Perturbation-evoked responses in primary motor cortex are modulated by behavioral context

Rapid Online Corrections for Proprioceptive and Visual Perturbations Recruit Similar Circuits in Primary Motor Cortex

An important aspect of motor function is our ability to rapidly generate goal-directed corrections for disturbances to the limb or behavioral goal. The primary motor cortex (M1) is a key region involved in processing feedback for rapid motor corrections, yet we know little about how M1 circuits are recruited by different sources of sensory feedback to make rapid corrections. We trained two male monkeys (Macaca mulatta) to make goal-directed reaches and on random trials introduced different sensory errors by either jumping the visual location of the goal (goal jump), jumping the visual location of the hand (cursor jump), or applying a mechanical load to displace the hand (proprioceptive feedback). Sensory perturbations evoked a broad response in M1 with ∼73% of neurons (n = 257) responding to at least one of the sensory perturbations. Feedback responses were also similar as response ranges between the goal and cursor jumps were highly correlated (range of r = [0.91, 0.97]) as were the response ranges between the mechanical loads and the visual perturbations (range of r = [0.68, 0.86]). Lastly, we identified the neural subspace each perturbation response resided in and found a strong overlap between the two visual perturbations (range of overlap index, 0.73-0.89) and between the mechanical loads and visual perturbations (range of overlap index, 0.36-0.47) indicating each perturbation evoked similar structure of activity at the population level. Collectively, our results indicate rapid responses to errors from different sensory sources target similar overlapping circuits in M1.

Proprioceptive and Visual Feedback Responses in Macaques Exploit Goal Redundancy

A common problem in motor control concerns how to generate patterns of muscle activity when there are redundant solutions to attain a behavioral goal. Optimal feedback control is a theory that has guided many behavioral studies exploring how the motor system incorporates task redundancy. This theory predicts that kinematic errors that deviate the limb should not be corrected if one can still attain the behavioral goal. Studies in humans demonstrate that the motor system can flexibly integrate visual and proprioceptive feedback of the limb with goal redundancy within 90 ms and 70 ms, respectively. Here, we show monkeys (Macaca mulatta) demonstrate similar abilities to exploit goal redundancy. We trained four male monkeys to reach for a goal that was either a narrow square or a wide, spatially redundant rectangle. Monkeys exhibited greater trial-by-trial variability when reaching to the wide goal consistent with exploiting goal redundancy. On random trials we jumped the visual feedback of the hand and found monkeys corrected for the jump when reaching to the narrow goal and largely ignored the jump when reaching for the wide goal. In a separate set of experiments, we applied mechanical loads to the arm of the monkey and found similar corrective responses based on goal shape. Muscle activity reflecting these different corrective responses were detected for the visual and mechanical perturbations starting at ∼90 and ∼70 ms, respectively. Thus, rapid motor responses in macaques can exploit goal redundancy similar to humans, creating a paradigm to study the neural basis of goal-directed motor action and motor redundancy.SIGNIFICANCE STATEMENT Moving in the world requires selecting from an infinite set of possible motor commands. Theories predict that motor commands are selected that exploit redundancies. Corrective responses in humans to either visual or proprioceptive disturbances of the limb can rapidly exploit redundant trajectories to a goal in <100 ms after a disturbance. However, uncovering the neural correlates generating these rapid motor corrections has been hampered by the absence of an animal model. We developed a behavioral paradigm in monkeys that incorporates redundancy in the form of the shape of the goal. Critically, monkeys exhibit corrective responses and timings similar to humans performing the same task. Our paradigm provides a model for investigating the neural correlates of sophisticated rapid motor corrections.

When intercepting moving targets, mid-movement error corrections reflect distinct responses to visual and haptic perturbations

We examined a key aspect of sensorimotor skill: the capability to correct performance errors that arise mid-movement. Participants grasped the handle of a robot that imposed a nominal viscous resistance to hand movement. They watched a target move pseudo-randomly just above the horizontal plane of hand motion and initiated quick interception movements when cued. On some trials, the robot's viscosity or the target's speed changed without warning coincident with the GO cue. We fit a sum-of-Gaussians model to mechanical power measured at the handle to determine the number, magnitude, and relative timing of submovements occurring in each interception attempt. When a single submovement successfully intercepted the target, capture times averaged 410 ms. Sometimes, two or more submovements were required. Initial error corrections typically occurred before feedback could indicate the target had been captured or missed. Error corrections occurred sooner after movement onset in response to mechanical viscosity increases (at 154 ms) than to unprovoked errors on control trials (215 ms). Corrections occurred later (272 ms) in response to viscosity decreases. The latency of corrections for target speed changes did not differ from those in control trials. Remarkably, these early error corrections accommodated the altered testing conditions; speed/viscosity increases elicited more vigorous corrections than in control trials with unprovoked errors; speed/viscosity decreases elicited less vigorous corrections. These results suggest that the brain monitors and predicts the outcome of evolving movements, rapidly infers causes of mid-movement errors, and plans and executes corrections-all within 300 ms of movement onset.

Rotational dynamics in motor cortex are consistent with a feedback controller

Recent studies have identified rotational dynamics in motor cortex (MC), which many assume arise from intrinsic connections in MC. However, behavioral and neurophysiological studies suggest that MC behaves like a feedback controller where continuous sensory feedback and interactions with other brain areas contribute substantially to MC processing. We investigated these apparently conflicting theories by building recurrent neural networks that controlled a model arm and received sensory feedback from the limb. Networks were trained to counteract perturbations to the limb and to reach toward spatial targets. Network activities and sensory feedback signals to the network exhibited rotational structure even when the recurrent connections were removed. Furthermore, neural recordings in monkeys performing similar tasks also exhibited rotational structure not only in MC but also in somatosensory cortex. Our results argue that rotational structure may also reflect dynamics throughout the voluntary motor system involved in online control of motor actions.

Unravelling the enigma of cortical tremor and other forms of cortical myoclonus

Cortical tremor is a fine rhythmic oscillation involving distal upper limbs, linked to increased sensorimotor cortex excitability, as seen in cortical myoclonus. Cortical tremor is the hallmark feature of autosomal dominant familial cortical myoclonic tremor and epilepsy (FCMTE), a syndrome not yet officially recognized and characterized by clinical and genetic heterogeneity. Non-coding repeat expansions in different genes have been recently recognized to play an essential role in its pathogenesis. Cortical tremor is considered a rhythmic variant of cortical myoclonus and is part of the 'spectrum of cortical myoclonus', i.e. a wide range of clinical motor phenomena, from reflex myoclonus to myoclonic epilepsy, caused by abnormal sensorimotor cortical discharges. The aim of this update is to provide a detailed analysis of the mechanisms defining cortical tremor, as seen in FCMTE. After reviewing the clinical and genetic features of FCMTE, we discuss the possible mechanisms generating the distinct elements of the cortical myoclonus spectrum, and how cortical tremor fits into it. We propose that the spectrum is due to the evolution from a spatially limited focus of excitability to recruitment of more complex mechanisms capable of sustaining repetitive activity, overcoming inhibitory mechanisms that restrict excitatory bursts, and engaging wide areas of cortex. Finally, we provide evidence for a possible common denominator of the elements of the spectrum, i.e. the cerebellum, and discuss its role in FCMTE, according to recent genetic findings.

Transient deactivation of dorsal premotor cortex or parietal area 5 impairs feedback control of the limb in macaques

We can generate goal-directed motor corrections with surprising speed, but their neural basis is poorly understood. Here, we show that temporary cooling of dorsal premotor cortex (PMd) impaired both spatial accuracy and the speed of corrective responses, whereas cooling parietal area 5 (A5) impaired only spatial accuracy. Simulations based on optimal feedback control (OFC) models demonstrated that "deactivation" of the control policy (reduction in feedback gain) and state estimation (reduction in Kalman gain) caused impairments similar to that observed for PMd and A5 cooling, respectively. Furthermore, combined deactivation of both cortical regions led to additive impairments of individual deactivations, whereas reducing the amount of cooling to PMd led to impairments in response speed but not spatial accuracy, both also predicted by OFC models. These results provide causal support that frontoparietal circuits beyond primary somatosensory and motor cortices are involved in generating goal-directed motor corrections.

Shared internal models for feedforward and feedback control of arm dynamics in non-human primates

Previous work has shown that humans account for and learn novel properties or the arm's dynamics, and that such learning causes changes in both the predictive (i.e., feedforward) control of reaching and reflex (i.e., feedback) responses to mechanical perturbations. Here we show that similar observations hold in old-world monkeys (Macaca fascicularis). Two monkeys were trained to use an exoskeleton to perform a single-joint elbow reaching and to respond to mechanical perturbations that created pure elbow motion. Both of these tasks engaged robust shoulder muscle activity as required to account for the torques that typically arise at the shoulder when the forearm rotates around the elbow joint (i.e., intersegmental dynamics). We altered these intersegmental arm dynamics by having the monkeys generate the same elbow movements with the shoulder joint either free to rotate, as normal, or fixed by the robotic manipulandum, which eliminates the shoulder torques caused by forearm rotation. After fixing the shoulder joint, we found a systematic reduction in shoulder muscle activity. In addition, after releasing the shoulder joint again, we found evidence of kinematic aftereffects (i.e., reach errors) in the direction predicted if failing to compensate for normal arm dynamics. We also tested whether such learning transfers to feedback responses evoked by mechanical perturbations and found a reduction in shoulder feedback responses, as appropriate for these altered arm intersegmental dynamics. Demonstrating this learning and transfer in non-human primates will allow the investigation of the neural mechanisms involved in feedforward and feedback control of the arm's dynamics.

Learning New Feedforward Motor Commands Based on Feedback Responses

Learning a new motor task modifies feedforward (i.e., voluntary) motor commands and such learning also changes the sensitivity of feedback responses (i.e., reflexes) to mechanical perturbations [1-9]. For example, after people learn to generate straight reaching movements in the presence of an external force field or learn to reduce shoulder muscle activity when generating pure elbow movements with shoulder fixation, evoked stretch reflex responses to mechanical perturbations reflect the learning expressed during self-initiated reaching. Such a transfer from feedforward motor commands to feedback responses is thought to take place because of shared neural circuits at the level of the spinal cord, brainstem, and cerebral cortex [10-13]. The presence of shared neural resources also predicts the transfer from feedback responses to feedforward motor commands. Little is known about such a transfer presumably because it is relatively hard to elicit learning in reflexes without engaging associated voluntary responses following mechanical perturbations. Here, we demonstrate such transfer by leveraging two approaches to elicit stretch reflexes while minimizing engagement of voluntary motor responses in the learning process: applying very short mechanical perturbations [14-19] and instructing participants to not respond to them [20-26]. Taken together, our work shows that transfer between feedforward and feedback control is bidirectional, furthering the notion that these processes share common neural circuits that underlie motor learning and transfer.

Long-latency Responses to a Mechanical Perturbation of the Index Finger Have a Spinal Component

In an uncertain external environment, the motor system may need to respond rapidly to an unexpected stimulus. Limb displacement causes muscle stretch; the corrective response has multiple activity bursts, which are suggested to originate from different parts of the neuraxis. The earliest response is so fast, it can only be produced by spinal circuits; this is followed by slower components thought to arise from primary motor cortex (M1) and other supraspinal areas.

In an uncertain external environment, the motor system may need to respond rapidly to an unexpected stimulus. Limb displacement causes muscle stretch; the corrective response has multiple activity bursts, which are suggested to originate from different parts of the neuraxis. The earliest response is so fast, it can only be produced by spinal circuits; this is followed by slower components thought to arise from primary motor cortex (M1) and other supraspinal areas. Spinal cord (SC) contributions to the slower components are rarely considered. To address this, we recorded neural activity in M1 and the cervical SC during a visuomotor tracking task, in which 2 female macaque monkeys moved their index finger against a resisting motor to track an on-screen target. Following the behavioral trial, an increase in motor torque rapidly returned the finger to its starting position (lever velocity >200°/s). Many cells responded to this passive mechanical perturbation (M1: 148 of 211 cells, 70%; SC: 67 of 119 cells, 56%). The neural onset latency was faster for SC compared with M1 cells (21.7 ± 11.2 ms vs 25.5 ± 10.7 ms, respectively, mean ± SD). Using spike-triggered averaging, some cells in both regions were identified as likely premotor cells, with monosynaptic connections to motoneurons. Response latencies for these cells were compatible with a contribution to the muscle responses following the perturbation. Comparable fractions of responding neurons in both areas were active up to 100 ms after the perturbation, suggesting that both SC circuits and supraspinal centers could contribute to later response components.

SIGNIFICANCE STATEMENT Following a limb perturbation, multiple reflexes help to restore limb position. Given conduction delays, the earliest part of these reflexes can only arise from spinal circuits. By contrast, long-latency reflex components are typically assumed to originate from supraspinal centers. We recorded from both spinal and motor cortical cells in monkeys responding to index finger perturbations. Many spinal interneurons, including those identified as projecting to motoneurons, responded to the perturbation; the timing of responses was compatible with a contribution to both short- and long-latency reflexes. We conclude that spinal circuits also contribute to long-latency reflexes in distal and forearm muscles, alongside supraspinal regions, such as the motor cortex and brainstem.

Generalizing movement patterns following shoulder fixation

Generalizing newly learned movement patterns beyond the training context is challenging for most motor learning situations. Here we tested whether learning of a new physical property of the arm during self-initiated reaching generalizes to new arm configurations. Human participants performed a single-joint elbow reaching task and/or countered mechanical perturbations that created pure elbow motion with the shoulder joint free to rotate or locked by the manipulandum. With the shoulder free, we found activation of shoulder extensor muscles for pure elbow extension trials, appropriate for countering torques that arise at the shoulder due to forearm rotation. After locking the shoulder joint, we found a partial reduction in shoulder muscle activity, appropriate because locking the shoulder joint cancels the torques that arise at the shoulder due to forearm rotation. In our first three experiments, we tested whether and to what extent this partial reduction in shoulder muscle activity generalizes when reaching in different situations: 1) different initial shoulder orientation, 2) different initial elbow orientation, and 3) different reach distance/speed. We found generalization for the different shoulder orientation and reach distance/speed as measured by a reliable reduction in shoulder activity in these situations but no generalization for the different elbow orientation. In our fourth experiment, we found that generalization is also transferred to feedback control by applying mechanical perturbations and observing reflex responses in a distinct shoulder orientation. These results indicate that partial learning of new intersegmental dynamics is not sufficient for modifying a general internal model of arm dynamics.NEW & NOTEWORTHY Here we show that partially learning to reduce shoulder muscle activity following shoulder fixation generalizes to other movement conditions, but it does not generalize globally. These findings suggest that the partial learning of new intersegmental dynamics is not sufficient for modifying a general internal model of the arm's dynamics.

Rapid crossed responses in an intrinsic hand muscle during perturbed bimanual movements

Mechanical perturbations in one upper limb often elicit corrective responses in both the perturbed as well as its contralateral and unperturbed counterpart. These crossed corrective responses have been shown to be sensitive to the bimanual requirements of the perturbation, but crossed responses (CRs) in hand muscles are far less well studied. Here, we investigate corrective CRs in an intrinsic hand muscle, the first dorsal interosseous (1DI), to clockwise and anticlockwise mechanical perturbations to the contralateral index finger while participants performed a bimanual finger abduction task. We found that the CRs in the unperturbed 1DI were sensitive to the direction of the perturbation of the contralateral index finger. However, the size of the CRs was not sensitive to the amplitude of the contralateral perturbation nor its context within the bimanual task. The onset latency of the CRs was too fast to be purely transcortical (<70 ms) in 12/12 participants. This confirms that during isolated bimanual finger movements, sensory feedback from one hand can influence the other, but the pathways mediating the earliest components of this interaction are likely to involve subcortical systems such as the brainstem or spinal cord, which may afford less flexibility to the task demands.NEW & NOTEWORTHY An intrinsic hand muscle shows a crossed response to a perturbation of the contralateral index finger. The crossed response is dependent on the direction of the contralateral perturbation but not on the amplitude or the bimanual requirements of the movement, suggesting a far less flexible control policy than those governing crossed responses in more proximal muscles. The crossed response is too fast to be purely mediated by transcortical pathways, suggesting subcortical contributions.

Considerations in using recurrent neural networks to probe neural dynamics

Recurrent neural networks (RNNs) are increasingly being used to model complex cognitive and motor tasks performed by behaving animals. RNNs are trained to reproduce animal behavior while also capturing key statistics of empirically recorded neural activity. In this manner, the RNN can be viewed as an in silico circuit whose computational elements share similar motifs with the cortical area it is modeling. Furthermore, because the RNN's governing equations and parameters are fully known, they can be analyzed to propose hypotheses for how neural populations compute. In this context, we present important considerations when using RNNs to model motor behavior in a delayed reach task. First, by varying the network's nonlinear activation and rate regularization, we show that RNNs reproducing single-neuron firing rate motifs may not adequately capture important population motifs. Second, we find that even when RNNs reproduce key neurophysiological features on both the single neuron and population levels, they can do so through distinctly different dynamical mechanisms. To distinguish between these mechanisms, we show that an RNN consistent with a previously proposed dynamical mechanism is more robust to input noise. Finally, we show that these dynamics are sufficient for the RNN to generalize to tasks it was not trained on. Together, these results emphasize important considerations when using RNN models to probe neural dynamics.NEW & NOTEWORTHY Artificial neurons in a recurrent neural network (RNN) may resemble empirical single-unit activity but not adequately capture important features on the neural population level. Dynamics of RNNs can be visualized in low-dimensional projections to provide insight into the RNN's dynamical mechanism. RNNs trained in different ways may reproduce neurophysiological motifs but do so with distinctly different mechanisms. RNNs trained to only perform a delayed reach task can generalize to perform tasks where the target is switched or the target location is changed.

Influence of kinesthetic motor imagery and effector specificity on the long-latency stretch response

The long-latency "reflexive" response (LLR) following an upper limb mechanical perturbation is generated by neural circuitry shared with voluntary control. This feedback response supports many task-dependent behaviors and permits the expression of goal-directed corrections at latencies shorter than voluntary reaction time. An extensive body of literature has demonstrated that the LLR shows flexibility akin to voluntary control, but it has not yet been tested whether instruction-dependent LLR changes can also occur in the absence of an overt voluntary response. The present study used kinesthetic motor imagery (experiment 1) and instructed participants to execute movement with the unperturbed contralateral limb (experiment 2) to explore the relationship between the overt production of a voluntary response and LLR facilitation. Activity in stretched right wrist flexors were compared with standard "do not-intervene" and "compensate" conditions. Our findings revealed that on ~40% of imagery and ~50% of contralateral trials, a response occurred during the voluntary epoch in the stretched right wrist flexors. On these "leaked" trials, the early portion of the LLR (R2) was facilitated and displayed a similar increase to compensate trials. The latter half of the LLR (R3) showed further modulation, mirroring the patterns of voluntary epoch activity. By contrast, the LLR on "non-leaked" imagery and contralateral trials did not modulate. We suggest that even though a hastened voluntary response cannot account for all instruction-dependent LLR modulation, the overt execution of a response during the voluntary epoch in the same muscle(s) as the LLR is a prerequisite for instruction-dependent facilitation of this feedback response.NEW & NOTEWORTHY Using motor imagery and contralateral responses, we provide novel evidence that facilitation of the long-latency reflex (LLR) requires the execution of a response during the voluntary epoch. A high proportion of overt response "leaks" were found where the mentally simulated or mirrored response appeared in stretched muscle. The first half of the LLR was categorically sensitive to the appearance of leaks, whereas the latter half displayed characteristics closely resembling activity in the ensuing voluntary period.

Independent representations of ipsilateral and contralateral limbs in primary motor cortex

Several lines of research demonstrate that primary motor cortex (M1) is principally involved in controlling the contralateral side of the body. However, M1 activity has been correlated with both contralateral and ipsilateral limb movements. Why does ipsilaterally-related activity not cause contralateral motor output? To address this question, we trained monkeys to counter mechanical loads applied to their right and left limbs. We found >50% of M1 neurons had load-related activity for both limbs. Contralateral loads evoked changes in activity ~10ms sooner than ipsilateral loads. We also found corresponding population activities were distinct, with contralateral activity residing in a subspace that was orthogonal to the ipsilateral activity. Thus, neural responses for the contralateral limb can be extracted without interference from the activity for the ipsilateral limb, and vice versa. Our results show that M1 activity unrelated to downstream motor targets can be segregated from activity related to the downstream motor output.

Feedforward and Feedback Control Share an Internal Model of the Arm's Dynamics

Recent work has shown that, when countering external forces, the nervous system adjusts not only predictive (i.e., feedforward) control of reaching but also reflex (i.e., feedback) responses to mechanical perturbations. Here we show that altering the physical properties of the arm (i.e., intersegmental dynamics) causes the nervous system to adjust feedforward control and that this learning transfers to feedback responses even though the latter were never directly trained. Forty-five human participants (30 females) performed a single-joint elbow reaching task and countered mechanical perturbations that created pure elbow motion. In our first experiment, we altered intersegmental dynamics by asking participants to generate pure elbow movements when the shoulder joint was either free to rotate or locked by the robotic manipulandum. With the shoulder unlocked, we found robust activation of shoulder flexor muscles for pure elbow flexion trials, as required to counter the interaction torques that arise at the shoulder because of forearm rotation. After locking the shoulder joint, which cancels these interaction torques, we found a substantial reduction in shoulder muscle activity over many trials. In our second experiment, we tested whether such learning transfers to feedback control. Mechanical perturbations applied to the arm with the shoulder unlocked revealed that feedback responses also account for intersegmental dynamics. After locking the shoulder joint, we found a substantial reduction in shoulder feedback responses, as appropriate for the altered intersegmental dynamics. Our work suggests that feedforward and feedback control share an internal model of the arm's dynamics.SIGNIFICANCE STATEMENT Here we show that altering the physical properties of the arm causes people to learn new motor commands and that this learning transfers to their reflex responses to unexpected mechanical perturbations, even though the reflex responses were never directly trained. Our results suggest that feedforward motor commands and reflex responses share an internal model of the arm's dynamics.

Context-dependent tactile texture-sensitivity in monkey M1 and S1 cortex

Caudal primary motor cortex (M1, area 4) is sensitive to cutaneous inputs, but the extent to which the physical details of complex stimuli are encoded is not known. We investigated the sensitivity of M1 neurons (4 Macaca mulatta monkeys) to textured stimuli (smooth/rough or rough/rougher) during the performance of a texture discrimination task and, for some cells, during a no-task condition (same surfaces; no response). The recordings were made from the hemisphere contralateral to the stimulated digits; the motor response (sensory decision) was made with the nonstimulated arm. Most M1 cells were modulated during surface scanning in the task (88%), but few of these were texture-related (24%). In contrast, 44% of M1 neurons were texture related in the no-task condition. Recordings from the neighboring primary somatosensory cortex (S1), the potential source of texture-related signals to M1, showed that S1 neurons were significantly more likely to be texture related during the task (57 vs 24%) than M1. No difference was observed in the no-task condition (52 vs. 44%). In these recordings, the details about surface texture were relevant for S1 but not for M1. We suggest that tactile inputs to M1 were selectively suppressed when the animals were engaged in the task. S1 was spared these controls, as the same inputs were task-relevant. Taken together, we suggest that the suppressive effects are most likely exerted directly at the level of M1, possibly through the activation of a top-down gating mechanism specific to motor set/intention. NEW & NOTEWORTHY Sensory feedback is important for motor control, but we have little knowledge of the contribution of sensory inputs to M1 discharge during behavior. We showed that M1 neurons signal changes in tactile texture, but mainly outside the context of a texture discrimination task. Tactile inputs to M1 were selectively suppressed during the task because this input was not relevant for the recorded hemisphere, which played no role in generating the discrimination response.

Correlations Between Primary Motor Cortex Activity with Recent Past and Future Limb Motion During Unperturbed Reaching

Many studies highlight that human movements are highly successful yet display a surprising amount of variability from trial to trial. There is a consistent pattern of variability throughout movement: initial motor errors are corrected by the end of movement, suggesting the presence of a powerful online control process. Here, we analyze the trial-by-trial variability of goal-directed reaching in nonhuman primates (five male Rhesus monkeys) and demonstrate that they display a similar pattern of variability during reaching, including a strong negative correlation between initial and late hand motion. We then demonstrate that trial-to-trial neural variability of primary motor cortex (M1) is positively correlated with variability of future hand motion (τ = ∼160 ms) during reaching. Furthermore, the variability of M1 activity is also correlated with variability of past hand motion (τ = ∼90 ms), but in the opposite polarity (i.e., negative correlation). Partial correlation analysis demonstrated that M1 activity independently reflects the variability of both past and future hand motions. These findings provide support for the hypothesis that M1 activity is involved in online feedback control of motor actions.SIGNIFICANCE STATEMENT Previous studies highlight that primary motor cortex (M1) rapidly responds to either visual or mechanical disturbances, suggesting its involvement in online feedback control. However, these studies required external disturbances to the motor system and it is not clear whether a similar feedback process addresses internal noise/errors generated by the motor system itself. Here, we introduce a novel analysis that evaluates how variations in the activity of M1 neurons covary with variations in hand motion on a trial-to-trial basis. The analyses demonstrate that M1 activity is correlated with hand motion in both the near future and the recent past, but with opposite polarity. These results suggest that M1 is involved in online feedback motor control to address errors/noise within the motor system.

An exploratory investigation of the effects of whole-head vibration on jaw movements

The perturbing effects of vibration applied to head and body structures are known to destabilize motor control and elicit corrective responses. Although such vibration response testing may be informative for identifying sensorimotor deficits, the effect of whole-head vibration has not been tested on oromotor control. The purpose of this study was to determine how jaw movements respond to the perturbing effects of whole-head vibration during jaw motor tasks. Ten healthy adults completed speech, chewing, and two syllable repetition tasks with and without whole-head vibration. Jaw movements were recorded using 3D optical motion capture. The results showed that the direction and magnitude of the response were dependent on the task. The two syllable repetition tasks responded to vibration, although the direction of the effect differed for the two tasks. Specifically, during vibration, jaw movements became slower and smaller during the syllable repetition task that imposed speed and spatial precision demands, whereas jaw movements became faster and larger during the syllable repetition task that only imposed speed demands. In contrast, jaw movements were unaffected by the vibration during speech and chewing. These findings suggest that the response to vibration may be dependent on spatiotemporal demands, the availability of residual afferent information, and robust feedforward models.

Rapid feedback responses are flexibly coordinated across arm muscles to support goal-directed reaching

A transcortical pathway helps support goal-directed reaching by processing somatosensory information to produce rapid feedback responses across multiple joints and muscles. Here, we tested whether such feedback responses can account for changes in arm configuration and for arbitrary visuomotor transformations-two manipulations that alter how muscles at the elbow and wrist need to be coordinated to achieve task success. Participants used a planar three degree-of-freedom exoskeleton robot to move a cursor to a target following a mechanical perturbation that flexed the elbow. In our first experiment, the cursor was mapped to the veridical position of the robot handle, but participants grasped the handle with two different hand orientations (thumb pointing upward or thumb pointing downward). We found that large rapid feedback responses were evoked in wrist extensor muscles when wrist extension helped move the cursor to the target (i.e., thumb upward), and in wrist flexor muscles when wrist flexion helped move the cursor to the target (i.e., thumb downward). In our second experiment, participants grasped the robot handle with their thumb pointing upward, but the cursor's movement was either veridical or was mirrored such that flexing the wrist moved the cursor as if the participant extended their wrist, and vice versa. After extensive practice, we found that rapid feedback responses were appropriately tuned to the wrist muscles that supported moving the cursor to the target when the cursor was mapped to the mirrored movement of the wrist, but were not tuned to the appropriate wrist muscles when the cursor was remapped to the wrist's veridical movement. NEW & NOTEWORTHY We show that rapid feedback responses were evoked in different wrist muscles depending on the arm's orientation, and this muscle activity was appropriate to generate the wrist motion that supported a reaching action. Notably, we also show that these rapid feedback responses can be evoked in wrist muscles that are detrimental to a reaching action if a nonveridical mapping between wrist and hand motion is extensively learned.

Perspectives on classical controversies about the motor cortex

Primary motor cortex has been studied for more than a century, yet a consensus on its functional contribution to movement control is still out of reach. In particular, there remains controversy as to the level of control produced by motor cortex ("low-level" movement dynamics vs. "high-level" movement kinematics) and the role of sensory feedback. In this review, we present different perspectives on the two following questions: What does activity in motor cortex reflect? and How do planned motor commands interact with incoming sensory feedback during movement? The four authors each present their independent views on how they think the primary motor cortex (M1) controls movement. At the end, we present a dialogue in which the authors synthesize their views and suggest possibilities for moving the field forward. While there is not yet a consensus on the role of M1 or sensory feedback in the control of upper limb movements, such dialogues are essential to take us closer to one.

Motor Cortical Visuomotor Feedback Activity Is Initially Isolated from Downstream Targets in Output-Null Neural State Space Dimensions

Neural circuits must transform new inputs into outputs without prematurely affecting downstream circuits while still maintaining other ongoing communication with these targets. We investigated how this isolation is achieved in the motor cortex when macaques received visual feedback signaling a movement perturbation. To overcome limitations in estimating the mapping from cortex to arm movements, we also conducted brain-machine interface (BMI) experiments where we could definitively identify neural firing patterns as output-null or output-potent. This revealed that perturbation-evoked responses were initially restricted to output-null patterns that cancelled out at the neural population code readout and only later entered output-potent neural dimensions. This mechanism was facilitated by the circuit's large null space and its ability to strongly modulate output-potent dimensions when generating corrective movements. These results show that the nervous system can temporarily isolate portions of a circuit's activity from its downstream targets by restricting this activity to the circuit's output-null neural dimensions.

Perturbation Predictability Can Influence the Long-Latency Stretch Response

Perturbations applied to the upper limbs elicit short (M1: 25–50 ms) and long-latency (M2: 50–100 ms) responses in the stretched muscle. M1 is produced by a spinal reflex loop, and M2 receives contribution from multiple spinal and supra-spinal pathways. While M1 is relatively immutable to voluntary intention, the remarkable feature of M2 is that its size can change based on intention or goal of the participant (e.g., increasing when resisting the perturbation and decreasing when asked to let-go or relax following the perturbation). While many studies have examined modulation of M2 between passive and various active conditions, through the use of constant foreperiods (interval between warning signal and a perturbation), it has also been shown that the magnitude of the M2 response in a passive condition can change based on factors such as habituation and anticipation of perturbation delivery. To prevent anticipation of a perturbation, most studies have used variable foreperiods; however, the range of possible foreperiod duration differs between experiments. The present study examined the influence of different variable foreperiods on modulation of the M2 response. Fifteen participants performed active and passive responses to a perturbation that stretched wrist flexors. Each block of trials had either a short (2.5–3.5 seconds; high predictability) or long (2.5–10.5 seconds; low predictability) variable foreperiod. As expected, no differences were found between any conditions for M1, while M2 was larger in the active rather than passive conditions. Interestingly, within the two passive conditions, the long variable foreperiods resulted in greater activity at the end of the M2 response than the trials with short foreperiods. These results suggest that perturbation predictability, even when using a variable foreperiod, can influence circuitry contributing to the long-latency stretch response.

An examination of the startle response during upper limb stretch perturbations

Unexpected presentation of a startling auditory stimulus (SAS>120 decibels) in a reaction time (RT) paradigm results in the startle reflex and an early release (<100ms) of the preplanned motor response (StartReact effect). Mechanical perturbations applied to the upper limbs elicit short- (M1) and long-latency (M2) stretch reflexes and have also been shown to initiate intended motor responses early (<100ms). Ravichandran et al. (2013) recently proposed that unexpected delivery of a perturbation could also elicit a startle response and therefore the StartReact effect may be responsible for the early trigger of a preplanned response. To investigate this further, we examined startle incidence, RT, and stretch reflex modulation for both expected and unexpected perturbations. In Experiment 1, participants performed active (ACT) and passive (DNI) conditions to an expected large perturbation (similar to previous studies examining M2). The startle response was not observed; however, the perturbation still elicited the voluntary response at short latency (<100ms) and goal-dependent modulation of the M2 response was observed. In Experiment 2, participants performed ACT and DNI conditions to a weak auditory stimulus or a small wrist perturbation. On unexpected trials we probed startle circuitry with a large perturbation or SAS. The SAS consistently elicited a startle response in both ACT and DNI conditions, but startle-like activity was only observed on 17.4% of ACT perturbation probe trials. Our findings suggest that while unexpected upper limb perturbations can be startling, startle triggering of the preplanned voluntary response is not the primary mechanism responsible for goal-dependent modulation of the M2 response.

Coordinating long-latency stretch responses across the shoulder, elbow, and wrist during goal-directed reaching

The long-latency stretch response (muscle activity 50-100 ms after a mechanical perturbation) can be coordinated across multiple joints to support goal-directed actions. Here we assessed the flexibility of such coordination and whether it serves to counteract intersegmental dynamics and exploit kinematic redundancy. In three experiments, participants made planar reaches to visual targets after elbow perturbations and we assessed the coordination of long-latency stretch responses across shoulder, elbow, and wrist muscles. Importantly, targets were placed such that elbow and wrist (but not shoulder) rotations could help transport the hand to the target-a simple form of kinematic redundancy. In experiment 1 we applied perturbations of different magnitudes to the elbow and found that long-latency stretch responses in shoulder, elbow, and wrist muscles scaled with perturbation magnitude. In experiment 2 we examined the trial-by-trial relationship between long-latency stretch responses at adjacent joints and found that the magnitudes of the responses in shoulder and elbow muscles, as well as elbow and wrist muscles, were positively correlated. In experiment 3 we explicitly instructed participants how to use their wrist to move their hand to the target after the perturbation. We found that long-latency stretch responses in wrist muscles were not sensitive to our instructions, despite the fact that participants incorporated these instructions into their voluntary behavior. Taken together, our results indicate that, during reaching, the coordination of long-latency stretch responses across multiple joints counteracts intersegmental dynamics but may not be able to exploit kinematic redundancy.

Distributed task-specific processing of somatosensory feedback for voluntary motor control

Corrective responses to limb disturbances are surprisingly complex, but the neural basis of these goal-directed responses is poorly understood. Here we show that somatosensory feedback is transmitted to many sensory and motor cortical regions within 25 ms of a mechanical disturbance applied to the monkey’s arm. When limb feedback was salient to an ongoing motor action (task engagement), neurons in parietal area 5 immediately (~25 ms) increased their response to limb disturbances, whereas neurons in other regions did not alter their response until 15 to 40 ms later. In contrast, initiation of a motor action elicited by a limb disturbance (target selection) altered neural responses in primary motor cortex ~65 ms after the limb disturbance, and then in dorsal premotor cortex, with no effect in parietal regions until 150 ms post-perturbation. Our findings highlight broad parietofrontal circuits that provide the neural substrate for goal-directed corrections, an essential aspect of highly skilled motor behaviors.

DOI:http://dx.doi.org/10.7554/eLife.13141.001

Humans and other animals can change a movement they are making in a split second, such as when a basketball player has to move around an unexpected opponent to shoot a ball through the hoop. These on-the-fly corrections rely on information about the movement that comes in from the senses. However, it was unclear how the brain and spinal cord process this sensory information to guide movement.

Omrani et al. have now recorded electrical activity from the brains of monkeys while the animals tried to keep their hand at a target. Each monkey wore a robotic exoskeleton that would occasionally move the monkey’s arm. Even if the monkey was not engaged in a motor task, a small nudge of the limb by the robot caused neural activity to spread rapidly throughout the sensory and motor regions of the cerebral cortex (the outer layer of the brain).

In some trials, when the monkey was actively trying to keep its hand at a target, the robot would again nudge the monkey’s arm slightly. Omrani et al. observed that within 25 milliseconds of this nudge, the activity in an area of the cortex called parietal area 5 responded even more, suggesting that this area was using information from the senses to actively deal with the change in arm position. This change in activity then spread to other parts of the brain.

In another set of trials, the monkey was trained to move to a second target if the robot nudged its arm. In this case, the activity in an area called the primary motor cortex increased even more, likely supporting the monkey’s ability to rapidly move to this second target. Overall, the study by Omrani et al. highlights the complex way that sensory feedback is processed in the cerebral cortex, supporting our ability to perform highly skilled motor actions.

DOI:http://dx.doi.org/10.7554/eLife.13141.002

Voluntary reaction time and long-latency reflex modulation

Stretching a muscle of the upper limb elicits short (M1) and long-latency (M2) reflexes. When the participant is instructed to actively compensate for a perturbation, M1 is usually unaffected and M2 increases in size and is followed by the voluntary response. It remains unclear if the observed increase in M2 is due to instruction-dependent gain modulation of the contributing reflex mechanism(s) or results from voluntary response superposition. The difficulty in delineating between these alternatives is due to the overlap between the voluntary response and the end of M2. The present study manipulated response accuracy and complexity to delay onset of the voluntary response and observed the corresponding influence on electromyographic activity during the M2 period. In all active conditions, M2 was larger compared with a passive condition where participants did not respond to the perturbation; moreover, these changes in M2 began early in the appearance of the response (∼ 50 ms), too early to be accounted for by voluntary overlap. Voluntary response latency influenced the latter portion of M2, with the largest activity seen when accuracy of limb position was not specified. However, when participants aimed for targets of different sizes or performed movements of various complexities, reaction time differences did not influence M2 period activity, suggesting voluntary activity was sufficiently delayed. Collectively, our results show that while a perturbation applied to the upper limbs can trigger a voluntary response at short latency (<100 ms), instruction-dependent reflex gain modulation remains an important contributor to EMG changes during the M2 period.

Goal-dependent modulation of the long-latency stretch response at the shoulder, elbow, and wrist

Many studies have demonstrated that muscle activity 50-100 ms after a mechanical perturbation (i.e., the long-latency stretch response) can be modulated in a manner that reflects voluntary motor control. These previous studies typically assessed modulation of the long-latency stretch response from individual muscles rather than how this response is concurrently modulated across multiple muscles. Here we investigated such concurrent modulation by having participants execute goal-directed reaches to visual targets after mechanical perturbations of the shoulder, elbow, or wrist while measuring activity from six muscles that articulate these joints. We found that shoulder, elbow, and wrist muscles displayed goal-dependent modulation of the long-latency stretch response, that the relative magnitude of participants' goal-dependent activity was similar across muscles, that the temporal onset of goal-dependent muscle activity was not reliably different across the three joints, and that shoulder muscles displayed goal-dependent activity appropriate for counteracting intersegmental dynamics. We also observed that the long-latency stretch response of wrist muscles displayed goal-dependent modulation after elbow perturbations and that the long-latency stretch response of elbow muscles displayed goal-dependent modulation after wrist perturbations. This pattern likely arises because motion at either joint could bring the hand to the visual target and suggests that the nervous system rapidly exploits such simple kinematic redundancy when processing sensory feedback to support goal-directed actions.

Neurons in red nucleus and primary motor cortex exhibit similar responses to mechanical perturbations applied to the upper-limb during posture

Primary motor cortex (M1) and red nucleus (RN) are brain regions involved in limb motor control. Both structures are highly interconnected with the cerebellum and project directly to the spinal cord, although the contribution of RN is smaller than M1. It remains uncertain whether RN and M1 serve similar or distinct roles during posture and movement. Many neurons in M1 respond rapidly to mechanical disturbances of the limb, but it remains unclear whether RN neurons also respond to such limb perturbations. We have compared discharges of single neurons in RN (n = 49) and M1 (n = 109) of one monkey during a postural perturbation task. Neural responses to whole-limb perturbations were examined by transiently applying (300 ms) flexor or extensor torques to the shoulder and/or elbow while the monkeys attempted to maintain a static hand posture. Relative to baseline discharges before perturbation onset, perturbations evoked rapid (<100 ms) changes of neural discharges in many RN (28 of 49, 57%) and M1 (43 of 109, 39%) neurons. In addition to exhibiting a greater proportion of perturbation-related neurons, RN neurons also tended to exhibit higher peak discharge frequencies in response to perturbations than M1 neurons. Importantly, neurons in both structures exhibited similar response latencies and tuning properties (preferred torque directions and tuning widths) in joint-torque space. Proximal arm muscles also displayed similar tuning properties in joint-torque space. These results suggest that RN is more sensitive than M1 to mechanical perturbations applied during postural control but both structures may play a similar role in feedback control of posture.

Long-latency reflexes account for limb biomechanics through several supraspinal pathways

Accurate control of body posture is enforced by a multitude of corrective actions operating over a range of time scales. The earliest correction is the short-latency reflex (SLR) which occurs between 20–45 ms following a sudden displacement of the limb and is generated entirely by spinal circuits. In contrast, voluntary reactions are generated by a highly distributed network but at a significantly longer delay after stimulus onset (greater than 100 ms). Between these two epochs is the long-latency reflex (LLR) (around 50–100 ms) which acts more rapidly than voluntary reactions but shares some supraspinal pathways and functional capabilities. In particular, the LLR accounts for the arm’s biomechanical properties rather than only responding to local muscle stretch like the SLR. This paper will review how the LLR accounts for the arm’s biomechanical properties and the supraspinal pathways supporting this ability. Relevant experimental paradigms include clinical studies, non-invasive brain stimulation, neural recordings in monkeys, and human behavioral studies. The sum of this effort indicates that primary motor cortex and reticular formation (RF) contribute to the LLR either by generating or scaling its structured response appropriate for the arm’s biomechanics whereas the cerebellum scales the magnitude of the feedback response. Additional putative pathways are discussed as well as potential research lines.

Interactions between stretch and startle reflexes produce task-appropriate rapid postural reactions

Neural pathways underpinning startle reflex and limb stretch reflexes evolved independently and have served vastly different purposes. In their most basic form, the pathways responsible for these reflex responses are relatively simple processing units that produce a motoric response that is proportional to the stimulus received. It is becoming clear however, that rapid responses to external stimuli produced by human and non-human primates are context-dependent in a manner similar to voluntary movements. This mini review discusses the nature of startle and stretch reflex interactions in human and non-human primates and the involvement of the primary motor cortex in their regulation.