Reversals of anticipatory postural adjustments during voluntary sway in humans
1 Department of Kinesiology, The Pennsylvania State University, University Park, PA 16802, USA
2 Department of Physical Therapy, University of Delaware, Newark, DE 19716, USA
Abstract
We describe reversals of anticipatory postural adjustments (APAs) with the phase of a voluntary cyclic whole-body sway movement. Subjects (n = 9) held a standard load in extended arms and released it by a bilateral shoulder abduction motion in a self-paced manner at different phases of the sway. The load release task was also performed during quiet stance in three positions: in the middle of the sway range and close to its extreme forward and backward positions. Larger APAs were seen during the sway task as compared to quiet stance. Although the direction of postural perturbation associated with the load release was always the same, the direction of the APAs in the leg muscles reversed when the subjects were close to the extreme forward position as compared to the APAs in other phases and during quiet stance. The trunk muscles showed smaller APA modulation at the extreme positions but larger modulation when passing through the middle position, depending on the direction of sway, forward or backward. The phenomenon of APA reversals emphasizes the important role of safety in the generation of postural adjustments associated with voluntary movements. Based on these findings, APAs could be defined as changes in the activity of postural muscles associated with a predictable perturbation that act to provide maximal safety of the postural task component.
, 百拇医药
Introduction
Upright human posture is inherently unstable with the challenge of balancing the body, which has a relatively high centre of mass (COM) and several joints along the body axis, on a narrow base of support (Hayes, 1982; Winter et al. 1996, 1998). When voluntary movements are made while standing, the redistribution of mass and inertial forces threaten balance. Anticipatory postural adjustments (APAs) are changes in the activity of postural muscles and associated shifts in the centre of pressure (COP; point of application of ground reaction forces) prior to the initiation of action (Belen'kii et al. 1967; Marsden et al. 1979; Cordo & Nashner, 1982; Bouisset & Zattara, 1987; for a review, see Massion, 1992).
, http://www.100md.com
It has been assumed that APAs are based on an estimate of the effects of an expected perturbation on standing balance. It has also been assumed that the purpose of APAs is to generate forces and moments opposing the mechanical effects of the expected postural perturbation (Bouisset & Zattara, 1987, 1990; Friedli et al. 1988).
Humans frequently perform fast arm movements and manipulate objects while moving the body, e.g. during walking, standing up, or leaning. In particular, subjects pulling on a handle while walking exhibit APAs that differ across the gait cycle (Nashner & Forssberg, 1986; Forssberg & Hirschfeld, 1988; Hirschfeld & Forssberg, 1991). In general, these findings have supported the idea that APAs act to oppose the effects of forthcoming perturbations.
, http://www.100md.com
A load release task has been used in many APA studies (Aruin & Latash, 1996; Slijper et al. 2002; Aruin & Shiratori, 2004; Slijper & Latash, 2004). This task leads to reproducible APAs in most postural muscles. It has a certain advantage as compared to voluntary movement tasks since the magnitude of the perturbation does not vary with the natural variability in the voluntary movement that triggers the load release. We used the load release task to examine APAs when subjects stood at different angles of body lean and while subjects voluntarily swayed. The results of our study show that APAs can reverse direction according to the lean of the body during the sway task. It therefore appears that APAs are not always organized to oppose the mechanical effects of expected postural perturbations.
, 百拇医药
Methods
Subjects
Nine healthy right-handed subjects, without any known neurological or motor disorder, participated in the experiment. Of these, five were males and four were females, with the mean age of 29.5 ± 4.6 years, mean weight of 63.6 ± 10.2 kg and mean height of 167.1 ± 9.6 cm (means ± S.D.). The subjects gave written informed consent according to the procedures approved by the Office for Regulatory Compliance of the Pennsylvania State University. The study was performed according to the Declaration of Helsinki.
, 百拇医药
Apparatus
A force platform (AMTI, OR-6) was used to record the moment around a frontal axis (My) and the vertical component of the reaction force (Fz). An oscilloscope (Tektronics TDS 210) showed the time pattern of My to the subject and the experimenter. A uni-directional accelerometer (Sensotec) was taped to the dorsal aspect of the subject's hand over the third metacarpal joint, such that the axis of sensitivity of the accelerometer was directed along the required motion of the hand. Disposable self-adhesive electrodes (3MRedDot, 3M Health Care, St. Paul, MN) were used to record the surface EMG activity of the following leg and trunk muscles: tibialis anterior (TA), soleus (SOL), rectus femoris (RF), biceps femoris (BF), rectus abdominis (RA) and erector spinae (ES). The electrodes were placed on both sides of the subject's body on the muscle bellies, with their centres approximately 3 cm apart. Data were recorded at the sampling frequency of 1000 Hz with 12-bit resolution. A Gateway 450 MHz PC with customized software based on the LabVIEW 4 package (National Instruments) was used to control the experiment and collect the data.
, 百拇医药
Procedure
During all experiments, the subjects held a 3 kg load (20 cm x 20 cm x 10 cm; approximately 4% of subject's body mass) in front of him/her by pressing on the sides of the load with both hands. Pilot studies showed that this load was large enough to bring about reproducible APAs without causing fatigue over repeated trials. The shoulders were at 90 deg flexion and the elbows were fully extended. The load was released under two task conditions: quiet stance (QS) and voluntary sway (VS).
, 百拇医药
In the QS condition, subjects initially stood on the force platform with their feet side-by-side, a hip width apart. They were required to release the load with a quick bilateral shoulder abduction movement. Subjects were instructed to stand as quietly as possible in the initial position before the beginning of the trial. Then, the computer generated a beep 500 ms after data collection began, which indicated that the subject should initiate the required action in a self-paced manner. Subjects were reminded not to initiate their actions immediately after the beep, but to wait for about a second, so that reaction time did not lead to APA modification (cf. De Wolf et al. 1998; Slijper et al. 2002). Data were collected over 3 s for each trial. Subjects performed three series, six repetitions each, at the QS condition in three positions: leaning forward (QSF), neutral standing upright (QSM) and leaning backward (QSB). The degree of body lean in QSF and QSB was such that the COP shift from the average value obtained during QSM was between 1 and 2 cm corresponding to typical COP shifts during the voluntary sway series (see below). For a given subject and for a given direction of the lean, the lean magnitude was reproduced across trials by marking the My position on the oscilloscope. Subjects occupied this initial position prior to every trial.
, 百拇医药
In the VS condition, the initial position of the subject was the same as in the QSM condition (vertical neutral stance). My at the neutral body position was marked on the oscilloscope, and the subject was asked to occupy this position prior to each trial. Then, subjects were instructed to shift their body weight towards their toes and heels rhythmically, thereby swaying forward and backward. They were asked to match the times of the peak forward and backward positions with the beats of the metronome. The metronome frequency was set at 1 Hz, so that the subject completed one full cycle of forward and backward sway in 2 s or at the frequency of 0.5 Hz. Subjects were asked to watch the oscilloscope, which showed them the current value of My. The required My shift was also marked on the oscilloscope (approximately 10 N m in each direction, which corresponded to COP shifts of about 1–2 cm depending on the subject's body weight). Thus, the degree of lean in the peak forward and backward My positions of the VS task corresponded to the degree of lean in the QSF and QSB tasks.
, http://www.100md.com
For each trial, data were collected over 8 s. On hearing the computer-generated beep, the subject was instructed to begin swaying to the beat of the metronome. After completing 2–3 full sway cycles, the subject released the load at a self-selected time. The subjects were instructed to release the load at different phases of the sway, randomly varying across the trials to have approximately equal numbers of trials with load release close to the two extreme positions (forward and backward maximal lean) and in the middle position while moving forward and backwards. The total of 32 trials were collected in three blocks of 8 trials with a rest period of at least a minute between the blocks. Pilot series showed that this instruction resulted in subjects releasing the load at different phases of sway while not forcing them to attend to particular cues. We wanted to keep the load release self-paced and did not provide specific cues at different phases of the cycle, as APAs are known to be modified in reaction time tasks (e.g. De Wolf et al. 1998; Slijper et al. 2002).
, 百拇医药
The order of the QS and VS conditions was balanced across subjects. Rest periods of 8 s between trials and 2 min between the two conditions were given. Fatigue was not an issue. Prior to each condition, two practice trials were given.
In addition to the main trials, two control trials were performed. The subject was asked to hold the load of 3 kg and sway to the beat of the metronome without releasing the load (VSbl). These data were used to compute the EMG activity of the leg and trunk muscles in the absence of the perturbation caused by the load release, as described in the next subsection.
, http://www.100md.com
Data processing
All signals were processed off-line and filtered with a 50 Hz low-pass, fourth-order, zero-lag Butterworth filter using LabVIEW 4 to avoid phase distortions in the signal. All EMG signals were rectified. Individual trials were viewed on the monitor screen and aligned according to the first visible change in the signal of the accelerometer (movement initiation). This moment will be referred to as ‘time zero’ (t0). Changes in the muscle activity associated with APAs were quantified as follows. In the QS trials, rectified EMG signals were integrated from 100 ms prior to t0 to t0 (EMG). Although in some conditions APAs were partly outside this time window, it was not practical to adjust the window of integration for different muscles, subjects, and conditions. So we used a conservative approach of EMG analysis within a standard 100 ms time window. Such an approach has been used in many earlier studies of APAs (e.g. Shiratori & Latash, 2000; Slijper & Latash, 2000). These integrals were corrected by subtracting integrated activity at rest from –500 to –450 ms prior to t0 multiplied by two to match the intervals of EMG integration (the baseline EMG activity, EMGblqs).
, 百拇医药
(1A)
In the VS trials, the most forward positions of the My signal were identified and marked. Then all the data were cut into cycles such that each cycle began at the most forward position of My and ended one sample before the next most forward My position. The time of load release within each trial was computed as a percentage of the sway cycle, with 0% representing the most forward position and 50% representing the most backward position. The following four ranges of the cycle time were used to assign a trial to one of the four groups:
, 百拇医药
Forward (VSF): 87–100% and 0–5%
Backward (VSB): 40–55%
Middle position while swaying forward: (VSMF): 70–78%
Middle position while swaying backward: (VSMB): 20–28%
In general, subjects were able to release the load close to the middle position far more often than close to the extreme positions. In order to get a reasonable number of trials from each subject for statistical purposes, we used a wider range of percentages for the extreme positions. These intervals were also slightly skewed in the direction leading to an extreme position rather than away from it, as subjects dropped the load more often moving towards these extreme positions than when moving away from them.
, http://www.100md.com
Similar to the data processing of QS trials, for the VS trials, rectified EMG signals were integrated from 100 ms prior to t0 to t0. These integrals were corrected by subtracting baseline EMG, which was computed from the unperturbed VS (VSbl) trial over a 100 ms interval prior to the phase of the sway cycle when the load was released (EMGblvs), that is, corresponding to the same time interval as in the VS trial.
(1B)
In order to compare the integrated EMG indices (EMG) across muscles and subjects, we normalized them by the highest absolute value of EMG values for each muscle, across trials for each subject (EMGcontrol) as follows: EMG indices for a muscle from a given trial were divided by EMGcontrol, such that all EMGs were between 1 and –1. Negative values corresponded to a decrease in the EMG activity during APAs.
, http://www.100md.com
To illustrate the change in pattern of IEMGs as a function of sway cycle, IEMGs were calculated at different phases of the VS task and arranged in order of their appearance in the sway cycle expressed as a percentage of the cycle duration. In order to minimize the effects of outliers, we used a 5-point moving average window, i.e. the averages over 5 consecutive data points within the sway cycle.
Coordinates of the centre of pressure (COP) in the anterior–posterior directions were calculated using the following approximation:
, 百拇医药
Statistics
Statistical methods included repeated-measures multivariate analysis of variance (MANOVA). To determine the effects of TASK (QS versus VS) and POSITION (leaning forward, F or backward, B) on IEMG indices of APAs, we performed two-way MANOVAs on the IEMG indices for each pair of muscles acting at each joint (TA/SOL, RF/BF, or RA/ES). Three separate two-way MANOVAs on APAs of muscle pairs from each joint were run: MANOVA-Ia, on IEMG of TA and SOL; MANOVA-Ib on IEMG of RF and BF and MANOVA-Ic on IEMG of RA and ES.
, http://www.100md.com
In order to test if there was an effect of different task conditions on the APAs when the load was released at the middle position, a one-way MANOVA (MANOVA-II) was performed with CONDITION as a factor having three levels: QSM, VSMB and VSMF. Again, we performed three MANOVAs on the IEMG indices for each pair of muscles acting at the three joints (TA/SOL, RF/BF or RA/ES). Planned comparisons were performed to compare the effects of QS and VS tasks. To test specific differences between pairs of conditions, Tukey's honest significant difference (HSD) tests were used.
, 百拇医药
Results
Muscles on the left and right sides of the body showed qualitatively similar results in all analyses. Therefore for simplicity, only the results of analyses on the muscles of the left side of the body are presented and discussed further.
General EMG patterns: quiet stance (QS)
When a person stood in a neutral upright position and held the load in extended arms, typically there was an increase in the activity of the dorsal muscles (SOL, BF and ES). When the person released the load with a quick bilateral shoulder abduction motion, prior to the load release, there was a drop in the activity of the dorsal muscles accompanied sometimes by an increase in the activity of the ventral muscles (TA, RF and RA). A typical pattern is illustrated for a representative subject in Fig. 1B. Time zero corresponds to the initiation of load release. APAs were quantified within a 100 ms time interval prior to this time (the time period between the two vertical lines in Fig. 1). Not all subjects showed changes in muscle activity in all muscles.
, http://www.100md.com
The load was released in the forward (A: QSF), middle (B: QSM) and backward (C: QSB) positions. Time zero corresponds to the time of load release, and the two vertical lines indicate the 100 ms time interval over which indices of anticipatory postural activity (APA) were quantified. EMG signals for soleus, biceps femoris and erector spinae are inverted for better visualization.
A similar pattern was seen when the subject stood quietly while leaning forward or backward, i.e. on the toes or the heels (Fig. 1A and C, respectively), but the early EMG changes were usually less marked and more variable across subjects. In particular, Fig. 1A shows no visible RF burst. In Fig. 1C, the RF burst was smaller than while standing in the neutral position (Fig. 1B) and there is no suppression of the SOL activity.
, 百拇医药
General EMG patterns: voluntary sway (VS)
EMG patterns associated with an unperturbed cycle (VSbl) of voluntary sway when holding a load with extended arms are presented in Fig. 2 for a representative subject. A cycle beginning from the most forward position followed by sway backward towards the heels and a return to the forward position is illustrated (see My in the top panel). Since the frequency of sway was 0.5 Hz, half a cycle took about 1 s in time. Therefore, times 0 and 2 s in Fig. 2 correspond to the extreme forward position, and time 1 s corresponds to the extreme backward position. Note the increase in the activity of the ventral muscles (TA, RF) when approaching the extreme backward position and a simultaneous drop in the activity of the dorsal muscles (SOL, BF, ES). Conversely, when approaching the extreme forward position, there is an increase in the activity of the dorsal muscles, accompanied by a drop in the ventral muscle activity.
, 百拇医药
The top panel shows the moment about the frontal axis (My), where the highest positive value corresponds to the extreme forward position and the smallest negative value corresponds to the extreme backward position during VS. The remaining three panels show EMG patterns in the lower leg, thigh and trunk muscles, respectively. EMG signals for soleus, biceps femoris and erector spinae muscles are inverted for better visualization. Note that the ventral muscles (tibialis anterior, rectus femoris, rectus abdominis) show an increase in the activity during motion towards the backward position and the dorsal muscles (soleus, biceps femoris, erector spinae) show an increase in activity during motion towards the forward position.
, 百拇医药
Changes in the muscle activity associated with release of the load (APAs) at four points during the sway cycle are illustrated in Fig. 3 for a representative subject. Similarly to Fig. 1, time zero corresponds to the initiation of load release, and APAs were quantified within a 100 ms time interval prior to this time, by subtracting the integrated activity at the same phase and over the same time interval during an unperturbed sway trail. Figure 3B and D shows changes in the activity of muscles when the load was released close to the neutral position during sway, while passing the neutral position during swaying in the backward and forward directions, respectively. Figure 3A and C shows EMGs when the load was released close to the extreme forward and backward positions, respectively, during sway. Note the anticipatory increase in the activity of ventral muscles (TA, RF, and to a lesser extent RA) when the load was released while moving towards the extreme backward position (Fig. 3B; VSMB) and while passing through this position (Fig. 3C; VSB). This was accompanied by an anticipatory drop in the activity of the dorsal muscles (SOL, BF and ES). In contrast, if the load was released while moving towards (Fig. 3D; VSMF) or passing through the extreme forward position (Fig. 3A; VSF), there was an anticipatory increase in the activity of the dorsal muscles and an anticipatory drop in the activity of the ventral muscles. These APAs also varied in different muscles across subjects.
, 百拇医药
The load was released in the forward (A: VSF), middle when moving backward (B: VSMB), backward (C: VSB) and middle when moving forward (D: VSMF) positions. Time zero corresponds to time of load release and the vertical lines indicate the 100 ms time interval before load release over which indices of anticipatory postural activity (APA) were quantified. EMG signals for soleus, biceps femoris and erector spinae are inverted for better visualization.
Figure 4 shows smoothed IEMG indices (see eqn (2) in the Methods) plotted as a function of phase of the sway cycle for a representative subject. Recall that 0 and 100% correspond to the most forward position and 50% to the most backward position. A third-order polynomial was fitted to the data in Fig. 4 and the squared regression coefficient (R2) values are presented. In general, the ankle (TA, SOL) and knee muscles (RF, BF) showed similar directions of changes in APAs over the sway cycle. The anterior muscles, TA and RF, showed an increase in APAs in the mid-backward position (passing through the middle position while moving backwards, 25% of the cycle), reaching the highest values at the extreme backward position (50% of the cycle) followed by a drop in APAs in the mid-forward position. (passing through the middle position while moving forward, 75% of the cycle) and the smallest APAs in the extreme forward positions (0/100% of the cycle). The posterior muscles, SOL and BF, showed out-of-phase APA changes with respect to APA changes in TA and RF. The trunk muscles, RA and ES, showed a different pattern of activity. Both RA and ES showed larger APAs in the mid-forward and mid-backward positions, and smaller APAs at the extreme positions.
, 百拇医药
Ventral muscles: TA, tibialis anterior; RF, rectus femoris; RA, rectus abdominis. Dorsal muscles: SOL, soleus; BF, biceps femoris; ES, erector spinae. 0 and 100%, extreme forward position; 50%, extreme backward position; 25%, middle position when moving backward; and 75%, middle position when moving forward. Data were smoothed with a 5-point moving average window and fitted with a third-order polynomial. R2 values are presented.
Effect of task on APAs in extreme body positions
, 百拇医药
Figure 5 shows IEMGs of individual muscles averaged across subjects with standard error bars for the QS and VS tasks in the extreme forward and backward positions. Note that the direction of APAs in the forward and backward positions is opposite in all the muscles. This effect was small for the QS task, but was more striking in the VS task. The lower leg and thigh muscles showed significant main effects of TASK (QS and VS) and POSITION (extreme forward and extreme backward; both main effects Wilks' lambda < 0.4; F2,7 > 6; P < 0.05). SOL and RF muscles, in particular showed significantly larger APAs in the VS condition as compared to QS (P < 0.05). The ventral muscles (TA and RF) showed an increase in IEMGs in the backward position (P < 0.05), while the dorsal muscles (particularly SOL) showed an increase in IEMGs in the forward position (P < 0.05; see also Figs 1 and 3). There was a significant TASK–POSITION interaction effect for the TA/SOL muscle pair (Wilks' lambda < 3; F2,7 > 10; P < 0.01), while this interaction effect was just below significance for the RF/BF pair (P = 0.064). Even though the trunk muscles, RA and ES, showed a difference in the magnitudes of APAs in the different phases of the VS cycle (see Fig. 4), these effects varied across subjects, and these muscles did not show any significant TASK or POSITION effects.
, 百拇医药
Average across subjects IEMG indices in the extreme forward (open columns) and backward (filled columns) positions are shown with standard error bars. Muscle names are abbreviated as in Fig. 4. Significant differences between the QS and VS tasks; * significant differences between the extreme forward and backward positions.
Effect of task on APAs in the middle body position
To analyse the effects of sway on APAs when the subjects were close to the middle position, IEMG indices for the APAs were compared across three conditions: (1) in the middle position during quiet stance (QSM), (2) while passing the middle position in the forward direction during VS (VSMF), and (3) while passing the middle position in the backward direction (VSMB). Averaged across subjects normalized IEMG indices are shown in Fig. 6 with standard error bars. Note the pronounced suppression of the muscle activity (negative values) in the dorsal muscles for QSM (see also Fig. 1). During voluntary sway, this suppression became less pronounced. In the ventral muscles, a tendency towards higher APAs was observed. MANOVA on IEMG over all pairs of muscles (TA/SOL, RF/BF and RA/ES) showed a significant main effect of CONDITION (QSM, VSMB and VSMF; Wilks' lambda < 0.2; F4,5 > 6; P < 0.05). Planned comparisons between the QSM and the two VS conditions revealed that APAs in SOL, BF, RA and ES were significantly larger in the VS conditions as compared to QS (F1,8 > 5; P < 0.05; indicated by in Fig. 6). Further, SOL and ES APAs were significantly higher during VSMF as compared to QSM. ES also showed significantly higher APAs during VSMF as compared to VSMB (indicated by * in Fig. 6).
, 百拇医药
Quiet stance middle position (QSM; filled columns), voluntary sway middle position when swaying backwards (VSMB; open columns) and when swaying forward (VSMF; hatched columns) are shown with standard error bars. Muscle names are abbreviated as in Fig. 4. Significant differences between the QS and VS tasks; * significant differences between VSMB and VSMF.
Discussion
The magnitude of anticipatory postural adjustments (APAs) is known to depend on the magnitude of the perturbation, its direction and postural stability (Nouillot et al. 1992; Aruin & Latash, 1995; Aruin et al. 1998; Dietz et al. 2000). In a series of studies, Forssberg and colleagues (Nashner & Forssberg, 1986; Forssberg & Hirschfeld, 1988; Hirschfeld & Forssberg, 1991) have shown that APA magnitude is modulated within the gait cycle. Their subjects were asked to respond to an audio signal that was presented in different phases of the step cycle with a strong pull on the handle. Lateral gastrocnemius and hamstring muscles were activated during APAs occurring in the early support phase whereas tibialis anterior and quadriceps muscles were activated when the pull was exerted just prior to and during the swing phase.
, 百拇医药
Interpretation of these findings is complicated, however, by the fact that the leg muscles were involved in two activities, locomotion and APAs. In particular, in the swing phase, the muscles of the leg could not contribute to postural stabilization since they had no contact with the support. It is hard therefore to assess the functional significance of the apparently reversed APAs seen in the swing phase. Our experiments have demonstrated reversals of APA action when both legs could contribute to postural stabilization in all phases of the sway cycle. APA reversals in such conditions represent a major novel finding of our study.
, 百拇医药
The role of postural stability: refining the APA definition
Typically, the magnitude of APAs decreases for both very stable conditions (e.g. standing while leaning against a wall, Friedli et al. 1984) and unstable conditions (e.g. standing on a narrow support area, Aruin et al. 1998). Our experiments confirmed observations by Aruin et al. (1998) that shifting the body weight forwards or backwards leads to a decrease in the magnitude of APAs associated with releasing a standard load from the extended arms.
, http://www.100md.com
Releasing a load held in the extended arms always tends to tilt the body backwards. Hence, in the most backward position in the sway cycle, this perturbation was more likely to push the body out of the limits of postural stability. In contrast, the same perturbation tended to move the body away from the dangerously close edge of stability when the subjects approached the most forward body posture.
The typical APA pattern associated with load release (Aruin et al. 1998; Fig. 1) always tends to tilt the trunk forward. When the body is close to the most forward position, this APA pattern would move the body closer to the limit of stability and, as such, might by itself violate the balance. In such a situation, the CNS reverses the direction of the APAs (Fig. 5) and moves the body away from the edge of the support area. Mechanical action of reversed APAs does not counteract the mechanical effects of the perturbation associated with the load release but acts in the same direction.
, 百拇医药
Traditionally APAs have been defined as changes in the patterns of muscle activity that produce forces and moments acting against the mechanical effect of an expected perturbation (Bouisset & Zattara, 1987; Friedli et al. 1988; Massion, 1992). Our experiments have shown that APAs can produce mechanical effects on posture acting together with the perturbation as long as such action is consistent with maximizing the safety of the posture. We therefore suggest that APAs could be defined as changes in the activity of postural muscles associated with a predictable perturbation that act to provide maximal safety of the postural task component. This definition is consistent with the findings of decreased APAs under postural threat or fear of falling (Adkin et al. 2002).
, 百拇医药
One view on APAs is that they result from an internal forward model that takes into account dynamic consequences of an expected perturbation and generates responses to counteract these consequences (Wing et al. 1997). However, we do not see how a forward-model-based APA would be able to account for the reversal of APA direction, which is based on safety concerns. An alternative interpretation is that predictive motor control mechanisms such as APAs are not based on computations of forces but ensure stability of balance using equilibrium-point control (Feldman & Latash, 2005); shifts in equilibrium position of the body during APAs may be based on a combination of factors, including closeness to the edge of the support.
, 百拇医药
The role of the magnitude of the perturbation
In our experiments, the actual perturbation associated with the load release could be expected to vary with the phase of the sway because of the modulation of inertial forces. This modulation was relatively small. Peak horizontal load acceleration during a sine sway at 2 Hz with the amplitude of load motion of about 0.5 m may be assessed as close to 0.5 m s–2. With respect to the ankle joint, the lever arm of the load was about 1 m. For the 3 kg load, the peak inertial torques at the ankle joints related to the load motion were of the order of 1.5 N m. The gravitational force acting on the load produced the moment about the ankle joint of about 20 N m. Hence, the motion-related modulation of the magnitude of the perturbation at the two extreme positions was under 10%.
, 百拇医药
Large differences in the APA magnitude were seen in the phases of moving through the middle posture in different directions (Fig. 6), particularly in the RA/ES muscle pair that has been viewed as providing a basic pattern of the APAs (Shiratori & Latash, 2000). At those phases, the body moved at the highest speed while the acceleration was close to zero. Hence, this modulation could not be related to the motion-related modulation of the magnitude of the inertial forces.
, http://www.100md.com
The observed differences in the APAs while moving through the mid-posture in opposite directions could be related to preparing a response while taking into account the fact that the body was moving towards one of the extreme positions. This makes it possible to relate changes in the APAs in all muscle pairs (cf. Fig. 5) to providing maximal safety for the postural task.
References
Adkin AL, Frank JS, Carpenter MG & Peysar GW (2002). Fear of falling modifies anticipatory postural control. Exp Brain Res 143, 160–170.
, 百拇医药
Aruin AS, Forrest WR & Latash ML (1998). Anticipatory postural adjustments in conditions of postural instability. Electroencephalogr Clin Neurophysiol 109, 350–359.
Aruin AS & Latash ML (1995). The role of motor action in anticipatory postural adjustments studied with self-induced and externally triggered perturbations. Exp Brain Res 106, 291–300.
Aruin AS & Latash ML (1996). Anticipatory postural adjustments during self-initiated perturbations of different magnitude triggered by a standard motor action. Electroencephalogr Clin Neurophysiol 101, 497–503.
, 百拇医药
Aruin AS & Shiratori T (2004). The effect of the amplitude of motor action on anticipatory postural adjustments. J Electromyogr Kinesiol 14, 455–462.
Belen'kii V, Gurfinkel VS & Pal'tsev YI (1967). Elements of control of voluntary movements. Biofizika 10, 135–141.
Bouisset S & Zattara M (1987). Biomechanical study of the programming of anticipatory postural adjustments associated with voluntary movement. J Biomech 20, 735–742.
, 百拇医药
Bouisset S & Zattara M (1990). Segmental movement as a perturbation to balance Facts and concepts. In Multiple Muscle Systems. Biomechanics and Movement Organization, ed. Winters J & Woo S-Y, pp. 498–506. Springer, New York.
Cordo PJ & Nashner LM (1982). Properties of postural adjustments associated with rapid arm movements. J Neurophysiol 47, 287–302.
De Wolf S, Slijper H & Latash ML (1998). Anticipatory postural adjustments during self-paced and reaction-time movements. Exp Brain Res 121, 7–19.
, 百拇医药
Dietz V, Kowalewski R, Nakazawa K & Colombo G (2000). Effects of changing stance conditions on anticipatory postural adjustment and reaction time to voluntary arm movement in humans. J Physiol 524, 617–627.
Feldman AG & Latash ML (2005). Testing hypotheses and the advancement of science: recent attempts to falsify the equilibrium point hypothesis. Exp Brain Res 161, 91–103.
Forssberg H & Hirschfeld H (1988). Phasic modulation of postural activation patterns during human walking. Prog Brain Res 76, 221–227.
, 百拇医药
Friedli WG, Cohen L, Hallett M, Stanhope S & Simon SR (1988). Postural adjustments associated with rapid voluntary arm movements. II. Biomechanical analysis. J Neurol Neurosurg Psychiatry 51, 232–243.
Friedli WG, Hallett M & Simon SR (1984). Postural adjustments associated with rapid voluntary arm movements I. Electromyographic data. J Neurol Neurosurg Psychiatry 47, 611–622.
Hayes KC (1982). Biomechanics of postural control. Exerc Sport Sci Rev 10, 363–339.
, http://www.100md.com
Hirschfeld H & Forssberg H (1991). Phase-dependent modulations of anticipatory postural activity during human locomotion. J Neurophysiol 66, 12–19.
Marsden CD, Merton PA & Morton HB (1979). Anticipatory postural responses in the human subject. J Physiol 275, 47P–48P.
Massion J (1992). Movement, posture and equilibrium: interaction and coordination. Prog Neurobiol 38, 35–56.
Nashner LM & Forssberg H (1986). Phase-dependent organization of postural adjustments associated with arm movements while walking. J Neurophysiol 55, 1382–1394.
, http://www.100md.com
Nouillot P, Bouisset S & Do MC (1992). Do fast voluntary movements necessitate anticipatory postural adjustments even if equilibrium is unstable Neurosci Lett 147, 1–4.
Shiratori T & Latash ML (2000). The roles of proximal and distal muscles in anticipatory postural adjustments under asymmetrical perturbations and during standing on rollerskates. Clin Neurophysiol 111, 613–623.
Slijper H & Latash ML (2000). The effects of instability and additional hand support on anticipatory postural adjustments in leg, trunk, and arm muscles during standing. Exp Brain Res 135, 81–93.
, 百拇医药
Slijper H & Latash ML (2004). The effects of muscle vibration on anticipatory postural adjustments. Brain Res 1015, 57–72.
Slijper H, Latash ML & Mordkoff JT (2002). Anticipatory postural adjustments under simple and choice reaction time conditions. Brain Res 924, 184–197.
Wing A, Flanagan RJ & Richardson J (1997). Anticipatory postural adjustments in stance and grip. Exp Brain Res 116, 122–130.
Winter DA, Patla AE, Prince F, Ishac M & Gielo-Perczak K (1998). Stiffness control of balance in quiet standing. J Neurophysiol 80, 1211–1221.
Winter DA, Prince F, Frank JS, Powell C & Zabjek KF (1996). Unified theory regarding A/P and M/L balance in quiet stance. J Neurophysiol 75, 2334–2343., http://www.100md.com(Vijaya Krishnamoorthy, an)
2 Department of Physical Therapy, University of Delaware, Newark, DE 19716, USA
Abstract
We describe reversals of anticipatory postural adjustments (APAs) with the phase of a voluntary cyclic whole-body sway movement. Subjects (n = 9) held a standard load in extended arms and released it by a bilateral shoulder abduction motion in a self-paced manner at different phases of the sway. The load release task was also performed during quiet stance in three positions: in the middle of the sway range and close to its extreme forward and backward positions. Larger APAs were seen during the sway task as compared to quiet stance. Although the direction of postural perturbation associated with the load release was always the same, the direction of the APAs in the leg muscles reversed when the subjects were close to the extreme forward position as compared to the APAs in other phases and during quiet stance. The trunk muscles showed smaller APA modulation at the extreme positions but larger modulation when passing through the middle position, depending on the direction of sway, forward or backward. The phenomenon of APA reversals emphasizes the important role of safety in the generation of postural adjustments associated with voluntary movements. Based on these findings, APAs could be defined as changes in the activity of postural muscles associated with a predictable perturbation that act to provide maximal safety of the postural task component.
, 百拇医药
Introduction
Upright human posture is inherently unstable with the challenge of balancing the body, which has a relatively high centre of mass (COM) and several joints along the body axis, on a narrow base of support (Hayes, 1982; Winter et al. 1996, 1998). When voluntary movements are made while standing, the redistribution of mass and inertial forces threaten balance. Anticipatory postural adjustments (APAs) are changes in the activity of postural muscles and associated shifts in the centre of pressure (COP; point of application of ground reaction forces) prior to the initiation of action (Belen'kii et al. 1967; Marsden et al. 1979; Cordo & Nashner, 1982; Bouisset & Zattara, 1987; for a review, see Massion, 1992).
, http://www.100md.com
It has been assumed that APAs are based on an estimate of the effects of an expected perturbation on standing balance. It has also been assumed that the purpose of APAs is to generate forces and moments opposing the mechanical effects of the expected postural perturbation (Bouisset & Zattara, 1987, 1990; Friedli et al. 1988).
Humans frequently perform fast arm movements and manipulate objects while moving the body, e.g. during walking, standing up, or leaning. In particular, subjects pulling on a handle while walking exhibit APAs that differ across the gait cycle (Nashner & Forssberg, 1986; Forssberg & Hirschfeld, 1988; Hirschfeld & Forssberg, 1991). In general, these findings have supported the idea that APAs act to oppose the effects of forthcoming perturbations.
, http://www.100md.com
A load release task has been used in many APA studies (Aruin & Latash, 1996; Slijper et al. 2002; Aruin & Shiratori, 2004; Slijper & Latash, 2004). This task leads to reproducible APAs in most postural muscles. It has a certain advantage as compared to voluntary movement tasks since the magnitude of the perturbation does not vary with the natural variability in the voluntary movement that triggers the load release. We used the load release task to examine APAs when subjects stood at different angles of body lean and while subjects voluntarily swayed. The results of our study show that APAs can reverse direction according to the lean of the body during the sway task. It therefore appears that APAs are not always organized to oppose the mechanical effects of expected postural perturbations.
, 百拇医药
Methods
Subjects
Nine healthy right-handed subjects, without any known neurological or motor disorder, participated in the experiment. Of these, five were males and four were females, with the mean age of 29.5 ± 4.6 years, mean weight of 63.6 ± 10.2 kg and mean height of 167.1 ± 9.6 cm (means ± S.D.). The subjects gave written informed consent according to the procedures approved by the Office for Regulatory Compliance of the Pennsylvania State University. The study was performed according to the Declaration of Helsinki.
, 百拇医药
Apparatus
A force platform (AMTI, OR-6) was used to record the moment around a frontal axis (My) and the vertical component of the reaction force (Fz). An oscilloscope (Tektronics TDS 210) showed the time pattern of My to the subject and the experimenter. A uni-directional accelerometer (Sensotec) was taped to the dorsal aspect of the subject's hand over the third metacarpal joint, such that the axis of sensitivity of the accelerometer was directed along the required motion of the hand. Disposable self-adhesive electrodes (3MRedDot, 3M Health Care, St. Paul, MN) were used to record the surface EMG activity of the following leg and trunk muscles: tibialis anterior (TA), soleus (SOL), rectus femoris (RF), biceps femoris (BF), rectus abdominis (RA) and erector spinae (ES). The electrodes were placed on both sides of the subject's body on the muscle bellies, with their centres approximately 3 cm apart. Data were recorded at the sampling frequency of 1000 Hz with 12-bit resolution. A Gateway 450 MHz PC with customized software based on the LabVIEW 4 package (National Instruments) was used to control the experiment and collect the data.
, 百拇医药
Procedure
During all experiments, the subjects held a 3 kg load (20 cm x 20 cm x 10 cm; approximately 4% of subject's body mass) in front of him/her by pressing on the sides of the load with both hands. Pilot studies showed that this load was large enough to bring about reproducible APAs without causing fatigue over repeated trials. The shoulders were at 90 deg flexion and the elbows were fully extended. The load was released under two task conditions: quiet stance (QS) and voluntary sway (VS).
, 百拇医药
In the QS condition, subjects initially stood on the force platform with their feet side-by-side, a hip width apart. They were required to release the load with a quick bilateral shoulder abduction movement. Subjects were instructed to stand as quietly as possible in the initial position before the beginning of the trial. Then, the computer generated a beep 500 ms after data collection began, which indicated that the subject should initiate the required action in a self-paced manner. Subjects were reminded not to initiate their actions immediately after the beep, but to wait for about a second, so that reaction time did not lead to APA modification (cf. De Wolf et al. 1998; Slijper et al. 2002). Data were collected over 3 s for each trial. Subjects performed three series, six repetitions each, at the QS condition in three positions: leaning forward (QSF), neutral standing upright (QSM) and leaning backward (QSB). The degree of body lean in QSF and QSB was such that the COP shift from the average value obtained during QSM was between 1 and 2 cm corresponding to typical COP shifts during the voluntary sway series (see below). For a given subject and for a given direction of the lean, the lean magnitude was reproduced across trials by marking the My position on the oscilloscope. Subjects occupied this initial position prior to every trial.
, 百拇医药
In the VS condition, the initial position of the subject was the same as in the QSM condition (vertical neutral stance). My at the neutral body position was marked on the oscilloscope, and the subject was asked to occupy this position prior to each trial. Then, subjects were instructed to shift their body weight towards their toes and heels rhythmically, thereby swaying forward and backward. They were asked to match the times of the peak forward and backward positions with the beats of the metronome. The metronome frequency was set at 1 Hz, so that the subject completed one full cycle of forward and backward sway in 2 s or at the frequency of 0.5 Hz. Subjects were asked to watch the oscilloscope, which showed them the current value of My. The required My shift was also marked on the oscilloscope (approximately 10 N m in each direction, which corresponded to COP shifts of about 1–2 cm depending on the subject's body weight). Thus, the degree of lean in the peak forward and backward My positions of the VS task corresponded to the degree of lean in the QSF and QSB tasks.
, http://www.100md.com
For each trial, data were collected over 8 s. On hearing the computer-generated beep, the subject was instructed to begin swaying to the beat of the metronome. After completing 2–3 full sway cycles, the subject released the load at a self-selected time. The subjects were instructed to release the load at different phases of the sway, randomly varying across the trials to have approximately equal numbers of trials with load release close to the two extreme positions (forward and backward maximal lean) and in the middle position while moving forward and backwards. The total of 32 trials were collected in three blocks of 8 trials with a rest period of at least a minute between the blocks. Pilot series showed that this instruction resulted in subjects releasing the load at different phases of sway while not forcing them to attend to particular cues. We wanted to keep the load release self-paced and did not provide specific cues at different phases of the cycle, as APAs are known to be modified in reaction time tasks (e.g. De Wolf et al. 1998; Slijper et al. 2002).
, 百拇医药
The order of the QS and VS conditions was balanced across subjects. Rest periods of 8 s between trials and 2 min between the two conditions were given. Fatigue was not an issue. Prior to each condition, two practice trials were given.
In addition to the main trials, two control trials were performed. The subject was asked to hold the load of 3 kg and sway to the beat of the metronome without releasing the load (VSbl). These data were used to compute the EMG activity of the leg and trunk muscles in the absence of the perturbation caused by the load release, as described in the next subsection.
, http://www.100md.com
Data processing
All signals were processed off-line and filtered with a 50 Hz low-pass, fourth-order, zero-lag Butterworth filter using LabVIEW 4 to avoid phase distortions in the signal. All EMG signals were rectified. Individual trials were viewed on the monitor screen and aligned according to the first visible change in the signal of the accelerometer (movement initiation). This moment will be referred to as ‘time zero’ (t0). Changes in the muscle activity associated with APAs were quantified as follows. In the QS trials, rectified EMG signals were integrated from 100 ms prior to t0 to t0 (EMG). Although in some conditions APAs were partly outside this time window, it was not practical to adjust the window of integration for different muscles, subjects, and conditions. So we used a conservative approach of EMG analysis within a standard 100 ms time window. Such an approach has been used in many earlier studies of APAs (e.g. Shiratori & Latash, 2000; Slijper & Latash, 2000). These integrals were corrected by subtracting integrated activity at rest from –500 to –450 ms prior to t0 multiplied by two to match the intervals of EMG integration (the baseline EMG activity, EMGblqs).
, 百拇医药
(1A)
In the VS trials, the most forward positions of the My signal were identified and marked. Then all the data were cut into cycles such that each cycle began at the most forward position of My and ended one sample before the next most forward My position. The time of load release within each trial was computed as a percentage of the sway cycle, with 0% representing the most forward position and 50% representing the most backward position. The following four ranges of the cycle time were used to assign a trial to one of the four groups:
, 百拇医药
Forward (VSF): 87–100% and 0–5%
Backward (VSB): 40–55%
Middle position while swaying forward: (VSMF): 70–78%
Middle position while swaying backward: (VSMB): 20–28%
In general, subjects were able to release the load close to the middle position far more often than close to the extreme positions. In order to get a reasonable number of trials from each subject for statistical purposes, we used a wider range of percentages for the extreme positions. These intervals were also slightly skewed in the direction leading to an extreme position rather than away from it, as subjects dropped the load more often moving towards these extreme positions than when moving away from them.
, http://www.100md.com
Similar to the data processing of QS trials, for the VS trials, rectified EMG signals were integrated from 100 ms prior to t0 to t0. These integrals were corrected by subtracting baseline EMG, which was computed from the unperturbed VS (VSbl) trial over a 100 ms interval prior to the phase of the sway cycle when the load was released (EMGblvs), that is, corresponding to the same time interval as in the VS trial.
(1B)
In order to compare the integrated EMG indices (EMG) across muscles and subjects, we normalized them by the highest absolute value of EMG values for each muscle, across trials for each subject (EMGcontrol) as follows: EMG indices for a muscle from a given trial were divided by EMGcontrol, such that all EMGs were between 1 and –1. Negative values corresponded to a decrease in the EMG activity during APAs.
, http://www.100md.com
To illustrate the change in pattern of IEMGs as a function of sway cycle, IEMGs were calculated at different phases of the VS task and arranged in order of their appearance in the sway cycle expressed as a percentage of the cycle duration. In order to minimize the effects of outliers, we used a 5-point moving average window, i.e. the averages over 5 consecutive data points within the sway cycle.
Coordinates of the centre of pressure (COP) in the anterior–posterior directions were calculated using the following approximation:
, 百拇医药
Statistics
Statistical methods included repeated-measures multivariate analysis of variance (MANOVA). To determine the effects of TASK (QS versus VS) and POSITION (leaning forward, F or backward, B) on IEMG indices of APAs, we performed two-way MANOVAs on the IEMG indices for each pair of muscles acting at each joint (TA/SOL, RF/BF, or RA/ES). Three separate two-way MANOVAs on APAs of muscle pairs from each joint were run: MANOVA-Ia, on IEMG of TA and SOL; MANOVA-Ib on IEMG of RF and BF and MANOVA-Ic on IEMG of RA and ES.
, http://www.100md.com
In order to test if there was an effect of different task conditions on the APAs when the load was released at the middle position, a one-way MANOVA (MANOVA-II) was performed with CONDITION as a factor having three levels: QSM, VSMB and VSMF. Again, we performed three MANOVAs on the IEMG indices for each pair of muscles acting at the three joints (TA/SOL, RF/BF or RA/ES). Planned comparisons were performed to compare the effects of QS and VS tasks. To test specific differences between pairs of conditions, Tukey's honest significant difference (HSD) tests were used.
, 百拇医药
Results
Muscles on the left and right sides of the body showed qualitatively similar results in all analyses. Therefore for simplicity, only the results of analyses on the muscles of the left side of the body are presented and discussed further.
General EMG patterns: quiet stance (QS)
When a person stood in a neutral upright position and held the load in extended arms, typically there was an increase in the activity of the dorsal muscles (SOL, BF and ES). When the person released the load with a quick bilateral shoulder abduction motion, prior to the load release, there was a drop in the activity of the dorsal muscles accompanied sometimes by an increase in the activity of the ventral muscles (TA, RF and RA). A typical pattern is illustrated for a representative subject in Fig. 1B. Time zero corresponds to the initiation of load release. APAs were quantified within a 100 ms time interval prior to this time (the time period between the two vertical lines in Fig. 1). Not all subjects showed changes in muscle activity in all muscles.
, http://www.100md.com
The load was released in the forward (A: QSF), middle (B: QSM) and backward (C: QSB) positions. Time zero corresponds to the time of load release, and the two vertical lines indicate the 100 ms time interval over which indices of anticipatory postural activity (APA) were quantified. EMG signals for soleus, biceps femoris and erector spinae are inverted for better visualization.
A similar pattern was seen when the subject stood quietly while leaning forward or backward, i.e. on the toes or the heels (Fig. 1A and C, respectively), but the early EMG changes were usually less marked and more variable across subjects. In particular, Fig. 1A shows no visible RF burst. In Fig. 1C, the RF burst was smaller than while standing in the neutral position (Fig. 1B) and there is no suppression of the SOL activity.
, 百拇医药
General EMG patterns: voluntary sway (VS)
EMG patterns associated with an unperturbed cycle (VSbl) of voluntary sway when holding a load with extended arms are presented in Fig. 2 for a representative subject. A cycle beginning from the most forward position followed by sway backward towards the heels and a return to the forward position is illustrated (see My in the top panel). Since the frequency of sway was 0.5 Hz, half a cycle took about 1 s in time. Therefore, times 0 and 2 s in Fig. 2 correspond to the extreme forward position, and time 1 s corresponds to the extreme backward position. Note the increase in the activity of the ventral muscles (TA, RF) when approaching the extreme backward position and a simultaneous drop in the activity of the dorsal muscles (SOL, BF, ES). Conversely, when approaching the extreme forward position, there is an increase in the activity of the dorsal muscles, accompanied by a drop in the ventral muscle activity.
, 百拇医药
The top panel shows the moment about the frontal axis (My), where the highest positive value corresponds to the extreme forward position and the smallest negative value corresponds to the extreme backward position during VS. The remaining three panels show EMG patterns in the lower leg, thigh and trunk muscles, respectively. EMG signals for soleus, biceps femoris and erector spinae muscles are inverted for better visualization. Note that the ventral muscles (tibialis anterior, rectus femoris, rectus abdominis) show an increase in the activity during motion towards the backward position and the dorsal muscles (soleus, biceps femoris, erector spinae) show an increase in activity during motion towards the forward position.
, 百拇医药
Changes in the muscle activity associated with release of the load (APAs) at four points during the sway cycle are illustrated in Fig. 3 for a representative subject. Similarly to Fig. 1, time zero corresponds to the initiation of load release, and APAs were quantified within a 100 ms time interval prior to this time, by subtracting the integrated activity at the same phase and over the same time interval during an unperturbed sway trail. Figure 3B and D shows changes in the activity of muscles when the load was released close to the neutral position during sway, while passing the neutral position during swaying in the backward and forward directions, respectively. Figure 3A and C shows EMGs when the load was released close to the extreme forward and backward positions, respectively, during sway. Note the anticipatory increase in the activity of ventral muscles (TA, RF, and to a lesser extent RA) when the load was released while moving towards the extreme backward position (Fig. 3B; VSMB) and while passing through this position (Fig. 3C; VSB). This was accompanied by an anticipatory drop in the activity of the dorsal muscles (SOL, BF and ES). In contrast, if the load was released while moving towards (Fig. 3D; VSMF) or passing through the extreme forward position (Fig. 3A; VSF), there was an anticipatory increase in the activity of the dorsal muscles and an anticipatory drop in the activity of the ventral muscles. These APAs also varied in different muscles across subjects.
, 百拇医药
The load was released in the forward (A: VSF), middle when moving backward (B: VSMB), backward (C: VSB) and middle when moving forward (D: VSMF) positions. Time zero corresponds to time of load release and the vertical lines indicate the 100 ms time interval before load release over which indices of anticipatory postural activity (APA) were quantified. EMG signals for soleus, biceps femoris and erector spinae are inverted for better visualization.
Figure 4 shows smoothed IEMG indices (see eqn (2) in the Methods) plotted as a function of phase of the sway cycle for a representative subject. Recall that 0 and 100% correspond to the most forward position and 50% to the most backward position. A third-order polynomial was fitted to the data in Fig. 4 and the squared regression coefficient (R2) values are presented. In general, the ankle (TA, SOL) and knee muscles (RF, BF) showed similar directions of changes in APAs over the sway cycle. The anterior muscles, TA and RF, showed an increase in APAs in the mid-backward position (passing through the middle position while moving backwards, 25% of the cycle), reaching the highest values at the extreme backward position (50% of the cycle) followed by a drop in APAs in the mid-forward position. (passing through the middle position while moving forward, 75% of the cycle) and the smallest APAs in the extreme forward positions (0/100% of the cycle). The posterior muscles, SOL and BF, showed out-of-phase APA changes with respect to APA changes in TA and RF. The trunk muscles, RA and ES, showed a different pattern of activity. Both RA and ES showed larger APAs in the mid-forward and mid-backward positions, and smaller APAs at the extreme positions.
, 百拇医药
Ventral muscles: TA, tibialis anterior; RF, rectus femoris; RA, rectus abdominis. Dorsal muscles: SOL, soleus; BF, biceps femoris; ES, erector spinae. 0 and 100%, extreme forward position; 50%, extreme backward position; 25%, middle position when moving backward; and 75%, middle position when moving forward. Data were smoothed with a 5-point moving average window and fitted with a third-order polynomial. R2 values are presented.
Effect of task on APAs in extreme body positions
, 百拇医药
Figure 5 shows IEMGs of individual muscles averaged across subjects with standard error bars for the QS and VS tasks in the extreme forward and backward positions. Note that the direction of APAs in the forward and backward positions is opposite in all the muscles. This effect was small for the QS task, but was more striking in the VS task. The lower leg and thigh muscles showed significant main effects of TASK (QS and VS) and POSITION (extreme forward and extreme backward; both main effects Wilks' lambda < 0.4; F2,7 > 6; P < 0.05). SOL and RF muscles, in particular showed significantly larger APAs in the VS condition as compared to QS (P < 0.05). The ventral muscles (TA and RF) showed an increase in IEMGs in the backward position (P < 0.05), while the dorsal muscles (particularly SOL) showed an increase in IEMGs in the forward position (P < 0.05; see also Figs 1 and 3). There was a significant TASK–POSITION interaction effect for the TA/SOL muscle pair (Wilks' lambda < 3; F2,7 > 10; P < 0.01), while this interaction effect was just below significance for the RF/BF pair (P = 0.064). Even though the trunk muscles, RA and ES, showed a difference in the magnitudes of APAs in the different phases of the VS cycle (see Fig. 4), these effects varied across subjects, and these muscles did not show any significant TASK or POSITION effects.
, 百拇医药
Average across subjects IEMG indices in the extreme forward (open columns) and backward (filled columns) positions are shown with standard error bars. Muscle names are abbreviated as in Fig. 4. Significant differences between the QS and VS tasks; * significant differences between the extreme forward and backward positions.
Effect of task on APAs in the middle body position
To analyse the effects of sway on APAs when the subjects were close to the middle position, IEMG indices for the APAs were compared across three conditions: (1) in the middle position during quiet stance (QSM), (2) while passing the middle position in the forward direction during VS (VSMF), and (3) while passing the middle position in the backward direction (VSMB). Averaged across subjects normalized IEMG indices are shown in Fig. 6 with standard error bars. Note the pronounced suppression of the muscle activity (negative values) in the dorsal muscles for QSM (see also Fig. 1). During voluntary sway, this suppression became less pronounced. In the ventral muscles, a tendency towards higher APAs was observed. MANOVA on IEMG over all pairs of muscles (TA/SOL, RF/BF and RA/ES) showed a significant main effect of CONDITION (QSM, VSMB and VSMF; Wilks' lambda < 0.2; F4,5 > 6; P < 0.05). Planned comparisons between the QSM and the two VS conditions revealed that APAs in SOL, BF, RA and ES were significantly larger in the VS conditions as compared to QS (F1,8 > 5; P < 0.05; indicated by in Fig. 6). Further, SOL and ES APAs were significantly higher during VSMF as compared to QSM. ES also showed significantly higher APAs during VSMF as compared to VSMB (indicated by * in Fig. 6).
, 百拇医药
Quiet stance middle position (QSM; filled columns), voluntary sway middle position when swaying backwards (VSMB; open columns) and when swaying forward (VSMF; hatched columns) are shown with standard error bars. Muscle names are abbreviated as in Fig. 4. Significant differences between the QS and VS tasks; * significant differences between VSMB and VSMF.
Discussion
The magnitude of anticipatory postural adjustments (APAs) is known to depend on the magnitude of the perturbation, its direction and postural stability (Nouillot et al. 1992; Aruin & Latash, 1995; Aruin et al. 1998; Dietz et al. 2000). In a series of studies, Forssberg and colleagues (Nashner & Forssberg, 1986; Forssberg & Hirschfeld, 1988; Hirschfeld & Forssberg, 1991) have shown that APA magnitude is modulated within the gait cycle. Their subjects were asked to respond to an audio signal that was presented in different phases of the step cycle with a strong pull on the handle. Lateral gastrocnemius and hamstring muscles were activated during APAs occurring in the early support phase whereas tibialis anterior and quadriceps muscles were activated when the pull was exerted just prior to and during the swing phase.
, 百拇医药
Interpretation of these findings is complicated, however, by the fact that the leg muscles were involved in two activities, locomotion and APAs. In particular, in the swing phase, the muscles of the leg could not contribute to postural stabilization since they had no contact with the support. It is hard therefore to assess the functional significance of the apparently reversed APAs seen in the swing phase. Our experiments have demonstrated reversals of APA action when both legs could contribute to postural stabilization in all phases of the sway cycle. APA reversals in such conditions represent a major novel finding of our study.
, 百拇医药
The role of postural stability: refining the APA definition
Typically, the magnitude of APAs decreases for both very stable conditions (e.g. standing while leaning against a wall, Friedli et al. 1984) and unstable conditions (e.g. standing on a narrow support area, Aruin et al. 1998). Our experiments confirmed observations by Aruin et al. (1998) that shifting the body weight forwards or backwards leads to a decrease in the magnitude of APAs associated with releasing a standard load from the extended arms.
, http://www.100md.com
Releasing a load held in the extended arms always tends to tilt the body backwards. Hence, in the most backward position in the sway cycle, this perturbation was more likely to push the body out of the limits of postural stability. In contrast, the same perturbation tended to move the body away from the dangerously close edge of stability when the subjects approached the most forward body posture.
The typical APA pattern associated with load release (Aruin et al. 1998; Fig. 1) always tends to tilt the trunk forward. When the body is close to the most forward position, this APA pattern would move the body closer to the limit of stability and, as such, might by itself violate the balance. In such a situation, the CNS reverses the direction of the APAs (Fig. 5) and moves the body away from the edge of the support area. Mechanical action of reversed APAs does not counteract the mechanical effects of the perturbation associated with the load release but acts in the same direction.
, 百拇医药
Traditionally APAs have been defined as changes in the patterns of muscle activity that produce forces and moments acting against the mechanical effect of an expected perturbation (Bouisset & Zattara, 1987; Friedli et al. 1988; Massion, 1992). Our experiments have shown that APAs can produce mechanical effects on posture acting together with the perturbation as long as such action is consistent with maximizing the safety of the posture. We therefore suggest that APAs could be defined as changes in the activity of postural muscles associated with a predictable perturbation that act to provide maximal safety of the postural task component. This definition is consistent with the findings of decreased APAs under postural threat or fear of falling (Adkin et al. 2002).
, 百拇医药
One view on APAs is that they result from an internal forward model that takes into account dynamic consequences of an expected perturbation and generates responses to counteract these consequences (Wing et al. 1997). However, we do not see how a forward-model-based APA would be able to account for the reversal of APA direction, which is based on safety concerns. An alternative interpretation is that predictive motor control mechanisms such as APAs are not based on computations of forces but ensure stability of balance using equilibrium-point control (Feldman & Latash, 2005); shifts in equilibrium position of the body during APAs may be based on a combination of factors, including closeness to the edge of the support.
, 百拇医药
The role of the magnitude of the perturbation
In our experiments, the actual perturbation associated with the load release could be expected to vary with the phase of the sway because of the modulation of inertial forces. This modulation was relatively small. Peak horizontal load acceleration during a sine sway at 2 Hz with the amplitude of load motion of about 0.5 m may be assessed as close to 0.5 m s–2. With respect to the ankle joint, the lever arm of the load was about 1 m. For the 3 kg load, the peak inertial torques at the ankle joints related to the load motion were of the order of 1.5 N m. The gravitational force acting on the load produced the moment about the ankle joint of about 20 N m. Hence, the motion-related modulation of the magnitude of the perturbation at the two extreme positions was under 10%.
, 百拇医药
Large differences in the APA magnitude were seen in the phases of moving through the middle posture in different directions (Fig. 6), particularly in the RA/ES muscle pair that has been viewed as providing a basic pattern of the APAs (Shiratori & Latash, 2000). At those phases, the body moved at the highest speed while the acceleration was close to zero. Hence, this modulation could not be related to the motion-related modulation of the magnitude of the inertial forces.
, http://www.100md.com
The observed differences in the APAs while moving through the mid-posture in opposite directions could be related to preparing a response while taking into account the fact that the body was moving towards one of the extreme positions. This makes it possible to relate changes in the APAs in all muscle pairs (cf. Fig. 5) to providing maximal safety for the postural task.
References
Adkin AL, Frank JS, Carpenter MG & Peysar GW (2002). Fear of falling modifies anticipatory postural control. Exp Brain Res 143, 160–170.
, 百拇医药
Aruin AS, Forrest WR & Latash ML (1998). Anticipatory postural adjustments in conditions of postural instability. Electroencephalogr Clin Neurophysiol 109, 350–359.
Aruin AS & Latash ML (1995). The role of motor action in anticipatory postural adjustments studied with self-induced and externally triggered perturbations. Exp Brain Res 106, 291–300.
Aruin AS & Latash ML (1996). Anticipatory postural adjustments during self-initiated perturbations of different magnitude triggered by a standard motor action. Electroencephalogr Clin Neurophysiol 101, 497–503.
, 百拇医药
Aruin AS & Shiratori T (2004). The effect of the amplitude of motor action on anticipatory postural adjustments. J Electromyogr Kinesiol 14, 455–462.
Belen'kii V, Gurfinkel VS & Pal'tsev YI (1967). Elements of control of voluntary movements. Biofizika 10, 135–141.
Bouisset S & Zattara M (1987). Biomechanical study of the programming of anticipatory postural adjustments associated with voluntary movement. J Biomech 20, 735–742.
, 百拇医药
Bouisset S & Zattara M (1990). Segmental movement as a perturbation to balance Facts and concepts. In Multiple Muscle Systems. Biomechanics and Movement Organization, ed. Winters J & Woo S-Y, pp. 498–506. Springer, New York.
Cordo PJ & Nashner LM (1982). Properties of postural adjustments associated with rapid arm movements. J Neurophysiol 47, 287–302.
De Wolf S, Slijper H & Latash ML (1998). Anticipatory postural adjustments during self-paced and reaction-time movements. Exp Brain Res 121, 7–19.
, 百拇医药
Dietz V, Kowalewski R, Nakazawa K & Colombo G (2000). Effects of changing stance conditions on anticipatory postural adjustment and reaction time to voluntary arm movement in humans. J Physiol 524, 617–627.
Feldman AG & Latash ML (2005). Testing hypotheses and the advancement of science: recent attempts to falsify the equilibrium point hypothesis. Exp Brain Res 161, 91–103.
Forssberg H & Hirschfeld H (1988). Phasic modulation of postural activation patterns during human walking. Prog Brain Res 76, 221–227.
, 百拇医药
Friedli WG, Cohen L, Hallett M, Stanhope S & Simon SR (1988). Postural adjustments associated with rapid voluntary arm movements. II. Biomechanical analysis. J Neurol Neurosurg Psychiatry 51, 232–243.
Friedli WG, Hallett M & Simon SR (1984). Postural adjustments associated with rapid voluntary arm movements I. Electromyographic data. J Neurol Neurosurg Psychiatry 47, 611–622.
Hayes KC (1982). Biomechanics of postural control. Exerc Sport Sci Rev 10, 363–339.
, http://www.100md.com
Hirschfeld H & Forssberg H (1991). Phase-dependent modulations of anticipatory postural activity during human locomotion. J Neurophysiol 66, 12–19.
Marsden CD, Merton PA & Morton HB (1979). Anticipatory postural responses in the human subject. J Physiol 275, 47P–48P.
Massion J (1992). Movement, posture and equilibrium: interaction and coordination. Prog Neurobiol 38, 35–56.
Nashner LM & Forssberg H (1986). Phase-dependent organization of postural adjustments associated with arm movements while walking. J Neurophysiol 55, 1382–1394.
, http://www.100md.com
Nouillot P, Bouisset S & Do MC (1992). Do fast voluntary movements necessitate anticipatory postural adjustments even if equilibrium is unstable Neurosci Lett 147, 1–4.
Shiratori T & Latash ML (2000). The roles of proximal and distal muscles in anticipatory postural adjustments under asymmetrical perturbations and during standing on rollerskates. Clin Neurophysiol 111, 613–623.
Slijper H & Latash ML (2000). The effects of instability and additional hand support on anticipatory postural adjustments in leg, trunk, and arm muscles during standing. Exp Brain Res 135, 81–93.
, 百拇医药
Slijper H & Latash ML (2004). The effects of muscle vibration on anticipatory postural adjustments. Brain Res 1015, 57–72.
Slijper H, Latash ML & Mordkoff JT (2002). Anticipatory postural adjustments under simple and choice reaction time conditions. Brain Res 924, 184–197.
Wing A, Flanagan RJ & Richardson J (1997). Anticipatory postural adjustments in stance and grip. Exp Brain Res 116, 122–130.
Winter DA, Patla AE, Prince F, Ishac M & Gielo-Perczak K (1998). Stiffness control of balance in quiet standing. J Neurophysiol 80, 1211–1221.
Winter DA, Prince F, Frank JS, Powell C & Zabjek KF (1996). Unified theory regarding A/P and M/L balance in quiet stance. J Neurophysiol 75, 2334–2343., http://www.100md.com(Vijaya Krishnamoorthy, an)