Motor Unit Recruitment for Dynamic Tasks: A Review (2009 PDF)

J Comp Physiol B (2009) 179:57–66 DOI 10.1007/s00360-008-0289-1 R EV IE W Motor unit recruitment for dynamic tasks: current understanding and future directions Emma F. Hodson-Tole · James M. Wakeling Received: 1 April 2008 / Revised: 12 June 2008 / Accepted: 19 June 2008 / Published online: 3 Ju...

J Comp Physiol B (2009) 179:57–66 DOI 10.1007/s00360-008-0289-1 R EV IE W Motor unit recruitment for dynamic tasks: current understanding and future directions Emma F. Hodson-Tole · James M. Wakeling Received: 1 April 2008 / Revised: 12 June 2008 / Accepted: 19 June 2008 / Published online: 3 July 2008 © Springer-Verlag 2008 Abstract Skeletal muscle contains many muscle Wbres in particular the recent wavelet-analysis approaches that that are functionally grouped into motor units. For any have allowed motor unit recruitment to be assessed during motor task there are many possible combinations of motor natural movements. Directions for future studies into motor units that could be recruited and it has been proposed that a recruitment within and between functional task groups and simple rule, the ‘size principle’, governs the selection of muscle compartments are suggested. motor units recruited for diVerent contractions. Motor units can be characterised by their diVerent contractile, energetic Keywords Neuromechanics · Electromyography · and fatigue properties and it is important that the selection Skeletal muscle of motor units recruited for given movements allows units with the appropriate properties to be activated. Here we review what is currently understood about motor unit Introduction recruitment patterns, and assess how diVerent recruitment patterns are more or less appropriate for diVerent move- The majority of mammalian muscles are composed of a ment tasks. During natural movements the motor unit mixture of muscle Wbre types, with a continuum of intrinsic recruitment patterns vary (not always holding to the size properties existing within each muscle (Bottinelli et al. principle) and it is proposed that motor unit recruitment is 1994a, b). The extrafusal muscle Wbres receive their nerve likely related to the mechanical function of the muscles. supply via
-motoneurons, with each
-motoneuron inner- Many factors such as mechanics, sensory feedback, and vating a number of muscle Wbres. Force production results central control inXuence recruitment patterns and conse- from a series of electrochemical processes, initiated in a quently an integrative approach (rather than reductionist) is muscle Wbre by the Wring of its associated
-motoneuron. In required to understand how recruitment is controlled during many studies of skeletal muscle the basic functional unit of diVerent movement tasks. Currently, the best way to muscle is considered to be the motor unit, consisting of an achieve this is through in vivo studies that relate recruit-
-motoneuron and all the muscle Wbres it innervates (Sher- ment to mechanics and behaviour. Various methods for rington 1929). A given level of sub-maximal stress and determining motor unit recruitment patterns are discussed, strain, generated over a speciWc time period, could be achieved by activating many diVerent combinations of motor units. The degrees-of-freedom within the system Communicated by I. D. Hume. suggest that there is likely to be one or more simplifying E. F. Hodson-Tole strategies that ensure the correct stress and strain are gener- School of Applied Physiology, ated over an appropriate time period for each given motor Georgia Institute of Technology, Atlanta, GE, USA task, without overloading the central nervous system. Iden- tifying recruitment strategies and determining their func- J. M. Wakeling (&) School of Kinesiology, Simon Fraser University, tional signiWcance has proved challenging. Technological Burnaby, BC, Canada advancements and modern analytical techniques do, how- e-mail: wakeling@sfu.ca ever, mean that current understanding of this topic is 123 58 J Comp Physiol B (2009) 179:57–66 improving. Here, we provide an overview of the patterns of proportional to the level of force at which the motor unit motor unit recruitment that have been observed historically was recruited (Henneman and Olson 1965; Zajac and Faden and more recently, and methods by which patterns may be 1985). identiWed and quantiWed. Since it was Wrst proposed a large amount of experimen- tal evidence has been produced that supports the predic- tions of the size principle (Fedde et al. 1969; Freund et al. Motor unit recruitment patterns: a review 1975; HoVer et al. 1987a; Hogrel 2003; Kier and Curtin 2002; Milner-Brown et al. 1973; Tanji and Kato 1973). A The corner stone of current understanding of motor unit growing number of examples, however, indicate that recruitment patterns was proposed by Henneman (1957). orderly recruitment of motor units may not always occur. His work on decerebrate cats indicated that, in response to Studies in humans have shown that auditory and visual greater stimulation, motoneurons were recruited in an feedback can alter recruitment orders (Basmajian 1963), as orderly fashion from the smallest through to the largest. can variations in proprioceptive inputs (Wagman et al. During derecruitment, the inverse order was seen so motor 1965) and cutaneous stimulations (Stephens et al. 1978). In units recruited Wrst were the last to be deactivated. Follow- addition, rapid shortening of the extensor digitorum brevis ing further work, involving stretch reXexes in decerebrate (Grimby and Hannerz 1977) and rapid lengthening of the cats, he termed this recruitment strategy ‘the size principle’ triceps surae (Nardone et al. 1989) also result in preferen- (Henneman et al. 1965a, b). Motor unit size has been tial recruitment of faster motor units. Non-orderly recruit- shown to vary both within and between muscles with a ment has also been reported from the results of glycogen trend for smaller units to be composed of slower twitch depletion studies of demanding activities such as supra- muscle Wbres and larger motor units to be composed of maximal cycling in humans (Gollnick et al. 1974) and faster twitch muscle Wbres (McPhedran et al. 1965a, b). In jumping in the bushbaby (Gillespie et al. 1974). More addition, smaller motor units have smaller diameter nerve recently it has been suggested that motor units, within an axons, which result in slower action potential conduction individual muscle, may form groups that can be indepen- velocities along them (Bawa et al. 1984). The size principle dently activated to fulWl speciWc functional roles (HoVer therefore predicts that, based on both contractile properties et al. 1987b; Loeb 1985). These pools of motor unit have and action potential conduction velocities, faster motor been termed ‘task groups’ and have been shown to be selec- units will be recruited after slower motor units have been tively recruited for diVerent kinematic conditions within a activated and will be the Wrst motor units to be derecruited. motor task such as a stride or a grasping movement (HoVer This pattern is facilitated by the relationship between the et al. 1987b; Loeb 1985; Riek and Bawa 1992; Von Tschar- motor unit
-motoneuron and its associated Ia aVerent neu- ner and Goepfert 2006; Wakeling and Rozitis 2004). It is, rons, from which it receives excitatory synaptic input. The therefore, likely that recruitment strategies other than those number of connections between the motoneuron and the Ia predicted by the size principle may be used during locomo- aVerent neuron is independent of motoneuron cell size, tion. It is possible that mechanical factors may inXuence a meaning that the synaptic density per unit area varies strategy for motor unit recruitment. In some instances it is inversely with the size of the motoneuron (Stein and Ber- counterintuitive for slower motor units to be recruited for toldi 1981). Smaller motoneurons therefore receive greater all motor tasks, speciWcally those that require shortening at synaptic inputs and will reach their depolarising threshold a fast strain rate or have high cycle frequencies, as the before larger motoneurons. mechanical properties of the slower motor units suggest Orderly recruitment of motor units has been proposed to they would contribute little or nothing to force development have several functional advantages. Firstly, it is thought to during rapid movements (Rome et al. 1988). Until recently simplify central nervous system control of muscle contrac- however, the diYculty has been in Wnding a reliable and tions (Henneman et al. 1974). Orderly recruitment also sensitive measure of in vivo activation of motor unit popu- ensures that the slowest, most fatigue resistant, motor units lations, and hence recruitment, that can be used to investi- are recruited Wrst for any given task (Henneman and Olson gate alternate motor unit recruitment strategies during 1965). The faster, fast fatiguing, motor units are therefore natural movements in intact individuals. reserved for infrequent, high intensity tasks such as jump- ing, where they can provide high forces for a short period Determining in vivo motor unit recruitment patterns of time. In addition, as faster motor units are composed of a greater number of muscle Wbres they are able to produce Activity within a motor unit can be detected by measuring greater force than slower motor units (Milner-Brown et al. the bioelectric signals that result from action potentials 1973). Orderly recruitment, therefore, facilitates a smooth travelling along the nerve axons or the muscle Wbres and force increment as it leads to a force increase that is roughly there are several methods through which this can be 123 J Comp Physiol B (2009) 179:57–66 59 achieved. Single unit activity can be recorded from chroni- diVerences in the muscle Wbre diameter (in a manner akin to cally implanted Wne-wire microelectrodes that ‘Xoat’ in the nerve axons; Hodgkin 1954). However, the only evidence spinal canal structures such as the dorsal root ganglia and for this is from in vitro preparations from the frog (Hakans- ventral roots; this procedure has been performed on freely son 1956) where it would be expected that lack of physio- moving cats (HoVer et al. 1981; Loeb et al. 1977; Proc- logical packing of the muscle Wbres would permit the hazka et al. 1976). The results from such studies have iden- conduction velocity to vary with Wbre diameter (Buchthal tiWed fundamental principles of motor control, and et al. 1955). To our knowledge there is no evidence from in particularly relevant to this review was the proposal that vivo data to support the assumption that variations in con- motor units were recruited in task groups according to the duction velocity are caused by muscle Wbre diameter, and in demands required during diVerent movements (Loeb 1985). contrast large variations in Wbre diameter have been shown An alternative approach to identify single motor unit to exist with very little variation in conduction velocity (or activity is using Wne-wire or needle electrodes inserted into the associated mean frequency of the myoelectric signal; the muscle (Olson et al. 1968). Motor unit action potentials Wakeling and Syme 2002). In contrast, it has been sug- (MUAPs) vary in magnitude due to the combination of a gested that diVerences in conduction velocity can occur due number of factors relating to: (1) the muscles anatomy, i.e. to diVerences in the electrical membrane properties of the the diameter of the muscle Wbres and the number of Wbres diVerent Wbre types (Buchthal et al. 1955) that are known to within each unit and (2) the recording technique, i.e. signal vary between muscle Wbre types (Albuquerque and ThesleV attenuation between the active Wbres and the recording 1968; LuV and Atwood 1972; Wallinga-De Jonge et al. electrodes. Additionally, the shape of the MUAP depends 1985). Conduction velocities can be determined for both on the intrinsic properties of the constituent muscle Wbres Wne-wire (Wakeling and Syme 2002) and surface EMG (Buchthal et al. 1973), the intracellular and extracellular recordings (typically using arrays of electrodes mounted to concentrations of ions (that can vary for instance with the surface of the skin (Farina et al. 2002; Hogrel 2003; fatigue; Brody et al. 1991) and the distance and orientation Houtman et al. 2003; Sadoyama et al. 1988). of the electrodes relative to the active Wbres (Lindstrom and If the purpose of the measurements is to identify patterns Magnusson 1977). Individual motor units can be character- of recruitment between populations of motor units, but the ised on the basis of the amplitude or, additionally, the shape identity of individual units is not important then a number of the recorded MUAPs. Implanting several electrodes to of further approaches can be used. In particular, the spectral give diVerent combinations of electrode orientations rela- properties of recorded MUAPs can give indications as to tive to the active motor units can increase the resolution of the Wbre types active at the time of recording. Each MUAP these techniques, and this is the basis of quadriWlar elec- is conducted along the muscle Wbres at a distinct velocity trodes (DeLuca 1993). These methods are particularly use- and thus can be detected by the recording electrodes for a ful for studying isometric contractions when there is a characteristic time and this gives the major feature of the stable relationship between the electrode position and the frequency of the MUAP spectrum. In addition, the shape of motor units. During dynamic tasks the shape changes that the MUAP also contributes to the spectral properties. The occur within the muscles can alter the relative positioning MUAP spectrum is additionally aVected by factors such as of the electrodes and thus the characteristics of the recorded the muscle temperature (Stalberg 1966), fatigue status MUAPs. Furthermore, during high intensity contractions (Brody et al. 1991; Mortimer et al. 1970), fascicle length the interference patterns generated by MUAPs from diVer- (Doud and Walsh 1995), cross-talk (Farina et al. 2002), ent motor units can change the size and shape of the degree of synchronisation and interference between MUAP recorded signal (DeLuca 1997) and this can hinder the clas- signals (DeLuca 1997) and the volume conductor proper- siWcation of individual units. ties of the myoelectric signal through the tissue (Roeleveld In some studies where individual motor units have been et al. 1997). The recorded myoelectric signal contains the characterised on the basis of the MUAPs their Wring pat- convolution of the spectral properties of the individual terns have been documented relative to their Wring thresh- MUAPs and their Wring statistics. With asynchronous and olds. An association exists between the Wring threshold and stochastic Wring, as is typically associated with myoelectric the type of muscle Wbre within the motor unit (Andreassen signals, there is very little inXuence of the Wring statistics in and Arendt-Nielsen 1987; Garnett et al. 1978), and there- the myoelectric spectrum. Indeed, motor unit synchronisa- fore diVerent types of motor unit can be distinguished in tion has been shown to contribute frequency components in this way. A further way to distinguish types of motor unit is the neighbourhood of the average Wring rate (15–25 Hz; through the conduction velocity of their MUAPs. MUAP DeLuca 1997) and occurs in only small bursts with conduction velocity increases for faster muscle Wbres fewer than 8% of Wrings (DeLuca 1993). In Wne-wire stud- (Kupa et al. 1995; Sadoyama et al. 1988; Wakeling and ies these low frequency components are usually Wltered Syme 2002). It is commonly thought that this is due to out (Hodson-Tole and Wakeling 2007). In studies where 123 60 J Comp Physiol B (2009) 179:57–66 signals have been collected using surface electrodes how- ume conductor eVect reshaping the spectra as the ever, these occur within the range of frequencies typically myoelectric signals pass through the tissues, the hybrid analysed (10–350 Hz) and so may result in a reduction of properties of many motor units resulting in intermediate the mean frequency values. However, if the spectra are contractile and presumably myoelectric properties and the assessed by their principal components (see below) then recruitment of diVerent motor units resulting in a subtle this already small eVect is further diminished. blend of these hybrid Wbres being activated. In order to Each MUAP occurs at a distinct time, leaving a charac- resolve the diVerences in spectra that occur when motor teristic, time-varying spectral signal. Traditionally, the unit recruitment patterns are varied it is important to use spectral properties of myoelectric signals have been charac- techniques that are powerful in resolving spectral proper- terised by their mean or median frequency and these mea- ties. Traditional measures such as the mean or median fre- sures have been associated with recruitment patterns and quency of the EMG spectrum consider all frequency diVerent proportions of motor unit types (Gerdle et al. components within the analysed frequency band; these may 1988; Kupa et al. 1995; Wretling et al. 1987). However, the contain unwanted elements such as signal noise, measure- mean and median frequency measures consider frequency ment variation or even the low frequency elements that components across the whole spectrum, and these may occur if synchronisation of the motor unit Wring occurs. include factors in addition of the types of motor unit active Instead, we have found that the alternative approach of (Wakeling 2007). Recently, a number of time-frequency using principal component analysis to resolve the major analysis techniques have been used to characterise the time- features of the intensity spectra provides a powerful tool to varying frequency spectra of myoelectric signals (Karlsson discriminate the Wne details of spectral shape that occur et al. 2000; von Tscharner 2000). The uncertainty principle when the motor unit recruitment patterns are varied (von of signal processing means that it is not possible to pre- Tscharner 2002; Wakeling and Rozitis 2004; Hodson-Tole cisely determine everything about the time and frequency and Wakeling 2007). Principal components identify the content of a signal simultaneously (Kaiser 1994; Calude major components of the EMG spectra and have loading and Stay 1995). A balance must therefore be reached scores that correlate with spectral shifts (Wakeling and between the time and frequency resolution of a resolved Rozitis 2004). With the appropriate experimental design signal. We have adopted a wavelet approach developed by (Hodson-Tole and Wakeling 2007; Wakeling and Rozitis von Tscharner (2000) that uses a set of wavelets that have 2004; Wakeling et al. 2006) the higher principal compo- their time resolutions speciWcally tuned to the physiological nents explain the variation in motor unit recruitment and response time of muscle twitches. Each wavelet acts as a the lower components explain smaller sources of variation band-pass Wlter and enables the intensity of the signal to be and measurement noise. It has been suggested that by con- calculated at diVerent times within that speciWc frequency Wning the Wnal analysis to the major principal components band. The combined wavelets extract the intensities across these smaller sources of variation are excluded (a process the physiological important range of frequencies (depend- that does not occur for mean frequency analysis) and this ing on the recording technique used). We have shown that can improve the resolution of the analysis technique the instantaneous spectra determined from wavelet decom- (Wakeling 2007). position of myoelectric signals are associated with activity A further development that can be adopted to resolve from diVerent types of muscle Wbre using a range of the speciWc frequency bands that indicate diVerent motor approaches: electrical stimulation to the nerve (Wakeling unit activity is to construct wavelets that have their prop- and Syme 2002), voluntary contractions (Wakeling and erties matched to the major frequency components within Rozitis 2004) and modelling (Wakeling 2007). Wavelet the signal, and correspondingly to the signals from diVer- approaches have shown that the spectral properties of the ent types of motor unit. These approaches involve an ini- myoelectric signal change along the course of a stride or tial wavelet analysis of the myoeletric signals, principal gait cycle (Hodson-Tole and Wakeling 2008a; von Tscharner component classiWcation of the major frequency proper- 2000; Wakeling and Rozitis 2004), and these results support ties followed by an optimisation to tune wavelets to those earlier reports of task-speciWc recruitment from the nerve frequencies. These approaches have been adopted by roots in the cat (Loeb 1985). both Von Tscharner and Goepfert (2006) and Hodson- Motor unit action potentials show substantial (two to Tole and Wakeling (2007) and can be used to directly tar- threefold) variation in both conduction velocity and spec- get the signals from diVerent types of motor unit within tral frequencies between fast- and slow-Wbre types (Wake- the analysis. It is therefore possible to study patterns of ling et al. 2001; Wakeling and Syme 2002). However, the motor unit recruitment using a number of diVerent tech- actual variations in spectral properties that are generated by niques, with the choice of approach often determined by a muscle during normal behaviours may be much more sub- the type of motor task to be studied and the question tle. This is due to a number of factors that include the vol- being addressed. 123 J Comp Physiol B (2009) 179:57–66 61 In vivo patterns of motor unit recruitment: new insights motor unit recruitment based on mechanical demand of the motor task should occur in some situations. If preferential Motor unit recruitment patterns can be studied on a number recruitment of faster motor units provides a mechanical of diVerent levels, i.e. within a group of synergistic muscles and/or energetic advantage it would be predicted that a sig- such as Xexor or extensor muscle groups acting as a unit, niWcant positive association exist between the recruitment within a single muscle or within speciWc regions of a single of faster motor units and faster muscle fascicle strain rates muscle (Riek and Bawa 1992). The size principle predicts during shortening. This has been shown to be true in man that faster motor units will be recruited after slower motor during cycling (Wakeling et al. 2006) and rats running on a units have been activated and will be the Wrst motor units to treadmill (Hodson-Tole and Wakeling 2008b). In both be deactivated (Henneman and Olson 1965; Henneman reports the association between strain rate and recruitment et al. 1965a) and studies of single motor units have shown of faster motor units was strongest in muscles with a large that orderly recruitment of motor units occurs within each population of faster Wbre types. Previous work has shown of the hierarchical levels described above (Riek and Bawa that orderly recruitment of motor units is harder to identify 1992). There is however, a growing body of evidence that in large, mixed Wbre type muscles, particularly when suggests orderly recruitment does not adequately describe numerous fast contracting Wbres are present (Burke et al. motor unit recruitment in all situations (Fedde et al. 1969; 1973; Burke and Rymer 1976; Fleshman et al. 1981; Proske Hodson-Tole and Wakeling 2007; 2008a, b; Hogrel 2003; and Waite 1976; Stephens and Stuart 1975), suggesting that Milner-Brown et al. 1973; Wakeling et al. 2001; 2006; motor unit recruitment based on mechanical demand of the Wakeling and Rozitis 2004). Further to this, recent work motor task may be more predominant in populations of quantifying patterns of motor unit recruitment in distinct faster motor units. Further work is required to test this sug- groups of Wbre type populations, indicates that not only do gestion. In addition, work should also consider whether the motor unit recruitment patterns change in response to spe- inXuence of mechanical factors is not only aVected by Wbre ciWc locomotor demands, but that recruitment patterns vary type population, but also by the motor task carried out and between muscles that are composed of diVerent proportions consideration should be made to the interaction between the of muscle Wbre types (Hodson-Tole and Wakeling 2008a). muscles and the musculoskeletal system as a whole. Spe- The functional signiWcance of these diVerences is currently ciWcally, the functional demands placed on the limb and unclear and is likely to remain so until general strategies individual muscles varies across the time course of a stride that govern patterns of motor unit recruitment during (Nigg and Wakeling 2001) and is reXected by changing dynamic tasks are identiWed. The exact nature of the other correlations between motor unit recruitment, myoelectric factors inXuencing patterns of motor unit recruitment must intensity and muscle fascicle strain rates (Hodson-Tole and therefore be investigated. Wakeling 2008b). It is therefore possible that changes occur in both functional demand and recruitment strategy. Motor unit recruitment strategies If this were found to be true it could be an indicator of the presence of functional task groups within a muscle. Such Motor unit recruitment for sustained, low force muscle con- groups have previously been identiWed in cat sartorius mus- tractions (e.g. postural tasks, load carrying) is well cle, where they have been shown to be active during diVer- explained by the orderly recruitment predicted by the size ent stride phases and to undergo diVerent strain trajectories principle. This strategy is considered to be beneWcial (HoVer et al. 1987b). DiVerent task groups may be sensitive because recruiting slower, fatigue-resistant motor units Wrst to diVerent stimuli and therefore be recruited in response to simpliWes central nervous system control (Henneman et al. speciWc locomotor events. Their presence provides a path- 1974; Zajac and Faden 1985), makes prolonged use of the way through which preferential recruitment of faster motor most fatigue-resistant Wbres and ensures a smooth force units within a muscle could be facilitated. increment (Henneman and Olson 1965; Zajac and Faden In addition to the theory that motor unit recruitment 1985). Such a strategy, however, is not necessarily favour- should be associated to the mechanical demands of the able for all motor tasks. In situations where rapid force motor task, it has been suggested that activation–deactiva- development is required (e.g. fast starts in swimming Wsh, tion kinetics might also be inXuential (Hodson-Tole and high speed locomotion such as supra-maximal cycling, cat Wakeling 2008a). It is well documented that faster motor paw-shake), preferential recruitment of faster motor units units have faster activation and deactivation rates than would be advantageous due to their faster activation and slower motor units (Burke et al. 1973). To ensure the relaxation rates (Burke et al. 1973) and their potential for appropriate force and/or mechanical work is produced over producing maximum mechanical power output and maxi- the course of a motor task, such as a stride, the timing of mum mechanical eYciency at faster strain rates (He et al. muscle activation and deactivation must accommodate 2000). Strong evidence therefore exists to suggest that these intrinsic properties. Activation–deactivation kinetics 123 62 J Comp Physiol B (2009) 179:57–66 determined from maximal activation of in situ muscle prep- plied to faster motor units, compared to feedback supplied arations do not however reXect rates which may occur in to slower motor units (Taylor and Gottlieb 1985). In addi- vivo. Activation–deactivation kinetics are inXuenced by a tion, a reduction in stretch reXex gain has been shown dur- wide range of factors such as changes in muscle fascicle ing locomotion at faster velocities (running vs. walking) strain (Brown et al. 1999; Close 1972; Josephson and (Capaday and Stein 1987). Indeed, Wring of muscle spindle Stokes 1999), motor unit Wring frequency (Roszek et al. aVerents varies between muscles and in response to diVer- 1994) and muscle fascicle strain rates (Brown and Loeb ent fascicle length changes, i.e. shortening versus lengthen- 2000). The complex relationship that undoubtedly exists ing contractions (Loeb 1984; Murphy and Martin 1993; between these intrinsic properties and the state/behaviour Prochazka 1996; Prochazka and Gorassini 1998), and is of the muscle suggest that this factor may be crucial in thought to inXuence muscle mechanical responses more determining muscle function and is highly likely to inXu- during lengthening contractions (Nichols and Houk 1976). ence motor unit recruitment patterns. It is well accepted This has led some researchers to suggest that diVerent that activation, and in particular, deactivation rates need to motor tasks may require diVerent fusimotor activation pat- be well matched to the cycle frequency of a movement terns in relation to skeletomotor activation patterns, and (Caiozzo and Baldwin 1997; Johnston 1991; Neptune and that this relationship maybe modulated by the direction the Kautz 2001). How strain, strain rate, motor unit Wring fre- muscle fascicle length changes (Murphy and Martin 1993). quency and any other relevant factor interact in vivo and As the work by Henneman et al. was conducted on modulate activation–deactivation rates is currently not well decerebrate cats many pathways, which normally inXuence understood. Gaining an insight into in vivo activation– motor unit recruitment, were not left intact and could there- deactivation rates during locomotion may therefore not fore not inXuence the results. One such factor would be only provide further insight into the intrinsic properties of activity and force generation in adjacent muscles. During muscles but may also improve current understanding of reXex contraction of the soleus and gastrocnemius muscles, motor control and the in vivo mechanical behaviour of reXex inhibition of the soleus muscle is mediated by stretch diVerent muscles. This is therefore an important avenue for of the gastrocnemius (Dacko et al. 1996). Selective inhibi- future work. tion of slower motor units have been recorded in the cat soleus muscle, but only when soleus inhibition occurred as Mechanisms determining motor unit recruitment a result of stretch of the medial gastrocnemius (SokoloV and Cope 1996). DiVerential recruitment was never demon- When identifying factors that may inXuence motor unit strated when the stretch reXex of the soleus muscle alone recruitment patterns it is important to consider the possible was assessed (SokoloV and Cope 1996). This work demon- mechanisms that control these relationships. Sensory feed- strates that it is important to consider the inXuence adjacent back from muscles spindles, Golgi tendon organs, joint muscles have on each other, and that to gain better under- receptors and cutaneous receptors can all potentially inXu- standing of motor control it is important to consider the ence recruitment patterns. The complex interactions musculoskeletal system as a whole, dynamic system rather between diVerent types of receptor,
-motoneurons and spi- than a set of independent units as proposed in a number of nal interneurons is, however, currently not well understood recent reviews (Chiel and Beer 1997; Pearson et al. 2006; and has been shown to vary between muscles and in Rossignol et al. 2006; Frigon and Rossignol 2006; Biew- response to diVerent motor tasks (Windhorst 2007). ener and Daley 2007). The size principle theory was developed following work In order to aid the control of diVerent motor units within on stretch reXexes in cats (Henneman 1957). The stretch a muscle it has been proposed that muscles that perform reXex causes a stretched muscle to contract and its primary more than one kinematic type of task may be functionally sources of input are mechanoreceptors, sensitive to muscle divided into task-oriented groups of motor units, or ‘task length changes, called muscle spindles. Muscle spindles groups’ (Loeb 1985). The presence of such independent contain a small number of intrafusal muscle Wbres, which groups of motor units means that an individual muscle has are supplied by sensory nerves (Ia and II aVerent Wbres) and the potential to fulWl multiple functions. Each group of -motoneurons. The correlation between motor unit axonal motor units could have diVerent central connections and conduction velocity and fused tetanic tension is weak when recruitment patterns and, it has also been suggested, may data are collected from large mixed Wbre muscles (Burke have intrinsic properties optimised for the performance of a et al. 1973; Burke and Rymer 1976; Fleshman et al. 1981; speciWc functional task (HoVer et al. 1987b). DiVerential Proske and Waite 1976; Stephens and Stuart 1975), with activation of diVerent muscle regions that have speciWc the lack of correlation most apparent when numerous faster mechanical roles has been described in a number of mus- contracting motor units are present. This Wnding may be a cles in several animal species, e.g. several cat muscles result of the weaker monosynaptic Ia aVerent feedback sup- (Chanaud and Macpherson 1991; Chanaud et al. 1991; 123 J Comp Physiol B (2009) 179:57–66 63 English and Weeks 1987; HoVer et al. 1987b; Pratt and Jordan 1980). They, therefore, provide a mechanism through Loeb 1991), pig masseter muscle (Herring et al. 1979), bird which preferential recruitment of faster motor units could pectoralis muscle (Dial et al. 1987), human common digital occur during locomotion. The selective inhibition of motor extensor muscle (Riek and Bawa 1992), monkey Xexor dig- units in the soleus muscle, demonstrated by SokoloV and itorum profundus (Schieber 1993) and guinea fowl lateral Cope (1996), was conducted on decerebrate cats and so gastrocnemius muscle (Higham et al. 2008). These Wndings diVerential motor unit recruitment can occur just from spi- indicate that functional task groups do exist within muscles, nal circuits (SokoloV and Cope 1996). However, recurrent and speciWcally highlights their existence in compartmenta- inhibition and the activation of Renshaw cells are addition- lised muscles. DiVerential activation of motor units is not, ally controlled by descending tracts from by supraspinal however, limited to such compartmentalised muscles: centres (for review see Katz and Pierrot-Deseilligny 1998). selective inhibition has been identiWed in the cat soleus DiVerential recruitment of motor units can thus be inXuenced (SokoloV and Cope 1996), and this muscle is considered to by circuits at both spinal and supraspinal levels. consist of a single neuromuscular compartment (SokoloV and Cope 1996) with motor units distributed across most of its cross-sectional area (Cope et al. 1986). When popula- Conclusion tions of motor units are studied, as with the analysis of myoelectric signals, the presence of functional task groups The size principle theory of motor unit recruitment pro- of motor units within the muscle is likely to result in diVer- vides a very robust framework with which motor unit ential activation of motor units being identiWed. Identifying recruitment patterns can be predicted, which persists when diVerential activation of motor units in such a way does not studies of some movements are made. The anatomical necessarily indicate that motor unit recruitment order structure of neuromuscular system shows that aVerents within a particular task group will diVer from that predicted from the spinal and brain levels can additionally modulate by the size principle. Indeed, orderly recruitment of motor motor unit activity, while motor unit task groups (deWned units within individual task groups has been identiWed from on the basis of anatomy or functionality) allow discrete studies of single motor units (Riek and Bawa 1992). populations of motor units to be diVerentially activated. One pathway that could lead to the diVerential recruit- Recent work has shown that diVerential activation of task ment of motor units is that involving Renshaw cells. Ren- groups does occur in response to the mechanical demands shaw cells are small neurons located between motor axon of the motor task. This represents a new perspective of neu- collaterals and motoneurons in the ventral spinal chord romuscular control of the musculoskeletal system and has (Renshaw 1941; 1946). They mediate recurrent inhibition important implications for those developing musculoskele- of
-motoneurons (Renshaw 1941) as well as interacting tal models, designing rehabilitation protocols and monitor- with other types of spinal neurons: -motoneurons (Ellaway ing musculoskeletal injury and disease, as well those 1971), Ia inhibitory interneurons (Hultborn 1972; Hultborn interested in control theory of robots. Analysis of myoelec- et al. 1971), ascending tract cells (Lindstrom and Schom- tric signals can now resolve activity from diVerent popula- burg 1973) and other Renshaw cells (Ryall 1970). Renshaw tions of motor units from signals collected using both Wne- cells are therefore thought to play an important role during wire and surface techniques, and provides a means by locomotion, although the exact nature of this role is which further investigation of motor unit recruitment pat- currently poorly understood. It has been shown that the terns may be pursued. recurrent inhibitory inXuence of Renshaw cells on
-moto- neurons diVers between motor unit types, with fast-fatigu- ing units being less inhibited than fatigue-resistant units, References which in turn are less inhibited than slow units (Friedman Albuquerque EX, ThesleV S (1968) A comparative study of membrane et al. 1981). As diVerences in cell size and excitability do properties of innervated and chronically denervated fast and slow not explain these diVerences, it has been proposed that the skeletal muscles of the rat. Acta Physiol Scand 73:471–480 number of Renshaw cell projections linked to
-motoneu- Andreassen S, Arendt-Nielsen L (1987) Muscle Wbre conduction rons of diVerent motor unit types diVers (Friedman et al. velocity in motor units of the human anterior tibial muscle: a new size principle parameter. J Physiol 391:561–571 1981). Renshaw cells themselves appear to receive more Basmajian JV (1963) Control and training of individual motor units. input from large
-motoneurons, and are stimulated less by Science 141:440–441 small
-motoneurons (Ryall et al. 1972). It is therefore Bawa P, Binder MD, Ruenzel P, Henneman E (1984) Recruitment or- apparent that Renshaw cells are activated by activity in der of motoneurons in stretch reXexes is highly correlated with their axonal conduction velocity. J Neurophysiol 52:410–420 faster motor units, which leads them to preferentially Biewener AA, Daley MA (2007) Unsteady locomotion: integrating inhibit activity in slower motor units, a pathway which has muscle function with whole body dynamics and neuromuscular been shown to be active during locomotion (Pratt and control. J Exp Biol 210:2949–2960 123 64 J Comp Physiol B (2009) 179:57–66 Bottinelli R, Betto R, SchiaYno S, Reggiani C (1994a) Maximum Ellaway PH (1971) Recurrent inhibition of fusimotor neurones exhib- shortening velocity and coexistence of myosin heavy chain iso- iting background discharges in the decerebrate and the spinal cat. forms in single skinned fast Wbres of rat skeletal muscle. J Muscle J Physiol 216:419–439 Res Cell Motil 15:413–419 English AW, Weeks OI (1987) An anatomical and functional analysis Bottinelli R, Betto R, SchiaYno S, Reggiani C (1994b) Unloaded of cat biceps femoris and semitendinosus muscles. J Morphol shortening velocity and myosin heavy chain and alkali light chain 191:161–175 isoform composition in rat skeletal muscle Wbres. J Physiol 478(Pt Farina D, Fosci M, Merletti R (2002) Motor unit recruitment strategies 2):341–349 investigated by surface EMG variables. J Appl Physiol 92:235– Brody LR, Pollock MT, Roy SH, De Luca CJ, Celli B (1991) pH-in- 247 duced eVects on median frequency and conduction velocity of the Fedde MR, DeWet PD, Kitchell RL (1969) Motor unit recruitment pat- myoelectric signal. J Appl Physiol 71:1878–1885 tern and tonic activity in respiratory muscles of Gallus domesti- Brown IE, Cheng EJ, Loeb GE (1999) Measured and modeled proper- cus. J Neurophysiol 32:995–1004 ties of mammalian skeletal muscle. II. The eVects of stimulus fre- Fleshman JW, Munson JB, Sypert GW, Friedman WA (1981) Rheo- quency on force-length and force-velocity relationships. J Muscle base, input resistance, and motor-unit type in medial gastrocne- Res Cell Motil 20:627–643 mius motoneurons in the cat. J Neurophysiol 46:1326–1338 Brown IE, Loeb GE (2000) Measured and modeled properties of mam- Freund HJ, Budingen HJ, Dietz V (1975) Activity of single motor units malian skeletal muscle: IV. Dynamics of activation and deactiva- from human forearm muscles during voluntary isometric contrac- tion. J Muscle Res Cell Motil 21:33–47 tions. J Neurophysiol 38:933–946 Buchthal F, Guld C, Rosenfalk P (1955) Innervation zone and propaga- Friedman WA, Sypert GW, Munson JB, Fleshman JW (1981) Recur- tion velocity in human muscles. Acta Physiol Scand 35:174–190 rent inhibition in type-identiWed motoneurons. J Neurophysiol Buchthal F, Dahl K, Rosenfalck P (1973) Rise time of the spike poten- 46:1349–1359 tial in fast and slowly contracting muscle of man. Acta Physiol Frigon A, Rossignol S (2006) Experiments and models of sensorimotor Scand 87:261–269 interactions during locomotion. Biol Cybern 95:607–627 Burke RE, Rymer WZ (1976) Relative strength of synaptic input from Garnett RAF, O’Donovan MJ, Stephens JA, Taylor A (1978) Motor short-latency pathways to motor units of deWned type in cat medi- units organization of human medial gastrocnemius. J Physiol al gastrocnemius. J Neurophysiol 39:447–458 287:33–43 Burke RE, Levine DN, Tsairis P, Zajac FEIII (1973) Physiological Gerdle B, Wretling ML, Henriksson-Larsen K (1988) Do the Wbre-type types and histochemical proWles in motor units of the cat gastroc- proportion and the angular velocity inXuence the mean power fre- nemius. J Physiol 234:723–748 quency of the electromyogram? Acta Physiol Scand 134:341–346 Calude CS, Stay MA (1995) From the Heisenberg to Godel via Chaitin. Gillespie CA, Simpson DR, Edgerton VR (1974) Motor unit recruit- Int J Theor Phys 44:1053–1065 ment as reXected by muscle Wbre glycogen loss in a prosimian Capaday C, Stein RB (1987) DiVerence in the amplitude of the human (bushbaby) after running and jumping. J Neurol Neurosurg Psy- soleus H reXex during walking and running. J Physiol 392:513– chiatry 37:817–824 522 Gollnick PD, Piehl K, Saltin B (1974) Selective glycogen depletion Caiozzo VJ, Baldwin KM (1997) Determinants of work produced by pattern in human muscle Wbres after exercise of varying intensity skeletal muscle: potential limitations of activation and relaxation. and at varying pedalling rates. J Physiol 241:45–57 Am J Physiol 273:C1049–C1056 Grimby L, Hannerz J (1977) Firing rate and recruitment order of toe Chanaud CM, Macpherson JM (1991) Functionally complex muscles extensor motor units in diVerent modes of voluntary contraction. of the cat hindlimb. III. DiVerential activation within biceps femo- J Physiol 264:865–879 ris during postural perturbations. Exp Brain Res 85:271–280 Hakansson CH (1956) Conduction velocity and amplitude of the action Chanaud CM, Pratt CA, Loeb GE (1991) Functionally complex mus- potential as related to circumference in the isolated Wbre of frog cles of the cat hindlimb. II. Mechanical and architectural heterog- muscle. Acta Physiol Scand 37:14–34 enity within the biceps femoris. Exp Brain Res 85:257–270 He Z-H, Bottinelli R, Pellegrino MA, Ferenczi MA, Reggiani C (2000) Chiel HJ, Beer RD (1997) The brain has a body: adaptive behaviour ATP consumption and eYciency of human single muscle Wbers emerges from interaction of nervous system, body and environ- with diVerent myosin isoform composition. Biophys J 79:945– ment. Trends Neurosci 20:553–557 961 Close RI (1972) The relations between sarcomere length and charac- Henneman E (1957) Relation between size of neurons and their sus- teristics of isometric twitch contractions of frog sartorius muscle. ceptibility to discharge. Science 126:1345–1347 J Physiol 220:745–762 Henneman E, Clamann HP, Gillies JD, Skinner RD (1974) Rank order Cope TC, Bodine SC, Fournier M, Edgerton VR (1986) Soleus motor of motoneurons within a pool: law of combination. J Neurophys- units in chronic spinal transected cats: physiological and morpho- iol 37:1338–1349 logical alterations. J Neurophysiol 55:1202–1220 Henneman E, Olson CB (1965) Relations between structure and func- Dacko SM, SokoloV AJ, Cope TC (1996) Recruitment of triceps surae tion in the design of skeletal muscles. J Neurophysiol 28:581–598 motor units in the decerebrate cat. I. Independence of type S units Henneman E, Somjen G, Carpenter DO (1965a) Excitability and inhi- in soleus and medial gastrocnemius muscles. J Neurophysiol bitability of motoneurons of diVerent sizes. J Neurophysiol 75:1997–2004 28:599–620 DeLuca CJ (1993) Precision decomposition of EMG signals. Methods Henneman E, Somjen G, Carpenter DO (1965b) Functional signiW- Clin Neurophysiol 4:1–28 cance of cell size in spinal motoneurons. J Neurophysiol 28:560– DeLuca CJ (1997) The use of surface electromyography in biomechan- 580 ics. J Appl Biomech 13:135–136 Herring SW, Grimm AF, Grimm BR (1979) Functional heterogeneity Dial KP, Kaplan SR, Goslow GE Jr, Jenkins FA Jr (1987) Structure in a multipinnate muscle. Am J Anat 154:563–576 and neural control of the pectoralis in pigeons: implications for Higham TE, Biewener A, Wakeling JM (2008) Functional diversiWca- Xight mechanics. Anat Rec 218:284–287 tion within and between muscle synergisists during locomotion. Doud JR, Walsh JM (1995) Muscle fatigue and muscle length interac- Biology Letters 4:41–44 tion: eVect on the EMG frequency components. Electromyogr Hodgkin AL (1954) A note on conduction velocity. J Physiol 125:221– Clin Neurophysiol 35:331–339 224 123 J Comp Physiol B (2009) 179:57–66 65 Hodson-Tole EF, Wakeling JM (2007) Variations in motor unit McPhedran AM, Wuerker RB, Henneman E (1965b) Properties of mo- recruitment patterns occur within and between muscles in the run- tor units in a homogeneous red muscle (Soleus) of the cat. J Neu- ning rat (Rattus norvegicus). J Exp Biol 210:2333–2345 rophysiol 28:71–84 Hodson-Tole EF, Wakeling JM (2008a) Motor unit recruitment pat- Milner-Brown HS, Stein RB, Yemm R (1973) The orderly recruitment terns 1: responses to changes in locomotor velocity and incline. J of human motor units during voluntary isometric contractions. J Exp Biol 211:1882–1892 Physiol 230:359–370 Hodson-Tole EF, Wakeling JM (2008b) Motor unit recruitment pat- Mortimer JT, Magnusson R, Petersen I (1970) Conduction velocity in terns 2: the inXuence on myoelectric intensity and muscle fascicle ischemic muscle: eVect on EMG frequency spectrum. Am J Phys- strain rate. J Exp Biol 211:1893–1902 iol 219:1324–1329 HoVer JA, Loeb GE, Marks WB, O’Donovan MJ, Pratt CA, Sugano N Murphy PR, Martin HA (1993) Fusimotor discharge patterns during (1987a) Cat hindlimb motoneurons during locomotion. I. Desti- rhythmic movements. Trends Neurosci 16:273–278 nation, axonal conduction velocity, and recruitment threshold. J Nardone A, Romano C, Schieppati M (1989) Selective recruitment of Neurophysiol 57:510–529 high-threshold human motor units during voluntary isotonic HoVer JA, Loeb GE, Sugano N, Marks WB, O’Donovan MJ, Pratt CA lengthening of active muscles. J Physiol 409:451–471 (1987b) Cat hindlimb motoneurons during locomotion. III. Func- Neptune RR, Kautz SA (2001) Muscle activation and deactivation tional segregation in sartorius. J Neurophysiol 57:554–562 dynamics: the governing properties in fast cyclical human move- HoVer JA, O’Donovan MJ, Pratt CA, Loeb GE (1981) Discharge pat- ment performance? Exerc Sport Sci Rev 29:76–81 terns of hindlimb motoneurons during normal cat locomotion. Nichols TR, Houk JC (1976) Improvement in linearity and regulation Science 213:466–467 of stiVness that results from the actions of the stretch reXex. J Hogrel JY (2003) Use of surface EMG for studying motor unit recruit- Neurophysiol 29:119–142 ment during isometric linear force ramp. J Electromyogr Kinesiol Nigg BM, Wakeling JM (2001) Impact forces and muscle tuning: a 13:417–423 new paradigm. Exerc Sport Sci Rev 29:37–41 Houtman CJ, Stegeman DF, Van Dijk JP, Zwarts MJ (2003) Changes Olson CB, Carpenter DO, Henneman E (1968) Orderly recruitment of in muscle Wber conduction velocity indicate recruitment of dis- muscle action potentials. Arch Neurol 19:591–597 tinct motor unit populations. J Appl Physiol 95:1045–1054 Pearson K, Ekeberg Ö, Büschges A (2006) Assessing sensory function Hultborn H (1972) Convergence on interneurones in the reciprocal Ia in locomotor systems using neuro-mechanical simulations. inhibitory pathway to motoneurones. Acta Physiol Scand Suppl Trends Neurosci 29:625–631 375:1–42 Pratt CA, Jordan LM (1980) Recurrent inhibition of motoneurons in Hultborn H, Jankowska E, Lindstrom S (1971) Recurrent inhibition decerebrate cats during controlled treadmill locomotion. J Neuro- from motor axon collaterals of transmission in the Ia inhibitory physiol 44:489–500 pathway to motoneurones. J Physiol 215:591–612 Pratt CA, Loeb GE (1991) Functionally complex muscles of the cat Johnston IA (1991) Muscle action during locomotion: a comparative hindlimb. I. Patterns of activation across sartorius. Exp Brain Res perspective. J Exp Biol 160:167–185 85:243–256 Josephson R, Stokes D (1999) Work-dependent deactivation of a crus- Prochazka A (1996) Proprioceptive feedback and movement regula- tacean muscle. J Exp Biol 202:2551–2565 tion. In: Rowell L, Shepherd JT (eds) Handbook of physiology, Kaiser G (1994) A friendly guide to wavelets. Birkhaeuser, Boston exercise: regulation and integration of multiple systems. Ameri- Karlsson S, Yu J, Akay M (2000) Time-frequency analysis of myo- can Physiological Society, New York, pp 89–127 electric signals during dynamic contractions: a comparative Prochazka A, Gorassini M (1998) Ensemble Wring of muscle aVerents re- study. IEEE Trans Biomed Eng 47:228–238 corded during normal locomotion in cats. J Physiol 507:293–304 Katz R, Pierrot-Deseilligny E (1998) Recurrent inhibition in humans. Prochazka A, Westerman RA, Ziccone SP (1976) Discharges of single Prog Neurobiol 27:325–355 hindlimb aVerents in the freely moving cat. J Neurophysiol Kier WM, Curtin NA (2002) Fast muscle in squid (Loligo pealei): con- 39:1090–1104 tractile properties of a specialized muscle Wbre type. J Exp Biol Proske U, Waite PM (1976) The relation between tension and axonal 205:1907–1916 conduction velocity for motor units in the medial gastrocnemius Kupa EJ, Roy SH, Kandarian SC, De Luca CJ (1995) EVects of muscle muscle of the cat. Exp Brain Res 26:325–328 Wber type and size on EMG median frequency and conduction Renshaw B (1941) InXuence of discharge of motoneurons upon exci- velocity. J Appl Physiol 79:23–32 tation of neighboring motoneurons. J Neurophysiol 4:167–183 Lindstrom LH, Magnusson RI (1977) Interpretation of myoelectric Renshaw B (1946) Central eVects of centripetal impulses in axons of power spectra: a model and its applications. Proc IEEE 65:653– spinal ventral roots. J Neurophysiol 9:191–204 662 Riek S, Bawa P (1992) Recruitment of motor units in human forearm Lindstrom S, Schomburg ED (1973) Recurrent inhibition from motor extensors. J Neurophysiol 68:100–108 axon collaterals of ventral spinocerebellar tract neurones. Acta Rossignol S, Dubuc R, Gossard JP (2006) Dynamic sensorimotor Physiol Scand 88:505–515 interactions in locomotion. Physiol Rev 86:89–154 Loeb GE (1984) The control responses of mammalian muscle spindles Roeleveld K, Blok JH, Stegeman DF, van Oosterom A (1997) Volume during normally executed tasks. Exerc Sport Sci Rev 12:157–204 conduction models for surface EMG; confrontation with mea- Loeb GE (1985) Motoneurone task groups: coping with kinematic het- surements. J Electromyogr Kinesiol 7:221–232 erogeneity. J Exp Biol 115:137–146 Rome LC, Funke RP, Alexander RM, Lutz GJ, Aldridge H, Scott F, Loeb GE, Bak MJ, Duysens J (1977) Long-term unit recording from Freadman M (1988) Why animals have diVerent muscle Wbre somatosensory neurons in the spinal ganglia of the freely walking types. Nature 335:824–827 cat. Science 197:1192–1194 Roszek B, Baan GC, Huijing PA (1994) Decreasing stimulation fre- LuV AR, Atwood HL (1972) Membrane properties and contraction of quency-dependent length-force characteristics of rat muscle. J single muscle Wbers in the mouse. Am J Physiol 222:1435–1440 Appl Physiol 77:2115–2124 McPhedran AM, Wuerker RB, Henneman E (1965a) Properties of mo- Ryall RW (1970) Renshaw cell mediated inhibition of Renshaw cells: tor units in a heterogeneous pale muscle (M. Gastrocnemius) of patterns of excitation and inhibition from impulses in motor axon the cat. J Neurophysiol 28:85–99 collaterals. J Neurophysiol 33:257–270 123 66 J Comp Physiol B (2009) 179:57–66 Ryall RW, Piercey MF, Polosa C, Goldfarb J (1972) Excitation of Ren- von Tscharner V (2002) Time-frequency and principal-component meth- shaw cells in relation to orthodromic and antidromic excitation of ods for the analysis of EMGs recorded during a mildly fatiguing motoneurons. J Neurophysiol 35:137–148 exercise on a cycle ergometer. J Electromyogr Kinesiol 12:479–492 Sadoyama T, Masuda T, Miyata H, Katsuta S (1988) Fibre conduction Von Tscharner V, Goepfert B (2006) Estimation of the interplay be- velocity and Wbre composition in human vastus lateralis. Eur J tween groups of fast and slow muscle Wbers of the tibialis anterior Appl Physiol Occup Physiol 57:767–771 and gastrocnemius muscle while running. J Electromyogr Kinesi- Schieber MH (1993) Electromyographic evidence of two functional ol 16:188–197 subdivisions in the rhesus monkey’s Xexor digitorum profundus. Wagman IH, Pierce DS, Burger RE (1965) Proprioceptive inXuence, in Exp Brain Res 95:251–260 volitional control of individual motor units. Nature 207:957–958 Sherrington C (1929) Ferrier lecture: some functional problems attach- Wakeling JM (2007) Patterns of motor recruitment can be determined ing to convergence. Proc R Soc B 105:332–362 using surface EMG. J Electromyogr Kinesiol. doi:10.1016/j.jele- SokoloV AJ, Cope TC (1996) Recruitment of triceps surae motor units kin.207.09.006 in the decerebrate cat. II. Heterogeneity among soleus motor Wakeling JM, Rozitis AI (2004) Spectral properties of myoelectric sig- units. J Neurophysiol 75:2005–2016 nals from diVerent motor units in the leg extensor muscles. J Exp Stalberg E (1966) Propagation velocity in human muscle Wbers in situ. Biol 207:2519–2528 Acta Physiol Scand Suppl 287:1–112 Wakeling JM, Syme DA (2002) Wave properties of action potentials Stein RB, Bertoldi R (1981) Unique contributions of slow and fast from fast and slow motor units of rats. Muscle Nerve 26:659–668 extensor muscles to the control of limb movements. In: Desmedt Wakeling JM, Pascual SA, Nigg BM, Von Tscharner V (2001) Surface JE (ed) Motor unit types, recruitment and plasticity in health and EMG shows distinct populations of muscle activity when mea- disease. Karger, Basel, pp 85–96 sured during sustained sub-maximal exercise. Eur J Appl Physiol Stephens JA, Stuart DG (1975) The motor units of cat medial gastroc- 86:40–47 nemius: speed–size relations and their signiWcance for the recruit- Wakeling JM, Uehli K, Rozitis AI (2006) Muscle Wbre recruitment can ment order of mortor units. Brain Res 91:177–195 respond to the mechanics of the muscle contraction. J R Soc Inter- Stephens JA, Garnett R, Buller NP (1978) Reversal of recruitment or- face 3:533–544 der of single motor units produced by cutaneous stimulation dur- Wallinga-De Jonge W, Gielen FL, Wirtz P, De Jong P, Broenink J ing voluntary muscle contraction in man. Nature 272:362–364 (1985) The diVerent intracellular action potentials of fast and slow Tanji J, Kato M (1973) Recruitment of motor units in voluntary con- muscle Wbres. Electroencephalogr Clin Neurophysiol 60:539–547 traction of a Wnger muscle in man. Exp Neurol 40:759–770 Windhorst U (2007) Muscle proprioceptive feedback and spinal net- Taylor A, Gottlieb S (1985) Convergence of several sensory modalities works. Brain Res Bull 73:155–202 in motor control. In: Barnes WJP, Gladden MH (eds) Feedback Wretling ML, Gerdle B, Henriksson-Larsen K (1987) EMG: a non- and motor control in invertebrates and vertebrates. Croon Helm, invasive method for determination of Wbre type proportion. Acta London, pp 77–91 Physiol Scand 131:627–628 von Tscharner V (2000) Intensity analysis in time-frequency space of Zajac FE, Faden JS (1985) Relationship among recruitment order, axonal surface myoelectric signals by wavelets of speciWed resolution. J conduction velocity and muscle-unit properties of type-identiWed Electromyogr Kinesiol 10:433–445 motor units in cat plantaris muscle. J Neurophysiol 53:1303–1322 123