Research in Developmental Disabilities 35 (2014) 3624–3631

Contents lists available at ScienceDirect

Research in Developmental Disabilities

Gait characteristics of children with cerebral palsy as they

walk with body weight unloading on a treadmill and overthe ground

Melissa L. Celestino, Gabriela L. Gama, Ana M.F. Barela *

Graduate Program in Human Movement Sciences, Institute of Physical Activity and Sport Sciences, Cruzeiro do Sul University, Brazil

A R T I C L E I N F O

Article history:

Received 3 June 2014

Received in revised form 31 August 2014

Accepted 2 September 2014

Available online

Keywords:

Body weight support system

Spatial–temporal parameters

Joint angles

A B S T R A C T

Body weight support (BWS) has become a typical strategy for gait training, in special with

children with cerebral palsy (CP). Although several findings have been reported in the

literature, it remains uncertain how different types of surfaces and gradual amount of BWS

can facilitate the mobility of children with CP. The aim of this study was to investigate gait

kinematic parameters of children with CP by manipulating BWS and two different types of

ground surfaces. Ten children (7.7� 2.1 years old) diagnosed with spastic CP and GMFCS

classification between levels II and IV were asked to walk on a treadmill and over the ground.

In both conditions, BWS was manipulated to minimize gravitational effects and spatial–

temporal gait parameters and lower limb joints were analyzed. The results revealed that the

type of ground surface causes greater impact on the gait pattern of children with CP as

compared to body weight unloading. This finding may provide new insights into the

behavioral heterogeneity of children with CP, and offers critical information to be considered

on interventional programs specifically designed to improve mobility on this population.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

While gait is an activity that most of us take for granted, it is a motor skill that requires an optimal pattern of motorcoordination and involves complex control mechanisms (Inman, Ralston, & Todd, 1994; Winter, 1991). The optimalperformance level might be specified by the interaction of three different categories of constraints (Newell, 1986): organism,including structural features (e.g., body mass and height) and functional features (e.g., synaptic connections); environment(e.g., gravity, ambient temperature, natural light, cultural backgrounds); and task, including goals of the task and rules andimplements specifying response dynamics. Constraints are considered as boundaries that limit an individual’s motion at thesame time that lead to alternative patterns of movement coordination and control; even if a movement is performed underthe same set of environmental and task constraints.

CP has been described as a group of disorders of the development of movement and posture that are permanent and causeactivity limitation, and are attributed to non-progressive disturbances that occurred in the developing fetal or infant brain(Rosenbaum, Paneth, Leviton, Goldstein, & Bax, 2007). The effect of CP on functional abilities varies greatly. Some people are

* Corresponding author at: Rua Galvao Bueno, 868, 138 andar, Bloco B, 01506-000 Sao Paulo, SP, Brazil. Tel.: +55 11 3385 3103.

E-mail addresses: ana.barela@cruzeirodosul.edu.br, ambarela@gmail.com (Ana M.F. Barela).

http://dx.doi.org/10.1016/j.ridd.2014.09.002

0891-4222/� 2014 Elsevier Ltd. All rights reserved.

M.L. Celestino et al. / Research in Developmental Disabilities 35 (2014) 3624–3631 3625

able to walk while others are not. Some people show normal to near typical intellectual function, but others may haveintellectual disabilities. Its motor severity is generally classified according to the Gross Motor Function Classification System(GMFCS) (Palisano et al., 1997; Palisano, Rosenbaum, Bartlett, & Livingston, 2008).

Among several movement deficiencies, the gait impairment is one of the major concerns for parents and caregivers.According to Chang, Rhodes, Flynn, and Carollo (2010), the gait impairment is due to spasticity (or abnormal muscle tone),reduced motor control ability and impaired balance. As these children grow with no simultaneous lengthening of skeletalmuscles, muscular tightness and muscle contractures, they eventually develop bony torsion. Consequently, children with CPon ambulatory condition usually present a decline in gait function over time (Bell, Ounpuu, DeLuca, & Romness, 2002;Johnson, Damiano, & Abel, 1997; Norlin & Odenrick, 1986). To overcome the inability to typically develop gait movementpatterns, different categories of treatment have been proposed and investigated, including therapeutic interventionsand neuromuscular invasive techniques.

For instance, the use of body weight support (BWS) is a typical therapeutic intervention technique that provides task-specific gait training. The rationale for using the BWS system is that the reduction of gravitational forces would reduce theload that should be overcome by the individual; facilitating the walking constraints. Consequently, this strategy mightpromote a gait pattern close to typical (Finch, Barbeau, & Arsenault, 1991). In general, the BWS systems consist of a mountingframe and a harness to support a percentage of the individual’s weight as s/he walks on a motorized treadmill (Mattern-Baxter, 2009; Mutlu, Krosschell, & Spira, 2009). The use of a treadmill is usually adopted to stimulates rhythmical andrepetitive steps (Visintin, Barbeau, Korner-Bitensky, & Mayo, 1998). In addition, the treadmill belt triggers interlimbsymmetry causing positive effects on the temporal parameters of walking (Harris-Love, Macko, Whitall, & Forrester, 2004),as well as diminishing the need for propulsive force generation at the end of the stance period (Norman, Pepin, Ladouceur, &Barbeau, 1995). More recently, the BWS systems have been employed during over ground walking in individuals with stroke(Lamontagne & Fung, 2004; Miller, Quinn, & Seddon, 2002; Prado-Medeiros et al., 2011; Sousa, Barela, Prado-Medeiros,Salvini, & Barela, 2009; Sousa, Barela, Prado-Medeiros, Salvini, & Barela, 2011) and children with CP (Matsuno, Camargo,Palma, Alveno, & Barela, 2010). However, to the best of our knowledge, it remains unclear whether the type of surface and/orthe amount of body weight unloaded could influence the way that children with CP walk. This study was designed toinvestigate gait kinematic parameters of children with CP by manipulating BWS and two different types of ground surfaces(treadmill and over ground).

2. Methods

2.1. Participants

Ten children (7.7 � 2.1 years old) diagnosed with spastic CP and GMFCS classification between levels II and IV wereselected to participate in this study (Table 1). The inclusion criteria also account for the individual’s ability to walkapproximately 7 m with or without assistance, and to understand the experimental instructions and procedures. Fourteenchildren were initially recruited and assessed. However, data from four participants were excluded due to experimentaldifficulties encountered either during data acquisition (n = 2) or data processing (n = 2). Prior to participation, each child’sparent or legal guardian provided informed consent. All experimental procedures were approved by the Institutional ReviewBoard at the University of Cruzeiro do Sul, Sao Paulo, Brazil (Protocol 020/2010).

2.2. Procedures

The children were asked to walk at a comfortable speed along a walkway (7 m) and on a treadmill under three differentconditions: walking wearing a harness and bearing full body weight (‘‘0% BWS’’ condition); walking wearing a harness and15% of full body weight unloaded (‘‘15% BWS’’ condition); and walking wearing a harness and 30% of full body weight

Table 1

General information about the children considered in the final sample.

ID Sex Age (years) Mass (kg) Height (cm) Diagnosis GMFCS

1 F 9.3 19.7 123 Diplegia III

2 M 8.3 23.8 132 Diplegia II

3 M 9.5 25.3 130 Diplegia II

4 F 3.2 13.1 93 Diplegia III

5 F 7.8 34.5 138 Hemiplegia III

6 F 7.5 30.0 108 Diplegia IV

7 M 8.9 30.7 122 Diplegia III

8 M 8.6 23.2 136 Diplegia III

9 F 4.9 15.5 110 Diplegia IV

10 F 9 19.1 104 Hemiplegia II

Mean – 7.7 23.5 119.6 – –

SD – 2.1 6.9 15.1 – –

Abbreviation: ID, child’s identification; GMFCS, Gross Motor Function Classification System (Palisano et al., 1997).

[(Fig._1)TD$FIG]

Fig. 1. Representation of the customized body weight support system, illustrating the suspended rail, moving cart and the two electrical servo motors, load

cell, drivers and computer that control the displacement, velocity and acceleration of the moving cart, and the harness. Note: more details are described in

the text.

M.L. Celestino et al. / Research in Developmental Disabilities 35 (2014) 3624–36313626

unloaded (‘‘30% BWS’’ condition). Fig. 1 exemplifies the customized BWS system (Finix Tecnologia). The system consists ofa suspended rail (7 m) mounted on the ceiling (3 m height) and sustained by two steel beams. A moving cart is attached tothe bottom of the rail allowing for backward and forward movements and controlled by a belt system linked to a servo motor.A customized program (LabView 2011, National Instruments Inc.) was developed to control displacement, velocity andacceleration of the moving cart. In addition, the moving cart was controlled by a second servo motor that controlled themechanical body weight supported provided by the harness. An instrumented load cell registered the body weight andallowed for weight unload manipulation. To accurately unload the individual’s body weight according to the desiredexperimental conditions, participants were asked to stay still until the belt’s length was properly adjusted.

A computerized gait analysis system (VICON Bonita 10) with seven infrared cameras was used to measure kinematic data.Based on the Vicon Plug-In Gait model (Vicon, 2010), reflective markers were placed on the sacrum, and bilaterally on theanterior superior iliac crest, the midpoint of the lateral femur, the lateral knee joint axis, the midpoint of the lateral tibia, thelateral malleolus, the calcaneus and the second metatarsal head. This reference point system has been broadly adoptee byprevious studies (e.g., Benedetti, Manca, Ferraresi, Boschi, & Leardini, 2011; Bohm & Doderlein, 2012). A calibration trial wasconducted prior to testing. Each participant stood upright to record the neutral position (baseline) of all joints and segments.

All participants were asked to walk barefoot at a self-determined speed, which was kept the same for all experimentalconditions. Before data acquisition, all children practiced for a few trials until they felt comfortable with the mechanicaldevice and the lab environment. Three trials, minimum, were recorded for each experimental condition. Theexperimental conditions were properly randomized within and across individuals.

2.3. Data processing

One intermediate stride per trial by each child, for a total of three selected trials for each condition, was analyzed. Thetrial selection was determined by the best visualization of the markers as the children walked with no interruption.Through visual inspection, a stride (walking cycle) was defined by two consecutive initial contacts of the same limb withthe ground along the progression line. In addition, walking events during a stride were identified for subsequentcalculation of the temporal organization of walking, such as initial and terminal double stance, single limb support andswing period (Perry, 1992). This procedure was carried out for both right and left sides of the body.

Joint and segmental angles were processed with Vicon Nexus software (version 1.8.5). Subsequent analyses wereperformed using Matlab software (MathWorks, Inc.). For joint and segmental angles, strides were normalized intime from 0% to 100% with a 1% step. These cycles were referred to the children’s neutral angles measured during thecalibration trial in each condition and were then averaged to obtain the mean cycle for each participant. The sameprocedure was repeated to obtain the mean cycle among participants.

The outcome measures analyzed in this study were: stride length (distance between two successive initial contacts of thesame foot to the surface, determined by the position of the calcaneus), speed (calculated as the ratio between stride lengthand duration), durations of total double stance and single limb support, maximum angles of pelvis (posterior tilt), hip(extension), knee (extension) and ankle (plantar flexion) and minimum angles of pelvis (anterior tilt), hip (flexion), knee(flexion) and ankle (dorsiflexion) at the sagittal plane during each stride. Data from right and left sides were averagedtogether before making comparisons among different experimental conditions.

2.4. Statistical analysis

For all variables, data from three trials under each condition were averaged for each child. Multivariate analyses ofvariance (MANOVAs) for repeated measures were employed, using surface (treadmill and over ground) and body weightunloading (0%, 15% and 30%) as factors. The dependent variables were stride length and stride speed for the first MANOVA;

Table 2

Mean (SD) values of the spatial–temporal parameters of walking during the stride cycle according to surface and body weight support (BWS).

Outcome measures/surface BWS p value

0% 15% 30% Surface BWS Surface*BWS

Stride length (m)

Over ground 0.56 (0,17) 0.60 (0.18) 0.61 (0.17) 0.743 0.905 0.028

Treadmill 0.60 (0.23) 0.57 (0.21) 0.56 (0.22)

Stride speed (m/s)

Over ground 0.44 (0.19) 0.45 (0.19) 0.46 (0.18) 0.877 0.585 0.001

Treadmill 0.47 (0.17) 0.45 (0.17) 0.42 (0.18)

Total double stance (%)

Over ground 34 (7.4) 37 (5.5) 31 (7.6) 0.002 0.075 0.480

Treadmill 42 (5.4) 40 (5.4) 39 (6.7)

Single limb support (%)

Over ground 32 (6.7) 27 (4.4) 31 (6.6) 0.005 0.378 0.065

Treadmill 23 (2.5) 25 (2.4) 24 (4.2)

M.L. Celestino et al. / Research in Developmental Disabilities 35 (2014) 3624–3631 3627

total double stance and single limb support for the second MANOVA; maximum angles of pelvis, hip, knee and ankle forthe third MANOVA; and minimum angles of pelvis, hip, knee and ankle for the fourth MANOVA.

When necessary, univariate analyses, comparisons with Bonferroni adjustments and Tukey post hoc tests wereemployed as necessary. Alpha level of 0.05 was adopted. All analyses were performed using the Statistical Packagefor Social Sciences software.

3. Results

Each child walked at the same mean speed in all conditions. The speed ranged from 0.29 to 0.83 (0.43, SD 0.19) m/s.Table 2 depicts mean (SD) values of walking cycle’s spatial–temporal parameters. MANOVA for stride length and speedrevealed interaction between the surface and BWS (Wilks’ Lambda = 0.42, F4,34 = 4.65, p< 0.005). The univariate analysisrevealed interaction between the surface and BWS for stride length (F2,18 = 4.37) and stride speed (F2,18 = 10.14). Tukey posthoc test revealed that only stride speed was lower when children walked with 30% BWS compared to 0% BWS on a treadmill.MANOVA for durations of total double stance and single limb support revealed a surface effect only (Wilks’ Lambda = 0.30,F2,8 = 9.40, p = 0.008). The univariate analysis revealed that children presented shorter double stance (F1,9 = 17.69) and longersingle limb support (F1,9 = 13.99) when they walked over ground compared to on a treadmill (Table 2).

Fig. 2 illustrates the mean (SD) values gait cycles of pelvis, hip, knee and ankle angles in the sagittal plane of all children asthey walked on both surfaces with 0%, 15% and 30% BWS. Qualitatively, the pelvis and all joints seemed to have roughlysimilar patterns in all conditions. Anterior and posterior pelvic tilt, hip and knee flexion and extension, and ankle dorsiflexionand plantar flexion were present in all conditions.

Table 3 depicts mean (SD) values of minimum and maximum angles for pelvis, hip, knee and ankle joints during the gaitcycle in all conditions. For minimum angles, MANOVA revealed an effect only for the surface (Wilks’ Lambda = 0.06,F4,6 = 22.68, p = 001). Univariate analyses revealed that children presented less pelvic posterior tilt (F1,9 = 7.67), hip extension(F1,9 = 7.92), knee extension (F1,9 = 14.52) and ankle plantar flexion (F1,9 = 13.57) when walking over ground compared to on atreadmill (Table 3).

For maximum angles, MANOVA revealed an effect for surface (Wilks’ Lambda = 0.13, F4,6 = 9.73, p = 0.009), and atendency for BWS (Wilks’ Lambda = 0.39, F8,30 = 2.24, p = 0.52). Univariate analyses for surface revealed that childrenpresented less pelvic anterior tilt (F1,9 = 7.92) and hip flexion (F1,9 = 16.18) when walking over ground compared to on atreadmill. Univariate analyses for BWS revealed differences for pelvic anterior tilt (F2,18 = 5.23), hip flexion (F2,18 = 5.47)and ankle dorsiflexion (F2,18 = 9.23). Comparisons with Bonferroni adjustment revealed that children presented higherexcursion for pelvis, hip and ankle with 0% BWS compared to 30% BWS, and higher excursion for ankle with 15% BWScompared to 30% BWS (Table 3).

4. Discussion

The aim of this study was to investigate gait kinematic parameters of children with CP by manipulating BWS and twodifferent types of ground surfaces. Overall, the results revealed that ground surfaces caused greater impact on gait patterns ofchildren with CP as compared to body weight unloading. Interestingly, even the children who could not walk independentlywere able to walk with a BWS system on both surfaces. Specifically, it was found that ground surface affected individual’stotal double stance, single limb support, minimum angles of pelvis, hip, knee and ankle and maximum angles of pelvisand hip, suggesting that walking over the ground promoted a gait pattern more similar to their typically developing peers.The body weight unloading affected the maximum angles of the pelvis, hip and ankle on both surfaces and stride speedon treadmill, suggesting that children with CP had difficulties overcoming the gravitational forces.

[(Fig._2)TD$FIG]

Fig. 2. Mean (SD) stride cycles of pelvis (A, E, I), hip (B, F, J), knee (C, G, K) and ankle (D, H, L) in the sagittal plane for the children with cerebral palsy as they

walked with 0% BWS (left panel), 15% BWS (middle panel) and 30% BWS (right panel) on a treadmill (gray area) and over ground (line).

M.L. Celestino et al. / Research in Developmental Disabilities 35 (2014) 3624–36313628

Table 3

Mean (SD) values of pelvis and joint angles (degrees) during the stride cycle according to surface and body weight support (BWS).

Outcome measures/surface BWS p value

0% 15% 30% Surface BWS Surface*BWS

Pelvic posterior tilt

Over ground 10 (7) 9 (8) 9 (9) 0.022 0.087 0.527

Treadmill 13 (8) 11 (7) 10 (10)

Hip extension

Over ground 2 (10) 2 (10) 3 (11) 0.020 0.105 0.353

Treadmill 7 (12) 5 (10) 6 (11)

Knee extension

Over ground 14 (13) 15 (14) 15 (13) 0.004 0.658 0.478

Treadmill 19 (15) 18 (13) 18 (13)

Ankle plantar flexion

Over ground �21 (25) �21 (18) �22 (22) 0.005 0.953 0.297

Treadmill �13 (17) �17 (18) �21 (22)

Pelvic anterior tilt

Over ground 19 (6) 18 (7) 18 (8) 0.020 0.016 0.248

Treadmill 22 (10) 20 (7) 19 (8)

Hip flexion

Over ground 42 (11) 42 (13) 42 (14) 0.003 0.014 0.199

Treadmill 49 (14) 44 (12) 43 (13)

Knee flexion

Over ground 59 (24) 61 (26) 61 (25) 0.383 0.884 0.498

Treadmill 62 (30) 57 (19) 58 (25)

Ankle dorsiflexion

Over ground 19 (11) 19 (16) 14 (9) 0.182 0.002 0.290

Treadmill 23 (12) 17 (12) 19 (15)

M.L. Celestino et al. / Research in Developmental Disabilities 35 (2014) 3624–3631 3629

Different from previous findings (Matsuno et al., 2010), similar stride length was observed in all experimentalconditions. This might be attributed to the mean walking speed that is known as a critical parameter that affect the gaitpattern (Winter, 1991). On the other hand, the children in this study presented greater instability while walking on atreadmill as compared to walking over the ground. This finding was supported by higher double stance duration andlower single limb support duration, respectively. For instance, it was observed that by increasing the base of support(both feet on the ground), balance and stability also increased. The children with CP took longer strides (keeping both feetin contact with the treadmill) as strategy to ensure greater walking stability. Furthermore, the tested children were morecapable of sustaining their limbs while walking over the ground as compared to walking on a treadmill, independent ofbody weight unloading. The duration of single limb support indicates the capacity for sustaining the limb (Perry, 1992).

Greater pelvic posterior tilt, hip and knee extension and ankle plantar flexion were observed when children walked ona treadmill as compared to walking over the ground. Similarly, greater pelvic anterior tilt and hip flexion while walkingon a treadmill as compared to over the ground. However, these findings should be carefully interpreted since thesedifferences did not indicate different ranges of motion of pelvis and lower limb joints. The children tested in thisexperiment kept their thigh closer to the shank (knee flexion), and the shank closer to their foot (ankle dorsiflexion) whilewalking on a treadmill as compared to over the ground, which suggests a more crouched gait. Interestingly, while walkingover the ground the children were able to keep these segments straighter as compared to walking on a treadmill. With theexcept of knee joint values that revealed greater inter-individual variability, all remaining values observed when thechildren with CP walked with BWS over the ground were pretty similar to their typically developing peers (Ounpuu, Gage,& Davis, 1991). In terms of BWS values, the children presented maximum angles similar to typically developing (Ounpuuet al., 1991) while walking with 30% BWS as compared to walking with 0% BWS.

The use of treadmill for gait intervention is very common. Among several well-known advantages are the restricted spacenecessary and the better control of the number of steps during a gait training session. Controversy, the use of the treadmillimposes a different context altering propulsion and balance patterns (Norman et al., 1995). For example, to walk over theground people must apply sufficient forces toward the floor to propel limbs forward. On a treadmill, the propulsiongenerated by the limbs is not necessarily proportional to velocity (Goldberg, Kautz, & Neptune, 2008), as the belt allows forthe limbs to passively move with minimized levels of muscle activation (Harris-Love et al., 2004). In terms of balance control,it seems that children with CP adopt a different strategy, as they would adopt while walking over the ground. This wasconfirmed on this study as the children extended duration of double stance and single limb support.

The results of this study clearly indicate the use of BWS system over the ground (instead of treadmill) would promotegreater kinematic benefits for children with CP. One additional advantage is that these children would not have to transfer oradapt their gait parameters from the treadmill to over ground, if the intervention program directly focuses on learning how

M.L. Celestino et al. / Research in Developmental Disabilities 35 (2014) 3624–36313630

to walk over the ground. This is particularly important because children with CP struggle with such type adaptation (Barelaet al., 2011). Importantly, the amount of BWS has to be set accordingly, so these children can sustain their body weightduring the single limb support of each lower limb without excessively flexing their hip and knee.

To present, only a few studies have assessed individuals with gait impairment as they walked with BWS (Lamontagne &Fung, 2004; Matsuno et al., 2010; Sousa et al., 2009). To the best of our knowledge only Matsuno et al. (2010) had used similarexperimental conditions; walking with BWS on a treadmill and over the ground. On this study we further Matsuno’s resultsas our innovative findings showed the consequences of manipulating practical conditions that therapists could eventuallyuse during training sessions with BWS. Additionally, this is the first study that assessed children with CP walking with acustomized BWS system on a treadmill and over ground with different body weight unloading.

Walking is a rather complex motor task and many other parameters beyond the spatial–temporal analysis and lowerbody joint angles must be considered. In addition, we recognize that a large number of participants would strengthen ourfindings. However, the challenges imposed by the inclusion criteria, as well as the difficulties faced on recordingkinematic data on this population limited our sample size. Nevertheless, our results provide new insights to therapistsand intervention programs targeting walking rehabilitation for this population.

5. Conclusion

We concluded that changing the ground surface in which children with CP walk positively impact their gaitperformance; more than body weight unloading. Therefore, therapists adopting BWS strategies for gait training ofchildren with CP, should be aware that although body unloading might facilitate gait execution the ground surface mightpromote greater impact on gait rehabilitation of children with CP.

Conflict of interest

The authors declare no competing interests.

Acknowledgments

This study was supported by the Sao Paulo Research Foundation – FAPESP (Grant #2010/15218-3) – and by a fellowshipby Coordenacao de Aperfeicoamento de Pessoal de Nıvel Superior – CAPES – for M.L. Celestino and by FAPEPS for G.L. Gama(#2013/01050-1). The study was conducted in the Laboratory of Movement Analysis, Institute of Physical Activity and SportSciences, Cruzeiro do Sul University. We would like to thank the children and their parents for participating in our study andP.B. Freitas for programming the routine in LabView.

References

Barela, J. A., Focks, G. M., Hilgeholt, T., Barela, A. M., Carvalho Rde, P., & Savelsbergh, G. J. (2011). Perception-action and adaptation in postural control of childrenand adolescents with cerebral palsy. Research in Developmental Disabilities, 32, 2075–2083.

Bell, K. J., Ounpuu, S., DeLuca, P. A., & Romness, M. J. (2002). Natural progression of gait in children with cerebral palsy. Journal of Pediatric Orthopaedics, 22,677–682.

Benedetti, M. G., Manca, M., Ferraresi, G., Boschi, M., & Leardini, A. (2011). A new protocol for 3D assessment of foot during gait: Application on patients withequinovarus foot. Clinical Biomechanics, 26, 1033–1038.

Bohm, H., & Doderlein, L. (2012). Gait asymmetries in children with cerebral palsy: Do they deteriorate with running? Gait and Posture, 35, 322–327.Chang, F. M., Rhodes, J. T., Flynn, K. M., & Carollo, J. J. (2010). The role of gait analysis in treating gait abnormalities in cerebral palsy. The Orthopedic Clinics of North

America, 41, 489–506.Finch, L., Barbeau, H., & Arsenault, B. (1991). Influence of body weight support on normal human gait: Development of a gait retraining strategy. Physical Therapy,

71, 842–856.Goldberg, E. J., Kautz, S. A., & Neptune, R. R. (2008). Can treadmill walking be used to assess propulsion generation? Journal of Biomechanics, 41, 1805–1808.Harris-Love, M. L., Macko, R. F., Whitall, J., & Forrester, L. W. (2004). Improved hemiparetic muscle activation in treadmill versus overground walking.

Neurorehabilitation and Neural Repair, 18, 154–160.Inman, V. T., Ralston, H. J., & Todd, F. (1994). Human locomotion. In J. Rose & J. G. Gamble (Eds.), Human walking (2nd ed., pp. 1–22). Baltimore: Williams & Wilkins.Johnson, D. C., Damiano, D. L., & Abel, M. F. (1997). The evolution of gait in childhood and adolescent cerebral palsy. Journal of Pediatric Orthopedics, 17, 392–396.Lamontagne, A., & Fung, J. (2004). Faster is better: Implications for speed-intensive gait training after stroke. Stroke, 35, 2543–2548.Matsuno, V. M., Camargo, M. R., Palma, G. C., Alveno, D., & Barela, A. M. (2010). Analysis of partial body weight support during treadmill and overground walking of

children with cerebral palsy. Revista Brasileira de Fisioterapia, 14, 404–410.Mattern-Baxter, K. (2009). Effects of partial body weight supported treadmill training on children with cerebral palsy. Pediatric Physical Therapy, 21, 12–22.Miller, E. W., Quinn, M. E., & Seddon, P. G. (2002). Body weight support treadmill and overground ambulation training for two patients with chronic disability

secondary to stroke. Physical Therapy, 82, 53–61.Mutlu, A., Krosschell, K., & Spira, D. G. (2009). Treadmill training with partial body-weight support in children with cerebral palsy: A systematic review.

Developmental Medicine and Child Neurology, 51, 268–275.Newell, K. M. (1986). Constraints on the development of coordination. In M. G. Wade & H. T. A. Whiting (Eds.), Motor development in children: Aspects of coordination

and control (pp. 341–360). Boston, MA: Martin Nighoff.Norlin, R., & Odenrick, P. (1986). Development of gait in spastic children with cerebral palsy. Journal of Pediatric Orthopedics, 6, 674–680.Norman, K. E., Pepin, A., Ladouceur, M., & Barbeau, H. (1995). A treadmill apparatus and harness support for evaluation and rehabilitation of gait. Archives of

Physical Medicine and Rehabilitation, 76, 772–778.Ounpuu, S., Gage, J. R., & Davis, R. B. (1991). Three-dimensional lower extremity joint kinetics in normal pediatric gait. Journal of Pediatric Orthopedics, 11, 341–349.Palisano, R., Rosenbaum, P., Walter, S., Russell, D., Wood, E., & Galuppi, B. (1997). Gross motor function classification system for cerebral palsy. Developmental

Medicine and Child Neurology, 39, 214–223.

M.L. Celestino et al. / Research in Developmental Disabilities 35 (2014) 3624–3631 3631

Palisano, R. J., Rosenbaum, P., Bartlett, D., & Livingston, M. H. (2008). Content validity of the expanded and revised gross motor function classification system.Developmental Medicine and Child Neurology, 50, 744–750.

Perry, J. (1992). Gait analysis. Throfare: Slack.Prado-Medeiros, C. L., Sousa, C. O., Souza, A. S., Soares, M. R., Barela, A. M., & Salvini, T. F. (2011). Effects of the addition of functional electrical stimulation to ground

level gait training with body weight support after chronic stroke. Revista Brasileira de Fisioterapia, 15, 436–444.Rosenbaum, P., Paneth, N., Leviton, A., Goldstein, M., & Bax, M. (2007). A report: The definition and classification of cerebral palsy April 2006. Developmental

Medicine and Child Neurology, 49, 8–14.Sousa, C. O., Barela, J. A., Prado-Medeiros, C. L., Salvini, T. F., & Barela, A. M. (2011). Gait training with partial body weight support during overground walking for

individuals with chronic stroke: A pilot study. Journal of NeuroEngineering and Rehabilitation, 8, 48.Sousa, C. O., Barela, J. A., Prado-Medeiros, C. L., Salvini, T. F., & Barela, A. M. F. (2009). The use of body weight support on ground level: An alternative strategy for gait

training of individuals with stroke. Journal of NeuroEngineering and Rehabilitation, 6, 43.Vicon (2010). Vicon plug-in-gait product guide – Foundation notes revision 2.0 March 2010. In Vicon Motion System. Vicon Motion System Limited.Visintin, M., Barbeau, H., Korner-Bitensky, N., & Mayo, N. E. (1998). A new approach to retrain gait in stroke patients through body weight support and treadmill

stimulation. Stroke, 29, 1122–1128.Winter, D. A. (1991). The biomechanics and motor control of human gait: Normal, elderly, and pathological (2nd ed.). Waterloo: University of Waterloo Press.