Postural Synergies Associated With a Stepping Task.

Recognizing the computational complexity involved in controlling the redundant degrees of freedom of the human body, Bernstein[1] suggested that muscles are constrained to act together in functional synergies. These synergies are believed to decrease the number of degrees of freedom that must be specified to achieve a position in space, thereby reducing the nervous system's computational burden. Most motor acts involve many different muscles, including those necessary to produce the desired joint displacements (focal movement) and those that act to maintain balance and posture during the movement. The question is whether there are identifiable synergies, reflecting central nervous system (CNS) control in the presence of particular constraints, for specific motor tasks.

The usefulness of the concept of muscle synergies has been hotly debated in the postural control literature[2,3] Some researchers[4-9] have argued for the existence of a repertoire of fixed, neurally based synergies for postural control. Other researchers[10-13] have suggested that synergies are coordinated patterns of muscle activation that are subject to various modulating influences and can be combined in various ways. Still others have concluded that the number of patterns of muscle activation used by the CNS is essentially unlimited and that the appearance of discrete synergies results from experimental, not neural, constraints.[3]

Although first described in detail in relation to the muscle activation patterns observed in response to external perturbations,[4,14] the synergy concept also is applicable to movements produced by voluntary effort. Researchers have described reproducible patterns of muscle activation associated with numerous activities performed in a standing position, including rapid elbow flexion and extension,[15] rapid arm raises,[10,12,16-19] pushes and pulls on a fixed handle,[5] unilateral lower-extremity movements,[20-22] and initiation of gait.[23-28] The usual finding is that a sequence of 2 or more postural muscles is activated in a feedforward manner, well in advance of focal muscle activation. For upper-extremity movements, the anticipatory postural muscle activity serves to counteract destabilizing forces generated by the forthcoming focal movement and minimize displacement of the body center of mass (COM).[29,30] For lower-extremity movements and initiation of gait, the postural activity produces controlled displacement of the COM over a changing base of support.[22,24,26]

Stepping presents a major challenge for the postural control system because the movement involves a limb that has been supporting a portion of the body's weight. The ability to maintain equilibrium while stepping or performing various unilateral lower-extremity movements in a standing position is essential for many functional activities. Knowing the extent to which postural muscles are recruited consistently in advance of these activities may be important for rehabilitation. If consistent synergy patterns associated with postural preparation for stepping activities can be identified, physical therapists may be able to use this information in evaluation and intervention for individuals with balance or gait disorders.

Although several investigators[24,26-28] have commented on the preprogrammed, stereotypic nature of the initiation of gait, few have examined the consistency of the order of postural muscle activation. Crenna and Frigo[26] reported a reproducible pattern involving inhibition of tonic soleus muscle activity followed by activation of the tibialis anterior (TA) muscle bilaterally. They considered this temporally invariant relationship to represent a motor program that may be used during tasks such as forward bending and rising up on the toes as well as during initiation of gait. Rogers and Pai[21] observed phasic activation of bilateral hip abductor and adductor musculature in consistent temporal relationships prior to both unilateral lower-extremity flexion during standing and forward-stepping activities. Only 6 subjects were included in each of these studies, however, and the proposed synergies were composed of only pairs of postural muscles.

Other researchers have noted considerable variability in performance of unilateral lower-extremity movements during standing. Beuter et al[31] analyzed the activity of the rectus femoris (RF) and biceps femoris muscles associated with stepping motions performed from quiet standing over obstacles of different heights. Although they did not examine electromyographic (EMG) activity prior to movement onset, they reported that the timing of EMG peaks during the stepping movement was highly variable across trials and conditions. Mouchnino et al[20] stated that the set of muscles activated in advance of unilateral hip abduction during standing varied among subjects.

Our study was undertaken as part of a broader investigation of age group differences in postural preparation for movement.[32] The purpose of this study was to examine the extent to which subjects without known impairment of the neuromuscular system in 3 age groups (children, young adults, and older adults) exhibit fixed sequences of postural muscle activation during performance of a stepping task. Specific research questions were: (1) Can a consistent postural synergy be identified for each subject?, (2) What sequences of postural muscle activation are most commonly observed in each age group?, (3) What is the effect of a change in task constraints on patterns of postural muscle activation?, and (4) Do age groups differ in the number of different sequences of postural muscle activation observed?

We selected a stepping task, a multijoint lower-extremity movement task requiring controlled movement of the COM over a changing base of support. We incorporated 2 movement contexts, "place" and "step," that involved different task constraints. In "place," the subject lifted one foot and placed it on the step. In "step," the subject lifted the foot, placed it on the step, and stepped up onto the step (bringing the other foot up).

We expected that subjects would demonstrate considerable variability in the order of onset of EMG activity during initiation of our stepping task, despite careful control of environmental and task conditions. We also anticipated that the difference in task constraints between "place" and "step," particularly the need to generate greater momentum during "step," would have an effect on postural synergies. Based on previous reports[33,34] suggesting that developmental changes in posture and movement coordination may continue well beyond 7 or 8 years of age and, on the assumption that the children had had less experience with the task than the other 2 age groups, we expected that the children would display less consistency within and among subjects.

Method

Subjects

Twenty subjects (10 male, 10 female) in each of 3 age groups--children (mean age=10.0 years, SD=1.3, range=8-12), young adults (mean age=30.4 years, SD=3.3, range=25-35), and older adults (mean age=68.3, SD=3.1, range=65-73)--participated in the study. Volunteers were accepted as subjects provided that they (1) were in good health, (2) were of an appropriate age (8-12 years, 25-35 years, or 65-75 years), (3) could ascend and descend stairs independently without upper-extremity support, (4) reported a preference for using the right foot for kicking a ball, (5) indicated on a screening questionnaire that they had no neurological or orthopedic problems (impairments) that might interfere with balance, and (6) were not taking any response-altering drugs or medications by self-report. The criterion related to foot preference was needed because our analysis included only right lower-extremity movement trials, and we speculated that subjects with a right lower-extremity preference might perform "place" and "step" with the right lower extremity differently than subjects with a left lower-extremity preference. Informed consent was obtained from each subject prior to testing. For the testing of children, consent from a parent or legal guardian also was obtained.

Testing Procedures

Each subject stood barefoot on a sheet of paper in a standardized location on a force platform, with the medial malleoli approximately 10 cm apart. Stance width was standardized because of its effect on center-of-pressure measurements, with the 10-cm distance chosen on the basis of previous investigations of standing balance in adults.[35-37] Because a stance width of 6 cm was used previously to study postural control in children between 2 and 14 years of age,[38] 10 cm was considered to provide adequate mediolateral stability for the 8- to 12-year-old children in our study. We made tracings of both of the subjects' feet to ensure consistent foot position during testing.

A wooden step, 20-cm high x 70-cm wide x 56-cm deep, was placed in front of the subject with the front edge 1.75 "foot-lengths" from the back of the subject's heels. This procedure adjusted for between-subject differences in length of the base of support that might have altered the biomechanical demands of the task. A pressure switch was taped to the subject's right heel to permit determination of movement onset and offset. Release of pressure on the switch resulted in opening of the circuit at the start of the movement, and contact of the foot with the step closed the circuit at the end of the movement.

Silver-silver chloride bipolar surface electrodes with onsite preamplifiers (gain x35) were placed over the TA, gastrocnemius-soleus (GS), hamstring (HS), and gluteus medius (GM) muscles of the left (stance) limb and the RF and GM muscles of the right (moving) limb. These muscles were selected because of their phasic modulation in advance of voluntary movement during pilot testing and as reported by other investigators[21,22,24,27,28] for similar lower-extremity tasks. Because pilot testing of 4 subjects indicated that the right TA was activated nearly synchronously with the left TA, only the left TA was monitored. The left quadriceps femoris muscle was not monitored because all 4 subjects in the pilot study demonstrated low-amplitude tonic activity in this muscle and in the single-joint knee extensors on the right during the interval of interest, with no identifiable bursts of muscle activity associated with initiation of the focal movement. The activity we observed in the quadriceps femoris muscle was similar to that described by Mouchnino et al.[20] The right RF, however, demonstrated close temporal correlation with movement onset and was selected as the focal muscle for the task.

A visual display consisting of 2 red light-emitting diodes (LEDs) mounted horizontally on a black background was placed approximately 1.2 m in front of and facing the subject at eye level. Illumination of the right or left LED indicated to the subject that initial movement should occur with the right or left lower extremity, respectively. We used a choice reaction time paradigm to ensure that subjects could not complete postural preparations for the task in advance of the data collection interval. A 500-Hz tone lasting 500 milliseconds was used as a warning signal. After a randomly varied delay of 1.5, 2, or 2.5 seconds, either the right or left LED was illuminated as the response signal. The timing of illumination of the LEDs was controlled by a BASIC program on a personal computer.

Subjects were instructed to move as rapidly as possible after the response signal, thereby attempting to minimize both reaction time and movement time. During right lower-extremity test trials, signals from the force platform, EMG device, and pressure switch were collected for a 2-second interval, beginning at the response signal (external trigger). To ensure that appropriate levels of electrical activity were recorded, EMG signals were monitored on an oscilloscope, and amplifier gain on all 6 channels was adjusted as needed. Electromyographic signals were amplified with an adjustable gain from 500 to 10,000, a common-mode rejection ratio of 87 dB at 60 Hz, and a frequency response of 40 to 4,000 Hz. The signals were high-pass filtered at 40 Hz and root-mean-square (RMS) processed with a time constant of 2.5 milliseconds. The manufacturer of the amplifier and processor module (EMG-67 amplifier(*)) provided an option of either 40 Hz or 75 Hz as the low cutoff frequency to maximize signal-to-noise ratios and minimize cable artifact. The 40-Hz cutoff was selected to avoid filtering major components of the EMG signals. A BEDAS-2 data acquisition and analysis package, ([dagger]) a Data Translation analog-to-digital converter, ([double dagger]) and a microcomputer system were used to sample EMG, force platform, and pressure switch signals online at 500 Hz. The data were stored on a disk for later analysis.

Subjects performed a block of 6 practice trials and 20 test trials in each of the 2 movement contexts (ie, "place" and "step"). The principal investigator (VSM) repeated the instructions prior to each block of practice and test trials. The order of presentation of "place" and "step" blocks was counterbalanced within each age group. Within each block, right and left lower-extremity trials were presented in pseudorandom order, so that 3 practice trials and 10 test trials were performed with each lower-extremity leading. There was an 8- to 10-second delay between trials for repositioning of the subject and storage of data files. Between blocks, subjects were given a 10-minute rest period. The entire testing procedure lasted approximately 1 hour.

Data Analysis

Right lower-extremity test trials were selected for analysis on the basis of movement times and patterns of center-of-pressure displacement. Movement times were determined from the pressure switch signals. Although within-subject variability in movement times was quite low, with within-subject standard deviations generally in the range of 40 to 80 milliseconds, the possibility of variations in the order of onset of EMG activity related to movement time differences was a concern. The 8 right lower-extremity test trials (4 out of the 10 for each "place" and "step" context) with movement times closest to the subject's mean movement time were selected as most representative of the subject's typical performance. In addition, the pattern of center-of-pressure displacement was examined for indications that the subject initially prepared to step with the left lower extremity rather than the right lower extremity. Trials in which this pattern occurred were excluded from the analysis. These trials were rare ([is less than] 10% of total test trials), posing no difficulties for identification of 8 acceptable test trials for each subject. A repeated-measures analysis of variance (ANOVA) with age group as a between-subjects factor and movement context as a within-subjects factor was used to identify any differences in mean movement times among age groups.

DATA-PAC II software ([sections]) was used for analysis of EMG data. All onsets of EMG activity were determined by the principal investigator. A muscle was considered "on" when the EMG tracing exceeded the established threshold (set at approximately 150% of baseline noise) for a period of 50 milliseconds or longer.[39] Baseline noise was determined by visual inspection of the EMG signal during quiet periods at the start of each trial. Onset latencies were calculated relative to the beginning of the data file for each trial, which coincided with presentation of the response signal. Based on measurement error associated with visual identification of onsets of bursts of muscle activity by the investigator, muscles activated within 15 milliseconds of each other were considered to have synchronous onsets.

Sequences of postural muscle activation were determined on the basis of EMG activity onset latencies. Postural muscles were operationally defined as muscles activated prior to the onset of EMG activity in the focal muscle (ie, the right RF). A preferred postural synergy was defined as the sequence observed during at least 3 of the 4 trials in each movement context for each subject. A Kruskal-Wallis test was used to identify differences among age groups in the number of different synergies observed. The percentage of subjects in each age group displaying the same preferred synergy in both movement contexts also was calculated.

Reliability of identification of onsets of bursts of muscle activity by the principal investigator was assessed during the course of the study. Data files containing a sample of 87 RMS-processed EMG recordings that had been analyzed following completion of data collection for each subject were randomly selected, arranged in random order, and given a coded file name by a research assistant. The principal investigator, who was not aware of the coded file names, then reanalyzed these data files to determine EMG activity onset latencies. Using procedures described by Shrout and Fleiss,[40] we obtained an intraclass correlation coefficient (ICC[3,1]) of .99 for intrarater reliability.

Results

Mean movement times did not differ among age groups (F=1.44; df=2,57; P=.245). The effect of movement context was significant (F= 16.47; df= 1,57; P [is less than] .001), but the group by context interaction was not (F=0.60; df=2,57; P=.551). Mean ([+ or -] SD) movement times were 552 [+ or -] 01 milliseconds for "place" and 525 [+ or -] 75 milliseconds for "step" in the group of children, 576 [+ or -] 118 milliseconds for "place" and 528 [+ or -] 96 milliseconds for "step" in the group of young adults, and 608 [+ or -] 139 milliseconds for "place" and 578 [+ or -] 121 milliseconds for "step" in the group of older adults. Across age groups, mean ([+ or -] SD) movement times were 579 [+ or -] 121 milliseconds for "place" and 544 [+ or -] 101 milliseconds for "step." Only 2 errors, in which the subject actually stepped with the left lower extremity, occurred on right lower-extremity test trials. These error trials, both by children, were excluded from the data analysis.

Mean onset latencies for all 6 muscles (5 postural and 1 focal) monitored in this study are displayed in Figure 1. Four postural muscles in the stance (left) lower extremity (TA, GS, HS, and GM) were included in the analysis of the sequences of postural muscle activation. Although activity of the GM of the moving (right) lower extremity also was monitored, EMG amplitudes of this muscle were too low in approximately 50% of the subjects to permit reliable identification of onsets of bursts of muscle activity. In those subjects who did exhibit phasic activity of the right GM in a majority of trials, this activity tended to occur synchronously with activation of the left TA for "place" and slightly later than left TA activation for "step."

[Figure 1 ILLUSTRATION OMITTED]

With regard to the first research question concerning within-subject consistency in sequences of postural muscle activation, a preferred postural synergy (consistency in at least 3 out of 4 trials) could be identified in only 30 subjects for "place" and 47 subjects for "step." The number of subjects in each age group displaying a preferred synergy is graphically represented in Figure 2.

[Figure 2 ILLUSTRATION OMITTED]

The second research question dealt with commonly observed sequences of muscle activation. An example of the EMG record for a young adult subject during "step" is displayed in Figure 3. The left TA typically was the first of the 4 postural muscles to be activated, regardless of movement context. Of the 480 total trials analyzed (4 "place" and 4 "step" for each of the 60 subjects), the TA was activated first in 447 trials (93%). Mean onset latency of the TA was 272 [+ or -] 113 milliseconds for "place" and 229 [+ or -] 80 milliseconds for "step."

[Figure 3 ILLUSTRATION OMITTED]

The sequence of muscles activated subsequent to the TA, but before focal movement onset, was more variable. Across subjects, a total of 14 different preferred synergies were observed for the 2 movement contexts. The number of subjects in each age group displaying specific preferred synergies during "place" and "step" is presented in Figure 4. The most common synergy for both "place" and "step" was TA followed by GM, with no other postural muscles active before movement onset (denoted by "TA-GM"). This sequence was seen in 9 subjects for "place" and in 27 subjects for "step." Another relatively common sequence, demonstrated by 4 subjects for "place" and 13 subjects for "step," was TA-HS-GM.

[Figure 4 ILLUSTRATION OMITTED]

The general pattern of initial activation of TA followed by later activation of GM, with or without intervening activity in other postural muscles, was exhibited by 48 subjects for "place" and 57 subjects for "step." Only 2 subjects, both older adults, had preferred synergies in which the TA was not the first muscle activated. These sequences, which occurred during "place," were GS-GM-TA for one subject (Fig. 5) and HS-TA-GM for the other subject. The pattern of displacement of the center of pressure for these 2 subjects was similar to that of other subjects in their age group.

[Figure 5 ILLUSTRATION OMITTED]

The third research question addressed the effects of task constraints, as represented by the 2 movement contexts, on patterns of muscle activation. The number of subjects having a preferred synergy was larger, and the number of different synergies observed was smaller for "step" than for "place." Only 9 of the 60 subjects (6 children and 3 older adults) used the same preferred synergy for both "place" and "step."

Age group differences in the number of different sequences of postural muscle activation were the focus of the fourth research question. The 3 age groups differed more in the number of different preferred synergies observed for "place" than for "step." For "place," the number of different preferred synergies was 2 for children, 8 for young adults, and 9 for older adults. For "step," children demonstrated 3 different preferred synergies, whereas the young and older adult groups each displayed 4 different preferred synergies. The Kruskal-Wallis procedure, the nonparametric equivalent of the one-way ANOVA, was used to test for differences among the age groups in the number of different preferred synergies observed. The test statistic was not significant (H=3.60, P=.165, assuming a chi-square distribution with 2 degrees of freedom).

Discussion

Subjects displayed considerable variability from one trial to the next in specific sequences of postural muscle activation during performance of a stepping task, particularly when instructed to place one foot on the step rather than to step up onto the step. Although almost all subjects demonstrated early TA activation and most subjects had 1 or 2 preferred patterns, a substantial percentage did not exhibit the same sequence on more than 2 of the 4 trials analyzed in each context. In addition, preferred synergies differed substantially from one subject to the next, both within and among age groups.

This high variability occurred despite our attempts to carefully control experimental conditions. Physical layout of the room, lighting conditions, instructions to the subjects, and number of practice trials remained constant. Movement times were controlled by instructing subjects to move as fast as possible and by selecting trials closest to each subject's mean movement time for inclusion in the analysis. A minimal number of muscle groups, identified as being important in postural preparations for stepping tasks by previous researchers[21,22,24,27,28] or on the basis of pilot work, were included in the search for clearly defined synergies. The subjects' initial position, although not as strictly controlled, was characterized by approximately symmetrical lower-extremity weight bearing, as indicated by the center-of-pressure data[32] and by consistency of foot position, as ensured by the use of foot tracings.

These results provide support for the idea that the neuromotor system can generate a variety of solutions for a given motor task.[3,41-43] The muscle sequence observed on a particular trial may have been affected by subtle changes in the subject's postural alignment, afferent inputs, perceived stability, or any of a host of different factors that somehow made that sequence most efficient for the individual at that time. The timing of the illumination of the response signal with respect to the periodic anterior-posterior sway that normally occurs in quiet standing also may have affected the sequencing of muscle activation. As Reed[44] has argued, the term "motor command" may be a misnomer. Central signals to muscles appear to influence, but do not determine, action.

Although use of a 2-choice paradigm rather than a simple reaction-time paradigm may have resulted in greater variability, we needed to ensure that we were able to detect preparatory postural muscle activity. If subjects had known in advance which lower extremity they were to move, then they might have performed some postural preparation prior to illumination of the response signal. We attempted to minimize any effects of the choice reaction-time paradigm on EMG activity onset order by discarding trials in which the pattern of center-of-pressure displacement suggested that the subject may have begun to initiate a step with the left lower extremity. Other researchers[5,8] who have used a choice reaction-time paradigm to investigate sequences of postural muscle activation accompanying rapid voluntary movements have reported that the relative temporal sequencing is comparable to that observed in simple reaction-time conditions.

Among those subjects in all age groups who exhibited preferred synergies, the majority displayed synergies involving a distal-to-proximal order of activation. The most commonly observed synergy, involving early activation of left TA (and probably the right TA as well, as indicated by pilot testing) followed by activation of left GM, may be understood in terms of biomechanical constraints on the degrees of freedom. Early TA activation may serve to bring the COM forward, as during the initiation of gait.[23-28] The fact that the TA was the first muscle activated on 93% of the trials is indicative of the central role of this muscle group in generating the necessary forward momentum for performance of the stepping task. The timing of this early activity is consistent with the initial backward shift of the center of pressure.

Although the latency of the TA activation relative to the response signal was longer in this study than the 163 to 173 milliseconds reported by Burleigh et al[25] for step initiation, this difference can be explained by the use of a visual cue in a choice reaction-time paradigm in our study and a proprioceptive cue in a simple reaction-time paradigm in their study. In addition, their subjects were all young adults, who are known to have shorter reaction times than children or older adults.[32,45,46] The mean time of TA activation reported by Elble et al[27] for gait initiation by older adults (aged 64-82 years) was 222 [+ or -] 80 milliseconds, closer to the values obtained in our study.

In addition to the early activation of the TA bilaterally, Elble et al[27] described subsequent activation of the HS and GS of the stance leg in preparation for the swing foot leaving the ground. In our study, some subjects demonstrated activation of either left HS or GS as part of a preferred synergy, but activation of both muscles in a synergy was seen in only 5 subjects and only during "place." However, we often observed phasic recruitment of the right GM at about the same time as initial TA activation, as reported by other researchers.[21,22,28] This GM activity may contribute to the displacement of the center of pressure toward the flexing (right) limb. Although the center of pressure moves to the right, the body's COM is displaced directly toward the stance (left) limb, as shown by Rogers and Pai[22] for a similar single-leg flexion task and Elble et al[27] for initiation of gait. Subsequent activation of the left GM acts to stabilize the pelvis and to control displacement of the COM to the left, over the stance limb.[21,28]

The relative contributions of biomechanical and neuromotor control variables to the observed differences in muscle sequences in the 2 movement contexts are difficult to determine. Only 9 subjects used the same preferred synergy for both "place" and "step." A switch from a preferred synergy involving 3 or more muscles for "place" to a synergy involving only 2 muscles for "step" was a fairly frequent occurrence among young and older adults. The observed differences may be clarified by examination of mean onset latencies (Fig. 1). Activation of left HS and GS tended to occur later for "step" than for "place." Because preferred synergies were determined on the basis of muscles activated prior to movement onset, these 2 muscles were less likely to be identified as part of a preferred synergy during "step."

One plausible explanation for the greater variability in sequences of muscle activation used during "place" is that the impulses (forces acting over time) are smaller than for "step" and therefore require less rigid programming of postural adjustments. Movement times were slower for "place" than for "step." One of the more robust findings in the literature on postural adjustments is that the timing and sequencing of postural muscle activity are more variable for slower or less forceful movements.[12,19,22,47] Another possibility is that the equilibrium requirements for "place" at the time of right foot contact with the step, involving complete deceleration of the right lower extremity and stabilization of the COM position over a new base of support, produced more complex postural muscle activity. Finally, subjects may not have had sufficient experience with "place" to establish or refine their motor patterns to the same degree as for "step."

Contrary to our expectations, the number of different preferred synergies tended to be smaller for children than for adults during "place" and were similar for the 2 groups during "step." Perhaps lack of experience with a task sometimes results in use of simpler, more consistent synergies rather than greater variability. If "place" was a less familiar and less practiced task for the children than "step," then they may have used a strategy designed to simplify the task as much as possible by minimizing the number of joints that were controlled.[1,33] Although adults, as a group, explored a number of different synergies, the children may have used a simplifying strategy involving activation of the minimal number of postural muscles necessary to meet the biomechanical demands of the task. During "step," however, both children and adults may have used the strategy that had proven most efficient in their extensive experience with the task.

This study had several limitations. We could not record from more than 6 muscles simultaneously. Activity in other muscles, particularly bilateral hip adductors and right GS and HS, may have played an important role in postural preparation for the stepping task. In addition, some subjects may have used muscles other than the right RF to initiate the stepping movement, as suggested by occasional activation of the right RF subsequent to the release of the pressure switch. Another limitation was the determination of movement onset and offset by use of the pressure switch under the subject's heel. Although subjects were instructed to land with the entire right foot on the step for both "place" and "step," they may have contacted the step with the forefoot at landing or maintained the contact of the forefoot with the ground after release of the pressure switch during movement initiation. This potential inconsistency in foot contact may have introduced error variability in the movement time measurements. The possibility also exists that our decision to analyze only those test trials with movement times closest to the subject's mean movement time substantially affected the results. We believe that this procedure yielded results representative of typical performance, but different results might have been obtained with analysis of the fastest trials. A final limitation of the study was the relatively small number of subjects in each age group, which decreased statistical power.

Despite these limitations, we believe that the order of muscle activation in preparation for our stepping task was not fixed. Specific sequences varied considerably within and among individuals. Biomechanical constraints related to the need to generate and control momentum of the COM in the sagittal and frontal planes, however, produced a general pattern of activation that was displayed by a majority of the subjects. This pattern was characterized by early activation of bilateral TA and right GM, with subsequent activation of left GM. The few subjects who did not use this pattern were still able to perform the task successfully, apparently using alternative muscle patterns to meet the biomechanical demands.

If sequences of postural muscle activation vary substantially during performance of a functional task by individuals without disabilities under relatively controlled laboratory conditions, then how useful is the synergy concept for evaluation and intervention in individuals with balance disorders? Many more variables are likely to affect the specific postural synergies displayed during a given task in the clinic than in the laboratory. We suggest that the postural accompaniments to discrete multijoint voluntary movements may be too variable for clinical usefulness. Perhaps intervention should focus not on improving the patient's execution of particular synergies, but on ameliorating underlying impairments and providing the patient with opportunities to practice maintenance of balance under a variety of environmental and functional task conditions. Because even relatively subtle changes in task constraints can produce marked changes in motor patterns, therapists should not expect improvements in one set of conditions to transfer to another.

Conclusions

We found considerable variability both within and among subjects in specific sequences of postural muscle activation during performance of a stepping task. Subjects in all 3 age groups were able to meet the biomechanical demands of the task in a variety of different ways, with the specific order of muscle activation modified in accordance with changes in the individual, the task, or the environment. Although a general pattern of early TA activation followed by left GM activation was observed in the majority of trials across different age groups and movement contexts, subjects who did not exhibit this pattern were still able to execute the task successfully. Physical therapists should recognize that a multitude of factors may affect the specific sequence of muscle activation exhibited during task performance. Consequently, patients need opportunities to practice tasks under the entire range of conditions they are likely to experience in their daily lives. Additional research is needed to determine whether any invariant characteristics of postural muscle activation, essential for performance of particular tasks, can be identified.

(*) Therapeutics Unlimited Inc, 2835 Friendship St, Iowa City, IA 52240.

([dagger]) Advanced Mechanical Technology Inc, 151 California St, Newton, MA 02158.

([double dagger]) Data Translation Inc, 100 Locke Dr, Marlborough, MA 01752.

([sections]) RUN Technologies Inc, 25622 Rolling Hills Rd, Laguna Hills, CA 92653.

References

[1] Bernstein N. Coordination and Regulation of Movements. New York, NY: Pergamon Press; 1967.

[2] Lee WA. Neuromotor synergies as a basis for coordinated intentional action. Journal of Motor Behavior. 1984;16:135-170.

[3] Macpherson JM. How flexible are muscle synergies? In: Humphrey DR, Freund HJ, eds. Motor Control: Concepts and Issues. New York, NY: John Wiley & Sons Inc; 1991:33-47.

[4] Nashner LM. Adapting reflexes controlling the human posture. Exp Brain Res. 1976;26:59-72.

[5] Cordo PJ, Nashner LM. Properties of postural adjustments associated with rapid arm movements. J Neurophysiol. 1982;47:287-302.

[6] Nashner LM, McCollum G. The organization of human postural movements: a formal basis and experimental synthesis. Behav Brain Sci. 1985;8:135-172.

[7] Horak FB, Nashner LM. Central programming of postural movements: adaptation to altered support-surface configurations. J Neurophysiol. 1986;55:1369-1381.

[8] Nardone A, Schieppati M. Postural adjustments associated with voluntary contraction of leg muscles in standing man. Exp Brain Res. 1988;69:469-480.

[9] Do MC, Nouillot P, Bouisset S. Is balance or posture at the end of a voluntary movement programmed? Neurosci Lett. 1991;130:9-11.

[10] Lee WA. Anticipatory control of postural and task muscles during rapid arm flexion. Journal of Motor Behavior. 1980;12:185-196.

[11] Nashner LM, Forssberg H. Phase-dependent organization of postural adjustments associated with arm movements while walking. J Neurophysiol. 1986;55:1382-1394.

[12] Lee WA, Buchanan TS, Rogers MW. Effects of arm acceleration and behavioral conditions on the organization of postural adjustments during arm flexion. Exp Brain Res. 1987;66:257-270.

[13] Hirschfeld H, Forssberg H. Phase-dependent modulations of anticipatory postural activity during human locomotion. J Neurophysiol. 1991;66:12-19.

[14] Nashner LM. Fixed patterns of rapid postural responses among leg muscles during stance. Exp Brain Res. 1977;30:13-24.

[15] Friedli WG, Hallett M, Simon SR. Postural adjustments associated with rapid vohmtary arm movements, I: electromyographic data. J Neurol Neurosurg Psychiatry. 1984;47:611-622.

[16] Aruin AS, Latash ML. Directional specificity of postural muscles in feed-forward postural reactions during fast voluntary arm movements. Exp Brain Res. 1995;103:323-332.

[17] Badke MB, Di Fabio RP. Effects of postural bias during support surface displacements and rapid arm movements. Phys Ther. 1985;65: 1490-1495.

[18] Bouisset S, Zattara M. A sequence of postural movements precedes voluntary movement. Neurosci Lett. 1981;22:263-270.

[19] Horak FB, Esselman P, Anderson ME, Lynch MK. The effects of movement velocity, mass displaced, and task certainty on associated postural adjustments made by normal and hemiplegic individuals. J Neurol Neurosurg Psychiatry. 1984;47:1020-1028.

[20] Mouchnino L, Aurenty R, Massion J, Pedotti A. Coordination between equilibrium and head-trunk orientation during leg movement: a new strategy built up by training. J Neurophysiol. 1992;67: 1587-1598.

[21] Rogers MW, Pai YC. Patterns of muscle activation accompanying transitions in stance during rapid leg flexion. Journal of Electromyography and Kinesiology. 1993;3:149-156.

[22] Rogers MW, Pai YC. Dynamic transitions in stance support accompanying leg flexion movements in man. Exp Brain Res. 1990;81: 398-402.

[23] Breniere Y, Do MC, Bouisset S. Are dynamic phenomena prior to stepping essential to walking? Journal of Motor Behavior. 1987;19:62-76.

[24] Breniere Y, Do MC, Sanchez. A biomechanical study of the gait initiation process. J Biophys Med Nucl. 1981;5:197-205.

[25] Burleigh AL, Horak FB, Malouin F. Modification of postural responses and step initiation: evidence for goal-directed postural interactions. J Neurophysiol. 1994;72:2892-2902.

[26] Crenna P, Frigo C. A motor programme for the initiation of forward-oriented movements in humans. J Physiol (Lond). 1991;437: 635-653.

[27] Elble RJ, Moody C, Leffler K, Sinha R. The initiation of normal walking. Mov Disord. 1994;9:139-146.

[28] Mann RA, Hagy JL, White V, Liddell D. The initiation of gait. J Bone Joint Surg Am. 1979;61:232-239.

[29] Frank JS, Earl M. Coordination of posture and movement. Phys Ther. 1990;70:855-863.

[30] Friedli WG, Cohen L, Hallett M, et al. Postural adjustments associated with rapid voluntary arm movements, II: biomechanical analysis. J Neurol Neurosurg Psychiatry. 1988;51:232-243.

[31] Beuter A, Carriere L, Boucher JP. Relationships between electromyography and kinematics in human stepping strategies. Neurosci Lett. 1987;77:119 -123.

[32] Mercer VS, Sahrmann SA, Abrams RA, et al. Age group differences in postural adjustments associated with a stepping task. Journal of Motor Behavior. 1997;29:243-253.

[33] Assaiante C, Amblard B. An ontogenetic model for the sensorimotor organization of balance control in humans. Human Movement Science. 1995;14:13-43.

[34] Haas G, Diener HC, Rapp H, Dichgans J. Development of feedback and feedforward control of upright stance. Dev Med Child Neurol. 1989;31:481-488.

[35] Day BL, Steiger MJ, Thompson PD, Marsden CD. Effect of vision and stance width on human body motion when standing: implications for afferent control of lateral sway. J Physiol (Lond). 1993;469:479-499.

[36] Goldie PA, Bach TM, Evans OM. Force platform measures for evaluating postural control: reliability and validity. Arch Phys Med Rehabil. 1989;70:510-517.

[37] Goldie PA, Matyas TA, Spencer KI, McGinley RB. Postural control in standing following stroke: test-retest reliability of some quantitative clinical tests. Phys Ther. 1990;70:234-243.

[38] Riach CL, Hayes KC. Maturation of postural sway in young children. Dev Med Child Neurol. 1987;29:650-658.

[39] Sabin C, Smith JL. Recovery and perturbation of paw-shake responses in spinal cats. J Neurophysiol. 1984;51:680-688.

[40] Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420-428.

[41] Berkinblit MB, Feldman AG, Fukson OI. Adaptability of innate motor patterns and motor control mechanisms. Behav Brain Sci. 1986;9:585-638.

[42] Latash ML, Anson JG. What are "normal movements" in atypical populations? Behav Brain Sci. 1996;19:55-106.

[43] Stelmach GE, Diggles VA. Control theories in motor behavior. Acta Psychol. 1982;50:83-105.

[44] Reed ES. From action gestalts to direct action. In: Whiting HTA, ed. Human Motor Actions: Bernstein Reassessed. Amsterdam, the Netherlands: Elsevier Science Publishers BV; 1984:157-168.

[45] Clark JE. Developmental differences in response processing. Journal of Motor Behavior. 1982;14:247-254.

[46] Salthouse TA, Somberg BL. Isolating the age deficit in speeded performance. J Gerontol. 1982;37:59-63.

[47] Moore S, Brunt D, Nesbitt ML, Juarez T. Investigation of evidence for anticipatory postural adjustments in seated subjects who performed a reaching task. Phys Ther. 1992;72:335-343.

VS Mercer, PT, PhD, is Assistant Professor, Division of Physical Therapy, CB #7135, Medical School Wing E, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7135 (USA) ([e-mail]). This study was completed in partial fulfillment of the requirements for Dr Mercer's doctoral degree at Washington University, St Louis, Mo. Address all correspondence to Dr Mercer.

SA Sahrmann, PT, PhD, FAPTA, is Director, Program in Movement Science, and Professor, Program in Physical Therapy and Department of Neurology, Washington University, St Louis, Mo.

Concept/research design and writing were provided by Dr Mercer and Dr Sahrmann; data collection and analysis, project management, and subjects were provided by Dr Mercer; and facilities/equipment and consultation were provided by Dr Sahrmann.

This study was approved by the Human Studies Committee of Washington University School of Medicine at Washington University Medical Center.

This study was supported, in part, by a grant from the Foundation for Physical Therapy.

Portions of this research were presented at the Joint Congress of the American Physical Therapy Association and the Canadian Physiotherapy Association; June 4-8, 1994; Toronto, Ontario, Canada.

This article was submitted September 23, 1997, and was accepted July 28, 1999.