PHYS THER
Vol. 79, No. 12, December 1999, pp. 1153-1162
Head Stability in Walking in Children With Cerebral Palsy and in Children and Adults Without Neurological Impairment
Kenneth G Holt,
Robert Ratcliffe and
Suh-Fang Jeng
KG Holt, PT, PhD, is Associate Professor, Applied Kinesiology Program, Sargent College, 635 Commonwealth Ave, Boston, MA 02215 (USA) (kgholt{at}bu.edu), and Research Fellow, Center for Ecological Study for Perception and Action, University of Connecticut. Address all correspondence to Dr Holt at Sargent College
R Ratcliffe, PT, is employed in the Physical Therapy Department, Lahey Clinic, Burlington, Mass
SF Jeng, PT, ScD, is Professor, School of Physical Therapy, Medical School, National Taiwan University, Taiwan
Submitted September 14, 1998;
Accepted July 28, 1999
 |
Abstract
|
|---|
Background and Purpose. The location of several sensory systems in the head implies that maintenance of head stability may be a potentially important part of locomotor activity. A limited amount of research, however, has been conducted to measure stability or to compare head stability among different groups. The purpose of this study was to determine whether a method for measuring head stability during walking could differentiate among 3 groups: (1) children with cerebral palsy, (2) children without neurological impairment, and (3) adults without neurological impairment. Subjects. Eight adults without known neurological impairment, 6 children without known neurological impairment, and 6 children with cerebral palsy and mild spastic hemiplegia were compared. Methods. Subjects walked on a treadmill at their preferred speed at a number of frequencies. Head stability was characterized by fluctuations in period and amplitude of head motion in the sagittal plane across walking cycles. Results. Mean period fluctuation was lower for the adults than for the children, and it was lower for the children without neurological impairments than for the children with cerebral palsy. Conclusion and Discussion. The method can be used to differentiate head stability among different groups during functional activities.
Key Words: Balance • Cerebral palsy • Head stability • Shock absorption
|
Introduction
|
|---|
Enormous financial and human costs are generated as a result of injuries sustained from falling. Elderly people and children with cerebral palsy, for example, face a greater risk of falls compared with young adults with no known neurological impairment.1 Falls may be caused by many environmental hazards and are made more likely by a large number of impairments that may prevent an individual from reacting appropriately. Impairments may be classified as those due to problems in the sensory system (eg, poor vision, poor proprioception) or those due to problems in the ability of the motor system to produce the appropriate muscular actions (eg, weakness, poor timing). In spite of this broad range of environmental and organismic hazards that may cause falls, there may be common movement characteristics among different groups of fallers that distinguish them as being at risk for falling. Because the head serves as the base for the auditory, vestibular, and visual systems, which are essential for navigation through a complex environment, we believe that the quantity and quality of head movement may be connected with a person's likelihood of falling. For example, the lower extremity appears to be organized in such a way as to ensure that the shock reaching the head after heel-strike is minimized across different walking conditions.2 The lower extremity is flexibly organized so that shock is absorbed through different joints under different walking conditions. Head motion may also be sensitive to pathology in the musculoskeletal system that supports it. For example, inability of a lower-extremity muscle or muscle group to respond to an environmental perturbation such as contact with a curb is likely to have repercussions at the head.
Attempts have been made to understand the role of head stability during a number of tasks. Berthoz and Pozzo3 studied head stability during postural tasks and found that the head is oriented to vertical and the amplitude of motion is kept to a minimum. The types of postural tasks included free walking, walking in place, running in place, and hopping4 and single-leg standing on a balance beam and bilateral stance on a rocking platform.5 Studies have also been conducted on subjects with bilateral vestibular deficits.6 It was concluded from these studies that stabilization of the head to minimize its motion may be a component of a postural control system. The dynamical systems approach suggests a different approach that concentrates not on the amount of motion, but on the consistency of motion as a measure of stability.7 The purposes of our research were (1) to explore the measurement of head stability during walking using a method derived from dynamical systems theory and (2) to determine whether the measure can distinguish between people assumed to have less head stability (children with cerebral palsy) and people assumed to have more head stability (adults and children without neurological impairment). This is the first of a series of experiments to determine whether the measure is capable of predicting whether individuals are at risk for falls.
The terms "stability" and "balance" have often been used interchangeably and defined, in biomechanical terms, as maintaining the center of gravity (COG) within the base of support (BOS).8 In contrast, a measure of stability drawn from nonlinear dynamics is defined in a way that may have more functional importance for physical therapists and may be more appropriate for locomotion. Stability is simply defined as the ability of a body to resist perturbations. Inability to resist perturbations may result in a fall or at least a change in the gait pattern, factors that are critically important to the functional well-being of an individual and the physical therapist who is analyzing gait. External influences such as changes in surface friction (eg, stepping from a cement walkway onto ice) and surface geometry (eg, a curb or space in the sidewalk) may or may not lead to a stumbling or falling, depending on the person's ability to react. Inability to resist the perturbation could be due to sensory-perceptual difficulties (eg, poor vision, poor proprioception) or inconsistency or weakness in muscle force production. Therefore, the dynamic definition captures the functional outcome of the external and internal constraints on the system. In contrast, a mechanical (balance) definition of stability may not be as functionally relevant. During locomotion, a person's COG falls outside the BOS during the portion of the gait cycle that the body weight is transferred from one lower extremity to the other lower extremity.9 In the mechanical sense, the person is "unstable" during single-leg support, a fairly large portion of the gait cycle. Nevertheless, the individual appears to be stable across walking cycles in that the pattern is repeated consistently, perturbations are resisted, and a fall does not often occur.
Mechanical stability (balance) and dynamic stability are related in the sense that as the person becomes mechanically unstable (unbalanced) during single-leg support, a method of resisting the "perturbation" is necessary. In walking, this involves weight acceptance on the opposite foot, at which time the COG moves back within the BOS formed in bilateral stance. If the individual has weakness in the lower extremity accepting the weight, a gait pattern may be adopted that maintains the COG within the BOS for as much of the gait cycle as possible (eg, short stride length), reduces the forces produced by the perturbation (eg, slow speed), or changes the muscles responsible for resisting the perturbation (eg, waddling gait).
One method of dynamic systems research is to assess stability using a "relaxation time" paradigm.10 Relaxation time refers to the time it takes to recover a stable cycle following a perturbation. For example, if a person trips on a curb, stumbles, and recovers, relaxation time would refer to the time it takes a body part to return to the normal trajectory. The use of the paradigm is problematic in gait research, particularly for assessing gait stability in potential fallers. Relaxation time is dependent on where in the gait cycle (from a heel-strike to the next heel-strike of the same foot) it is imposed, but a precisely timed perturbation in gait is difficult to achieve. Furthermore, it has been our experience that many potential fallers are fearful of relatively minor perturbations imposed while they are walking on a treadmill.
An alternative to the relaxation time paradigm is to measure the variability of the trajectory of the head across locomotor cycles. In adults without neuromuscular impairments, the head assumes a near-sinusoidal trajectory while the individual walks on a smooth, level surface at constant speed.9 Internal and external influences have the potential to cause a fall and to upset the regular, sinusoidal path of the head. There are a number of methods for assessing stability through variability. Pozzo et al5 measured the standard deviation of the amplitude of head motion from the vertical during standing balance tasks on a beam or rocking bar. Two methods derived from nonlinear dynamics have also been used. One method measures variability of the trajectory of the head over repeated cycles in the phase plane (a plot of the displacement of a body segment versus its velocity). During walking, the phase plane for the head follows an elliptical curve for each cycle. Over a number of cycles, the ellipses overlap to a certain extent, but not perfectly. The resulting plot (Fig. 1) is known as a "limit cycle" and has a "bandwidth," the thickness of which is indicative of the stability of the trajectory over cycles.11,12 Essentially, the less stable a system, the greater the bandwidth, or variability around the mean, it will display. The other method, and the one used in this experiment, measures the amount of fluctuation in the period (1/frequency) and the amplitude of the sinusoidal path of a body part (Fig. 2).13 The methods capture the idea that stability is characterized by the consistency of motion rather than the magnitude of motion assumed by other investigators and that it can be assessed during functional tasks.
View larger version (39K):
[in this window]
[in a new window]
|
Figure 1. Phase plane trajectories for individuals walking at their preferred frequency. Each ellipse represents one walking cycle.
|
|
View larger version (44K):
[in this window]
[in a new window]
|
Figure 2. Time series plots for individuals walking at their preferred frequency. Hand-selected peaks and valleys for the children with cerebral palsy are marked with a vertical dashed line.
|
|
We studied head stability for 3 groups of subjects: (1) children with cerebral palsy and mild spastic hemiplegia, (2) children without known neurological impairment, and (3) adults without known neurological impairment. All subjects were tested at their preferred walking frequency and at frequencies above and below the preferred walking frequency. Frequencies other than the preferred frequency were chosen to reflect the types of challenges that might occur in everyday activities. The hypotheses were based on the findings that even though children without neurological impairments show an adult gait pattern at a relatively early age,14 limb coordination patterns continue to stabilize as the children grow older.15 In addition, children with cerebral palsy show greater variability in interlimb coordination patterns than do their age-, sex-, height-, and weight-matched peers without neurological impairments.16 We hypothesized that the adults without neurological impairments would demonstrate greater head stability at each frequency compared with the children in both groups and that the children without neurological impairments would demonstrate greater head stability than the children with cerebral palsy. We also hypothesized that preferred patterns would be more stable than other patterns, consistent with the claims for dynamic systems.7 Thus, preferred speed and frequency of walking should have greater stability than metronome-driven frequency conditions.
|
Method
|
|---|
The data were collected under 2 slightly different experimental protocols for adults and children. Differences were mainly due to the need for greater rest periods for the children between conditions and are noted in the text.
Subjects
In our study, we had 8 adult volunteers from the Boston University community (4 male and 4 female, age range=24–38 years), 6 children without neurological impairments, and 6 age- and sex-matched children with cerebral palsy. All adult subjects had no prior disability or injury. Each children's group was composed of 5 boys and 1 girl, aged 7 to 12 years. The children with cerebral palsy were recruited from the pediatric orthopedic unit of Massachusetts General Hospital and from the public schools of Massachusetts and Rhode Island. All children with cerebral palsy were functionally successful ambulators in that they usually walked without assistive devices during daily activities (except for 2 subjects who used ankle-foot orthoses during daily activities; however, these subjects did not use the orthoses during the experiment). Scores on the Ashworth scale17 ranged from 0 to 2 for the upper and lower extremities (Tab. 1), as measured by a pediatric physical therapist with 10 years of experience.
Instrumentation
To calculate head stability, we used the Peak Performance Motion Analysis System * to collect kinematic data. Reflective markers (2.5 cm in diameter) were placed at the anatomical landmark formed by the junction of the zygomatic arch and temporal bone for videotaped analysis of the head in the sagittal plane. Data were collected on the right side of the body for the adults and for the children. Although not critical, the right-side head marker was used for stability analysis to be consistent with the data collection procedure used for the adults. All subjects were filmed at 60 Hz (Panasonic model WV-D5100
) as they walked on a treadmill (Quinton model Q65
). Digitization, filtering (fourth-order Butterworth filter, cutoff frequency of 5 Hz), and calculation of vertical head displacement in the sagittal plane were completed using the automatic module of the Peak Performance 2D system (version 4.2.1*). To allow appropriate rest periods for recovery to baseline levels, an electrocardiograph was used to monitor heart rate in both groups of children.
Procedure
Preferred speed and frequency.
The task in this experiment was to walk on a treadmill at a preferred speed established prior to testing. To obtain preferred speed and frequency of treadmill walking, subjects were asked to walk at a rate that was comfortable and to direct the experimenter to increase or decrease the speed of the treadmill until a comfortable speed had been reached. The experimenter then raised or lowered the speed, and the subject was asked again to direct the experimenter to a comfortable speed. When the subject had directed the experimenter to the same speed (±3%) on 3 consecutive occasions, it was deemed that the subject had achieved the preferred mode. The treadmill speed was kept constant for all conditions. Preferred frequency was calculated by measuring the time taken for 20 strides and was expressed as period (1/frequency). The consistency of the adult subjects' speed and period was assessed in a previous study and was shown to be repeatable.18 Preferred speed and period on the treadmill for both groups of children were assessed on 2 separate days. The average difference in preferred speed between days was 0.08 m/s (approximately 6%), and the average difference in preferred period between days was 0.017 s (approximately 2%). The experimental manipulations were conducted on the second day after the preferred speed had been established.
Testing protocol.
All subjects walked at their preferred frequency and the preselected constant preferred treadmill speed. Adult subjects were required to walk in 8 conditions: at the preferred frequency, at the frequency predicted by a hybrid pendulum and spring model (for details, see Holt et al19), and at frequencies that were ±15%, ±25%, and ±35% of the preferred frequency. Both groups of children were required to walk in 6 conditions (ie, at the preferred and predicted frequencies and at frequencies that were ±15% and ±25% of the preferred frequency) due to difficulty some children experienced at ±35% of the preferred frequency. In the final analysis, comparisons between groups were made only at the preferred frequency and the ±15% and ±25% conditions. Data for the predicted frequency condition were also eliminated from the analysis because they were not different from data for the preferred frequency condition for the adults and children without neurological impairments. A metronome was used to encourage subjects to walk at the preferred frequency, and subjects were instructed to keep time to the metronome by making the heel-strike of each foot coincide with each beat. Each adult without neurological impairments walked for 8 minutes under each of the 7 conditions, which were ordered according to an 8 x 8 Latin square design. Both groups of children walked for 5 minutes in each frequency condition and rested for 3 minutes between conditions to ensure that heart rate returned to ±5% of the baseline value. Conditions were administered according to a 6 x 6 Latin square design. The adult subjects were videotaped from the right side in the sagittal plane during the last 3 minutes of each condition, and a minimum of 20 continuous stride cycles were used for analysis. Videotaped data from the right side of both groups of children in the sagittal plane were collected for a 5-minute period in each condition, and a minimum of 17 continuous strides were used for analysis. Differences in the length of videotaping reflected a longer time period required to obtain a sufficient number of strides for analysis in the children with cerebral palsy and mild spastic hemiplegia.
Data Analysis
Each time series data plot for vertical head displacement had a roughly sinusoidal shape (Fig. 2), with the head displacement for the children with cerebral palsy showing most variability. The variability of the amplitude and period of each waveform was calculated using a modified root-mean-square index13:
where 
T is the relative fluctuation magnitude (ie, the magnitude of amplitude or period fluctuations) for a particular trial, Tpv is the peak-to-valley magnitude of a particular cycle, MTpv is the mean peak-to-valley magnitude across all cycles of a given trial, Tvp is the valley-to-peak magnitude of a particular cycle, MTvp is the mean valley-to-peak magnitude across all cycles of a given trial, and N is the number of cycles in a given trial. This measure takes into account the fact that the head trajectory during the mid-stance to double support phase of gait (peak to valley) may differ from the head trajectory during the double support to mid-stance phase of gait (valley to peak).5 To determine the location and magnitude of peaks and valleys, a peak-picking computer algorithm custom-written by the second author was used. All peaks picked were marked on the time series plot and checked visually. In circumstances where the algorithm failed to accurately pick peaks, peaks were handpicked if possible or the trial was discarded. This algorithm worked well for the adults and children without neurological impairments, but head trajectories for the children with cerebral palsy were more irregular, and most peaks and valleys had to be handpicked. When a data point was handpicked, the major peak and valley for each walking cycle were chosen. Small-amplitude and high-frequency changes in trajectory direction were ignored. A trial was discarded if there appeared to be more than one major peak or valley within a walking cycle. Thus, the picking criterion was very conservative with respect to frequency variation. An example of a handpicked trial is shown in Figure 2 for the children with cerebral palsy.
A mixed analysis of variance was performed for each dependent variable (amplitude and period fluctuation), with separate between- and within-group factors. All analyses were done using a P
.05 level of significance.
|
Results
|
|---|
Between-Group Comparisons
Group comparisons indicated that there were differences among the 3 groups in both period fluctuation (F=15.27, P=.0002) and amplitude fluctuation (F= 11.82, P=.0006). Planned contrasts showed that mean period fluctuation for the adults without neurological impairments (
=0.060 second) versus the children with cerebral palsy (=0.220 second) was less at each frequency condition. Mean period fluctuation for the children without neurological impairments (=0.121 second) versus the children with cerebral palsy was also less for each frequency condition. Mean period fluctuation for the adults without neurological impairments versus the children without impairments was different only at the –15% and +25% conditions (Tab. 2, Fig. 3). Mean amplitude fluctuation for the adults without neurological impairments (=1.084 cm) versus the children with cerebral palsy (=2.264 cm) and the children without neurological impairments (=2.144 cm) was different for each frequency condition. Mean amplitude fluctuation for both groups of children was different only at the +25% frequency condition (Tab. 2, Fig. 4).
View this table:
[in this window]
[in a new window]
|
Table 2. Results for Between-Group Comparisons for Subjects Walking at Preferred Frequency and at ±15% and ±25% of Preferred Frequencya
|
|
Within-Group Comparisons
Preferred contrast analysis showed that the adults without neurological impairments demonstrated differences in amplitude fluctuation only at the –35% condition (F=43.342, P=.0001) and in period fluctuation only at the –35% condition (F=28.021, P=.0001). There were no differences between conditions in either group of children for either dependent variable.
|
Discussion
|
|---|
The measurement of stability during functional activities has been difficult to quantify, but we believe that it should be considered a critical component of physical therapy evaluation. For example, gait width (the medio-lateral distance between consecutive foot placements of opposite feet) is taught to students and is often used clinically to assess stability. The underlying assumption is that increased gait width is a compensation to enable an individual to keep the COG within the BOS. This assumption may be true in the mediolateral direction, but cannot capture the stability required for forward progression as the center of mass moves outside the BOS.
In a previous study,11 we claimed that maintaining stability of the head is an important task in preventing falls during locomotion. In that study, we showed that the head trajectory was more stable than the lower-extremity body segments during walking in adults without neurological impairments. We have also shown that the lower-limb segments show flexibility in ensuring that the shock experienced at the feet does not reach the head.2 Those results also added support to the claims of other investigators3–6 with regard to importance of minimizing head motion during stabilization of gaze during postural tasks. In this experiment, we hypothesized that the measure of head stability we used should be able to differentiate among 3 groups expected to show differences based on previously reported differences in lower-extremity limb-phasing stability.16 In particular, the measure of period fluctuation for the head showed marked decreases in stability from adults without neurological impairments to children without neurological impairments to children with cerebral palsy. Group comparisons indicated that period fluctuation for children without neurological impairments was twice as high as for adults and half that of children with cerebral palsy. The measure was particularly sensitive in distinguishing head stability between children with cerebral palsy and children without neurological impairments in all frequency conditions. The measure of amplitude fluctuation was not as sensitive in that even though there were group differences between the children with cerebral palsy and the children without neurological impairments, the differences were small, and post hoc analysis revealed that there were differences only in one frequency condition.
The reason that the period fluctuation measure shows greater sensitivity in distinguishing among groups may be related to the findings that show that frequency (1/period) acts as control factor that influences both the stability of interlimb coordination and the occurrence of transitions. Stabilization of amplitude may not be a task requirement because it does not have an important effect on the stability of, and potential to drive transitions in, the desired movement pattern. In children without neurological impairments and in children with cerebral palsy, it may be that the ability to control period may be limited, whereas the necessity to rigidly control amplitude is not as critical for any of the groups studied.
Falls are often the result of an inability to respond appropriately to perturbations. Tripping on a curb, slipping on ice, and falling up or down stairs are examples of perturbations that, if not responded to appropriately, produce falls. Factors such as weakness, poor sensory information, increased stiffness, and fear of falling may all contribute to poor recovery from perturbation. At this time, however, there is no measure to predict the potential for falling, regardless of the etiology. For example, a walking study that included assessment of 21 quantitative kinematic variables, including a mechanical measure based on gait width, was unable to discern a predictor of prospective falls.20 Our study is a very tentative first step in obtaining such a measure using a dynamic systems approach to stability, but without the necessity of actually perturbing a subject and incurring the inherent dangers of such a perturbation. The relationship between variability of the head trajectory and falls must remain largely hypothetical at this time. We have no data on the falling history of our subjects. It could be the case that there are boundaries in which a certain amount of variability is safe and that the subjects in all 3 groups fell within the boundaries.
Head trajectory may not be the control variable during walking. Scholz and Brandt21 reported that the center of mass showed less variability than the head when subjects stood from a sitting position. Because we made no estimates of the center of mass trajectory, it is not possible to make the same comparison for walking. It is possible, however, that the task constraints in going from a sitting position to a standing position are quite different than in walking. In standing from a seated position, it is mechanically important that the center of mass be placed over the feet as the action occurs. If the feet are placed too far forward or backward, it is impossible to lift the body mass without either falling or failing the task, or using a great deal of strength. The boundaries through which the center of mass can traverse are quite limited, and variability in the center of mass trajectory is probably constrained in that way. Conversely, the act of standing up is unlikely to be a highly visually dependent task. Thus, head variability, within certain limits, may be permissible within the task. In contrast, locomotion in a complex environment is highly visually dependent, whereas maintaining the center of mass within the BOS is undesirable and almost impossible to achieve under normal walking conditions. To the contrary, the act of the body falling during single-leg stance onto the opposite limb may serve to provide pendular and elastic energy conservation between limbs. The reality is that, in both tasks, both constraints probably have more importance during certain phases than during other phases and that, in general, rising from a sitting position to a standing position successfully is more mechanically than perceptually dependent.
The next step in experimentation is to show that the method can actually differentiate between individuals with a known history of falls and individuals without a known history of falls. We are planning a study of elderly people without neurological impairments and known elderly fallers. If this method is capable of distinguishing between the 2 groups, the clinical implications would be important. For example, the method could potentially be used as a screening tool for people who are at risk for falling and as an evaluation tool to determine the success of treatments that address the causes of falling.
|
Conclusion
|
|---|
Generally, the results of our study support the claim that preferred states of dynamic systems will demonstrate greatest stability. We used a curve fitting procedure to investigate trends across conditions. Inverted U-shaped curves were observed for both period and amplitude fluctuations, with the least variability at or close to preferred conditions. Regression values (r2) for the adults without neurological impairments, the children without neurological impairment, and the children with cerebral palsy were .814, .438, and .753, respectively, for period fluctuation and .900, .973, and .646, respectively, for amplitude fluctuation. These results are consistent with those of our previous work.11, 18 This finding suggests that, although (from a clinical perspective) we may want to improve locomotor speed, there is no need (from a stability perspective) to attempt to manipulate stride length or frequency independently. Individuals, regardless of disease, will hone in on the optimal configuration of stride length and frequency for them. Nevertheless, this suggestion should be taken cautiously. Although there were differences between some conditions, the actual amount of difference in variability between conditions was quite small (Figs. 3 and 4), particularly for the adults without neurological impairments, and brings to question the functional importance of the finding.
|
Footnotes
|
|---|
This study was approved by the Charles River Institutional Review Board, Boston University.
* Peak Performance Technologies Inc, 7388 S Revere Pkwy, Ste 603, Englewood, CO 80112. 
Panasonic Co, 1 Panasonic Way, Secaucus, NJ 07094.
Quinton Instrument Co, 2121 Terry Ave, Seattle, WA 98121.
|
References
|
|---|
- National Safety Council Accident Facts. Chicago, Ill: National Safety Council,1988
.
- Ratcliffe RJ, Holt KG. Low frequency shock absorption in human walking.
Gait & Posture.1997; 5:93–100.
- Berthoz A, Pozzo T. Intermittent head stabilization during postural and locomotory tasks in humans. In: Amblard B, Berthoz A, Clarac F, eds.
Posture and Gait: Development, Adaptation, and Modulation. New York, NY: Elsevier Science Inc,1988
.
- Pozzo T, Berthoz A, Lefort L. Head stabilization during various locomotor tasks in humans, I: normal subjects.
Exp Brain Res.1990; 82:97–106.[Web of Science][Medline]
- Pozzo T, Levik Y, Berthoz A. Head and trunk movements in the frontal plane during complex dynamic equilibrium tasks in humans.
Exp Brain Res.1995; 106:327–328.[Web of Science][Medline]
- Pozzo T, Berthoz A, Vitte E, Lefort L. Head stabilization during locomotion: perturbations induced by vestibular disorders.
Acta Otolaryngol Suppl (Stockh).1991; 481:322–327.
- Haken H, Kelso JAS, Bunz H. A theoretical model of phase transitions in human hand movements.
Biol Cybern.1985; 51:347–356.[Web of Science][Medline]
- Perry J.
Gait Analysis: Normal and Pathological Function. Thorofare, NJ: Slack Inc:1993
.
- Inman VT, Ralston HJ, Todd F.
Human Walking. Baltimore, Md: Williams & Wilkins,1979
.
- Scholz JP, Kelso JAS. A quantitative approach to understanding the formation and change of coordinated movement patterns.
Journal of Motor Behavior.1989; 21:122–124.[Web of Science][Medline]
- Holt KG, Jeng SF, Ratcliffe R, Hamill J. Energetic cost and stability in preferred human walking.
Journal of Motor Behavior.1995; 27:164–179.[Web of Science][Medline]
- Kay BA, Saltzman EL, Kelso JAS. Steady-state and perturbed rhythmical movements: a dynamical analysis.
J Exp Psychol Hum Percept Perform.1991; 17:183–197.[Web of Science][Medline]
- Rosenblum LO, Turvey MT. Maintenance tendency in co-ordinated rhythmic movements: relative fluctuations and phase.
Neuroscience.1988; 27:289–300.[Web of Science][Medline]
- Sutherland DH, Olshen R, Cooper L, Woo SL. The development of mature gait.
J Bone Joint Surg Am.1980; 62:336–353.[Free Full Text]
- Jeng SF, Liao HF, Lai JS, Hou JW. Optimization of walking in children.
Med Sci Sports Exerc.1996; 29:370–376.
- Jeng SF, Holt KG, Fetters L, Certo C. Self-optimization of walking in non-disabled children and children with spastic hemiplegic cerebral palsy.
Journal of Motor Behavior.1996; 28:15–27.[Web of Science][Medline]
- Bohannon RW, Smith MB. Interrater reliability of a modified Ashworth scale of muscle spasticity.
Phys Ther.1987; 67:206–207.[Abstract/Free Full Text]
- Holt KG, Hamill J, Andres RO. Predicting the minimal energy costs of human walking.
Med Sci Sports Exerc.1991; 23:491–498.[Web of Science][Medline]
- Holt KG, Hamill J, Andres RO. The force-driven harmonic oscillator as a model for human locomotion.
Human Movement Sciences.1990; 9:55–68.
- Feltner ME, MacRae PG, McNitt-Gray JL. Quantitative gait assessment as a predictor of prospective and retrospective falls in community-dwelling older women.
Arch Phys Med Rehabil.1994; 75:447–453.[Web of Science][Medline]
- Scholz JP, Brandt LC. Trajectory formation and development of the control of standing up from sitting.
Motor Control.1997; 1:314–339.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
|
|
 
J. J. Kavanagh, S. Morrison, and R. S. Barrett
Lumbar and cervical erector spinae fatigue elicit compensatory postural responses to assist in maintaining head stability during walking
J Appl Physiol,
October 1, 2006;
101(4):
1118 - 1126.
[Abstract]
[Full Text]
[PDF]
|
|
|
Copyright © 1999 by the American Physical Therapy Association.
Top
Introduction
Method
