Flexibility of motor pattern generation across stimulation conditions by the neonatal rat spinal cord

Generating variability from motor primitives during infant locomotor development

Motor variability is a fundamental feature of developing systems allowing motor exploration and learning. In human infants, leg movements involve a small number of basic coordination patterns called locomotor primitives, but whether and when motor variability could emerge from these primitives remains unknown. Here we longitudinally followed 18 infants on 2-3 time points between birth (~4 days old) and walking onset (~14 months old) and recorded the activity of their leg muscles during locomotor or rhythmic movements. Using unsupervised machine learning, we show that the structure of trial-to-trial variability changes during early development. In the neonatal period, infants own a minimal number of motor primitives but generate a maximal motor variability across trials thanks to variable activations of these primitives. A few months later, toddlers generate significantly less variability despite the existence of more primitives due to more regularity within their activation. These results suggest that human neonates initiate motor exploration as soon as birth by variably activating a few basic locomotor primitives that later fraction and become more consistently activated by the motor system.

Development of Locomotor-Related Movements in Early Infancy

This mini-review focuses on the emergence of locomotor-related movements in early infancy. In particular, we consider multiples precursor behaviors of locomotion as a manifestation of the development of the neuronal networks and their link in the establishment of precocious locomotor skills. Despite the large variability of motor behavior observed in human babies, as in animals, afferent information is already processed to shape the behavior to specific situations and environments. Specifically, we argue that the closed-loop interaction between the neural output and the physical dynamics of the mechanical system should be considered to explore the complexity and flexibility of pattern generation in human and animal neonates.

Distinct locomotor precursors in newborn babies

Mature locomotion involves modular spinal drives generating a set of fundamental patterns of motoneuron activation, each timed at a specific phase of locomotor cycles and associated with a stable muscle synergy. How locomotor modules develop and to what extent they depend on prior experience or intrinsic programs remains unclear. To address these issues, we herein leverage the presence at birth of two types of locomotor-like movements, spontaneous kicking and weight-bearing stepping. The former is expressed thousands of times in utero and postnatally, whereas the latter is elicited de novo by placing the newborn on the ground for the first time. We found that the neuromuscular modules of stepping and kicking differ substantially. Neonates kicked with an adult-like number of temporal activation patterns, which lacked a stable association with systematic muscle synergies across movements. However, on the ground neonates stepped with fewer temporal patterns but all structured in stable synergies. Since kicking and ground-stepping coexist at birth, switching between the two behaviors may depend on a dynamic reconfiguration of the underlying neural circuits as a function of sensory feedback from surface contact. We tracked the development of ground-stepping in 4- to 48-mo-old infants and found that, after the age of 6 mo, the number of temporal patterns increased progressively, reaching adult-like conformation only after independent walking was established. We surmise that mature locomotor modules may derive by combining the multiple patterns of repeated kicking, on the one hand, with synergies resulting from fractionation of those revealed by sporadic weight-bearing stepping, on the other hand.

Chronic electromyograms in treadmill running SOD1 mice reveal early changes in muscle activation

KEY POINTS

The present study demonstrates that electromyograms (EMGs) obtained during locomotor activity in mice were effective for identification of early physiological markers of amyotrophic lateral sclerosis (ALS). These measures could be used to evaluate therapeutic intervention strategies in animal models of ALS. Several parameters of locomotor activity were shifted early in the disease time course in SOD1G93A mice, especially when the treadmill was inclined, including intermuscular phase, burst skew and amplitude of the locomotor bursts. The results of the present study indicate that early compensatory changes may be taking place within the neural network controlling locomotor activity, including spinal interneurons. Locomotor EMGs could have potential use as a clinical diagnostic tool.

ABSTRACT

To improve our understanding of early disease mechanisms and to identify reliable biomarkers of amyotrophic lateral sclerosis (ALS), a progressive neurodegenerative disease, we measured electromyogram (EMG) activity in hind limb muscles of SOD1G93A mice. By contrast to clinical diagnostic measures using EMGs, which are performed on quiescent patients, we monitored activity during treadmill running aiming to detect presymptomatic changes in motor patterning. Chronic EMG electrodes were implanted into vastus lateralis, biceps femoris posterior, lateral gastrocnemius and tibialis anterior in mice from postnatal day 55 to 100 and the results obtained were assessed using linear mixed models. We evaluated differences in parameters related to EMG amplitude (peak and area) and timing (phase and skew, a measure of burst shape) when animals ran on level and inclined treadmills. There were significant changes in both the timing of activity and the amplitude of EMG bursts in SOD1G93A mice. Significant differences between wild-type and SOD1G93A mice were mainly observed when animals locomoted on inclined treadmills. All muscles had significant effects of mutation that were independent of age. These novel results indicate (i) locomotor EMG activity might be an early measure of disease onset; (ii) alterations in locomotor patterning may reflect changes in neuronal drive and compensation at the network level including altered activity of spinal interneurons; and (iii) the increased power output necessary on an inclined treadmill was important in revealing altered activity in SOD1G93A mice.

Spinal Locomotor Circuits Develop Using Hierarchical Rules Based on Motorneuron Position and Identity

The coordination of multi-muscle movements originates in the circuitry that regulates the firing patterns of spinal motorneurons. Sensory neurons rely on the musculotopic organization of motorneurons to establish orderly connections, prompting us to examine whether the intraspinal circuitry that coordinates motor activity likewise uses cell position as an internal wiring reference. We generated a motorneuron-specific GCaMP6f mouse line and employed two-photon imaging to monitor the activity of lumbar motorneurons. We show that the central pattern generator neural network coordinately drives rhythmic columnar-specific motorneuron bursts at distinct phases of the locomotor cycle. Using multiple genetic strategies to perturb the subtype identity and orderly position of motorneurons, we found that neurons retained their rhythmic activity-but cell position was decoupled from the normal phasing pattern underlying flexion and extension. These findings suggest a hierarchical basis of motor circuit formation that relies on increasingly stringent matching of neuronal identity and position.

From spontaneous motor activity to coordinated behaviour: a developmental model

In mammals, the developmental path that links the primary behaviours observed during foetal stages to the full fledged behaviours observed in adults is still beyond our understanding. Often theories of motor control try to deal with the process of incremental learning in an abstract and modular way without establishing any correspondence with the mammalian developmental stages. In this paper, we propose a computational model that links three distinct behaviours which appear at three different stages of development. In order of appearance, these behaviours are: spontaneous motor activity (SMA), reflexes, and coordinated behaviours, such as locomotion. The goal of our model is to address in silico four hypotheses that are currently hard to verify in vivo: First, the hypothesis that spinal reflex circuits can be self-organized from the sensor and motor activity induced by SMA. Second, the hypothesis that supraspinal systems can modulate reflex circuits to achieve coordinated behaviour. Third, the hypothesis that, since SMA is observed in an organism throughout its entire lifetime, it provides a mechanism suitable to maintain the reflex circuits aligned with the musculoskeletal system, and thus adapt to changes in body morphology. And fourth, the hypothesis that by changing the modulation of the reflex circuits over time, one can switch between different coordinated behaviours. Our model is tested in a simulated musculoskeletal leg actuated by six muscles arranged in a number of different ways. Hopping is used as a case study of coordinated behaviour. Our results show that reflex circuits can be self-organized from SMA, and that, once these circuits are in place, they can be modulated to achieve coordinated behaviour. In addition, our results show that our model can naturally adapt to different morphological changes and perform behavioural transitions.

Force-sensitive afferents recruited during stance encode sensory depression in the contralateral swinging limb during locomotion

Afferent feedback alters muscle activity during locomotion and must be tightly controlled. As primary afferent depolarization-induced presynaptic inhibition (PAD-PSI) regulates afferent signaling, we investigated hindlimb PAD-PSI during locomotion in an in vitro rat spinal cord-hindlimb preparation. We compared the relation of PAD-PSI, measured as dorsal root potentials (DRPs), to observed ipsilateral and contralateral limb endpoint forces. Afferents activated during stance-phase force strongly and proportionately influenced DRP magnitude in the swinging limb. Responses increased with locomotor frequency. Electrical stimulation of contralateral afferents also preferentially evoked DRPs in the opposite limb during swing (flexion). Nerve lesioning, in conjunction with kinematic results, support a prominent contribution from toe Golgi tendon organ afferents. Thus, force-dependent afferent feedback during stance binds interlimb sensorimotor state to a proportional PAD-PSI in the swinging limb, presumably to optimize interlimb coordination. These results complement known actions of ipsilateral afferents on PAD-PSI during locomotion.

Defining feasible bounds on muscle activation in a redundant biomechanical task: practical implications of redundancy

Measured muscle activation patterns often vary significantly from musculoskeletal model predictions that use optimization to resolve redundancy. Although experimental muscle activity exhibits both inter- and intra-subject variability we lack adequate tools to quantify the biomechanical latitude that the nervous system has when selecting muscle activation patterns. Here, we identified feasible ranges of individual muscle activity during force production in a musculoskeletal model to quantify the degree to which biomechanical redundancy allows for variability in muscle activation patterns. In a detailed cat hindlimb model matched to the posture of three cats, we identified the lower and upper bounds on muscle activity in each of 31 muscles during static endpoint force production across different force directions and magnitudes. Feasible ranges of muscle activation were relatively unconstrained across force magnitudes such that only a few (0-13%) muscles were found to be truly "necessary" (e.g. exhibited non-zero lower bounds) at physiological force ranges. Most of the muscles were "optional", having zero lower bounds, and frequently had "maximal" upper bounds as well. Moreover, "optional" muscles were never selected by optimization methods that either minimized muscle stress, or that scaled the pattern required for maximum force generation. Therefore, biomechanical constraints were generally insufficient to restrict or specify muscle activation levels for producing a force in a given direction, and many muscle patterns exist that could deviate substantially from one another but still achieve the task. Our approach could be extended to identify the feasible limits of variability in muscle activation patterns in dynamic tasks such as walking.

Dopamine exerts activation-dependent modulation of spinal locomotor circuits in the neonatal mouse

Monoamines can modulate the output of a variety of invertebrate and vertebrate networks, including the spinal cord networks that control walking. Here we examined the multiple changes in the output of locomotor networks induced by dopamine (DA). We found that DA can depress the activation of locomotor networks in the neonatal mouse spinal cord following ventral root stimulation. By examining disinhibited rhythms, where the Renshaw cell pathway was blocked, we found that DA depresses a putative recurrent excitatory pathway that projects onto rhythm-generating circuitry of the spinal cord. This depression was D(2) but not D(1) receptor dependent and was not due exclusively to depression of excitatory drive to motoneurons. Furthermore, the depression in excitation was not dependent on network activity. We next compared the modulatory effects of DA on network function by focusing on a serotonin and a N-methyl-dl-aspartate-evoked rhythm. In contrast to the depressive effects on a ventral root-evoked rhythm, we found that DA stabilized a drug-evoked rhythm, reduced the frequency of bursting, and increased amplitude. Overall, these data demonstrate that DA can potentiate network activity while at the same time reducing the gain of recurrent excitatory feedback loops from motoneurons onto the network.

Enabling techniques for in vitro studies on mammalian spinal locomotor mechanisms

The neonatal rodent spinal cord maintained in vitro is a powerful model system to understand the central properties of spinal circuits generating mammalian locomotion. We describe three enabling approaches that incorporate afferent input and attached hindlimbs. (i) Sacral dorsal column stimulation recruits and strengthens ongoing locomotor-like activity, and implementation of a closed positive-feedback paradigm is shown to support its stimulation as an untapped therapeutic site for locomotor modulation. (ii) The spinal cord hindlimbs-restrained preparation allows suction electrode electromyographic recordings from many muscles. Inducible complex motor patterns resemble natural locomotion, and insights into circuit organization are demonstrated during spontaneous motor burst 'deletions', or following sensory stimuli such as tail and paw pinch. (iii) The spinal cord hindlimbs-pendant preparation produces unrestrained hindlimb stepping. It incorporates mechanical limb perturbations, kinematic analyses, ground reaction force monitoring, and the use of treadmills to study spinal circuit operation with movement-related patterns of sensory feedback while providing for stable whole-cell recordings from spinal neurons. Such techniques promise to provide important additional insights into locomotor circuit organization.

Development of human locomotion

Neural control of locomotion in human adults involves the generation of a small set of basic patterned commands directed to the leg muscles. The commands are generated sequentially in time during each step by neural networks located in the spinal cord, called Central Pattern Generators. This review outlines recent advances in understanding how motor commands are expressed at different stages of human development. Similar commands are found in several other vertebrates, indicating that locomotion development follows common principles of organization of the control networks. Movements show a high degree of flexibility at all stages of development, which is instrumental for learning and exploration of variable interactions with the environment.

Cervicolumbar coordination in mammalian quadrupedal locomotion: role of spinal thoracic circuitry and limb sensory inputs

Effective quadrupedal locomotion requires a close coordination between the spatially distant central pattern generators (CPGs) controlling forelimb and hindlimb movements. Using isolated preparations of the neonatal rat spinal cord, we explore the role of intervening thoracic circuitry in cervicolumbar CPG coordination and the contribution to this remote coupling of limb somatosensory inputs. In preparations activated with bath-applied N-methyl-D,L-aspartate, serotonin, and dopamine, the coordination between locomotor-related bursts recorded in cervical and lumbar ventral roots was substantially weakened, although not abolished, when the thoracic segments were selectively withheld from neurochemical stimulation or were exposed to a low Ca(2+) solution to block synaptic transmission. Moreover, cervicolumbar CPG coordination was reduced after a thoracic midsagittal section, suggesting that cross-cord projections participate in the anteroposterior coupling. In quiescent preparations, either cyclic or tonic electrical stimulation of low-threshold afferent pathways in C8 or L2 dorsal roots (DRs) could elicit coordinated ventral root bursting at both cervical and lumbar levels via an activation of the underlying CPG networks. When lumbar rhythmogenesis was prevented by local synaptic transmission blockade, L2 DR stimulation could still drive left-right alternating cervical bursting in preparations otherwise exposed to normal bathing medium. In contrast, when the cervical generators were selectively blocked, C8 DR stimulation was unable to activate the lumbar CPGs. Thus, in the newborn rat, anteroposterior limb coordination relies on active burst generation within midcord thoracic circuitry that additionally conveys ascending and weaker descending coupling influences of distant limb proprioceptive inputs to the cervical and lumbar generators, respectively.

Motoneuronal and muscle synergies involved in cat hindlimb control during fictive and real locomotion: a comparison study

We compared the activity profiles and synergies of spinal motoneurons recorded during fictive locomotion evoked in immobilized decerebrate cat preparations by midbrain stimulation to the activity profiles and synergies of the corresponding hindlimb muscles obtained during forward level walking in cats. The fictive locomotion data were collected in the Spinal Cord Research Centre, University of Manitoba, and provided by Dr. David McCrea; the real locomotion data were obtained in the laboratories of M. A. Lemay and B. I. Prilutsky. Scatterplot representation and minimum spanning tree clustering algorithm were used to identify the possible motoneuronal and muscle synergies operating during both fictive and real locomotion. We found a close similarity between the activity profiles and synergies of motoneurons innervating one-joint muscles during fictive locomotion and the profiles and synergies of the corresponding muscles during real locomotion. However, the activity patterns of proximal nerves controlling two-joint muscles, such as posterior biceps and semitendinosus (PBSt) and rectus femoris (RF), were not uniform in fictive locomotion preparations and differed from the activity profiles of the corresponding two-joint muscles recorded during forward level walking. Moreover, the activity profiles of these nerves and the corresponding muscles were unique and could not be included in the synergies identified in fictive and real locomotion. We suggest that afferent feedback is involved in the regulation of locomotion via motoneuronal synergies controlled by the spinal central pattern generator (CPG) but may also directly affect the activity of motoneuronal pools serving two-joint muscles (e.g., PBSt and RF). These findings provide important insights into the organization of the spinal CPG in mammals, the motoneuronal and muscle synergies engaged during locomotion, and their afferent control.

The same core rhythm generator underlies different rhythmic motor patterns

Rhythmically active motor circuits can generate different activity patterns in response to different inputs. In most systems, however, it is not known whether the same neurons generate the underlying rhythm for each different pattern. Thus far, information regarding the degree of conservation of rhythm generator neurons is limited to a few pacemaker-driven circuits, in most of which the core rhythm generator is unchanged across different output patterns. We are addressing this issue in the network-driven, gastric mill (chewing) circuit in the crab stomatogastric nervous system. We first establish that distinct gastric mill motor patterns are triggered by separate stimulation of two extrinsic input pathways, the ventral cardiac neurons (VCNs) and postoesophageal commissure (POC) neurons. A prominent feature that distinguishes these gastric mill motor patterns is the LG (lateral gastric) protractor motor neuron activity pattern, which is tonic during the VCN rhythm and exhibits fast rhythmic bursting during the POC rhythm. These two motor patterns also differed in their cycle period and some motor neuron phase relationships, duty cycles, and burst durations. Despite the POC and VCN motor patterns being distinct, rhythm generation during each motor pattern required the activity of the same two, reciprocally inhibitory gastric mill neurons [LG, Int1 (interneuron 1)]. Specifically, reversibly hyperpolarizing LG or Int1, but no other gastric mill neuron, delayed the start of the next gastric mill cycle until after the imposed hyperpolarization. Thus, the same circuit neurons can comprise the core rhythm generator during different versions of a network-driven rhythmic motor pattern.

Estimation of musculoskeletal models from in situ measurements of muscle action in the rat hindlimb

Musculoskeletal models are often created by making detailed anatomical measurements of muscle properties. These measurements can then be used to determine the parameters of canonical models of muscle action. We describe here a complementary approach for developing and validating muscle models, using in situ measurements of muscle actions. We characterized the actions of two rat hindlimb muscles: the gracilis posticus (GRp) and the posterior head of biceps femoris (BFp; excluding the anterior head with vertebral origin). The GRp is a relatively simple muscle, with a circumscribed origin and insertion. The BFp is more complex, with an insertion distributed along the tibia. We measured the six-dimensional isometric forces and moments at the ankle evoked from stimulating each muscle at a range of limb configurations. The variation of forces and moments across the workspace provides a succinct characterization of muscle action. We then used this data to create a simple muscle model with a single point insertion and origin. The model parameters were optimized to best explain the observed force-moment data. This model explained the relatively simple muscle, GRp, very well (R(2)>0.85). Surprisingly, this simple model was also able to explain the action of the BFp, despite its greater complexity (R(2)>0.84). We then compared the actions observed here with those predicted using recently published anatomical measurements. Although the forces and moments predicted for the GRp were very similar to those observed here, the predictions for the BFp differed. These results show the potential utility of the approach described here for the development and refinement of musculoskeletal models based on in situ measurements of muscle actions.

Central Pattern Generators of the Mammalian Spinal Cord

Neuronal networks within the spinal cord of mammals are responsible for generating various rhythmic movements, such as walking, running, swimming, and scratching. The ability to generate multiple rhythmic movements highlights the complexity and flexibility of the mammalian spinal circuitry. The present review describes features of some rhythmic motor behaviors generated by the mammalian spinal cord and discusses how the spinal circuitry is able to produce different rhythmic movements with their own sets of goals and demands.

Control of breathing by raphe obscurus serotonergic neurons in mice

We used optogenetics to determine the global respiratory effects produced by selectively stimulating raphe obscurus (RO) serotonergic neurons in anesthetized mice and to test whether these neurons detect changes in the partial pressure of CO(2), and hence function as central respiratory chemoreceptors. Channelrhodopsin-2 (ChR2) was selectively (∼97%) incorporated into ∼50% of RO serotonergic neurons by injecting AAV2 DIO ChR2-mCherry (adeno-associated viral vector double-floxed inverse open reading frame of ChR2-mCherry) into the RO of ePet-Cre mice. The transfected neurons heavily innervated lower brainstem and spinal cord regions involved in autonomic and somatic motor control plus breathing but eschewed sensory related regions. Pulsed laser photostimulation of ChR2-transfected serotonergic neurons increased respiratory frequency (fR) and diaphragmatic EMG (dEMG) amplitude in relation to the duration and frequency of the light pulses (half saturation, 1 ms; 5-10 Hz). dEMG amplitude and fR increased slowly (half saturation after 10-15 s) and relaxed monoexponentially (tau, 13-15 s). The breathing stimulation was reduced ∼55% by methysergide (broad spectrum serotonin antagonist) and potentiated (∼16%) at elevated levels of inspired CO(2) (8%). RO serotonergic neurons, identified by their entrainment to short light pulses (threshold, 0.1-1 ms) were silent (nine cells) or had a low and regular level of activity (2.1 ± 0.4 Hz; 11 cells) that was not synchronized with respiration. These and nine surrounding neurons with similar characteristics were unaffected by adding up to 10% CO(2) to the breathing mixture. In conclusion, RO serotonergic neurons activate breathing frequency and amplitude and potentiate the central respiratory chemoreflex but do not appear to have a central respiratory chemoreceptor function.

Spatiotemporal dynamics of rhythmic spinal interneurons measured with two-photon calcium imaging and coherence analysis

In rhythmic neural circuits, a neuron often fires action potentials with a constant phase to the rhythm, a timing relationship that can be functionally significant. To characterize these phase preferences in a large-scale, cell type-specific manner, we adapted multitaper coherence analysis for two-photon calcium imaging. Analysis of simulated data showed that coherence is a simple and robust measure of rhythmicity for calcium imaging data. When applied to the neonatal mouse hindlimb spinal locomotor network, the phase relationships between peak activity of >1,000 ventral spinal interneurons and motor output were characterized. Most interneurons showed rhythmic activity that was coherent and in phase with the ipsilateral motor output during fictive locomotion. The phase distributions of two genetically identified classes of interneurons were distinct from the ensemble population and from each other. There was no obvious spatial clustering of interneurons with similar phase preferences. Together, these results suggest that cell type, not neighboring neuron activity, is a better indicator of an interneuron's response during fictive locomotion. The ability to measure the phase preferences of many neurons with cell type and spatial information should be widely applicable for studying other rhythmic neural circuits.

Specificity of intramuscular activation during rhythms produced by spinal patterning systems in the in vitro neonatal rat with hindlimb attached preparation

In intact adult vertebrates, muscles can be activated with a high degree of specificity, so that even within a single traditionally defined muscle, groups of motor units can be differentially activated. Such differential activation might reflect detailed control by descending systems, potentially resulting from postnatal experience such that activation of motor units is precisely tailored to their mechanical actions. Here we examine the degree to which such specific activation can be seen in the rhythmic patterns produced by isolated spinal motor systems in neonates. We examined motor output produced by the in vitro neonatal rat spinal cord with hindlimb attached. We recorded the activity of different regions within the posterior portion of biceps femoris (BFp; i.e., excluding the anterior/vertebral head). We found that in the rhythms evoked by bath application of serotonin/N-methyl-d-aspartate (5-HT/NMDA), all regions of BFp were active during extension. However, the regions of BFp were activated in a specific sequence, with the activation of rostral regions consistently preceding those of more caudal regions in both afferented and deafferented preparations. In the rhythms evoked by cauda equina (CE) stimulation, rostral and middle regions of BFp remained active in extension, but the caudal region of BFp was usually active in flexion. Stimulation of L5 and S2 dorsal roots typically evoked rhythms with all regions of BFp active during extension; although the same rostral to caudal sequence of activation observed in 5-HT/NMDA evoked rhythms could also be observed in these rhythms, we also observed cases with reversed sequences, with activity proceeding from caudal to rostral. S2 dorsal root stimulation occasionally evoked rhythms with flexor activity in caudal BFp, similar to CE-evoked rhythms. Taken together, these results suggest a high degree of individuated control of muscles by spinal pattern generating networks, even at birth.

Understanding complex muscles in the rat hindlimb: Activations and actions

We present research examining the function of complex muscles in the rat hindlimb. Two related sets of experiments are described. In the first, we examine the degree of specificity in spinal pattern generators, assessing whether the pattern generators at birth are capable of differentially activating intramuscular subdivisions in the complex hindlimb muscle biceps femoris. In the second, we describe a novel approach for creating a musculoskeletal model to capture the mechanical actions of individual muscles and evaluate its ability to capture the action of both simple and complex muscles in the rat hindlimb.