| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Department of Rehabilitation Medicine, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands; 2Gait and Posture Laboratory, Department of Kinesiology, University of Waterloo, Waterloo, Canada; and 3Sint Maartenskliniek Research Development & Education, Nijmegen, The Netherlands
Submitted 5 December 2006; accepted in final form 23 March 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
). Can we use such a model to understand how we ambulate over ground conditions that are constantly changing? An important part within the step cycle is the moment of touchdown, because at that point, one has to match the leg stiffness with the characteristics of the ground surface. If there is a mismatch, the CNS has to react.
The mismatch at the moment of touchdown can be caused by hitting the support surface earlier or later than expected. The involved processes have been studied most extensively during jumps or falls from various heights. The usual strategy is to activate support muscles before landing. Most stiffness at landing is caused by proactive, centrally organized anticipatory muscle activation, which is timed to the expected instant of touch down (Santello 2005). The anticipatory muscle activation is also scaled to the expected stiffness of the landing surface (Kamibayashi and Muro 2006). If the touchdown is earlier than expected, the leg is not sufficiently stiff and the leg yields (Verdini et al. 2006). This brief yield of the leg seems to be related to reduced muscle activity of the quadriceps and tibialis anterior muscles before foot contact during walking. When hopping on an unexpectedly stiff surface, the ankle and knee become more flexed after landing because the extensor activations before landing were insufficient (Moritz and Farley 2004). After foot contact, however, leg muscle activations occur with latencies of 68–188 ms, thereby restoring leg stiffness.
If foot contact occurs later than expected, reflex activity after foot contact enhances. In drop falls through a false platform, the postlanding reflex component was increased when the interval between the false and solid floor was 
50 ms (McDonagh and Duncan 2002). The short interval needed to induce the corrections indicated that the reactions might be mediated subcortically. Do similar mechanisms operate when an unexpected surface height is encountered during walking? When a cat steps in an unexpected hole, there is an absence of loading at the expected time. As a consequence, there is a short latency silencing of the extensor muscle activity (Gorassini et al. 1994). This supports the contention that load feedback provides reinforcing feedback during the stance phase of walking (Duysens and Pearson 1980; for review see Duysens et al. 2000). Grey et al. (2004) described results, which most closely resembled the cat data, with a decrease in extensor activation occurring shortly after the onset of unloading during midstance caused by an induced plantarflexion movement of the ankle. In humans, the importance of load feedback during gait has been emphasized by other authors as well (Dietz and Duysens 2000), but the corrective responses to an unexpected step down or sudden loss of ground support have only recently been examined.
Several studies have used a platform that dropped down after foot contact. In a study by Nakazawa et al. (2004), a platformdropped 1 cm after the vertical component of the subject's load had reached 60% of body weight. Marigold and Patla (2005) studied the response to a step on an unexpectedly compliant surface with a drop in ground support of 
2 cm. Furthermore, Nieuwenhuijzen et al. (2002) have studied steps on an inverting platform, which also lowered the ground support surface but in a medial-lateral direction only. All these studies found long-latency muscle responses (later than 100 ms) in a range of lower limb muscles that could stabilize the ankle. In contrast to the cat experiments, the individuals made contact with the support surface before and during the lowering of the surface, thereby providing afferent feedback. Furthermore, it is questionable whether a 1- to 2-cm drop in ground support is sufficient to compare with the cat "foot-in-hole" studies. Therefore the following questions were asked to identify the responses of healthy subjects to an unexpected change in support surface height during walking. 1) Does the absence of expected load feedback in an unexpected step-down trigger fast reflex activity? 2) Does stepping down unexpectedly result in enhanced postlanding muscle activity compared with stepping down expectedly? 3) Does an unexpectedly high support surface result in fast reflex activity to compensate for the premature onset of loading?
To answer these questions, subjects stepped down and onto a level surface, both with and without awareness of the surface height. The conditions can be characterized by three important time periods: 1) before expected foot contact, 2) between expected and actual foot contact, and 3) postlanding. In the case of an unexpected step-down, the subjects approach the surface without awareness of the surface condition (period 1). The absence of expected foot contact is likely to trigger an error signal prior to the actual foot contact (period 2), followed by a strong delayed foot contact, which may induce enhanced postlanding activity (period 3). To test this paradigm, the unexpected step-downs are studied under two contexts: first, with respect to unexpected level walking, and second, with respect to stepping down expectedly. In unexpected level walking, the subjects are not aware of the support surface condition. Therefore the surface is approached similarly as in the unexpected step-down condition. As a consequence, the preprogrammed anticipatory muscle activity before the moment of the expected foot contact is the same. This allows us to determine the exact onset latency and muscle synergy triggered by the absence of the expected load feedback to answer question 1). This comparison is an analog to the comparison used in the cat foot-in-hole experiments.
In the second comparison, the physical conditions of the task are the same (step-down), whereas the level of expectancy differs. This allows us to answer question 2); whether postlanding muscle activity (period 3) is enhanced as a consequence of the unexpectedness of the step-down. To answer question 3), unexpected level walking is compared with expected level walking. In unexpected level walking, subjects are expecting a step-down (period 1). Hence, their preparation of a step-down is interrupted by a foot contact that arrives earlier than expected. This allows us to study the effects of premature onset of loading on postlanding reflex activity (period 3).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Twelve young adults (5 male) participated in the study [age, 24 ± 3 (SD) yr; height, 1.76 ± 0.12 m; body weight, 67 ± 13 kg]. None of the subjects had a neurological or musculoskeletal disease that could affect the behavioral responses. The experiment was performed in accordance with the Declaration of Helsinki and approved by the regional ethics committee.
Experimental set-up and instrumentation
The experimental set-up is shown systematically in Fig. 1A. It consisted of a 2.6-m long wooden walkway ending on a gravity driven platform of 1 m2, supported at each corner by two electromagnets (Commissaris et al. 2002). By releasing the electromagnets, the platform could lower the ground support surface by 5.0 cm. Subjects were instructed to walk four steps, starting with their right foot, and come to a full stop with their feet placed next to each other. The third step, with the right foot, was the step on the platform. This step is referred to as ipsilateral (i). The step before the step on the platform, with the left foot, is referred to as contralateral (c; see Fig. 1A). All subjects wore a safety harness attached to the ceiling. Furthermore, subjects wore glasses that blocked the lower part of their visual field to deprive them from visual feedback of the platform position. During the trials, subjects were instructed to look at a marker on a wall straight ahead to control for head position. Step length and cadence were standardized by foot placement instructions and the use of a metronome. Subjects performed several practice trials to become familiar with the correct foot placement and cadence before data collection.
|
Protocol
The experimental definitions and protocol are schematically shown in Fig. 1C. The series of trials were divided in two categories: expected (E) and unexpected (U; see Fig. 1C, left, definitions). The U trials were preceded by the warning that the platform might be lowered 5 cm during each of the following trials. Strictly speaking, the lowered platform trials were not 100% unexpected, because the subjects were aware of the possibility of a lowered platform. Nevertheless, this terminology was used to distinguish these from the second group, the expected trials (E). In the latter trials, the subjects were told beforehand whether the platform was lowered or not. Hence we use the labels E and U in the sense of aware and unaware, but we prefer the expectancy label (E, U) because it better describes the anticipatory state of the subjects, which was the topic of this study. Furthermore, the trials received a second label because they were also divided in two categories based on the physical condition of the landing surface. The level trials were termed level (L), and the lowered platform trials were termed down (D). Hence, for example, an ED trial was a trial in which the subjects expected (E) the platform to be down (D).
Subjects started the protocol with 10 trials of expected level walking (EL; Fig. 1C, top right, protocol). Next, the platform was unexpectedly lowered nine times (UD, hence UD 1, UD 2, etc.) in a series in which each UD was followed by 8–10 trials of level walking (UL, hence UL 1, UL 2, etc.). The exact number of UL trials varied to prevent the subjects from being able to predict the next UD (Fig. 1C, 2nd line, protocol). In the last part of the experiment, the subjects performed 10 expected step-down trials (ED, Fig. 1C, 3rd line, protocol). All UL trials were intrinsically unexpected, because the subjects were always in anticipation of a possible down trial. However, the level of anticipation differed within the series of UL trials. The first UL trial after a UD (UL 1) was characterized by a strong expectation of another step-down. This induced a brief yield of the leg at the moment of foot contact, as was characterized by a short unloading phase. Hence, the UL 1 trials represent the level walking trials where subjects were most biased toward expecting a step-down. Therefore they were grouped separately for further analysis on the concomitant EMG responses. In the subsequent UL trials, the subjects gradually adapted, which was characterized by a decrease in the amplitude and duration of the unloading phase. During the UL 7 trials, the duration of the unloading phase was significantly decreased compared with the UL 1 trials, whereas it was still increased compared with the EL trials. Apparently, the walking pattern of the subjects had stabilized and was no longer strongly biased toward expecting a step-down. However, this walking pattern was still slightly different from expected level walking. Because of these two reasons, the trials UL 7 were selected as control trials to compare with the UD condition (UD-UL).
Data analysis
The EMG data of all subjects were normalized with respect to the maximum EMG that was measured during a UL 7 trial. Maximum EMG was determined for each muscle by calculating a moving average over a window of 30 ms. The responses of the UD trials were studied with regard to two conditions: UL and ED. For the first comparison, UD – UL, the data were aligned such that the EHC of the UD trials coincided with the AHC of the UL trials. For the second comparison, UD – ED, the data of both conditions were aligned at the moment of AHC.
The amplitude of the difference in EMG activity was calculated by a subtraction method. For each subject, the average normalized EMG data of the UL and ED trials (for the 1st and 2nd comparison, respectively) was subtracted from the average of the UD 2–9 trials. The subtracted data were divided in 29 bins of 10 ms, starting at EHC of the UD trials and AHC of the UL trials. The EHC of the UD trials was on average 90 ms before AHC. Afterward, the muscle activity per bin was averaged for all subjects.
The onset latency of the EMG responses was determined by visual inspection. For each subject and muscle, the mean of the UD trials was plotted together with the mean ± 2 SD of the UL condition. For graphic use only, the EMG trials were filtered (0-lag, 2nd-order Butterworth filter; low-pass, 50 Hz). Using a custom-made MATLAB program, onset latencies were determined by placing a cursor at the moment where the UD curve exceeded the mean ± 2 SD of the UL curve for a period of 30 ms. Afterward, group averages of the onset latencies were calculated for all muscles.
As appeared from analysis of the landing strategies, subjects occasionally landed flatfooted or on their forefoot instead of with their heel first in the UD condition. Nevertheless, the heel landing was the most commonly used strategy (72.5%). Therefore only heel landings were selected for further analysis.
Statistical analysis
To verify if the average EMG amplitudes of the 10-ms bins were significantly different from zero, multiple one-sample t-tests were performed. A one-way ANOVA was used to test whether the onset latencies differed across muscles and whether the onset latencies of the first trial differed from the onset latencies of the subsequent trials. An 
level of 0.05 was chosen for the level of significance.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
The reflex activity as a consequence of stepping down unexpectedly was studied by comparing the UD with the UL condition. Shown in the top left panel of Fig. 3 is the average MGi activity in the UD and UL condition for one subject. At the moment of EHC, the subjects' preprogrammed muscle activity was similar for both conditions. In contrast, a large increase in MGi activity could be observed well before AHC. To further explore this difference, the EMG responses of both conditions were subtracted from each other and averaged for the whole population. The absence of foot contact at EHC induced reflex activity in the ipsilateral antigravity muscles (MGi, RFi) and contralateral flexor muscles (TAc, BFc), well before AHC (Fig. 3, horizontal gray bars). Simultaneously, there was a slight but significant decrease in BFi activity. The decrease was the result of silencing of the BF activity that usually occurred soon after AHC, as was the case for the UL trials. The onset latencies ranged from 47 to 69 ms after EHC (Fig. 3, arrows), which corresponds to 21–43 ms before AHC. The onset latencies did not differ significantly from each other. Nevertheless, a general pattern was observed. The MGi, BFi, TAc, and BFc muscles showed the first responses at 47–57 ms after EHC, followed by the RFi at 69 ms after EHC. In most subjects, the RFc was activated as well, with an average onset of 69 ms after EHC. Because of a larger variability between the subjects, this response did not reach the level of significance for the whole group, except for the very late part of the response. The extra muscle activity continued for 200 ms in all ipsilateral muscles and to a lesser extent in the contralateral muscles.
|
|
To study the effects of expectancy on postlanding muscle activity, the UD trials were compared with the ED trials, in which the physical condition of the task was the same (step-down), whereas the level of expectancy differed (U vs. E). Three subjects had to be excluded from the ED condition because they accidentally had vision of the support surface by removing the blocking glasses. The UD versus ED comparison showed increased postlanding EMG activity, as is shown in Fig. 5. Increased activity was found in the extensor muscles on the ipsilateral side (MGi and RFi), as well as in the contralateral TAc and to a lesser extent BFc. This increased postlanding activity had durations of 70–180 ms (Fig. 5) starting already at the moment of AHC. In contrast, postlanding EMG activity in the UD versus UL comparison (Fig. 3) was observed in other muscles as well (TAi, BFi, RFc). This was presumably caused by the difference in the mechanical situation (step-down vs. level walking).
|
Because the EMG of the UD and ED condition differed from each other, changes might also be found in the underlying kinesiology, as will be examined next. The kinesiology of the UD and ED condition differed from each other with respect to certain aspects (Fig. 6, left column). In the UD trials, the ankle was slightly more plantar flexed before AHC, suggestive of a more flat-footed landing, which may have been caused by the increased MGi activity. After AHC, the ipsilateral ankle dorsiflexed more rapidly compared with the ED condition. Furthermore, the ipsilateral knee extended, instead of the knee flexion that was observed after AHC in the ED condition. This might be caused by the increased RFi activity before and after AHC. On the contralateral side, the knee rapidly flexed, probably because of the early activity of the contralateral TAc and BFc.
|
7 trials. Whether the results were different for the level walking trials that were most unexpected (UL 1) was studied next. Does an unexpectedly high support surface result in fast reflex activity to compensate for the premature onset of loading?
In the UL 1 trials, immediately after a UD trial, the subjects clearly anticipated another step down because they consistently showed an increased load rate during the first 30 ms of foot contact, followed by a brief yield in the vertical ground reaction force (vGRF). This is shown at the top of Fig. 7 for a typical subject, in which the average of the UL 1 trials was compared with the average of the expected level walking trials (EL).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Unexpected step-down triggers fast reflexive braking reaction
In the case of a lower than expected surface, the main finding was that EMG responses in various leg muscles occurred very soon after passing the level surface (47–69 ms) and well before the actual heel contact (21–43 ms). The muscle synergy involved a brief braking reaction, which was achieved through short-latency activations in a number of muscles. On the ipsilateral side, the MGi was the first muscle activated followed closely by the RFi. Simultaneously, the ongoing activity of the BFi was suppressed. This induced ankle plantar flexion and knee extension. On the contralateral side, the TAc and BFc were activated, resulting in a large knee flexion. The latencies were barely longer than the one observed in the ipsilateral MGi, indicating that the response had a bilateral nature. The observed synergy showed a remarkable resemblance with the one described by Hase and Stein (1998), with latencies of 150–200 ms, in an experiment in which subjects were required to walk and stop abruptly after they received a nonnociceptive electrical stimulation of the peroneal nerve above the ankle (resemblance schematically shown in Fig. 9, A and B). In this study, it is shown that a similar synergy can also be elicited as a natural response to an unexpected absence of ground support surface as soon as 47 ms after passing the level surface (EHC).
|
The question arises as to which pathways may have been responsible for delivering the error signal to the CNS that triggered the fast braking reaction. The trigger for this response must have been the absence of foot contact at the point of expected landing. This implies that there are very fast pathways signaling the absence of expected foot contact. Fast responses during perturbations of walking might be attributable to monosynaptic stretch reflexes. However, the most obvious responses in this study are seen in the MGi and TAc. These muscles were not lengthened but shortened because of the delayed heel contact and ankle plantarflexion on the ipsilateral side and forward propulsion of the body on the contralateral side. This makes involvement of stretch reflexes in the braking reaction not plausible.
Second, the reaction might be caused by the absence of the load-reinforcing reflex. For loading it is known from cat work that there are extensor reinforcing reflexes that appear preferentially during locomotion (Duysens and Pearson 1980; Duysens et al. 2000). When cats step in an unexpected hole, it was shown that there is a short latency (35 ms) decrease in the ongoing activity of the ankle extensor muscles (Gorassini et al. 1994). Later experiments by Hiebert and Pearson (1999) showed that it was possible to compensate for the EMG silence by stimulating extensor nerves, thereby providing evidence for the contention that the EMG silence was caused by absence of proprioceptive input from these extensor muscles. In this study, a suppression of the extensor muscles as a consequence of the absence of foot contact was not found. However, this experiment differed from the cat studies in several important aspects. First, there is no extensor activity around heel contact in humans during normal walking, which makes it impossible to show EMG silencing after passing the expected heel contact. Grey et al. (2004) were able to show short latency extensor muscle silencing during a rapid plantar flexion movement in humans, but this was done during mid-stance when the extensor muscles were active. A second important difference is that cats are quadrupeds, and therefore balance is much less threatened than in the biped human. Moreover, a different corrective response is needed. The cats stepped on a treadmill and hence had to continue stepping after the perturbation, which required a substantial activation of the flexor muscles. In contrast, these subjects stopped walking briefly after stepping on the lowered surface. Altogether, the absence of the extensor-reinforcing reflex is not likely to have caused the fast braking reaction that was observed.
Next, it might be argued that the load receptors on the contralateral side (the standing leg) experienced an increased load at the transition of stance to swing. Because of the forward propulsion and downward movement of the body and the delay of the ipsilateral foot contact and thus double support phase, the load underneath the forefoot of the contralateral leg increased and was lengthened with respect to a normal phase transition. Moreover, onset latencies of the ipsilateral and contralateral leg did not differ significantly from each other, indicating that the response might have originated from the contralateral leg as well.
Another contributor to the corrective response might have been the vestibular system, detecting the increased forward rotation of the body and the drop in surface height. It is unlikely, however, that the fastest responses were vestibularly induced, because the vestibulo-spinal tract requires 60 ms to reach the peripheral muscles (Britton et al. 1993). Nevertheless, the vestibular system might have played a role in the later parts of the corrective response.
Braking reaction as a result of an internal model comparison
In this study, the CNS expected strong sensory feedback at the moment of predicted foot contact. The absence of the expected feedback acted as a very powerful stimulus to initiate the fast braking reaction that was observed. These findings show great resemblance with the paradigms of several studies of reaching, grasping, and aiming (Blakemore et al. 1998; Wolpert and Miall 1996). These studies are in favor of a theoretical model in which there is a continuous comparison between the expected feedback and the real-time sensory feedback. An efference copy of the motor commend is used to predict the sensory consequences of the ongoing action. This is compared with the actual sensory feedback of the movement. A discrepancy between the two will lead to a fast adjustment of the motor command.
Several suggestions are made as to where this comparison is made: at a spinal, cerebellar, or transcortical level. In cat, the responses to the absence of expected load suggested spinal involvement. To determine the spinal contribution to the corrective response, Hiebert et al. (1994) studied the responses to stepping in an unexpected hole in spinal cat. The responses were delayed and of smaller amplitude, indicating that spinal pathways could generate at least part of the response, but that supraspinal input was necessary for a fast onset and an appropriately scaled response. The ability to generate complex reactions as a response to postural perturbations and loading was also shown in cats with a high level of decerebration (so-called "premammilary cats," Nichols et al. 1978), indicating that the presence of the brain stem or cerebellum in addition to the spinal pathways was sufficient to evoke these types of responses. In human infants, who are generally believed to have an immature corticospinal tract, and therefore reduced supraspinal input, widespread task, and phase-dependent reactions were found as well (Lam et al. 2003). After stroke, which damages the supraspinal structures, it was found that extensor muscles were still modulated based on loading (Marigold et al. 2004). Moreover, in humans with a spinal cord lesion, it was shown that the load compensating mechanisms are still operational (Dietz and Duysens 2000; Dietz et al. 2002). These studies indicate that a large proportion of the responses to postural perturbations in humans might be stored at a spinal level. Could supraspinal structures play a role as well?
The cerebellum is though to be an important structure involved in comparing the sensory reafference with the predicted sensory feedback (Wolpert et al. 1998). Animal and human studies have shown that cerebellar integrity is needed to respond in time and accurately to unexpected perturbations of arm movements (Shimansky et al. 2004; Timmann et al. 2000). However, the fast latencies that were observed in this study (47–69 ms) are too short for transcerebellar loops. Cerebellar cooling in monkeys has shown a decrease in the M2 amplitude (Hore and Villis 1984), which occurs at an average latency of 69 ms in human (Peterson et al. 1998). The presently observed latencies are also too short for transcortical loops. Evidence from sensory evoked potential (SEP) registration and magnetic evoked potential (MEP) stimulation showed that the transcortical pathway for the TA muscle is 70 ms (Nielsen et al. 1997).
In this study, the subjects were able to predict the moment at which the perturbation occurred (always at right heel contact). Therefore the cerebellum might have primed the spinal structures of the possibility of an upcoming perturbation, as previously suggested by Shimansky et al. (2004) for a different paradigm. This abolishes the need for a transcerebellar loop, thereby reducing the onset latency of the corrective response to spinal levels. The fast braking reaction observed in this study had the same onset latencies during the first compared with the subsequent trials. This suggests that the spinal structures were already correctly primed before the subjects had experienced the nature of the perturbation. Because the moment of heel contact is a very important phase within the step cycle, the cerebellum might always preprogram the spinal structures before the moment of foot contact with the expected sensory feedback. As a consequence, a fast modification can be generated if necessary. The actual sensory feedback is likely to be responsible for shaping the compensatory synergies, such that they result in appropriate responses to the given perturbation.
Future studies, with unexpected perturbations during different phases of the step cycle can shed light on the question of whether these fast latencies can be evoked during other phases of the step cycle as well. Moreover, the exact role of the cerebellum in this reaction might be clarified by testing the current paradigm with patients suffering from cerebellar damage.
Stepping down unexpectedly results in enhanced postlanding muscle activity
The postlanding muscle activity of stepping down unexpectedly was higher compared with stepping down expectedly. Apparently, the absence of expected foot contact in the unexpected step down triggered enhanced muscle activity, partly similar to what was observed in passing a false floor during jumping (McDonagh and Duncan 2002). The latter study found enhanced activation of the antigravity muscles (MG, RF, and BF) soon after impact (35–80 ms). In this study, enhanced postlanding activity was found in the ipsilateral MG and RF as well. However, the increased activity did not show sharp onsets at an equivalent latency. This discrepancy may have been caused by several factors. First, landing during walking is much smoother than after a jump. Furthermore, the priority in jumping is to damp the impact, whereas in the current setting, maintaining balance and rapidly terminate the forward propulsion of the body is more important. Therefore the postlanding reflex activity may have been concealed by the fast braking reaction that was already generated shortly after EHC.
Unexpected level walking triggers fast flexor response
In the unexpected level walking condition, the support surface was higher than expected. At the moment of foot contact, the limb was not yet prepared for impact as was shown by the increased plantar flexion and diminished activation of the TAi. As a consequence, the leg yielded, resulting in a brief unloading period at 40–78 ms after touch down. The unloading period was similar to the one observed by Verdini et al. (2006), who found it to be related with reduced muscle activity before touch down as well.
The yield of the leg coincided with reflex responses at 46–81 ms after foot contact. Apparently, the unexpected early foot contact acted as a very powerful stimulus to trigger fast reflexive responses. The muscle synergy consisted of an early activation of the ipsilateral BFi muscle (46 ms), followed by activations of the TAi, RFi, and MGc (63–66 ms) and BFc (81 ms). The onset latencies were slightly faster than those observed after hopping on an unexpectedly hard surface (68–188 ms; Moritz and Farley 2004). The reflexes in TAi, RFi, and BFi might be a result of stretch. Although not all of these muscles were stretched as a result of the early impact, it has been shown in catching that stretch can induce simultaneous reactions in the antagonists as well as the agonists, probably to stabilize the joints (Lacquaniti and Maioli 1987). Alternatively, the reflex responses may have been induced by the cutaneous receptors of the foot sole. The muscle synergy of both the ipsilateral and contralateral leg resembled muscle activations that were found during stumbling (Eng et al. 1994; Schillings et al. 2000; cf. Fig. 9, C and D). The ipsilateral responses generated an ankle and knee flexion, whereas the activity of the contralateral muscles (MGc, BFc) generated a knee extension that delayed the moment of contralateral toe-off. The reflexive bursts, particularly of the BF muscles, ended simultaneously with the end of the unloading phase, indicating that active reloading of the leg resumed immediately after the flexion response had ended.
In conclusion, a difference between expected and actual loading at foot contact can trigger fast functionally relevant corrective muscle synergies, both on an unexpectedly high and low support surface. The muscle synergies and kinematic responses depend strongly on the perturbation and are almost the exact opposite of each other, which can be seen when comparing Fig. 9, A and C. However, the onset latencies are very similar. This finding is in favor of the presence of an internal model, continuously comparing the expected with the actual sensory feedback. The resulting error signal is able to trigger fast responses, which are appropriately coordinated to ensure stability in the light of very different perturbations.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: J. Duysens, Dept. of Rehabilitation Medicine, 898, Radboud Univ. Nijmegen Medical Centre, PO Box 9101, 6500 HB Nijmegen, The Netherlands (E-mail: j.duysens{at}reval.umcn.nl)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Britton TC, Day BL, Brown P, Rothwell JC, Thompson PD, Marsden CD. Postural electromyographic responses in the arm and leg following galvanic vestibular stimulation in man. Exp Brain Res 94: 143–151, 1993.[ISI][Medline]
Commissaris DA, Nieuwenhuijzen PH, Overeem S, de Vos A, Duysens JE, Bloem BR. Dynamic posturography using a new movable multidirectional platform driven by gravity. J Neurosci Methods 113: 73–84, 2002.[CrossRef][ISI][Medline]
Dietz V, Duysens J. Significance of load receptor input during locomotion: a review. Gait Posture 11: 102–110, 2000.[CrossRef][ISI][Medline]
Dietz V, Müller R, Colombo G. Locomotor activity in spinal man: significance of afferent input from joint and load receptors. Brain 125: 2626–2634, 2002.
Drew T, Dubuc R, Rossignol S. Discharge patterns of reticulospinal and other reticular neurons in chronic, unrestrained cats walking on a treadmill. J Neurophysiol 55: 375–401, 1986.
Duysens J, Clarac F, Cruse H. Load-regulating mechanisms in gait and posture: comparative aspects. Physiol Rev 80: 83–133, 2000.
Duysens J, Pearson KG. Inhibition of flexor burst generation by loading ankle extensor muscles in walking cats. Brain Res 187: 321–332, 1980.[CrossRef][ISI][Medline]
Duysens J, Trippel M, Horstmann GA, Dietz V. Gating and reversal of reflexes in ankle muscles during human walking. Exp Brain Res 82: 351–358, 1990.[ISI][Medline]
Eng JJ, Winter DA, Patla AE. Strategies for recovery from a trip in early and late swing during human walking. Exp Brain Res 102: 339–349, 1994.[ISI][Medline]
Gorassini MA, Prochazka A, Hiebert GW, Gauthier MJ. Corrective responses to loss of ground support during walking. I. Intact cats. J Neurophysiol 71: 603–610, 1994.
Grey MJ, Mazzaro N, Nielsen JB, Sinkjaer T. Ankle extensor proprioceptors contribute to the enhancement of the soleus EMG during the stance phase of human walking. Can J Physiol Pharmacol 82: 610–616, 2004.[CrossRef][ISI][Medline]
Grüneberg C, Nieuwenhuijzen PH, Duysens J. Reflex responses in the lower leg following landing impact on an inverting and non-inverting platform. J Physiol 550: 9859–9893, 2003.
Hase K, Stein RB. Analysis of rapid stopping during human walking. J Neurophysiol 80: 255–261, 1998.
Hiebert GW, Gorassini MA, Jiang W, Prochazka A, Pearson KG. Corrective responses to loss of ground support during walking. II. Comparison of intact and chronic spinal cats. J Neurophysiol 71: 611–622, 1994.
Hiebert GW, Pearson KG. Contribution of sensory feedback to the generation of extensor activity during walking in the decerebrate cat. J Neurophysiol 81: 758–770, 1999.
Hore J, Villis T. Loss of set in muscle responses to limb perturbations during cerebellar dysfunction. J Neurophysiol 51: 1137–1148, 1984.
Kamibayashi K, Muro M. Modulation of pre-programmed muscle activation and stretch reflex to changes of contact surface and visual input during movement to absorb impact. J Electromyogr Kinesiol 16: 432–439, 2006.[CrossRef][ISI][Medline]
Lacquaniti F, Maioli C. Anticipatory and reflex coactivation of antagonist muscles in catching. Brain Res 406: 373–378, 1987.[CrossRef][ISI][Medline]
Lam T, Wolstenholme C, van der Linden M, Pang MY, Yang JF. Stumbling corrective responses during treadmill-elicited stepping in human infants. J Physiol 553: 319–331, 2003.
Marigold DS, Eng JJ, Inglis JT. Modulation of ankle muscle postural reflexes in stroke: influence of weight-bearing load. Clin Neurophysiol 115: 2789–2797, 2004.[CrossRef][ISI][Medline]
Marigold DS, Patla AE. Adapting locomotion to different surface compliances: neuromuscular responses and changes in movement dynamics. J Neurophysiol 94: 1733–1750, 2005.
Marigold DS, Patla AE. Strategies for dynamic stability during locomotion on a slippery surface: effects of prior experience and knowledge. J Neurophysiol 88: 339–353, 2002.
McDonagh MJ, Duncan A. Interaction of pre-programmed control and natural stretch reflexes in human landing movements. J Physiol 544: 985–994, 2002.
Moritz CT, Farley CT. Passive dynamics change leg mechanics for an unexpected surface during human hopping. J Appl Physiol 97: 1313–1322, 2004.
Nakazawa K, Kawashima N, Akai M, Yano H. On the reflex coactivation of ankle flexor and extensor muscles induced by a sudden drop of support surface during walking in humans. J Appl Physiol 96: 604–611, 2004.
Nichols TR, Stein RB, Bawa P. Spinal reflexes as a basis for tremor in the premammillary cat. Can J Physiol Pharmacol 56: 375–383, 1978.[ISI][Medline]
Nielsen J, Petersen N, Fedirchuk B. Evidence suggesting a transcortical pathway from cutaneous foot afferents to tibialis anterior motoneurones in man. J Physiol 2: 473–484, 1997.
Nieuwenhuijzen PH, Grüneberg G, Van Galen GP, Duysens J. Modulation of the startle response during human gait. J Neurophysiol 84: 65–74, 2000.
Nieuwenhuijzen PHJA, Grüneberg C, Duysens J. Mechanically induced ankle inversion during human walking and jumping. J Neurosci Methods 117: 133–140, 2002.[CrossRef][ISI][Medline]
Petersen N, Christensen LOD, Morita H, Sinkjaer T, Nielsen J. Evidence that a transcortical pathway contributes to stretch reflexes in the tibialis anterior muscle in man. J Physiol 1: 267–276, 1998.
Santello M. Review of motor control mechanisms underlying impact absorption from falls. Gait Posture 21: 85–94, 2005.[CrossRef][ISI][Medline]
Schepens B, Delwaide PJ. Modifications of audio-spinal facilitation during gait in normal man. Electroencephalogr Clin Neurophysiol 97: 424–431, 1995.[CrossRef][Medline]
Schillings AM, van Wezel BM, Mulder T, Duysens J. Muscular responses and movement strategies during stumbling over obstacles. J Neurophysiol 83: 2093–2102, 2000.
Shimansky Y, Wang JJ, Bauer RA, Bracha V, Bloedel JR. On-line comensation for perturbations of a reaching movement is cerebellar dependent: support for the task dependency hypothesis. Exp Brain Res 155: 156–172, 2004.[CrossRef][ISI][Medline]
Tang PF, Woollacott MH, Chong RK. Control of reactive balance adjustments in perturbed human walking: roles of proximal and distal postural muscle activity. Exp Brain Res 119: 141–152, 1998.[CrossRef][ISI][Medline]
Timmann D, Richter S, Bestmann S, Kalveram KT, Konczak J. Predictive control of muscle responses to arm perturbations in cerebellar patients. J Neurol Neurosurg Psychiatry 69: 345–352, 2000.
Verdini F, Marcucci M, Benedetti MG, Leo T. Identification and characterisation of heel strike transient. Gait Posture 24: 77–84, 2006.[CrossRef][ISI][Medline]
Wolpert DM, Miall RC. Forward models for physiological motor control. Neural Netw 9: 1265–1279, 1996.[CrossRef][ISI][Medline]
Wolpert DM, Miall RC, Kawato M. Internal models in the cerebellum. Trends Cogn Sci 2: 338–347, 1998.[CrossRef][ISI]
Yang JF, Stein RB. Phase-dependent reflex reversal in human leg muscles during walking. J Neurophysiol 63: 1109–1117, 1990.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME |
