Biology subjects

Hug, F.

Publications and source records attributed to Hug, F..

At least 19 recordsLinked to original sources

bioRxiv2026

Brain Oscillations Extend Beyond Task-Relevant Motor Neuron Pools and Contribute to Shaping the Functional State of the Motor System

It remains unknown whether oscillatory brain activity associated with sensorimotor behavior is routed selectively to task-relevant muscles or expressed more broadly, including in task-irrelevant muscles. Here we combined electroencephalography with large-scale recordings of spinal motor neurons innervating the tibialis anterior. Participants maintained a submaximal dorsiflexion while performing a Go/No-Go task in which the instructed response was either a ballistic dorsiflexion or a ballistic handgrip contraction, making the tibialis anterior task-relevant or task-irrelevant, respectively. Alpha- and beta-band modulations observed at the cortical level were largely expressed in motor neuron output, including in the task-irrelevant motor neuron pool. The peripheral expression of these modulations differed across frequency bands: alpha was partly effector-dependent, consistent with more selective transmission to the task-relevant pool, whereas beta was largely effector-independent, consistent with broader expression across motor neuron pools. Using simulation-based inference, we found that task-related changes in motor output were best explained by modulations in net excitatory drive, whereas alpha- and beta-band inputs contributed primarily to motor neuron synchronization. A complementary simulation showed that this synchronization may facilitate rapid changes in motor output. These results support a parallel control architecture in which low-frequency drive determines motor output, whereas higher-frequency oscillatory inputs shape synchronization within motor pools more broadly, potentially setting the motor system in a state that favours rapid adjustments in output.

neuroscience↗

bioRxiv2026

Disrupted Inhibitory-Excitatory Balance Underlies Spinal Motoneuron Dysfunction in Incomplete Spinal Cord Injury

Incomplete spinal cord injury disrupts voluntary movement, in part through motoneuron dysfunction, yet the mechanisms underlying this dysfunction remain poorly understood. Using a non-invasive approach to decode the spiking activity of large populations of spinal motoneurons, we quantified the relative contributions of excitatory, inhibitory, and neuromodulatory inputs to motoneuron rate coding after chronic incomplete spinal cord injury. Eighteen participants with incomplete spinal cord injury and 18 age- and sex-matched control participants performed submaximal isometric plantar flexion tasks while high-density surface electromyography was recorded from the soleus and gastrocnemius medialis muscles. Motoneuron firing behaviour was analysed to estimate neuromodulatory drive and the balance between inhibitory and excitatory inputs. Participants with incomplete spinal cord injury exhibited lower rate coding, characterised by lower firing rates at recruitment, lower firing rate modulation, and lower peak firing rates compared with healthy controls. Although estimates of neuromodulatory drive did not differ between groups, individuals with spinal cord injury showed a shift in the inhibition-excitation balance toward greater inhibition compared with controls. Furthermore, increasing inhibitory input through muscle length changes and antagonist tendon vibration modulated motoneuron firing in controls, but not in individuals with incomplete spinal cord injury. Together, these findings suggest that impaired rate coding after incomplete spinal cord injury arises from an altered inhibitory-excitatory balance rather than reduced neuromodulatory drive. Taking advantage of methodological advances to decode spinal motor neuron activity during voluntary contraction, this study identified excessive inhibitory input to spinal motoneurons as a key neural mechanism contributing to muscle weakness and impaired motor function in individuals with incomplete spinal cord injury.

neuroscience↗

bioRxiv2025

Probabilistic inference of Homonymous and Heteronymous Recurrent Inhibition in Human Muscles from Large-Scale Motor Neuron Recordings

Understanding how spinal circuits shape motor neuron behavior during muscle contractions remains a major challenge. Here, we combined large-scale motor unit recordings with simulation-based inference to generate probabilistic estimates of homonymous and heteronymous recurrent inhibition, a key spinal circuit that has remained largely inaccessible during natural voluntary contractions. We constructed synchronization cross-histograms from motor neuron spike trains and extracted features representative of recurrent inhibition. Because these features are also influenced by higher-frequency components of common synaptic input, we developed a simulation-based inference framework to disentangle these effects. Following validation, we applied this framework to experimental data from six muscles at two contraction intensities, revealing previously uncharacterized muscle-and intensity-dependent patterns: recurrent inhibition decreased with contraction intensity in most muscles but increased in the vastus lateralis and medialis. The pipeline is openly available and designed for reuse on comparable datasets and for adaptation to diverse experimental contexts, including other spinal circuits. TeaserLarge-scale motor-unit recordings enable in-vivo probabilistic mapping of spinal recurrent inhibition.

neuroscience↗

bioRxiv2025

The impact of mechanical requirements on the neural control of skeletal muscle and subsequent energetic rates

1The energetic cost of skeletal muscle contraction is a fundamental driver in the selection of locomotor strategies. Muscle energy consumption depends on muscle-fibre typology, neural drive, and mechanical state. While the influence of fibre-type and mechanical state on energy use has been extensively characterised in isolated muscle preparations, these experiments fail to capture the influence of realistic in vivo motor unit recruitment strategies. Hence, this study aims to capture neural drive in the tibialis anterior, in particular motor unit recruitment thresholds and rate coding, and the corresponding changes in energy use across a range of mechanical demands. Fixed ankle angle dorsiflexion contractions were performed at varying rates of torque development, while high density electromyography characterized motor unit spiking activity, B-mode ultrasound captured muscle fascicle dynamics, and indirect calorimetry measured energetic rates. Faster rates of torque development required earlier recruitment of motor units with increased motor unit firing rates which coincided with increased fascicle strain rates. At higher rates of torque development, these altered motor unit discharge patterns and muscle mechanics coincided with increased muscle energy use. Together, these findings highlight that in vivo muscle energetics emerge from the dynamic interplay between neural control and mechanical demands.

physiology↗

bioRxiv2025

Diurnal Modulation of Persistent Inward Current Contribution to Spinal Motor Neuron Behaviour

Despite the critical role of persistent inward currents (PICs) in modulating motor neuron output, and thus neuromuscular performance, it remains unknown whether their contribution to motor neuron discharge behaviour varies throughout the day. This study aimed to determine whether PIC-related effects on motor neuron activity during submaximal dorsiflexion tasks differ between the early morning and late afternoon. Eighteen healthy adults (4 females; 27.4{+/-}5.6 years) performed triangular isometric contractions at two time-points: early morning (7:00-8:30 a.m.) and late afternoon (5:00-7:30 p.m.). Two conditions were tested: (1) a relative condition, where the target force corresponded to 40% of the maximal voluntary force (MVF) measured during that session, and (2) an absolute condition, where the target force was 40% of the MVF recorded during the first session. High-density surface electromyography signals were recorded from the tibialis anterior and decomposed into motor unit spike trains. The prolongation effect of PICs, estimated via {Delta}F, was significantly greater in the late afternoon in both the relative and absolute force conditions. The amplification effect of PICs, estimated by the acceleration phase of the discharge trajectory, was also higher in the late afternoon, but only in the relative force condition. No time-of-day differences were found for brace height, while attenuation was reduced in the late afternoon in the relative force condition. Collectively, these findings provide evidence for a diurnal modulation of the influence of PICs on motor neuron discharge behaviour, likely mediated by reduced inhibitory input in the late afternoon rather than by changes in neuromodulatory drive. Key pointsO_LI- Human neuromuscular performance fluctuates throughout the day, with higher force output typically observed in the late afternoon. C_LIO_LI- While circadian effects on peripheral determinants of muscle function (e.g., muscle- tendon properties, metabolic efficiency) are well documented, the neural mechanisms underlying these diurnal variations remain poorly understood. C_LIO_LI- Persistent inward currents (PICs) play a critical role in regulating motor neuron behaviour, yet their potential diurnal modulation has not been previously investigated. C_LIO_LI- Our results provide novel evidence for a diurnal modulation of the influence of PICs on motor neuron discharge behaviour, likely mediated by reduced inhibitory input in the late afternoon rather than by changes in neuromodulatory drive. C_LIO_LI- These findings suggest that intrinsic motor neuron excitability (and therefore spinal gain modulation) fluctuates across the day. C_LI

neuroscience↗

bioRxiv2024

Heterogeneous distribution of inhibitory inputs among motor units as a key mechanism for motor adaptations to pain

Pain significantly influences movement, yet the neural mechanisms underlying the range of observed motor adaptations remain unclear. This study combined experimental data and in silico models to investigate the contribution of inhibitory and neuromodulatory inputs to motor unit behaviour in response to nociceptive stimulation during contractions at 30% of maximal torque. Specifically, we aimed to unravel the distribution pattern of inhibitory inputs to the motor unit pool. Seventeen participants performed isometric knee extension tasks under three conditions: Control, Pain (induced by injecting hypertonic saline into the infra-patellar fat pad), and Washout. We identified large samples of motor units in the vastus lateralis (up to 53/participant) from high-density electromyographic signals, leading to three key observations. First, while motor unit discharge rates significantly decreased during Pain, a substantial proportion of motor units (14.8-24.8%) did not show this decrease and, in some cases, even exhibited an increase. Second, using complementary approaches, we found that pain did not significantly affect neuromodulation, making it unlikely to be a major contributor to the observed changes in motor unit behaviour. Third, we observed a significant reduction in the proportion of common inputs to motor units during Pain. To explore potential neurophysiological mechanisms underlying these results, we simulated the behaviour of motor unit pools with varying distribution patterns of inhibitory inputs. Our simulations support the hypothesis that a non-homogeneous distribution of inhibitory inputs, not strictly organised according to motor unit size, is a key mechanism underlying the motor response to nociceptive stimulation during moderate contraction intensity. Key pointsO_LIPain affects movement, but the neural mechanisms underlying these motor adaptations are not well defined. C_LIO_LIThe traditional view is that pain causes uniform (homogeneous) inhibition among motor units. C_LIO_LIRecent research has observed differential motor unit responses to experimental pain - some with decreased discharge rates and others with increased discharge rates. C_LIO_LICombining experimental data with modelling, we provide compelling evidence of increased inhibition that is non-uniformly distributed across motor units, regardless of their size. C_LI Legend of the abstract figureWe combined experimental data and in silico models to investigate the contribution of inhibitory and neuromodulatory inputs to motor unit behaviour in response to nociceptive stimulation during submaximal isometric contractions at 30% of maximal voluntary contraction. We identified large samples of motor units in the vastus lateralis, leading to three key observations. First, while motor unit discharge rates significantly decreased during Pain, a substantial proportion of motor units did not show this decrease and, in some cases, even exhibited an increase. Second, using complementary approaches, we found that pain did not significantly affect neuromodulation, making it unlikely to be a major contributor to the observed changes in motor unit behaviour. Third, we observed a significant reduction in the proportion of common inputs to motor units during Pain. Together with our simulations, these results provide evidence of increased inhibition that is non-uniformly distributed across motor units, regardless of their size. Ppp, pulses per second; MVC, maximal voluntary contraction.

neuroscience↗

bioRxiv2024

Flexible Control of Motor Units: Is the Multidimensionality of Motor Unit Manifolds a Sufficient Condition?

Understanding flexibility in the neural control of movement requires identifying the distribution of common inputs to the motor units. In this study, we identified large samples of motor units from two lower limb muscles: the vastus lateralis (VL; up to 60 motor units/participant) and the gastrocnemius medialis (GM; up to 67 motor units/participant). First, we applied a linear dimensionality reduction method to assess the dimensionality of the manifolds underlying the motor unit activity. We subsequently investigated the flexibility in motor unit control under two conditions: sinusoidal contractions with torque feedback, and online control with visual feedback on motor unit firing rates. Overall, we found that the activity of GM motor units was effectively captured by a single latent factor defining a unidimensional manifold, whereas the VL motor units were better represented by three latent factors defining a multidimensional manifold. Despite this difference in dimensionality, the recruitment of motor units in the two muscles exhibited similarly low levels of flexibility. Using a spiking network model, we tested the hypothesis that dimensionality derived from factorization does not solely represent descending cortical commands but is also influenced by spinal circuitry. We demonstrated that a heterogeneous distribution of inputs to motor units, or specific configurations of recurrent inhibitory circuits, could produce a multidimensional manifold. This study clarifies an important debated issue, demonstrating that while motor unit firings of a non-compartmentalised muscle can lie in a multidimensional manifold, the central nervous system may still have limited capacity for flexible control of these units. Key pointsO_LITo generate movement, the central nervous system distributes both excitatory and inhibitory inputs to the motor units. C_LIO_LIThe level of flexibility in the neural control of these motor units remains a topic of debate with significant implications for identifying the smallest unit of movement control. C_LIO_LIBy combining experimental data and in silico models, we demonstrated that the activity of a large sample of motor units from a single muscle can be represented by a multidimensional linear manifold; however, these units show very limited flexibility in their recruitment. C_LIO_LIThe dimensionality of the linear manifold may not directly reflect the dimensionality of descending inputs but could instead relate to the organisation of local spinal circuits. C_LI

neuroscience↗

bioRxiv2024

The volitional control of individual motor units is constrained within low-dimensional manifolds by common inputs

AbstractThe implementation of low-dimensional movement control by the central nervous system has been debated for decades. In this study, we investigated the dimensionality of the control signals received by spinal motor neurons when controlling either the ankle or knee joint torque. We first identified the low-dimensional latent factors underlying motor unit activity during torque- matched isometric contractions. Subsequently, we evaluated the extent to which motor units could be independently controlled. To this aim, we used an online control paradigm in which participants received the corresponding motor unit firing rates as visual feedback. We identified two main latent factors, regardless of the muscle group (vastus lateralis-medialis and gastrocnemius lateralis-medialis). The motor units of the gastrocnemius lateralis could be controlled largely independently from those of the gastrocnemius medialis during ankle plantarflexion. This dissociation of motor unit activity imposed similar behavior to the motor units that were not displayed in the feedback. Conversely, it was not possible to dissociate the activity of the motor units between the vastus lateralis and medialis muscles during the knee extension tasks. These results demonstrate that the number of latent factors estimated from linear dimensionality reduction algorithms does not necessarily reflect the dimensionality of volitional control of motor units. Overall, individual motor units were never controlled independently of all others but rather belonged to synergistic groups. Together, these findings provide evidence for a low-dimensional control of motor units constrained by common inputs, with notable differences between muscle groups. Significance statementIn this study, we initially examined the latent factors underlying motor unit activity in the vastii or gastrocnemii muscles during torque-matched isometric contractions. We then explored the extent to which these motor units could be controlled independently, using an online control paradigm where participants received visual information on motor unit firing rates. Although participants were able to dissociate the activity of a few motor unit pairs (from the gastrocnemius medialis and lateralis muscles), our results provide direct evidence of a low- dimensional control, constrained by common inputs, limiting flexibility in motor unit recruitment. Furthermore, we show that the number of latent factors identified by dimensionality reduction algorithms does not necessarily reflect the dimensionality of volitional control of motor units.

neuroscience↗

bioRxiv2023

The decoding of extensive samples of motor units in human muscles reveals the rate coding of entire motoneuron pools

Movements are performed by motoneurons transforming synaptic inputs into an activation signal that controls muscle force. The control signal emerges from interactions between ionotropic and neuromodulatory inputs to motoneurons. Critically, these interactions vary across motoneuron pools and differ between muscles. To provide the most comprehensive framework to date of motor unit activity during isometric contractions, we identified the firing activity of extensive samples of motor units in the Tibialis Anterior (129{+/-}44 per participant; n=8) and the Vastus Lateralis (130{+/-}63 per participant; n=8) muscles during isometric contractions of up to 80% of maximal force. From this unique dataset, the rate coding of each motor unit was characterised as the relation between its instantaneous firing rate and the applied force, with the assumption that the linear increase in isometric force reflects a proportional increase in the net synaptic excitatory inputs received by the motoneuron. This relation was characterised with a natural logarithm function that comprised two stages. The initial stage was marked by a steep acceleration of firing rate, which was greater for low- than medium- and high-threshold motor units. The second stage comprised a linear increase in firing rate, which was greater for high- than medium- and low-threshold motor units. Changes in firing rate were largely non-linear during the ramp-up and ramp-down phases of the task, but with significant prolonged firing activity only evident for medium-threshold motor units. Contrary to what is usually assumed, our results demonstrate that the firing rate of each motor unit can follow a large variety of trends with force across the pool. From a neural control perspective, these findings indicate how motor unit pools use gain control to transform inputs with limited bandwidths into an intended muscle force.

neuroscience↗

bioRxiv2023

Tutorial on MUedit: An open-source software for identifying and analysing the discharge timing of motor units from electromyographic signals

We introduce the open-source software MUedit and we describe its use for identifying the discharge timing of motor units from all types of electromyographic (EMG) signals recorded with multi-channel systems. MUedit performs EMG decomposition using a blind-source separation approach. Following this, users can display the estimated motor unit pulse trains and inspect the accuracy of the automatic detection of discharge times. When necessary, users can correct the automatic detection of discharge times and recalculate the motor unit pulse train with an updated separation vector. Here, we provide an open-source software and a tutorial that guides the user through i) the parameters and steps of the decomposition algorithm, and ii) the manual editing of motor unit pulse trains. Further, we provide simulated and experimental EMG signals recorded with grids of surface electrodes and intramuscular electrode arrays to benchmark the performance of MUedit. Finally, we discuss advantages and limitations of the blind-source separation approach for the study of motor unit behaviour during muscle contractions in humans.

neuroscience↗

bioRxiv2023

Estimates of persistent inward currents in lower limb muscles are not different between inactive, resistance-trained and endurance-trained young individuals

Persistent inward currents (PICs) increase the intrinsic excitability of -motoneurons. The main objective of this study was to determine whether estimates of -motoneuronal PIC magnitude is influenced by chronic endurance and resistance training. We also aimed to investigate whether there is a relationship in the estimates of -motoneuronal PIC magnitude between muscles. Estimates of PIC magnitude were obtained in three groups of young individuals: resistance-trained (n=12), endurance-trained (n=12), and inactive (n=13). We recorded high-density surface electromyography (HDsEMG) signals from tibialis anterior, gastrocnemius medialis, soleus, vastus medialis, and vastus lateralis. Then, signals were decomposed with convolutive blind source separation to identify motor units spike trains. Participants performed triangular isometric contractions to a peak of 20% of their maximum voluntary contraction. A paired-motor-unit analysis was used to calculate {Delta}F, which is assumed to be proportional to PIC magnitude. Despite the substantial differences in physical training experience between groups, we found no differences in {Delta}F, regardless of the muscle. Significant correlations of estimates of PICs magnitude were found between muscles of the same group (VL-VM, SOL-GM). Only one correlation (out of 8) between muscles of different groups was found (GM and TA). Overall, our findings suggest that estimates of PIC magnitude in the lower limb muscles are not influenced by physical training experience in healthy young individuals. They also suggest muscle-specific and muscle group-specific regulations of the diffuse monoamine inputs.

neuroscience↗

bioRxiv2023

I-Spin live: An open-source software based on blind-source separation for decoding the activity of spinal alpha motor neurons in real-time

Decoding the activity of individual neural cells during natural behaviours allows neuroscientists to study how the nervous system generates and controls movements. Contrary to other neural cells, the activity of spinal motor neurons can be determined non-invasively (or minimally invasively) from the decomposition of electromyographic (EMG) signals into motor unit firing activities. For some interfacing and neuro-feedback investigations, EMG decomposition needs to be performed in real-time. Here, we introduce an open-source software that performs real-time decoding of motor neurons using a blind-source separation approach for multichannel EMG signal processing. Separation vectors (motor unit filters) are optimised for each motor unit from baseline contractions and then re-applied in real-time during test contractions. In this way, the firing activity of multiple motor neurons can be provided through different forms of visual feedback. We provide a complete framework with guidelines and examples of recordings to guide researchers who aim to study movement control at the motor neuron level. We first validated the software with synthetic EMG signals generated during a range of isometric contraction patterns. We then tested the software on data collected using either surface or intramuscular electrode arrays from five lower limb muscles (gastrocnemius lateralis and medialis, vastus lateralis and medialis, and tibialis anterior). We assessed how the muscle or variation of contraction intensity between the baseline contraction and the test contraction impacted the accuracy of the real-time decomposition. This open-source software provides a set of tools for neuroscientists to design experimental paradigms where participants can receive real-time feedback on the output of the spinal cord circuits.

neuroscience↗

bioRxiv2023

A graph-based approach to identify motor neuron synergies

Multiple studies have experimentally observed common fluctuations in the discharge rates of spinal motor neurons, which have been classically interpreted as generated by correlated synaptic inputs. However, so far it has not been possible to identify the number of inputs, nor their relative strength, received by each motor neuron. This information would reveal the distribution of inputs and dimensionality of the neural control of movement at the motor neuron level. Here, we propose a method that generates networks of correlation between motor neuron outputs to estimate the number of common inputs to motor neurons and their relative strengths. The method is based on force-directed graphs, the hierarchical clustering of motor neurons in the graphs, and the estimation of input strengths based on the graph structure. To evaluate the accuracy and robustness of the method, we simulated 100 motor neurons driven by a known number of inputs with fixed weights. The simulation results showed that 99.2 {+/-} 0.6%, 94.3 {+/-} 2.2 %, and 95.1 {+/-} 2.7 % of the motor neurons were accurately assigned to the input source with the highest weight for simulations with 2, 3, and 4 inputs, respectively. Moreover, the normalised weigths (range 0 to 1) with which each input was transmitted to individual motor neurons were estimated with a root-mean-squared error of 0.11, 0.18, and 0.28 for simulations with 2, 3, and 4 inputs, respectively. These results were robust to errors introduced in the discharge times (as they may occur due to errors by decomposition algorithms), with up to 5% of missing spikes or false positives. We finally applied this method on various experimental datasets to demonstrate typical case scenario when studying the neural control of movement. Overall, these results show that the proposed graph-based method accurately describes the distribution of inputs across motor neurons. Authors summaryAn important characteristics for our understanding of the neural control of natural behaviors if the dimensionality in neural control signals to the musculoskeletal system. This dimensionality in turn depends on the number of synaptic inputs transmitted to the elementary units of this control, i.e., the spinal motor neurons, and on their correlation. We propose a graph-based approach applied to the discharge times of motor neurons to estimate the number of inputs and associated strength transmitted to each motor neuron. For this purpose, we first calculated the correlation between motor neuron smoothed discharge rates, assuming that correlated discharge rates result from the reception of a correlated inputs. Then, we derived networks/graphs in which each node represented a motor neuron and where the nodes were positioned close to each or further apart, depending on the level of correlated activities of the corresponding motor neurons. Using simulations of motor neuron behaviour, we showed that the spatial information embedded in the proposed graphs can be used to accurately estimate the number and the relative strengths of the inputs received by each motor neurons. This method allows to reconstruct the distribution of synaptic inputs to motor neurons from the observed motor neuron activity.

physiology↗

bioRxiv2022

Two motor neuron synergies, invariant across ankle joint angles, activate the triceps surae during plantarflexion

Recent studies have suggested that the nervous system generates movements by controlling groups of motor neurons (synergies) that do not always align with muscle anatomy. In this study, we determined whether these synergies are robust across tasks with different mechanical constraints. We identified motor neuron synergies using principal component analysis (PCA) and cross-correlations between smoothed discharge rates of motor neurons. In Part 1, we used simulations to validate these methods. The results suggested that PCA can accurately identify the number of common inputs and their distribution across active motor neurons. Moreover, the results confirmed that cross-correlation can separate pairs of motor neurons that receive common inputs from those that do not receive common inputs. In Part 2, sixteen individuals performed plantarflexion at three ankle angles while we recorded electromyographic signals from the gastrocnemius lateralis (GL) and medialis (GM) and the soleus (SOL) with grids of surface electrodes. PCA revealed two motor neuron synergies. These motor neuron synergies were relatively stable with no significant differences in the distribution of motor neuron weights across ankle angles (p=0.62). When the cross-correlation was calculated for pairs of motor units tracked across ankle angles, we observed that only 13.0% of pairs of motor units from GL and GM exhibited significant correlations of their smoothed discharge rates across angles, confirming the low level of common inputs between these muscles. Overall, these results highlight the modularity of movement control at the motor neuron level, suggesting a sensible reduction of computational resources for movement control. Key points summaryO_LIThe central nervous system may generate movements by activating groups of motor neurons (synergies) with common inputs. C_LIO_LIWe show here that two main sources of common inputs drive the motor neurons innervating the triceps surae muscles during isometric ankle plantarflexions. C_LIO_LIWe report that the distribution of these common inputs is globally invariant despite changing the mechanical constraints of the tasks, i.e., the ankle angle. C_LIO_LIThese results suggest the functional relevance of the modular organization of the central nervous system to control movements. C_LI

neuroscience↗

bioRxiv2022

Toward a generalizable deep CNN for neural drive estimation across muscles and participants

High-density electromyography (HD-EMG) decomposition algorithms are used to identify individual motor unit spike trains, which collectively constitute the neural code of movements, to predict motor intent. This approach has advanced from offline to online decomposition, from isometric to dynamic contractions, leading to a wide range of neural-machine interface applications. However, current online methods need offline retraining when applied to the same muscle on a different day or to a different person, which limits their applications in a real-time neural-machine interface. We proposed a deep convolutional neural network (CNN) framework for neural drive estimation, which captures general spatiotemporal properties of motor unit action potentials to generalize its application without retraining to HD-EMG data recorded in separate sessions, muscles, and participants. We recorded HD-EMG signals from the vastus medialis and vastus lateralis muscles while participants performed isometric contractions during two sessions separated by approximately 20 months. We identified motor unit spike trains from HD-EMG signals using a blind source separation (BSS) method, and then used the cumulative spike train (CST) of these motor units and the HD-EMG signals to train and validate the deep CNN. On average, the correlation coefficients between CST from BSS and that from deep CNN were 0.977{+/-}0.007 for leave-one-out across-sessions-and-muscles validation and 0.985{+/-}0.005 for leave-one-out across-participants validation. When trained with more than four datasets, the performance of deep CNN saturated at 0.979{+/-}0.001 for cross validations across muscles, sessions, and participants. Therefore, we can conclude that the deep CNN is generalizable across the afore-mentioned conditions without retraining. We could potentially generate a robust deep CNN to estimate neural drive to muscles for neural-machine interfaces.

neuroscience↗

bioRxiv2022

Handedness is associated with less common input to spinal motor neurons innervating different hand muscles

Whether the neural control of manual behaviours differs between the dominant and non-dominant hand is poorly understood. This study aimed to determine whether the level of common synaptic input to motor neurons innervating the same or different muscles differs between the dominant and the non-dominant hand. Seventeen participants performed two motor tasks with distinct mechanical requirements: an isometric pinch and an isometric rotation of a pinched dial. Each task was performed at 30% of maximum effort and was repeated with the dominant and non-dominant hand. Motor units were identified from two intrinsic (flexor digitorum interosseous and thenar) and one extrinsic muscle (flexor digitorum superficialis) from high-density surface electromyography recordings. Two complementary approaches were used to estimate common synaptic inputs. First, we calculated the coherence between groups of motor neurons from the same and from different muscles. Then, we estimated the common input for all pairs of motor neurons by correlating the low-frequency oscillations of their discharge rate. Both analyses led to the same conclusion, indicating less common synaptic input between motor neurons innervating different muscles in the dominant hand than in the non-dominant hand, which was only observed during the isometric rotation task. No differences in common input were observed between motor neurons of the same muscle. This lower level of common input could confer higher flexibility in the recruitment of motor units, and therefore, in mechanical outputs. Whether this difference between the dominant and non-dominant arm is the cause or the consequence of handedness remains to be determined. Key points- How the neural control of manual behaviours differs between the dominant and non-dominant hand remains poorly understood. - We decoded the spiking activities of spinal motor neurons innervating one extrinsic and two intrinsic hand muscles during isometric tasks. - We estimated the common synaptic input to motor neurons innervating the same or different muscles. - There is less common synaptic input between motor neurons innervating different muscles in the dominant than in the non-dominant hand during isometric rotation tasks. - No differences in common input were observed between motor neurons of the same muscle. - Lower level of common input could confer higher flexibility in the recruitment of motor units.

neuroscience↗

bioRxiv2022

Common synaptic inputs are not distributed homogeneously among the motor neurons that innervate synergistic muscles

The force generated by the muscles involved in an action is produced by common synaptic inputs received by the engaged motor neurons. The purpose of our study was to identify the low-dimensional latent components, defined hereafter as neural modules, underlying the discharge rates of the motor units from two knee extensors (vastus medialis and lateralis) and two hand muscles (index and thumb muscles) during isometric contractions. The neural modules were extracted by factor analysis from the pooled motor units and no assumptions were made regarding the orthogonality of the modules or the association between the modules and each muscle. Factor analysis identified two independent neural modules that captured most of the covariance in the discharge rates of the motor units in the synergistic muscles. Although the neural modules were strongly correlated with the discharge rates of motor units in each of the synergistic pair of muscles, not all motor units in a muscle were correlated with the neural module for that muscle. The distribution of motor units across the pair of neural modules differed for each muscle: 80% of the motor units in first dorsal interosseous were more strongly correlated with the neural module for that muscle, whereas the proportion was 70%, 60%, and 45% for the thenar, vastus medialis, and vastus lateralis muscles. All other motor units either belonged to both modules or to the module for the other muscle (15% for vastus lateralis). Based on a simulation of 480 integrate-and-fire neurons receiving independent and common inputs, we demonstrate that factor analysis identifies the three neural modules with high levels of accuracy. Our results indicate that the correlated discharge rates of motor units arise from at least two sources of common synaptic input that are not distributed homogeneously among the motor neurons innervating synergistic muscles.

neuroscience↗

bioRxiv2022

A deep CNN framework for neural drive estimation from HD-EMG across contraction intensities and joint angles

ObjectivePrevious studies have demonstrated promising results in estimating the neural drive to muscles, the net output of all motoneurons that innervate the muscle, using high-density electromyography (HD-EMG) for the purpose of interfacing with assistive technologies. Despite the high estimation accuracy, current methods based on neural networks need to be trained with specific motor unit action potential (MUAP) shapes updated for each condition (i.e., varying muscle contraction intensities or joint angles). This preliminary step dramatically limits the potential generalization of these algorithms across tasks. We propose a novel approach to estimate the neural drive using a deep convolutional neural network (CNN), which can identify the cumulative spike train (CST) through general features of MUAPs from a pool of motor units. MethodsWe recorded HD-EMG signals from the gastrocnemius medialis muscle under three isometric contraction scenarios: 1) trapezoidal contraction tasks with different intensities, 2) contraction tasks with a trapezoidal or sinusoidal torque target, and 3) trapezoidal contraction tasks at different ankle angles. We applied a convolutive blind source separation (BSS) method to decompose HD-EMG signals to CST and segmented both signals into windows to train and validate the deep CNN. Then, we optimized the structure of the deep CNN and validated its generalizability across contraction tasks within each scenario. ResultsWith the optimal configuration for the HD-EMG data window (overlap of 20 data points and window length of 40 data points), the deep CNN estimated the CST close to that from BSS, with a correlation coefficient higher than 0.96 and normalized root-mean-square-error lower than 7% with respect to the BSS (golden standard) within each scenario. ConclusionThe proposed deep CNN framework can utilize data from different contraction tasks (e.g., different intensities), learn general features of MUAP variants, and estimate the neural drive for other contraction tasks. SignificanceWith the proposed deep CNN, we could potentially build a neuraldrive-based human-machine interface that is generalizable to different contraction tasks without retraining.

bioengineering↗