Training BIG to move faster: the application of the speed–amplitude relation as a rehabilitation strategy for people with Parkinson’s disease

We have used the phenomenon that speed increases with movement amplitude as a rehabilitation strategy. We tested the hypothesis that the generalized training of amplitude in the limb motor system may reduce bradykinesia and hypokinesia in the upper and lower limbs in subjects with Parkinson's disease (PD) across disease severity (Stage I, n=6; Stage II, n=7; Stage III, n=5). While studies have separately examined the relationship of amplitude to speed in reaching and gait, the same study has not reported the relationship for both limb systems. Moreover, the rehabilitation intervention, Training BIG, is unique in that it applies well-established treatment concepts from a proven treatment for the speech motor system in PD [Lee Silverman Voice Treatment (LSVT)] to the limb motor system. Subjects (n=18) participated in intense practice (1-h sessions/4x week/4 weeks) of large amplitude movements involving the whole body (i.e., head, arm, trunk, and leg) while focusing on the sensory awareness of "movement bigness." Testing procedures were designed to demonstrate the transfer of generalized amplitude practice to speed improvements during functional "untrained" tasks in "uncued" conditions with blinded testers. After therapy, the subjects significantly increased their speed of reaching and gait for the preferred speed condition. This effect was greater when the severity of the disease was less. The results support further application and efficacy studies of Training BIG. Amplitude-based behavioral intervention in people with PD appears to be a simple target that may be applied in different contexts for multiple tasks and results in improved speed-amplitude scaling relations across the upper and lower limbs.

Novel muscle patterns for reaching after cervical spinal cord injury: a case for motor redundancy

A fundamental issue in the neuromotor control of arm movements is whether the nervous system can use distinctly different muscle activity patterns to obtain similar kinematic outcomes. Although computer simulations have demonstrated several possible mechanical and torque solutions, there is little empirical evidence that the nervous system actually employs fundamentally different muscle patterns for the same movement, such as activating a muscle one time and not the next, or switching from a flexor to an extensor. Under typical conditions, subjects choose the same muscles for any given movement, which suggests that in order to see the capacity of the nervous system to make a different choice of muscles, the nervous system must be pushed beyond the normal circumstances. The purpose of this study, then, was to examine an atypical condition, reaching of cervical spinal cord injured (SCI) subjects who have a reduced repertoire of available distal arm muscles but otherwise a normal nervous system above the level of lesion. Electromyography and kinematics of the shoulder and elbow were examined in the SCI subjects performing a center-out task and then compared to neurologically normal control subjects. The findings showed that the SCI-injured subjects produced reaches with typical global kinematic features, such as straight finger paths, bell-shaped velocities, and joint excursions similar to control subjects. The SCI subjects, however, activated only the shoulder agonist muscle for all directions, unlike the control pattern that involved a reciprocal pattern at each joint (shoulder, elbow, and wrist). Nonetheless, the SCI subjects could activate their shoulder antagonist muscles, elbow flexors, and wrist extensor (extensor carpi radialis) for isometric tasks, but did not activate them during the reaching movements. These results demonstrate that for reaching movements, the SCI subjects used a strikingly different pattern of intact muscle activities than control subjects. Hence, the findings imply that the nervous system is capable of choosing either the control pattern or the SCI pattern. We would speculate that control subjects do not select the SCI pattern because the different choice of muscles results in kinematic features (reduced fingertip speed, multiple shoulder accelerations) other than the global features that are somehow less advantageous or efficient.

Shoulder muscle activity in Parkinson's disease during multijoint arm movements across a range of speeds

Bradykinesia is one of the primary symptoms of Parkinson disease and leads to significant functional limitations for patients. Single joint movement studies, that have investigated the mechanism of bradykinesia, suggest that several features of muscle activity are disrupted, including modulation of burst amplitude and duration, and the number of bursts. It has been proposed that it is the blending of these different burst deficits that collectively defines bradykinesia. This study adds two new approaches to the study of bradykinesia. First, we examined the features of shoulder muscle activities during multijoint arm movement in bradykinetic and control subjects, such that previously reported single joint hypotheses could be tested for generalized arm movement. Second, we directly manipulated speed while keeping distance constant for a large range of speeds. In this manner, we could compare individual trials of muscle activity between controls and subjects with Parkinson's disease (PD) for movements matched for both speed and movement duration. Our results showed that while a multiple burst pattern of shoulder muscles was a common strategy for all subjects (young, elderly controls and PD), subjects with PD showed several burst abnormalities, including deficits in initial agonist burst amplitude and duration at both fast and slow speeds. Subjects with PD also (1) failed to produce a one-burst pattern at fast speeds and, instead, produced a predominance of multiple burst patterns and (2) showed a relationship between the number of burst deficits and the severity of disease. These results extend the findings of single joint studies to multi-joint and similarly indicate that a combination of burst modulation abnormalities correlate with bradykinesia and disease severity.

General coordination of shoulder, elbow and wrist dynamics during multijoint arm movements

Studies of multijoint arm movements have demonstrated that the nervous system anticipates and plans for the mechanical effects that arise from motion of the linked limb segments. The general rules by which the nervous system selects appropriate muscle activities and torques to best deal with these intersegmental effects are largely unknown. In order to reveal possible rules, this study examined the relationship of muscle and interaction torques to joint acceleration at the shoulder, elbow and wrist during point-to-point arm movements to a range of targets in the horizontal plane. Results showed that, in general, dynamics differed between the joints. For most movements, shoulder muscle torque primarily determined net torque and joint acceleration, while interaction torque was minimal. In contrast, elbow and wrist net torque were determined by a combination of muscle and interaction torque that varied systematically with target direction and joint excursion. This "shoulder-centered pattern" occurred whether subjects reached targets using straight or curved finger paths. The prevalence of a shoulder-centered pattern extends findings from a range of arm movement studies including movement of healthy adults, neurological patients, and simulations with altered interaction effects. The shoulder-centered pattern occurred for most but not all movements. The majority of the remaining movements displayed an "elbow-centered pattern," in which muscle torque determined initial acceleration at the elbow and not at the shoulder. This occurred for movements when shoulder excursion was <50% of elbow excursion. Thus, both shoulder- and elbow-centered movements displayed a difference between joints but with reversed dynamics. Overall, these findings suggest that a difference in dynamics between joints is a general feature of horizontal plane arm movements, and this difference is most commonly reflected in a shoulder-centered pattern. This feature fits well with other general shoulder-elbow differences suggested in the literature on arm movements, namely that: (a) agonist muscle activity appears more closely related to certain joint kinematics at the shoulder than at the elbow, (b) adults with neurological damage display less disruption of shoulder motion than elbow motion, and (c) infants display adult-like motion first in the shoulder and last at the wrist.

Trunk muscle activity during the simultaneous performance of two motor tasks

A unique feature of trunk muscles is that they can be activated to meet functional requirements for combined behaviors, including those related to posture and breathing. Trunk muscles therefore may have developed mechanisms for dealing with simultaneous inputs for different task requirements. This study was designed to test the hypothesis that a linear addition in trunk muscle activities would occur when an isometric trunk task and a pulsed expiration task was performed simultaneously. Surface electromyograms (EMG) were recorded from four trunk regions (medial and lateral back, upper and lower lateral abdomen) in sitting during the performance of the individual isometric trunk task, the individual pressure task, and the combined task (isometric trunk and pressure task). The direction of static holding for the isometric trunk task was varied between flexion and extension positions. For the pressure task subjects produced two consecutive pressure pulses (2/s) to a target oral pressure. For each muscle recording, a linear prediction was calculated from the mathematical addition of the EMG recorded from the individual trunk and pressure tasks. This linear prediction was compared to the actual muscle activity recorded during the combined task. Typically the EMG from two muscles showed linear addition, such that the relative contribution of muscle activity did not change for the combined task. This suggests that the motor commands for each task reached these motor neuron pools essentially unmodified. The other two muscles showed nonlinear combination of two EMG patterns. That is, qualitatively both EMG patterns, specific to each command, were evident in the measured EMG traces for the combined task, but quantitatively the muscle did not meet all criteria for linear addition. Linear addition may provide a simple mechanism for combining breathing-related behaviors (expiratory efforts) with other trunk behaviors (holding against gravity). This suggests that some muscles can be shared for two different voluntary tasks without changing their contribution to either component task. At the same time, nonlinear combination suggests that some muscles are shared, but their contribution to either component task may be modulated, thus avoiding the construction of a third new and unique plan.

Electromyographic responses to a mechanical perturbation applied during impending arm movements in different directions: one-joint and two-joint conditions

Directional tuning is a common finding for many physiological features of arm movements and related neuronal activity. We investigated whether the electromyographic response to a brief (30 ms) torque perturbation prior to voluntary movement depends on the direction of the impending movement. Pointing movements with the elbow joint alone and those involving both the shoulder and elbow joints were studied in separate experiments. Target direction was varied between flexion and extension for the one-joint experiments and among four spatial directions for the two-joint experiments. Movement trials in which a perturbation stretched the flexor muscles just prior to the pointing movement were randomly interspersed among unperturbed movement trials in each experiment. A small pre-load ensured some background activity of the flexor muscles. Results were remarkably similar for the one- and two-joint conditions. The short-latency reflex response of the stretched muscles (in a 30-60 ms window after perturbation onset) was not modulated with direction of target-reaching movement in a statistically significant manner, which confirms earlier findings for one-joint movements and extends these to the two-joint condition. Beyond the short-latency window, the perturbation provoked earlier onsets of target-reaching muscle activities for the agonist muscles, whether or not the muscle had been stretched by the perturbation. The onset of the braking activity of the antagonist muscles also occurred earlier in the presence of the brief perturbation prior to movement, irrespective of whether the muscle had been stretched or not. The magnitude of target-reaching muscle activity, in general, was greater for the perturbed trials, though not consistently for all muscles or all directions. These results suggest that, when movement is about to be initiated, in either single- or multi-joint conditions, the long-latency effects of the stretch strongly depend on the intended direction of movement. The dependence is such that the response serves to hasten and augment the intended movement, but not necessarily to oppose the perturbation.

Control of the wrist in three-joint arm movements to multiple directions in the horizontal plane

In a reaching movement, the wrist joint is subject to inertial effects from proximal joint motion. However, precise control of the wrist is important for reaching accuracy. Studies of three-joint arm movements report that the wrist joint moves little during point-to-point reaches, but muscle activities and kinetics have not yet been described across a range of movement directions. We hypothesized that to minimize wrist motion, muscle torques at the wrist must perfectly counteract inertial effects arising from proximal joint motion. Subjects were given no instructions regarding joint movement and were observed to keep the wrist nearly motionless during center-out reaches to directions throughout the horizontal plane. Consistent with this, wrist muscle torques exactly mirrored interaction torques, in contrast to muscle torques at proximal joints. These findings suggest that in this reaching task the nervous system chooses to minimize wrist motion by anticipating dynamic inertial effects. The wrist muscle torques were associated with a direction-dependent choice of muscles, also characterized by initial reciprocal activation rather than initial coactivation to stiffen the wrist joint. In a second experiment, the same pattern of muscle activities persisted even after many trials reaching with the wrist joint immobilized. These results, combined with similar features at the three joints, such as cosine-like tuning of muscle torques and of muscle onsets across direction, suggest that the nervous system uses similar rules for muscles at each joint, as part of one plan for the arm during a point-to-point reach.

Hyper-reflexia without spasticity after unilateral infarct of the medullary pyramid

Whether or not a lesion confined to the pyramidal tract produces spasticity in humans remains an unresolved controversy. We have studied a patient with an ischemic lesion of the right medullary pyramid, using objective measures of hyper-reflexia, spasticity, and weakness. Electromyographic activity (EMG) of the biceps muscles was recorded under the following conditions: (1) in response to a tendon tap with an instrumental reflex hammer, (2) in response to imposed quick stretch with motion analysis, and (3) during an isometric holding task. Hyper-reflexia of the involved arm in response to tendon tap was shown to be due primarily to an increase in the gain of the reflex arc. No velocity-dependent increase in the response to quick stretch of the involved arm was present. This was consistent with the absence of detectable spasticity on the clinical exam. These findings suggest that a lesion confined to the medullary pyramid can give rise to weakness and hyper-reflexia without causing spasticity. Moreover, these findings suggest that different anatomical substrates may underlie the clinical phenomena of hyper-reflexia and spasticity.

Directional effects of changes in muscle torques on initial path during simulated reaching movements

Adults are able to reach for an object for the first time with appropriate direction, speed, and accuracy. The rules by which the nervous system is able to set muscle activities to accomplish these outcomes are still debated and, indeed, the sensitivity of kinematics to variations in muscle torques is unknown for complex arm movements. As a result, this study used computer simulations to characterize the effects of change in muscle torque on initial hand path. The same change was applied to movements towards 12 directions in the horizontal plane, and changes were systematically manipulated such that: (1) torque amplitude was changed at one joint, (2) timing of torque was changed at one joint, and (3) amplitude and/or timing was changed at two joints. Results showed that simultaneous changes in torque amplitude at shoulder and elbow joints affected initial speed uniformly across direction. These results add to conclusions from previous experimental and modeling work that the simplest rule to produce a desired change in speed for any direction is to scale torque amplitude at both joints. In contrast, all simulations showed nonuniform effects on initial path direction. For some regions of the workspace, initial path direction was little affected by either a +/-30% change in amplitude or a +/-100-ms change in timing, whereas for other regions the same changes produced large effects on initial path direction. These findings suggest that the range of possible torque solutions to achieve a particular initial path direction varies within the workspace and, consequently, the requirements for an accurate initial path will vary within the workspace.

Selection of muscles for initiation of planar, three-joint arm movements with different final orientations of the hand

Muscle activities and joint rotations were examined at the shoulder, elbow, and wrist joints for pointing movements to targets in the horizontal plane. In such movements, multiple arm configurations are possible for a given target location. Thus, starting from the same initial configuration and for the same target location in space, the joint excursions could be varied. When no constraints were placed on the final orientation of the hand, the choice of muscles initially activated at the wrist joint was consistent with a function to resist inertial effects of proximal segment motion on the wrist joint. When subjects were asked to produce different final orientations of the hand for the same target location, the initial choice of muscles at the three joints was preserved in most trials, whether wrist flexion or extension was required to reach the final hand orientation. The relative onset times of muscle activity at the different joints were also not correlated with wrist excursion. This suggests a predetermined initial selection of muscles that is related to target location, not to joint angular excursion. The fact that the required final hand orientation was nevertheless achieved suggests that the planning of these pointing movements is not a unitary process, but is comprised of two components: a fixed initial muscle selection for a given target location in space, and a selection appropriate for the required joint excursions.

Activity of Wrist Muscles Elicited during Imposed or Voluntary Movements about the Elbow Joint

To examine the coordination of muscles during multijoint movement, we compared the response of wrist muscles to perturbations about the elbow joint with their activation during a volitional elbow movement. The purpose was to test the following two predictions: (a) Responses can occur in muscles not stretched by the perturbation, as has been reported for other multijoint systems; and (b) the motor pattern in response to a perturbation mimics an opposing volitional motor pattern across the two joints. We recorded the electromyographic (EMG) activity of elbow and wrist muscles as well as the flexion/extension motions at the elbow and wrist joints during individual trials that either involved a response to a torque perturbation that extended the elbow or required volitional elbow flexion. The results of this study confirmed that responses were elicited in the nonstretched wrist muscles when the elbow joint was perturbed. The same motor sequence of elbow and wrist flexors was present for both the volitional and perturbation task (with the forearm supinated), regardless of whether the wrist joint was immobilized or freely moving. The findings suggest that the nervous system relies on the purposeful coupling of elbow and wrist flexors to counter the inertial effects during the unrestricted voluntary movement, even though the coupling does not appear to be purposeful during the perturbation or with the wrist immobilized. The coupling of elbow and wrist flexors, however, was not rigidly fixed, as evidenced by muscle onsets that adapted over repeated perturbation trials and a reversal of the wrist muscle activated (wrist extensor) when the forearm was pronated. Hence, the coupling of muscle activities can be modified quantitatively when not beneficial and can be altered qualitatively with different initial configurations of the arm.

Coupled and uncoupled limb oscillations during paw-shake response

Intersegmental limb dynamics and muscle activities were analyzed for consecutive cycles of paw-shake responses from chronic-spinalized cats to investigate how hindlimb trajectories organize into a pattern with regular oscillations, a steady-state response, or alternatively, into a pattern with irregular oscillations, a nonsteady-state response. In the spinalized preparation, steady-state and nonsteady-state responses have an equal likelihood of emerging from the initial cycles of a paw-shake response, suggesting that regular coupling of joint oscillations is not planned by pattern-generating networks within lumbosacral segments. To examine the characteristics of coupled and uncoupled limb oscillations during paw-shake responses, we assessed patterns of muscle activity and hindlimb kinematics of six adult chronic-spinalized cats. Additionally, we used inverse-dynamics techniques to quantify the intersegmental dynamics of the paw, leg, and thigh. Our data indicate that by the second cycle of both steady-state and nonsteady-state responses, the basic pattern of interaction between muscle and motion-dependent torques at the ankle and knee joints was established. During subsequent cycles of steady-state responses, a consistent sequence of timing changes occurred, such that, just prior to steady-state oscillations, torque maximums peaked simultaneously at each joint and joint reversals occurred simultaneously. Although nonsteady-state responses showed a similar sequence during beginning cycles, increased ankle muscle and net torques during middle cycles created larger inertial torques at the knee joint that were not counteracted and resulted in irregular and uncoupled knee oscillations. It is likely that neither steady-state nor nonsteady-state oscillations are planned by pattern-generating networks within lumbosacral segments, but that patterns of interjoint coordination emerge from the coupling among oscillators. For paw-shake responses in the spinalized preparation, coupling may depend on interactions between central circuits and motion-dependent feedback that is necessary to stabilize inertial effects due to large ankle joint accelerations.

Paw-shake responses with joint immobilization: EMG changes with atypical feedback

To determine the effects of atypical motion-related feedback on motor patterns of the paw shake, EMG patterns of selected flexor and extensor muscles were recorded under four conditions of joint immobilization (hip and ankle alone, hip-knee, hip-knee-ankle) and compared to responses evoked in the freely-moving hindlimb of the chronic-spinal cat. With only the ankle joint casted, paw shaking was easily evoked by applying tape to the paw, and cyclic characteristics were not altered. However, under the three conditions with hip-joint immobilization (hip alone, hip-knee, hip-knee-ankle), responses were difficult to obtain, and if elicited, the number of cycles within a response decreased and cycle periods were prolonged. The temporal organization of consecutive cycles, however, was not altered by immobilization of any joint(s). Ankle (LG) and hip (GM) extensor activity was relatively unaffected by conditions of joint immobilization. In contrast, hip flexor (IP) and knee extensor (VL) bursts were often absent under all three conditions of hip-joint immobilization, and if present, VL burst durations decreased under the casted hip-knee-ankle condition, while the onset of IP activity occurred early in the cycle with prolonged bursts under casted ankle and casted hip-knee-ankle conditions. The coactivity of the knee extensor (VL) and ankle flexor (TA) was disrupted by conditions of hip-joint immobilization: VL onset was dissociated from TA onset and coincident with LG onset. These results suggest that motion-related feedback from the hip joint is particularly important in the initiation, cycle frequency, and the number of cycles of paw-shake responses. The presence of atypical motion-dependent feedback from the hip joint altered activity of knee and ankle anterior muscles, while motion-dependent feedback from the ankle joint changed activity of the anterior hip muscle. Moreover, the results suggest a differential control of posterior and anterior muscles of the hindlimb, consistent with paw-shake limb dynamics.

Mutable and immutable features of paw-shake responses after hindlimb deafferentation in the cat

1. Hindlimb paw-shake responses were assessed before and after unilateral deafferentation (L3-S1) in chronic-spinal cats (n = 5), spinalized at the T12 level 1 yr earlier. Selected ankle flexor [tibialis anterior (TA)] and extensor [lateral gastrocnemius (LG)] and knee extensor [vastus lateralis (VL)] muscles were surgically implanted with chronic electromyographic (EMG) electrodes to determine mutable features of cycle characteristics and muscle synergies that are modulated by motion-dependent feedback as opposed to immutable features that are centrally programmed and not modulated by limb afference. 2. Paw-shake responses were difficult to elicit in the extensively deafferented hindlimb; this was true particularly during the first recovery weeks after deafferentation. By the end of the first month, however, brief responses of 1 or 2 cycles were commonly elicited in four of five cats, and responses of 3-7 cycles were common by the end of the second month in three of five cats. Initially, responses in the deafferented limb were elicited by stimuli applied to the dorsolateral thigh, an oval patch of skin innervated by intact S2 afferents. Over the 4-mo recovery period, however, the receptive field of the largely denervated skin expanded, and responses were also elicited by stimuli applied to the lateral aspect of the knee and shank, but usually not the paw. 3. In addition to fewer average cycles per response (5 vs. 10 cycles), paw shaking evoked in the deafferented hindlimb was characterized by longer-than-average cycle periods (124 vs. 98 ms), but the average cycle period varied widely among responses, ranging from 99 to 239 ms. Before deafferentation, the temporal organization of consecutive cycles was stereotypic; cycle periods increased linearly throughout a response. After deafferentation, however, there was no systematic relationship between cycle period and cycle number, and approximately 14% of the records with greater than or equal to 3 cycles were characterized by arhythmical sequences of EMG bursts. 4. At the ankle, LG burst duration was not altered by deafferentation, but TA onset and burst duration were affected. Before deafferentation, TA onset was invariant with respect to the beginning of the cycle, and burst duration increased linearly with cycle period. After deafferentation, however, TA onset was delayed, and the delay increased linearly with cycle period. Consequently, the TA burst duration was brief and unrelated to cycle period.(ABSTRACT TRUNCATED AT 400 WORDS)

Intralimb coordination of the paw-shake response: a novel mixed synergy

Intralimb coordination of the paw-shake response (PSR) was studied in five normal and eleven spinal adult cats. Representative extensor and flexor muscles that function at the hip, knee, and ankle joints were recorded, and in six spinal cats the kinematics of these joints were determined from high-speed cinefilm. The PSR was characterized uniquely by mixed (flexor-extensor) synergies. Knee extensor (VL) and ankle flexor (TA) coactivity constituted one synergy, while the second synergy included hip extensors (GM, BF), knee flexors (BF, LG), and ankle extensor (LG). Joint displacements reflected the mixed synergy. Motions at the knee and ankle were out of phase, while motions at the hip were in phase with movements of the knee. Electromyographic burst durations and onset latencies were similar for normal and spinal cats, and in all cycles of a given PSR, the recruitment pattern was consistent for all muscles, except VL. High variability and missing bursts marked the activity of VL in some spinal cats. In PSRs with missing VL bursts, oscillations at the knee joint were not coordinated with cyclic actions at the hip and ankle. From the kinematic records three distinct phases of the PSR were identified: start-up consisted of the initial four to six cycles during which hip, knee, and ankle actions progressively became organized; steady-state included the middle three to five cycles that were characterized by consistent displacement at all three joints; and slow-down comprised the last three to four cycles during which the rate of oscillations slowed, and joint excursions decreased. During steady-state cycles, muscle contractions acted to reverse joint motions at the knee and ankle joints. Thus, knee and ankle extensor recruitment coincided with joint flexion, while joint flexors were recruited during joint extension. Muscle activity at the hip, however, was in phase with displacement. While neural input to muscle is consistent throughout the three phases of the PSR, segment motions can become progressively organized during start-up to achieve stable oscillations. Whether the PSR attains steady-state or not may hinge on the sensitive interplay that occurs between muscle activities and intersegmental mechanical interactions. That kinetic interplay is detailed in the following paper.

Effect of auditory rhythm on muscle activity

the changes in the electromyogram patterns of two antagonist muscles were studied when female subjects performed a motor task with and without an auditory rhythm. During the performance of the motor task without the rhythm, the subjects demonstrated a common and consistent personal tempo and a common electromyogram pattern. With the imposed timing of an even or uneven rhythm, the initiation and duration of the electromyograms changed significantly for subjects in both rhythm groups. Variations of electromyograms decreased significantly for subjects inthe even-rhythm group and increased significantly for subjects in the uneven-rhythm group. The authors suggest specific ways in which rhythm can be used in rehabilitative techniques to modify the onset, duration, and inconsistency of muscular activity.