravitational torques of the simple pendulum (i.e., right lower limb) were more pronounced throughout the period of this study than the changes represented by the gravitational torque of the inverted pendulum (i.e., whole body). The CCIs normalized by gravitational torque reflected the motor strategies of the children participants of the present study during the process of acquisition and development of independent gait. The CCI/mLg of the hip and the total normalized CCI of the lower limb decreased significantly. A possible explanation for the decrease in the CCI/mLg of the hip could be the inability of the participants to generate a gait pattern with characteristic propulsion at the ankle joint. Conversely, these children generate a propulsive force at the hip joint in order to produce continuous locomotion. The reduction in the relation between CCI and mLg at the hip region allows these children to use the energy conserved by the inertia of this segment. The progressive changes in the CCI/mLg suggest that the children exploited the properties of their musculoskeletal system, allowing the emergency of more effective strategies that take advantage of their available dynamic resources [6]. These progressive decreases in the CCI/mLg occurred in different points in time in each phase of gait. In the first month postacquisition, a decrease in the CCI/mLg of the hip was

observed in the stance phase of gait. In the second month, this decrease was also verified in the swing phase. The hip was used as a source for propulsion both in the stance and swing phases of gait, which reflects the great demand in terms of generating and conserving energy at the hip, in the initial phases of gait acquisition. This result is consistent with those presented by Okamoto et al. [3], who observed a decrease in the patterns of muscular co-contraction, analyzed qualitatively, 1 month after acquisition of independent gait, allowing the children to walk for more prolonged periods. A similar study performed by Clark and Phillips [4], also showed, through kinematic analyses, the unfreezing of degrees of freedom of the lower limbs of three children, in the period of 2–3 months after the beginning of independent gait. In the present study, the muscular co-contraction was considered as one of the mechanisms used by the children to diminish the high dimensionality of learning to walk, enabling the emergence of coordinated motor activity [10,17]. The relation between CCIs and mLg at the knee and ankle joints, both in the stance and swing phases of gait, did not present significant differences throughout the seven longitudinal evaluations. Possible explanation for these results includes the fact that the period of this study might not have been sufficient to show modifications in these variables. In addition, the relatively higher variability observed in the ankle joint compared to the hip, may have made it difficult to demonstrate longitudinal differences in this joint. However, it is possible that the absence of a significant effect in these two joints illustrates the need of the children to maintain them stabilized in the initial stages of acquisition of gait and/ or the lack of participation of these joints in gait propulsion during this period. Therefore, the absence of significant differences in the magnitude of these CCI/mLg during the 2 months post-acquisition period suggests that the children did not take advantage of the inertial forces generated at the knee and ankle joints, as observed at the hip. The results of this study inform that typical children, during gait acquisition, seem to maintain the knee and ankle joints more stiffer, in order to stabilize these joints and diminish the complexity of this activity. On the other hand, the decrease in the relation between CCI and mLg at the hip indicates that these children exploit their dynamic resources, taking advantage of the inertial properties of the lower limbs during the swing phase and of the propulsive activity of the hip during stance. Such results suggest that, during the initial phase of gait acquisition, the typical children learn the dynamics of the most relevant joints for the execution of walking; maintaining restricted the joint movements of the lower limbs that might impose an excessive demand relative to the capability of the children. References [1] Bernstein N. The co-ordination and regulation of movements. London: Pergamon; 1967.

P.S.C. Chagas et al. / Gait & Posture 24 (2006) 375–381 [2] Okamoto T, Okamoto K. Electromyographic characteristics at the onset of independent walking in infancy. Electromyogr Clin Neurophysiol 2001;41:33–41. [3] Okamoto T, Okamoto K, Andrew PD. Electromyographic developmental changes in one individual from newborn stepping to mature walking. Gait Posture 2003;17:18–27. [4] Clark JE, Phillips SJ. A longitudinal study of intralimb coordination in the first year of independent walking: a dynamic systems analysis. Child Develop 1993;64:1143–57. [5] Smith LB, Thelen E. Development as a dynamic system. Trends Cognitive Sci 2003;7(8):343–8. [6] Fonseca ST, Holt KG, Fetters L, Saltzman E. Dynamic resources used in ambulation by children with spastic hemiplegic cerebral palsy: relationship to kinematics, energetics, and asymmetries. Phys Ther 2004;84(4):344–58. [7] Adolph KE, Avolio AM. Walking Infants adapt locomotion to changing body dimensions. J Exp Psychol 2000;26(3):1148–66. [8] Bartlett DJ. Relationship between selected anthropometric characteristics and gross motor development among infants developing typically. Pediatr Phys Ther 1998;10:114–9. [9] Schneider K, Zernicke RF. Mass, center of mass and moment of inertia estimates for infant limb segments. J Biomech 1992;25(2): 145–8. [10] Turvey MT, Fitch HL, Tuller B. The Bernstein perspective I: the problems of degrees of freedom and context-conditioned variability. In: Kelso JA.., editor. Human motor control. New Jersey: Lawrence Erlbaum Associates Publishing; 1982. p. 239–52.

381

[11] Piper MC, Pinnell LE, Darrah J, Maguire T, Byrne PJ. Construction and validation of the alberta infant motor scale (AIMS). Can J Public Health 1992;83(2):46–50. [12] Cram JR, Kasman GS, Holtz J. Introduction to surface electromyography. Maryland: Aspen Publishers; 1998. [13] Sutherland DH, Olshen RA, Biden EN, Hyatt MP. The development of the mature gait. London: MacKeith Press; 1988. [14] Holt KG, Obuseck JP, Fonseca ST. Constrains on disordered locomotion. A dynamical systems perspective on spastic cerebral palsy. Hum Mov Sci 1996;15:177–202. [15] Holt KG, Butcher R, Fonseca ST. Limb stiffness in active leg swinging of children with spastic hemiplegic cerebral palsy. Pediatr Phys Ther 2000;12:50–61. [16] Jensen RK. Body segment mass, radius and radius gyration proportions of children. J Biomech 1986;19:359–68. [17] Fonseca ST, Silva PLP, Ocarino JM, Ursine PGS. Ana´lise de um me´todo eletromiogra´fico para quantificac¸a˜o de co-contrac¸a˜o muscular. Revista Brasileira Cieˆncia e Movimento 2001;9(3):23–30. [18] Yang JF, Winter DA. Eletromyographic amplitude normalization methods: improving their sensitivity as diagnostic tools in gait analysis. Arch Phys Med Rehab 1984;65:517–21. [19] Unnithan VB, Dowling JJ, Frost G, Volpe AB, Bar-Or O. Cocontraction and phasic activity during gait in children with cerebral palsy. Electromyogr Clin Neurophysiol 1996;36:487–94. [20] Obusek J, Holt KG, Rosenstein R. The hybrid mass-spring pendulum model of leg swinging: stiffness as a determinant of cycle period. Biol Cybern 1995;73:139–47.