Muscle spindle and fusimotor activity in locomotion

Abstract

Mammals may exhibit different forms of locomotion even within a species. A particular form of locomotion (e.g. walk, run, bound) appears to be selected by supraspinal commands, but the precise pattern, i.e. phasing of limbs and muscles, is generated within the spinal cord by so-called central pattern generators. Peripheral sense organs, particularly the muscle spindle, play a crucial role in modulating the central pattern generator output. In turn, the feedback from muscle spindles is itself modulated by static and dynamic fusimotor (gamma) neurons. The activity of muscle spindle afferents and fusimotor neurons during locomotion in the cat is reviewed here. There is evidence for some alpha–gamma co-activation during locomotion involving static gamma motoneurons. However, both static and dynamic gamma motoneurons show patterns of modulation that are distinct from alpha motoneuron activity. It has been proposed that static gamma activity may drive muscle spindle secondary endings to signal the intended movement to the central nervous system. Dynamic gamma motoneuron drive appears to prime muscle spindle primary endings to signal transitions in phase of the locomotor cycle. These findings come largely from reduced animal preparations (decerebrate) and require confirmation in freely moving intact animals.

Introduction

There is a considerable element of automatic behaviour in the process of locomotion as has been evident from the earliest observations of headless chickens running in a farmyard to modern-day sights of people carrying out cerebral and other complex motor tasks while running or walking. The principal components of the automatic aspects of locomotion lie in the spinal pattern generators first postulated by Brown (1911) and investigated extensively by others (see Grillner, 1981). Superimposed on these pattern generators are supraspinal (including cerebral) control and sensory feedback, again researched and reviewed extensively elsewhere (Shik & Orlovsky, 1976; Pearson, 2008; Prochazka & Ellaway, 2012). This review concentrates on presenting what is known of the sensory feedback from a specific peripheral receptor system, the muscle spindles, that can be observed during mammalian locomotion, and discusses ideas as to how that feedback might be used in modulating locomotor rhythms. The review is based on a presentation Symposium held in honour of the life time work of Robert (Bob) Banks in Durham, UK, 2014, sponsored by the Physiological Society and the Biophysical Sciences Institute (Durham University).

The muscle spindle

Muscle spindles are present in large numbers in skeletal muscles and are the most frequently found sense organs in the musculoskeletal tissues of mammalian limbs. They possess two types of sensory nerve endings innervated by fast conducting myelinated nerve fibres, the two types having very different sensitivities to the dynamic phase of any muscle length change. Muscle spindles are unique among musculoskeletal sense organs in having a motor innervation. The motor innervation is capable of controlling the dynamic and static sensitivity of spindle afferent discharge separately through different types of fusimotor neurones. As a result of their complexity and prominence in signalling afferent information to the central nervous system (CNS) about the passive and active states of skeletal muscles, the muscle spindle has received a very large amount of investigation and generated a literature too vast to review here. As a starting point to gain insight into this remarkable sense organ, the Physiological Society Monograph by Matthews (1972, 1981) and Hulliger (1984) are comprehensive review articles.

Much of the detailed anatomy and physiology of the muscle spindle was established over an intense period of research in the 30 years leading up to 1990. The details that emerged from that time are summarized in Fig.1, a diagram of the muscle spindle with its afferent and efferent innervation drawn by Yves Laporte. Laporte’s group in Toulouse (France) had collaborated with that of David Barker in Durham (UK) to establish one of the key debatable features of spindle innervation, namely which of the two distinct types of motor terminals on intrafusal muscle fibres, plate and trail endings, were innervated by which of the two distinct types of fusimotor fibre, the dynamic and static gamma efferent axons. The research was one of the most remarkable pieces of work that has been accomplished in the field of motor control, combining electrophysiology and histology (Barker et al. 1973). In a delicate initial operation in Toulouse, the motor innervation of the tenuissimus muscle of a cat was reduced to a single gamma efferent. After allowing several days for degeneration of all cut efferent axons, the identity of the surviving gamma efferent was established in a second experiment through stimulation of the axon and recordings from tenuissimus spindle endings. The muscle was then excised, fixed appropriately and sent to Barker’s laboratory for histological investigation. In the event, in all 10 muscles investigated the function of the surviving axon was determined as a static gamma efferent and the motor innervation of the spindles was of the trail type. By omission, it seemed most likely that dynamic axons would likely provide the efferent innervation of plate endings. The work also established clearly that the two types of intrafusal fibre (bag and chain) then known are both supplied by static gamma efferents through trail endings. The clear link between anatomy and physiology was a pivotal point in advancing understanding of the working of the muscle spindle, but the move away from the previously held idea (Boyd, 1971; Matthews, 1972) in which just two types of intrafusal fibres (bag and chain) were thought to be separately controlled by the two types of gamma efferent raised more questions and fostered further research.

Fig 1.

Diagrammatic representation of the mammalian muscle spindle. From left to right: parent muscle; capsule (c.) nerve supply (n.), intrafusal muscle fibre bundle (i.m.f.); typical repertory of intrafusal muscle fibres; sensory innervation showing the primary afferent (group Ia axon) innervating all intrafusal fibres and the secondary afferent (group II axon) innervating chain and bag₂ fibres; motor innervation showing dynamic gamma axon restricted to the bag₁ fibre and static gamma axon innervating chain and bag₂ fibres. Beta innervations are alpha motoneurons innervating both extrafusal skeletal muscle fibres and intrafusal fibres. Personal communication from Yves Laporte.

A key contributor to this ongoing investigation was the work of Bob Banks in whose honour the articles in this edition of the Journal of Anatomy have been assembled. By 1980 it was well established that the mammalian muscle spindle actually contained three types of intrafusal muscle fibre, two types of bag fibre (bag₁ and bag₂) and chain fibres, that could be identified both histologically (Banks et al. 1977) and electrophysiologically (Boyd, 1976). It had also emerged that the length-dependent discharges and the responses to length changes of the primary and secondary endings of spindles were in large part independently controlled by static and dynamic gamma motoneurons (Emonet-Denand et al. 1977). What was not so clear was whether there was a clear division in terms of fusimotor innervation of a dynamic bag₁ fibre solely by dynamic fusimotor neurons and a static bag₂ and chain fibres by static fusimotor neurons (Barker et al. 1976b; Boyd et al. 1977). The histological study of the motor innervation of the spindle carried out by Banks (1981) established that static fusimotor neurons rarely innervated the dynamic bag₁ intrafusal muscle fibre. As an example of the painstaking work that allowed such a conclusion, Fig.2 (taken from Banks, 1981) shows a schematic reconstruction of a muscle spindle made from some 8500 sections. Efferent axons may be seen dividing and innervating both chain and bag₂ fibres. The separate innervation of the bag₁ fibre in this case by a fusimotor axon that did not innervate either the bag₂ or chain fibres reinforced the idea of separate innervation by static and dynamic axons in that it caused only a non-twitch-like contraction of sarcomeres of the bag₁ fibre and had separately been identified electrophysiologically as a dynamic axon (Banks et al. 1978). Further work (Banks, 1991), again using a powerful combination of electrophysiological experiments and histological observation, helped resolve the debate as to whether static fusimotor neurons should have two sub-classifications. It was shown that the biasing effect on spindle afferent discharge was produced by bag₂ activity alone, or in combination with chain fibres, whereas the driving action was a purely chain fibre effect.

Fig 2.

Distal and proximal poles of a mammalian muscle spindle. Primary (P) and secondary (S₁) afferent endings are shown hatched. Bag₁ (b₁), bag₂ (b₂) and four chain (c) intrafusal muscle fibres are identified. Motor nerve endings are shown as filled ovals; all of the endings on the b1 fibre are innervated by the same motor axon that was electrophysiologically identified as being a dynamic axon. Horizontal arrows indicate movements of sarcomeres on the b₁ fibre that were visible during stimulation of the dynamic axon. Upper scale – length (mm). Adapted from fig. 2 from Banks (1981) with permission.

In conclusion, the detailed anatomical and physiological knowledge of the mammalian muscle spindle reviewed here and elsewhere has been crucial in developing our understanding of the roles that spindles play in the control of movement. In the following sections, the behaviour of muscle spindles, including their efferent fusimotor innervations, during locomotion is reviewed.

Spindle afferent activity during locomotion

Afferent feedback from somatic sensory systems is essential for effective outcome of both automatic and voluntary movements. The principal systems contributing to that control come from fast-conducting afferent axons innervating muscle spindles, musculo-skeletal tendinous junctions, joints and skin. The over-riding importance of the muscle spindle in this feedback is emphasized by the fact that, special senses apart, it is the only peripheral sensory (proprioceptive, cutaneous) system that is itself controlled by the CNS.

Recordings of spindle activity during locomotion have not been achieved in humans. This has been due largely to the restrictions imposed by the only available human peripheral nerve recording technique of microneurography (Hagbarth & Vallbo, 1967). The use of microneurography has been limited to small or weak muscle contractions that would not displace the intraneural recording electrode. Nevertheless, recordings from the nerve supply to finger muscles have shown that the fusimotor drive during voluntary movements is not routinely linked to that of alpha motoneurons (Wessberg & Vallbo, 1995; Kakuda et al. 1996; Jones et al. 2001) as had been proposed by the concept of alpha–gamma linkage (Matthews, 1972). Our knowledge of the patterns of firing of muscle spindle afferents during locomotion has been drawn almost exclusively from the cat, both from intact, freely moving animals and reduced (decorticate, decerebrate) preparations.

It was clear from the earliest recordings of muscle spindle afferents during locomotion that discharges were highly modulated but the behaviour was not explicable if the receptors were acting as simple passive detectors of muscle length and velocity. Spindle recordings in decerebrate and decorticate cats (Severin, 1970; Perret & Buser, 1972; Perret & Berthoz, 1973) showed marked fluctuations in spindle activity during locomotor movements that were best interpreted as being driven by a combination of dynamic and static fusimotor activity. A major advance came from the first recordings in an intact animal. Figure3 is taken from Prochazka et al. (1977) who recorded the discharges of identified muscle spindle primary afferents (Ia) from in-dwelling wires in the dorsal roots of freely moving cats. They compared the afferent discharges during active step cycles, recorded as length of an ankle extensor muscle, with those recorded during passive movements designed to simulate the active step cycle but under deep anaesthesia to block fusimotor activity. The Ia discharge patterns are clearly different in the passive and active states with pronounced extra firing during the extensor phases and an increased duration and temporal advance during flexion. The authors were able to conclude that increased rates of Ia discharge indicated the presence of fusimotor drive that was modulated at different times during the gait cycle. Further experiments on chronically implanted freely walking cats (Loeb, 1981) revealed that spindle afferent behaviour did not conform to any single functional pattern for all muscles but depended on muscle type (flexor, extensor, bi-articular). It was proposed that dynamic fusimotor bias promoted detection of degree and rate of stretch sensed by spindles in extensor muscles and rapidly modulated static fusimotor activity preserved spindle activity during rapidly shortening contractions of flexor muscles. Although work at the time had revealed evidence of independent alpha and gamma motoneuron activity during locomotion, the concept of a rather tight alpha–gamma linkage or co-activation continued to be promoted (Cabelguen et al. 1984). A substantial degree of independent gamma motoneuron control, however, received further support with the introduction of a new concept, that of ‘fusimotor set’ (Prochazka et al. 1985). Prochazka et al. (1985) stimulated individual gamma axons, identified as either static or dynamic, while recording discharges of spindle afferents responding to locomotor movements recorded previously in other animals. They then varied the efferent stimulation patterns until they obtained the best fit to the discharge patterns from those previous experiments. This allowed them to discount the idea of predominantly alpha–gamma co-activation and conclude that the CNS ‘set’ a substantial tonic static fusimotor drive during locomotion that could be changed to a largely dynamic fusimotor drive during more unpredictable and novel movements that required patterns of loading that deviated from those occurring during stereotypic stepping. Due to continuing uncertainties over interpreting fusimotor drive from spindle afferent recordings, an attempt was made (Prochazka & Gorassini, 1998) to create a retrospective ensemble of afferent discharges from a large population of primary endings during normal stepping. Among several conclusions was the admission that components of fusimotor drive were linked to alpha motoneuron activity, at least in extensor muscles of the hind limb, possibly more so than thought previously. The review did establish first, that primary afferent discharge rates were indeed very high during normal locomotion with average discharge rates of 80 s⁻¹, reinforcing the dominant role that spindles play in afferent feedback during locomotion, and second that Ia afferents primarily signalled muscle length and (particularly) muscle velocity.

Fig 3.

Muscle spindle primary afferent discharges during active and passive (simulated) step cycles during locomotion in a freely moving cat. (a and b) Top trace: EMG of lateral gastrocnemius; middle trace: length of ankle extensors; lower trace: instantaneous firing rate of a spindle primary afferent of the ankle extensors. (A) Passive flexion and extension of ankle during anaesthesia. (B) Active step cycle with length signals well-matched to the passive movements. Locomotor cycle duration approximately 1.19 s. Adapted from fig. 1 from Prochazka et al. (1977) with permission.

The lack of a consensus of opinion in previous work suggested that a further refinement of technique was required to sort out some anomalies of interpretation of spindle discharge during locomotion that were arising from the separate impact on spindle behaviour caused by the extent and rate of mechanical stretch of the parent muscle, unloading of spindles caused by skeletal muscle contraction, and the separate influences of static and dynamic fusimotor drive. The initial approach of our team (Taylor et al. 2000a,b) was to compare the behaviour of spindles during active movements with the behaviour of the same spindles when the locomotor movements were exactly reproduced passively. We also set out to record from a substantial number of secondary spindle endings (2⁰) in addition to primaries (1⁰), as recording from 1⁰ endings had previously dominated the literature. Finally, we decided to identify the connections of both 1⁰ and 2⁰ endings to the three types of intrafusal fibre (bag₁, bag₂ and chain) using succinylcholine (Taylor et al. 1992) so as not to overlook the possibility of an influence on spindle discharge of any subdivision of the static fusimotor system (Banks, 1991; Taylor et al. 1999a). Experiments were carried out on locomoting decerebrate cats in which a number of spindle afferents from hind limb flexor and extensor muscles were fully characterized in terms of their intrafusal fibre contacts. Figure4 shows discharges recorded during active treadmill walking and, again, from the same set of afferents, during the identical ankle movements applied passively. It is clear by comparison with the modest discharge patterns under passive movements that there is a strong fusimotor drive during active locomotion. It can also be observed that extensor and flexor afferents do not discharge in a clearly reciprocal manner, suggesting that the patterns of fusimotor discharges are different to flexor and extensor muscles. The study did reveal subtle differences in the way afferent endings with specific combinations of terminals on bag₁, bag₂ and chain fibres behave during locomotion; just two findings will be highlighted here.

Fig 4.

Spindle discharges recorded in a decerebrate cat during treadmill locomotion Three legs walked on the treadmill, while the left hind limb was fixed and allowed to rotate only at the ankle. The two top records are smoothed rectified EMG signals from left tibialis anterior (TA) and medial gastrocnemius (MG). Successively below these are instantaneous frequency records (in impulses s⁻¹) from an MG and a TA spindle afferent. In each case the afferent type in terms of intrafusal fibre contacts is indicated together with afferent conduction velocity. The next record below these is of ankle rotation. The next sequence of recordings is a repeat of the above while the ankle is moved passively through the exact same set of excursions with the animal deeply anaesthetized. Vertical dashed and solid lines, respectively, indicate the beginning and end of ankle extension. Figure 2 from Taylor et al. (2000a) with permission.

First, Fig.5 shows results from 2⁰ afferents (bag₂ and chain) obtained by subtracting the firing rates recorded under passive rotation of the ankle, exactly mimicking the active movement, from those recorded under active stepping. It was anticipated that the result would reflect the profile of static fusimotor drive during active locomotion. The difference records from the tibialis anterior units show changes in firing rate inversely linked to muscle length changes, i.e. firing rate increases as the muscle shortens, and vice versa. The same is the case for the gastrocnemius medialis unit, but with some phase advance. There was a very poor relationship between these difference records and the electromyogram (EMG) generated by the muscles. The concept was proposed that a modulated fusimotor drive might represent a temporal template of the required movement that is fed back to the CNS through 2⁰ afferent discharges (Taylor et al. 1999b). Any unexpected changes in external loading could then be signalled by a departure from this template. It should be stressed that the expected movement in these experiments was not the same as the movement that would be expected during weight-bearing locomotion. The parent limb for the afferents was largely denervated. It was supported above the treadmill and the ankle simply allowed to flex and extend, bearing only its own weight. During normal locomotion the ankle extensor, for example, would undergo different periods of unloaded and loaded (weight-bearing) shortening and lengthening. Nevertheless, the drive to the test muscles, which behaved largely isotonically, may well not have been simply what was required to achieve such movement. The opposite hind limb and, indeed, the forelimbs were weight-bearing to some extent. The drive to the test muscles might well have been influenced by the central drive required to provide the movements of the other three limbs, i.e. not simply that required to move the ankle in the relative absence of weight-bearing.

Fig 5.

Difference records of the ensemble mean firing rates of two secondary spindle afferent firing for the tibialis anterior (TA) and one secondary ending for the gastrocnemius medialis (GM) muscles from a decerebrate cat related to movement during treadmill locomotion. (A and B) The solid dots represent firing rates calculated as differences between the discharges during active and (equivalent) passive movements. Solid lines are the angular rotation of the ankle with arrowheads showing the direction of muscle lengthening. Adapted from figs 4 and 9 from Taylor et al. (2000a) with permission.

A second striking finding was the pattern of firing of b₁b₂c 1⁰ endings, for both ankle flexor and extensor muscles, in relation to the start of muscle lengthening. Afferent difference (active minus passive) records showed a distinct burst of activity attributable to fusimotor drive (see figs 7B and 9C from Taylor et al. 2000a). This burst of firing at the start of muscle lengthening was not seen for b₂c 1⁰ endings (e.g. see fig. 7A from Taylor et al. 2000a). Our speculation at the time was that this burst was supported by increased dynamic fusimotor drive, although there was no evidence as to whether that dynamic drive was steady or fluctuated during the locomotor cycle. Further experiments recording directly from gamma efferents, identified as either static or dynamic (see below), were required to establish the exact pattern.

Fig 7.

Pattern of gastrocnemius medialis dynamic gamma motoneuron firing during locomotion. (A) An ensemble of the discharges from four gastrocnemius medialis b₁b₂c primary afferents (conduction velocities all > 95 ms⁻¹) recorded during locomotion. Ankle angle signalled by solid trace (superimposed arrowhead shows gastrocnemius medialis lengthening). The afferent firings are the difference records between active and passive movements during locomotion. Arrows show two peaks of afferent firing that can be attributed to gamma efferent discharge during rapid lengthening (dynamic gamma support) and at the end of lengthening (static gamma support; see text for details). (B) Data from an experiment in which two gastrocnemius medialis dynamic gamma units were recorded simultaneously. Top: 15 superimposed ankle movement traces normalized with respect to locomotor cycle length (0.96 s). Below: the firing of two dynamic gamma motoneurons shown as 15-dot rasters, which reveals the interrupted nature of their firing. Adapted from fig. 9 from Taylor et al. (2000a) and fig. 8 from Taylor et al. (2000b) with permission.

Fusimotor activity during locomotion

The lingering uncertainties of identifying static and dynamic fusimotor patterns from spindle afferent recordings that were also influenced by muscle length changes could only be resolved by making recordings from electrophysiologically identified gamma motoneurons. The first direct recordings from gamma motoneurons, identified by axonal conduction velocity (< 45 ms⁻¹), in locomoting decerebrate cats (Murphy et al. 1984) revealed two different patterns for the ankle extensor muscles, one highly modulated and the other essentially tonic. The firing of the phasic units tended to be linked to the discharge of alpha motoneurons. When the cat was not locomoting, the two types of gamma motoneuron had different and largely non-overlapping resting discharge rates. By indirect means, an association was made between high resting rates, highly modulated discharge during locomotion and dynamic nature of gamma efferents, whereas low resting rates and non-modulated discharge during locomotion were ascribed to static efferents. When Murphy & Hammond (1993) recorded from gamma efferents supplying a denervated ankle flexor muscle, they found the converse behaviour for putative dynamic (tonic) and static (phasic) efferents, and neither type exhibited a resting discharge before locomotion commenced. The conundrum of why static and dynamic gamma efferents should apparently have reversed roles in extensor and flexor muscles acting at the same joint was discussed but not resolved.

The experiments conducted by Taylor et al. (2000b) used a more robust technique to identify gamma motoneurons recorded during decerebrate locomotion, by noting simultaneously the behaviour of both gamma efferents and spindle afferents subjected to ramp and hold stretch of the parent muscle when certain areas of the midbrain close to the mesencephalic locomotor region were stimulated. Figure6 shows one result using that method, clearly identifying the static or dynamic type of the gamma motoneurons. Three patterns of gamma motoneuron activity emerged from the study. All static efferents were modulated in a time-dependent fashion with the locomotor cycle, but with two distinct classes. Approximately two-thirds fired with a modulation increasing in parallel with the active muscle shortening in ankle extension and which was linearly well-matched to the time course of the difference of active minus passive firing of the spindle secondary afferents, as had been postulated from earlier analysis of spindle secondary afferent recordings alone (Taylor et al. 2000a). A phase advance of this pattern of static firing, superimposed on a tonically raised level of static activity, in the case of the ankle extensor was also confirmed by the direct gamma recordings. The remaining one-third of static efferents again showed firing throughout the locomotor cycle but with a more modest fluctuation that increased smoothly during ankle flexion, for the ankle extensor muscle. Arguments were presented that these static gamma patterns might support the notion of separate control over the bag₂ and chain intrafusal muscle fibres. The pattern of discharge of identified dynamic gamma efferents was very different to that of the static efferents. They had an on–off type or interrupted type of discharge that turned on abruptly at the onset of muscle shortening and was sustained at a high level until just after the start of muscle lengthening. The pattern is illustrated in Fig.7B. As can be seen from the accompanying primary afferent recordings (Fig.7A), the dynamic discharge is appropriately timed to support a burst of primary afferent discharge that would signal the onset of muscle stretch. The importance of timing for this effect was strengthened by further experiments monitoring b₁b₂c 1⁰ afferent discharges to sinusoidal stretch of the parent muscle during interrupted electrical stimulation of identified dynamic efferents. Figure 9 in Taylor et al. (2000b) shows prominent peaks in 1⁰ afferent discharge when dynamic efferent stimulation continued just into the start of lengthening. These patterns of discharge of static and dynamic gamma efferents in an ankle extensor muscle were subsequently found also for the ankle flexor, tibialis anterior (Taylor et al. 2006), the only difference being the lack of a clear separation of two classes of static efferent discharge.

Fig 6.

Identification of gamma axons by stimulation of the midbrain. The upper three traces show the firing of three single gamma axons from a small filament of the gastrocnemius medialis nerve, which was otherwise intact (static or dynamic type and conduction velocity, in ms⁻¹, are shown above each trace). The bottom two traces show the firing of two gastrocnemius medialis spindle afferents; their intrafusal fibre contacts and conduction velocity (in ms⁻¹) are shown to the left, above the traces. All records are instantaneous frequency plots. Ramp and hold stretches of gastrocnemius medialis (not shown) were applied continuously every 6 s. The spindle records show that repetitive midbrain stimulation signalled by the horizontal bar increases dynamic and decreases static spindle afferent responses to stretch, and so allows identification of the three gamma axons. Figure2 from Taylor et al. (2000b) with permission.

Conclusions

In the studies described above (Taylor et al. 2000a,b, 2006), we postulate that the patterns of static gamma efferents during locomotion in the decerebrate cat may involve selective innervation of bag₂ and chain intrafusal muscle fibres, and that their combination is able to provide a ‘temporal template’ of the expected movement for this particular type of locomotion (decerebrate preparation) and in the largely absent conditions of weight-bearing for the test muscles. This would act to expand the dynamic response range of the spindles. The dynamic gamma efferent pattern appears to sensitize primary afferents critically to detect the onset of muscle lengthening and to detect departures from the trajectory of the intended movement. These patterns are similar for antagonist extensor and flexor muscles acting at the ankle, despite the rather different roles that such muscles play during locomotion. The findings of different patterns of activity for dynamic and two types of static gamma efferents argue strongly against a classical alpha–gamma co-activation principle in the generation of locomotion.

Conflict of interest

The authors declare that they have no conflict of interest.