Decoding single motor unit (MU) spiking activity, that is, action potentials discharged by motor neurons and their innervated muscle fibres, non-invasively from the surface electromyogram (EMG), represents a viable alternative to invasive brain recordings for neural human–machine interfaces (Holobar and Farina, 2021; Formento et al., 2021; Farina and Holobar, 2015). One potential application for such non-invasive neural interfaces is to augment the number of degrees of freedom a person can control by exploiting the fact that hundreds to thousands of motor neurons innervate muscles (Heckman and Enoka, 2012). If single MUs could be individually controlled, they could potentially be activated in a multitude of ways, leading to an enormous potential for additional information transfer. New MU activation patterns made of independent control of units could then be used to provide the basis for augmented control signals without impeding the original function of the innervated muscles for natural limb control (Eden et al., 2022).

The possibility of controlling part of the MUs in a muscle independently would indeed provide a separation from the natural control of force, mainly provided by the population behaviour, and a vast augmentation resource from independent control. Nonetheless, while being a very attractive potential mechanism for artificial augmentation, the independent control of MUs would imply an increased computational load by the central nervous system (CNS) without any known natural functional benefit for the human motor system. Indeed, most previous observations indicate that, contrary to independent control, single MUs tend to be activated in a very stereotyped way which is determined by the common input received by functional groups of MUs (Negro et al., 2016b; Tanzarella et al., 2021; Farina et al., 2014) and by their biophysical properties (i.e. Henneman’s size principle stating a dependency between the exerted force at which an MU starts to contribute and the neuron’s size; Henneman, 1957; Henneman et al., 1974). Since the size of the soma is inversely related to the membrane resistance, smaller motor neurons discharge action potentials earlier and at higher rates than larger neurons for the same net excitatory synaptic input (Stein et al., 2005). For this reason, if a pool of MUs receives the same common input, MU recruitment is solely dependent on the MU anatomy and on the intrinsic excitability of the motor neurons. The size principle has been observed in several muscles (Desmedt and Godaux, 1977a; Oya et al., 2009; van Zuylen et al., 1988; Thomas et al., 1987; Thomas et al., 1986; Monster and Chan, 1977) and appears to remain robust in various scenarios (Thomas et al., 1987; Thomas et al., 1986; Adam and De Luca, 2003; Fling et al., 2009; Desmedt and Godaux, 1977b; Thomas et al., 1978; Jones et al., 1994).

Previous works have tried to challenge the perspective of single MUs only being activated in a predetermined fashion (Formento et al., 2021; Basmajian, 1963). A recent study even provided evidence that indeed there is, to some degree, a neural substrate that would allow for the selective cortical control of MUs via descending pathways (Marshall et al., 2021). However, so far, it is unclear whether humans can learn to leverage such a potential neural structure for selectively activating MUs by converting the common neural input received by an MU pool to independent inputs to individual MUs and thus change the original MU recruitment.

This study examined whether humans could control pairs of MUs innervating the same muscle flexibly. Further, it addressed how this potential ability depends on the similarity of the MU pairs in size or, equivalently, in recruitment threshold. For this purpose, we used a neural interface that provided subjects with biofeedback on the activity of individual MUs (Barsakcioglu et al., 2021). Subjects were encouraged to navigate a cursor inside a 2D space into different targets as quickly as possible by selectively recruiting different MUs using feedback received on both MU activity and resulting cursor movement. This allowed us to assess if subjects were able to leverage potential selective descending pathways that would facilitate independent synaptic input to individual MUs or if, instead, they used control strategies based on a common input to the MU pair. After several days of training, all subjects achieved the target-tracking task. However, the control strategies used that allowed individual MU activation did not leverage potential selective inputs to single MUs. Instead, subjects strongly favoured control strategies based on a common input signal combined with changes in intrinsic motor neuron excitability due to history-dependent physiological properties of the activated MUs.

Volitional and flexible control of single MUs could revolutionise neural-interface applications. Here, we used real-time biofeedback on single MU activity to encourage subjects to learn independent control of pairs of MUs. Our results showed that subjects could gain control over four MUs from a single muscle. The control strategies that emerged, allowing for selective MU control, were limited by the presence of a common input to the MU pool. Therefore, subjects did not exploit potential neural structures with selective inputs to individual MUs.

In this study, the identified MU pools were ranked based on their recruitment order. We have shown that the de-recruitment threshold does not always match the recruitment one. While small deviations between recruitment and de-recruitment threshold may be explained by a slight increase in co-contraction during the declining force ramp (De Luca and Mambrito, 1987), larger mismatches might be due to intrinsic neuromodulatory and inhibitory mechanisms. The latter are triggered following the activation of an MU, which can disrupt the simple dependency of recruitment order from neuron size and input received (Heckman and Enoka, 2012). One potential mechanism contributing to this neuromodulation in motor neurons is persistent inward current (PIC) (Binder et al., 2020). PICs, that is, descending neuromodulation from the brainstem to the spinal cord, can alter the excitability of motor neurons which may lead to a self-sustained firing of action potentials and thus to de-recruitment at a force level lower than the recruitment threshold, that is, recruitment de-recruitment hysteresis (Binder et al., 2020). In this study, we examined the behaviour of pairs of MUs at very low forces, that is, less than 10% MVC. These low-threshold MUs active at such force levels may be more prone to the effect of sustained PICs (Heckman et al., 2008b), which might explain the hysteresis between recruitment and de-recruitment threshold observed for most MUs examined in this experiment.

An inhibitory input is needed to extinguish the impact of PICs on the MU discharge behaviour (Heckmann et al., 2005). Such an inhibitory signal leads to the reversal of MU activity to the initial state, once an MU stops firing for a prolonged amount of time. This explains why MU recruitment was not critically altered even after extensive single MU control during this experiment. The very rare cases in which an activation of the higher threshold MU before the lower threshold one occurred may be explained by an incomplete extinction of the PIC effect since in the vast majority of these cases (>90%) the preceding trial required an activation of MU2.

PICs act diffuse across the entire MU pool (Johnson and Heckman, 2014). Therefore, the excitability of MU1 and MU2 should be affected similarly by this neuromodulatory mechanism. However, it may be possible that the broad influence of PICs is focalised by local inhibition, which allows suppression of the PIC effect on selective MU pools and sub-pools (Heckman et al., 2008a). Inhibitory signals to motor neurons can be facilitated via various mechanisms, including descending pathways, spinal circuitries, or pre-synaptic inhibition (Heckman et al., 2009). One intensively studied form of inhibition directly influencing PICs is reciprocal inhibition (Johnson and Heckman, 2014; Heckman et al., 2008a; Hultborn et al., 2004), triggered by changes in the length of antagonist muscles, which can occur, for example, by moderately altering joint angles (Hyngstrom et al., 2007). Further, descending inputs may additionally tune interneurons mediating reciprocal inhibition and thus also influence PICs and the excitability of single motor neurons (Jankowska, 1992). Synaptic inhibition received by motor neurons follows the size principle. Hence, smaller motor neurons tend to show a greater hyperpolarisation than larger ones for the same inhibitory input (Henneman et al., 1965; Heckman and Binder, 1988). Consequently, a higher threshold MU may continue firing after a smaller threshold MU already became silent. In our experiment, we did not measure PICs nor inhibitory inputs to MUs. Hence, we cannot pinpoint the exact mechanisms explored by subjects to generate the necessary activation patterns to move the cursor towards TIII. However, we consider that it is likely that a mixture of broadly distributed neuromodulatory mechanisms, such as PICs, and locally acting inhibition may have produced the variability in MU activity that was observed in this study. Similar findings on the increasing variability in MU activation by incorporating neuromodulation and inhibition have been shown in a simulation study (Powers et al., 2012).

During a progressive increase in force, recruitment depends only on the MU anatomy and the input received. If humans can learn to leverage potential structures in the CNS that allow selective inputs to MUs (Marshall et al., 2021), changes in the recruitment order during this initial phase should be expected. However, it is important to underline that a conclusion of flexible control based on changes in MU recruitment cannot be drawn for time intervals that follow an activation of the MU. In these cases, the de-recruitment of an MU at a force level different from the recruitment threshold could be incorrectly interpreted as an alteration of the recruitment order. Presumably, such changes result from the relative intrinsic excitability of the motor neurons which override the sole impact of the received synaptic input on the recruitment order. Therefore, a direct proof of altered MU recruitment as a consequence of independent input to different MUs needs to be provided during the initial activation phase, that is, an MU with higher recruitment threshold activated before an MU with lower threshold without preceding activations. It is also worth mentioning that this proof should further include MUs with sufficiently different recruitment thresholds since synaptic noise may influence the relative recruitment order for MUs of very similar thresholds (Heckman and Enoka, 2012).

We did not get results supporting a general flexible control of MUs, that is, volitional activation of higher threshold MUs before lower threshold ones at initial recruitment. However, flexible control of individual MUs could still be a framework explaining how subjects were able to reach the different targets in the 2D space. If this was the case then, since control would be achieved only after MUs were recruited, this would imply that flexible control is state-dependent: it can only be achieved in the context of previously contracted muscle fibres. Such state dependency restricts the possible neural strategies that could allow flexible control of MUs. One possible strategy consistent with such state-dependent control of individual MUs could be relying on an input signal to MUs not directly linked to motor function and non-homogeneously distributed among the MU pool. For example, cortical oscillations could meet these criteria if descending projections of this activity to large and small MUs in a pool differed (Ibáñez et al., 2021; Bräcklein et al., 2021; Bräcklein et al., 2022). A different alternative that may allow for flexible MU control after initial recruitment is volitional modulation of reciprocal inhibition (Chen et al., 2006; Thompson et al., 2013) presumably via direct descending commands (Jankowska, 1992; Nielsen et al., 1995). Such augmentation of synaptic inhibition could potentially lead to changes in the excitability of specific MUs. Such inputs to MUs could provide a certain degree of flexibility to control subgroups of MUs in a muscle volitionally. Future studies are needed to test this hypothesis.

Throughout the 14 days of training, subjects were asked to modulate the discharge rate of MU pairs independently to navigate a cursor as quickly as possible into different targets inside a 2D space. The target and angle hit rates indicate that subjects could achieve control over these single MUs. Moreover, the angle hit analysis (e.g. see Figure 5A) revealed that during the last 5 days of training, subjects were able to produce the necessary activation pattern in most trials despite different properties of the selected MUs. Although subjects repeated the target-tracking task with different sets of MUs every day, they consistently reported the use of the same control strategies across days. Hence, these findings suggest that subjects learned to establish universal control strategies that allowed for the achievement of the target task for various combinations of MUs. To reach TI or TII, subjects used precise force control to exert either a low-enough force that only the low threshold MU (MU1) turned active (TI) or a force above the recruitment threshold of both MU1 and MU2 (TII). To reach TII, subjects used strategies that corresponded to a physiological activation of the two MUs similar to the initial force ramp.

As could be expected, subjects had difficulties in navigating the cursor towards TI when the difference in recruitment threshold between MU1 and MU2 was small (≤1% MVC; see Figure 3D). In these cases, subjects could not selectively activate MU1 while keeping MU2 deactivated and still reach a discharge rate of MU1 high enough to place the cursor near the centre of TI. When the difference in recruitment threshold was larger (≥3% MVC), the discharge rate of MU1 reached the desired value before the onset of MU2. It seems the discharge rate of MU1 saturated (Fuglevand et al., 1993) due to its intrinsic properties, as recently discussed in Fuglevand et al., 2015. Unlike for TI and TII, TIII required a more cumbersome approach. To reach TIII, subjects mainly mimicked the trajectory of TII first, that is, activating both MUs, followed by the second stage of control in which the force level was reduced until the lower threshold MU turned off by leveraging the mismatch in the de-recruitment thresholds. In order to then place and hold the cursor inside TIII, subjects gradually increased the force again without reactivating MU1. These second and third stages of control were possible in principle by maintaining a common ionotropic input to the MU pair combined with neuromodulatory input and synaptic inhibition, as described above, even when MU2 was initially de-recruited before MU1. If a direct activation of MU2 would have been possible as it was for MU1, subjects would have chosen to mimic a cursor trajectory along the direct path from the origin to the target centre, as observed for both TI and TII (see Figure 4). This almost never occurred in the hundreds of trials tested. Therefore, the sole activation of a higher threshold MU was only possible by exploiting the history-dependent activation of MUs, that is, exclusive firing of MU2 follows the combined activation of MU1 and MU2. The results suggest that this three-stage approach to achieve a hit in TIII as quickly as possible was feasible for the subjects while a more efficient strategy of directly activating MU2 without a preceding activation of MU1 was not.

It has been previously shown that subjects can learn to control MUs independently when exposed to biofeedback on the discharge behaviour (Formento et al., 2021; Basmajian, 1963). In these previous investigations, the subjects were allowed for contractions along multiple directions. Such variations in force directions (ter Haar Romeny et al., 1984; Desnedt and Gidaux, 1981) but also other motor behavioural changes, including alternations in postures (Nardone et al., 1989), and contraction speed (Desmedt and Godaux, 1977a), potentially have an impact on the recruitment order. Similar changes in an MU pool’s discharge activity imposed by such behavioural changes were recently suggested in non-human primates (Marshall et al., 2021). Therefore, individual MU control may be triggered by small compensatory movements rather than being the result of a dedicated and volitionally controllable individual synaptic inputs. Indeed, independent control of individual MUs would imply that an MU can be controlled independently of all other MUs. The fact that a pair of MUs can be controlled independently when varying the task does not imply that the two MUs are independently controlled in absolute terms. They are simply independently controlled with respect to each other. For example, in some tasks or in some conditions, they may be part of different groups of MUs receiving two different common inputs (Tanzarella et al., 2021).

A recent study in humans provided evidence for the existence of MU pool synergies similar to the functional grouping of muscles involved in a single movement (Tanzarella et al., 2021). The CNS may send a common input to these MU pool synergies, which are not per se limited to innervating only a single muscle (Laine et al., 2015). In our experiment, we chose a simple case of an MU pool constituting a functional group during ankle dorsiflexion, that is, MUs innervating the TA. During more complex tasks, for example, movement along multiple directions, the CNS would send different common inputs to certain numbers of groups of MUs. While the synergistic organisation of MUs might be flexible across tasks, for example, movement along multiple directions, it remains yet to be explored if the input to a single functional MU group can be changed volitionally from common to individual while the performed task is maintained. Hence, it is crucial that initial conditions during MU recruitment, such as posture, contraction speed, and force direction, are kept constant when flexible MU recruitment is investigated. MU activation based on changes in behaviour, or the performed motor task, does not indicate that subjects can volitionally trigger MU activity by a selective synaptic input as it is possible for their cortical counterparts. Furthermore, these constraints need to be considered in neural-machine interface applications relying on flexible MU control. Possible extracted control signals may depend on behavioural changes and not on a designated descending control command, and therefore, may not satisfy the intended application. Similar constraints effectively apply for augmentation when the aim is to extend the degrees of freedom that a human can volitional control, that is, adding supernumerary degrees of freedom to the natural ones (Eden et al., 2022; Dominijanni et al., 2021). In such cases, if the control of a supernumerary degree of freedom is based on single MU activation, this activation must be uncoupled from motor behaviour to ensure coordination between natural and supernumerary effectors. This would correspond to breaking common input into multiple inputs. Nevertheless, even based on behavioural changes, single MU control can be a resource for specific neural-interface applications, for example, in the absence of any additional motor information or to augment a specific motor task (see Eden et al., 2022 for augmentation by transfer).

Previous studies investigating flexible MU control focused on upper-limb muscles (Formento et al., 2021; Basmajian, 1963; Marshall et al., 2021). Here, we extracted single MU activity from the TA since this muscle has properties (e.g. muscle fibres arrangement, proximity to the skin, distribution of innervation zones) that facilitate a reliable decomposition of MU activity using surface recordings (Barsakcioglu et al., 2021; Bräcklein et al., 2021; Bräcklein et al., 2022; Negro et al., 2016a; Dideriksen et al., 2018; Del Vecchio et al., 2020) and a relatively large number of monosynaptic connections with the motor cortex that could potentially be leveraged for direct independent MU control (Ibáñez et al., 2021; Dideriksen et al., 2018). Our results strongly suggest that subjects do not tend to find or opt for a control strategy that relies on flexible MU recruitment order under constraint isometric conditions. In contrast, the established control strategy seemed to be based on a common synaptic input to the MU pair combined with intrinsic changes of MU excitability due to neuromodulation and inhibition. Although the complexity of tasks in which humans use upper and lower limbs may differ, we are not aware of any evidence suggesting that the CNS employs fundamentally different strategies to orchestrate MU activity in a functional motor neuron pool differently across limbs. Future research on this topic may help to better understand if the CNS may allow for variability in upper-limb MU pools.

Throughout the experiment, subjects were restricted by an ankle dynamometer and instructed to perform dorsiflexion of the ankle only. We measured force due to dorsiflexion but not, for example, rotational forces due to ankle supination. Modest rotations of the ankle can increase the effect of inhibition in specific MUs and thus might affect their excitability (Hyngstrom et al., 2007). Such effects were only observed for modest changes in joint angles (~20°) (Hyngstrom et al., 2007), which were not possible to perform when the dynamometer restricted the foot. It may be that even very subtle rotations or slight movements of other body parts, such as the knee or hip, may have influenced the discharge activity of specific MUs (Jankowska, 1992), potentially a helpful approach for achieving TIII-instructed trials. However, none of the subjects reported the use of compensatory movements as a control strategy. Furthermore, even if the subjects unknowingly performed very subtle compensatory movements in TIII-instructed trials, they would have discovered and chosen this strategy over directly utilising potential descending pathways triggering independent MU control.

To summarise, we have demonstrated the ability to control up to four MUs from a single muscle using real-time feedback on single MU discharge behaviour. Furthermore, we have shown by operant conditioning that subjects learn concrete control strategies to recruit and de-recruit several MUs volitionally. These strategies exploit orderly recruitment in agreement with the Henneman’s size principle and a common input to their motor neurons. Conversely, the observed strategies do not leverage potential pathways that may provide selective inputs to single MUs. It is concluded that converting common input to a (synergistic) pool of motor neurons into independent input to single MUs within the same task seems extremely challenging for the CNS.

All data generated or analysed during this study are included in the manuscript.

1. 1. Adam A
   2. De Luca CJ

   (2003) Recruitment order of motor units in human vastus lateralis muscle is maintained during fatiguing contractions

   Journal of Neurophysiology 90:2919–2927.

   https://doi.org/10.1152/jn.00179.2003

   • PubMed
   • Google Scholar
2. 1. Barsakcioglu DY
   2. Bracklein M
   3. Holobar A
   4. Farina D

   (2021) Control of Spinal Motoneurons by Feedback From a Non-Invasive Real-Time Interface

   IEEE Transactions on Bio-Medical Engineering 68:926–935.

   https://doi.org/10.1109/TBME.2020.3001942

   • PubMed
   • Google Scholar
3. 1. Basmajian JV

   (1963) Control and training of individual motor units

   Science (New York, N.Y.) 141:440–441.

   https://doi.org/10.1126/science.141.3579.440

   • PubMed
   • Google Scholar
4. 1. Binder MD
   2. Powers RK
   3. Heckman CJ

   (2020) Nonlinear Input-Output Functions of Motoneurons

   Physiology (Bethesda, Md.) 35:31–39.

   https://doi.org/10.1152/physiol.00026.2019

   • PubMed
   • Google Scholar
5. 1. Bräcklein M
   2. Ibáñez J
   3. Barsakcioglu DY
   4. Farina D

   (2021) Towards human motor augmentation by voluntary decoupling beta activity in the neural drive to muscle and force production

   Journal of Neural Engineering 18:016001.

   https://doi.org/10.1088/1741-2552/abcdbf

   • PubMed
   • Google Scholar
6. 1. Bräcklein M
   2. Barsakcioglu DY
   3. Del Vecchio A
   4. Ibáñez J
   5. Farina D
   6. M. Bräcklein, D. Y. Barsakcioglu, A. Del Vecchio, J. Ibáñez, and D. Farina

   (2022) Reading and Modulating Cortical β Bursts from Motor Unit Spiking Activity

   The Journal of Neuroscience 42:3611–3621.

   https://doi.org/10.1523/JNEUROSCI.1885-21.2022

   • PubMed
   • Google Scholar
7. 1. Chen XY
   2. Chen L
   3. Chen Y
   4. Wolpaw JR

   (2006) Operant conditioning of reciprocal inhibition in rat soleus muscle

   Journal of Neurophysiology 96:2144–2150.

   https://doi.org/10.1152/jn.00253.2006

   • PubMed
   • Google Scholar
8. 1. De Luca CJ
   2. Mambrito B

   (1987) Voluntary control of motor units in human antagonist muscles: coactivation and reciprocal activation

   Journal of Neurophysiology 58:525–542.

   https://doi.org/10.1152/jn.1987.58.3.525

   • PubMed
   • Google Scholar
9. 1. Del Vecchio A
   2. Holobar A
   3. Falla D
   4. Felici F
   5. Enoka RM
   6. Farina D

   (2020) Tutorial: Analysis of motor unit discharge characteristics from high-density surface EMG signals

   Journal of Electromyography and Kinesiology 53:102426.

   https://doi.org/10.1016/j.jelekin.2020.102426

   • PubMed
   • Google Scholar
10. 1. Desmedt JE
    2. Godaux E

    (1977a) Fast motor units are not preferentially activated in rapid voluntary contractions in man

    Nature 267:717–719.

    https://doi.org/10.1038/267717a0

    • PubMed
    • Google Scholar
11. 1. Desmedt JE
    2. Godaux E

    (1977b) Ballistic contractions in man: characteristic recruitment pattern of single motor units of the tibialis anterior muscle

    The Journal of Physiology 264:673–693.

    https://doi.org/10.1113/jphysiol.1977.sp011689

    • PubMed
    • Google Scholar
12. 1. Desnedt JE
    2. Gidaux E

    (1981) Spinal Motoneuron Recruitment in Man: Rank Deordering with Direction But Not with Speed of Voluntary Movement

    Science 214:933–936.

    https://doi.org/10.1126/science.7302570

    • Google Scholar
13. 1. Dideriksen JL
    2. Negro F
    3. Falla D
    4. Kristensen SR
    5. Mrachacz-Kersting N
    6. Farina D

    (2018) Coherence of the Surface EMG and Common Synaptic Input to Motor Neurons

    Frontiers in Human Neuroscience 12:207.

    https://doi.org/10.3389/fnhum.2018.00207

    • PubMed
    • Google Scholar
14. 1. Dominijanni G
    2. Shokur S
    3. Salvietti G
    4. Buehler S
    5. Palmerini E
    6. Rossi S
    7. De Vignemont F
    8. d’Avella A
    9. Makin TR
    10. Prattichizzo D
    11. Micera S

    (2021) The neural resource allocation problem when enhancing human bodies with extra robotic limbs

    Nature Machine Intelligence 3:850–860.

    https://doi.org/10.1038/s42256-021-00398-9

    • Google Scholar
15. 1. Eden J
    2. Bräcklein M
    3. Ibáñez J
    4. Barsakcioglu DY
    5. Di Pino G
    6. Farina D
    7. Burdet E
    8. Mehring C

    (2022) Principles of human movement augmentation and the challenges in making it a reality

    Nature Communications 13:1–13.

    https://doi.org/10.1038/s41467-022-28725-7

    • PubMed
    • Google Scholar
16. 1. Farina D
    2. Negro F
    3. Dideriksen JL

    (2014) The effective neural drive to muscles is the common synaptic input to motor neurons

    The Journal of Physiology 592:3427–3441.

    https://doi.org/10.1113/jphysiol.2014.273581

    • PubMed
    • Google Scholar
17. 1. Farina D
    2. Holobar A

    (2015) Human?Machine Interfacing by Decoding the Surface Electromyogram [Life Sciences

    IEEE Signal Processing Magazine 32:115–120.

    https://doi.org/10.1109/MSP.2014.2359242

    • Google Scholar
18. 1. Fling BW
    2. Knight CA
    3. Kamen G

    (2009) Relationships between motor unit size and recruitment threshold in older adults: implications for size principle

    Experimental Brain Research 197:125–133.

    https://doi.org/10.1007/s00221-009-1898-y

    • PubMed
    • Google Scholar
19. 1. Formento E
    2. Botros P
    3. Carmena JM

    (2021) Skilled independent control of individual motor units via a non-invasive neuromuscular-machine interface

    Journal of Neural Engineering 18:35.

    https://doi.org/10.1088/1741-2552/ac35ac

    • PubMed
    • Google Scholar
20. 1. Fuglevand AJ
    2. Winter DA
    3. Patla AE

    (1993) Models of recruitment and rate coding organization in motor-unit pools

    Journal of Neurophysiology 70:2470–2488.

    https://doi.org/10.1152/jn.1993.70.6.2470

    • Google Scholar
21. 1. Fuglevand AJ
    2. Lester RA
    3. Johns RK

    (2015) Distinguishing intrinsic from extrinsic factors underlying firing rate saturation in human motor units

    Journal of Neurophysiology 113:1310–1322.

    https://doi.org/10.1152/jn.00777.2014

    • PubMed
    • Google Scholar
22. 1. Gorassini M
    2. Yang JF
    3. Siu M
    4. Bennett DJ

    (2002) Intrinsic activation of human motoneurons: reduction of motor unit recruitment thresholds by repeated contractions

    Journal of Neurophysiology 87:1859–1866.

    https://doi.org/10.1152/jn.00025.2001

    • PubMed
    • Google Scholar
23. 1. Heckman CJ
    2. Binder MD

    (1988) Analysis of effective synaptic currents generated by homonymous Ia afferent fibers in motoneurons of the cat

    Journal of Neurophysiology 60:1946–1966.

    https://doi.org/10.1152/jn.1988.60.6.1946

    • PubMed
    • Google Scholar
24. 1. Heckman CJ
    2. Hyngstrom AS
    3. Johnson MD

    (2008a) Active properties of motoneurone dendrites: diffuse descending neuromodulation, focused local inhibition

    The Journal of Physiology 586:1225–1231.

    https://doi.org/10.1113/jphysiol.2007.145078

    • PubMed
    • Google Scholar
25. 1. Heckman CJ
    2. Johnson M
    3. Mottram C
    4. Schuster J

    (2008b) Persistent inward currents in spinal motoneurons and their influence on human motoneuron firing patterns

    The Neuroscientist 14:264–275.

    https://doi.org/10.1177/1073858408314986

    • PubMed
    • Google Scholar
26. 1. Heckman CJ
    2. Mottram C
    3. Quinlan K
    4. Theiss R
    5. Schuster J

    (2009) Motoneuron excitability: the importance of neuromodulatory inputs

    Clinical Neurophysiology 120:2040–2054.

    https://doi.org/10.1016/j.clinph.2009.08.009

    • PubMed
    • Google Scholar
27. 1. Heckman CJ
    2. Enoka RM

    (2012) Motor unit

    Comprehensive Physiology 2:2629–2682.

    https://doi.org/10.1002/cphy.c100087

    • PubMed
    • Google Scholar
28. 1. Heckmann CJ
    2. Gorassini MA
    3. Bennett DJ

    (2005) Persistent inward currents in motoneuron dendrites: implications for motor output

    Muscle & Nerve 31:135–156.

    https://doi.org/10.1002/mus.20261

    • PubMed
    • Google Scholar
29. 1. Henneman E

    (1957) Relation between size of neurons and their susceptibility to discharge

    Science (New York, N.Y.) 126:1345–1347.

    https://doi.org/10.1126/science.126.3287.1345

    • PubMed
    • Google Scholar
30. 1. Henneman E
    2. Somjen G
    3. Carpenter DO

    (1965) Excitability and inhibitibility of motoneurons of different sizes

    Journal of Neurophysiology 28:599–620.

    https://doi.org/10.1152/jn.1965.28.3.599

    • PubMed
    • Google Scholar
31. 1. Henneman E
    2. Clamann HP
    3. Gillies JD
    4. Skinner RD

    (1974) Rank order of motoneurons within a pool: law of combination

    Journal of Neurophysiology 37:1338–1349.

    https://doi.org/10.1152/jn.1974.37.6.1338

    • PubMed
    • Google Scholar
32. 1. Holobar A
    2. Farina D

    (2021) Noninvasive Neural Interfacing With Wearable Muscle Sensors: Combining Convolutive Blind Source Separation Methods and Deep Learning Techniques for Neural Decoding

    IEEE Signal Processing Magazine 38:103–118.

    https://doi.org/10.1109/MSP.2021.3057051

    • Google Scholar
33. 1. Hultborn H
    2. Brownstone RB
    3. Toth TI
    4. Gossard JP

    (2004) Key mechanisms for setting the input–output gain across the motoneuron pool

    Progress in Brain Research 143:75–95.

    https://doi.org/10.1016/s0079-6123(03)43008-2

    • Google Scholar
34. 1. Hyngstrom AS
    2. Johnson MD
    3. Miller JF
    4. Heckman CJ

    (2007) Intrinsic electrical properties of spinal motoneurons vary with joint angle

    Nature Neuroscience 10:363–369.

    https://doi.org/10.1038/nn1852

    • PubMed
    • Google Scholar
35. 1. Ibáñez J
    2. Del Vecchio A
    3. Rothwell JC
    4. Baker SN
    5. Farina D

    (2021) Only the Fastest Corticospinal Fibers Contribute to β Corticomuscular Coherence

    The Journal of Neuroscience 41:4867–4879.

    https://doi.org/10.1523/JNEUROSCI.2908-20.2021

    • PubMed
    • Google Scholar
36. 1. Jankowska E

    (1992) Interneuronal relay in spinal pathways from proprioceptors

    Progress in Neurobiology 38:335–378.

    https://doi.org/10.1016/0301-0082(92)90024-9

    • PubMed
    • Google Scholar
37. 1. Johnson MD
    2. Heckman CJ

    (2014) Gain control mechanisms in spinal motoneurons

    Frontiers in Neural Circuits 8:81.

    https://doi.org/10.3389/fncir.2014.00081

    • PubMed
    • Google Scholar
38. 1. Jones KE
    2. Lyons M
    3. Bawa P
    4. Lemon RN

    (1994) Recruitment order of motoneurons during functional tasks

    Experimental Brain Research 100:503–508.

    https://doi.org/10.1007/BF02738409

    • PubMed
    • Google Scholar
39. 1. Laine CM
    2. Martinez-Valdes E
    3. Falla D
    4. Mayer F
    5. Farina D

    (2015) Motor Neuron Pools of Synergistic Thigh Muscles Share Most of Their Synaptic Input

    The Journal of Neuroscience 35:12207–12216.

    https://doi.org/10.1523/JNEUROSCI.0240-15.2015

    • PubMed
    • Google Scholar
40. Preprint

    1. Marshall NJ
    2. Glaser JI
    3. Trautmann EM
    4. Amematsro EA
    5. Perkins SM
    6. Shadlen MN
    7. Abbott LF
    8. Cunningham JP
    9. Churchland MM

    (2021) Flexible Neural Control of Motor Units

    bioRxiv.

    https://doi.org/10.1101/2021.05.05.442653

    • Google Scholar
41. 1. McGill KC
    2. Lateva ZC
    3. Marateb HR

    (2005) EMGLAB: an interactive EMG decomposition program

    Journal of Neuroscience Methods 149:121–133.

    https://doi.org/10.1016/j.jneumeth.2005.05.015

    • PubMed
    • Google Scholar
42. 1. Monster AW
    2. Chan H

    (1977) Isometric force production by motor units of extensor digitorum communis muscle in man

    Journal of Neurophysiology 40:1432–1443.

    https://doi.org/10.1152/jn.1977.40.6.1432

    • PubMed
    • Google Scholar
43. 1. Nardone A
    2. Romanò C
    3. Schieppati M

    (1989) Selective recruitment of high-threshold human motor units during voluntary isotonic lengthening of active muscles

    The Journal of Physiology 409:451–471.

    https://doi.org/10.1113/jphysiol.1989.sp017507

    • PubMed
    • Google Scholar
44. 1. Negro F
    2. Muceli S
    3. Castronovo AM
    4. Holobar A
    5. Farina D

    (2016a) Multi-channel intramuscular and surface EMG decomposition by convolutive blind source separation

    Journal of Neural Engineering 13:2.

    https://doi.org/10.1088/1741-2560/13/2/026027

    • PubMed
    • Google Scholar
45. 1. Negro F
    2. Yavuz UŞ
    3. Farina D

    (2016b) The human motor neuron pools receive a dominant slow-varying common synaptic input

    The Journal of Physiology 594:5491–5505.

    https://doi.org/10.1113/JP271748

    • PubMed
    • Google Scholar
46. 1. Nielsen J
    2. Crone C
    3. Sinkjaer T
    4. Toft E
    5. Hultborn H

    (1995) Central control of reciprocal inhibition during fictive dorsiflexion in man

    Experimental Brain Research 104:99–106.

    https://doi.org/10.1007/BF00229859

    • PubMed
    • Google Scholar
47. 1. Oya T
    2. Riek S
    3. Cresswell AG

    (2009) Recruitment and rate coding organisation for soleus motor units across entire range of voluntary isometric plantar flexions

    The Journal of Physiology 587:4737–4748.

    https://doi.org/10.1113/jphysiol.2009.175695

    • PubMed
    • Google Scholar
48. 1. Patikas D
    2. Michailidis C
    3. Bassa H
    4. Kotzamanidis C
    5. Tokmakidis S
    6. Alexiou S
    7. Koceja DM

    (2002) Electromyographic changes of agonist and antagonist calf muscles during maximum isometric induced fatigue

    International Journal of Sports Medicine 23:285–289.

    https://doi.org/10.1055/s-2002-29079

    • PubMed
    • Google Scholar
49. 1. Powers RK
    2. Elbasiouny SM
    3. Rymer WZ
    4. Heckman CJ

    (2012) Contribution of intrinsic properties and synaptic inputs to motoneuron discharge patterns: a simulation study

    Journal of Neurophysiology 107:808–823.

    https://doi.org/10.1152/jn.00510.2011

    • PubMed
    • Google Scholar
50. 1. Rochon J
    2. Gondan M
    3. Kieser M

    (2012) To test or not to test: Preliminary assessment of normality when comparing two independent samples

    BMC Medical Research Methodology 12:1–11.

    https://doi.org/10.1186/1471-2288-12-81

    • Google Scholar
51. 1. Stein RB
    2. Gossen ER
    3. Jones KE

    (2005) Neuronal variability: noise or part of the signal?

    Nature Reviews. Neuroscience 6:389–397.

    https://doi.org/10.1038/nrn1668

    • PubMed
    • Google Scholar
52. 1. Tanzarella S
    2. Muceli S
    3. Santello M
    4. Farina D

    (2021) Synergistic Organization of Neural Inputs from Spinal Motor Neurons to Extrinsic and Intrinsic Hand Muscles

    The Journal of Neuroscience 41:6878–6891.

    https://doi.org/10.1523/JNEUROSCI.0419-21.2021

    • PubMed
    • Google Scholar
53. 1. ter Haar Romeny BM
    2. van der Gon JJ
    3. Gielen CC

    (1984) Relation between location of a motor unit in the human biceps brachii and its critical firing levels for different tasks

    Experimental Neurology 85:631–650.

    https://doi.org/10.1016/0014-4886(84)90036-0

    • PubMed
    • Google Scholar
54. Book

    1. Terjung R

    (1981)

    Comprehensive Physiology

    New Jersey, United States: John Wiley & Sons, Inc.

    • Google Scholar
55. 1. Thomas JS
    2. Schmidt EM
    3. Hambrecht FT

    (1978) Facility of motor unit control during tasks defined directly in terms of unit behaviors

    Experimental Neurology 59:384–397.

    https://doi.org/10.1016/0014-4886(78)90230-3

    • PubMed
    • Google Scholar
56. 1. Thomas CK
    2. Ross BH
    3. Stein RB

    (1986) Motor-unit recruitment in human first dorsal interosseous muscle for static contractions in three different directions

    Journal of Neurophysiology 55:1017–1029.

    https://doi.org/10.1152/jn.1986.55.5.1017

    • PubMed
    • Google Scholar
57. 1. Thomas CK
    2. Ross BH
    3. Calancie B

    (1987) Human motor-unit recruitment during isometric contractions and repeated dynamic movements

    Journal of Neurophysiology 57:311–324.

    https://doi.org/10.1152/jn.1987.57.1.311

    • PubMed
    • Google Scholar
58. 1. Thompson AK
    2. Pomerantz FR
    3. Wolpaw JR

    (2013) Operant conditioning of a spinal reflex can improve locomotion after spinal cord injury in humans

    The Journal of Neuroscience 33:2365–2375.

    https://doi.org/10.1523/JNEUROSCI.3968-12.2013

    • PubMed
    • Google Scholar
59. 1. van Zuylen EJ
    2. Gielen CC
    3. Denier van der Gon JJ

    (1988) Coordination and inhomogeneous activation of human arm muscles during isometric torques

    Journal of Neurophysiology 60:1523–1548.

    https://doi.org/10.1152/jn.1988.60.5.1523

    • PubMed
    • Google Scholar

Author details

1. Mario Bräcklein

   Department of Bioengineering, Imperial College London, London, United Kingdom

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-1537-7495

2. Deren Yusuf Barsakcioglu

   Department of Bioengineering, Imperial College London, London, United Kingdom

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – review and editing

   Competing interests

   No competing interests declared

3. Jaime Ibáñez

   1. Department of Bioengineering, Imperial College London, London, United Kingdom
   2. Department of Clinical and Movement Disorders, Institute of Neurology, University College London, London, United Kingdom
   3. BSICoS, IIS Aragón, Universidad de Zaragoza, Zaragoza, Spain

   Contribution

   Conceptualization, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-8439-151X

4. Jonathan Eden

   Department of Bioengineering, Imperial College London, London, United Kingdom

   Contribution

   Conceptualization, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

5. Etienne Burdet

   Department of Bioengineering, Imperial College London, London, United Kingdom

   Contribution

   Conceptualization, Funding acquisition, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

6. Carsten Mehring

   1. Bernstein Center Freiburg, University of Freiburg, Freiburg im Breisgau, Germany
   2. Faculty of Biology, University of Freiburg, Freiburg im Breisgau, Germany

   Contribution

   Conceptualization, Funding acquisition, Methodology, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-8125-5205

7. Dario Farina

   Department of Bioengineering, Imperial College London, London, United Kingdom

   Contribution

   Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing – review and editing

   For correspondence

   d.farina@imperial.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-7883-2697

• 1,512

  Page views

• 393

  Downloads

• 11

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.