Central Processes Involved in Arm Movement Control · Central Processes Involved in Arm Movement Control 5 Figure 1-1. Typical visually triggered head movements in chronically vestibulecto

Chapter 1

Central Processes Involved in Arm Movement Control

EMILIO BIZZI

1. Introduction

The mechanism whereby the central nervous system (CNS) controls, si­multaneously, the large number of degrees of freedom of the multijoint musculoskeletal system has long been one of the major questions in motor physiology. This incredibly complex task must be dependent upon "rules" which permit original and creative motor solutions to problems such as reaching for objects in a structured space. As a step toward understanding these rules, a number of investigators have studied relatively simple motions involving a limited number of joints (Asatryan & Feldman, 1965; Bizzi, Polit, & Morasso, 1976; Cooke, 1979; Kelso & Holt, 1980). The results have indicated some of the processes that subserve these movements and have given us a picture of the elementary building blocks from which more complex movements may be constructed. Several points have emerged from these investigations. First, it has become apparent that the forces that control the muscles result from "commands" that are, to a great extent, precomputed in some part of the CNS. These observations are based on studies made in deafferented animals that have demonstrated open-loop reaching (Bizzi et aI., 1976; Polit & Bizzi, 1979; Taub, Goldberg, & Taub, 1975). Second, recent investigations have convincingly demonstrated that muscles, to a first approximation, can be thought of as behaving like tunable springs (Rack & Westbury, 1974). In fact, springs and muscles have a fundamental property in common: they produce force as a function oflength (Feldman, 1966; Partridge, 1979; Rack & Westbury, 1974).

P. F. MacNeilage (ed.), The Production of Speech© Springer-Verlag New York Inc. 1983

4 Emilio Bizzi

These findings have suggested some possible control strategies whereby the CNS takes advantage of the mechanical properties of the muscles.

In this chapter I will discuss the foregoing points by examining the results obtained from monkeys performing visually evoked head and arm movements. In general, the questions of open-loop control and the role of feedback during movements were studied in intact animals that were later deprived of sensory feedback. In addition, both intact and sensory deprived animals were subjected to various force disturbances. The combination of these two approaches has allowed us to make some inferences on the way in which the CNS controls movement.

II. Experiments and Results

I will review first the results obtained by applying a constant-torque load to the head. When this type of load was applied, a constant degree of head undershoot was observed. In the intact animal, while a constant load was being applied there was an increase in electromyographic (EMG) activity, presumably due to an increase in muscle spindle and tendon organ activity. As shown in Figure 1-1, in spite of these changes in the flow of proprioceptive activity the head reached its "intended" final position after the constant load was removed. In fact, the final head position was equal (on average) to that reached when the load had not been applied, which suggests that the program for final position was maintained during load application and was not readjusted by pro­prioceptive afferents generated during the movement, but is preprogrammed. It should be noted that the load disturbances were totally unexpected and that the monkeys were not trained to move their head to a certain position, but chose to program a head movement together with an eye movement in order to perform a visual discrimination task (Bizzi, 1974; Bizzi, Kalil, & Tagliasco, 1971; Bizzi et al, 1976).

In a second set of experiments Bizzi, Polit, and Morasso (1976) examined the effect of stimulating proprioceptors only during the dynamic phase. To this end they used as a stimulus a load that modified the trajectory but did not represent a steady-state disturbance. This was done by using an inertial load. As a result of the sudden and unexpected increase in inertia during centrally initiated head movement, the following changes in head trajectory, relative to unloaded movement, were observed: an initial slowing down of the head, followed by a relative increase in velocity (due to the kinetic energy acquired by the load being transmitted to the decelerating head) that culminated in an overshoot; finally, the head returned to the intended position (Fig. 1-2).

The changes in head trajectory brought about by the sudden and unexpected increase in head inertia induced corresponding modifications in the length and tension of neck muscles. The agonist muscles were, in fact, first subjected to

Central Processes Involved in Arm Movement Control 5

Figure 1-1. Typical visually triggered head movements in chronically vestibulecto­mized monkey to appearance of target at 40° but performed in total darkness. A shows an unloaded movement. In B, a constant-force load (315 g/cm) was applied at the start of the movement, resulting in an undershoot of final position relative to A despite an increase in EMG activity . In C, a constant-force load (726 g/cm) was applied. Note that the head returns to the same final position after removal of the load. Vertical calibration in degrees ; time marker is I sec; EMG recorded from left splenius capitis. (From Bizzi et aI., 1976.)

increased tension because the application of the load slowed down the process of muscle shortening; then the shortening of the same muscles was facilitated during the overshoot phase of the head movement induced by the kinetic energy of the load. Such loading and unloading did, of course, provoke the classical muscle spindle response (presumably mediated by group IA and group II afferent fibers) which, in tum, affected the agonist EMG activity. Figure 1-2B shows that there was first a greater increase in motor unit discharge during muscle stretch than would have occurred if no load were applied, followed by a sudden decrease in activity at the beginning of the overshoot phase. Therefore, during a head movement, an unexpected inertial load induced a series of waxing and waning proprioceptive signals from muscle spindles, tendons, and joints, but the intended head position was eventually reached, even in the complete absence of other sensory (visual and vestibular) cues. This observation, together with those on the effect of constant-torque loads, suggests that the central program establishing final head position is not dependent on a readout of proprioceptive afferents generated during the movement, but instead is pre­programmed.

To provide a further test of the hypothesis that final pOSItIon is pre­programmed, Bizzi et al. (1976) investigated how chronically vestibulectomized monkeys reached final head position without visual feedback when they were deprived, in addition, of neck proprioceptive feedback. Following the un­expected application of a constant-torque load at the beginning of a visually

6 Emilio Bizzi

triggered movement, the posture attained by the head was short of intended final position. After removal of the constant torque, the head attained a position that was found to be equal to the one reached by the head in the no-load case. These results indicate that the head motor system behaved qualitatively in the same way both before and after deafferentation with respect to head position.

I believe that the results of these experiments contribute to our understanding of the mechanism whereby movement is terminated and a newly acquired position is maintained. If one assumes that the "program" for head movements and posture specifies a given level of alpha motoneuron activity to both agonist and antagonist muscles, and that the firing of these neurons will determine a particular length-tension curve in each muscle, then it must be concluded that the final resting position of the head is determined by the length-tension properties of all of the muscles involved. This hypothesis explains both the head undershoot when a constant load is applied and the attainment of the intended final head position following the removal of the load, as shown in Figures 1-1 and 1-2, respectively.

In a complementary set of experiments involving arm movements, Pol it and Bizzi ( 1979) extended the previously described findings on the final position of the head. Adult rhesus monkeys were trained to point to a target light with the forearm and to hold the arm in that position for about 1 sec in order to obtain a reward (Figure 1-3). The monkey was seated in a primate chair, and its forearm was fastened to an elbow apparatus that permitted flexion and extension of the forearm about the elbow in the horizontal plane. A torque motor in series with

A

Figure 1-2. Typical head responses of a chronically vestibulectomized monkey to sudden appearance of target at -40°. A represents an unloaded movement, whereas in B a load of approximately six times the inertia of the head was applied at the start of the movement, as indicated by the force record. Both movements were performed in total darkness, the light having been turned off by the increase in EMG (splenius capitis). Peak force exerted by the monkey is approximately 750 g/cm; head calibration is in degrees ; time marker is 1 sec . (From Bizzi et aI. , 1976.)

Central Processes Involved in Arm Movement Control 7

Figure 1-3. Monkey set up in arm apparatus. Arm is strapped to splint, which pivots at elbow. Target lights are mounted in perimeter arc at 5 0 intervals. During experimental session, the monkey was not permitted to see its arm, and the room was darkened. (From Polit & Bizzi, 1979.)

the shaft of this apparatus was used to apply positional disturbances to the arm. The experiments were conducted in a dark room to minimize visual cues; at no time during an experiment was the animal able to see its forearm. At random times, the initial position of the forearm was displaced. In most cases, the positional disturbance was applied immediately after the appearance of the target light and was stopped just prior to the activation of the motor units in the agonist muscle. Hence, when the motor command specifying a given forearm movement occurred, the positional disturbance had altered the length of the agonist and antagonist muscles, and the proprioceptive stimulation resulting from this disturbance had altered their state of activation. In spite of thef>e changes, the intended final arm position was always reached; this was true whether the torque motor had displaced the forearm farther away from, closer to, or even beyond the intended final position. There were no significant differences among the final positions achieved in these three conditions. Naturally, the attainment of the intended arm position in this experiment could be explained by assuming that afferent proprioceptive information modified the original motor command. However, the results of previous work on final head position suggest an alternative hypothesis: The motor program underlying arm movement specifies, through the selection of a new set of length-tension curves, an equilibrium point between agonists and antagonists that correctly positions the arm in relation to the visual target. To investigate this hypothesis, Polit and

8 Emilio Bizzi

Bizzi (1979) retested the monkey's pointing performance after it had undergone a bilateral C r-T3 dorsal rhizotomy. They could elicit the pointing response very soon after the surgery (within 2 days in some of the animals). The forearm was again displaced (at random times) immediately after the appearance of the target light and released just prior to the activation of motor units in the agonist muscles. Even though the arm was not visible to the animal and the pro­prioceptive activity could not reach the spinal cord, the arm reached its intended final position "open loop" (a fact corroborated by lack of any sign of reflex response or reprogramming in the EMG activity). For each target position, t tests were performed to test for differences between the average final position of movements in the undisturbed and the disturbed conditions. No significant differences were found.

These findings, together with those derived from head movement experi­ments, suggest that arm and head movements depend on neural patterns that are programmed prior to movement initiation. What is being preprogrammed is a process that is capable of controlling final head and arm position as an equilibrium point resulting from the interaction of agonist and antagonist muscles. This view may be illustrated by reference to a simple mechanical analogue. Assume that the muscles moving a body segment can be represented by a pair of springs acting across a hinge in the agonist-antagonist configuration. If the CNS were to specify a new length-tension relationship for one of the springs, movements would occur until a new equilibrium point of the two opposing springs was reached. However, the specification of a new set of length-tension settings does not occur in a step like fashion. In fact, recent experimental evidence addressed to the question of the time course of the neural signal executing the transition from an initial to a final posture has demonstrated the existence of a gradually changing control signal during arm movement performed at moderate speed (Bizzi, Accornero, Chapple, & Hogan, 1982 ).

III. Conclusion

The experimental evidence that has been summarized here indicates that for certain visuomotor responses the "commands" to the musculature are pre­programmed. What is being specified by these commands is a trajectory as well as a final position. Whether the process underlying final arm position and the process specifying movement duration and velocity can be thought of as parallel and relatively independent processes or as a single process specifying, in time, a series of positions remains to be ascertained.

A second goal of these studies was to develop some perspectives on the role of afferent feedback during voluntary movements (Bizzi et ai., 1976; Bizzi, Dev, Morasso, & Polit, 1978; Pol it & Bizzi, 1979). The studies of the deafferented animal showed that the successful execution of forearm programs released by target presentation was contingent on the animal's knowing the position of its

Central Processes Involved in Arm Movement Control 9

arm relative to its body. Whenever the usual spatial relationship between the animal and the arm apparatus was changed or a constant-bias load was applied, the monkey's pointing response was inaccurate. The intact monkeys, in contrast, were able to compensate quickly for any variations in their ac­customed position with respect to the arm apparatus. The dramatic inability of the deafferented monkey to execute accurate pointing responses in an unusual postural setting or when a constant-bias load was applied underscores the great importance of afferent feedback in updating and adjusting the execution of learned motor patterns when posture is changed. These findings emphasize the widespread influence and importance of afferent impulses in the control of voluntary movement. They suggest that, in addition to contributing to the classical spinal and supraspinal reflex loops, which may servo-assist movement (Marsden, Merton, & Morton, 1976a, 1976b; Vallbo, 1973; Wilson, 1961), provide a small contribution to load compensation (Allum, 1975; Bizzi et aI., 1978; Conrad, Matsunami, Meyer-Lohmann, Weisendanger, & Brooks, 1974), and/or linearize muscle properties (Nichols & Houk, 1973, 1976), the afferent system may affect, in a manner that is not yet understood, a reorganization of the central processes that are released when targets are presented. It is perhaps of interest to comment that although servo assistance or load compensation can occur during a single centrally driven movement, the postulated reorganization has a longer time scale, encompassing a few movements.

Acknowledgments. This research was supported by National Institute of Neurological Disease and Stroke Research Grant NS09 343, National Institute of Arthritis, Metabolism and Digestive Diseases Research Grant AM 26710, and National Eye Institute Grant EY02621.

References

Allum, 1. H. J. Responses to load disturbances in human shoulder muscles: The hypothesis that one component is a pulse test information signal. Experimental Brain Research, 1975,22,307-326.

Asatryan, D. G., & Feldman, A. G. Biophysics of complex systems and mathematical models. Functional tuning of nervous system with control of movement or maintenance of a steady posture. I. Mechanographic analysis of the work of the joint on execution of a postural task. Biophysics, 1965, 10, 925-935.

Bizzi, E. The coordination of eye-head movement. Scientific American, 1974, 231, 100-106.

Bizzi, E., Accornero, N., Chapple, W., & Hogan, N. Arm trajectory formation in monkeys. Experimental Brain Research, 1982,46,139-143.

Bizzi, E., Dev, P., Morasso, P., & Polit, A. The effect of load disturbances during centrally initiated movements. Journal of Neurophysiology, 1978,41,542-556.

Bizzi, E., Kalil, R. E., & Tagliasco, V. Eye-head coordination in monkeys: Evidence for centrally patterned organization. Science, 1971, 173, 452-454.

10 Emilio Bizzi

Bizzi, E., Polit, A., & Morasso, P. Mechanisms underlying achievement of final head position. Journal oj Neurophysiology, 1976,39,435-444.

Conrad, B., Matsunami, C., Meyer-Lohmann, J., Wiesendanger, M., & Brooks, V. B. Cortical load compensation during voluntary elbow movements. Brain Research, 1974.81,507-514.

Cooke, J. D. Dependence of human arm movements on limb mechanical properties. Brain Research, 1979, 165, 366-369.

Feldman, A. G. Functional tuning of the nervous system during control of movement or maintenance of a steady posture. III. Mechanographic analysis of the execution by man of the simplest motor tasks. Biophysics, 1966, 11, 766-775.

Kelso, J. A. S., & Holt, K. G. Exploring a vibratory system's analysis of human movement production. Journal oj Neurophysiology, 1980,43, 1183-1196.

Marsden, C. D., Merton, P. A., & Morton, H. B. Servo action in the human thumb. Journal oj Physiology, London, 1976,257, 1-44. (a)

Marsden, C. D., Merton, P. A., & Morton, H. B. Stretch reflexes and servo actions in a variety of human muscles. Journal oj Physiology, London, 1976,257, 531-560.(b)

Nichols, T. R. & Houk, J. C. Reflex compensation for variations in the mechanical properties of a muscle. Science, 1973, 181, 182-184.

Nichols, T. R., & Houk, J. C. The improvement in linearity and the regulation of stiffness that results from the actions of the stretch reflex. Journal oj Neuro­physiology, 1976,39, 119-142.

Partridge, L. D. Muscle properties: A problem for the motor controller physiologist. In R. E. Talbott & D. R. Humphrey (Eds.), Posture and movement. New York: Raven Press, 1979.

Polit, A., & Bizzi, E. Characteristics of motor programs underlying arm movements in monkeys. Journal oJNeurophysiology, 1979,42,183-194.

Rack, P. M. H., & Westbury, D. R. The short range stiffness of active mammalian muscle and its effect on mechanical properties. Journal oj Physiology, London, 1974, 240,331-350.

Taub, E., Goldberg, I. A., & Taub, P. Deafferentation in monkeys: Pointing at a target without visual feedback. Experimental Neurology, 1975, 46, 178-186.

VaHbo, A. B. The significance of intramuscular receptors in load compensation during voluntary contractions in man. In R. B. Stein, K. G. Pearson, R. S. Smith, & J. B. Redford, (Eds.), Control oJposture and locomotion. New York: Plenum, 1973.

Wilson, D. M. The central nervous control of flight in a locust. Journal oJExperimental Biology, 1961,38,471-490.