WHAT ARE "NORMAL MOVEMENTS" IN ATYPICAL POPULATIONS?

Keywords

Abstract

1. Introduction

Studies of populations whose ability to perform voluntary movements is impaired due to natural reasons (e.g., aging), inborn deficiency (e.g., Down syndrome), trauma (e.g., spinal cord injury), or illness (e.g., Parkinson's disease), frequently result in a basic question: Are the observed motor patterns, which may be rather different from those observed in unimpaired persons, actually abnormal and should they be corrected? Analysis of this question is important not only for deeper understanding of the mechanisms of control of normal and disordered movements but also to assess the effectiveness of existing therapeutic approaches and to provide a focus within which development of new therapies can be considered.

A common misconception is that any major deviation from motor patterns seen in the general unimpaired population is bad. This misconception is revealed in the way the research findings are presented and interpreted, and in the prescriptions to correct the "wrong" motor patterns (Larkin & Hoare, 1991; Lockwood, 1987; Morris, Iansek, Matyas & Summers, in press; Shepherd, 1984). This view is based on a number of questionable premises:

1. Patterns of voluntary movements seen in the general population are the only correct ones;

2. Deviations from the "normal" patterns reflect failure of the central nervous system (CNS) to behave "correctly";

3. We know more about motor control than the CNS does.

The principal objective of this paper is to suggest alternatives to these premises and to exemplify the alternatives with a number of examples from motor research involving atypical populations.

Figure 1 illustrates our approach. This figure could be viewed as a red herring, but we feel that it is necessary to start with a general scheme, even if it is imprecise and purposefully provocative. We consider "normal movement patterns" (Average Person in Fig. 1) to represent a spectrum that merges at one end with clumsiness and impaired movements and, at the other end, with perfection and uniquely specified movements. Clumsy children and elite athletes are at the opposite ends of the spectrum and signify observable limits of natural variability. The whole spectrum is associated with basically the same CNS priorities (or coordinative rules) that are used to solve the problem of motor redundancy (see Section 3 for details). However, motor patterns within a segment of this spectrum may be quite different, e.g. if we compare walking of a heavyweight weight-lifter and of a ballet dancer.

When we move beyond these limits into an area that may be considered pathological or otherwise special, the CNS priorities may change and lead to apparently atypical motor patterns. This may happen in the absence of any gross neurological or motor pathology, e.g., due to changes in cognition and/or decision-making (examples being movements of Down syndrome individuals and persons with schizophrenia). Further to the left, we encounter morphological, biochemical, or structural CNS changes that may induce differences in motor patterns by themselves and also by changing the CNS priorities, e.g. in Parkinson's disease and spinal cord injury. Toward the left end of the axis, we see peripheral changes, as in cases of amputation, which certainly limit motor patterns by themselves, but may also lead to reorganization within the CNS (taking advantage of its plasticity), and to resultant changes in CNS priorities. We will try to persuade the reader that changed CNS priorities may play an important role in defining movement patterns that look abnormal and that these patterns reflect control signals that have been elaborated by the CNS based on the current state of its priorities and on the state of the neural and motor structures. In this sense, these movement patterns may be considered optimal.

Let us start with a brief sortie into an area of movement patterns that are commonly considered "normal". This will define, in a descriptive fashion, part of the spectrum in Fig. 1 that is labeled "Average Person".

2. Basic Features of "Normal Motor Patterns"

"Normal", "normality", "normal motor patterns" are misnomers and have been the scourge of scholars who have attempted to generate movement typologies (e.g., Sheridan, 1985). So, what is "normal" and what is a "normal movement pattern"? If "normal" is identified by the presence of consistency and the absence of differences within a subpopulation, then the slowness in Down syndrome and shuffling gait in Parkinson's disease are normal. If "normal" is related to movement outcome then the movement patterns seen in the individual with cerebral palsy or the wheelchair-bound person with a spinal cord injury who satisfactorily negotiates the entry to a building via steps or ramp are normal.

Let us continue by stating up front that, like normality itself, the mechanisms of normal motor control are generally unknown. We have only a few tentative, marginally specific hypotheses about the role of different structures within the CNS in control of voluntary movements. These hypotheses may look relatively successful when applied to a narrow class of phenomena associated with voluntary movements; however, commonly it is easy to imagine a motor phenomenon that would contradict any of them. We don't know which variables are used for communication within the CNS during decision-making and movement execution processes. We are unable to give an unambiguous, clear definition of "motor command" or "coordination". There is not even consensus on what is perceived by our own brains as a "motor task". After this fit of (self)-flagellation, we would like to acknowledge the considerable recent progress in understanding of a number of particular aspects of motor control. This progress has not led, however, to a major breakthrough that would result in the formulation of a general, universally acceptable theory of motor coordination.

Because of the lack of a general theory of motor control, the only remaining option is to present a description of the regularities in motor behavior. We will review briefly some of the basic features of single-joint and multi-joint voluntary movements, postural reactions associated with voluntary movements, and some issues of motor variability. In order to avoid a lengthy list of references, the reader is referred to the following recent sources which contain references to the original publications: Gottlieb, Corcos & Agarwal 1989; Winters & Woo 1990; Massion 1992; Latash 1993; Newell & Corcos 1993; Swinnen, Massion, Heuer & Casaer, 1994.

2.1. Single-Joint Movements.

First of all, let us state that for all practical purposes in human motor control, single-joint movements exist principally as a helpful laboratory abstraction but not as a real-life phenomena. This is due to a number of factors including the polyfunctionality of many muscles (e.g., Buchanan, Rovai & Rymer 1989), the specificity of biarticular muscles (e.g., van Ingen Schenau 1989; Nichols 1989), and mechanical joint interaction. Having said this, let us now review the basic features of movements that are mostly confined to one degree of freedom of a single joint and which are frequently described as single-joint movements.

If a subject is asked to perform a fast, smooth, discrete movement to a target, the joint trajectory demonstrates a very typical pattern with "bell-shaped" velocity and double-peaked acceleration (Fig. 2). Slow movements demonstrate an initial increase in velocity, a period of relatively constant velocity, and a decrease in velocity while approaching the target.

Recording muscle activity during fast single-joint movements reveals a typical electromyographic (EMG) pattern usually termed the "tri-phasic pattern" (Fig. 2). The beginning of the agonist* EMG burst (triceps in Fig. 2) is usually the first detectable event accompanying fast voluntary movement. It precedes first detectable kinematic changes by several tens of milliseconds. The initial agonist burst is accompanied by a relatively low coactivation of the antagonist muscle and is followed by an antagonist EMG burst (biceps in Fig. 2) during which the agonist is relatively quiescent. Later, a second agonist burst may be observed, whose function is rather obscure and is sometimes associated with damping terminal oscillations of the moving limb.

This triphasic EMG pattern is relatively robust although characteristics within each of the three components may alter with changes in movement parameters. In particular, an increase in velocity of movements over a constant amplitude against a constant load leads to an increase in the initial slope of the EMG rise, peak value and area of the first agonist burst, a decrease in the delay before the antagonist burst, and an increase in the antagonist burst amplitude and area. An increase in movement amplitude without changes in external load and instructions concerning movement velocity (e.g., "as fast as possible") leads to relatively uniform initial slopes of agonist EMG rise, higher and longer first EMG burst with a corresponding increase in its area, longer delays before the antagonist burst, and inconsistent changes in the antagonist burst amplitude and duration. An increase in inertial load without changing movement amplitude leads to higher and longer agonist EMG activity, no changes in the rate of the EMG rise, longer delay before the antagonist burst and no apparent changes in the antagonist burst characteristics or an increase in its amplitude and area.

2.2. Multi-Joint Movements.

If a task is to move a working point (e.g., the tip of a finger) from a certain initial position to a specific final position on a plane, subjects prefer nearly straight trajectories of the working point independent of movement speed, load, and some other factors. Such movements are characterized by bell-shaped velocity profiles for the working point while speed profiles for individual joints can have a more complex, multi-phasic form and depend upon the area of external space where the movement is performed. Deviations of working point trajectories from straight lines have been described. These deviations are more pronounced for movements in certain areas of the working space. Flash (1987) has ascribed them to changes in compliance of the working point in different areas of the working space and has modeled curved working point trajectories in the framework of the "equilibrium trajectory hypothesis" (Flash & Hogan 1985).

When the subjects are asked to perform curvilinear movements, the working point trajectories demonstrate clear segmentation giving the appearance of being glued together of separate fragments. Movement velocity increases with a decrease in the radius of curved trajectories. Minimal movement time has been reported for straight trajectories. A quantitative relation between movement speed of the working point and radius of the trajectory has been proposed in a form of the "law of 2/3":

V = b*r2/3, (1) where V is speed of the working point, r is radius of the trajectory, and b is a constant.

If a subject is asked to draw simple planar figures like straight lines, triangles, and squares, movement time is relatively constant for drawing figures of different size but belonging to the same class (e.g., a big and a small triangle). This effect has been addressed as the principle of isochrony. An increase in the number of segments leads to a nearly standard increase in movement time by approximately 200 ms per segment. That is, drawing a triangle takes approximately 400 ms more than drawing a straight line, and drawing a square takes 200 ms more than drawing a triangle.

During three-dimensional, unconstrained movements, subjects try to restrict movements to a plane. They demonstrate apparent segmentation of movement trajectories when trying to change from one movement plane to another.

2.3. Postural Adjustments and Pre-programmed Reactions.

Virtually all natural human movements, e.g., pointing, reaching, grasping, throwing, catching, etc., involve two distinct peripheral components. One component involves changes in activation patterns of muscles directly participating in the movement. The other component provides postural stability for the limb directly involved in the primary movement and/or equilibrium for the trunk and for the head. Any voluntary movement is by itself a postural perturbation mostly because of the mechanical coupling of joints. The transmission of forces and torques from the moving segment through the body's linked segments is the primary reason for postural perturbations. Corrections to postural perturbations induced by a voluntary movement are termed "anticipatory" if they are released prior to the movement. These corrections represent feedforward postural control associated with the movement control which prevents or alleviates the posture and equilibrium disturbances. They are pre-programmed, triggered internally, and time-locked to the future movement's initiation. Their mechanical effect counteracts the postural disturbance expected from the planned movement. For example, in normal standing, a fast movement of the shoulder is preceded by changes in the levels of activation of leg and trunk muscles that lead to a shift of the center of mass in a direction opposite to the one expected from action of the reactive forces induced by the primary movement. Similarly, at the level of single limb movements, a fast movement of the elbow is accompanied by phasic changes in the levels of activation of muscles acting at the wrist joint which prevent flapping of the palm.

Voluntary limb movements are also associated with pre-programming of urgent corrections to possible external perturbations that may endanger successful execution of the task. These pre-programmed reactions (alias triggered reactions, alias M2-3, alias functional stretch reflex) come at a latency of about 70 ms which is greater than the latency of monosynaptic responses and is less than the simple reaction time. If a perturbation (e.g., a decrease or an increase in load) is applied to the joint in which the primary movement is occurring, predictable changes in muscle activity will result. An increase in external load usually leads to increased activity in the agonist muscles, while decreased load leads to diminished agonist and/or increased antagonist activity. An important, distinctive feature of pre-programmed reactions is their dependence upon the context of a motor task and, particularly, upon the instruction to the subject. If a subject is asked not to correct limb position in cases of external perturbations, these reactions decrease and may even disappear. In contrast, if the subject is asked to always preserve a position (or a trajectory) in cases of perturbations, pre-programmed reactions are increased. A similar context dependence is demonstrated by postural corrections. For example, if a subject is standing in a train holding a cup in an extended hand and the train starts to move, the pattern of postural corrections will be different if the cup is full of hot tea than if the cup contains play dough (modeling clay).

2.4. Normal Motor Variability.

Variability is a nuisance for some motor studies and a source of excitement for others. It may be considered a consequence of "noise" within the system for movement production implying its suboptimal performance, or an inherent component of a perfectly functioning non-equilibrium, dynamic system. The phenomena of variability of voluntary movements by themselves indicate that "correct" peripheral motor patterns may form a rather wide spectrum.

The most famous general law of motor variability is probably Fitts' Law which deals with a specific aspect of variability, namely with the speed-accuracy trade-off. It was originally introduced for describing the dependence of movement time upon target size and movement amplitude during repetitive and, later, discrete arm movements performed under visual control with explicitly presented targets. Fitts proposed a logarithmic relation among movement time (MT), movement amplitude (A), and target width (W):

MT = a + b log2 f(2A,W) (1) where a and b are empirically defined constants.

Since the original works by Fitts, logarithmic relations like Eq. (1) have been used for describing experimental findings in a variety of conditions with movements of different complexity. A different relation between movement time, movement distance, and variability assessed as the standard deviation (SD) of final position or steady state force has been reported based on the studies of very fast force pulses or movements:

SD = a + b f(A,MT) (2) where a and b are constants. In order to compare the results of these experiments with the findings by Fitts, a variable has been introduced characterizing dispersion of the final position termed "effective target width" We. Eq. (2) has been compared to Eq. (1) many times, although Eq. (2) relates movement distance and time to a variable characterizing actual performance (SD, or We), while Eq. (1) relates the same variables to a variable characterizing externally imposed task requirements (W).

During repetitions of fast movements to a target, variability of the movement trajectory increases during the first phase of the movement and decreases later on. This finding has been interpreted by some authors as resulting from a smart compensating action of the central nervous system, while other researchers have shown that this fact may follow the laws of kinematics assuming an underlying error in timing and no corrective actions by the CNS.

2.5. Dynamic Pattern Generation

Until now, our story has followed a number of implicit assumptions, most conspicuously that the system of motor control is based on an hierarchical model so that a smart decision-making homunculus solves a motor task taking into account all the relevant information about the body and environment. The homunculus issues patterns of control variables that are supplied to functionally "lower" structures that implement the message.

This assumption is made in over 95% of motor control studies. However, an alternative explanation has become progressively stronger and is attracting more and more sympathizers. Different names have been used to describe this approach, in particular "dynamic pattern generation". Dynamic pattern generation rejects the hierarchical principle in motor control. According to it, there is no such a thing as "motor program", and maintenance of the vertical posture (as well as other coordinated motor activities) is achieved by introducing a certain dynamic order into a non-equilibrium system involving neural structures, effectors, external force fields, as well as intentional and perceptual variables. This approach emphasizes the importance of a group of variables that are frequently addressed as "collective variables" or "order parameters". These variables are assumed to define general patterns of coordination and, ultimately, shape behavior of the system including output signals to individual limbs, joints, and muscles. When participating in a postural task, the system fluctuates around an equilibrium defined by the task and the dynamic properties of the system. Postural synergies induced by a perturbation, whether external or modeled within the system based on planned actions, may be considered emergent properties of the system defined by its dynamic organization and, therefore, do not require any explicit planning or pre-programming.

Within this scheme, an important role is assigned to perceptual and intentional variables. They are supposed to be coupled with neural and mechanical variables by common equations that are the essential equations of motion. Unfortunately, these equations are generally unknown. If, as we hope, control of movement relies on general physical principles then the derivation and solution for such equations would seem to be essential for the development of the dynamic pattern generation approach.

3. The Problem of Choice (The Bernstein Problem)

Let us start, ad absurdum, from the third of the basic assumptions (section 1) that we know about motor control more than the CNS does. This assumption is certainly wrong because the CNS "knows" how to develop and control movements while we are just starting to formulate viable hypotheses in this field. This is especially true with respect to unconstrained multi-joint movements. Even the residual CNS of the spinal frog is able to bring about beautifully coordinated, multi-joint targeted movements, for example, during the wiping reflex (Fukson, Berkinblit & Feldman 1980).

In 1936, N. A. Bernstein suggested that the basic problem of motor control may be addressed as overcoming redundant degrees-of-freedom (DOFs). This problem, known as "the Bernstein problem" (for reviews see Whiting 1984; Turvey 1990; Latash 1993), seems obvious when considered at the level of the periphery of the motor system, since the number of DOFs summed over the joints is usually more than three, which is the number of dimensions of the space in which we happen to live. Note, however, that the Bernstein problem was formulated at the level of control rather than at the level of peripheral motor patterns: How does the CNS choose among patterns of (hypothetical) control variables that may lead equivalently to a desired goal?

Let us keep in mind, however, that there are likely to be constraints that may decrease the number of actual DOFs of the endpoint of a multi-joint limb. These constraints may be biomechanical (e.g., biarticular muscles; see van Ingen Schenau 1989; Zajac & Gordon 1989; Gielen, van Ingen Schenau, Tax & Theeuwen 1990) and physiological (e.g., inter-muscular reflexes; see Nichols, 1989). There may also be constraints at the level of control that we have not yet discovered. As a result, the problem may disappear and degenerate into a unique solution for a given motor task.

For example, each wheel of a car has two degrees of freedom. Therefore, four independent wheels have eight DOFs. However, we know that a car with four wheels has only two DOFs that allow it to travel across a surface. The remaining DOFs are eliminated by the mechanical design of the car and its controls (e.g., the system of steering).

Even if there is no redundancy (in the sense that each location of an endpoint corresponds unambiguously to one and only one combination of values of hypothetical central variables) there may still be redundancy in choosing time patterns of these variables that lead from an initial to a final set of values. The choice of a particular time pattern may be affected by factors related to dynamic stability of the movement, changes in activity of seemingly non-involved muscles that assure stable posture of the limb, and/or equilibrium of the whole body (for a recent review see Massion 1992), etc.

The Bernstein problem can occur at the level of single-joint, single DOF movements where there is seemingly no room for peripheral redundancy. Commonly, there are multiple "agonists" and "antagonists" for virtually any movement direction, and also muscles whose action cannot be easily classified into "flexors" or "extensors", "adductors" or "abductors" (Buchanan et al. 1986, 1989; van Zuylen, Gielen & Denier van der Gon 1988; Flanders & Soechting 1990; Tax, Denier van der Gon & Erkelens 1990). Even for a static motor task, the question of which muscles should be activated and to what levels to match a required joint torque may not have a unique solution. Thus, the Bernstein problem at this level of control may emerge for single-joint movements.

There is even an analog of the Bernstein problem at the level of single muscle control. Level of activation of a muscle is a result of summation of the activity of numerous motor units that may fire at different frequencies. Consider the question: "Which motor units should the CNS recruit and at what frequencies in order to match a required level of activation?" This is a typical problem of choice that is likely to have only a probabilistic solution. Henneman's size principle of motor unit recruitment (motor unit recruitment proceeds from small motor unit to progressively larger ones; Henneman, Somjen & Carpenter 1965) restricts accessible patterns of activation but does not unambiguously solve the problem. This principle may be termed a coordinative rule, a rule that restricts the area of possible solutions and makes not all the physically-possible solutions equivalent. Perhaps, at other levels, the Bernstein problem also has only probabilistic solutions, and the main quest of the motor control research should be considered as searching for coordinative rules rather than for unique solutions of motor control problems.

There are two qualitatively different ways of making a choice: First, to consider all the details of how a desired outcome is achieved, or second, to consider only the ultimate outcome. Imagine a group of 100 drummers. If the conductor wants to get a level of noise of 100 dB, he/she may tell each and every drummer how strongly to beat. On the other hand, he/she may provide an instruction: "Listen to the level of noise. If it is over 100 dB, beat lighter; if it is less than 100 dB, beat harder." Then, individual drummers will demonstrate a certain distribution in their performance which may fluctuate while keeping the ultimate outcome rather stable. So, when analyzing the behavior of a decision-making system (e.g., the brain), the first questions to ask are: What does the system care about? What kind of rules or criteria may be used in making a choice?

4. Priorities of the Central Nervous System

Let us consider the following schematic drawing representing the process of generation of a voluntary movement (Fig. 3). This scheme identifies three steps in this process: Understanding the motor task, generation of time patterns of control variables, and execution. The last step is assumed to involve neural mechanisms that are usually addressed as "reflex", as well as mechanical properties of muscles, tendons, and joints and their interaction with the external force field. The drawing also includes a box labeled "coordination" which defines values of order parameters (see above). We assume that the subject participating in a mental experiment understands the instruction and correctly identifies the explicit motor goal. This assumption leaves room for variability at the first step in Fig. 3: For example, a person with a cognitive disorder may understand the explicit motor task but, at the same time, be equally concerned with other factors that are not perceived as significant by unimpaired subjects (e.g., personal safety, not breaking the experimental setup, etc.). This may lead to the generation of different time functions of control variables at the next level and, consequently, to different movements.

However, even if there is no variability at the first step, the variability and flexibility of movement patterns during repetition of the same motor task by control (unimpaired) subjects suggests that the second step (generation of control patterns) virtually always involves choice. In other words, CNS engages in multiple strategies of muscle coordination to achieve a single movement outcome (e.g., Keshner, 1994). In other words, we assume that the number of control variables at this step is larger than the number of parameters defining a motor task. Thus, the problem of generation of appropriate values of control variables is similar, for example, to solving a system of two equations with four unknowns. Such "ill-posed" problems cannot be solved without adding additional equations that are not imposed upon the system and can be chosen by the CNS based on some secondary considerations. According to our terminology, such additional equations reflect priorities of the CNS. Within the suggested scheme, CNS priorities participate in the process of generation of the hypothetical control functions. We do not know which neural structures participate in this process and suspect that any reductionist attempts at this point are likely to be unsuccessful and misleading.

So, the CNS seems to be confronted with a problem of choice for virtually each and every movement including the most simple, frequently studied, single-joint movements. Apparently, it is able to make this choice or to allow the choice to be made "by itself" based on criteria and considerations that are unknown to us at this time. A few attempts at deciphering these CNS priorities have been made. Most frequently, the researchers tried to guess at CNS's internal solutions by investigating the consequences of optimizing certain functions of performance (for a review see Nelson 1983; Seif-Naraghi & Winters 1990). Attempts at minimizing or maximizing certain cost functions based on movement kinematics (peak velocity, peak acceleration, jerk), movement dynamics (joint torques), energy, or functions related to such notions as "comfort" and "effort" have not led to a breakthrough in understanding how natural movements are actually controlled. Some of these approaches, particularly the minimum-jerk model (Hogan 1984; Flash & Hogan 1985) and a kinematic model by Gutman (Gutman, Gottlieb & Corcos 1992), have demonstrated an impressive correspondence to the actual movement kinematics observed in the experiments. This does not mean, however, that the intact CNS is minimizing a function of jerk (it rather implies that the CNS does not violate the minimum-jerk principle too much) or calculating an exponential function based on a non-linear "internal time". Therefore, the actual CNS priorities still remain unknown.

The existence of choice (theoretically,at least) suggests that the CNS may wish to reconsider its priorities in certain situations wherein the components of the system for movement production are grossly changed. A change in the priorities may lead to a corresponding change in the externally observed patterns of voluntary movements. For example, Fosbury-flop (a commonly used high-jump technique) is apparently not a coordinative pattern the CNS prefers to use for jumping on an everyday basis. But, in the artificial conditions of track-and-field competition, when there are no unexpected changes in the external force field, no hidden obstacles, and there is just one priority (to clear the bar at the greatest height possible), the CNS may be "persuaded" to use a new, quite unusual pattern of coordination. So, by changing the external conditions of movement execution (the context of a motor task in a broad sense) in combination with extensive practice, one may be able to alter the CNS priorities in unimpaired individuals and force the CNS to demonstrate movement patterns that are quite different from those commonly seen in the general population, e.g., as previously mentioned, walking as seen in ballet dancers and weight-lifters.

Let us illustrate this point with a diagram that is somewhat similar to a figure by Karl Newell (e.g., Newell, van Emmerik & McDonald 1989) (Fig. 4). Control patterns are produced, according to this diagram, with equally important contributions by the system's structure, motor task (including environmental factors and intentions), and priorities of the CNS. The structure and environment will also influence the process of generation of movement patterns on the basis of the control patterns. We have not drawn feedback loops since that would have distracted the reader and harmed the nearly gothic symmetry and beauty of the Figure.

A change in the CNS priorities is likely to occur during the early stages of human life. An impressive series of studies by Thelen and her colleagues (Thelen & Fogel , 1989; Thelen, Zernicke, Schneider, Jensen, Kamm & Corbetta, 1992; Lockman & Thelen, 1993) has demonstrated rather abrupt changes in the coordination patterns during the development of infant reaching and locomotion. These developmental jumps seem to follow the discovery by the CNS of the biomechanics of its own effectors and the basic physical properties of the external force fields.

Within the dynamic pattern generation approach, the emergence of CNS priorities during normal development or their modification during adaptation to an impairment may be associated with the elaboration of equations that include functionally significant variables (state variables), neural, mechanical, perceptual, and intentional. One may think about this process as of self-organization within a complex system. The alternative, hierarchical approach would consider the same process as being controlled or led by the smart "higher" levels of a hypothetical motor control hierarchy.

Consider now the system of motor control of a chronically impaired or otherwise atypical person. His/her lifetime experience is filled with everyday voluntary movements in conditions of frequently changing goals and external forces. If the differences between this person and an average "control subject" are large enough, there is a fair chance that his/her CNS will reconsider its priorities and elaborate atypical movement patterns for solving everyday motor tasks. Thus, motor patterns in impaired or otherwise atypical populations may serve, on the one hand, to provide reflections of the changed CNS priorities, and, on the other hand, as a window into the basic, original priorities of the CNS that make the choice of unique motor patterns possible. We certainly do not wish to claim that changed CNS priorities are the only important factor defining abnormal motor patterns. An impaired system may well be genuinely unable to display movement patterns seen in the general population, amputation and complete spinal cord injury being probably the most obvious examples. However, we would like to draw attention to changes in movement patterns that may not be forced upon the system by a major chronic impairment but instead result from a person's reaction to a primary impairment.

Let us consider one more example. A car with a faulty spark plug does not reconsider its priorities and does not switch to alternative strategies. We see the ability of our body to adapt to pathological changes as a reflection of the basic differences between the design of a car and that of our body. First, the design of the car does not involve redundancy. Second, the car does not have a brain. Brain plus redundancy make the design of our body, including the system for production of voluntary movements, flexible and able to adapt not only to changes in external conditions but also, at least to some extent, to changes within the body itself.

5. Adaptive Changes in Motor Patterns of Atypical Individuals

Any major stable difference between impaired and unimpaired groups of human beings (e.g., impaired decision-making in Down syndrome; changed biomechanics in amputees; changed reflexes in spasticity; impaired pre-programming in Parkinson's disease, etc.) makes the whole system of movement generation different so that its priorities are likely to be reconsidered, and the patterns that used to be optimal for an unimpaired system are no longer optimal. Actually, even learning a simple task may lead to plastic changes within the CNS as reflected by the dynamics of a modulation of the cortical motor output (Pascual-Leone, Grafman & Hallett, 1994).

The basic assumption: The CNS may solve the problem of redundancy differently for different states of the system for movement production.

In other words, the CNS may elaborate new optimal patterns taking into account the actual (changed) state of the CNS and/or peripheral motor system and its as yet, unrevealed goals. This does not necessarily mean that the system cannot perform "normally", i.e. within the normally observed variability of motor patterns. It may be able to demonstrate "normal" patterns but prefers to behave differently. That is, "Fosbury floppish" motor behavior may well be a reflection of CNS prioritized goals in a motor impaired individual. Deciphering the peculiarities of these adaptive behaviors may be a way to understand the underlying basic goals of the motor control system.

Our basic assumption is indirectly corroborated by the observations of changes in motor patterns that may be stable and reproducible across an atypical population without being forced upon the system by a primary disorder. In particular, extensive pratice in standardized laboratory conditions or in a rehabilitation setting may make these apparently abnormal patterns look similar to those observed in unimpaired subjects. We shall consider examples of such movements as well as possible therapeutic value of teaching patients "correct movement patterns" later in this paper.

One may use another set of notions to convey the same general message. If a complex dynamic system participating in the production of voluntary movements undergoes a major change in one or more of its components, its formerly stable patterns of behavior may become unstable and lead to global changes in the behavior. For example, if a downhill skier adapts to a certain terrain, a local change in a part of the terrain may change her preferred trajectory not only at the exact location of the change but also below it. After a few runs, one is likely to see adaptive changes in the upper part of the trajectory as well. These changes may not be consciously perceived by the skier.

We believe that, for any apparently abnormal motor pattern, the first question to be asked is: What does the CNS perceive to be its primary goal during the execution of this particular motor task? A straightforward answer, e.g. to follow the exact instruction within a given task setting, may be true for motivated, unimpaired subjects, although, even in unimpaired subjects, considerations such as minimizing discomfort, or making sure that the experimental setup does not break may be as important as optimizing the performance.

Other, frequently ignored components of a motor task that may be considered important by the CNS are, for example, those related to maintaining gaze fixation, equilibrium of the head and body, and posture of the limbs with respect to the trunk during the required movements. Central control of movement involves multiple parallel commands coordinated in order to fulfill a certain goal (Arbib 1981). In particular, control of a voluntary movement requires maintaining a reference frame in conditions of possible external and internal perturbations (Massion 1992). This reference frame may refer to the position of a segment, or an extremity, or the whole body. It may also refer to a more general notion of equilibrium, i.e. keeping the projection of the body center of gravity within the support area.

A target-centered reference frame may coexist with an egocentric one. We think, however, that the egocentric frame usually plays the dominant role, and that the CNS primarily tries to preserve it. Consider, however, an interesting example described by Marsden, Rothwell & Traub (1979) when apparently the preservation of a target-centered frame was the purpose of pre-programmed responses. In these experiments, subjects performed flexion of a thumb from a certain initial to a certain final position against a constant load provided by a torque-motor. The load could be increased or decreased unexpectedly during the movement. An increase in the load led to an increase in flexor activity with a characteristic latency; a decrease in the load led to a flexor activity subsidence (see also Marsden, Merton & Morton 1972; Marsden, Merton, Morton, Adam & Hallett 1978). However, if the unloading was provided not by a decrease in the motor torque but by lifting the hand by the wrist so that the thumb was moved away from the lever, flexor activity was increased with the same characteristic latency bringing about a movement that tried to restore the thumb position on the lever. This inversion of the pre-programmed reactions from a decrease to an increase in the flexor activity may be considered an example of an action within a device-centered reference frame. The signals from the proximal limb segments did not by themselves give rise to the inverted pattern of the flexor activity changes, as demonstrated in the experiments when a subject held the device in his hand so that a similar lift by the wrist did not lead to moving the thumb from the lever.

Any voluntary movement, especially a fast one, induces a postural perturbation because of dynamic, intersegmental forces, and shifts in the center of gravity. Thus, voluntary movements may be considered self-inflicted postural perturbations that may, to a certain degree, be predicted by the CNS which may try to introduce corrections both prior to the actual perturbation (in a feedforward manner) and in response to it (via proprioceptive feedback loops).

Indeed, voluntary limb movements are nearly always associated with changes in activity of postural muscles (Belenkii, Gurfinkel & Paltsev 1967; Cordo & Nashner 1982; Bouisset & Zattara 1983, 1990; Dufosse, Hugon & Massion 1985; Brown & Frank 1987; Crenna, Frigo, Massion & Pedotti 1987; Massion 1992). Some of these changes occur prior to the movement and can be described as anticipatory. Their assumed role is to minimize perturbation of the limb or body posture that would otherwise be induced by the movement. The inability of some of the impaired persons to properly modulate these anticipatory postural corrections may force them to alter the primary movement patterns as well, e.g., to slow down. Such slowing down could be regarded as a deliberate (although not necessarily consciously perceived) strategy of the CNS, and should be considered adaptive to a primary deficit in anticipatory pre-programming.

In impaired subjects, we may expect less obvious factors to play an important role in making a choice of movement strategy. Changes in the central motor patterns may be considered secondary (adaptive) to the apparent primary changes in the systems related to movement production. Let us now discuss in more detail several examples of motor pathologies when some of the apparent abnormalities in motor patterns are likely to be consequences of adaptive changes within the CNS.

6. Motor Reorganization after a Limb Amputation

Let us start with the most straightforward example we have been able to come up with, that is amputation of a part of a limb. Here, the primary cause of the apparent motor disorders is unambiguously clear. Limb amputation leads to a major disruption of the biomechanical and neurophysiological relations developed during the lifetime. There is evidence, however, that the consequences of limb amputation may involve a major reorganization of both afferent and efferent projections that by themselves may contribute to the difference of the motor patterns from those seen in unimpaired persons.

Considerable changes in biomechanics of walking after amputation have been described by Czerniecki, Gitter & Munro (1991). In healthy subjects, ankle plantarflexors are the major energy generators while knee extensors are the major energy absorbers. The role of hip extensors is relatively small. But, in below knee amputees, ankle plantarflexors are obviously unavailable, and hip extensors become the main source of energy absorption and generation. This rearrangement should be considered adaptive since it allows amputees to walk even though the gait may occur in a less energy efficient manner than observed in unimpaired walking.

Consequences of amputation also involve neurological reorganization at both segmental and suprasegmental levels. Obviously, the elimination of a considerable number of proprioceptors residing in the amputated portion of the leg leads to an abrupt change in the patterns of afferent inflow and is likely to lead to changes in relative weight of contribution of other, seemingly unaffected reflex projections. Reflex contribution is considered an important factor in the natural patterns of voluntary movements (Feldman 1986; Latash 1993), with inter-muscular and inter-joint reflexes playing an important role (Nichols 1989). Descending motor commands should apparently take into account the existing state of reflex connections. Besides, proprioceptive inflow is used in the process of generation of automatic, pre-programmed adjustments in the activity of muscles providing postural stability during voluntary movements (for a review see section 2.3 and also Massion 1992). Thus, amputation of a distal portion of a leg may be expected to lead to both a rearrangement of descending motor commands and a shift of postural control from predominantly proprioception-based to other modalities, e.g., visual and vestibular signals.

The flexibility (plasticity) of the projections within the CNS represents one of its most remarkable features and is likely to contribute to the processes of motor learning and adaptation to trauma. Since the classical works of Lashley (1933), it has been proposed that brain injury may lead to a dramatic topographic reorganization in the adjacent areas which may significantly contribute to recovery after stroke (also see Luria 1973; Jenkins & Merzenich 1987). Changes in peripheral afferent flow have been shown to induce changes in the receptor field sizes and locations in brain cortex of the cat (Kalaska & Pomeranz 1979; Metzler & Marks 1979; Franck 1980; Recanzone, Allard, Jenkins & Merzenich 1990). Somatosensory cortical representations (area 3b) in monkeys have been shown to change after a specific training of one hand (Jenkins , Merzenich & Recanzone 1990; Recanzone, Merzenich, Jenkins, Grajski & Dinse 1992), and after digit amputation or fusion (Merzenich, Nelson, Stryker, Cynader, Schoppman & Zook 1984; Jenkins et al., 1990). CNS plasticity is not limited to grossly changed pathological states and supraspinal structures as demonstrated by the pioneering experiments by Wolpaw and colleagues (Wolpaw 1983; Wolpaw, Braitman & Seegal 1983).

Neurological reorganization of descending control signals after a below-knee amputation in humans was studied with transcranial magnetic stimulation by Fuhr, Cohen, Dang, Findley, Haghighi, Oro & Hallett (1992). In this study, stimuli at optimal positions of the coil recruited a larger percentage of a-motoneurons controlling the muscles in the residual leg. These muscles could also be activated from larger areas of the scalp than the muscles at the intact side. Similar results, also in human subjects, have been reported after upper limb amputation (Cohen, Bandinelli, Findley & Hallett 1991). More recently, Kew, Ridding, Rothwell, Passingham, Leigh, Soorikumaran, Frackowiak, and Brooks (1994) studied cortical blood flow during attempts at voluntary shoulder movements as well as corticospinal excitability with transcranial magnetic stimulation in traumatic and congenital upper limb amputees. This study has shown, in particular, an increase in cortical blood flow in the partially deafferented M1/S1 during movements of the ipsilateral, intact arm in traumatic amputees but not in congenital amputees. Moreover, transcranial magnetic stimulation of the partially deafferented M1 revealed increased corticospinal excitatibility only in traumatic amputees but not in congenital amputees. These observations suggest that adaptive changes in the human cortex after an amputation may proceed differently during early development as compared to adult life.

The necessity to adapt the learned motor patterns after a lower limb amputation is relatively obvious and has not been seriously challenged. However, in cases when the primary cause is not so apparent to an outside observer, e.g., when it represents an imbalance within the hypothetical central controller of voluntary movements or an impairment at a "higher" level of decision-making, the primary aims of rehabilitation and physical therapy are frequently seen as bringing the peripheral patterns of movements as close to those seen in unimpaired persons as possible (Larkin & Hoare, 1991; Lieber & Bodine-Fowler, 1993; Lockwood, 1987; Shepherd, 1984; Stein, Yang, Belanger & Pearson, 1993). The next several examples will try to persuade the reader that the role of adaptive changes may be very significant in populations characterized by motor patterns that differ from the commonly observed ones.

7. Pre-Programming in Parkinson's Disease

There are four basic clinical features of Parkinson's disease: tremor, bradykinesia, rigidity, and deficit in postural reflexes (Fahn 1990). Tremor is characterized by a 5-6 Hz alternating activity of antagonist muscles controlling a joint leading to alternating joint movements that can be seen both at rest and during voluntary movements in the joint. Bradykinesia usually refers to slowness of voluntary movements and difficulty in movement initiation. It can affect any part of the body and be more or less generalized. Rigidity is a sustained increase in the resistance to externally imposed joint movements. Deficits in postural reflexes reveal themselves as poor modulation of anticipatory and/or pre-programmed changes in activity of postural muscles associated with voluntary movements or in response to external perturbation.

When all the parameters of a simple movement are known in advance, patients with Parkinson's disease initiate and perform the movement slower. In particular, they demonstrate an increase in reaction time (Heilman et al., 1976; Evarts et al., 1981; Stelmach et al., 1986; Brown & Marsden, 1991) which increases with an increase in movement complexity (Sanes, 1985). Movement time in Parkinson's disease demonstrates a linear increase with an increase in the ratio of target size to movement distance (Halsband et al., 1990) conforming to the Fitts' law. However, the slope of this relation is considerably greater than in control subjects leading to longer movement times at similar levels of index of difficulty. Contrary to these findings, Draper and Johns (1964) have reported nearly constant peak velocity in Parkinson's disease when movement amplitude was changed. An increase in movement time in Parkinson's disease is accompanied by a considerable asymmetry of the acceleration and deceleration phases (Inzelberg et al., 1990) and is mostly due to a prolongation of the deceleration phase. Isenberg & Conrad (1994) also reported a change in the shape of movement trajectories and an impaired synchronization of vertical and horizontal velocity components.

Both temporal and spatial motor variability of targeted limb movements has been shown to be higher in Parkinson's disease (Sheridan et al., 1987) which the authors attribute to an increased "inherent" variability in muscle force production (see also Wierzbicka et al., 1991). Sheridan and Flowers (1990) have suggested that Parkinson's bradykinesia may in part result from the increased variability in order to preserve an acceptable level of accuracy, i.e. be a result of a compensation rather than a primary deficit .

Fast single-joint voluntary limb movements of patients with Parkinson's disease are characterized by relations between load, kinematics, and EMGs similar to those observed in the control population. However, the entire targeted movement is frequently constructed of several discernible segments with the EMG patterns demonstrating a number of repeated cycles of agonist-antagonist bursts (Hallett & Khoshbin, 1980; Berardelli et al., 1984, 1986; Inzelberg et al., 1990). A considerable amount of cocontraction of antagonist muscles (Hayashi et al., 1988) can also be a factor disrupting both EMG and kinematic patterns during voluntary movements.

Berardelli et al. (1986) report that, during wrist flexions, the EMG does not saturate but can be modulated, within limits, to the amplitude of the movement. Patients with Parkinson's disease are able to produce accurate force levels (Stelmach & Worringham, 1988b) although the control of the rate of force increase and decrease seems to be more affected (Wing, 1988). These observations suggests that the problem is not in achieving absolute levels of EMGs but rather in the dynamics of muscle activation which may depend considerably upon the action of reflex feedback loops and changes within the segmental apparatus.

Patients with Parkinson's disease demonstrate profoundly different pre-programmed reactions. Stretching of a quiescent or voluntarily activated muscle of a patient with Parkinson's disease leads to long-latency muscle responses whose amplitude is considerably higher than in the control population (Berardelli et al., 1983; Rothwell et al., 1983; Cody et al., 1986; Hunter et al., 1988). This increase has been attributed to an "overcompensation" in the transmission in a hypothetical receptor-motor cortex-muscle loop (trans-cortical loop, Lee & Tatton, 1975) and has been considered a possible mechanism of parkinsonian rigidity (Tatton & Lee, 1975; Mortimer & Webster, 1979; Berardelli et al., 1983; Rothwell et al., 1983). The capacity to suppress the pre-programmed reactions induced by postural perturbations has been shown to be impaired in Parkinson's disease (Schieppati & Nardone, 1991). There is substantial variability among the studies of anticipatory postural adjustments in Parkinson's disease. These reports vary from minor changes in the anticipatory reactions in Parkinson's disease (Diener, Dichgans, Guschlbauer, Bacher, & Langenbach, 1989; Latash, Aruin, Neyman, & Nicholas, 1995), and include, no differences in the timing of the early EMG bursts in postural muscles while the amplitude of the bursts is decreased (Dick, Rothwell, Berardelli, Thompson, Gioux, Benecke, Day, & Marsden, 1986), smaller EMG changes in muscles involved in a postural component of a bi-manual task, often, lacking anticipatory EMG changes (Viallet, Massion, Massarino, & Khalil, 1987), and the lack of anticipatory postural adjustments in 95% of patients with Parkinson's disease (Bazalgette et al., 1986; cf. Traub et al., 1980a).

Since the primary cause of Parkinson's disease is undoubtedly supraspinal, it has been suggested that motor disorders in Parkinson's disease are due to changes in the descending motor commands. In particular, a basic deficit in the composition of complex sequences of motor programs has been hypothesized (Evarts, Teravainen, & Calne, 1981; Sanes, 1985; Isenberg & Conrad, 1994). The segmental apparatus has been assumed to be generally intact (e.g., Hallett, 1993). This view is corroborated, in particular, by observations of unchanged tendon jerk reflexes in PD patients (Rothwell et al., 1983) and generally normal short-latency action of Ia muscle afferents (Matthews et al., 1990). However, a number of changes in presumably segmental mechanisms have been reported. They include a deficit in reciprocal inhibition that could, in particular, contribute to the co-contraction of antagonist muscles during voluntary movements, increased reflex activity during tracking phases in which the muscle is lengthening (Johnson et al., 1991), and a "paradoxical" Westphal phenomenon (Andrews et al., 1972; Lee et al., 1983; Berardelli & Hallett, 1984; Matthews et al., 1990). The Westphal phenomenon represents an abrupt reflex excitation of a muscle in response to an externally imposed movement leading to a decrease in the muscle length. In a sense, it is an inverse of the stretch-reflex. Also, Glendinning and Enoka (1994) described secondary changes in motor neuron firing patterns secondary to disuse in Parkinson's disease.

We are now going to suggest an extreme hypothesis that most of the motor problems in Parkinson's disease are adaptive to a single primary cause. We are aware of the data suggesting different mechanisms for different symptoms of Parkinson's disease; however, our goal is not to prove our hypothesis but rather to demonstrate how apparently different motor abnormalities may result from adaptive processes within the CNS. Let us assume that the primary dysfunction in Parkinson's disease includes problems in pre-programming. If the ability to pre-program motor corrections is impaired and there is no compensation, the most commonly used motor programs may become useless since any external perturbation would lead to their global disruption. Let us suppose that the CNS still wants to use some of the programs that require corrections on the basis of pre-programming like walking and maintenance of vertical posture. The necessary pre-programmed reactions are stored in the memory but the mechanism of their triggering is defective. What kind of adaptive behavior may be expected from such a system?

First the CNS is likely to prefer to move at lower speeds (bradykinesia). Second, it may try to compensate for the impaired ability to adequately pre-program by decreasing the triggering threshold for the pre-programmed corrections and/or increasing their gain thus leading to a likely overcompensation. One of the consequences of an excessively strong corrective movement can be a new perturbation leading to a pre-programmed reaction in the opposite direction.

Several results can be expected from such a compensatory mechanism. First, resistance of the system to externally imposed movements will increase (cf. rigidity). Second, oscillations can occur with a period corresponding to slightly more than double the latency of the pre-programmed reactions due to the time necessary for the peripheral receptors to react to a perturbation induced by a preceding pre-programmed reaction. This assessment leads to a value of about 5-6 Hz (cf. parkinsonian tremor). Note, however, that there is a strong evidence for the central nature of the Parkinsonian resting tremor (Lee & Stein 1981; Lee, Murphy & Tatton 1983). Third, walking and standing will be possible although they are likely to look awkwardly "rigid". In this framework, only the loss of postural reflexes seems to be a reflection of the primary dysfunction in pre-programming.

Bouisset & Zattara (1990) formulated the following question: Is bradykinesia a consequence of the pathological postural reactions or are the postural reactions reduced secondarily to the decrease in movement speed? They favor the first hypothesis which fits well with our general scheme. The system may prefer to function sub-optimally (but relatively reliably) rather than risk the total failure. The view that patients with Parkinson's disease prefer safer motor strategies is also supported by the observation of a much higher safety margin demonstrated by patients with Parkinson's disease in precision grip tasks (Muller & Abbs 1990; cf. with the results of Cole, Abbs & Turner 1988 in subjects with Down syndrome described in section 10).

8. Muscle Coactivation in Dystonia

Dystonia is a poorly understood supraspinal pathology which is not accompanied by any gross anatomical or neurophysiological changes (Fahn 1988; Ghez, Gordon & Hening 1988; Marsden 1988; Hallett, 1993). Its most prominent features are seen at the level of multi-joint movements, frequently involving neck and trunk movements with pronounced rotational components (Fahn 1988; Marsden 1988). Single-joint movements in dystonic patients are characterized by trajectory oscillations, hesitations, reversals, and multiple EMG bursts (Ghez et al. 1988; Marsden 1988). Recently, Latash and Gutman (1994) have suggested that dystonia is a problem of imbalance within the descending signals that may not be correlated with any discernible pathology in the supraspinal or spinal structures. These authors modeled dystonic movements in the framework of the equilibrium-point hypothesis (Feldman 1986; Latash 1993) assuming that the hypothetical imbalance of the control signals may lead to transient changes in the gain of the tonic stretch reflex for certain muscles. It was shown that small variations in the patterns of the assumed changes may lead to considerably different trajectories and EMG patterns thus accounting for the notoriously high variability, and even uniqueness of dystonic movements (Fahn 1988; Ghez et al. 1988; Marsden 1988). Within this approach, some of the reported segmental abnormalities in dystonia (e.g., a deficit in the Ia-mediated inhibition and the Westphal phenomenon; Rothwell, Obeso, Day & Marsden 1983; Nakashima, Rothwell, Day, Thompson, Shannon & Marsden 1989) may represent secondary changes of the system in response to a chronic primary disorder of motor control.

According to our basic idea, a hypothetical supraspinal pathology that underlies dystonia is likely to be accompanied by both primary and adaptive changes in patterns of voluntary movements. In particular, we suggest that the coactivation of antagonist and remote muscles (Herz 1944; Hoefer & Putnam 1940; Hughes & McLelland 1985; Fahn 1988) frequently seen in dystonic patients represents an attempt to create a compensatory, although ineffective strategy. One may call it a "strategy of desperation". According to the basic assumption by Latash & Gutman (1994), patients with dystonia use normal control patterns that should have lead to smooth, monotonic trajectories. However, the resulting trajectories are each time very different from what the person is expecting. From our everyday experience, we know that if a limb trajectory is perturbed by an external force, an effective strategy to counteract the perturbation is to increase joint or limb stiffness by co-contracting antagonist muscles or muscles controlling other joints of the limb. The consequences of the hypothesized imbalance in descending signals may be thought of as perturbations generated within the system. Such perturbations are so unusual, that the CNS may still try to use corrective strategies that are effective against external perturbations, i.e. to coactivate all the limb muscles. Unfortunately for the patients, this strategy is ineffective against trajectory perturbations generated by the assumed mechanism, and additional muscle coactivation fails to stabilize joint trajectory.

9. Changed Postural and Locomotor Patterns in Different Populations

Patients with relatively mild vestibular disorders demonstrate a peculiar pattern of walking resembling cross country-skiing (or Groucho-type walking) in which head displacements in the vertical direction are minimized. Since head stabilization is a major component of the postural control system during walking (e.g., Pozzo, Berthoz & Lefort 1990), occurrences of such peculiar walking patterns might have been the consequences of an impaired ability of these patients to stabilize the head. Attempts to return the walking pattern of these patients to approximate "normal" may have negative influences. In particular the head may start to exhibit high-amplitude movements resembling those seen in patients with severe vestibular disorders.

A similar conclusion can be drawn from the observations of patients with vestibular disorders who use relatively large displacements in their ankles for postural corrections (an "ankle strategy") but lack compensatory hip movements (a "hip strategy") which are commonly seen in unimpaired subjects (Shumway-Cook & Horak 1989; Horak, Nashner & Diener 1990). One reason for this may be that a shift in the body center of gravity performed by changing position in the ankle joints leads to a smaller displacement of the head than a similar shift performed by a movement in the hip joints.

A recent paper by Winter, Ruder & MacKinnon (1990) presents a number of cases of abnormal walking patterns in patients with various "primary causes" including amputation, joint replacement, and spastic hemi-syndromes following brain injury. Analysis of gait biomechanics has led the authors to conclude that many of the atypical features represent results of adaptation and should not be considered pathological. Winter et al. write: "...in cases of major surgery...or in long-term therapy it is important not to treat the adaptations (secondary problems) but to treat the primary problems." (p. 692).

If a primary deficit is inborn, e.g., Down syndrome and mild cerebral palsy, atypical locomotor and/or postural patterns develop differently from the very beginning of the individual's life rather than emerge as a result of an adaptation to a primary deficit.

10. Slow Movements and Changed Pre-Programming in Down Syndrome

Motor disorders in Down syndrome are commonly addressed as "clumsiness". The word "clumsiness" is used to indicate movements that look different from and less efficient than those observed in the general population. Two major components of clumsiness in DS include slowness of the movements and the inability to rapidly respond to the changing environment (Kerr and Blais 1985; Lincoln, Courchesne, Kilman & Galambos 1985; Shumway-Cook and Woollacott 1985; Cole et al., 1988; Nativ and Abbs 1989). The latter factor is manifested in the laboratory studies as a deficit in pre-programming and longer reaction times (Shumway-Cook & Woollacott 1985; Anson 1989, 1992; Latash & Corcos 1991). Other differences in motor performance of DS persons (see Anson, 1992; Latash, 1992; Sugden & Keogh 1990 for reviews) include low muscle tone and correlated low voluntary muscle contraction force (Rarick et al., 1976; Morris et al., 1982), excessive forces in such tests as finger tapping or handgrip (Frith & Frith 1974; Henderson et al. 1981; Cole et al. 1988), increased variability in various aspects of motor performance (Frith & Frith 1974; Henderson 1985; Elliott et al. 1986; Elliott & Weeks 1990; Shumway-Cook & Woollacott 1985), and the lack of adaptation to changes in sensory information (Shumway-Cook & Woollacott 1985; Cole et al. 1988; Nativ & Abbs 1989).

Discrete single-joint movements of individuals with DS are typically slow and frequently consist of several distinct submovements (Latash & Corcos 1991). In some trials, movement kinematics may be characterized by a normal-looking bell-shaped velocity profile. Other trials in the same series, however, can demonstrate irregular trajectories with visible "bumps" and possible reversals of movement direction, and multiple bursts of activity in the agonist and antagonist muscles.

Commonly, clumsiness has been assumed to originate from a malfunction of a subsystem within the general system of movement production (e.g., Larkin & Hoare 1992). One function of this subsystem is likely to be information processing including decision-making. Mental retardation which is almost always associated with DS, could affect decision-making by delaying the accumulation and translation of information specific to the stimulus and motor response. Our basic assumption implies that the CNS of a person whose decision-making component of the system for movement production is in some way impaired, may "prefer" to facilitate clumsy movements rather than risk total failure during motor task performance. During a lifetime, the CNS accumulates experiences that would allow it to predict that unexpected changes in external conditions occur rather frequently and can include changes in the movement goal, in the external forces, in the inertial loading, etc. Therefore, if the CNS is aware of its impaired ability to make quick, adequate decisions, it may be reluctant to produce motor commands leading to very fast movements in order to have more time for evasive actions or corrections in response to a change in the environment (perturbation) and/or to attenuate potentially damaging effects of the perturbation.

In children with DS, pre-programmed postural reactions in response to perturbations during standing demonstrate less coupling among the synergetic muscles (Shumway-Cook & Woollacott 1985). These reactions have been reported as a distal-to-proximal sequencing of joint involvement, similar to unimpaired children but with greater delays. Onset latencies for postural reactions were considerably longer in DS.

Pre-programming during unidirectional single-joint movements in unimpaired control subjects usually involves a "reciprocal pattern" of muscle activation, i.e. an unexpected loading leads to an increase in the agonist activity, while an unexpected unloading leads to a decrease in the agonist activity with a possible increase in the activity of the antagonist (Houk 1976; Marsden, Merton, Morton, Rothwell & Traub 1981; Bonnet 1983; Rothwell, Day, Berardelli & Marsden 1986). Subjects with Down syndrome have frequently demonstrated a coactivation pattern of preprogramming that involves an increase in activity of both agonist and antagonist muscles irrespective of the direction of a perturbation (Latash 1992; Latash, Almeida & Corcos 1993; Almeida, Aruin & Latash, 1994). Should this difference be considered a sign of an inability of the system of pre-programming to behave "correctly" or is this a sign of the preferred strategy for a changed CNS?

Coactivation could represent the consequence of an altered (impaired) mechanism of pre-programming or it could represent a "safety-catch" imposed by the CNS to allow movement to be controlled within the constraints of its impaired operating capacity. Novices in the early stages of acquiring a new motor skill frequently demonstrate greater than optimal levels of cocontraction which appear to increase stability and reduce the likelihood of error (Rosenbaum, 1991). This cocontraction commonly disappears after the skill is well learned. Thus, we think that muscle coactivation is likely to reflect active intervention by the CNS rather than its inability to use "more normal" patterns of muscle activation.

If the reciprocal strategy is used, pre-programming an increase in activity of a "wrong" muscle group can lead to exacerbation of the effects of the perturbation. The coactivation strategy is more universal in the sense that it leads to an attenuation of the effects of perturbations independently of the perturbation direction. On the other hand, it is always suboptimal, since it cannot, in principle lead to total compensation of the effects of perturbation. This may be the reason why this strategy has not been reported for highly practiced control subjects who prefer to use the more effective although more challenging reciprocal strategy. Apparently, this strategy is within a safety zone established by their unimpaired CNS.

If a person is confident in his/her abilities to make quick judgments concerning concurrent motor tasks, current conditions of their execution, and their possible future changes, pre-programming is a strong albeit suboptimally effective mechanism for introducing urgent corrections (Houk 1978; Marsden et al. 1981; Chan & Kearney 1982; Bonnet 1983). If the individual's ability to make these decisions is impaired, it may become preferable to adjust the control mechanisms suppressing potentially harmful ones that are likely to include very fast movements and modulation of pre-programmed reactions. In the reproducible and friendly conditions of the laboratory, these internal restrictions may be lifted leading to virtually normal performance in motor tests. In particular, in our experiments (Latash et al., 1993), subjects with Down syndrome who were well-practiced frequently demonstrated a mixture of reciprocal and coactivation patterns of pre-programming in different trials within the same series or block of trials.

In another study (Almeida, Corcos & Latash 1994), prolonged practice of single-joint elbow flexion movements "as fast as possible" led to a striking improvement in the performance transferable to different distances, and different initial and final positions. The question of whether an improvement acquired in a standardized laboratory environment may benefit everyday movements performed in much less reproducible conditions remains open. We are cautiously pessimistic. When the CNS for the first time encounters unpredictable perturbations, it may quickly return to the old, reliable, safe patterns. It is possible that practice with an element of uncertainty may be successful in persuading the CNS that it is able to reconsider its priorities and shift to more effective albeit more challenging modes of control.

11. The Order of Joint Involvement in a Simple Reaction Time, Multi-Joint Reaching Task

In "normal" subjects, a proximal-to-distal order is commonly observed (e.g., Joris, Edwards van Muyen, van Ingen Schenau & Kemper 1985; van Ingen Schenau 1989) which is assumed to be mechanically more efficient for a unidirectional reaching movement. In a recent study, Anson (1992) has described an inverted order of joint involvement in a simple reaction time task in young adults with Down syndrome. Control subjects who performed an identical task with the same instructions as received by the Down syndrome subjects produced proximal-to-distal sequencing as reflected in both reaction time and premotor time (Mawston & Anson, in press). Should one consider the inverted order of joint involvement in Down syndrome individuals as a sign of an inability of the CNS to recruit joints in the "correct" proximal-to-distal order?

Note, that in an earlier study, Kaminski & Gentile (1986) reported that the order of joint involvement in healthy subjects could depend on relative amplitude of displacement in individual joints, so that the joint with bigger amplitude started earlier. In conditions where joint amplitudes were the same, the order of involvement was proximal-to-distal. If amplitude in a distal joint was greater, a distal-to-proximal sequencing could be observed. This finding may, however, reflect the method for detecting the beginning of joint motion used by Kaminski and Gentile, a moment when joint speed exceeded 5% of the peak value observed during the movement. This would be considerably later than commonly used measures of reaction time. Their finding, therefore, may reflect a difference in relative initial acceleration in the joints, so that they might reach the threshold at different times while their actual involvement could happen simultaneously or in a fixed proximal-to-distal order. These methodological problems are similar to those encountered when studying electromechanical delay (for a review see Corcos, Gottlieb, Latash, Almeida & Agarwal 1992).

Let us consider these findings from the viewpoint of our basic assumption. When a subjects is given an instruction "to reach a target as quickly as possible after a signal", he/she may, in fact, understand it in two different ways:

1). Directly, i.e. get to the target as fast as possible (minimize movement time).

2). Initiate the movement as quickly as possible and get to the target.

Note that distal-to-proximal sequencing of joint involvement is optimal for the second form of the instruction because of the low inertia of the distal segment but not for the first one. On the other hand, proximal-to-distal sequencing is mechanically more efficient for the first form of the instruction but not for the second one.

Why should a subject who has Down syndrome understand the instruction in two different ways? Down syndrome commonly results in a degree of mild to moderate mental retardation (Epstein, 1987). Decision-making in a simple reaction time task, although apparently trivial for control subjects (GO when you hear the tone), may be difficult for subjects with Down syndrome. Let us also take into account that an experimenter is likely to watch the movements during practice and during the actual experiment, and to express (willingly or not) his/her satisfaction/dissatisfaction with the performance. In particular, the signs of "happiness" on the face of the experimenter may correlate with quickly initiated movements rather than with those having the shortest movement time. In this case, there is a conflict between the formal instruction and reinforcement during the experiment. It is up to the subject to decide which way to go.

Down syndrome subjects may be more responsive to verbal and non-verbal reinforcement during the training and experimental procedure rather than to a formal instruction. Therefore, they are more likely to act according to the experimenter's reaction to their performance, i.e. to optimize the signs of happiness that correlate with quickly initiated movements rather than with quickly completed ones. This may represent an important difference between control subjects who stubbornly stick to the formal instruction and more flexible Down syndrome individuals. In this sense, it could be argued that the Down syndrome subjects understand the instruction better!

There is a relatively straightforward way to check this hypothesis. The subjects (control and Down syndrome) should be given one of two instructions: "To initiate a movement as quickly as possible" or "To touch the target as quickly as possible". Reinforcement during practice should closely follow the instruction and preferably be based on an objective measure. The prediction is that in both groups one may expect a reversal of the order of joint involvement when the instruction is changed from an emphasis on initiation of movement to an emphasis on hitting the target.

We would like to finish the main body of our paper with an example of an experiment that could be used to test our main hypothesis that a rearrangement of CNS priorities may play a major role in generating apparently abnormal motor patters. Mauerberg, Schuller, and Fantucci (1994) have shown that stride length during both walking and running in persons with Down syndrome is shorter than in control subjects. There are two basic ways of trying to train a person with Down syndrome to run faster. First, one can start with teaching them how to make strides of a "normal" length, for example by drawing lines on the floor. Then, they may be encouraged to run faster while keeping the "normal" relation between running speed and stride length. An alternative way is to encourage them to run faster using their preferred stride length. Apparently, the second strategy is based on considering naturally occurring changes in stride length in persons with Down sydrome as adaptive. We predict (other factors being equal), that the second group will beat the first one in a race. Moreover, we predict that the second group will demonstrate a sustained effect of practice transferable to different conditions (e.g., additional loading, different properties of the running surface, fatigue, etc.) while the first group will have major difficulties in keeping the "normal" running pattern under a minimal change in the external conditions.

12. Concluding Comments

The basic message is as follows:

If you see, in a clinical setting or in an experiment, a motor pattern that is very different from what is observed in unimpaired, control subjects, do not jump to the conclusion that it is a sign of inability of the CNS to behave correctly, and, as such, should be corrected.

Rather, an apparently abnormal motor pattern should be viewed as a sign that the CNS has rearranged its priorities and crafted a new subset of solutions from an infinite stock supplied through the redundancy of the motor system which exists for most of the natural and experimentally studied voluntary movements. Getting back to the title, we would like to suggest that adaptive changes in motor patterns should be considered normal and, as such, should not be corrected.

In course of the present paper, we have purposefully downplayed the role of primary disorders as compared to the role of adaptive changes within the CNS. Certainly, if the spinal cord of a person is cut, most of the disabling changes in voluntary movements and reflexes are consequences of the primary disorder. However, in cases of less crude and acute structural changes or changes at a functional rather than structural level, the relative role of the adaptive changes is likely to increase considerably. This may be particularly true for cognitive deficits when motor behavior may demonstrate consistent deviations from "normal" without any clear neurological or structural deficits, movements of persons with schizophrenia being a good example (Rosen et al., 1993). So, our purpose has been to draw attention to the frequently overlooked adaptive factor contributing to disordered movements.

Our present lack of understanding of the basic principles of motor control prevents us from making recommendations about how to distinguish between primary and adaptive changes in motor patterns. Obviously, one needs to know the original basic priorities of the CNS in order to be able to study deviations from these priorities. By the same token, any therapeutic implications are tentative and limited to the friendly "hands-off" advice formulated at the beginning of these concluding comments.

Acknowledgments

This work was in part supported by a grant HD 30128 from the National Center for Medical Rehabilitation Research, NIH and by a grant from the Down's Syndrome Research Fund (Chicago, USA).

References

Almeida, G. L., Aruin, A. S., & Latash, M. L. (1994) Organization of a simple two-joint synergy in individuals with Down syndrome. Brazilian International Journal of Adapted Physical Education Research 1: 141-142.

Almeida, G. L., Corcos, D. M. & Latash, M. L. (1994) Practice and transfer effects during fast single joint elbow movements in individuals with Down syndrome. Physical Therapy 74: 1000-1016.

Andrews, C. J., Burke, D. & Lance, J. W. (1972) The response to muscle stretch and shortening in Parkinsonian rigidity. Brain 95: 795-812

Anson, J. G. (1989) Down Syndrome: Neuromotor programming and fractionated reaction time. In: Motor Control in Down Syndrome. ed. M. L. Latash, pp. 6-13, Rush Medical Center, Chicago

Anson, J. G. (1992) Neuromotor control and Down syndrome. In: (ed.) Approaches to the Study of Motor Control and Learning. ed. J. J. Summers, pp. 387-412, N-Holland: Amsterdam.

Arbib, M. A. (1981) Perceptual structures and distributed motor control. In: Handbook of Physiology, Section 1, The Nervous System, vol. II, part 2, ed. V. B. Brooks, pp. 1449-1480, Amer Physiol Soc: Bethesda.

Bazalgette, D., Zattara, M., Bathien, N., Bouisset, S., Rondot, P. (1986) Postural adjustments associated with rapid voluntary arm movements in patients with Parkinson's disease. Advances in Neurology 45: 371-374

Belenkii, V. Y., Gurfinkel, V. S. & Paltsev, Y. I. (1967) Elements of control of voluntary movements. Biophysics 12: 135-141.

Berardelli, A., Dick, J. P. R., Rothwell, J. C., Day, B. L. & Marsden, C. D. (1986) Scaling of the size of the first agonist EMG burst during rapid wrist movements in patients with Parkinson's disease. Journal of Neurology, Neurosurgery, and Psychiatry 49: 1273-1279

Berardelli, A. & Hallett, M. (1984) Shortening reaction of human tibialis anterior. Neurology 34: 242-246

Berardelli, A., Sabra, A. F., Hallett, M., Berenberg, W. & Simon, S. R. (1983) Stretch reflexes of triceps surae in patients with upper motor neuron syndromes. Journal of Neurology, Neurosurgery, and Psychiatry 46: 54-60

Berardelli, A., Rothwell, J. C., Day, B. L. & Marsden, C. D. (1984) Movements not involved in posture are abnormal in Parkinson's disease. Neuroscience Letters 47: 47-50

Bonnet, M. (1983) Anticipatory changes of long-latency stretch responses during preparation for directional hand movements. Brain Research 280: 51-62.

Bouisset, S. & Zattara, M. (1983) Anticipatory postural movements related to a voluntary movement. In: Physiologie Spatiale, pp. 137-141, Toulouse: Cepadues Editions

Bouisset, S. & Zattara, M. (1990) Segmental movement as a perturbation to balance? Facts and concepts. In: Multiple Muscle Systems. Biomechanics and Movement Organization, eds. J. M. Winters & S. L.-Y. Woo, pp. 498-506, New York e.a.:Springer-Verlag

Brown, R. G. & Marsden, C. D. (1991) Dual task performance and processing resources in normal subjects and patients with Parkinson's disease. Brain 114: 215-231

Brown, J. E. & Frank, F. S. (1987) Influence of event anticipation of postural actions accompanying voluntary movement. Experimental Brain Research 67: 645-650

Buchanan, T. S., Almdale, D. P. J., Lewis, J. L. & Rymer, W. Z. (1986) Characteristics of synergetic relations during isometric contractions of human elbow muscles. Journal of Neurophysiology 56: 1225-1241

Buchanan, T. S., Rovai, G. P. & Rymer, W. Z. (1989) Strategies for muscle activation during isometric torque generation at the human elbow. Journal of Neurophysiology 39: 925-935

Chan, C. W. Y. & Kearney, R. E. (1982) Is the functional stretch reflex servo controlled or preprogrammed? Electroencephalography Clinical Neurophysiology 53: 310-324

Cody, F. W. J., MacDermott, N., Matthews, P. B. C. & Richardson, H. C. (1986) Observations on the genesis of the stretch reflex in Parkinson's disease. Brain 109: 229-249

Cohen, L. G., Bandinelli, S., Findley, T. W. & Hallett, M. (1991) Motor reorganization after upper limb amputation in humans: a study with focal magnetic stimulation. Brain 114: 615-627

Cole, K. J., Abbs, J. H. & Turner, G. S. (1988) Deficits in the production of grip force in Down Syndrome. Developmental Medicine and Child Neurology 30: 752-758

Corcos, D. M., Gottlieb, G. L., Latash, M. L., Almeida, G. L. & Agarwal, G. C. (1992) Electromechanical delay: An experimental artifact. Journal of Electromyography and Kinesiology 2: 59-68.

Cordo, P. J. & Nashner, L. M. (1982) Properties of postural adjustments associated with rapid arm movements. Journal of Neurophysiology 47: 287-302

Crenna, P., Frigo, C., Massion, J. & Pedotti, A. (1987) Forward and backward axial synergies in man. Experimenal Brain Research 65: 538-548

Czerniecki, J. M., Gitter, A. & Munro, C. (1991) Joint moment and muscle power output characteristics of below knee amputees during running: The influence of energy storing prosthetic feet. Journal of Biomechanics 24: 63-75

Dick, J. P. R., Rothwell, J. C., Berardelli, A., Thompson, P. D., Gioux, M., Benecke, R., Day, B. L., & Marsden, C. D. (1986) Associated postural adjustments in Parkinson's disease. Journal of Neurology, Neurosurgery, and Psychiatry 49: 1378-1385.

Diener, H. C., Dichgans, J., Guschlbauer, B., Bacher, M., & Langenbach, P. (1989) Disturbances of motor preparation in basal ganglia and cerebellar disorders. Progress in Brain Research 80: 481-488.

Draper, I. T. & Johns, R. S. (1964) The disorders of movement in Parkinsonism and the effect of drug treatment. John Hopkins Hospital Bulletin 115: 465-480.

Dufosse, M., Hugon, M. & Massion, J. (1985) Postural forearm changes induced by predictable in time or voluntary triggered unloading in man. Experimenal Brain Research 60: 330-334.

Elliott, D., Weeks, D. J. (1990) Cerebral specialization and the control of oral and limb movements for individuals with Down's syndrome. Journal of Motor Behavior 22: 6-18.

Elliott, D., Weeks, D. J., Jones, R. (1986) Lateral asymmetries in finger-tapping by adolescents and young adults with Down's syndrome. American Journal of Mental Deficiency 90: 472-475.

Epstein, C. J. (1987). Down syndrome. In: Adelman, G. (ed.) Encyclopedia of Neuroscience, Birkhauser: Boston, MS, pp. 336-337.

Evarts, E. V., Teravainen, H. & Calne, D. B. (1981) Reaction time in Parkinson's disease. Brain 104: 167-186

Fahn, S. (1988) Concept and classification of dystonia. In: Fahn, S., Marsden, C.D. and Calne, D.B. (eds.) Dystonia 2, Advances in Neurology, vol. 50, Raven Press: New York, pp. 1-8.

Feldman, A. G. (1986) Once more on the equilibrium-point hypothesis (l model) for motor control. Journal of Motor Behavior 18: 17-54

Flanders, M. & Soechting, J. F. (1990) Arm muscle activation for static forces in three-dimensional space. Journal of Neurophysiology 64: 1818-1837.

Flash, T. (1987) The control of hand equilibrium trajectories in multi-joint arm movements. Biol Cybern 57: 257-274

Flash, T. & Hogan, N. (1985) The coordination of arm movements: An experimentally confirmed mathematical model. Journal of Neuroscience 5: 1688-1703

Franck, J. I. (1980) Functional reorganization of cat somatic sensory-motor cortex (SmI) after selective dorsal root rhisotomies. Brain Research 186: 458-462.

Frith, U. & Frith, C. D. (1974) Specific motor disabilities in Down's syndrome. Journal of Child Psychology and Psychiatry 15: 293-301

Fuhr, P., Cohen, L. G., Dang, N., Findley, T. W., Haghighi, S., Oro, J. & Hallett, M. (1992) Physiological analysis of motor reorganization following lower limb amputation. Electroencephalography Clinical Neurophysiology 85: 53-60

Fukson, O. I., Berkinblit, M. B. & Feldman, A. G. (1980) The spinal frog takes into account the scheme of its body during the wiping reflex. Science, 209: 1261-1263.

Ghez, C., Gordon, J. & Hening, W. (1988) Trajectory control in dystonia. In: Fahn, S., Marsden, C. D. & Calne, D.B. (eds.) Dystonia 2, Advances in Neurology, vol. 50, Raven Press: New York, pp. 141-155.

Glendinning, D. S. & Enoka, R. M. (1994) Motor unit behavior in Parkinson's disease. Physical Therapy, 74: 61-70.

Gielen, S., van Ingen Schenau G. J., Tax, T. & Theeuwen, M. (1990) The activation of mono- and bi-articular muscles in multi-joint movements. In: Multiple Muscle Systems. Biomechanics and Movement Organization. eds. J. M. Winters & S. L.-Y. Woo, pp. 302-311, New York e.a.:Springer-Verlag

Gottlieb GL, Corcos DM, Agarwal GC (1989) Strategies for the control of voluntary movements with one mechanical degree of freedom. Behavioral and Brain Sciences 12: 189-250.

Gutman, S. R., Gottlieb, G. L. & Corcos, D. M. (1992) Exponential model of a reaching movement trajectory with non-linear time. Comments in Theoretical Biology 2: 357-384.

Hallett, M. (1993) Physiology of basal ganglia disorders: an overview. Canadian Journal of Neurological Science, 20: 177-183.

Hallett, M. & Khoshbin, S. (1980) A physiological mechanism of bradykinesia. Brain 103: 301-314

Halsband, U., Homberg, V. & Lange, H. J. (1990) Slowing of different types of voluntary movement in extrapyramidal disease: Fitts' law and idiographic writing. In: Berardelli A, Benecke R, Manfredi M, Marsden CD (Eds.) Motor Disturbances II, pp. 182-190, London e.a.: Academic Press

Hayashi, A., Kagamihara, Y., Nakajima, Y., Narabayashi, H., Okuma, Y. & Tanaka, R. (1988) Disorder in reciprocal innervation upon initiation of voluntary movement in patients with Parkinson's disease. Experimental Brain Research 70: 437-440

Heilman KM, Bowers D, Watson RT, Greer M (1976) Reaction time in Parkinson disease. Arch Neurol 33: 139-140

Henderson, S. E. (1985) Motor skill development. In: Lane, D. & Stratford, B. (Eds.) Current Approaches to Down Syndrome, pp. 187-218, London: Holt, Rhinehart & Winstone

Henderson, S. E., Morris, J. & Frith, V. (1981) The motor deficit in Downs syndrome children: A problem of timing? Journal of Child Psychology and Psychiatry 22: 233-245

Henneman, E., Somjen, G. & Carpenter, D. O. (1965) Excitability and inhibitibility of motoneurones of different sizes. Journal of Neurophysiology 28: 599-620

Herz, E. (1944) Dystonia I. Historical review, analysis of dystonic symptoms and physiologic mechanisms involved. Archives of Neurology and Psychiatry (Chicago) 51: 305-318.

Hoefer, P. F. A. & Putnam, T. J. (1940) Action potentials of muscles in athetosis and Syndenham chorea. Archives of Neurology and Psychiatry (Chicago) 44: 417-431.

Hogan, N. (1984) An organizational principle for a class of voluntary movements. Journal of Neuroscience 4: 2745-2754

Horak, F. B., Nashner, L. M. & Diener, H. C. (1990) Postural strategies associated with somatosensory and vestibular loss. Experimenal Brain Research 82: 167-177

Houk, J. C. (1976) An assessment of stretch reflex function. Progress in Brain Research 44: 303-314.

Houk, J. C. (1979) Regulation of stiffness by skeletomotor reflexes. Annual Review of Physiology 41: 99-114.

Hughes, M. & McLellan, D. L. (1985) Increased co-activation of the upper limb muscles in writer's cramp. Journal of Neurology, Neurosurgery, and Psychiatry 48: 782-787.

Hunter, J. P., Ashby, P. & Lang, A. E. (1988) Afferents contributing to the exaggerated long latency reflex response to electrical stimulation in Parkinson's disease. Journal of Neurology, Neurosurgery, and Psychiatry 51: 1405-1410

Inzelberg, R., Flash, T. & Korczyn, A. D. (1990) Kinematic properties of upper-limb trajectories in Parkisnon's disease and idiopathic torsion dystonia. Advances in Neurology 53: 183-189

Isenberg, C. & Conrad, B. (1994) Kinematic properties of slow arm movements in Parkinson's disease. Journal of Neurology 241: 323-330.

Jenkins, W. M. & Merzenich, M. M. (1987) Reorganization of neocortical representations after brain injury: a neurophysiological model of the bases of recovery from stroke. In: Progress in Brain Research. eds. F. J. Seil, E. Herbert & B. M. Carlson, pp. 249-266, Elsevier: Amsterdam.

Jenkins, W. M., Merzenich, M. M. & Recanzone, G. (1990) Neocortical representation dynamics in adult primates: Implications for neuropsychology. Neuropsychologia 28: 573-584.

Johnson, M. T. V., Kipnis, A. N., Lee, M. C., Loewenson, R. B. & Ebner, T. J. (1991) Modulation of the stretch reflex during volitional sinusoidal tracking in Parkinson's disease. Brain 114: 443-460

Joris, H. J. J., Edwards van Muyen, A. J., van Ingen Schenau, G. J. & Kemper, H. C. G. (1985) Force, velocity, and energy flow during the overarm throw in female handball players. Journal of Biomechanics 18: 409-414.

Ingen Schenau van, 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.

Kalaska, J. & Pomeranz, B. (1979) Chronic paw denervation causes an age-dependent appearance of novel responses from forearm in "paw cortex" of kittens and adult cats. Journal of Neurophysiology 42: 618-633.

Kaminski, T. & Gentile, A. M. (1986) Joint control strategies and hand trajectories in multijoint pointing movements. Journal of Motor Behavior 18: 261-278

Kerr, R. & Blais, C. (1985) Motor skill acquisition by individuals with Down Syndrome. American Journal of Mentat Deficiency 90: 313-318

Keshner, E. A. (1994) Vertebral orientation and muscle activation patterns during controlled head movements in cats. Experimental Brain Research 98: 546-550.

Kew, J. J. M., Ridding, M. C., Rothwell, J. C., Passingham, R. E., Leigh, P. N., Sooriakumaran, S., Frackowiak, R. S. J., Brooks, D. J. (1994) Reorganization of cortical blood flow and transcranial magnetic stimulation maps in human subjects after upper limb amputation. Journal of Neurophysiology 72: 2517-2524.

Larkin, D. & Hoare, D. (1992) The movement approach: a window to understanding the clumsy child. In: Approaches to the Study of Motor Control and Learning. ed. J. J. Summers, pp. 413-439, N-Holland: Amsterdam.

Lashley, K. S. (1933) Integrative functions of the cerebral cortex. Physiological Reviews, 13: 1-42.

Latash, M. L. (1992) Motor control in Down syndrome: The role of adaptation and practice. Journal of Developmental and Physical Disability 4: 227-261.

Latash, M. L. (1993) Control of Human Movement. Human Kinetics: Urbana, IL.

Latash, M. L., Almeida, G. L. & Corcos, D. M. (1993) Pre-programmed reactions in individuals with Down syndrome: The effects of instruction and predictability of the perturbation. Archives of Physical Medicine and Rehabilitation 73: 391-399.

Latash, M. L., Aruin, A. S., Neyman, I., & Nicholas, J. J. (1995) Anticipatory postural adjustments during self-inflicted and predictable perturbations in Parkinson's disease. Journal of Neurology, Neurosurgery, and Psychiatry ( in press).

Latash, M. L. & Corcos, D. M. (1991) Kinematic and electromyographic characteristics of single-joint movements of individuals with Down Syndrome. American Journal of Mental Retardation 96: 189-201.

Latash, M.L. & Gutman, S.R. (1994) Abnormal motor patterns in the framework of the equilibrium-point hypothesis: A cause for dystonic movements? Biological Cybernetics 71: 87-94.

Lee, R. G., Murphy, J. T. & Tatton, W. G. (1983) Long-latency myotatic reflexes in man" Mechanisms, functional significance, and changes in patients with Parkinson's disease or hemiplegia. Advances in Neurology 39: 489-508.

Lee, R. G. & Stein, R. B. (1981) Resetting of tremor by mechanical perturbations: a comparison of essential tremor and Parkinsonian tremor. Annals of Neurology, 10: 523-531.

Lee, R. G. & Tatton, W. G. (1975) Motor responses to sudden limb displacements in primates with specific CNS lesions and in human patients with motor system disorders. Canadian Journal of Neurological Science 2: 285-293

Lee, W. A. (1980) Anticipatory control of posture and task muscles during rapid arm flexion. Journal of Motor Behavior 12: 185-196

Lieber, R. L. & Bodine-Fowler, S. C. (1993). Skeletal muscle mechanics: Implications for rehabilitation. Physical Therapy 73: 844-856.

Lincoln, A. J., Courchesne, E., Kilman, B. A. & Galambos, R. (1985) Neuropsychological correlates of information-processing by children with Down syndrome. American Journal of Mental Deficiency 89: 403-414.

Lockwood, R. (1987). Rehabilitation. In: Lockwodd, R. (ed.) Physical Education and Disability, The Australian Council for Health, Physical Education and Recreation: Parkside, South Australia, pp. 14-23.

Lockman, J. J. & Thelen, E. (1993) Developmental biodynamics: Brain, body, behavior connections. Child Development 64: 953-959.

Luria, A. R. (1973) The Working Brain: An Introduction to Neuropsychology. Basic Books Inc.: New York.

Marsden, C. D. (1988) Investigation of dystonia. In: Fahn, S., Marsden, C.D. & Calne, D.B. (eds.) Dystonia 2, Advances in Neurology, vol. 50, Raven Press: New York, pp. 35-44

Marsden, C. D., Merton, R. A. & Morton, H. B. (1972) Servo action in human voluntary movement. Nature 238: 140-143

Marsden, C. D., Merton, R. A., Morton, H. B., Adam, J. E. R. & Hallett, M. (1978) Automatic and voluntary responses to muscle stretch in man. Progress in Clinical Neurophysiology 4: 167-177

Marsden, C. D., Merton, R. A., Morton, H. B., Rothwell, J. C. & Traub, M. M. (1981) Reliability and efficacy of the long-latency stretch reflex in the human thumb. Journal of Physiology 316: 47-60

Marsden, C. D., Rothwell, J. C. & Traub, M. (1979) Long latency stretch reflex of the human thumb can be reversed if the task is changed. Journal of Physiology 293: 41P-42P

Massion, J. (1992) Movement, posture and equilibrium: Interaction and Coordination. Progress in Neurobiology 38: 35-56.

Matthews, P. B. C., Cody, F. W. J., Richardson, H. C. & MacDermott, N. (1990) Observations on the reflex effects seen in Parkinson's disease on terminating a period of tendon vibration. Journal of Neurology, Neurosurgery, and Psychiatry 53: 215-219

Mauerberg E., Schuller J., & Fantucci I. (1994) Phase portrait description of walking pattern of severely mentally retarded subjects. Brazilian International Journal of Adapted Physical Education Research 1: 19-50.

Mawston, G. A. & Anson, J. G. (in press). Down syndrome: Attention and neuromotor reaction time. Abstracts of the 11-th Annual Australasian Winter Conference on Brain Research. International Journal of Neuroscience (in press)

Merzenich, M. M., Nelson, R. J., Stryker M. S., Cynader, M. S., Schoppman, A. & Zook, J. M. (1984) Somatosensory cortical map changes following digit amputation in adult monkeys. Journal of comparative Neurology 224: 591-605

Metzler, J. & Marks, P. S. (1979) Functional changes in cat somatic sensory-motor cortex during short-term reversible epidural blocks. Brain Research 177: 379-383.

Mortimer, J. A. & Webster, D. D. (1979) Evidence for a quantitative association between EMG stretch responses and Parkinsonian rigidity. Brain Research 162: 169-173

Morris, A. F., Vaughan, S. E. & Vaccaro, P. (1982) Measurements of neuromuscular tone and strength in Down's syndrome children. Journal of Mental Deficiency Research 26: 41-46

Morris, M. E., Iansek, R., Matyas, T. A. & Summers, J. J. (in press) Motor control of gait in Parkinson's disease. Abstracts of the 12-th International Australasian Winter Conference on Brain Research. International Journal of Neuroscience

Muller, F. & Abbs, J. H. (1990) Precision grip in parkinsonian patients. Advances in Neurology 53: 191-195

Nakashima, K., Rothwell, J. C., Day, B. L., Thompson, P. D., Shannon, K., Marsden, C. D. (1989) Reciprocal inhibition between forearm muscles in patients with writer's cramp and other occupational cramps, symptomatic hemidystonia and hemiparesis due to stroke. Brain 112: 681-697.

Nativ, J. & Abbs, J. H. (1989) Goal-directed arm movements in Down syndrome. In: Motor Control in Down Syndrome. ed. M. L. Latash, pp. 36-42, Rush Medical Center, Chicago

Nelson, W. (1983) Physical principles for economies of skilled movements. Biological Cybernetics 46: 135-147

Newell, K. M., Corcos, D. M. (Eds.) (1993) Variability in Motor Control. Human Kinetics: Urbana, IL

Newell, K. M., Van Emmerik, R. E. A. & McDonald, P. V. (1989) Biomechanical constraints and action theory. Human Movement Science 8: 403-409.

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

Pascual-Leone, A., Grafman, J. & Hallett M. (1994) Modulation of cortical motor output maps during development of implicit and explicit knowledge. Science 263: 1287-1289.

Pozzo, T., Berthoz, A. & Lefort, L. (1990) Head stabilization during various locomotor tasks in humans. I. Normal subjects. Experimental Brain Research 82: 97-106

Rarick, G. L., Dobbins, D. A. & Broadhead, G. G. (1976) The Motor Domain and its Correlates in Educated Handicapped Children. Englewood Cliffs, N.J.: Prentice Hall.

Recanzone, G. H., Allard, T. T., Jenkins, W. M. & Merzenich, M. M. (1990) Receptive-field changes induced by peripheral nerve stimulation in SI of adult cats. Journal of Neurophysiology 63: 1213-1225.

Recanzone, G. H., Merzenich, M. M., Jenkins, W. M., Grajski, K. A. & Dinse, H. R. (1992) Topographic reorganization of the hand representation in cortical area 3b of owl monkeys trained in a frequency-discrimination task. Journal of Neurophysiology 67: 1031-1056.

Rosen, A. J., Lockhart, J. J., Gants E. S. & Westergaard, C. K. (1993). Maintenance of grip-induced muscle tension: a behavioral marker of schizophrenia. Journal of Abnormal Psychology 100: 583-593

Rosenbaum, D. A. (1991) Human Motor Control. Academic Press: San Diego, CA.

Rothwell, J. C., Day, B. L., Berardelli, A. & Marsden, C. D. (1986) Habituation and conditioning of the human long latency stretch reflex. Experimental Brain Research 63: 197-204

Rothwell, J. C., Obeso, J. A., Traub, M. M. & Marsden, C. D. (1983) The behavior of the long-latency stretch reflex in patients with Parkinson's disease. Journal of Neurology Neurosurgery and Psychiatry 46: 35-44

Sanes, J. N. (1985) Information processing deficits in Parkinson's disease during movement. Neuropsychology 23: 381-392

Schieppati, M. & Nardone, A. (1991) Free and supported stance in Parkinson's disease. The effect of posture and "postural set" on leg muscle responses to perturbation, and its relation to the severety of the disease. Brain 114: 1227-1244

Seif-Naraghi, A. H. & Winters, J. M. (1990) Optimized strategies for scaling goal-directed dynamic limb movements. In: Multiple Muscle Systems. Biomechanics and Movement Organization. eds. J. M. Winters & S. L.-Y. Woo, pp. 312-334, New York e.a.:Springer-Verlag

Shepherd, R. (1984). Physiotherapy in Paediatrics. 2-nd edition, William Heinemann Medical Books Ltd.: London, England.

Sheridan, M. R. (1985). Individual differences in voluntary movement. In: Kirklady, B. D. (ed.) Individual Differences in Movement, MTP Press: Lancaster, England, pp. 3-26.

Sheridan, M. R. & Flowers, K. A. (1990) Movement variability and bradykinesia in Parkinson's disease. Brain, 113: 1149-1161

Sheridan, M. R., Flowers, K. A. & Hurrell, J. (1987) Programming and execution of movement in Parkinson's disease. Brain 110: 1247-1271

Shumway-Cook, A. & Woollacott, M. H. (1985) Dynamics of postural control in the child with Down Syndrome. Physical Therapy 65: 1315-1322

Shumway-Cook, A. & Horak, F. B. (1989) Vestibular rehabilitation: An exercise approach to managing symptoms of vestibular dysfunction. Seminars in Hearing 10: 196-208

Stein, R. B., Yang, J. F., Belanger, M. & Pearson, K. G. (1993). Modification of reflexes in normal and abnormal movements. Progress in Brain Research 97: 189-196.

Stelmach, G. E. & Worringham, C. J. (1988) The preparation and production of isometric force in Parkinson's disease. Neuropsychology 26: 93-103

Stelmach, G. E., Worringham, C. J. & Strand, E. A. (1986) Movement preparation in Parkinson's disease: The use of advance information. Brain 109: 1179-1194

Sugden, D. A., Keogh, J. F. (1990) Problems in Movement Skill Development. University of South Carolina Press: Columbia.

Swinnen, S. P., Massion, J., Heuer, H & Casaer, P. (eds.) (1994). Interlimb Coordination: Neural, Dynamical, and Cognitive Constraints. Academic Press, Inc.: San Diego e.a.

Tatton, W. G. & Lee, R. G. (1975) Evidence for abnormal long-loop reflexes in rigid Parkinsonian patients. Brain Research 100: 671-676

Tax, A. A. M., Denier van der Gon, J. J. & Erkelens, C. J. (1990) Differences in coordination of elbow flexor muscles in force tasks and in movement tasks. Experimental Brain Research 81: 567-572

Teasdale, N., Phillips, J., Stelmach, G. E. (1990) Temporal movement control in patient with Parkinson's disease. Journal of Neurology, Neurosurgery, and Psychiatry 53: 862-868

Thelen E., & Fogel A. (1989) Toward an action-based theory of infant development. In: Lockman J. J. & Hazen N. L. (eds.) Action in Social Context: Perspectives on Early Development, pp. 23-63, New York: Plenum.

Thelen E., Zernicke R., Schneider K., Jensen J., Kamm K., & Corbetta D. (1992) The role of intersegmental dynamics in infant neuromotor development. In: Stelmach G. E. & Requin J. (eds.) Tutorials in Motor Behavior II, pp. 533-548, Amsterdam: Elsevier.

Traub, M. M., Rothwell, J. C. & Marsden, C. D. (1980) Anticipatory postural reflexes in Parkinson's disease and other akinetic-rigid syndromes and in cerebellar ataxia. Brain 103: 393-412

Turvey, M. T. (1990) Coordination. American Psychologist 45: 938-953

van Zuylen, E. J., Gielen, C. C. A. M. & Denier van der Gon, J. J. (1988) Coordination and inhomogenous activation of human arm muscles during isometric torques. Journal of Neurophysiology 60: 1523-1548

Viallet, F., Massion, J., Massarino, R., & Khalil R. (1987) Performance of a bimanual load-lifting task by Parkinsonian patients. Journal of Neurology, Neurosurgery, and Psychiatry 50: 1274-1283.

Whiting, H. T. A. (1984) Human Motor Actions: Bernstein Reassessed. Amsterdam: Elsevier

Wierzbicka, M., Wiegner, A. W., Logigian, E. L. & Young, R. R. (1991) Abnormal most-rapid isometric contractions in patients with Parkinson's disease. Journal of Neurology, Neurosurgery, and Psychiatry 54: 210-216

Wing, A. M. (1988) A comparison of the rate of pinch grip force increases and decreases in Parkinsonian bradykinesia. Neuropsychology 26: 479-482

Winter, D., Ruder, G. K. & MacKinnon, C. D. (1990) Control of balance of upper body during gait. In: In: Multiple Muscle Systems. Biomechanics and Movement Organization. eds. J. M. Winters & S. L.-Y. Woo, pp. 534-541, New York e.a.:Springer-Verlag

Winters, J. M. & Woo, S. L.-Y. (Eds.) (1990) Multiple Muscle Systems. Biomechanics and Movement Organization. New York e.a.:Springer-Verlag

Wolpaw, J. R. (1983) Adaptive plasticity in the primate spinal stretch reflex: reversal and redevelopment. Brain Research 278: 299-304

Wolpaw, J. R., Braitman, D. J. & Seegal, R. F. (1983) Adaptive plasticity in primate spinal stretch reflex: initial development. Journal of Neurophysiology 50: 1296-1311

Zajac, F. E. & Gordon, M. E. (1989) Determining muscle's force and action in multi-articular movements. Exercise and Sport Science Reviews 17: 187-230

Figure Captions

Figure 1

We consider movements of all unimpaired individuals to be controlled by coordinative rules (CNS priorities) that allow a choice to be made in the presence of motor redundancy. CNS priorities are assumed to be generally the same for all unimpaired individuals including those that are commonly considered control subjects (Average Person), clumsy children at one end of the spectrum and elite athletes (e.g., Michael Jordan) at the other end. Changes in CNS priorities may occur following a change in any of the components taking part in the generation of voluntary movements. These may include impaired decision-making or altered world perception, CNS structural changes, peripheral structural changes, and maybe, at the other end of the spectrum, very specialized training (e.g., top sumo wrestlers).

Figure 2

An example of kinematic and electromyographic (EMG) patterns during a fast single-joint movement. Averaged across 6 trials data are illustrated. The subject performed elbow extension movements over 30. Note the "bell-shaped" velocity, the "double-peaked" acceleration, and the "tri-phasic" EMG pattern. The antagonist (biceps) EMG is inverted for better visualization.

Figure 3

The process of generation of a voluntary movement is assumed to involve three major steps: Understanding the task, Generation of control patterns, and Execution. We consider the step of generation of control patterns as a major source of flexibility and variability of voluntary movements. At this level, the CNS can generally manipulate more variables than the number of parameters defining a motor task. This requires the CNS to add constraints on the control variables. Thus, at this level, the hypothetical CNS priorities are assumed to emerge and to play their role in shaping movement patterns.

Figure 4

Movement patterns are defined by control patterns, by the structure of the body including the CNS structure), and by the external forces (environment). Control patterns are defined by the body structure, task, environment, and CNS priorities that may also be addressed as coordinative rules.

* We will not consume time trying to come up with universally acceptable definitions for "agonist" and "antagonist". In most single-joint movement experiments, one can address a muscle accelerating the joint in a desired direction as "agonist", and a muscle braking this movement as an "antagonist". This definition is as good (or as bad) as any other.