Research Report Trunk and Hip Muscle Activation Patterns Are Different During Walking in Young Children With and Without Cerebral Palsy Laura A. Prosser, Samuel C.K. Lee, Ann F. VanSant, Mary F. Barbe, Richard T. Lauer L.A. Prosser, PT, PhD, is Postdoctoral Fellow, Rehabilitation Medicine Department, National Institutes of Health Clinical Center, Bldg 10-CRC, 1-1469, 10 Center Dr, Bethesda MD 20892 (USA). Address all correspondence to Dr Prosser at:
[email protected]. S.C.K. Lee, PT, PhD, is Assistant Professor, Physical Therapy Department, University of Delaware, Newark, Delaware, and Research Associate, Shriners Hospitals for Children, Philadelphia, Pennsylvania. A.F. VanSant, PT, PhD, FAPTA, is Professor, Department of Physical Therapy, Temple University, Philadelphia, Pennsylvania. M.F. Barbe, PhD, is Professor, Departments of Physical Therapy and Anatomy and Cell Biology, Temple University. R.T. Lauer, PhD, is Assistant Professor, Departments of Physical Therapy and Electrical and Computer Engineering, Temple University. [Prosser LA, Lee SCK, VanSant AF, et al. Trunk and hip muscle activation patterns are different during walking in young children with and without cerebral palsy. Phys Ther. 2010;90:986 –997.]
Background. Poor control of postural muscles is a primary impairment in people with cerebral palsy (CP). Objective. The purpose of this study was to investigate differences in the timing characteristics of trunk and hip muscle activity during walking in young children with CP compared with children with typical development (TD).
Methods. Thirty-one children (16 with TD, 15 with CP) with an average of 28.5 months of walking experience participated in this observational study. Electromyographic data were collected from 16 trunk and hip muscles as participants walked at a self-selected pace. A custom-written computer program determined onset and offset of activity. Activation and coactivation data were analyzed for group differences.
Results. The children with CP had greater total activation and coactivation for all muscles except the external oblique muscle and differences in the timing of activation for all muscles compared with the TD group. The implications of the observed muscle activation patterns are discussed in reference to existing postural control literature.
Limitations. The potential influence of recording activity from adjacent deep trunk muscles is discussed, as well as the influence of the use of an assistive device by some children with CP. Conclusions. Young children with CP demonstrate excessive, nonreciprocal trunk and hip muscle activation during walking compared with children with TD. Future studies should investigate the efficacy of treatments to reduce excessive muscle activity and improve coordination of postural muscles in CP.
© 2010 American Physical Therapy Association
Post a Rapid Response to this article at: ptjournal.apta.org 986
f
Physical Therapy
Volume 90
Number 7
July 2010
Muscle Activation Patterns During Walking in Children
P
muscles in individuals with CP may guide the design of postural control interventions.
oor control of trunk postural muscles is a primary impairment in people with cerebral palsy (CP).1–3 To date, the majority of work investigating postural control in CP has done so indirectly by studying lower-extremity muscle responses to balance perturbations and center-of-pressure (COP) trajectories during standing and walking or focused primarily on sitting postural control.3,4 Direct study of trunk and hip muscle activity during upright movement is desirable in CP because impairments in postural muscle function are associated with upright functional activity limitations and participation restrictions in various populations5–7 and treatments to improve trunk and hip muscle activation may increase functional ability.
The majority of gait research in CP includes participants who are between the ages of 6 years and adolescence. Recent neurorehabilitation research emphasizes the importance of early and intensive intervention after central nervous system damage to best promote neuroplasticity.12–14 If these concepts are to be applied to individuals with CP, with interventions targeted earlier after the neural insult, more knowledge is needed about motor control and the neural activation of muscles in younger children with CP than are typically studied. For this study, the function of trunk and hip muscles in young children is of particular interest.
Poor postural control in people with CP may be a direct expression of the neural insult or the result of compensations for other primary impairments, such as altered muscle tone (resistance to stretch) and deficits in neuromuscular activation. Poor postural control can cause secondary compensation by other muscles to assist in providing postural stability, reducing the effectiveness of muscles that typically function as primary movers of the extremities.8 Evidence to support this notion includes observations that children with CP have greater ambulatory ability when the distal limb musculature is primarily affected and proximal limb musculature is less affected.9,10 In addition, compared with knee and ankle muscles, the strength (force-generating capacity) of the hip abductors in children with CP explained the largest variance in gait and gross motor function.11 Given the importance of postural control for functional mobility and the limitations associated with impairments in trunk and hip muscles in people with CP, direct investigation of the activation patterns of trunk and hip
Electromyographic (EMG) analysis is an important component in the examination of muscle function in people with CP. Several factors can be extracted from the EMG signal to provide insight into muscle activation patterns.9,15,16 The most common clinical use of muscle EMG analysis is to determine the onset and offset timing of muscle activity during movement. This type of temporal information identifies periods of muscle activity and inactivity throughout the gait cycle and can be used to determine coactivation of antagonistic muscle groups. The primary objective of this study was to investigate differences in the timing characteristics of trunk and hip muscle activity during walking in young children with CP compared with children with typical development (TD). Identifying early deficits in postural muscle function in individuals with CP may lead to the development of more direct interventions that have the potential to improve postural control before compensatory postural motor strategies are reinforced. A secondary objective was to report typical activity patterns in
July 2010
the trunk and hip muscles in young children with TD.
Method Participants Participants with CP were recruited through the CP clinics at the Shriners Hospital for Children in Philadelphia, Pennsylvania, the Children’s Specialized Hospital in Mountainside, New Jersey, and other local rehabilitation facilities. Participants with TD were recruited from siblings of the participants with CP, from children of people known to the investigators, and from a local day care center. All data collection procedures were explained, and parents gave their informed consent to the research and to publication of the results. Assent of a minor also was obtained from participants who were 7 years of age or older. The inclusion criteria for all children were: 0.5 to 60 months of walking experience, ability to ambulate barefoot at least 4.6 m (15 ft) with supervision (children with CP could use their usual assistive device [ie, crutches or walker without pelvic guide] if it did not stabilize or restrict movement of the trunk or pelvis), and the ability to follow 1-step verbal directions. Children with CP additionally had a diagnosis of spastic diplegia or quadriplegia and a Gross Motor Function Classification System (GMFCS) level of II or III.17 Children were not considered for the study if they had a lower-extremity fracture or surgery in the previous 12
Available With This Article at ptjournal.apta.org • The Bottom Line Podcast • Audio Abstracts Podcast This article was published ahead of print on April 29, 2010, at ptjournal.apta.org.
Volume 90
Number 7
Physical Therapy f
987
Muscle Activation Patterns During Walking in Children months, a botulinum toxin injection in the previous 6 months, or a history of dorsal rhizotomy or lowerextremity tendon transfer.
months, was calculated as the difference between the participant’s age on the day of the study and the age at onset of walking.
The selection of months of walking experience rather than age as a primary inclusion criterion was based on reports that experience is a stronger predictor of walking and balance skill than age in early walkers.18,19 The onset of walking was operationally defined as the age when the child was able to take at least 3 continuous, independent steps (without assistance of another person, but may have been with an assistive device).20 Walking experience, in
Procedure Anthropometric measurement. All anthropometric measurements were taken by the same pediatric physical therapist and included height, seated height, weight, and bilateral leg length (anterior superior iliac spine to the apex of the medial malleolus). All anthropometric lengths were measured with a Harpenden anthropometer,* with the exception of those in 2 children who were fearful of the device. For these children, a
The Bottom Line What do we already know about this topic? Individuals with spastic cerebral palsy (CP) demonstrate excessive activation of lower-extremity muscles during walking. Although impairment of postural control is a primary characteristic of CP, the muscle activation patterns of the postural muscles during walking are unknown in both children with CP and children with typical development.
What new information does this study offer? Young children with CP demonstrated differences in the timing patterns of muscle activation during walking for the trapezius, erector spinae, rectus abdominis, external oblique, gluteus maximus, gluteus medius, rectus femoris, and semitendinosus muscles. Children with CP also demonstrated greater total muscle activation and coactivation in all muscles studied except the external oblique. This excessive activity was characterized by early onset and late cessation of activity in the external oblique, gluteus maximus, and gluteus medius muscles, and by continuous activation in the trapezius, erector spinae, rectus femoris, and semitendinosus muscles.
If you’re a patient, what might these findings mean for you? Young children with CP may benefit from treatments that focus on improving the coordination of trunk and hip muscles during walking, and from treatments that focus on reducing the amount of time those muscles are active during the walking cycle. However, specific treatments with these goals have not yet been investigated.
988
f
Physical Therapy
Volume 90
Number 7
standard tape measure was used in lieu of the anthropometer. EMG instrumentation. Surface EMG data from trunk, gluteal, and thigh muscles were acquired using a 16-channel recording system (Myomonitor III and trigger module†) with preamplified silver-silver chloride parallel bar surface electrodes with a 10.0-mm interelectrode distance. The EMG data were collected at 1,200 Hz, preamplified with a gain of 10, and band-pass filtered between 20 and 450 Hz. The EMG data were collected from 8 muscles bilaterally: trapezius (TZ), erector spinae (ES), rectus abdominis (RA), external oblique (EO), gluteus maximus (GMx), gluteus medius (GMd), rectus femoris (RF), and semitendinosus (ST) (Tab. 1). The RF and ST were chosen in addition to the trunk and gluteal muscles because these muscles anatomically cross the hip joint. Sensor placement for the abdominal muscles was determined using the methods described by Ng et al.21 Sensor placement for all other muscles (back, gluteal, and thigh) was determined in accordance with the SENIAM project recommendations (Tab. 1).22 The skin areas were cleaned with alcohol, and the sensors were affixed to the skin with double-sided adhesive. The electrodes were further secured using hypoallergenic tape or a flexible, latex-free, nonadhesive wrap (Coflex-NL‡) encircling the waist and thighs. Self-adhesive reference electrodes§ were placed on the skin over the patella bilaterally. A
* Holtain Ltd, Crosswell, Crymych, Pembs, SA41 3UF United Kingdom. † Delsys Inc, PO Box 15734, Boston, MA 02215. ‡ Andover Healthcare Inc, 9 Fanaras Dr, Salisbury, MA 01952. § Axelgaard Manufacturing Co Ltd, Lystrup 8520, Denmark.
July 2010
Muscle Activation Patterns During Walking in Children volitional contraction of each muscle, when possible, was elicited to verify placement and confirm the absence of electrical activity from an adjacent muscle. The children were asked to perform specific movements to elicit the corresponding muscle contractions, such as leaning backward in a sitting position to activate the RA muscles or standing on one leg (with hands held, if needed) to activate the GMd muscles. Data from 2 static baseline EMG trials were collected prior to the walking trials to establish baseline muscle activity. For these trials, the child lay still in a supine position for 5 seconds. The trial with the least muscle activity was selected for postprocessing. Walking trials. Children walked barefoot down an instrumented walkway (GAITRite㛳) at a selfselected pace. Data from 3 to 5 trials, each consisting of at least 4 consecutive footfalls, were collected, depending on participant tolerance and fatigue. Start and stop targets were placed on the floor approximately 1.5 m (5 ft) beyond either end of the instrumented walkway to minimize acceleration or deceleration while walking on the walkway. A walking trial was started by having the child stand on the start target. Data collection was initiated through the GAITRite software, which triggered EMG collection through a trigger module† for synchronous recording. The child then was instructed to walk to the target beyond the opposite end of the walkway. Children had the opportunity to sit on a chair between walking trials to minimize fatigue. During the walking trials, the EMG preamplification unit, which typically is worn on a backpack, was 㛳
CIR Systems, 60 Garlor Dr, Havertown, PA 19083.
July 2010
Table 1. Electromyogram Sensor Locations Muscle Trapezius
(middle)22
Sensor Location 50% of the distance from the medial border of the scapula to the T3 spinous process, parallel to the line from T5 to the acromion
Erector spinae (longissimus)22
1–2 finger widths lateral from the L1 spinous process, oriented vertically
Rectus abdominis21
At the level of the anterior superior iliac spine, 1–2 cm lateral to the midline, oriented vertically
External oblique21
Just below the rib cage at the inferior angle of the ribs, oriented obliquely
Gluteus maximus22
50% of the distance from the sacral vertebrae to the greater trochanter, at the greatest prominence of the middle buttocks, parallel to a line from the posterior superior iliac spine to the middle posterior thigh
Gluteus medius22
In a side-lying position, 50% of the distance from the iliac crest to the greater trochanter, parallel to this axis
Quadriceps femoris (rectus femoris)22
50% of the distance from the anterior superior iliac spine to the superior patella, parallel to this axis
Semitendinosus22
50% of the distance from the ischial tuberosity to the medial tibial epicondyle, parallel to this axis
carried behind all participants by an assistant so as not to add weight that could affect muscle activity in the smaller children. If needed, children were motivated and rewarded with stickers, small snacks, or favorite toys. Walking trials were videotaped for later gait cycle selection, and parents or caregivers signed a separate consent statement to allow videotaping. Because this was an observational study that attempted to describe the natural pattern of muscle activation at equivalent levels of walking experience, walking speed was not controlled. The majority of children with CP are unable to match the walking speed of children with TD; therefore, the only reasonable method to control for speed would be for the children with TD to walk at greatly reduced speeds. In addition to being a difficult task to have young children (TD group’s mean age⫽39.7 months) walk at a steady state at target slow speeds, purposely altering walking speed to walk slowly (within individuals) has been shown
to significantly alter typical muscle activation patterns, to a greater extent than walking at fast speeds and particularly in proximal muscles.23,24 This method would have confounded the secondary objective of this study (ie, to report typical, unaltered patterns in trunk and hip muscles for young children with TD). Data Analysis Video footage of each trial was reviewed to determine the most appropriate gait cycles to select for data analysis. Ten gait cycles (5 left, 5 right) were selected based on the observation of each individual’s typical walking pattern (walking without distractions, pauses, or moving arms toward an object). The 10 selected gait cycles were extracted from the EMG files using the timesynchronized marker data (initial foot contact for consecutive footfalls) collected from the instrumented walkway. Stance phase was identified as the period from initial foot contact to ipsilateral toe-off. Swing phase was identified as the period from toe-off to subsequent
Volume 90
Number 7
Physical Therapy f
989
1.10 (0.14), 0.75–1.32 16.5 (2.2), 11.2–22.9 0.56 (0.06), 0.46–0.69 17.3 (5.4), 10.0–31.2 28.5 (18.1), 1.0–60.0 31 22.9 (13.8), 8.0–55.0 Total
a
15 34.8 (10.2),* 18.0–55.0
Data reported as mean (SD), range of values. TD⫽typically developing, CP⫽cerebral palsy, M⫽male, F⫽female, BMI⫽body mass index. Asterisks indicate a significant difference from the TD group (P⬍.05).
1.06 (0.14), 0.83–1.32 17.2 (2.4), 14.7–22.9 0.56 (0.06), 0.48–0.65 19.6 (5.9),* 10.9–31.2
14 F, 17 M 51.0 (24.1), 13.0–108.0
0.97 (0.13), 0.75–1.18 15.9 (1.8), 11.2–18.8 0.55 (0.06), 0.46–0.69 15.1 (3.9), 10.0–21.9
63.1 (23.2),* 25.0–108.0
39.7 (19.5), 13.0–67.5 9 F, 7 M
5 F, 10 M 28.4 (17.0), 2.0–60.0
28.6 (19.6), 1.0–58.0 16 11.7 (3.1), 8.0–20.0
CP
Onset of Walking (mo)
TD
Physical Therapy
N
Demographic and Anthropometric Dataa
f
Group
Table 2. 990
Walking Experience (mo)
Sex
Age (mo)
Weight (kg)
Height (m)
BMI (kg/m2)
Seated Height (m)
Muscle Activation Patterns During Walking in Children
Volume 90
foot contact. The EMG data were processed using custom-written programs in MATLAB software.# All signals were normalized to 1,000 points, representing the gait cycle from 0% to 100% in 0.1% increments.
muscle was active in both the CP and TD groups. Assuming unequal variance between groups, MannWhitney U tests were used to determine differences in percent activation and coactivation (P⬍.05).
To identify muscle activity throughout the selected gait cycles, the Teager-Kaiser energy operator was applied to the EMG data.25 This method uses both the amplitude and frequency components of the signal and was found to better detect the onset of muscle activity than a standard amplitude threshold method.26,27 The EMG data from the selected gait cycles for each participant were filtered with a second-order, low-pass Butterworth filter with phase correction and a cutoff frequency of 10 Hz and averaged across cycles. To determine the onset/offset threshold, the TeagerKaiser energy operator was applied to the static EMG baseline signal. The resulting output then was rectified, and the mean and standard deviation were calculated. The mean, plus 1 standard deviation, was used as the threshold above which defined muscle activity during walking.
An additional analysis was performed to determine when during the gait cycle muscle activity varied between groups. For this analysis, the gait cycle was reduced to 100 points (1% increments), and the number of children in each group who had activity in the muscle at each point in the gait cycle was determined. The chi-square test was performed at each point in the gait cycle to determine whether significant differences (P⬍.05) existed between groups in the number of children who had activity in the particular muscle.28 All statistical analyses were performed using SPSS software, version 11.0.**
Total activation and coactivation first were analyzed as a percent of the gait cycle. A percent of the gait cycle for activation was calculated for each muscle by summing the duration (in percents) of all periods of muscle activity. Coactivation was determined by calculating the total time (in percents) antagonistic muscles were simultaneously active. Coactivation was calculated for the ipsilateral RA-ES and RF-ST muscle pairs. Group means were calculated for percent muscle activation and coactivation, from which 95% confidence intervals (CIs) were determined. These measures allowed for comparison of total relative time a particular # The Mathworks Inc, 3 Apple Hill Dr, Natick, MA 01760.
Number 7
Role of the Funding Sources Direct costs for this study, including data acquisition equipment, supplies, travel to data collection sites, and consultant fees, were funded by a Clinical Research Grant to Dr Prosser from the Section on Pediatrics, American Physical Therapy Association. A National Institute of Neurological Disorders and Stroke grant (R03NS048875) to Dr Lauer funded the time of the principal investigator and a coinvestigator and costs related to dissemination. A National Institute of Child Health and Human Development grant (R01HD043859) to Dr Lee funded the time of a coinvestigator and research aides. This research also was supported, in part, by the Intramural Research Program of the Clinical Center, National Institutes of Health.
** SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.
July 2010
Muscle Activation Patterns During Walking in Children
Results A power analysis was performed after the initial 9 children (6 with TD, 3 with CP) completed the study. A sample size of 8 to 28 individuals (4 –14 in each group) was predicted to power individual muscle activity variables to 0.80 or greater at the P⬍.05 level of significance. The maximum sample was used to adequately power all variables, and data were collected from an additional 20% (6 children) to account for anticipated difficulties with participant tolerance and for cases of unusable data. Thus, a total of 34 children were enrolled in this study. Data from 3 children were excluded (1 child had a questionable diagnosis of CP, and 2 children were unable to walk without additional assistance from an investigator during the testing session). Data for the remaining 31 children (17 male, 14 female; 15 with CP, 16 with TD) were used for analysis. Walking experience did not differ between groups (P⫽.969). The children with CP were heavier (P⫽.017) and had longer legs (P⫽.029) than the children with TD due to a later onset of walking and, therefore, were older at the time of testing. In the group of children with CP, 7 were classified as GMFCS level II, and 8 were classified as level III. One child was classified as spastic quadriplegic, and 14 children were classified as spastic diplegic. Three children walked without assistive devices, 9 used posterior rolling walkers, 1 used bilateral forearm crutches, and 2 used unilateral forearm crutches. Data for range of motion were not obtained from 1 child with CP due to time constraints. Demographic and anthropometric data are provided in Table 2. Walking speed was normalized using leg length as per Hof29 and was 0.22 (SD⫽0.10) for the CP group and 0.42 (SD⫽0.06) for the TD group.
July 2010
Figure 1. Mean total percent activation of trunk and hip muscles for children with cerebral palsy (CP) and children with typical development (TD). Bars represent upper and lower bounds of 95% confidence intervals. Asterisks indicate muscles that were significantly higher in the CP group compared with the TD group. TZ⫽trapezius, ES⫽erector spinae, RA⫽rectus abdominis, EO⫽external oblique, GMx⫽gluteus maximus, GMd⫽gluteus medius, RF⫽rectus femoris, and ST⫽semitendinosus.
Muscle activation from left and right sides did not differ within groups; thus, left-side and right-side data were pooled for percent total activation and coactivation. The CP group had significantly more total activation time for each muscle (ranging from P⬍.001 to P⫽.024), except for the EO, which was not different from that of the TD group (P⫽.593). Group means for activation, including 95% CIs, are shown in Figure 1. The CP group also had significantly more total coactivation time for both the RA-ES muscle pair (P⫽.007) and the RF-ST muscle pair (P⬍.001). Coactivation for the RA-ES muscle pair averaged 20% (95% CI⫽5%–36%) for the CP group and 1% (95% CI⫽0%– 3%) for the TD group. Coactivation for the RF-ST muscle pair averaged 75% (95% CI⫽61%– 89%) for the CP group and 20% (95% CI⫽10%–30%) for the TD group. To determine where in the gait cycle the children with CP had excessive muscle activity, histograms were generated to show the number of active muscles at each point in the gait cycle. No meaningful differences existed between left and right sides in number of active muscles or in
percent stance time (57% in the TD group, 59% in the CP group) or percent swing time. However, because of potential differences in left and right symmetry within individuals (eg, left side active, right side not active) and the inability to average these nominal data, each side was counted individually, resulting in a maximum count of 30 muscles for the CP group and 32 muscles for the TD group. Figure 2 shows histograms for each muscle for both groups. The asterisks in the figure identify the ranges in the gait cycle when the CP group had significantly more active muscles compared with the TD group, as determined by the chi-square tests. The TZ was active in more children in the CP group compared with the TD group throughout the majority of the gait cycle, except for the period from mid-stance to late stance. The ES was active in more children in the CP group from just prior to initial contact through mid-stance. The RA also was active in more children in the CP group throughout most of the gait cycle. There were no differences in EO activity between groups at any point in the stride. The GMx was active in more children in the CP group during
Volume 90
Number 7
Physical Therapy f
991
Muscle Activation Patterns During Walking in Children both the stance and swing phases of gait, but not during the transitions between the phases. The GMd and ST were active in more children in the CP group from mid-stance through early swing. Except for a short period of time around initial contact, the RF was active in more children in the CP group throughout the gait cycle. A summary figure was generated from the histograms to summarize the periods of activity for each muscle in each group. Each group was considered to have activity at a particular point in the gait cycle if the number of active muscles was at least one half of the total number of muscles at that point (15/30 for the CP group, 16/32 for the TD group). Figure 3 shows the generalized periods of activity for each muscle in each group. The EO had several points in the gait cycle for each group when activity was present for one less than one half of the total number of muscles. Because these points were in the midst of large periods of activity, they were included within those periods of activation. The TZ, ES, RF, and ST demonstrated activity throughout the gait cycle in the majority of children with CP. The EO, GMx, and GMd demonstrated similar phases of activity in the CP group compared with the TD group, but these muscles demonstrated longer periods of activation in the CP group, including both earlier onset and delayed offset of activity. Although the RA demonstrated more total muscle activation in the CP group compared with the TD group, it was not active in most children with CP throughout the gait cycle. Figure 2. Histograms for number of active muscles at each point in the gait cycle (1% increments) in children with typical development (TD) and children with cerebral palsy (CP). Left and right sides were counted individually, for a maximum of 32 in the TD group and 30 in the CP group. Thick vertical lines indicate toe-off, the transition from stance phase to swing phase (57% of the gait cycle in the TD group, 59% of the gait cycle in the CP group). Asterisks indicate periods of activity where the CP group had significantly more active muscles than the TD group. TZ⫽trapezius, ES⫽erector spinae, RA⫽rectus abdominis, EO⫽external oblique, GMx⫽gluteus maximus, GMd⫽gluteus medius, RF⫽rectus femoris, and ST⫽semitendinosus.
992
f
Physical Therapy
Volume 90
Number 7
Discussion This study is the first to investigate muscle activation patterns of the trunk and hip muscles during walking in children with TD and children with CP. It also demonstrates the use of quantitative methods of analyzing EMG signals to determine periods of July 2010
Muscle Activation Patterns During Walking in Children activity and inactivity across the gait cycle. Increased activity for the CP group was observed as prolonged durations of bursts of activity and muscles that were continuously activated throughout the gait cycle compared with the TD group. Even for the EO, which was not significantly different between groups for total muscle activity, more than half of the CP group had activity during 80% of the gait cycle compared with 39% of the gait cycle in the TD group. Excessive muscle activity has been reported in the lower-extremity muscles in children with CP.10,30 In these studies, however, the children with CP were compared with children with TD of the same age who had more walking experience because walking onset is delayed in children with CP. Children with TD who have less walking experience are known to have more muscle activity than children with TD who have more walking experience.31 Using walking experience, not age, for inclusion in this study is a novel approach to control for the improvement in walking ability that occurs with practice after the onset of walking. The increased muscle activity in the CP group in this study, however, was present even as the groups were compared by walking experience rather than age. The patterns of muscle activity in the GMd and ST in the TD group were consistent with those reported previously by Sutherland and colleagues31 in their study of more than 300 children between the ages of 1 and 7 years. The period of activity in the GMx in our study was slightly shorter than that reported in the study by Sutherland and colleagues. This difference can likely be attributed to differences between the 2 studies in the methods used to determine activation. Sutherland and colleagues used visual inspection of the raw EMG signals, whereas we used July 2010
Figure 3. Muscle activity during the gait cycle in children with typical development (TD) and children with cerebral palsy (CP). Dashed vertical lines indicate toe-off, the transition from stance phase to swing phase (57% of the gait cycle in the TD group, 59% of the gait cycle in the CP group). TZ⫽trapezius, ES⫽erector spinae, RA⫽rectus abdominis, EO⫽external oblique, GMx⫽gluteus maximus, GMd⫽gluteus medius, RF⫽rectus femoris, and ST⫽semitendinosus.
advanced processing methods and objective rules to determine activation patterns. Although the GMd, ST, and GMx were the only muscles in the current study for which activation patterns have been reported previously for young children with TD, muscle activity patterns during walking have been shown to approximate adult patterns after the age of 3 years.31 For this reason, comparison of data for the other muscles included in this study with data from adults is justified. Timing patterns of activity for the RF, as well as the gluteal muscles, in the TD group are consistent with those reported in adults.32 White and McNair33 investigated patterns of activity in the ES, RA, and EO in adults. They did not use a threshold to determine the timing of muscle activity onset and offset, but rather identified different patterns by averaging normalized amplitude curves across participants. Therefore, exact comparison is not possible, but areas of increased normalized amplitude for the ES and EO in their study do cor-
respond to periods of activity for the ES and EO in the TD group in the current study. Also similar to the TD group in the current study, the majority of adult participants in the study by White and McNair did not have periods of activity above baseline in the RA. The comparison of TD muscle activity data from the current study with data from the studies mentioned above demonstrates that, similar to lower-extremity muscles, activation patterns of the trunk muscles approximate those of adult patterns by a young age. As expected, self-selected walking speed was slower in the CP group than in the TD group. It is possible that differences in walking speed between the groups had some effect on the current results. However, previous studies24,34 have shown significant changes in EMG signal amplitude with changes in walking speed, but little effect on the timing of muscle activation at speeds similar to that of the current CP group (0.50 m/s). In contrast, one study23 showed increased proximal muscle
Volume 90
Number 7
Physical Therapy f
993
Muscle Activation Patterns During Walking in Children coactivation at “very slow” speeds, slower than that of the current CP group. However, comparison of the current data with the data from these studies using adults walking at fixed speeds is problematic because the speeds were not normalized for body size. Similar muscle activity timing patterns were reported in children 3 years of age and older compared with adults, despite a significantly slower walking speed.31 If the fixed speeds in these comparison studies were adjusted for children, it is likely that even fewer or no differences in activation and coactivation timing would have been observed at speeds comparable to those of the current CP group. Furthermore, even with normalization of walking speeds, caution must be taken because studies to date have investigated only differences within individuals and not across specific diagnostic groups. To our knowledge, only Schwartz et al35 have investigated the effect of walking speed on muscle activity patterns in children with TD. As in our work, Schwartz and colleagues did not enforce fixed speeds due to concerns about alterations in typical gait patterns that may occur if the individuals were paced. Instead, they instructed children to walk at “very slow,” “slow,” and “fast” speeds, and they later classified normalized walking speeds based on the number of standard deviations from each individual’s self-selected speed. In our study, the CP group walked at speeds equivalent to the slow speed in the study by Schwartz and colleagues. Although EMG amplitude generally decreased at slow and very slow speeds in the study by Schwartz and colleagues, the most notable difference in a comparable muscle was seen in the RF, which demonstrated relative inactivity at the slow speeds in the study by Schwartz and colleagues. This result is counter to the results of the current study, in which 994
f
Physical Therapy
Volume 90
constant activity was observed in the RF in the slow-walking CP group, further emphasizing the excessive activity in the CP group. However, research is needed to understand the effects of slow and fast walking speeds specifically in children with CP. Despite our use of automated EMG processing techniques and objective rules to determine the duration of muscle activity in each group, there remains no flawless method to analyze EMG signals, particularly in children with neurological impairments. Unlike the larger and thicker muscles of the thigh and gluteal region, the superficial muscles of the trunk are thin, and the recording sensors may have recorded some activity from the underlying muscles. The internal oblique, rhomboids, and transverse abdominis muscles are directly deep to the sensor location for the EO, TZ, and RA, respectively. The ES was recorded from a location that is deep to the broad superficial fascia of the latissimus dorsi muscle. The use of fine-wire needle EMG electrodes would avoid this potential issue, but application in young children has clear feasibility and ethical limitations. Additionally, needle electrodes record only from a single or small group of motor units, and as a result, the recorded signal may not be representative of the activity of the entire muscle.36 An additional limitation exists in the comparison of muscle activity between the TD and CP groups, especially for the TZ, because the majority (12/15) of the children in the CP group used an assistive device for walking. The mean percent activation of the TZ in the children with CP who used an assistive device was 90%. Of the 3 children with CP who did not use an assistive device, 1 child had activation similar to the device-using mean (88%), and the other 2 children had far less activa-
Number 7
tion (22% and 11%) than both the device-using mean and the TD group mean (45%). Although these limited data are inconclusive, it is possible that use of an assistive device alone may have contributed to greater activation of the TZ in some children in the CP group. The TZ may be activated during forward movement of and bearing weight through the assistive device. Use of a device may have additionally affected the activation timing of other muscles. This issue is difficult to avoid when studying young children with CP. According to the GMFCS classification,17 only children classified as level I (least impaired) begin to walk without the use of any assistive device. Children classified as level II walk with an assistive device for the first few years of walking, then later do not require a device. Children classified as level III require an assistive device for functional ambulation. Therefore, to study any children with greater severity of CP than those who are least impaired (GMFCS level I) during the early years of walking, the use of walking aids must be allowed. We briefly considered collecting data from trials with the TD group using assistive devices to control for this difference. Walking with an assistive device, however, would have been a novel task for the children with TD, and there are well-known differences in the timing and magnitude of muscle activation during novel tasks.37 Future work could include several CP groups, classified by GMFCS level, gait pattern, the particular assistive device used, and investigate various walking speeds. Finally, although the use of an assistive device may affect muscle activation in addition to motor control differences, it is important to recognize that the children with CP use more postural muscle effort in daily walking than their peers, indifferent of the cause, which is important from cardiovasJuly 2010
Muscle Activation Patterns During Walking in Children cular endurance, muscular fatigue, shoulder kinetics, and biomechanical efficiency perspectives. A final limitation is the consideration that, although the groups had equal time since the onset of walking (walking experience), the amount of walking practice was still likely to be higher in the TD group. After the onset of walking, walking quickly becomes the primary means of mobility for children with TD. Children with CP often practice walking only in therapy sessions or during more structured periods of practice at home. Even when of school age, children with CP take fewer steps per day than children with TD.38 As a measure of neural activation of skeletal muscle, surface EMG is a noninvasive, indirect measure of motor control. An understanding of neuromuscular control is useful to inform clinicians and researchers of the specific impairments that contribute to functional activity limitations and participation restrictions. Specifically, making clinical decisions to treat patients with deficits in postural control requires an understanding of the causes of poor postural control. Reducing excessive postural muscle activity and improving coordination and reciprocation among postural muscles in people with CP might be accomplished through task-specific, repeated practice paradigms, biofeedback, electrical stimulation, pharmacological agents, or exercise. Several other studies39 – 43 have examined postural control in people with CP during static standing, during reaching, and during external perturbation balance testing. These studies demonstrated increased coactivation, prolonged latency of activation, altered muscle recruitment order following perturbations, and continuous activation of lowerextremity and postural muscles. The July 2010
present study demonstrated similar findings for muscle activation during walking, including increased coactivation and continuous activation of postural muscles. With the exception of the RA and GMx, all muscles in the CP group were active over 75% of the gait cycle. This excess activation may create a functionally rigid trunk, which may restrict the child’s ability to make fine adjustments to trunk position relative to the lower extremities and the environment and constrain the therapist’s ability to grade muscle activity in response to external perturbations.4 Roerdink and colleagues44 reported that, after a stroke, individuals had less stability but also more regularity in frontalplane COP trajectories during standing compared with their peers who were healthy. With recovery and rehabilitation, COP trajectories became less regular. They suggested that, after stroke, the participants attempted to limit variations in COP in order to decrease the degrees of freedom that they must control and that the individuals were better able to control multiple degrees of freedom after rehabilitation. A similar strategy may occur in young children with CP. Hsue and reported reduced colleagues45 anterior-posterior displacements of COP and center of mass (COM) during walking in children with CP. Limiting excursion and variability in COP and COM by excessively activating muscles of the trunk and hips may be a strategy that children with CP use to maintain upright posture against gravity and move the body forward, despite the multitude of neurological impairments limiting typical movement patterns. Excessive muscle activation may be a compensation for poor control of postural muscles and may limit the ability to precisely control changes in their body’s COM during dynamic
movements. If increased irregularity of COP trajectories is a favorable result of rehabilitation after stroke (as in the study by Roerdink and colleagues), perhaps interventions that aim to completely reduce static standing postural sway in people with CP should be closely reexamined, and training to encourage controlled postural sway in all directions should be investigated. Current therapeutic, medical, and surgical treatment for people with CP focuses on upper- and lowerextremity interventions for spasticity (hypertonicity) management, musculoskeletal abnormalities, and functional training.46,47 The results of this study suggest that postural muscle control training should be investigated to address impairments of the trunk and to encourage the development of more functional, reciprocal, and efficient movement strategies in children with CP. Core muscle control is related to athletic performance and function in adults who are healthy,48,49 and improving core muscle control may have promise in people with CP.50 Strategies to increase the child’s ability to control greater variations in trunk movement through phasic trunk muscle coordination, rather than constant muscle activity, may encourage more effective and efficient patterns of postural muscle control, which, in turn, may encourage more efficient patterns of movement in the upper and lower extremities. All authors provided concept/idea/research design and reviewed the manuscript prior to submission. Dr Prosser, Dr Lee, Dr VanSant, and Dr Lauer provided writing and facilities/ equipment. Dr Prosser provided data collection, project management, and participants. Dr Prosser, Dr Lee, and Dr Lauer provided data analysis and fund procurement. The authors thank Steve Capella and Jenny Lee for their assistance with data collection and Diana Deshefy, PT, DPT, Samuel Pierce, PT, PhD, and Erin Sheeder, PT, DPT, for assistance with participant recruitment.
Volume 90
Number 7
Physical Therapy f
995
Muscle Activation Patterns During Walking in Children Preliminary results of this study were included in a poster presentation at the Combined Sections Meeting of the American Physical Therapy Association; February 9 –12, 2009; Las Vegas, Nevada. The Institutional Review Board of Temple University Hospital (for Shriners Hospital) and the institutional review boards of the additional data collection sites approved all procedures. Direct costs for this study, incuding data acquisition equipment, supplies, travel to data collection sites, and consultant fees, were funded by a Clinical Research Grant to Dr Prosser from the Section on Pediatrics, American Physical Therapy Association. A National Institute of Neurological Disorders and Stroke grant (R03NS048875) to Dr Lauer funded the time of the principal investigator and a coinvestigator and costs related to dissemination. A National Institute of Child Health and Human Development grant (R01HD043859) to Dr Lee funded the time of a coinvestigator and research aides. This research also was supported, in part, by the Intramural Research Program of the Clinical Center, National Institutes of Health. This article was submitted May 15, 2009, and was accepted March 22, 2010. DOI: 10.2522/ptj.20090161
References 1 Rosenbaum P, Paneth N, Leviton A, et al. A report: the definition and classification of cerebral palsy, April 2006. Dev Med Child Neurol Suppl. 2007;109:8 –14. 2 Davis MF, Worden K, Clawson D, et al. Confirmatory factor analysis in osteopathic medicine: fascial and spinal motion restrictions as correlates of muscle spasticity in children with cerebral palsy. J Am Osteopath Assoc. 2007;107:226 – 232. 3 van der Heide JC, Hadders-Algra M. Postural muscle dyscoordination in children with cerebral palsy. Neural Plasticity. 2005;12:197–203. 4 Woollacott MH, Crenna P. Postural control in standing and walking in children with cerebral palsy. In: Hadders-Algra M, Carlberg EB, eds. Postural Control: A Key Issue in Developmental Disorders. London, United Kingdom: Mac Keith Press; 2008: 97–130. 5 Lin SI, Woollacott MH, Jensen JL. Postural response in older adults with different levels of functional balance capacity. Aging Clin Exp Res. 2004;16:369 –374. 6 Era P, Avlund K, Jokela J, et al. Postural balance and self-reported functional ability in 75-year-old men and women: a crossnational comparative study. J Am Geriatr Soc. 1997;45:21–29.
996
f
Physical Therapy
Volume 90
7 Liao HF, Hwang AW. Relations of balance function and gross motor ability for children with cerebral palsy. Percept Mot Skills. 2003;96(3 pt 2):1173–1184. 8 Nicholson JH, Morton RE, Attfield S, Rennie D. Assessment of upper-limb function and movement in children with cerebral palsy wearing lycra garments. Dev Med Child Neurol. 2001;43:384 –391. 9 Lauer RT, Stackhouse CA, Shewokis PA, et al. A time-frequency based electromyographic analysis technique for use in cerebral palsy. Gait Posture. 2007;26: 420 – 427. 10 Policy JF, Torburn L, Rinsky LA, Rose J. Electromyographic test to differentiate mild diplegic cerebral palsy and idiopathic toe-walking. J Pediatr Orthop. 2001;21: 784 –789. 11 Ross SA, Engsberg JR. Relationships between spasticity, strength, gait, and the GMFM-66 in persons with spastic diplegia cerebral palsy. Arch Phys Med Rehabil. 2007;88:1114 –1120. 12 Nudo RJ. Adaptive plasticity in motor cortex: implications for rehabilitation after brain injury. J Rehabil Med. 2003; 41(suppl):7–10. 13 Horn SD, DeJong G, Smout RJ, et al. Stroke rehabilitation patients, practice, and outcomes: Is earlier and more aggressive therapy better? Arch Phys Med Rehabil. 2005; 86(12 suppl 2):S101–S114. 14 Dobkin B, Barbeau H, Deforge D, et al. The evolution of walking-related outcomes over the first 12 weeks of rehabilitation for incomplete traumatic spinal cord injury: the multicenter randomized Spinal Cord Injury Locomotor Trial. Neurorehabil Neural Repair. 2007;21:25–35. 15 Hodges PW, Bui BH. A comparison of computer-based methods for the determination of onset of muscle contraction using electromyography. Electroencephalogr Clin Neurophysiol. 1996;101:511– 519. 16 Roetenberg D, Buurke JH, Veltink PH, et al. Surface electromyography analysis for variable gait. Gait Posture. 2003;18: 109 –117. 17 Palisano R, Rosenbaum P, Walter S, et al. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol. 1997;39:214 –223. 18 Adolph KE, Vereijken B, Shrout PE. What changes in infant walking and why. Child Dev. 2003;74:475– 497. 19 Sundermier L, Woollacott MH, Roncesvalles N, Jensen J. The development of balance control in children: comparisons of EMG and kinetic variables and chronological and developmental groupings. Exp Brain Res. 2001;136:340 –350. 20 Wu J, Looper J, Ulrich BD, et al. Exploring effects of different treadmill interventions on walking onset and gait patterns in infants with Down syndrome. Dev Med Child Neurol. 2007;49:839 –945.
Number 7
21 Ng JK, Kippers V, Richardson CA. Muscle fibre orientation of abdominal muscles and suggested surface EMG electrode positions. Electromyogr Clin Neurophysiol. 1998;38:51–58. 22 Hermens HJ, Freriks B, Disselhorst-Klug C, Rau G. Development of recommendations for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol. 2000;10:361–374. 23 Nymark JR, Balmer SJ, Melis EH, et al. Electromyographic and kinematic nondisabled gait differences at extremely slow overground and treadmill walking speeds. J Rehabil Res Dev. 2005;42:523–534. 24 van Hedel HJ, Tomatis L, Muller R. Modulation of leg muscle activity and gait kinematics by walking speed and bodyweight unloading. Gait Posture. 2006;24:35– 45. 25 Li X, Zhou P, Aruin AS. Teager-Kaiser energy operation of surface EMG improves muscle activity onset detection. Ann Biomed Eng. 2007;35:1532–1538. 26 Solnik S, DeVita P, Rider P, et al. TeagerKaiser operator improves the accuracy of EMG onset detection independent of signal-to-noise ratio. Acta Bioeng Biomech. 2008;10:65– 68. 27 Lauer RT, Prosser LA. Use of the TeagerKaiser energy operator for muscle activity detection in children. Ann Biomed Eng. 2009;37:1584 –1593. 28 Portney LG, Watkins MP. Foundations of Clinical Research, Applications to Practice. Stamford, CT: Appleton & Lange; 1993. 29 Hof AL. Scaling gait data to body size. Gait Posture. 1996;4:222–223. 30 Unnithan VB, Dowling JJ, Frost G, et al. Co-contraction and phasic activity during gait in children with cerebral palsy. Electromyogr Clin Neurophysiol. 1996;36: 487– 494. 31 Sutherland D, Olshen R, Biden E, Wyatt M. The Development of Mature Walking. London, United Kingdom: Cambridge University Press; 1988. 32 Soderberg GL, Knutson LM. EMG methodology. In: Craik RL, Oatis CA, eds. Gait Analysis: Theory and Application. St Louis, MO: Mosby; 1995:293–306. 33 White SG, McNair PJ. Abdominal and erector spinae muscle activity during gait: the use of cluster analysis to identify patterns of activity. Clin Biomech (Bristol, Avon). 2002;17:177–184. 34 Den Otter AR, Geurts AC, Mulder T, Duysens J. Speed related changes in muscle activity from normal to very slow walking speeds. Gait Posture. 2004;19:270 – 278. 35 Schwartz MH, Rozumalski A, Trost JP. The effect of walking speed on the gait of typically developing children. J Biomech. 2008;41:1639 –1650. 36 Soderberg GL, Knutson LM. A guide for use and interpretation of kinesiologic electromyographic data. Phys Ther. 2000;80: 485– 498.
July 2010
Muscle Activation Patterns During Walking in Children 37 Vorro J, Hobart D. Kinematic and myoelectric analysis of skill acquisition, I: 90-cm subject group. Arch Phys Med Rehabil. 1981;62:575–582. 38 Bjornson KF, Belza B, Kartin D, et al. Ambulatory physical activity performance in youth with cerebral palsy and youth who are developing typically. Phys Ther. 2007; 87:248 –257. 39 Ferdjallah M, Harris GF, Smith P, Wertsch JJ. Analysis of postural control synergies during quiet standing in healthy children and children with cerebral palsy. Clin Biomech (Bristol, Avon). 2002;17:203–210. 40 van der Heide JC, Begeer C, Fock JM, et al. Postural control during reaching in preterm children with cerebral palsy. Dev Med Child Neurol. 2004;46:253–266. 41 Hadders-Algra M, van der Fits IB, Stremmelaar EF, Touwen BC. Development of postural adjustments during reaching in infants with CP. Dev Med Child Neurol. 1999;41:766 –776.
Invited Commentary During walking at preferred speeds, young children with cerebral palsy (CP)—in comparison with children with typical development—show differences in timing of trunk and hip muscle activation, marked by excessive muscle activation during almost the entire stride cycle and with increased coactivation between the ipsilateral rectus abdominis and erector spinae muscle pair and the rectus femoris and semitendinosus muscle pair. This conclusion by Prosser and colleagues1 was based on surface electromyogphic recordings of 8 trunk, gluteal, and thigh muscles on both body sides. They hypothesize that the excessive muscle activation may create a functionally rigid trunk, limiting “the child’s ability to make fine adjustments to trunk position relative to the lower extremities and the environment and the therapist’s ability to grade muscle activity in response to external perturbations.” Similarly, the ability to control the body’s center of mass during walking will be hampered. Thus, the authors continue, the physical therapy July 2010
42 Woollacott MH, Burtner P, Jensen J, et al. Development of postural responses during standing in healthy children and children with spastic diplegia. Neurosci Biobehav Rev. 1998;22:583–589. 43 Nashner LM, Shumway-Cook A, Marin O. Stance posture control in select groups of children with cerebral palsy: deficits in sensory organization and muscular coordination. Exp Brain Res. 1983;49:393– 409. 44 Roerdink M, De Haart M, Daffertshofer A, et al. Dynamical structure of center-ofpressure trajectories in patients recovering from stroke. Exp Brain Res. 2006;174: 256 –269. 45 Hsue BJ, Miller F, Su FC. The dynamic balance of the children with cerebral palsy and typical developing during gait, part I: spatial relationship between COM and COP trajectories. Gait Posture. 2009;29: 465– 470. 46 Boyd RN, Morris ME, Graham HK. Management of upper limb dysfunction in children with cerebral palsy: a systematic review. Eur J Neurol. 2001;8(suppl 5):150 – 166.
47 Damiano DL, Alter KE, Chambers H. New clinical and research trends in lower extremity management for ambulatory children with cerebral palsy. Phys Med Rehabil Clin North Am. 2009;20:469 – 491. 48 Abt JP, Smoliga JM, Brick MJ, et al. Relationship between cycling mechanics and core stability. J Strength Cond Res. 2007; 21:1300 –1304. 49 Willson JD, Dougherty CP, Ireland ML, Davis IM. Core stability and its relationship to lower extremity function and injury. J Am Acad Orthop Surg. 2005;13:316 –325. 50 Shurtleff TL, Standeven JW, Engsberg JR. Changes in dynamic trunk/head stability and functional reach after hippotherapy. Arch Phys Med Rehabil. 2009;90:1185– 1195.
Robert C. Wagenaar
interventions should focus on the reduction of excessive trunk and hip muscle activation and the improvement of the coordination of trunk movements. Prosser and colleagues’ study of CP gait is unique in recognizing the importance of evaluating the impaired control of trunk and hip muscle activity during posture, gait, and upright movement in general. The authors emphasize that the literature on CP gait has addressed mainly the kinematics and biomechanics of the lower extremities, which is a general concern in the study of pathological gait.2 Another remarkable feature of Prosser and colleagues’ study is the inclusion of children with CP and children with TD with similar walking experience, ranging from 1 month to 5 years. According to the authors, the best evidence indicates that walking experience is a stronger predictor of walking and balance skill than age in early walkers. Walking experience was estimated by the difference between the child’s age
on the day of study and the age at onset of walking. The major concern with the outcomes and the conclusions by Prosser and colleagues is whether the excessive trunk and hip muscle activity observed in the children with CP during walking is the result of their neurological disorder or the low speed at which they prefer to walk. In addition, it is open to further investigation how the excessive trunk and hip muscle activity affects the coordination dynamics and biomechanics of the pelvic, thoracic, and trunk rotations. For example, Wagenaar and Beek3 demonstrated that systematically increasing walking speed from 0.25 to 1.50 m/s results in a change in the coordination of trunk rotation in the transversal plane from an in-phase relationship between pelvic and thoracic rotations (pelvis and thorax move in the same direction) in the lower speed range (0.25– 0.75 m/s) to an out-ofphase relationship (counter-rotation between pelvis and thorax) in the
Volume 90
Number 7
Physical Therapy f
997
