Muscle synergies encoded within the spinal cord: evidence from focal intraspinal NMDA iontophoresis in the frog

Identifying representative synergy matrices for describing muscular activation patterns during multidirectional reaching in the horizontal plane

Muscle synergies have been proposed as a simplifying principle of generation of movements based on a low-dimensional control by the CNS. This principle may be useful for movement restoration by, e.g., functional electrical stimulation (FES), if a limited set of synergies can describe several functional tasks. This study investigates the possibility of describing a multijoint reaching task of the upper limb by a linear combination of one set of muscle synergies common to multiple directions. Surface electromyographic (EMG) signals were recorded from 12 muscles of the dominant upper limb of eight healthy men during single-joint movements and a multijoint reaching task in 12 directions in the horizontal plane. The movement kinematics was recorded by a motion analysis system. Muscle synergies were extracted with nonnegative matrix factorization of the EMG envelopes. Synergies were computed either from the single-joint movements to describe the two degrees of freedom independently or from the multijoint movements. On average, the multijoint reaching task could be accurately described in all the directions (coefficient of determination >0.85) by a linear combination of either four synergies extracted from the individual degrees of freedom or three synergies extracted from multijoint movements in at least three reaching directions. These results indicate that a large set of multijoint movements can be generated by a synergy matrix of limited dimensionality and common to all directions if the synergies are extracted from a representative number of directions. The linear combination of synergies may thus be used in strategies for restoring functions, such as FES.

Multi-muscle synergies in a dual postural task: evidence for the principle of superposition

We used the framework of the uncontrolled manifold hypothesis to quantify multi-muscle synergies stabilizing the moment of force about the frontal axis (M(Y)) and the shear force in the anterior-posterior direction (F(X)) during voluntary body sway performed by standing subjects. We tested a hypothesis whether the controller could stabilize both M(Y) and F(X) at the same time when the task and the visual feedback was provided only on one of the variables (M(Y)). Healthy young subjects performed voluntary body sway in the anterior-posterior direction while different loads were attached at the ankle level producing horizontal forces acting forward or backwards. Principal component analysis was used to identify three M-modes within the space of integrated indices of muscle activation. Variance in the M-mode space across sway cycles was partitioned into two components, one that did not affect a selected performance variable (M(Y) or F(X)) and the other that did. Under all loading conditions and for each performance variable, a higher value for the former variance component was found. We interpret these results as reflections of two multi-M-mode synergies stabilizing both F(X) and M(Y). The indices of synergies were modulated within the sway cycle; both performance variables were better stabilized when the body moved forward than when it moved backward. The results show that the controller can use a set of three elemental variables (M-modes) to stabilize two performance variables at the same time. No negative interference was seen between the synergy indices computed for the two performance variables supporting the principle of superposition with respect to multi-muscle postural control.

A simple experimentally based model using proprioceptive regulation of motor primitives captures adjusted trajectory formation in spinal frogs

Spinal circuits may organize trajectories using pattern generators and synergies. In frogs, prior work supports fixed-duration pulses of fixed composition synergies, forming primitives. In wiping behaviors, spinal frogs adjust their motor activity according to the starting limb position and generate fairly straight and accurate isochronous trajectories across the workspace. To test whether a compact description using primitives modulated by proprioceptive feedback could reproduce such trajectory formation, we built a biomechanical model based on physiological data. We recorded from hindlimb muscle spindles to evaluate possible proprioceptive input. As movement was initiated, early skeletofusimotor activity enhanced many muscle spindles firing rates. Before movement began, a rapid estimate of the limb position from simple combinations of spindle rates was possible. Three primitives were used in the model with muscle compositions based on those observed in frogs. Our simulations showed that simple gain and phase shifts of primitives based on published feedback mechanisms could generate accurate isochronous trajectories and motor patterns that matched those observed. Although on-line feedback effects were omitted from the model after movement onset, our primitive-based model reproduced the wiping behavior across a range of starting positions. Without modifications from proprioceptive feedback, the model behaviors missed the target in a manner similar to that in deafferented frogs. These data show how early proprioception might be used to make a simple estimate initial limb state and to implicitly plan a movement using observed spinal motor primitives. Simulations showed that choice of synergy composition played a role in this simplicity. To generate froglike trajectories, a hip flexor synergy without sartorius required motor patterns with more proprioceptive knee flexor control than did patterns built with a more natural synergy including sartorius. Such synergy choices and control strategies may simplify the circuitry required for reflex trajectory construction and adaptation.

Stability of muscle synergies for voluntary actions after cortical stroke in humans

Production of voluntary movements relies critically on the functional integration of several motor cortical areas, such as the primary motor cortex, and the spinal circuitries. Surprisingly, after almost 40 years of research, how the motor cortices specify descending neural signals destined for the downstream interneurons and motoneurons has remained elusive. In light of the many recent experimental demonstrations that the motor system may coordinate muscle activations through a linear combination of muscle synergies, we hypothesize that the motor cortices may function to select and activate fixed muscle synergies specified by the spinal or brainstem networks. To test this hypothesis, we recorded electromyograms (EMGs) from 12-16 upper arm and shoulder muscles from both the unaffected and the stroke-affected arms of stroke patients having moderate-to-severe unilateral ischemic lesions in the frontal motor cortical areas. Analyses of EMGs using a nonnegative matrix factorization algorithm revealed that in seven of eight patients the muscular compositions of the synergies for both the unaffected and the affected arms were strikingly similar to each other despite differences in motor performance between the arms, and differences in cerebral lesion sizes and locations between patients. This robustness of muscle synergies that we observed supports the notion that descending cortical signals represent neuronal drives that select, activate, and flexibly combine muscle synergies specified by networks in the spinal cord and/or brainstem. Our conclusion also suggests an approach to stroke rehabilitation by focusing on those synergies with altered activations after stroke.

Muscle synergies as a predictive framework for the EMG patterns of new hand postures

Synchronous muscle synergies have been suggested as a framework for dimensionality reduction in muscle coordination. Many studies have shown that synergies form a descriptive framework for a wide variety of tasks. We examined if a muscle synergy framework could accurately predict the EMG patterns associated with untrained static hand postures, in essence, if they formed a predictive framework. Hand and forearm muscle activities were recorded while subjects statically mimed 33 postures of the American Sign Language alphabet. Synergies were extracted from a subset of training postures using non-negative matrix factorization and used to predict the EMG patterns of the remaining postures. Across the subject population, as few as 11 postures could form an eight-dimensional synergy framework that allowed for at least 90% prediction of the EMG patterns of all 33 postures, including trial-to-trial variations. Synergies were quite robust despite using different postures in the training set, and also despite using a varied number of postures. Estimated synergies were categorized into those which were subject-specific and those which were general to the population. Population synergies were sparser than the subject-specific synergies, typically being dominated by a single muscle. Subject-specific synergies were more balanced in the coactivation of multiple muscles. We suggest as a result that global muscle coordination may be a combination of higher order control of robust subject-specific muscle synergies and lower order control of individuated muscles, and that this control paradigm may be useful in the control of EMG-based technologies, such as artificial limbs and functional electrical stimulation systems.

Multi-finger prehension: control of a redundant mechanical system

The human hand has been a fascinating object of study for researchers in both biomechanics and motor control. Studies of human prehension have contributed significantly to the progress in addressing the famous problem of motor redundancy. After a brief review of the hand mechanics, we present results of recent studies that support a general view that the apparently redundant design of the hand is not a source of computational problems but a rich apparatus that allows performing a variety of tasks in a reliable and flexible way (the principle of abundance). Multi-digit synergies have been analyzed at two levels of a hypothetical hierarchy involved in the control of prehensile actions. At the upper level, forces and moments produced by the thumb and virtual finger (an imagined finger with a mechanical action equal to the combined mechanical action of all four fingers of the hand) co-vary to stabilize the gripping action and the orientation of the hand-held object. These results support the principle of superposition suggested earlier in robotics with respect to the control of artificial grippers. At the lower level of the hierarchy, forces and moments produced by individual fingers co-vary to stabilize the magnitude and direction of the force vector and the moment of force produced by the virtual finger. Adjustments to changes in task constraints (such as, for example, friction under individual digits) may be local and synergic. The latter reflect multi-digit prehension synergies and may be analyzed with the so-called chain effects: Sequences of relatively straightforward cause-effect links directly related to mechanical constraints leading to non-trivial strong co-variation between pairs of elemental variables. Analysis of grip force adjustments during motion of hand-held objects suggests that the central nervous system adjusts to gravitational and inertial loads differently. The human hand is a gold mine for researchers interested in the control of natural human movements.

Adjustments of motor pattern for load compensation via modulated activations of muscle synergies during natural behaviors

It has been suggested that the motor system may circumvent the difficulty of controlling many degrees of freedom in the musculoskeletal apparatus by generating motor outputs through a combination of discrete muscle synergies. How a discretely organized motor system compensates for diverse perturbations has remained elusive. Here, we investigate whether motor responses observed after an inertial-load perturbation can be generated by altering the recruitment of synergies normally used for constructing unperturbed movements. Electromyographic (EMG, 13 muscles) data were collected from the bullfrog hindlimb during natural behaviors before, during, and after the same limb was loaded by a weight attached to the calf. Kinematic analysis reveals the absence of aftereffect on load removal, suggesting that load-related EMG changes were results of immediate motor pattern adjustments. We then extracted synergies from EMGs using the nonnegative matrix factorization algorithm and developed a procedure for assessing the extent of synergy sharing across different loading conditions. Most synergies extracted were found to be activated in all loaded and unloaded conditions. However, for certain synergies, the amplitude, duration, and/or onset time of their activation bursts were up- or down-modulated during loading. Behavioral parameterizations reveal that load-related modulation of synergy activations depended on the behavioral variety (e.g., kick direction and amplitude) and the movement phase performed. Our results suggest that muscle synergies are robust across different dynamic conditions and immediate motor adjustments can be accomplished by modulating synergy activations. An appendix describes the novel procedure we developed, useful for discovering shared and specific features from multiple data sets.

Endpoint force fluctuations reveal flexible rather than synergistic patterns of muscle cooperation

We developed a new approach to investigate how the nervous system activates multiple redundant muscles by studying the endpoint force fluctuations during isometric force generation at a multi-degree-of-freedom joint. We hypothesized that, due to signal-dependent muscle force noise, endpoint force fluctuations would depend on the target direction of index finger force and that this dependence could be used to distinguish flexible from synergistic activation of the musculature. We made high-gain measurements of isometric forces generated to different target magnitudes and directions, in the plane of index finger metacarpophalangeal joint abduction-adduction/flexion-extension. Force fluctuations from each target were used to calculate a covariance ellipse, the shape of which varied as a function of target direction. Directions with narrow ellipses were approximately aligned with the estimated mechanical actions of key muscles. For example, targets directed along the mechanical action of the first dorsal interosseous (FDI) yielded narrow ellipses, with 88% of the variance directed along those target directions. It follows the FDI is likely a prime mover in this target direction and that, at most, 12% of the force variance could be explained by synergistic coupling with other muscles. In contrast, other target directions exhibited broader covariance ellipses with as little as 30% of force variance directed along those target directions. This is the result of cooperation among multiple muscles, based on independent electromyographic recordings. However, the pattern of cooperation across target directions indicates that muscles are recruited flexibly in accordance with their mechanical action, rather than in fixed groupings.

Extracting simultaneous and proportional neural control information for multiple-DOF prostheses from the surface electromyographic signal

A novel signal processing algorithm for the surface electromyogram (EMG) is proposed to extract simultaneous and proportional control information for multiple DOFs. The algorithm is based on a generative model for the surface EMG. The model assumes that synergistic muscles share spinal neural drives, which correspond to the intended activations of different DOFs of natural movements and are embedded within the surface EMG. A DOF-wise nonnegative matrix factorization (NMF) is developed to estimate neural control information from the multichannel surface EMG. It is shown, both by simulation and experimental studies, that the proposed algorithm is able to extract the multidimensional control information simultaneously. A direct application of the proposed method would be providing simultaneous and proportional control of multifunction myoelectric prostheses.

Modulation of Phasic and Tonic Muscle Synergies With Reaching Direction and Speed

How the CNS masters the many degrees of freedom of the musculoskeletal system to control goal-directed movements is a long-standing question. We have recently provided support to the hypothesis that the CNS relies on a modular control architecture by showing that the phasic muscle patterns for fast reaching movements in different directions are generated by combinations of a few time-varying muscle synergies: coordinated recruitment of groups of muscles with specific activation profiles. However, natural reaching movements occur at different speeds and require the control of both movement and posture. Thus we have investigated whether muscle synergies also underlie reaching at different speeds as well as the maintenance of stable arm postures. Hand kinematics and shoulder and elbow muscle surface EMGs were recorded in five subjects during reaches to eight targets in the frontal plane at different speeds. We found that the amplitude modulation of three time-invariant synergies captured the variations in the postural muscle patterns at the end of the movement. During movement, three phasic and three tonic time-varying synergies could reconstruct the time-normalized muscle pattern in all conditions. Phasic synergies were modulated in both amplitude and timing by direction and speed. Tonic synergies were modulated only in amplitude by direction. The directional tuning of both types of synergies was well described by a single or a double cosine function. These results suggest that muscle synergies are basic control modules that allow generating the appropriate muscle patterns through simple modulation and combination rules.

Flexible muscle modes and synergies in challenging whole-body tasks

We used the idea of hierarchical control to study multi-muscle synergies during a whole-body sway task performed by a standing person. Within this view, at the lower level of the hierarchy, muscles are united into groups (M-modes). At the higher level, gains at the M-modes are co-varied by the controller in a task-specific way to ensure low variability of important physical variables. In particular, we hypothesized that (1) the composition of M-modes could adjust and (2) an index of M-mode co-variation would become weaker in more challenging conditions. Subjects were required to perform a whole-body sway at 0.5 Hz paced by a metronome. They performed the task with eyes open and closed, while standing on both feet or on one foot only, with and without vibration applied to the Achilles tendons. Integrated indices of muscle activation were subjected to principal component analysis to identify M-modes. An increase in the task complexity led to an increase in the number of principal components that contained significantly loaded indices of muscle activation from 3 to 5. Hence, in more challenging tasks, the controller manipulated a larger number of variables. Multiple regression analysis was used to define the Jacobian of the system mapping small changes in M-mode gains onto shifts of the center of pressure (COP) in the anterior-posterior direction. Further, the variance in the M-mode space across sway cycles was partitioned into two components, one that did not affect an average across cycles COP coordinate and the other that did (good and bad variance, respectively). Under all conditions, the subjects showed substantially more good variance than bad variance interpreted as a multi-M-mode synergy stabilizing the COP trajectory. An index of the strength of the synergy was comparable across all conditions, and there was no modulation of this index over the sway cycle. Hence, our first hypothesis that the composition of M-modes could adjust under challenging conditions has been confirmed while the second hypothesis stating that the index of M-mode co-variation would become weaker in more challenging conditions has been falsified. We interpret the observations as suggesting that adjustments at the lower level of the hierarchy-in the M-mode composition-allowed the subjects to maintain a comparable level of stabilization of the COP trajectory in more challenging tasks. The findings support the (at least) two-level hierarchical control scheme of whole-body movements.

Multi-muscle synergies in an unusual postural task: quick shear force production

We considered a hypothetical two-level hierarchy participating in the control of vertical posture. The framework of the uncontrolled manifold (UCM) hypothesis was used to explore the muscle groupings (M-modes) and multi-M-mode synergies involved in the stabilization of a time profile of the shear force in the anterior-posterior direction. Standing subjects were asked to produce pulses of shear force into a target using visual feedback while trying to minimize the shift of the center of pressure (COP). Principal component analysis applied to integrated muscle activation indices identified three M-modes. The composition of the M-modes was similar across subjects and the two directions of the shear force pulse. It differed from the composition of M-modes described in earlier studies of more natural actions associated with large COP shifts. Further, the trial-to-trial M-mode variance was partitioned into two components: one component that does not affect a particular performance variable (V(UCM)), and its orthogonal component (V(ORT)). We argued that there is a multi-M-mode synergy stabilizing this particular performance variable if V(UCM) is higher than V(ORT). Overall, we found a multi-M-mode synergy stabilizing both shear force and COP coordinate. For the shear force, this synergy was strong for the backward force pulses and nonsignificant for the forward pulses. An opposite result was found for the COP coordinate: the synergy was stronger for the forward force pulses. The study shows that M-mode composition can change in a task-specific way and that two different performance variables can be stabilized using the same set of elemental variables (M-modes). The different dependences of the ΔV indices for the shear force and COP coordinate on the force pulse direction supports applicability of the principle of superposition (separate controllers for different performance variables) to the control of different mechanical variables in postural tasks. The M-mode composition allows a natural mechanical interpretation.

Modulation of muscle synergy recruitment in primate grasping

In grasping, the CNS controls a particularly large number of degrees of freedom. We tested the idea that this control is facilitated by the presence of muscle synergies. According to the strong version of this concept, these synergies are invariant, hard-wired patterns of activation across muscles. Synergies may serve as modules that linearly sum, each with specific amplitude and timing coefficients, to generate a large array of muscle patterns. We tested two predictions of the synergy model. A small number of synergies should (1) account for a large fraction of variation in muscle activity, and (2) be modulated in their recruitment by task variables, even in novel behavioral contexts. We also examined whether the synergies would (3) have broadly similar structures across animals. We recorded from 15 to 19 electrodes implanted in forelimb muscles of two rhesus macaques as they grasped and transported 25 objects of variable shape and size. We show that three synergies accounted for 81% of the electromyographic data variation in each monkey. Each synergy was modulated in its recruitment strength and/or timing by object shape and/or size. Even when synergies were extracted from a small subset of object shape and size conditions and then used to reconstruct the entire dataset, we observed highly similar synergies and patterns of modulation. The synergies were well conserved between monkeys, with two of the synergies exceeding chance structural similarity, and the third being recruited, in both animals, in proportion to the size of the object handled.

Muscle synergies during locomotion in the cat: a model for motor cortex control

It is well established that the motor cortex makes an important contribution to the control of visually guided gait modifications, such as those required to step over an obstacle. However, it is less clear how the descending cortical signal interacts with the interneuronal networks in the spinal cord to ensure that precise changes in limb trajectory are appropriately incorporated into the base locomotor rhythm. Here we suggest that subpopulations of motor cortical neurones, active sequentially during the step cycle, may regulate the activity of small groups of synergistic muscles, likewise active sequentially throughout the step cycle. These synergies, identified by a novel associative cluster analysis, are defined by periods of muscle activity that are coextensive with respect to the onset and offset of the EMG activity. Moreover, the synergies are sparse and are frequently composed of muscles acting around more than one joint. During gait modifications, we suggest that subpopulations of motor cortical neurones may modify the magnitude and phase of the EMG activity of all muscles contained within a given synergy. Different limb trajectories would be produced by differentially modifying the activity in each synergy thus providing a flexible substrate for the control of intralimb coordination during locomotion.

Extracting neural drives from surface EMG: a generative model and simulation studies

A generative model for the surface EMG is presented. The model is built on the assumption that motor units in synergistic muscle share neural drives from spinal level, which correspond to the activation of different degrees of freedom (DOF) of natural movements, and are embedded within surface EMG. An artificial neural network (ANN) is developed to extract these drives simultaneously from the multi-channel surface EMG. A direct application of this technique would be providing control signals to prosthetic devices that are capable of simultaneous control of multiple DOF. It also has potential applications in the diagnosis and rehabilitation of spinal cord injuries and other neuromuscular disorders.

Muscle Synergies Characterizing Human Postural Responses

Postural control is a natural behavior that requires the spatial and temporal coordination of multiple muscles. Complex muscle activation patterns characterizing postural responses suggest the need for independent muscle control. However, our previous work shows that postural responses in cats can be robustly reproduced by the activation of a few muscle synergies. We now investigate whether a similar neural strategy is used for human postural control. We hypothesized that a few muscle synergies could account for the intertrial variability in automatic postural responses from different perturbation directions, as well as different postural strategies. Postural responses to multidirectional support-surface translations in 16 muscles of the lower back and leg were analyzed in nine healthy subjects. Six or fewer muscle synergies were required to reproduce the postural responses of each subject. The composition and temporal activation of several muscle synergies identified across all subjects were consistent with the previously identified “ankle” and “hip” strategies in human postural responses. Moreover, intertrial variability in muscle activation patterns was successfully reproduced by modulating the activity of the various muscle synergies. This suggests that trial-to-trial variations in the activation of individual muscles are correlated and, moreover, represent variations in the amplitude of descending neural commands that activate individual muscle synergies. Finally, composition and temporal activation of most of the muscle synergies were similar across subjects. These results suggest that muscle synergies represent a general neural strategy underlying muscle coordination in postural tasks.

Combining modules for movement

We review experiments supporting the hypothesis that the vertebrate motor system produces movements by combining a small number of units of motor output. Using a variety of approaches such as microstimulation of the spinal cord, NMDA iontophoresis, and an examination of natural behaviors in intact and deafferented animals we have provided evidence for a modular organization of the spinal cord. A module is a functional unit in the spinal cord that generates a specific motor output by imposing a specific pattern of muscle activation. Such an organization might help to simplify the production of movements by reducing the degrees of freedom that need to be specified.

Action-based sensory encoding in spinal sensorimotor circuits

The concept of a modular organisation of the spinal withdrawal reflex circuits has proven to be fundamental for the understanding of how the spinal cord is organised and how the sensorimotor circuits translate sensory information into adequate movement corrections. Recent studies indicate that a task-related body representation is engraved at the network level through learning-dependent mechanisms involving an active probing procedure termed 'somatosensory imprinting' during development. It was found that somatosensory imprinting depends on the tactile input that is associated with spontaneous movements that occur during sleep and results in elimination of erroneous connections and establishment of correct connections. In parallel studies it was found that the strength of the first order tactile synapses in rostrocaudally elongated zones in the adult dorsal horn in the lower lumbar cord is related to the modular organisation of the withdrawal reflexes. Hence, the topographical organisation of the tactile input to this spinal area seems to be action-based rather than a simple body map as previously thought. Far from being innate and adult like at birth, the adult organisation seems to emerge from an initial 'floating' and diffuse body representation with many inappropriate connections through profound activity-dependent rearrangements of afferent synaptic connections. It is suggested that somatosensory imprinting plays a key role in the self-organisation of the spinal cord during development.

Learning effects on muscle modes and multi-mode postural synergies

We used the framework of the uncontrolled manifold hypothesis to explore the effects of practice on the composition of muscle groups (M-modes) and multi-M-mode synergies stabilizing the location of the center of pressure (COP). In particular, we tested a hypothesis that practice could lead to a transition from co-contraction muscle activation patterns to reciprocal patterns. We also tested a hypothesis that new sets of M-modes would form stronger synergies stabilizing COP location. Subjects practiced load release tasks for five days while standing on a board with a narrow support surface (unstable board). Their M-modes and indices of multi-M-mode synergies were computed during standing without instability and during standing on an unstable board before practice, in the middle of practice, and at the end of practice. During standing without instability, subjects showed two consistent M-modes uniting dorsal and ventral muscles of the body respectively (reciprocal modes). While standing on an unstable board, prior to practice, subjects commonly showed M-modes uniting muscle pairs with opposing actions at major leg joints-co-contraction modes. Such sets of M-modes failed to stabilize the COP location in the anterior-posterior direction. Practice led to better task performance reflected in fewer incidences of lost balance. This was accompanied by a drop in the occurrence of co-contraction M-modes and the emergence of multi-mode synergies stabilizing the COP location. We conclude that the central nervous system uses flexible sets of elemental variables (modes) to ensure stable trajectories of important performance variables (such as COP location). Practice can lead to adjustments in both the composition of M-modes and M-mode co-variation patterns resulting in stronger synergies stabilizing COP location.

Role of biomechanics and muscle activation strategy in the production of endpoint force patterns in the cat hindlimb

We used a musculoskeletal model of the cat hindlimb to compare the patterns of endpoint forces generated by all possible combination of 12 hindlimb muscles under three different muscle activation rules: homogeneous activation of muscles based on uniform activation levels, homogeneous activation of muscles based on uniform (normalized) force production, and activation based on the topography of spinal motoneuron pools. Force patterns were compared with the patterns obtained experimentally by microstimulation of the lumbar spinal cord in spinal intact cats. Magnitude and orientation of the force patterns were compared, as well as the proportion of the types found, and the proportions of patterns exhibiting points of zero force (equilibrium points). The force patterns obtained with the homogenous activation and motoneuron topography models were quite similar to those measured experimentally, with the differences being larger for the patterns from the normalized endpoint forces model. Differences in the proportions of types of force patterns between the three models and the experimental results were significant for each model. Both homogeneous activation and normalized endpoint force models produced similar proportions of equilibrium points as found experimentally. The results suggest that muscle biomechanics play an important role in limiting the number of endpoint force pattern types, and that muscle combinations activated at similar levels reproduced best the experimental results obtained with intraspinal microstimulation.

Humeral retractor EMG during quadrupedal walking in primates

The mammalian humeral retractors latissimus dorsi, teres major and caudal parts of the pectoral muscles are commonly thought to contribute to forward impulse during quadrupedal locomotion by pulling the body over the supporting forelimb. While most electromyographic studies on recruitment patterns for these muscles tend to support this functional interpretation, data on muscle use in chimpanzees and vervet monkeys have suggested that the humeral retractors of nonhuman primates are largely inactive during the support phase of quadrupedal locomotion. In the chimpanzee and vervet monkey, in contrast to what has been documented for other mammals, the contributions of latissimus dorsi, caudal pectoralis major, and teres major during quadrupedal locomotion are restricted to slowing down the swinging forelimb in preparation for hand touchdown and/or retracting the humerus to help lift the hand off the substrate at the initiation of swing phase. Based on these results, it has been proposed that unique patterns of shoulder muscle recruitment are among a set of characteristics that distinguish the form of quadrupedalism displayed by nonhuman primates from that of other nonprimate mammals. However, two primate taxa is a limited sample upon which to base such far-reaching conclusions. Here we report on the activity patterns for the humeral retractors during quadrupedal walking in an additional eight species of nonhuman primates. There is some variability in the activity patterns for latissimus dorsi, caudal pectoralis major and teres major, both between and within species, but in general the results confirm that the humeral retractors of primate quadrupeds do not contribute to forward impulse by pulling the body over the supporting forelimb.

Muscle modes and synergies during voluntary body sway

We studied the coordination of muscle activity during voluntary body sway performed by human subjects at different frequencies. Subjects stood on the force platform and performed cyclic shifts of the center of pressure (COP) while being paced by the metronome. A major question was: does the makeup of muscle synergies and their ability to assure reproducible sway trajectory vary with the speed of the sway? Principal component analysis was used to identify three muscle groups (M-modes) within the space of integrated indices of muscle activity. M-mode vectors were similar across both subjects and sway frequencies. There were also similar relations between changes in the magnitudes of all three M-modes and COP shifts (the Jacobians) across the sway frequencies. Variance in the M-mode space across sway cycles was partitioned into two components, one that did not affect the average value of COP shift ("good variance") and the other that did. An index (DeltaV) was computed reflecting the relative amount of the "good variance"; this index has been interpreted as reflecting a multi-M-mode synergy stabilizing the COP trajectory. The average value of DeltaV was similar across all sway frequencies; DeltaV showed a within-a-cycle modulation at low but not at high sway frequencies. The modulation was mostly due to variations in the "good variance". We conclude that muscle modes and their mapping on COP shifts are robust across a wide range of rates of COP shifts. Multi-M-mode synergies stabilize COP shifts (assure its reproducibility) within a wide range of its speeds, but only during cyclic COP changes. Taken together with earlier studies that showed weak or absent multi-M-mode synergies during fast discrete COP shifts, the results suggest a basic difference between the neural control assuring stability of steady-state processes (postural or oscillatory) and transient processes (such as discrete actions). Current results provide the most comprehensive support for the notion of multi-M-mode synergies stabilizing time profiles of important performance variables in motor tasks involving large muscle groups.

The use of flexible arm muscle synergies to perform an isometric stabilization task

OBJECTIVE

To evaluate if functional synergies are comprised of flexible combinations of a small number of ensembles of upper limb muscles to stabilize a particular performance variable during a force production task.

METHODS

Electromyographic (EMG) signals of wrist, elbow and shoulder muscles were recorded. Linear combinations of indices of the integrated EMG of nine muscles (muscle modes or M-modes) and their relationship to changes in the moment of force produced by the right arm of subjects about the horizontal axis of a handle were first determined. Uncontrolled manifold (UCM) analysis was performed to determine the extent to which variance of the M-modes acted to produce a consistent change in the moment of force.

RESULTS

Subjects exhibited three M-modes, two of which were 'reciprocal' while the third mode was considered a co-contraction mode. The three M-modes were found to be combined to form synergies that produced a consistent change in the moment of force across repetitive trials. Variance in the M-mode space that led to consistent changes in the moment of force across repetitions was significantly higher than variance that tended to produce inconsistent changes in the moment of force.

CONCLUSIONS

Flexible combinations of activations of ensembles of muscles are organized to stabilize the value of or produce consistent changes in the value of important performance variables.

SIGNIFICANCE

The study of flexible muscle synergies in healthy individuals paves the way to understanding abnormal postural and movement patterns in individuals with neurological disorders.

Running over rough terrain reveals limb control for intrinsic stability

Legged animals routinely negotiate rough, unpredictable terrain with agility and stability that outmatches any human-built machine. Yet, we know surprisingly little about how animals accomplish this. Current knowledge is largely limited to studies of steady movement. These studies have revealed fundamental mechanisms used by terrestrial animals for steady locomotion. However, it is unclear whether these models provide an appropriate framework for the neuromuscular and mechanical strategies used to achieve dynamic stability over rough terrain. Perturbation experiments shed light on this issue, revealing the interplay between mechanics and neuromuscular control. We measured limb mechanics of helmeted guinea fowl (Numida meleagris) running over an unexpected drop in terrain, comparing their response to predictions of the mass-spring running model. Adjustment of limb contact angle explains 80% of the variation in stance-phase limb loading following the perturbation. Surprisingly, although limb stiffness varies dramatically, it does not influence the response. This result agrees with a mass-spring model, although it differs from previous findings on humans running over surfaces of varying compliance. However, guinea fowl sometimes deviate from mass-spring dynamics through posture-dependent work performance of the limb, leading to substantial energy absorption following the perturbation. This posture-dependent actuation allows the animal to absorb energy and maintain desired velocity on a sudden substrate drop. Thus, posture-dependent work performance of the limb provides inherent velocity control over rough terrain. These findings highlight how simple mechanical models extend to unsteady conditions, providing fundamental insights into neuromuscular control of movement and the design of dynamically stable legged robots and prosthetic devices.

Intraspinal microstimulation excites multisegmental sensory afferents at lower stimulus levels than local alpha-motoneuron responses

Microstimulation within the motor regions of the spinal cord is often assumed to activate motoneurons and propriospinal neurons close to the electrode tip. However, previous work has shown that intraspinal microstimulation (ISMS) in the gray matter activates sensory afferent axons as well as alpha-motoneurons (MNs). Here we report on the recruitment of sensory afferent axons and MNs as ISMS amplitudes increased. Intraspinal microstimulation was applied through microwires implanted in the dorsal horn, intermediate region and ventral horn of the L(5)-L(7) segments of the spinal cord in four acutely decerebrated cats, two of which had been chronically spinalized. Activation of sensory axons was detected with electroneurographic recordings from dorsal roots. Activation of MNs was detected with electromyographic (EMG) recordings from hindlimb muscles. Sensory axons were nearly always activated at lower stimulus levels than MNs irrespective of the stimulating electrode location. EMG response latencies decreased as ISMS stimulus intensities increased, suggesting that MNs were first activated transsynaptically and then directly as intensity increased. ISMS elicited antidromic activity in dorsal root filaments with entry zones up to 17 mm rostral and caudal to the stimulation sites. We posit that action potentials elicited in localized terminal branches of afferents spread antidromically to all terminal branches of the afferents and transsynaptically excite MNs and interneurons far removed from the stimulation site. This may help explain how focal ISMS can activate many MNs of a muscle even though they are distributed in long thin columns.

Control of fast-reaching movements by muscle synergy combinations

How the CNS selects the appropriate muscle patterns to achieve a behavioral goal is an open question. To gain insight into this process, we characterized the spatiotemporal organization of the muscle patterns for fast-reaching movements. We recorded electromyographic activity from up to 19 shoulder and arm muscles during point-to-point movements between a central location and 8 peripheral targets in each of 2 vertical planes. We used an optimization algorithm to identify a set of time-varying muscle synergies, i.e., the coordinated activations of groups of muscles with specific time-varying profiles. For each one of nine subjects, we extracted four or five synergies whose combinations, after scaling in amplitude and shifting in time each synergy independently for each movement condition, explained 73-82% of the data variation. We then tested whether these synergies could reconstruct the muscle patterns for point-to-point movements with different loads or forearm postures and for reversal and via-point movements. We found that reconstruction accuracy remained high, indicating generalization across these conditions. Finally, the synergy amplitude coefficients were directionally tuned according to a cosine function with a preferred direction that showed a smaller variability with changes of load, posture, and endpoint than the preferred direction of individual muscles. Thus the complex spatiotemporal characteristics of the muscles patterns for reaching were captured by the combinations of a small number of components, suggesting that the mechanisms involved in the generation of the muscle patterns exploit this low dimensionality to simplify control.

The acquisition and implementation of the smoothness maximization motion strategy is dependent on spatial accuracy demands

We recently showed that extensive training on a sequence of planar hand trajectories passing through several targets resulted in the co-articulation of movement components and in the formation of new movement elements (primitives) (Sosnik et al. in Exp Brain Res 156(4):422-438, 2004). Reduction in movement duration was accompanied by the gradual replacing of a piecewise combination of rectilinear trajectories with a single, longer curved one, the latter affording the maximization of movement smoothness ("global motion planning"). The results from transfer experiments, conducted by the end of the last training session, have suggested that the participants have acquired movement elements whose attributes were solely dictated by the figural (i.e., geometrical) form of the path, rather than by both path geometry and its time derivatives. Here we show that the acquired movement generation strategy ("global motion planning") was not specific to the trained configuration or total movement duration. Performance gain (i.e., movement smoothness, defined by the fit of the data to the behavior, predicted by the "global planning" model) transferred to non-trained configurations in which the targets were spatially co-aligned or when participants were instructed to perform the task in a definite amount of time. Surprisingly, stringent accuracy demands, in transfer conditions, resulted not only in an increased movement duration but also in reverting to the straight trajectories (loss of co-articulation), implying that the performance gain was dependent on accuracy constraints. Only 28.5% of the participants (two out of seven) who were trained in the absence of visual feedback from the hand (dark condition) co-articulated by the end of the last training session compared to 75% (six out of eight) who were trained in the light, and none of them has acquired a geometrical motion primitive. Furthermore, six naive participants who trained in dark condition on large size targets have all co-articulated by the end of the last training session, still, none of them has acquired a geometrical motion primitive. Taken together, our results indicate that the acquisition of a geometrical motion primitive is dependent on the existence of visual feedback from the hand and that the implementation of the smoothness-maximization motion strategy is dependent on spatial accuracy demands. These findings imply that the specific features of the training experience (i.e., temporal or spatial task demands) determine the attributes of an acquired motion planning strategy and primitive.

Sequential activation of muscle synergies during locomotion in the intact cat as revealed by cluster analysis and direct decomposition

During goal-directed locomotion, descending signals from supraspinal structures act through spinal interneuron pathways to effect modifications of muscle activity that are appropriate to the task requirements. Recent studies using decomposition methods suggest that this control might be facilitated by activating synergies organized at the level of the spinal cord. However, it is difficult to directly relate these mathematically defined synergies to the patterns of electromyographic activity observed in the original recordings. To address this issue, we have used a novel cluster analysis to make a detailed study of the organization of the synergistic patterns of muscle activity observed in the fore- and hindlimb during treadmill locomotion. The results show that the activity of a large number of forelimb muscles (26 bursts of activity from 18 muscles) can be grouped into 11 clusters on the basis of synchronous co-activation. Nine (9/11) of these clusters defined muscle activity during the swing phase of locomotion; these clusters were distributed in a sequential manner and were related to discrete behavioral events. A comparison with the synergies identified by linear decomposition methods showed some striking similarities between the synergies identified by the different methods. In the hindlimb, a simpler organization was observed, and a sequential activation of muscles similar to that observed in the forelimb during swing was less clear. We suggest that this organization of synergistic muscles provides a means by which descending signals could provide the detailed control of different muscle groups that is necessary for the flexible control of multi-articular movements.

Muscle synergy organization is robust across a variety of postural perturbations

We recently showed that four muscle synergies can reproduce multiple muscle activation patterns in cats during postural responses to support surface translations. We now test the robustness of functional muscle synergies, which specify muscle groupings and the active force vectors produced during postural responses under several biomechanically distinct conditions. We aimed to determine whether such synergies represent a generalized control strategy for postural control or if they are merely specific to each postural task. Postural responses to multidirectional translations at different fore-hind paw distances and to multidirectional rotations at the preferred stance distance were analyzed. Five synergies were required to adequately reconstruct responses to translation at the preferred stance distance-four were similar to our previous analysis of translation, whereas the fifth accounted for the newly added background activity during quiet stance. These five control synergies could account for > 80% total variability or r2 > 0.6 of the electromyographic and force tuning curves for all other experimental conditions. Forces were successfully reconstructed but only when they were referenced to a coordinate system that rotated with the limb axis as stance distance changed. Finally, most of the functional muscle synergies were similar across all of the six cats in terms of muscle synergy number, synergy activation patterns, and synergy force vectors. The robustness of synergy organization across perturbation types, postures, and animals suggests that muscle synergies controlling task-variables are a general construct used by the CNS for balance control.

Anticipatory adjustments of multi-finger synergies in preparation for self-triggered perturbations

We studied changes in multi-finger synergies associated with predictable and unpredictable force perturbations applied to a finger during a multi-finger constant total force production task. The main hypothesis was that indices of multi-finger synergies can show anticipatory changes in preparation for a predictable perturbation. Subjects sat in a chair and pressed on force sensors with the four fingers of the right hand. The task was to produce a constant level of total force. The fingers acted against loads that produced upward directed forces. The loads (applied either to the index or to the ring finger) could be disengaged either by the subject or by the experimenter. An index of finger co-variation, DeltaV was computed across sets of 12 trials at each time sample and for all tasks separately. During steady-state force production, all subjects showed positive DeltaV values corresponding to strong negative covariation among finger forces interpreted as a force-stabilizing synergy. Prior to self-triggered unloading, subjects showed an anticipatory drop in DeltaV that started 100-125 ms prior to the unloading time. Such early changes were absent in trials with experimenter-triggered unloading. After an unloading, subjects changed forces of both perturbed and unperturbed fingers and reached a new sharing pattern of the total force. In experimenter-triggered conditions, changes in the forces of unperturbed fingers could be seen as early as 120 ms following an unloading. The index DeltaV dropped following a perturbation and then recovered; the recovery occurred faster in self-triggered conditions. We conclude that humans can use feed-forward changes in multi-finger synergies (anticipatory synergy adjustments) in anticipation of a predictable perturbation. These changes may help avoid prolonged weakening of a multi-digit force-stabilizing synergy. We discuss a possibility that anticipatory postural adjustments may represent a particular case of the phenomenon of anticipatory synergy adjustments and suggest a hierarchical control scheme that incorporates a possibility of independent control over the output of a multi-element system and covariation patterns among outputs of its elements.

Muscle synergies during voluntary body sway: combining across-trials and within-a-trial analyses

We investigated co-varied changes in muscle activity during voluntary sway tasks that required a quick shift of the center of pressure (COP). We hypothesized that multi-muscle synergies (defined as task-specific covariation of elemental variables, muscle modes) stabilize a COP location in the anterior-posterior direction prior to a voluntary COP shift and that during the shift the synergies would weaken. Standing subjects performed two tasks, a cyclic COP shift over a range corresponding to 80% of the maximal amplitude of voluntary COP shift at 1 Hz and a unidirectional quick COP shift over the same nominal amplitude. The cyclic sway task was used to define muscle modes (M-modes, leg and trunk muscle groups with parallel scaling of muscle activation level within a group) and the relations between small changes in the magnitudes of M-modes [in the principal component analysis (PCA), the M-mode magnitudes are equivalent to PC scores] and COP shifts. A novel approach was used involving PCA applied to indices of muscle integrated activity measured both within a trial and across trials. The unidirectional sway task was performed in a self-paced (SP) manner and under a typical simple reaction time (RT) instruction. M-modes were also defined along trials at those tasks; they have been shown to be similar across tasks. Integrated indices of muscle activity in the SP-sway and RT-sway tasks were transformed into the M-modes. Variance in the M-mode space was partitioned into two components, one that did not affect the average value of COP shift (V (UCM)) and the other that did (V (ORT)). An index (DeltaV) corresponding to the normalized difference between V (UCM) and V (ORT) was computed. During steady-state posture, DeltaV was positive corresponding to most M-mode variance lying in a sub-space corresponding to a stable COP location across trials. Positive DeltaV values have been interpreted as reflecting a multi-M-mode synergy stabilizing the COP location. The magnitude of DeltaV was larger in SP trials than in RT trials. During voluntary COP shifts, the DeltaV magnitude dropped to zero or even became negative. We conclude that M-mode synergies stabilize COP location during quiet standing, while these synergies weaken or disappear during fast voluntary COP shifts. Under RT conditions, the COP stabilizing synergies were weaker supposedly to facilitate a quick COP shift without time for preparation. The suggested method of M-mode identification may potentially be applied to analysis of postural synergies in persons with impaired postural control such as elderly persons, persons with atypical development, or in the course of rehabilitation after an injury.

Muscle synergies during voluntary body sway: combining across-trials and within-a-trial analyses

We investigated co-varied changes in muscle activity during voluntary sway tasks that required a quick shift of the center of pressure (COP). We hypothesized that multi-muscle synergies (defined as task-specific covariation of elemental variables, muscle modes) stabilize a COP location in the anterior-posterior direction prior to a voluntary COP shift and that during the shift the synergies would weaken. Standing subjects performed two tasks, a cyclic COP shift over a range corresponding to 80% of the maximal amplitude of voluntary COP shift at 1 Hz and a unidirectional quick COP shift over the same nominal amplitude. The cyclic sway task was used to define muscle modes (M-modes, leg and trunk muscle groups with parallel scaling of muscle activation level within a group) and the relations between small changes in the magnitudes of M-modes [in the principal component analysis (PCA), the M-mode magnitudes are equivalent to PC scores] and COP shifts. A novel approach was used involving PCA applied to indices of muscle integrated activity measured both within a trial and across trials. The unidirectional sway task was performed in a self-paced (SP) manner and under a typical simple reaction time (RT) instruction. M-modes were also defined along trials at those tasks; they have been shown to be similar across tasks. Integrated indices of muscle activity in the SP-sway and RT-sway tasks were transformed into the M-modes. Variance in the M-mode space was partitioned into two components, one that did not affect the average value of COP shift (V (UCM)) and the other that did (V (ORT)). An index (DeltaV) corresponding to the normalized difference between V (UCM) and V (ORT) was computed. During steady-state posture, DeltaV was positive corresponding to most M-mode variance lying in a sub-space corresponding to a stable COP location across trials. Positive DeltaV values have been interpreted as reflecting a multi-M-mode synergy stabilizing the COP location. The magnitude of DeltaV was larger in SP trials than in RT trials. During voluntary COP shifts, the DeltaV magnitude dropped to zero or even became negative. We conclude that M-mode synergies stabilize COP location during quiet standing, while these synergies weaken or disappear during fast voluntary COP shifts. Under RT conditions, the COP stabilizing synergies were weaker supposedly to facilitate a quick COP shift without time for preparation. The suggested method of M-mode identification may potentially be applied to analysis of postural synergies in persons with impaired postural control such as elderly persons, persons with atypical development, or in the course of rehabilitation after an injury.

Matrix Factorization Algorithms for the Identification of Muscle Synergies: Evaluation on Simulated and Experimental Data Sets

Several recent studies have used matrix factorization algorithms to assess the hypothesis that behaviors might be produced through the combination of a small number of muscle synergies. Although generally agreeing in their basic conclusions, these studies have used a range of different algorithms, making their interpretation and integration difficult. We therefore compared the performance of these different algorithms on both simulated and experimental data sets. We focused on the ability of these algorithms to identify the set of synergies underlying a data set. All data sets consisted of nonnegative values, reflecting the nonnegative data of muscle activation patterns. We found that the performance of principal component analysis (PCA) was generally lower than that of the other algorithms in identifying muscle synergies. Factor analysis (FA) with varimax rotation was better than PCA, and was generally at the same levels as independent component analysis (ICA) and nonnegative matrix factorization (NMF). ICA performed very well on data sets corrupted by constant variance Gaussian noise, but was impaired on data sets with signal-dependent noise and when synergy activation coefficients were correlated. Nonnegative matrix factorization (NMF) performed similarly to ICA and FA on data sets with signal-dependent noise and was generally robust across data sets. The best algorithms were ICA applied to the subspace defined by PCA (ICAPCA) and a version of probabilistic ICA with nonnegativity constraints (pICA). We also evaluated some commonly used criteria to identify the number of synergies underlying a data set, finding that only likelihood ratios based on factor analysis identified the correct number of synergies for data sets with signal-dependent noise in some cases. We then proposed an ad hoc procedure, finding that it was able to identify the correct number in a larger number of cases. Finally, we applied these methods to an experimentally obtained data set. The best performing algorithms (FA, ICA, NMF, ICAPCA, pICA) identified synergies very similar to one another. Based on these results, we discuss guidelines for using factorization algorithms to analyze muscle activation patterns. More generally, the ability of several algorithms to identify the correct muscle synergies and activation coefficients in simulated data, combined with their consistency when applied to physiological data sets, suggests that the muscle synergies found by a particular algorithm are not an artifact of that algorithm, but reflect basic aspects of the organization of muscle activation patterns underlying behaviors.

Central and sensory contributions to the activation and organization of muscle synergies during natural motor behaviors

Previous studies have suggested that the motor system may simplify control by combining a small number of muscle synergies represented as activation profiles across a set of muscles. The role of sensory feedback in the activation and organization of synergies has remained an open question. Here, we assess to what extent the motor system relies on centrally organized synergies activated by spinal and/or supraspinal commands to generate motor outputs by analyzing electromyographic (EMG) signals collected from 13 hindlimb muscles of the bullfrog during swimming and jumping, before and after deafferentation. We first established that, for both behaviors, the intact and deafferented data sets possess low and similar dimensionalities. Subsequently, we used a novel reformulation of the non-negative matrix factorization algorithm to simultaneously search for synergies shared by, and synergies specific to, the intact and deafferented data sets. Most muscle synergies were identified as shared synergies, suggesting that EMGs of locomotor behaviors are generated primarily by centrally organized synergies. Both the amplitude and temporal patterns of the activation coefficients of most shared synergies, however, were altered by deafferentation, suggesting that sensory inflow modulates activation of those centrally organized synergies. For most synergies, effects of deafferentation on the activation coefficients were not consistent across frogs, indicating substantial interanimal variability of feedback actions. We speculate that sensory feedback might adapt recruitment of muscle synergies to behavioral constraints, and the few synergies specific to the intact or deafferented states might represent afferent-specific modules or feedback reorganization of spinal neuronal networks.

Localization and connectivity in spinal interneuronal networks: the adduction-caudal extension-flexion rhythm in the frog

We have previously reported that focal intraspinal N-methyl-d-aspartate (NMDA) iontophoresis in the frog elicits a motor output, which is organized in terms of its constituent isometric force directions at the ipsilateral ankle and its topography. Furthermore, the associated EMG patterns can be reconstructed as the linear combinations of seven muscle synergies, labeled A to G. We now focus on one of the most common NMDA-elicited outputs, the adduction-caudal extension-flexion rhythm, and examine the relationship between the different force phases in terms of synergies and topography. Two distinct EMG patterns produce caudal extensions, and only one of the two patterns is used at most sites. The key synergy combinations for the two patterns are B + e and D + c (strongest synergies capitalized). These two patterns map at distinct locations in the lumbar cord. Within individual sites rhythms, we find linkages among the synergies used to produce adductions, the onsets of flexions after caudal extensions, and the synergy pattern producing the caudal extensions. For example, the synergy composition of adductions at B + e caudal extension sites is dominated by E + b and at D + c caudal extension sites by C + d. The two types of adductions map at distinct locations, situated between the two caudal extension regions. Specifically the linked patterns of caudal extension-adduction interleave rostrocaudally in a CE2-ADD1-ADD2-CE1 sequence, where 1 and 2 refer to the two pattern types. The implications of this topography and connectivity with respect to motor systems organization and behaviors are discussed.

A central back-coupling hypothesis on the organization of motor synergies: a physical metaphor and a neural model

We offer a hypothesis on the organization of multi-effector motor synergies and illustrate it with the task of force production with a set of fingers. A physical metaphor, a leaking bucket, is analyzed to demonstrate that an inanimate structure can show apparent error compensation among its elements. A neural model is developed using tunable back-coupling loops as means of assuring error compensation in a task-specific way. The model demonstrates non-trivial features of multi-finger interaction such as delayed emergence of force stabilizing synergies and simultaneous stabilization of the total force and total moment produced by the fingers. The hypothesis suggests that neurophysiological structures involving short-latency feedback may play a central role in the formation of motor synergies.

Shared and specific muscle synergies in natural motor behaviors

Selecting the appropriate muscle pattern to achieve a given goal is an extremely complex task because of the dimensionality of the search space and because of the nonlinear and dynamical nature of the transformation between muscle activity and movement. To investigate whether the central nervous system uses a modular architecture to achieve motor coordination we characterized the motor output over a large set of movements. We recorded electromyographic activity from 13 muscles of the hind limb of intact and freely moving frogs during jumping, swimming, and walking in naturalistic conditions. We used multidimensional factorization techniques to extract invariant amplitude and timing relationships among the muscle activations. A decomposition of the instantaneous muscle activations as combinations of nonnegative vectors, synchronous muscle synergies, revealed a spatial organization in the muscle patterns. A decomposition of the same activations as a combination of temporal sequences of nonnegative vectors, time-varying muscle synergies, further uncovered specific characteristics in the muscle activation waveforms. A mixture of synergies shared across behaviors and synergies for specific behaviors captured the invariances across the entire dataset. These results support the hypothesis that the motor controller has a modular organization.

A limited set of muscle synergies for force control during a postural task

Recently developed computational techniques have been used to reduce muscle activation patterns of high complexity to a simple synergy organization and to bring new insights to the long-standing degrees of freedom problem in motor control. We used a nonnegative factorization approach to identify muscle synergies during postural responses in the cat and to examine the functional significance of such synergies for natural behaviors. We hypothesized that the simplification of neural control afforded by muscle synergies must be matched by a similar reduction in degrees of freedom at the biomechanical level. Electromyographic data were recorded from 8-15 hindlimb muscles of cats exposed to 16 directions of support surface translation. Results showed that as few as four synergies could account for >95% of the automatic postural response across all muscles and all directions. Each synergy was activated for a specific set of perturbation directions, and moreover, each was correlated with a unique vector of endpoint force under the limb. We suggest that, within the context of active balance control, postural synergies reflect a neural command signal that specifies endpoint force of a limb.

Modular premotor drives and unit bursts as primitives for frog motor behaviors

Spinal cord modularity impacts on our understanding of reflexes, development, descending systems in normal motor control, and recovery from injury. We used independent component analysis and best-basis or matching pursuit wavepacket analysis to extract the composition and temporal structure of bursts in hindlimb muscles of frogs. These techniques make minimal a priori assumptions about drive and motor pattern structure. We compared premotor drive and burst structures in spinal frogs with less reduced frogs with a fuller repertoire of locomotory, kicking, and scratching behaviors. Six multimuscle drives explain most of the variance of motor patterns (approximately 80%). Each extracted drive was activated with pulses at a single time scale or common duration (approximately 275 msec) burst structure. The data show that complex behaviors in brainstem frogs arise as a result of focusing drives to smaller core groups of muscles. Brainstem drives were subsets of the muscle groups from spinal frogs. The 275 msec burst duration was preserved across all behaviors and was most precise in brainstem frogs. These data support a modular decomposition of frog behaviors into a small collection of unit burst generators and associated muscle drives in spinal cord. Our data also show that the modular organization of drives seen in isolated spinal cord is fine-tuned by descending controls to enable a fuller movement repertoire. The unit burst generators and their associated muscle synergies extracted here link the biomechanical "primitives," described earlier in the frog, rat, and cat, and to the elements of pattern generation examined in fictive preparations.

Biological arm motion through reinforcement learning

The present paper discusses an optimal learning control method using reinforcement learning for biological systems with a redundant actuator. It is difficult to apply reinforcement learning to biological control systems because of the redundancy in muscle activation space. We solve this problem with the following method. First, we divide the control input space into two subspaces according to a priority order of learning and restrict the search noise for reinforcement learning to the first priority subspace. Then the constraint is reduced as the learning progresses, with the search space extending to the second priority subspace. The higher priority subspace is designed so that the impedance of the arm can be high. A smooth reaching motion is obtained through reinforcement learning without any previous knowledge of the arm's dynamics.

Mechanical aspects of legged locomotion control

We review the mechanical components of an approach to motion science that enlists recent progress in neurophysiology, biomechanics, control systems engineering, and non-linear dynamical systems to explore the integration of muscular, skeletal, and neural mechanics that creates effective locomotor behavior. We use rapid arthropod terrestrial locomotion as the model system because of the wealth of experimental data available. With this foundation, we list a set of hypotheses for the control of movement, outline their mathematical underpinning and show how they have inspired the design of the hexapedal robot, RHex.

Distinct temporal activity patterns in the rat M1 and red nucleus during skilled versus unskilled limb movement

The mammalian motor system contains multiple interconnected supraspinal networks, but little is known about their relative roles in producing different movements and behaviors, particularly given their apparently fused activity in primates. We tested whether the task context, as well as using a phylogenetically older mammal, rats, could distinguish the separate contributions of these networks. We obtained simultaneous multi-single neuron recordings from the forelimb motor cortex and magnocellular red nucleus as rats performed two contextually different, but kinematically similar, forelimb reach-like tasks: highly learned, skilled reaching for food through a narrow slot, a task requiring extensive training, versus the swing phases of treadmill locomotion. In both the M1 and the mRN, large subpopulations of neurons peaked in their spike firing rates near the onset and the end of the swing phase during treadmill locomotion. In contrast, neural subgroups in the two areas displayed different temporal sequences of activity during the skilled reaching task. In the mRN, the majority of task-modulated neurons peaked in their firing rate in the middle of the reach when the rat was preparing to project the arm through the slot, whereas large subgroups of M1 neurons displayed elevated firing rates during the initial and terminal phases of the reach. These results suggest that motor-behavioral context can alter the degree of overlapping activity in different supraspinal sensorimotor networks. Moreover, results for the skilled reaching task in rats may have highlighted a distinct processing role of the rubral complex: adapting natural muscle synergies across joints and limbs to novel task demands, in concert with cortically based learning.

Modelling the control of interceptive actions

In recent years, several phenomenological dynamical models have been formulated that describe how perceptual variables are incorporated in the control of motor variables. We call these short-route models as they do not address how perception-action patterns might be constrained by the dynamical properties of the sensory, neural and musculoskeletal subsystems of the human action system. As an alternative, we advocate a long-route modelling approach in which the dynamics of these subsystems are explicitly addressed and integrated to reproduce interceptive actions. The approach is exemplified through a discussion of a recently developed model for interceptive actions consisting of a neural network architecture for the online generation of motor outflow commands, based on time-to-contact information and information about the relative positions and velocities of hand and ball. This network is shown to be consistent with both behavioural and neurophysiological data. Finally, some problems are discussed with regard to the question of how the motor outflow commands (i.e. the intended movement) might be modulated in view of the musculoskeletal dynamics.

Quantifying generalization from trial-by-trial behavior of adaptive systems that learn with basis functions: theory and experiments in human motor control

During reaching movements, the brain's internal models map desired limb motion into predicted forces. When the forces in the task change, these models adapt. Adaptation is guided by generalization: errors in one movement influence prediction in other types of movement. If the mapping is accomplished with population coding, combining basis elements that encode different regions of movement space, then generalization can reveal the encoding of the basis elements. We present a theory that relates encoding to generalization using trial-by-trial changes in behavior during adaptation. We consider adaptation during reaching movements in various velocity-dependent force fields and quantify how errors generalize across direction. We find that the measurement of error is critical to the theory. A typical assumption in motor control is that error is the difference between a current trajectory and a desired trajectory (DJ) that does not change during adaptation. Under this assumption, in all force fields that we examined, including one in which force randomly changes from trial to trial, we found a bimodal generalization pattern, perhaps reflecting basis elements that encode direction bimodally. If the DJ was allowed to vary, bimodality was reduced or eliminated, but the generalization function accounted for nearly twice as much variance. We suggest, therefore, that basis elements representing the internal model of dynamics are sensitive to limb velocity with bimodal tuning; however, it is also possible that during adaptation the error metric itself adapts, which affects the implied shape of the basis elements.

Differential spatial organization of otolith signals in frog vestibular nuclei

Activation maps of pre- and postsynaptic field potential components evoked by separate electrical stimulation of utricular, lagenar, and saccular nerve branches in the isolated frog hindbrain were recorded within a stereotactic outline of the vestibular nuclei. Utricular and lagenar nerve-evoked activation maps overlapped strongly in the lateral and descending vestibular nuclei, whereas lagenar amplitudes were greater in the superior vestibular nucleus. In contrast, the saccular nerve-evoked activation map coincided largely with the dorsal nucleus and the adjacent dorsal part of the lateral vestibular nucleus, corroborating a major auditory and lesser vestibular function of the frog saccule. The stereotactic position of individual second-order otolith neurons matched the distribution of the corresponding otolith nerve-evoked activation maps. Furthermore, particular types of second-order utricular and lagenar neurons were clustered with particular types of second-order canal neurons in a topology that anatomically mirrored the preferred convergence pattern of afferent otolith and canal signals in second-order vestibular neurons. Similarities in the spatial organization of functionally equivalent types of second-order otolith and canal neurons between frog and other vertebrates indicated conservation of a common topographical organization principle. However, the absence of a precise afferent sensory topography combined with the presence of spatially segregated groups of particular second-order vestibular neurons suggests that the vestibular circuitry is organized as a premotor map rather than an organotypical sensory map. Moreover, the conserved segmental location of individual vestibular neuronal phenotypes shows linkage of individual components of vestibulomotor pathways with the underlying genetically specified rhombomeric framework.

Tongue-surface movement patterns during speech and swallowing

The tongue has been frequently characterized as being composed of several functionally independent articulators. The question of functional regionality within the tongue was examined by quantifying the strength of coupling among four different tongue locations across a large number of consonantal contexts and participants. Tongue behavior during swallowing was also described. Vertical displacements of pellets affixed to the tongue were extracted from the x-ray microbeam database. Forty-six participants recited 20 vowel-consonant-vowel (VCV) combinations and swallowed 10 ccs of water. Tongue-surface movement patterns were quantitatively described by computing the covariance between the vertical time-histories of all possible pellet pairs. Phonemic differentiation in vertical tongue motions was observed as coupling varied predictably across pellet pairs with place of articulation. Moreover, tongue displacements for speech and swallowing clustered into distinct groups based on their coupling profiles. Functional independence of anterior tongue regions was evidenced by a wide range of movement coupling relations between anterior tongue pellets. The strengths and weaknesses of the covariance-based analysis for characterizing tongue movement are considered.