and biarticular muscles

502. G.J. uan Ingen Schenau et al. /&man Movement Science I3 (1994) 495-517 ..... Bethesda, MD: Geiger American Physiological Society. Bekoff, A., 1989.
2MB taille 17 téléchargements 242 vues
ELSEVIER

Human Movement Science 13 (1994) 495-517

Differential

use and control of mono- and biarticular muscles

Gerrit Jan van Ingen Schenau ‘,*, Carol A. Pratt b, Jane M. Macpherson

b

aDept. of Functional Anatomy, Faculty of Human Movement Sciences, Free University, van der Boechorststraat 9, 1081 BT Amsterdam, The Netherlands b R.S. Dow Neurological Sciences Institute, Legacy Good Samaritan Hospital and Medical Center, Portland, OR, USA

Abstract Most purposive movements are multi-joint actions in which specific patterns of joint moments are necessary to achieve the required segmental accelerations and to exert the appropriate direction and magnitude of force on the environment. Both human and animal studies suggest that mono- and biarticular muscles have different roles in these complex movements. Monoarticular muscles appear to show simple flexor or extensor activation patterns closely related to the required joint displacements, whereas biarticular muscles often exhibit complex multiple bursts of activity aimed at a fine-regulation of the distribution of net moments over the joints of the limb. Different patterns of activation of monoand biarticular muscles have been observed during natural movements and in response to sudden perturbations or peripheral afferent inputs. Evidence is provided that the control of mono- and biarticular muscles is based on different processes and different sources of information. Monoarticular muscles may be more rigidly grouped into simple flexor or extensor synergies, whereas biarticular muscle activity may be more flexibly sculpted by motion-related feedback from peripheral afferents. Little is known about where and how the integration between perceptual information and the control of net joint moments is realized.

as in a commonwealth; when order is once established in it there is no need for a separate monarch to preside over each separate task.” “

. . .

Aristotle

* Corresponding

(384-322

BC), De motu animalium

author.

0167-9457/94/$07.00

0 1994 Elsevier Science B.V. All rights reserved

SSDI 0167-9457(94)00022-7

(p. 703).

496

G.J. ran lngen Schenau et al. /Human

Mot~ement Science 13 (1994) 495-517

1. Introduction The relation between perception and action has received considerable attention ever since the time of Galen (131-201 AD), who identified the role of the nervous system, including afferent and efferent pathways, in movement control. Until the first half of the present century that attention was, and often still is, mainly focussed only on the motor, or efferent, system. Triggered by the convincing arguments of Bernstein (e.g. Bernstein, 1967) and his many co-workers (see Greene, 1972, for an overview of the “Moscow group”), contemporary scientists of various disciplines appear to be fascinated by the question of how the complex action system of biological organisms can become organized to produce functionally effective motor behaviors. Many of the theories on movement control forwarded in the past decades deal with possible strategies that the central nervous system may use to reduce the computational complexity of movement generation. However, experimental evidence for such theories is mainly based on a limited repertoire of tasks such as monoarticular movements and constrained reaching movements (for further details see Bizzi et al., 1992; Hogan and Winters, 1990; Macpherson, 1991). Many tasks in the real world involve multi-joint movements that require specific mechanical interactions with the environment (e.g. the ground or objects). The complexity of such tasks is illustrated by two problems: (i) the transformation of joint rotations into the desired translation of objects or of one’s own body (discussed previously, e.g. Van Ingen Schenau, 19891, and (ii> the controlled application of force against the environment, termed contact control tasks. The latter problem is illustrated by the difficulty that exists in stabilizing robotic devices during contact control tasks, as opposed to reaching and grasping tasks. Bizzi and colleagues (1992) refer to the observation by roboticists that these devices “... break into a pathologically uncontrollable chattering instability” in contact control tasks, a phenomenon referred to as “contact instability”. Unlike robots, biological organisms have little problem with such tasks. The purpose of this paper is to illustrate the specific requirements of contact control tasks and to discuss (and speculate about) possible strategies the nervous system may use in the perception-action coupling(s) necessary to organize these tasks. It is hypothesized that biarticular muscles may play a unique role in contact control tasks. The discussion will focus upon the possible role and neural control of one class of biarticular muscles. those that are bifunctional. Bifunctional biarticulars are defined

G.J. can Ingen Schenau et al. /Human

Movement Science 13 (1994) 495-517

497

as those muscles that exert opposite actions at the two spanned joints (i.e. flexion at one joint and extension at the other); whereas unifunctional biarticulars exert the same action at both joints (flexion/flexion, or extension/extension).

2. On the mechanics

of contact control tasks

Contemporary textbooks on functional anatomy suggest a strong relation between the action of a muscle and the degree of freedom it is supposed to control about the joint(s) crossed. In other words, it is suggested that for an extension one needs to activate extensors, for an abduction abductors, etc. Some simple examples will show that although this is generally true for a single-joint movement, the more realistic multi-joint case is considerably more complicated. Since Newton launched a formal definition of force as the cause of acceleration, the existence of a causal relation between muscle forces and the resulting movement is beyond dispute. The technique of “inverse dynamics”, introduced by Elftman in 1939, can be used to calculate the net moments at each joint as the sum of the moments of the force of all muscles and other structures that cross the joint. Nevertheless, such a powerful mechanical analysis cannot parse out the contributions of each individual muscle acting at a joint. Even when the pattern of muscle activation is known, it is difficult to assess the actions of individual muscles in multi-joint tasks (for examples, see Karst and Hasan, 1991a, b; Buchanan et al., 1986; 1989; Flanders and Soechting, 1990; Macpherson, 1991). Despite this “indeterminacy” problem (Bernstein, 19671, calculation of the net joint moments that are produced in contact control tasks has revealed that the required muscle actions are considerably more complicated than what might be predicted on the basis of gross muscle anatomy. As previously described (Van Ingen Schenau et al., 1992a; 1992b), the complexity of the mechanics of a contact control task can be easily demonstrated with the help of the simple arm task illustrated in Fig. 1A. Imagine that a subject seated at a horizontal table is asked to move a heavy object to the left along the indicated trajectory. Such a movement requires an extension about the elbow joint combined with a horizontal adduction about the shoulder (compare initial and final positions of the arm in Fig. 1A). However, inverse dynamic computation demonstrates that the generation of the external force, Fe, requires the combination of a net flexion moment at the elbow joint and a horizontal adduction moment at the

498

G.J. van Ingen Schenau et al. /Human

A

Mocement Science 13 (1994) 495-517

I3

Fig. 1. Contact control task. A: Subject (viewed from above) is required to move a heavy block from the initial position on the right (darker drawing) to the final position on the left (lighter drawing), with a force indicated by vector F,. The reaction force on the hand is indicated by F,. The required joint angle changes are extension at the elbow and adduction at the shoulder. The required joint moments are flexor at the elbow (M,,) and adductor at the shoulder (M,,). B: Same task as in A, but the table is tilted to add a force component due to gravity (Fs). The required joint angle changes are again, extension at the elbow and adduction at the shoulder. The required joint moments are now extensor at the elbow (Me*) and adductor at the shoulder (Ms2). The net force on the object (Fr) is the sum of the force exerted by the subject (F,) and the force due to gravity (F,).

shoulder. This can be clarified by noting that the reaction force on the hand, F,, has the effect of extending the lower arm and hand about the elbow; thus, this (quasi-static) situation can only be stable if the reaction force is opposed by a flexor moment about the elbow joint. If the subject were to activate the elbow extensors and shoulder adductors, according to the required joint angular displacements, the object would move not to the left, but rather toward the top of the diagram in Fig. 1A. This simple example illustrates clearly that it may be necessary to activate a flexor muscle in a task which requires extension of the same joint. This theoretical analysis is supported by several experimental observations of electromyographic (EMG) activity patterns during contact control tasks (Van Ingen Schenau, 1989; Van Ingen Schenau et al., 1992a; Gielen et al., 1990). Different directions of force exerted by the hand require different combinations of joint moments. As shown in Fig. lB, let us now imagine that the table with the heavy object is tilted in such a way that the subject also has to account for the influence of gravity (F,) on the object. In order to produce the resultant force, F,, necessary to move the object to the left, the subject now must exert a force on the object, F,, that is upward and to

G.J. Sian Ingen Schenau et al./Human

Movement Science 13 (1994) 495-517

499

the left. This force has a different magnitude and direction than in the task of Fig. lA, and requires a completely different distribution of the net moments over the shoulder and elbow joints. The joint displacements are the same as in Fig. lA, namely elbow extension and shoulder adduction. But now there is a small extension moment about the elbow combined with a horizontal adduction moment about the shoulder. These examples serve to illustrate that the required muscle actions in contact control tasks can by no means be predicted on the basis of the required changes in hand (and joint> positions, but depend strongly on the mechanical interaction with the environment. Although the actual relation between the magnitude and direction of external forces and the net joint moments necessary to realize these forces may be complicated by the additional factors of inertial and gravitational forces, it is obvious that in all types of movements the control of the external forces requires a specific adjustment of the distribution of net moments over the joints. This paradoxical conflict between the direction of joint motion and net joint moment also occurs during tasks in which the leg exerts force against an object or the ground, such as in postural control tasks (Macpherson, 1988a1, cycling Wan Ingen Schenau et al., 1992a), walking, sprinting, jumping, (Jacobs and Van Ingen Schenau, 1992a), and lifting (Toussaint et al., 1994). It has been observed in analyses of jumping, sprinting, and lifting in humans that the ground reaction force is always directed more or less toward the body center of gravity (see Fig. 2A), as long as gravity is the only other significant external force. This observation is associated with the requirement to preserve the angular momentum of the system (Van Ingen Schenau, 1989; Jacobs and Van Ingen Schenau, 1992a). Since gravity acts at the body’s center of mass, the resultant of the ground reaction force acting on the legs should have no net moment about the body’s center of mass in order to avoid unwanted rotations of the entire system. In such cases, the line of action of the ground reaction force will run mostly between the hip and knee joints in humans. For the quadruped, the situation for quiet stance is quite similar. Even though the line of action of the ground reaction force under each hind limb does not pass through the center of mass of the animal, it does pass close to the hip joint, and remains behind the knee (Lacquaniti et al., 1990; Fung et al., 1992). It is the resultant of the four ground reaction forces that passes through the center of mass, with no net moments about the center of mass under static conditions. The muscle activation pattern that is predicted by the kinematics of the limbs clearly corresponds to that based on the calculation of joint

500

G.J. ~banIngen Schenau et al. /Human

A

Mmement Science 13 (1994) 495-517

B

Fig. 2. A: Human subject exerting a force against the ground. Note that the ground reaction force vector passes close to, or through, the body center of gravity (beg), and behind the knee. 8: Quadruped hind limb during the stance phase of a gallop or jump. As in A, the ground reaction force passes near the body center of gravity, but in front of the knee.

moments (Fung et al., 1992). In contrast, during galloping or jumping in the quadruped, where the push-off of the hindlimbs is often performed without ground contact of the forelimbs, the line of action of the ground reaction force often must pass in front of the knee since the center of gravity is located about mid-way between forelimbs and hind limbs (Fig. 2B). This action requires a large extending moment in the hip combined with a flexing moment in the knee, as the knee is extending. A similar conflict between joint angular motion and net moment occurs in humans during cycling. This problem will be discussed in more detail in the following section. In particular, evidence for a role of bifunctional muscles in resolving these conflicts between angular displacement and net moment will be presented. 2.1. Coordination of monoarticular and bifunctional muscles: Co-contraction. The specific requirements for the distribution of net joint moments in contact control tasks were originally identified in an inverse dynamical analysis of human cycling Wan Ingen Schenau, 1989; Van Ingen Schenau et al., 1992a). Based on the simultaneous recording of the electromyograms of a number of monoarticular and bifunctional muscles in this and other leg extension tasks, it became clear that bifunctional muscles may play a unique role in these complex tasks. For cycling this is visualised in Fig. 3. In

G.J. c;an Ingen Schenau et al. /Human

Mooement Science 13 (1994) 495-517

501

Fig. 3. Muscle activation patterns during a cycling task. A: First part of leg extension: the force against the pedal (F,) is down and forward. Subjects use the monoarticular extensors, gluteus medius and the vasti, and the bifunctional rectus femoris. B: Second part of leg extension: the force against the pedal (F,) is down and backward. Subjects use the same monoarticular extensors as in A, but now use the bifunctional hamstrings. See text for details.

order to be able to do work on the pedal, cyclists appear to push the pedal slightly forward during the first part of leg extension (Fig. 3A) and slightly backward during the last part of the extension phase (Fig. 3B). This is accomplished through distinct changes in the net moments during leg extension. The hip extension moment is zero at a pedal position of top dead center, and increases to a large value at a pedal position of bottom dead center. The net moment in the knee, however, has a large positive (extension) value around top dead center, and decreases to negative (flexion) values at a position where the crank is about 130” from top dead center. These observations have been reported by many other groups as well (e.g. Ericson et al., 1986; Gregor et al., 198.5; Redfield and Hull, 1986). The upper leg muscle activity patterns found during the leg extension phase can be summarized as follows: the activity of the monoarticular hip and knee extensors is maintained throughout the extension phase, whereas the activity of the bifunctionals changes reciprocally. Rectus femoris decreases in activity and the hamstrings simultaneously increase. Thus, there are two distinct phases of co-activation of antagonists. During the first phase of leg extension the monoarticular gluteus maximus is co-active with the bifunctional rectus femoris (Fig. 3A), whereas during the last phase the

502

G.J. uan Ingen Schenau et al. /&man

Movement Science I3 (1994) 495-517

monoarticular vasti are co-active with the bifunctional hamstrings (Fig. 3B). These muscle patterns agree with those reported by other groups (Andrews, 1987; Ericson et al., 1986; Gregor et al., 1985; Redfield and Hull, 1986). The co-activation of vasti and hamstrings, in particular, is often described as uneconomical, and a non-optimal result of neuronal constraints (e.g. Andrews, 1987; Gregor et al., 1985; Ong et al., 1990). In contrast, we believe that the co-activation of these apparent antagonists should be judged as a highly effective and efficient solution for the specific requirements of contact control tasks. Let us consider the consequences of not activating biarticular muscles. In the limb position shown in Fig. 3A, the force exerted on the pedal (F,) is achieved by a small (almost zero) moment about the hip combined with a large extending moment about the knee. With monoarticulars alone, such joint moments can only be accomplished by reducing the activity of gluteus maximus and maintaining the activity of the vasti. The consequences are that the gluteus maximus, which is in a position to shorten as the hip is extending and thus contribute to work, would have to remain passive since activation of this muscle would cause a less effective force direction on the pedal (more vertically downward). The activation of a bifunctional muscle, the rectus femoris, can solve this problem by reducing the net hip moment while allowing gluteus maximus to be active and contribute to work (work which appears computationally as extra knee joint work). Calculation of muscle length changes showed that the rectus femoris mainly shortens during this phase of hip and knee extension (Van Ingen Schenau et al., 1992a), thus no energy is lost in eccentric contractions. During the second phase of leg extension, a comparable solution is provided by the bifunctional hamstring muscles (Fig. 3B). If the limb had only monoarticular muscles, the vasti would have to be de-activated during this phase of cycling, since the force exerted against the pedal (F,) requires a flexion moment at the knee in combination with a large extension moment at the hip. This knee flexion moment could only be accomplished by an eccentric contraction of a monoarticular knee flexor. Such an action would be highly ineffective and inefficient since the vasti, which are shortening, would have to remain unemployed while energy is wasted in the flexor. Co-activation of vasti and hamstrings solves this problem since the vasti can contribute to work while the hamstrings are creating a net flexion moment at the knee. No energy is wasted through this co-activation of apparent antagonists, since the hamstrings do not appear to be lengthened significantly throughout this phase of pedalling.

G.J. van Ingen Schenau et al./Human

Movement Science 13 (1994) 495-517

503

Using cross-correlation techniques it was found that the phases of muscle shortening in the monoarticulars coincide completely with the phases in which these muscles are presumed to exert force (determined by accounting for the electromechanical delay between the EMG signal and the development of force in the muscle). Such correlations were not found for the bifunctionals. It was therefore hypothesized that bifunctional muscles are primarily responsible for the (fine) regulation of the distribution of net moments over the hip and knee joints in contact control tasks. This hypothesis was further supported by a recent study of static leg extensions in which subjects were required to exert force against the ground in various predetermined directions (Jacobs and Van Ingen Schenau, 1992b). The difference in activity between hamstrings and rectus femoris was strongly correlated with the difference in net moments at hip and knee (r = 0.97) during various joint positions and external force directions. The global differences between the actions of monoarticular and bifunctional muscles were observed in other leg tasks as well (Bobbert and Van Ingen Schenau, 1988; Jacobs and Van Ingen Schenau, 1992a). These findings suggest that the monoarticular muscles are primarily responsible for the generation of force and work, whereas the bifunctional muscles control the direction of external forces by regulating the distribution of the net moments across the joints. The question arises as to the generality of this apparent functional distinction between monoarticular and bifunctional muscles as an organizing principle within the CNS. Experiments analyzing postural control in the quadrupedal cat indicate that a similar mechanism may play a role in the ground reaction forces that are produced as the animal maintains balance during movements of the support surface (Macpherson, 1988a, b). Although a complete inverse dynamical study is not yet complete, the pattern of EMG activity indicates that bifunctional muscles may play an important role in determining the direction of the ground reaction force. Cats were trained to stand on a supporting surface with each paw on a force plate, and were subjected to small linear translations of the support in each of 16 directions in the horizontal plane. In order to maintain balance, the animal had to produce a net force against the support that accelerated its body center of mass in the direction of the displacement. EMG responses were typically evoked in the appropriate muscles at a latency of 40-60 ms. Fig. 4 illustrates the pattern of EMG activity evoked in the proximal muscles of the left hind limb in response to a translation either backward and to the left (back/left in Fig. 4A), or forward and to the right (forward/right in

504

G.J. uan Ingen Schenau et al. /Human

A. Back/left Translation

Mouement Science 13 (1994) 495-517

B. Forward/right Translation

RF SAa

-

AX

Ax

Fig. 4. Response of the cat left hind limb to small linear translation of the support surface. A: In response to a translation back and to the left, the sagittal plane component of the ground reaction force (F,) is backward and upward. Activity is evoked in the hip flexor iliopsoas (IP), and the bifunctional muscles rectus femoris (RF), anterior sartorius (SAa), and semitendinosus (ST). B: In response to a translation forward and to the right, the sagittal plane component of the ground reaction force (F,) is forward and upward. Activity is evoked in the monoarticular vasti, and the bifunctional anterior (BFa) and medial (BFm) biceps femoris.

Fig. 4B); note that only the sagittal plane force components are shown. In response to the back/left translation, both hip and knee joints appeared to flex slightly, while the net moment was flexor at the hip and extensor at the knee (Fig. 4A). The hip flexor, iliopsoas was activated, along with the bifunctionals rectus femoris and anterior sartorius (hip flexor and knee extensor), as well as semitendinosus (hip extensor and knee flexor). It is likely that the rectus and sartorius provided the extensor moment at the knee and contributed to the flexor moment at the hip. Semitendinosus probably played a role in allowing the knee to flex. One would predict that the hip extensor moment of semitendinosus would be offset by the hip flexor moment from rectus, sartorius, and iliopsoas. The response to a forward/right translation (Fig. 4B) is perhaps clearer than the previous example. The angular displacement at hip and knee appeared to be extension while the net moment was extensor at the hip, and flexor at the knee. The monoarticulars gluteus medius (hip extensor) and vasti (knee extensors) were recruited during the postural response, in addition to the

G.J. can Ingen Schenau et al. /Human

Movement Science 13 (1994) 495-517

505

bifunctionals anterior and medial biceps femoris (hip extensor, knee flexor) (see Chanaud and Macpherson 1991). It is likely that the vasti produced work during shortening as the knee extended, whereas the biceps contributed to the flexor moment at the knee. These data are consistent with the hypothesis that bifunctional muscles play a major role in the control of the direction of external forces, even in the short latency responses evoked by postural disturbances. As argued from the examples above, co-activation of monoarticular agonists and bifunctional antagonists occurs in most multi-joint movements and can be judged as a highly effective solution to the problems associated with the necessity to control external forces. From a mechanical point of view it would be inefficient to activate monoarticular antagonists in tasks which require powerful joint extensions, although during the learning of new skills such co-activations may be useful in order to limit the number of degrees of freedom (e.g. Vereijken et al., 1992). It would be interesting to investigate whether powerful eccentric contractions ever occur in skilled subjects (apart from tasks that require the absorption of mechanical energy). Co-activation of antagonists is often seen when the joint is mainly responsible for maintaining the stability of the system, as is the case for the ankle joint in speed skating. Koning et al. (1991) found a distinct co-activation of soleus and tibialis anterior in a speed skating push-off (which requires a suppression of an explosive plantar flexion). In unrestrained sprinting, however, tibialis anterior activity is largely absent during the (powerful) plantar flexion (Jacobs and Van Ingen Schenau, 1992a).

3. Differential

control of monoarticular

and biarticular

muscles

Animal studies have demonstrated that the activity of bifunctional muscles is intimately associated with the distribution of intersegmental forces, as recently shown in a series of studies by Smith and her colleagues of the activity of semitendinosus (ST) during treadmill locomotion (Buford and Smith, 1990; Smith and Zernicke, 1987; Smith et al., 1993). As shown in Fig. 5A, the cat ST, a hip extensor and knee flexor, typically exhibits two bursts of activity during walking on a treadmill (Buford and Smith, 1990; Chanaud et al., 1991; Pratt et al., 1991; Smith et al., 1993) or over ground (Engberg and Lundberg, 1969; Rasmussen et al., 1978; English and Weeks, 1987). A flexor-related burst, which occurs when the paw leaves the ground

506

G.J. L‘UIE Ingen Schenau et al. /Human

Mol;ement

Science I3 (1994) 495-517

(paw off), most likely contributes to knee flexion at the onset of the swing phase. A second, extensor-related burst occurs just before the end of the stance phase and may be involved in decelerating hip flexion and knee extension prior to paw contact. Interestingly, the two ST bursts vary in their amplitude and consistency during different speeds and forms of locomotion, and recent evidence suggests that the relative strength of the two ST bursts is dependent on gait-related intersegmental dynamics within the hindlimb. During walking, the amplitude of both bursts is speed-related, but the paw-contact burst is always significantly smaller than the paw-off burst, and can even be absent at the slowest walking speeds (Chanaud et al., 1991; Engberg and Lundberg, 1969; Smith et al., 1993). In contrast, during the gallop, the paw-contact burst is larger and less variable than the paw-off burst (Smith et al., 1993). The shift in the relative strength of the flexor- and extensor-related ST bursts can be explained on the basis of differences in gait-associated kinetics. Unique to the gallop is an inertial torque related to hip linear acceleration (Wisleder et al., 19901, that is probably associated with large extensions of the lumbar spine at the end of stance (Goslow et al., 19731, and that flexes the knee, thereby reducing the need for a knee flexor (ST) muscle activation. The gait-related alterations in ST activity suggest that bifunctional muscle activity is finely tuned to the inertial interactions associated with multi-articular movements. This may be an effective strategy by which the neuro-muscular system can take full advantage of inertial forces to reduce unnecessary muscle activations, especially energetically wasteful lengthening contractions. Many studies have shown that bifunctional muscles characteristically have more complicated and task-dependent behavior than unifunctional muscles. Examples include studies comparing locomotion, jumping, scratching and paw shaking in intact cats (Abraham and Loeb, 1985; Carlson-Kuhta and Smith, 1990; Hoy and Zernicke, 1985; Hoy et al., 1985; Loeb et al., 1985; Pratt and Loeb, 1991; Pratt et al., 1991; Prochazka et al., 1977; Prochazka et al., 1989; Smith et al., 1977; Smith et al., 1980; Smith et al., 1985; Smith and Zernicke, 1987) and hatching, swimming, and walking in unrestrained chicks (Bekoff, 1989; Bekoff et al., 1987; Johnston and Bekoff, 1989). The greater flexibility in bifunctional muscle activity is consistent with the idea that the recruitment of bifunctional muscles is influenced more by motion-related feedback from the periphery than by central programs. Results from recent studies of forward and backward treadmill walking in Smith’s laboratory support this proposal (Buford et al., 1990;

G.J. uan lngen Schenau et al. /Human

Movement Science 13 (1994) 495-517

507

Buford and Smith, 1990; Perell et al., 1993; Pratt et al., 1992). Unlike forward walking in which the hip, knee, and ankle flex and extend synergically during the step cycle, during backward walking there is a mixed synergy of hip flexion combined with knee and ankle extension during the stance phase (Buford et al., 1990). Despite the different hindlimb kinematics (Buford and Smith, 1990) and kinetics (Perell et al., 19931 associated with the two forms of walking, the activation patterns of monoarticular (Buford and Smith, 1990) and unifunctional (i.e., same action at each joint) biarticular muscles (Pratt et al., 19921 were similar during forward and backward walking. However, all bifunctional muscles studied thus far were found to have different activation patterns during forward and backward walking (Pratt et al., 1992). These results suggest that unifunctional muscles, whether they cross one or more joints, are organized into a basic flexor/extensor synergy by the spinal locomotor pattern-generating network (Buford and Smith, 1990; Perret and Cabelguen, 1976; Perret and Cabelguen, 1980; Pratt, 1992; Pratt et al., 1991; Pratt et al., 1992); this basic synergy is used in both forward and backward walking. The adaptation of activity in bifunctional muscles to altered limb dynamics, along with changes in posture (Buford et al., 19901, determines the direction of progression, forward or backward. The observation that bifunctional muscles have more flexible activation patterns that appear to be closely linked to intersegmental dynamics and external forces suggests that motion-and force-related feedback plays a more critical role in the recruitment of bi- than unifunctional muscles. Indeed, it appears that normal locomotor activation patterns in bifunctional muscles require peripheral inputs and are not entirely centrally programmed. For example, in contrast to unifunctional muscles, the bifunctional ST only exhibits the normal two-burst pattern in the presence of limb movement and intact afferentation. In the absence of phasic inputs during fictive locomotion (locomotor activity evoked in paralyzed animals), ST motoneurons exhibit only a single, abnormally long flexor burst (Andersson and Grillner, 1981; Conway et al., 1987; Grillner and Zangger, 1984; Pearson and Rossignol, 1991; Perret and Cabelguen, 1976). In contrast, deafferentation moves the single ST burst from the flexor to the extensor phase of locomotion (Grillner and Zangger, 1984; Perret and Cabelguen, 1980). These results demonstrate that the activity of ST during locomotion is strongly dependent on peripheral inputs. Another indication that peripheral inputs are particularly important in shaping the activity of bifunctional muscles is their specialized relationship

G.J. cm Ingen Schrnau et al. /Human

508

BFa

e

SM

m--0

hkwement Science 13 (1994) 495-517

_

1.0

-0 -7 Stance

I

I

-30

0

I 100 ms

I

I

-30

0

I 100 ms

Fig. 5. A: Normal activation patterns of two bifunctional (semitendinosus, ST and semimembranosus, SM) and one unifunctional (anterior biceps, BFa) hip extensors during treadmill walking in the intact cat. ST and SM also flex the knee. Two normalized step cycles are shown with the stance phase indicated by the bar in the bottom trace. Both traces are displayed at the same gain. Note that both bifunctional muscles have some activity during both the stance and swing phases of the step cycle. B and C: Cutaneous reflexes evoked in BFa and SM by electrical stimulation (4 X threshold, 4T) of the saphenous (SAPH) cutaneous nerve displayed in peristimulus raster format. Each trace presents 130 ms of rectified and bin-integrated (2 ms bins) EMG activity, 30 ms before to 100 ms after the stimulus (time 0). The stimuli were delivered randomly at various times in the step cycle. The traces were re-ordered by the computer and are displayed according to the phase of the step cycle during which the stimulus occurred. All rasters are arranged with the onset of the swing phase at the top; the line along the bottom left border of each raster indicates the stance phase. The top bold traces in each raster display are a summary peri-stimulus histogram showing the reflex responses averaged over the period when the muscle was active (stance only for BFa, separate stance and swing averages for SM). Note the excitatory response in SM during swing (top histogram in 5C) and the inhibition during stance (bottom histogram in 50. Data are from Pratt et al. (1991).

with low-threshold cutaneous afferents. It is now known that most, if not all, segmental afferent systems have multiple parallel excitatory and inhibitory spinal pathways connecting them to motoneurons innervating 1952; Jankowska, hindlimb muscles (Bald&era et al., 1981; Hagbarth, 1992; McCrea, 1986, 1992). In a recent study of biarticular thigh muscles in intact cats (Pratt et al., 19911, electrical stimulation of low-threshold (non-noxious) hindlimb cutaneous afferents during treadmill locomotion typically evoked excitatory responses in unifunctional flexor muscles and inhibitory responses in unifunctional extensors. In contrast, bifunctional muscles exhibited not only combinations of excitation and inhibition, but also shifts from excitation to inhibition at different phases of the step cycle. An example of different connections between cutaneous afferents and unifunctional vs. bifunctional hip extensors is illustrated in Fig. 5. During treadmill walking, as shown in Fig. 5A, the bifunctional semimembranosus (SM), a hip extensor and knee flexor, was primarily active during stance but

G.J. van Ingen Schenau et al. /Human

Movement Science 13 (1994) 495-517

509

was also active during the swing phase. Note that the uniarticular hip extensor, anterior biceps (BFa), was only active during the stance phase. Stimulation of a cutaneous nerve (saphenous, SAPH) produced the classic extensor inhibition in both hip extensors during the stance phase of the step cycle (lower half of raster displays in Figs. 5B and Cl, but there was also a distinctive excitatory response evoked in SM during the swing phase (upper half of raster display and top peri-stimulus histogram in Fig. 50. In general, this study showed a close correspondence between the profiles of cutaneous reflex responses and the locomotor synergies recorded during locomotion. That is, unifunctional muscles had simple (flexor or extensor) activations both during locomotion and in response to cutaneous afferent stimulation. In contrast, bifunctional muscles often exhibited complex, multiple bursts of activity during a single step cycle and displayed complex mixtures of excitatory and inhibitory responses to low-threshold cutaneous stimuli. Recent evidence suggests that some proprioceptive inputs also may be differentially distributed to uni- and bifunctional muscles. Unlike spindle afferents, Golgi tendon organ (GTO) afferents appear to have more global actions since they can affect muscles acting at the same and distant joints (Eccles and Lundberg, 1959; Harrison and Jankowska, 198.5; Lundberg et al., 1975; McCrea, 1986). The widespread reflex connections of GTO afferents would seem to make them well-suited to coordinating intersegmental forces. Historically, GTO’s have been thought to produce reflex inhibition in the muscle of origin and its synergists (Eccles et al., 1957). However, more recently it has been shown that under certain conditions, including locomotion (Conway et al., 1987; Duysens and Pearson, 1980; Gossard et al., 1994; Guertin et al., 1993; Pearson and Collins, 1993; Pearson et al., 19921, quiet stance (Pratt et al., 19931, and in non-behaving, non-barbiturate-anesthetized cats (Nichols, 1989; Powers and Binder, 19851, extensor GTO afferents can produce reflex excitation in hindlimb extensors, i.e. a positive force feedback. There is some evidence that at the ankle this positive force feedback is preferentially distributed to bifunctional extensors. Nichols (1989) showed that stretch of a relaxed (unloaded) MG in a decerebrate cat produced the classic excitatory stretch reflex in its uniarticular synergist, soleus, and its bifunctional synergist, lateral gastrocnemius (LG). However, when the same stretch was applied to an active (loaded) MG, the soleus was inhibited, whereas LG continued to be excited. Nichols concluded that this selective reflex reversal could be due to the action of MG GTO afferents,

510

G.J. uan Ingen Schenau et al/Human

Mooement Science 13 (1994) 495-517

since it was load rather than length dependent, and since muscle spindle connections among synergists are excitatory. He also suggested that this could be a mechanism for “enhancing the relative contributions of biarticular muscles to resultant limb stiffness”; the inhibition of soleus would be functionally appropriate under these conditions because the actions of uniarticular muscles could impede rather than promote mechanical coupling among the joints (Nichols, 1989). Interestingly, the reflex effects from cutaneous afferents in the sural nerve, which innervates the skin over the heel and lateral surface of the foot (Hagbarth, 19521, are also differentially distributed to the bifunctional MG and the unifunctional soleus - MG is primarily excited whereas soleus is typically inhibited (Abraham et al., 1985; Aniss et al., 1992; LaBella et al., 1989), and the differential effects appear to be task-dependent (Duysens et al., 1991). Sural effects on LG are more variable and can be similar to those recorded in MG or soleus depending on the experimental preparation. The similar differential distribution of excitatory and inhibitory effects from sural and ankle extensor GTO afferents may contribute to the task-dependent, differential recruitment of the uniarticular (soleus) and bifunctional (MG and LG) heads of the triceps surae that has been observed in both humans and cats. During lengthening contractions in humans (Nardone and Schieppati, 1988) and during paw shaking in cats (Smith et al., 19801 there is a selective recruitment of MG and LG and not soleus. Likewise, during the preparatory period preceding a postural adjustment, rapid dorsiflexion of the ankle in standing humans facilitated MG activity but inhibited activity in soleus (Woollacott et al., 1984). In the cat during the rapid response to translation of the support surface, LG increased in activity whereas soleus was simultaneously inhibited for directions of translation in which the limb was loaded; both muscles showed simultaneous inhibition during unloading of the limb (Fig. 4 in Dunbar and Macpherson, 1993). Similar results were obtained in freely standing cats exposed to a sudden drop of the support surface under one paw (Pratt et al., 1993). As shown schematically in Fig. 6 and elsewhere (Rushmer et al., 19871, when the force plate under a cat’s paw is suddenly dropped, the unsupported limb and its diagonal partner are unloaded whereas the contralateral limb and its diagonal partner are loaded (increase in vertical force). Fig. 6 shows force (LH Fv) and EMG recordings from the left hind limb, which was loaded following a drop of the left forelimb (Pratt et al., 1993). There is a differential response of the soleus and LG to the drop -LG was excited, whereas the initial response in soleus was inhibition.

G.J. uan Ingen Schenau et al. /Human

Movement Science 13 (1994) 495-517

LF Drop

# r-

511

:

Time (ms) Fig. 6. Responses in two ankle extensors, lateral gastrocnemius (LG) and soleus (SOL), evoked by a vertical drop (40 mm, onset indicated by a dashed line) of the left forelimb (LF) in an intact standing cat. In the schematic at the top of the figure, the dropped limb is indicated by the arrow, and the limbs that are loaded (increase in vertical force) are indicated by shaded boxes. The top trace shows the increase in vertical force (LH Fv, upward deflection) exerted by the left hind limb. The bottom two traces show the EMGs recorded in the left hindlimb. Note that the initial response in soleus is inhibition. Data are from Pratt et al. (1993).

Because there was no background activity in LG, it cannot be determined conclusively that the excitation in LG was not also preceded by an initial inhibition. However, two observations suggest that the inhibition was preferentially directed to the soleus: (1) the onset of the excitatory response in LG occurs during the trough of inhibition in soleus; and (2) the onset and peak of the excitatory response occurred later in soleus than in LG. Furthermore, intracellular recordings of soleus motoneurons would be needed to determine whether the increased activity in soleus following the inhibition is indeed a real excitation or a post-inhibitory rebound. 4. Conclusion In summary, there is substantial evidence to indicate that bifunctional muscle activity is a critical element in the smooth execution of multiarticular movements. We have presented data indicating that bifunctional muscles have specialized connections with at least some populations of segmental afferents that could contribute to the precise matching of muscle activity with intersegmental dynamics. Though we presented evidence that proprioceptive and cutaneous afferents are important candidates, little is

512

G.J. r:an Ingen Schenau et al./Human

Movement Science 13 (1994) 495-517

known about what roles these or other afferents might play in monitoring limb dynamics and external forces. We are far from understanding how this information is perceived by central circuits and coupled to the skeletomotor action system to result in smoothly coordinated motor behaviors. Similarly, little is understood of how spinal interneuronal circuits and supra-spinal circuits might participate and cooperate in the perception-action couplings necessary to organise these complex tasks (Gandevia and Burke, 1992; McCrea, 1992). A better understanding of the role of sensory inputs in coordinating muscle activity with limb dynamics and external forces will require studies that map the actions of individual afferent systems onto the activations of muscles during natural, multi-articular movements. Although it is hypothesized here that the control of monoarticular and bifunctional muscles might be based on different parallel processes, it should be noted that this can only partly be the case since, from a mechanical point of view, such a separation is simply impossible. Even if activation of monoarticular muscles is based primarily on position information, nevertheless these muscles clearly influence the magnitude of the net moments about the joints that are crossed. Fine-regulation of the distribution of the net moments by bifunctional antagonists is not possible without knowledge regarding the activity of the monoarticulars. On the other hand, bifunctional muscles may influence the total work production since they often exhibit length changes. It remains to be determined how integration in the control of these different muscles and their differential functions is achieved. Clearly, all of these considerations, observations (and speculations) in favor of parallel mechanisms for the control of mono- and bifunctional muscles need considerably more attention before they can result in a sound theoretical framework. Moreover, this paper has dealt with only a few, albeit complex aspects of motor control. Nevertheless, the arguments presented here seem to justify advocating that future studies should more explicitly account for the possible differences in the organization of the control of mono- and bi- (or poly-) articular muscles. Acknowledgments CAP was supported by the National Science Foundation 9120818) and JMM by the National Institute of Neurological Stroke (grant NS29025).

(grant IBNDiseases and

G.J. LUGIngen Schenau

et al. /Human

Mocement

Science 13 (1994) 495-517

513

References Abraham L.D. and G.E. Loeb, 1985. The distal musculature of the cat: Patterns of normal use. Experimental Brain Research 58, 580-593. Abraham, L.D., W.M. Marks and G.E. Loeb, 1985. The distal hindlimb musculature of the cat. Cutaneous reflexes during locomotion. Experimental Brain Research 58, 594-603. Andersson, 0. and S. Grillner, 1981. Peripheral control of the cat’s step cycle. I. Phase dependent effects of ramp-movements of the hip during ‘fictive locomotion’. Acta Physiologica Scandinavia 113, 899101. Andrews, J.G., 1987. The functional roles of the hamstrings and quadriceps during cycling. Lombard’s paradox revisited. Journal of Biomechanics 20, 565-575. Aniss, A.M., SC. Gandevia and D. Burke, 1992. Reflex responses in active muscles elicited by stimulation of low-threshold afferents from the human foot. Journal of Neurophysiology 67, 1375-1384. Baldissera, F., H. Hultborn and M. Illert, 1981. ‘Integration in spinal neuronal systems’. In: J.M. Brookhart, V.B. Mountcastle and S.R. Brooks (Eds.), Handbook of physiology - The nervous system II (pp. 509-595). Bethesda, MD: Geiger American Physiological Society. Bekoff, A., 1989. ‘Continuity of pattern generating mechanisms in embryonic and posthatching chicks’. In: C. von Euler, H. Forssberg, C. von Hofsten, H. Lagerkranz and R. Zetterstrom (Eds.), Neurobiology of early infant behavior (pp.79-86). New York: Macmillan. Bekoff, A., M.P. Nusbaum, A.L. Sabichi and M. Clifford, 1987. Neural control of limb coordination. I. Comparison of hatching and walking motor output patterns in normal and deafferented chicks. Journal of Neuroscience 7, 2320-2330. Bernstein, N.A., 1967. The coordination and regulation of movements. London: Pergamon. Bizzi, E., N. Hogan, F.A. Mussa-Ivaldi and S. Giszter, 1992. Does the nervous system use equilibrium point control to guide single and multiple joint movements? Behavioral and Brain Sciences 15, 603-613. Bobbert, M.F. and G.J. van Ingen Schenau, 1988. Coordination in vertical jumping. Journal of Biomechanics 21, 2499262. Buchanan, T.S., D.P.J. Almdale, J.L. Lewis and W.Z. Rymer, 1986. Characteristics of synergic relations during isometric contractions of human elbow muscles. Journal of Neurophysiology 56, 1225-1241. Buchanan, T.S., G.P. Rovai and W.Z. Rymer, 1989. Strategies for muscle activation during isometric torque generation at the human elbow. Journal of Neurophysiology 62, 1201-1212. Buford, J.A. and J.L. Smith, 1990. Adaptive control for backward quadrupedal walking. II. Hindlimb muscle synergies. Journal of Neurophysiology 64, 7566766. Buford, J.A., R.F. Zernicke and J.L. Smith, 1990. Adaptive control for backward quadrupedal walking. I. Posture and hindlimb kinematics. Journal of Neurophysiology 64, 745-755. Carlson-Kuhta, P. and J.L. Smith, 1990. Scratch responses in normal cats: Hindlimb kinematics and muscle synergies. Journal of Neurophysiology 64, 1653-1667. Chanaud, C.M. and J.M. Macpherson, 1991. Functionally complex muscles of the cat hindlimb. III. Differential activation within biceps femoris during postural perturbations. Experimental Brain Research 85, 271-280. Chanaud, C.M., CA. Pratt and G.E. Loeb, 1991. Functionally complex muscles of the cat hindlimb. V. The roles of histochemical fiber-type regionalization and mechanical heterogeneity in differential muscle activation. Experimental Brain Research 85, 300-313. Conway, B.A., H. Hultborn and 0. Kiehn, 1987. Proprioceptive input resets central locomotor rhythm in the spinal cat. Experimental Brain Research 68, 643-656. Dunbar, D.C. and J.M. Macpherson, 1993. Activity of neuromuscular compartments in lateral gastrocnemius evoked by postural corrections during stance. Journal of Neurophysiology 70, 2337-2349.

514

G.J. can Ingen Schenau et ai./Human

Movement Science 13 (1994) 495-517

Duysens, J. and K.G. Pearson, 1980. Inhibition of flexor burst generation by loading ankle extensor muscles in walking cats. Brain Research 187, 321-332. Duysens, J., A.A.M. Tax, B. van der Doelen, M. Trippel and V. Dietz, 1991. Selective activation of human soleus or gastrocnemius in reflex responses during walking and running. Experimental Brain Research 87, 193-204. Eccles, J.C., R.M. Eccles and A. Lundberg, 1957. Synaptic actions in motoneurons caused by impulses in Golgi tendon organs. Journal of Physiology 138, 227-252. Eccles, R.M. and A. Lundberg, 1959. Synaptic action in motoneurones by afferents which may evoke the flexion reflex. Archives of Italian Biology 97, 199-221. Elftman, H., 1939. Forces and energy changes in the leg during walking. American Journal of Physiology 125, 339-356. Engberg, I. and A. Lundberg, 1969. An electromyographic analysis of muscular activity in the hindlimb of the cat during unrestrained locomotion. Acta Physiologica Scandinavia 75, 614-630. English, A.W. and 0.1. Weeks, 1987. An anatomical and functional analysis of cat biceps femoris and semitendinosus muscles. Journal of Morphology 191, 161-175. Ericson, M.O., A. Bratt, R. Nisell, U.P. Arborelius and J. Ekholm, 1986. Power output and work in different muscle groups during ergometer cycling. European Journal of Applied Physiology 55, 229-235. Flanders, M. and J.F. Soechting, 1990. Arm muscle activation for static forces in three-dimensional space. Journal of Neurophysiology 64, 1818-1837. Fung, J., P. Rougier, CA. Pratt and J.M. Macpherson, 1992. Influence of interlimb distance on the control of quiet stance in the cat. Neuroscience Abstracts 18, 1408. Gandevia, S.C. and D. Burke, 1992. Does the nervous system depend on kinesthetic information to control natural limb movements? Behavioral and Brain Sciences 15, 614-632. Gielen, C.C.A.M., G.J. van Ingen Schenau, A.A.M. Tax and M. Theeuwen, 1990. ‘The activation of mono- and biarticular muscles in multi-joint movements’. In: J.M. Winter and S.L.Y. Woo (Eds.), Multiple muscle systems (pp 302-311). New York: Springer-Verlag. Goslow, G.E. Jr., R.M. Reinking and D.G. Stuart, 1973. The cat step cycle: Hind limb joint angles and muscle lengths during unrestrained locomotion. Journal of Morphology 141, l-42. Gossard, J.-P., R.M. Brownstone, I. Barajon and H. H&born, 1994. Transmission in a locomotor-related group Ib pathway from hindlimb extensor muscles in the cat. Experimental Brain Research (in press). Greene, P.H., 1972. ‘Problems of organization of motor systems’. In: R. Rosen and F.M. Snell (Eds.), Progress in theoretical biology (pp. 303-338). New York: Academic Press. Gregor, R.J., P.R. Cavanagh and M. Lafortune, 1985. Knee flexor moments during propulsion in cycling - A creative solution to Lombard’s Paradox. Journal of Biomechanics 18, 307-316. Grillner. S. and P. Zangger, 1984. The effect of dorsal root transection on the efferent motor pattern in the cat’s hindlimb during locomotion. Acta Physiologica Scandinavia 120, 393-405. Guertin, P., M.J. Angel, M.-C. Perreault, P.A. Carr and D.A. McCrea, 1993. Ankle extensor group I afferents excite extensors throughout the hindlimb during MLR-evoked fictive locomotion. Neuroscience Abstracts 19, 142. Hagbarth, K.-E., 1952. Excitatory and inhibitory skin areas for flexor and extensor motoneurones. Acta Physiologica Scandinavia 94, l-58. Harrison, P. and E. Jankowski, 1985. Organization of input to the interneurones mediating group I non-reciprocal inhibition of motoneurones in the cat. Journal of Physiology 361, 403-418. Hogan, N. and J.M. Winters, 1990. ‘Principles underlying movement organisation: upper limb’. In: J.M. Winters and S.L.Y. Woo (Eds.), Multiple muscle systems (pp. 182-193). New York: Springer-Verlag. Hoy, M. and R.F. Zernicke, 1985. Modulation of limb dynamics in the swing phase of locomotion. Journal of Biomechanics 18, 49-60.

G.J. can Ingen Schenau et ai./Human

Mcwetnent Science 13 (1994) 495-517

515

Hoy, M.G., R.F. Zernicke and J.L. Smith, 1985. Contrasting roles of inertial and muscle movements at knee and ankle during paw-shake response. Journal of Neurophysiology 54, 1282-1294. Jacobs, R. and G.J. van Ingen Schenau, 1992a. Intermuscular coordination in a sprint push off. Journal of Biomechanics 25, 953-965. Jacobs, R. and G.J. van Ingen Schenau, 1992b. Control of an external force in leg extensions in humans. Journal of Physiology 457, 611-626. Jankowska, E., 1992. Interneuronal relay in spinal pathways from proprioceptors. Progress in Neurobiology 38, 335-37s. Johnston, R.M. and A. Bekoff, 1989. Differential modulation of double bursting muscles in the chick. Sot. Neuroscience Abstracts 15, 1044. Karst, G.M. and Z. Hasan, 1991a. Initiation rules for planar, two-joint arm movements - Agonist selection for movements throughout the work space. Journal of Neurophysiology 66, 1579-1593. Karst, G.M. and Z. Hasan, 1991b. Timing and magnitude of electromyographic activity for two-joint arm movements in different directions. Journal of Neurophysiology 66, 1594-1604. Koning, J.J., G. de Groot and G.J. van Ingen Schenau, 1991. Coordination of leg muscles in speed skating. Journal of Biomechanics 24, 137-146. LaBella, L.A., J.P. Kehler and D.A. McCrea, 1989. A differential synaptic input to the motor nuclei of triceps surae from the caudal and lateral cutaneous sural nerves. Journal of Neurophysiology 61, 291-301. Lacquaniti, F., M. le Taillanter, L. Lopiano and C. Maioli, 1990. The control of limb geometry in cat posture. Journal of Physiology 426, 177-192. Loeb, G.E., J.A. Hoffer and W.G. Marks, 198.5. Activity of spindle afferents from cat anterior thigh muscles. III. Effects of external stimuli. Journal of Neurophysiology 54, 578-591. Lundberg, A., K. Malmgren and E.D. Schomburg, 1975. Convergence from Ib, cutaneous and joint afferents in reflex pathways to motoneurones. Brain Research 87, 81-84. Macpherson, J.M., 1988a. Strategies that simplify the control of quadrupedal stance. I. Forces at the ground. Journal of Neurophysiology 60, 204-217. Macpherson, J.M., 1988b. Strategies that simplify the control of quadrupedal stance. II. Electromyographic activity. Journal of Neurophysiology 60, 218-231. Macpherson, J.M., 1991. ‘How flexible are muscle synergies?’ In: D.R. Humphrey and H.-J. Freund (Eds.), Freedom to move: Dissolving boundaries in motor control. Dahlem Konferenzen (pp. 33-47). Chichester: Wiley. McCrea, D.A., 1986. ‘Spinal cord circuitry and motor reflexes’. In: K.B. Pandolf (Ed.), Exercise and sport science review (pp. 105-141). New York: Macmillan. McCrea, D.A., 1992. Can sense be made of spinal interneuron circuits? Behavioral and Brain Sciences 15, 633-643. Nardone, A. and M. Schieppati, 1988. Shift of activity from slow to fast muscle during voluntary lengthening contractions of the triceps surae muscles in humans. Journal of Physiology 395, 363-381. Nichols, T.R., 1989. The organization of heterogenic reflexes among muscles crossing the ankle joint in the decerebrate cat. Journal of Physiology 410, 463-477. Ong, H.D., H. Hemami and S. Simon, 1990. ‘Simulation studies of musculo-skeletal dynamics in cycling and sitt ing on a chair’. In: J.M. Winter and S.L.Y. Woo (Eds), Multiple muscle systems (pp. 518-533). New York: Springer-Verlag. Pearson, K.G. and D.F. Collins, 1993. Reversal of the influence of group Ib afferents from plantaris on activity in medial gastrocnemius muscle during locomotor activity. Journal of Neurophysiology 70, 1009-1017. Pearson, K.G., J.M. Ramirez and W. Jiang, 1992. Entrainment of the locomotor rhythm by group Ib afferents from ankle extensor muscles in spinal cats. Experimental Brain Research 90, 557-566. Pearson, K.G. and S. Rossignol, 1991. Fictive motor patterns in chronic spinal cats. Journal of Neurophysiology 66, 1874-1887.

516

G.J. mn Ingen Schenau et al/Human

Mocement Science I3 (1994) 495-517

Perell, K.D., J. Gregor, J.A. Buford and J.L. Smith, 1993. Adaptive control of backward quadrupedal walking: IV. Hindlimb kinetics during stance and swing. Journal of Neurophysiology 70, 22262240. Perret, C. and J.-M. Cabelguen, 1976. Central and reflex participation in the timing of locomotor activations of a bifunctional muscle, the semi-tendinosus, in the cat. Brain Research 106, 390-395. Perret, C. and J.-M. Cabelguen, 1980. Main characteristics of the hindlimb locomotor cycle in the decorticate cat with special reference to bifunctional muscles. Brain Research 187, 333-352. Powers, R.K. and M.D. Binder, 1985. Distribution of oligosynaptic group I input to the cat medial gastrocnemius motoneuron pool. Journal of Neurophysiology 53, 497-517. Pratt, C.A., 1992. ‘Specializations in the outputs of low-threshold cutaneous reflex pathways during normal locomotion’. In: M. Woollacott and F. Horak (Eds.), Posture and gait: Control mechanisms, Vol. II (pp. 32-35). Eugene, OR: University of Oregon Books. Pratt, C.A., J.A. Buford and J.L. Smith, 1992. Mutable activation of bifunctional muscles during forward and backward walking. Neuroscience Abstracts 18, 1555. Pratt, C.A., CM. Chanaud and G.E. Loeb, 1991. Functionally complex muscles of the cat hindlimb. IV. Intramuscular distribution of movement command signals and cutaneous reflexes in broad, bifunctional thigh muscles. Experimental Brain Research 85, 281-299. Pratt, C.A., J. Fung and R. Jacobs, 1993. Evidence of positive force feedback during standing in intact cats. Neuroscience Abstracts 19, 142. Pratt, C.A. and G.E. Loeb, 1991. Functionally complex muscles of the cat hindlimb. I. Patterns of activation across sartorius. Experimental Brain Research 85, 243-256. Prochazka, A., P. Trend, M. Hulliger and S. Vincent, 1989. Ensemble proprioceptive activity in the cat step cycle: Towards a representative look-up chart. Progress in Brain Research 80, 61-74. Prochazka, A., R.A. Westerman and S.P. Ziccone, 1977. Ia afferent activity during a variety of voluntary movements in the cat. Journal of Physiology 268, 423-448. Rasmussen, S., A.K. Chan and G.E. Goslow, Jr., 1978. The cat step cycle: Electromyographic patterns for hindlimb muscles during posture and locomotion. Journal of Morphology 155, 253-270. Redfield, R. and M.L. Hull, 1986. On the relation between joint moments and pedalling rates at constant power in bicycling. Journal of Biomechanics 19, 317-329. Rushmer, D.S., J.M. Macpherson, D.C. Dunbar, C.J. Russell and S.L. Windus, 1987. Automatic postural responses in the cat: Responses of proximal and distal hindlimb muscles to drop of support from a single hind- or forelimb. Experimental Brain Research 65, 527-537. Smith, J.L., B. Betts, V.R. Edgerton and R.F. Zernicke, 1980. Rapid ankle extension during paw shakes. Selective recruitment of fast ankle extensors. Journal of Neurophysiology 43, 612-620. Smith, J.L., S.H. Chung and R.F. Zernicke, 1993. Gait-related motor patterns and hindlimb kinetics for the cat trot and gallop. Experimental Brain Research 94, 308-322. Smith, J.L., V.R. Edgerton, B. Betts and T.C. Collates, 1977. EMG of slow and fast ankle extensors of cat during posture, locomotion and jumping. Journal of Neurophysiology 40, 503-513. Smith, J.L., M.G. Hoy, G.F. Koshland, D.M. Phillips and R.F. Zernicke, 1985. Intralimb coordination of the paw-shake response: A novel mixed synergy. Journal of Neurophysiology 54, 1271-1281. Smith, J.L. and R.F. Zernicke, 1987. Predictions for neural control based on limb dynamics. Trends in Neurosciences 10, 123-128. Van Ingen Schenau, G.J., 1989. From rotation to translation: Constraints on multi-joint movements and the unique action of bi-articular muscles. Human Movement Science 8, 301-337. Van Ingen Schenau, G.J., P.J.M. Boots, G. de Groot, R.J. Snackers and W.W.L.M. van Woensel, 1992a. The constrained control of force and position in multi-joint movements. Neuroscience 46, 197-207. Van Ingen Schenau, G.J., P.J. Beek and R.J. Bootsma, 1992b. Is position information alone sufficient for the control of external forces? Behavioral and Brain Sciences 15, 804-805. Vereijken, B.R., E.E.A. van Emmerik, H.T.A. Whiting and K.M. Newell, 1992. Freetzhng degrees of freedom in skill aquisition. Journal of Motor Behavior 24, 133-142.

G.J. van Ingen Schenau et al. /Human

Movement Science 13 (1994) 495-517

517

Wisleder, D., R.F. Zernicke and J.L. Smith, 1990. Speed-related changes in hindlimb intersegmental dynamics during the swing phase of cat locomotion. Experimental Brain Research 79, 651-660. Woollacott, M.H., M. Bonnet and K. Yabe, 1984. Preparatory process for anticipatory postural adjustments: Modulation of leg muscles reflex pathways during preparation for arm movements in standing man. Experimental Brain Research 55, 263-271.