Electromyographic patterns in children with cerebral palsy: Do they change after surgery?

Gait & Posture 26 (2007) 362–371 www.elsevier.com/locate/gaitpost

Electromyographic patterns in children with cerebral palsy: Do they change after surgery? D. Patikas *, S.I. Wolf, W. Schuster, P. Armbrust, T. Dreher, L. Do¨derlein Department of Orthopaedic Surgery, University of Heidelberg, Germany Received 13 April 2006; received in revised form 4 October 2006; accepted 5 October 2006

Abstract The purpose of this study was to investigate the changes in electromyographic (EMG) patterns after multilevel surgical treatment in children with spastic cerebral palsy. Children with diplegia (n = 18) and hemiplegia (n = 16) aging from 6 to 16 years participated in the study. Twenty healthy children within the same age span are presented as reference. Gait analysis and surface electromyograms of seven major lower limb muscles were assessed before and 1–5 years after the multilevel surgery. The most frequent procedures were equinus correction, distal rectus femoris transfer, femoral derotation osteotomy and hamstrings lengthening. The results showed that the EMG pattern of the soleus, lateral gastrocnemius and tibialis anterior muscles became closer to normal after the surgery, while no differences were detected between diplegic and hemiplegic patients. Furthermore, a subgroup of 10 patients showed an increase in medial hamstrings activation during preswing that decreased postoperatively. These findings indicate that changes in EMG patterns should not be ruled out after surgical treatment, although the extent of these changes is limited compared to changes in the kinematics. Abnormal muscle activation before the operation can be related to a compensatory response in some patients and this can be manipulated after surgery. # 2006 Elsevier B.V. All rights reserved. Keywords: Electromyography; Cerebral palsy; Gait; Surgery; Child

1. Introduction Cerebral palsy (CP) is an irreversible and nonprogressive disorder caused by brain injury before, during or shortly after birth. Impaired motor control affecting movement and posture are common consequences of CP. In spastic CP with no involvement of the extrapyramidal system, several secondary abnormalities develop during growth, such as bony deformities and changes in the joint and muscle properties [1,2]. As a result, daily activities become challenging for children with CP and therefore treatment is often indicated. The type of treatment, which CP patients receive, depends on their individual requirements. The aim of multilevel orthopaedic surgery is to correct bony and soft * Corresponding author at: Orthopaedic University Clinic of Heidelberg, Schlierbacher Landstr. 200a, 69118 Heidelberg, Germany. Tel.: +49 6221 966720; fax: +49 6221 966725. E-mail address: [email protected] (D. Patikas). 0966-6362/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.gaitpost.2006.10.012

tissue deformities, all in one session [3]. Although this intervention generally improves the patient’s effectiveness in everyday tasks such as walking [3,4], the causes of the disorder are still present. Electromyography (EMG) has been shown to be a reliable method of evaluating patients with CP during gait [5]. These patients show abnormal EMG patterns and timing during gait [6]. Earlier reports showed that muscular activation patterns do not change after a surgical intervention [6–10]. These studies also suggested that no alteration in the muscle activation pattern should be expected since the surgical intervention is targeted to the periphery (bones, muscles and tendons). Some authors argue that the EMG is a ‘fingerprint’ in patients with CP preventing them from adapting to the new biomechanical status after the surgery [7]. Taking this into consideration, EMG measurement after the operation would be redundant. However, these studies either assessed a small number of subjects or their conclusions were drawn from visual observation of the raw EMG signals, which is subjective.

D. Patikas et al. / Gait & Posture 26 (2007) 362–371

More recent studies showed that the activity of the tibialis anterior muscle reduced when patients with an excessive plantarflexion at initial contact used ankle–foot orthoses [11]. Even short-term intensive balance training improved muscle coordination, but the level of adaptation depended on the severity of the child’s disability [12]. These studies also suggested that muscle recruitment in patients with CP may change after an intervention. It has been shown that the activation timing of the semitendinosus and vastus lateralis muscles changes after hamstring lengthening procedures [13]. These changes were attributed not only to biomechanical changes but also to changes in the recruitment of muscles that were not surgically treated. This indicates that other multilevel orthopaedic surgical procedures may also have an effect on muscular activation during gait. The existence of such changes might be of interest to orthopaedic surgeons and may help to more precisely define treatment planning, the extent of surgery and postoperative rehabilitation treatment. Therefore, the purpose of this study was to evaluate the EMG patterns during walking before and after multilevel surgical treatment in a group of children with diplegic and hemiplegic CP. Moreover, we tested for differences between children with diplegia and hemiplegia.

363

2.2. Gait analysis All subjects walked barefoot at a self-selected speed along a lane 15 m in length during data acquisition. Kinematics and kinetics were recorded using a ninecamera 3D motion analysis system (VICON 612, Oxford Metrics Limited, UK) operating at 120 Hz and two forceplates (Kistler Instruments Co.), respectively. The Euler angles for the hip, knee and ankle were calculated using 15 passive infrared reflecting markers placed in standardized positions according to the Helen Hayes marker set [14]. The Gillette Gait Index [15] was calculated for both sides in these patients. According to this index each side was assigned as being more or less involved. To preserve a more homogeneous group with regards to severity, we evaluated only the more involved side in each patient. Kinematic and kinetic parameters were analyzed in the sagittal, coronal, and transverse planes for the pelvis, hip, knee and ankle to document the pre- and postoperative status of the patients. For reasons of simplicity, the presentation of the kinematic and kinetic results was focused on the sagittal plane only. 2.3. EMG measurement

2. Methods 2.1. Subjects Thirty-four patients with CP (18 with diplegia and 16 with hemiplegia) aging from 6 to 16 years (mean  standard deviation [S.D.] 10.1  3.0 years) participated in the present study (Table 1). They were able to walk unsupported and had no previous surgery, casting, or medical treatment (e.g., botulinum toxin or baclofen). The study was approved by the local Ethical Committee. Furthermore, data from 20 healthy children of the same age range were taken as a reference. 3D-gait analysis, including kinematic, kinetic and EMG data recordings, was carried out in all patients 1–3 days before and 1–5 years (2.53  1.23) after multilevel surgery. As shown in Table 2, the most frequent surgical procedures were equinus correction, distal rectus femoris transfer, femoral derotation osteotomy and hamstrings lengthening.

We collected data from seven superficial muscles responsible mainly for hip, knee and ankle extension/ flexion: the vastus lateralis, the rectus femoris, the biceps femoris, the medial hamstrings, the tibialis anterior, the lateral gastrocnemius and the soleus muscle. Bipolar surface adhesive electrodes (Blue Sensor, Ambu Inc., Glen Burnie, MD, USA) with an inter-electrode distance of 2 cm were applied over each of the examined muscles according to the SENIAM guidelines [16]. The EMG signal was preamplified (5000) using the Biovision EMG apparatus (Biovision Inc., Wehrheim, Germany). 2.4. Signal processing and data analysis The analog forceplate and EMG data were digitized using a 12-bit A/D card, with a sample frequency of 1080 Hz. The EMG signal was fully rectified off-line and

Table 1 Demographic and descriptive data of the participants during the pre- and postoperative examination

Age (years) Height (cm) Body mass (kg) Sex (boys/girls) Interval between examinations (years)

Children with diplegia (n = 18)

Children with hemiplegia (n = 16)

Pre-operative

Post-operative

Pre-operative

Post-operative

10.5  2.76 (6.0–15.8) 139  11.4 (118–163) 33.3  9.4 (19.8–51.0) 12/6 2.31  0.98 (0.9–4.3)

12.8  2.9 (7.9–18.1) 152  11.9 (131–169) 44.1  12.3 (25.9–67.1)

9.6  3.2 (6.0–16.2) 138  20.5 (115–192) 32.2  12.6 (17.6–60.4) 10/6 2.77  1.45 (1.0–4.9)

12.4  2.5 (7.7–17.4) 153  18.9 (120–192) 42.1  13.9 (20.6–68.8)

Healthy children (n = 20)

11.7  2.4 (8.1–15.7) 151  14.4 (125–175) 44.6  12.3 (25.6–64.7) 8/12 –

364

D. Patikas et al. / Gait & Posture 26 (2007) 362–371

Table 2 List of surgical procedures performed for each patient Patient ID

Sex

Age

Diagnosis

0434 0514* 0566 0670* 0747 0858* 0886 1153 1170 1415 1425* 1440 1442 1468 1486* 1492* 1493* 1504* 1155 1529 1585 1642 0170 0638 0878 1120 1128 1292* 1430 1625 0573 0911 1084 1446*

Girl Girl Girl Boy Boy Boy Boy Boy Boy Boy Boy Boy Boy Girl Girl Boy Boy Girl Boy Girl Boy Boy Girl Girl Boy Boy Boy Boy Boy Girl Boy Boy Girl Boy

8.9 10.3 13.3 11.3 9.3 11.8 8.8 7.1 15.8 9.9 6.0 15.7 10.5 10.9 11.1 6.1 12.5 9.4 8.5 9.8 11.9 16.0 7.2 8.0 6.1 6.5 8.6 7.7 6.7 16.2 10.8 11.6 6.0 11.8

Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Diplegia Type I Type I Type I Type I Type II Type II Type II Type II Type II Type II Type II Type II Type IV Type IV Type IV Type IV

POTB

B

FDO

DRFT

B B B

B B B B B B B B B B B B B R B B B B

B B R R B B B B B B B B

HAM

TDO

B R B B B B B

TAL

R L

B R

ST-FDC

B B

B-FDC

B B L B

L B

B

R B

B B R R

B L B R

B B

R

B B B L

B L

B B

R R L R R

R L R R

L R

R

L R R

L R R

R

R

R

L R R

R

B B R L R L

Strayer

B R

B B

L

R B

Baumann

R

R

R R

R

R B

L

L

B

R

R R R L R

R R

The diagnosis column for the hemiplegic patients contains the type of classification according to the kinematics in the sagittal plane [30]. Note: POTB, psoas over the brim; FDO, femoral derotation osteotomy; DRFT, distal rectus femoris transfer; HAM, hamstrings lengthening; TDO, tibia derotation osteotomy; TAL, tendo-Achilles lengthening; ST–FDC, soft-tissue foot deformity correction; B-FDC, bony foot deformity correction; B, both sides; R, right side; L, left side. Asterisks (*) indicate the subgroup of patients that had increased medial hamstrings activity preoperatively.

the linear envelope was calculated at a cut-off frequency of 9 Hz [17,18]. The EMG amplitude was normalized to the average value of each stride for the respective muscle and subject. Hence, the EMG signals were expressed as percentage of the mean. Each gait cycle was divided into seven subphases: loading response, mid stance, terminal stance, preswing, initial swing, mid swing and terminal swing. These subphases were defined by taking into consideration the foot strike and foot off of both feet [19]. Each subphase was interpolated to 30 datapoints and the average of 3–8 gait cycles for each dependent variable and condition (subject, examination and more involved side) was calculated. To determine a magnitude that describes the similarity of the EMG pattern of a patient to the mean EMG pattern of the norm group, the norm-distance (ND) was calculated, defined as the absolute difference between the EMG of the patient (p) and the mean value of the norm group (N) at a specific time (t), normalized to

the standard deviation (s) of the norm group at the same time t [20]. NDpt ¼

jxpt  xNt j sNt

To quantify the deviation from normal for each parameter examined, we calculated the mean norm-distance for the whole gait cycle. Two-way ANOVA was used to determine the effect of the factor DIAGNOSIS (levels: diplegia and hemiplegia), the factor TIME for repeated measurements (levels: examination before and after the surgery) and the interaction between these factors for the norm-distance of the spatiotemporal, kinematic, kinetic and EMG parameters. If a significant effect was shown in the ANOVA, a second two-way ANOVA (factors DIAGNOSIS and TIME) was applied to the mean of each of the seven subphases. The level of significance was set at p < 0.05 and the 95% confidence intervals (95%CI)

D. Patikas et al. / Gait & Posture 26 (2007) 362–371

were calculated. The statistical analysis was performed using SPSS Version 13 software (The Apache Software Foundation, IL, USA).

3. Results As shown in Table 3, patients with hemiplegia walked significantly faster (95%CI 0.05–0.24 m s1), had a shorter stance phase (95%CI 1.9–5.3% of gait cycle), and longer stride length (95%CI 0.08–0.26 m) than the patients with diplegia. Concerning the changes between pre- and postoperative data, both groups showed a statistically significant decrease in stance phase duration (95%CI 0.2–2.8% of gait cycle), increase in stride length (95%CI 0.03–0.14 m) and decrease in GGI (95%CI 137–338). The kinematic and kinetic parameters in the sagittal plane showed an overall improvement after the surgery. Postoperatively, the mean norm-distance of all patients was significantly lower ( p < 0.05) for all examined parameters, except for the hip flexor/extensor power, indicating that the kinematic and kinetic curves came closer to the average of the control group of healthy children. Furthermore, a significant difference in the norm-distance between diplegic and hemiplegic patients was observed for the hip and knee kinematics and kinetics (except the hip flexion/extension angle) in the sagittal plane, whereby values were always higher in the diplegic patients (Fig. 1). A more detailed assessment of the average kinematic and kinetic data in the sagittal plane of all patients (Fig. 2) showed that, after the surgery, the ankle was more dorsiflexed (mean plantar flexion reduced from 9.9  13.9 to 0.6  4.38), the peak dorsiflexion moment during opposite foot off was reduced from 0.89  0.35 to 0.41  0.30 Nm/kg, and the plantar flexor power generation increased from 1.32  0.88 to 1.73  0.66 W/kg.

365

Concerning the EMG data, the norm-distance decreased significantly for the soleus (95%CI 0.134–0.367), lateral gastrocnemius (95%CI 0.032–0.294) and tibialis anterior muscles (95% 0.014–0.275) (Fig. 3). For the factor DIAGNOSIS, the norm-distance of the biceps femoris showed marginally higher values for the diplegic than for the hemiplegic group ( p = 0.048). The 95%CI of this difference was 0.001–0.282. All other muscles failed to show significant differences for the factor DIAGNOSIS, TIME and DIAGNOSIS  TIME interaction ( p > 0.05). As shown in Fig. 4, the activation level of the soleus and lateral gastrocnemius muscles decreased during the terminal swing (soleus 95%CI 8.9–47.4% and lateral gastrocnemius 95%CI 5.6–30.9%) and increased during the terminal stance phase (soleus 95%CI 2.0–20.4% and lateral gastrocnemius 95%CI 1.1–19.3%). The lateral gastrocnemius muscle activity also increased during the preswing phase (95%CI 0.1–11.3%) and a decrease was seen during the loading response (95%CI 2.2–23.6%). Furthermore, the tibialis anterior muscle showed a significant decrease during the terminal stance phase (95%CI 2.4–13.7%). These changes were not group specific (no significant DIAGNOSIS  TIME interaction) and corresponded to an EMG pattern closer to that of the group of healthy children (Fig. 4). As mentioned above, no systematic differences in the EMG pattern were observed in the thigh muscles after the operation (Figs. 3 and 4). However, this was not the case for all examined muscles and patients. After a subjective visual inspection of the signals for these muscles, 15–25% of the EMG patterns showed changes after surgical treatment. One example is shown in Fig. 5: the abnormal activity of the medial hamstrings before and after the foot off was reduced 1 year after surgery in this patient. The increased internal hip rotation and the abnormal ankle kinematics and kinetics before the surgery represented indications for femoral

Table 3 Spatiotemporal parameters and Gillette Gait Index for the diplegic and hemiplegic patients before (pre) and after (post) the operation 1 y

Diplegia (n = 18)

Hemiplegia (n = 16)

Norm (n = 20)

Speed (m s )

Pre Post

0.95  0.22 1.04  0.19

1.11  0.15 1.16  0.12

1.30  0.18

Cadence (strides min1)

Pre Post

123  12 124  14

124  14 120  15

126  13

Stance phase (% gait cycle)*,y

Pre Post

64.2  4.0 62.4  3.1

60.3  2.5 59.1  2.2

60.2  1.5

Stride time (s)

Pre Post

0.98  0.12 0.98  0.11

0.97  0.11 1.01  0.12

0.96  0.10

Stride length (m)*,y

Pre Post

0.92  0.16 1.00  0.11

1.08  0.18 1.17  0.12

1.25  0.16

Gillette Gait Index*

Pre Post

399  331 120  97

317  260 117  54

17  6

The last column shows the respective values of the age-matched group of healthy children. * Significant difference ( p < 0.05) between the examinations (pre- vs. postoperative). y Significant difference ( p < 0.05) between the patients’ diagnosis (diplegia vs. hemiplegia).

366

D. Patikas et al. / Gait & Posture 26 (2007) 362–371

Fig. 1. Pre- and postoperative norm-distance for the kinematic and kinetic parameters in the sagittal plane of the group with diplegia (D), hemiplegia (H) and for all patients together (D + H). Symbols designate significant main effect ( p < 0.05) between the patient groups (factor DIAGNOSIS) and the examinations (factor TIME).

Fig. 2. Average pre- and postoperative joint angle, moment, and power for the hip, knee and ankle in the sagittal plane. Dashed lines show the standard deviation from the mean. Asterisks indicate significant difference between the pre- and postoperative values for the mean of the gait cycle phase.

D. Patikas et al. / Gait & Posture 26 (2007) 362–371

367

Fig. 3. Pre- and postoperative norm-distance for the examined muscles of the group with diplegia (D), hemiplegia (H) and for all patients together (D + H). Symbols designate significant main effect ( p < 0.05) between the patient groups (factor DIAGNOSIS) and the examinations (factor TIME).

derotation, combined with equinus and soft-tissue foot deformity corrections. A similar decrement in the medial hamstrings activity was observed in 10 of the 34 patients (Table 2). Data from these 10 patients were grouped and plotted against data from the rest of the patients: the patients with increased medial hamstrings activity during the preswing preoperatively had the tendency to walk with a more internally rotated hip position, with their knees more flexed during stance, and in a more dorsiflexed position during stance (Fig. 6).

4. Discussion This study confirms the presence of systematic changes in the EMG patterns after multilevel surgical treatment in ambulatory children with CP. However, taking into consideration the changes in norm-distance for the kinematic and kinetic parameters that occur after surgery, the observed changes in norm-distance for the EMG were relatively small and restricted to the shank muscles (lateral gastrocnemius, soleus and tibialis anterior). The majority of the examined children (30 out of 34) had equinus foot deformity in the more involved side, which was corrected by lengthening procedures for the triceps surae. Preoperatively, enhanced EMG activity on the lateral gastrocnemius and soleus muscles at initial contact may

result in a stretch reflex response. Postoperatively this response was less marked, resulting in a more normal EMG pattern. The cause for this decrease may be the reduced equinus, which allowed a more dorsiflexed position during the initial contact. Furthermore, the increased activation level of the soleus and lateral gastrocnemius during the terminal stance, combined with the increment in dorsiflexion, can be beneficial for an increased power generation of the plantar flexors, which is important for the forward progression of the body [21]. The mean EMG patterns for the thigh muscles examined were retained after the surgery. Observing each patient separately and visually comparing the preoperative with the postoperative EMG patterns, we often observed the ‘fingerprint’ character of the EMG that has been previously reported [7]: the pre- and postoperative EMG patterns were very similar. However, this was not the case for all patients and muscles. After visual inspection 15–25% of the examined EMG patterns showed changes after the surgical treatment. This suggests that certain muscles may have the potential to modify their activity, as previously reported [13]. Although these changes were not systematic for all patients, such data indicate that preoperative recruitment or inhibition of some muscles, which results in a deviation from normal, might be compensatory in origin. Variations in the firing patterns in healthy subjects allow the nervous system to develop new ways of organizing [22], and this cannot be

368

D. Patikas et al. / Gait & Posture 26 (2007) 362–371

Fig. 4. Average pre- and postoperative EMG patterns of all patients. Dashed lines show the standard deviation from the mean. Asterisks indicate significant difference between the pre- and postoperative values for the mean of the gait cycle phase.

Fig. 5. Pre- and postoperative kinematic and medial hamstrings EMG data for one patient (#1446). Fully rectified (illuminated) and linear envelope (thick lines) of the EMG signals is shown. The vertical axis on the right of the EMG signals corresponds to the linear envelope. Upward filled arrows and downward empty arrows represent the initial contact and the toe off, respectively.

D. Patikas et al. / Gait & Posture 26 (2007) 362–371

369

Fig. 6. Average pre- and postoperative data of the biceps femoris, medial hamstrings EMG and the hip rotation, knee flexion/extension and ankle dorsiflexion angle for (A) patients who showed a normal medial hamstrings activity during the preswing (n = 24) and (B) patients who showed an increased medial hamstrings EMG activity during the same gait phase (n = 10).

ruled out a priori for children with diplegic and hemiplegic CP. As stated in a former study [23], we observed a small subgroup of patients with an increased activation of the medial hamstrings during the preswing phase, which was decreased after the operation. Nevertheless, the size of this subgroup (n = 10, see Table 2) was not large enough to account for an effect for the whole group of patients (n = 34); such a trend can be seen in Fig. 4, however. Interestingly, patients of this subgroup walked on average with a more internally rotated hip, a more flexed knee, and in a more dorsiflexed position before the operation. When these attributes improved after the surgical intervention,

the abnormal increase in medial hamstrings activity during preswing was reduced. Although the contribution of the medial hamstrings as internal hip rotators has been characterized as ‘negligible’ according to musculoskeletal models [24], an abnormal activation could be disadvantageous to the already internally rotated hip [25] and hence this could be an indication for treatment. However, since the activation of this muscle is reduced after the surgery, it cannot be ruled out that this ‘abnormal’ activation is driven voluntarily and has a functional purpose, such as assistance for the initiation of knee flexion. Moreover, it could be argued that the activation of the medial hamstrings becomes unnecessary after the surgery, since the lateral

370

D. Patikas et al. / Gait & Posture 26 (2007) 362–371

gastrocnemius, which becomes more active after the surgery (Fig. 4), can contribute more to the initiation of the knee flexion during preswing. Furthermore, it is worth noting that this phenomenon was only observed for the medial hamstrings but not for the biceps femoris, a finding that supports the functional differentiation of these muscles in CP. In agreement with previous studies [26,27], the patients with diplegia demonstrated more abnormal gait patterns than the patients with hemiplegia, as reflected by kinematic and kinetic sagittal parameters of the hip and knee. This fact is also reflected in the surgical procedures that were carried out: femoral derotation osteotomy, distal rectus femoris transfer, and hamstrings lengthening in children with diplegia have been assessed more often. However, differences between the diplegic and hemiplegic patients in terms of EMG data were not observed, except for a marginally higher norm-distance for the biceps femoris of the diplegic patients. This difference could be attributed to the fact that children with hemiplegia walked faster than the ones with diplegia (mean difference 0.15 m s1, see also Table 3), taking into account that EMG is sensitive to changes in speed [28]. However, such small differences in speed may not be enough to modify the EMG [28] and therefore this could explain the absence of more differences in the EMG findings between patients with diplegia and hemiplegia. On the other hand, it has not been proven yet whether children with CP are able to adapt their muscular activity to changes in speed in the same manner as healthy children do. Another factor that may have influenced the results is the aging of the patients. Although the gait pattern of a healthy child at the age of 8 years or older is very similar to that of an adult [29], this has not been shown in longitudinal studies in children with CP. It is yet not known how well the neuromuscular system of children with CP can adapt to changes in body mass and/or height. For example, body weight gain in the presence of anatomical and structural abnormalities may require the recruitment or inhibition of muscles that normally do not change their activation pattern during maturation. This issue that has not been clarified yet for children with CP, and may have biased the results. In conclusion, the findings of the present study support the notion that EMG measurement before and after surgery can describe the clinical condition of the patient and can assist in planning the surgery and the postoperative treatment. When analyzing the preoperative EMG data to make a clinical decision, it should not be taken for granted that the activation patterns will remain the same no matter what biomechanical changes occur after orthopaedic surgery. Further controlled studies with more homogeneous samples and surgical treatment may elucidate the ongoing compensatory mechanisms that are in operation before surgery. Understanding such strategies will help improve our prediction of the surgical intervention.

References ˜ unpuu S, DeLuca PA, Romness MJ. Natural progression [1] Bell KJ, O of gait in children with cerebral palsy. J Pediatr Orthop 2002;22:677– 82. [2] Gage JR, Novacheck TF. An update on the treatment of gait problems in cerebral palsy. J Pediatr Orthop B 2001;10(4):265–74. [3] Nene AV, Evans GA, Petrick JH. Simultaneous multiple operations for spastic diplegia: outcome and functional assessment of walking in 18 patients. J Bone Joint Surg (Br) 1993;75-B(3):488–94. [4] Gage JR, Perry JP, Hicks RR, Koop S, Werntz JR. Rectus femoris transfer to improve knee function of children with cerebral palsy. Dev Med Child Neurol 1987;29:159–66. [5] Granata KP, Padua DA, Abel MF. Repeatability of surface EMG during gait in children. Gait Posture 2005;22:346–50. [6] Perry JP, Hoffer MM. Preoperative and postoperative dynamic electromyography as an aid in planning tendon transfers in children with cerebral palsy. J Bone Joint Surg (Am) 1977;59A:531–7. [7] Gueth V, Abbink F, Reuken R. Comparison of pre and post-operative electromyographs in children with cerebral palsy. Electromyogr Clin Neurophysiol 1985;25:233–43. [8] Brunt D, Scarborough N. Ankle muscle activity during gait in children with cerebral palsy and equinovarus deformity. Arch Phys Med Rehabil 1988;69:115–7. [9] DeLuca PA. The use of gait analysis and dynamic EMG in the assessment of the child with cerebral palsy. Hum Mov Sci 1991;10:543–54. [10] Lee EH, Goh JCH, Bose K. Value of gait analysis in assessment of surgery in cerebral palsy. Arch Phys Med Rehabil 1992;73:642–6. [11] Romkes J, Hell AK, Brunner R. Changes in muscle activity in children with hemiplegic cerebral palsy while walking with and without ankle– foot orthoses. Gait Posture 2006;24:467–74. [12] Woollacott M, Shumway-Cook A, Hutchinson S, Ciol M, Price R, Katrin D. Effect of balance training on muscle activity used in recovery of stability in children with cerebral palsy: a pilot study. Dev Med Child Neurol 2005;47(7):455–61. [13] Buurke JH, Hermens HJ, Roetenberg D, Harlaar J, Rosenbaum D, Kleissen RFM. Influence of hamstrings lengthening on muscle activation timing. Gait Posture 2004;20:48–53. [14] Kadaba MP, Ramakrishnan HK, Wootten ME. Measurement of lower extremity kinematics during level walking. J Orthop Res 1990;8(3): 383–92. [15] Schutte LM, Narayanan U, Stout JL, Selber P, Gage JR, Schwartz MH. An index for quantifying deviation from normal gait. Gait Posture 2000;11:25–31. [16] Hermens HJ, Freriks B, Merletti R, Stegeman DF, Blok JH, Raw G, et al. European recommendations for surface electromyography. Enshede, The Netherlands: Roessingh Research and Development; 1999. [17] Gabel RH, Brank RA. The effects of signal conditioning on the statistical analyses of gait EMG. Electroencephalogr Clin Neurophysiol 1994;93:188–201. [18] Shiavi R, Frigo C, Pedotti A. Electromyographic signals during gait: criteria for envelope filtering and number of strides. Med Biol Eng Comput 1998;35:171–8. [19] Perry JP. Gait analysis: normal and pathological function. Thorofare, NJ: Slack; 1992. [20] Wolf SI, Loose T, Schablowski M, Do¨derlein L, Rupp R, Gerner HJ, et al. Automated feature assessment in instrumented gait analysis. Gait Posture 2006;23:331–8. [21] Winter DA. Energy generation and absorption at the ankle and knee during fast, natural and slow cadences. Clin Orthop 1983;175:147–54. [22] Sporns O, Edelman GM. Solving Bernstein’s problem: a proposal for the development of coordinated movement by selection. Child Dev 1993;64(4):960–81. [23] Patikas D, Wolf SI, Do¨derlein L. Electromyographic evaluation of the sound and involved side during gait of spastic hemiplegic children with cerebral palsy. Eur J Neurol 2005;12:691–9.

D. Patikas et al. / Gait & Posture 26 (2007) 362–371 [24] Arnold AS, Asakawa DS, Delp SL. Do the hamstrings and adductors contribute to excessive internal rotation of the hip in persons with cerebral palsy? Gait Posture 2000;11(3):181–90. [25] Sutherland DH, Schottstaedt ER, Larsen LJ, Ashley RK, Callander JN, James PM. Clinical and electromyographic study of seven spastic children with internal rotation gait. J Bone Joint Surg (Am) 1969;51A(6):1070–82. [26] Rodda JM, Graham HK. Classification of gait patterns in spastic hemiplegia and spastic diplegia: a basis for a management algorithm. Eur J Neurol 2001;8(5):S98–108.

371

[27] Winters TF, Gage JR, Hicks RR. Gait patterns in spastic hemiplegia in children and young adults. J Bone Joint Surg (Am) 1987;69(3):437– 41. [28] Hof AL, Elzinga H, Grimmius W, Halbertsma JPK. Speed dependence of averaged EMG profiles in walking. Gait Posture 2002;16:78–86. [29] Sutherland DH. The development of mature gait. Gait Posture 1997;6(2):163–70. [30] Winter DA, Yack HJ. EMG profiles during normal human walking: stride-to-stride and inter-subject variability. Electroencephalogr Clin Neurophysiol 1987;67(5):402–11.