Skip to main content
NCBI home page
The Journal of Physiology logoLink to The Journal of Physiology
. 2002 Aug 15;543(Pt 1):273–288. doi: 10.1113/jphysiol.2001.013229

Static γ-motoneurones couple group Ia and II afferents of single muscle spindles in anaesthetised and decerebrate cats

M H Gladden 1, H Matsuzaki 1
PMCID: PMC2290485  PMID: 12181298

Abstract

Ideas about the functions of static γ-motoneurones are based on the responses of primary and secondary endings to electrical stimulation of single static γ-axons, usually at high frequencies. We compared these effects with the actions of spontaneously active γ-motoneurones. In anaesthetised cats, afferents and efferents were recorded in intramuscular nerve branches to single muscle spindles. The occurrence of γ-spikes, identified by a spike shape recognition system, was linked to video-taped contractions of type-identified intrafusal fibres in the dissected muscle spindles. When some static γ-motoneurones were active at low frequency (< 15 Hz) they coupled the firing of group Ia and II afferents. Activity of other static γ-motoneurones which tensed the intrafusal fibres appeared to enhance this effect. Under these conditions the secondary ending responded at shorter latency than the primary ending. In another series of experiments on decerebrate cats, responses of primary and secondary endings of single muscle spindles to activation of γ-motoneurones by natural stimuli were compared with their responses to electrical stimulation of single γ-axons supplying the same spindle. Electrical stimulation mimicked the natural actions of γ-motoneurones on either the primary or the secondary ending, but not on both together. However, γ-activity evoked by natural stimuli coupled the firing of afferents with the muscle at constant length, and also when it was stretched. Analysis showed that the timing and tightness of this coupling determined the degree of summation of excitatory postsynaptic potentials (EPSPs) evoked by each afferent in α-motoneurones and interneurones contacted by terminals of both endings, and thus the degree of facilitation of reflex actions of group II afferents.

Ruffini (1898) discovered that muscle spindles possess two types of sensory ending enclosed in the one receptor, and called them primary and secondary sensory endings. These endings provide information about muscle length that is essential for the normal control of movement. γ-Motoneurones change the responses of the sensory endings to muscle stretch by modifying the mechanical conditions within muscle spindles. The forces that deform the endings are transmitted to the sensory endings along three types of intrafusal muscle fibre, each possessing unique mechanical properties and innervated by γ-motoneurones. The changes made to the sensory output when each type of fibre contracts were established by stimulating single γ-axons that innervated one fibre type only (see Boyd, 1981). In brief, the primary sensory ending responds to contraction of all three fibre types (bag1, bag2 and chain fibres), whereas the secondary ending responds to contractions of chain fibres but not, or only very weakly, to contractions of bag1 and bag2 fibres. Contractions of chain fibres can ‘drive’ the primary sensory ending – each γ-spike sets up a twitch contraction that extends the primary ending and elicits a spike – but do not drive secondary endings. Our aim was to find out how natural activity of γ-motoneurones affects the output of the sensory endings. We recorded the responses of the two kinds of ending from the same spindle simultaneously since this could provide a better insight into how fusimotor activity is governing the mechanical state of the spindle than is provided by analysing changes in responses of primary and sensory endings from different spindles. We found that when γ-motoneurones are naturally active they can break most of the rules that we had learnt from stimulating single γ-axons. Under these conditions, not only could the secondary endings be driven, but they were driven more securely than the primary endings and at shorter latency. In consequence, pairs of afferent spikes left the muscle spindle with the secondary spikes in the lead. If the spikes from the primary endings are in the lead, the lower conduction velocity of group II axons from secondary endings makes it impossible for spikes from the two endings to arrive approximately simultaneously at common target cells in the spinal cord, so their respective EPSPs cannot summate. However, as will be demonstrated, when the output of the endings is coupled with the spikes from the secondary endings in the lead, an effective summation of EPSPs can occur, and the secondary endings can potently facilitate the responses of α-motoneurones. Importantly, this type of coupling was also recorded during changes of muscle length.

The role of secondary sensory endings in movement control has been difficult to define. It has often been assessed by manoeuvres that remove the contribution of the primary sensory ending, for example by pressure block, or by relying on the timing of the earlier arrival of group Ia impulses from stretched muscles. This investigation shows that neither of these approaches is reliable since they do not take into account the coupling actions of static γ-motoneurones. Preliminary reports of some of the present work have been published (Gladden & Matsuzaki, 2001, 2002).

Methods

Two preparations were used to investigate the actions of γ-motoneurones excited by natural stimuli. Firstly, the impulse traffic to and from single muscle spindles in the tenuissimus muscle was recorded in anaesthetised cats with intact innervation (Fig. 1A). In this preparation γ-motoneurones were excited by small stretches applied to other muscles of the same hindlimb. In the second preparation (Fig. 1B), the responses of the primary and a secondary ending of a single muscle spindle were recorded in decerebrate cats, initially with the ventral roots intact. In order to record from single axons in the spinal roots it was necessary to denervate the whole hindlimb except for the muscle containing the single muscle spindle. Therefore the γ-motoneurones could not be excited by stretching other muscles of this limb. Instead they were excited by stimulating nociceptors in a forelimb. The ventral roots were then cut, and single γ-motor axons innervating the same spindle were stimulated to compare the fusimotor effects with those caused by natural excitation. All procedures accorded with current UK legislation.

Figure 1. The two experimental arrangements.

Figure 1

Open in a new tab

A, nerve impulses to and from a single muscle spindle were recorded with a pair of electrodes on the intramuscular nerve branch to the spindle. Afferent and efferent spikes were separated by a spike shape recognition system. Conduction velocities were estimated by spike-triggered averaging records from a second pair of electrodes on the main muscle nerve. γ-Motoneurones were excited by pulling on muscles in the ipsilateral hindlimb. B, afferents from a single muscle spindle were recorded in dorsal root filaments. γ-Motoneurones were activated by stimulating nociceptors in a forelimb. Later, after cutting the ventral roots, single γ-axons that supplied the spindle were stimulated to find out how closely this would mimic the previously recorded natural γ-activity.

Anaesthetised preparation with intact innervation

Cats weighing 2.1-3.0 kg of either sex were anaesthetised initially with pentobarbitone sodium (Sagatal, Rhône Mérieux Ltd, Harlow, UK; 45 mg kg−1) intraperitoneally. In all experiments anaesthesia was assessed regularly by observing pupil size, and testing for reflex withdrawal, and supplemented intravenously whenever required. Operative trauma was kept to a minimum because it may depress spinal reflexes through the release of endogenous opiates in barbiturate anaesthesia (see Duggan et al. 1985). The tenuissimus muscle was exposed by reflecting the biceps muscle which was kept undamaged by cutting along its fascial insertion. The lower end of the tenuissimus muscle was dissected free from the biceps and exteriorised into a bath containing Krebs solution of composition (mm): NaCl 115; KCl 4.6; CaCl2 2.5; MgSO4 1.2; NaHCO3 24.1; KH2PO4 1.2; and glucose 1 g l−1 at room temperature. This solution was equilibrated with a 95 % O2, 5 % CO2 gas mixture. The last spindle in the tenuissimus muscle was identified (see below) and exposed. Sufficient length of the intramuscular nerve branch just proximal to this spindle was dissected to drape over bipolar recording electrodes. The main nerve to the tenuissimus muscle was also dissected free and placed on another pair of electrodes. Movement of the intrafusal fibres was viewed with a Leitz Diavert (Wetzlar, Germany) inverted microscope and recorded on videotape.

The decerebrate preparation

Anaesthesia was induced with xylazine equivalent (Rompun, Bayer UK Ltd, Bury St Edmunds, UK; 1.1 mg kg−1) and ketamine (Vetelar, Parke, Davis & Co., Pontypool, UK; 22 mg kg−1) intramuscularly or with pentobarbitone sodium (Sagatal; 45 mg kg−1) intraperitoneally. Until the animal was decerebrated (which rendered it insentient) anaesthesia was maintained with either chloralose, 40 mg kg−1 dissolved in distilled water at 70 °C, injected slowly intravenously, or pentobarbitone sodium i.v. For the decerebration, both carotid arteries were ligated, and the midbrain sectioned immediately anterior to the superior colliculi and nucleus ruber. All brain tissue rostral to this section was removed. The physiological state was monitored from the blood pressure through a cannula in the left carotid artery, the fractional end-tidal CO2, and the core temperature. Between 5 and 10ml of 10 % dextran (Rheomacrodex) was given slowly intravenously every 8–12 h. After both types of experiment all animals were killed by giving an overdose of anaesthetic intravenously.

In general, for the identification of muscle afferents and efferents the methods described by Boyd et al. (1977) were followed, but adapted where necessary. The left hindlimb was denervated except for the tenuissimus muscle, and the nerve of this muscle was dissected free of connective tissue and placed on stimulating electrodes. Following a laminectomy to expose the cauda equina, the L7 and S1 dorsal roots were searched for afferents. All afferents considered to be from primary endings had conduction velocities above 80 m s−1. All secondary endings studied had conduction velocities below 60 m s−1. About as many group Ia as II axons were found usually, but the total number of fibres per cat varied, the highest number found being 17 Ia and 19 group II afferents.

The afferent endings were located by moving a stimulating electrode along the length of the muscle in 1 mm steps and testing the electrical stimulus threshold. In the strap-like tenuissimus muscle, spindles are arranged in a line end-to-end along the middle, so as the electrode passed beyond the spindle which contained the endings of the afferents the stimulus threshold began to increase. This method differs slightly from that devised by Bessou and Laporte (1962) who used electrical stimulation and very localised pressure to locate the spindles, without measuring changes in electrical threshold.

The responses of the primary and a secondary ending of a single spindle to γ-excitation were recorded simultaneously. After having excited the γ-motoneurones innervating the chosen spindle (see below) by natural stimuli the ventral roots were cut, and single fusimotor axons in ventral root filaments that innervated the tenuissimus muscle were identified by stimulation of the muscle nerve. Fusimotor axons which elicited a response from the chosen primary and secondary spindle afferents were then stimulated. The frequencies used were 10, 20, 35, 50, 75, 100 and 150 Hz, or 10, 20, 30, 40, 50, 75, 100 and 150 Hz. A ramp frequency (Dickson et al. 1993) was also applied at constant length. Pulses from a Geiger counter acted as triggers for random stimulation. When two γ-axons were stimulated simultaneously, separate Geiger counters were used. The axons were also stimulated at 50, 75 100 and 150 Hz during a ramp-and-hold stretch of the muscle. This served to identify the axons as dynamic or static.

Muscle length

The left hindlimb was placed in an intermediate position during experiments, being neither fully flexed nor fully extended, nor abducted at the hip, as abduction shortens the tenuissimus muscle because it crosses the hip joint. The tenuissimus is very compliant; a 4 cm-long portion close to the nerve entry where the spindles studied were located extended by 30–37 % between full flexion and full extension of the hindlimb. Stretches applied to the ends of the muscle were therefore not well transmitted to these spindles. To apply stretch, the muscle was gripped with a metal chuck close to the chosen spindle. The chuck was attached to a steel rod projecting from the shaft of an electromagnetic length servo. For ramp-and-hold stretches the stretch amplitude was 2 mm, and the rate of pull was 4, 6, or 8 mm s−1. In some experiments random length changes were applied in the same way and with the same maximum amplitude. The relation between log plots of power spectral density against frequency of stretching was flat up to a certain frequency – that is, all frequencies below this frequency were equally represented. In the experiments illustrated in Fig. 6 and Fig. 7 this frequency was 36 and 55 Hz respectively.

Figure 6. Afferents coupled during stretching.

Figure 6

Open in a new tab

A, instantaneous frequencies of spikes from primary (P) and secondary (S) endings with random length input beginning at 70 s (same record as in Fig. 5A before 70 s). B, cross-correlogram between S and P spikes during random stretching (70-250 s in A) shows that the majority of P spikes followed S spikes. C, representative section of record with expanded time-scale (taken from about 170 s) showing P and S spikes in relation to muscle stretches. D, cross-correlograms between the peaks of the stretches and each afferent. E, same as D but for test spindle stretched passively. F, cross-correlograms between the peaks of stretches (at dotted line) and first P and S spike in response to each stretch, showing relative timing as spikes left the spindle.

Figure 7. Random length change and random stimulation of static γ-axons in test spindle.

Figure 7

Open in a new tab

A, cross-correlograms between spikes from secondary (S) and primary (P) endings (upper row), and S auto-correlograms (lower row), during length change alone and with random stimulation of two γs-axons: γ1 innervated fibres in the opposite pole to the location of the secondary ending, and γ2 innervated fibres at the same pole (same spindle as in Fig. 4). B, cross-correlograms constructed between the peaks of the stretches and P and S spikes (length), and during random stimulation of γ1 (length + γ1) and γ2 (length + γ2). The cross-correlograms in columns 4 and 5 are between the stimuli applied to the γ-axons and P and S spikes during random length change. All correlograms were normalised for spike rate. Bin widths 0.5 ms.

Identification of contracting intrafusal fibres

At the end of some experiments the spindle was exteriorised and superfused in a bath as described below for the anaesthetised preparation. After exposing the spindle, single γ-axons that supplied it were stimulated to find out which intrafusal fibres they innervated. It was important to establish the relation of the contracting fibres to the secondary ending, whether at the same or the opposite side of the primary sensory ending. Chain fibres were distinguished from bag fibres by their smaller diameters, and their tendency to buckle at the end of the fluid space when the spindle is held at short length (Gladden, 1976). Dynamic bag1 fibres were distinguished from static bag2 by their differing mechanical behaviour; the amplitude of movement of bag1 fibres is much less than that of bag2 fibres within the fluid space in these slow tonic-type fibres (see Boyd et al. 1977). The same criteria were used to identify intrafusal fibres in the anaesthetised preparation, when the intrafusal fibres were contracting under the control of γ-motoneurones (see below). Under those conditions it is difficult to detect activity if the γ-motoneurones discharge at steady frequencies. The movement of dynamic bag1 fibres characteristically appears jerky during natural γ-excitation, although they contract smoothly when γ-axons innervating them are stimulated at constant frequency. This jerkiness may reflect the breaking of resting actin-myosin bonds by contractions induced by sporadic changes in fusimotor excitation.

Stimulation of nociceptive afferents

The aim of this stimulation was to excite γ-motoneurones by pinching the forepaws and thus eliciting a withdrawal reflex. Under these conditions γ-motoneurones could have been responding to nociception, to arousal, or as a part of postural adjustments.

Histology

At the end of some experiments spindles were labelled with threads to facilitate the later identification of proximal and distal poles. They were fixed in 3 % glutaraldehyde in phosphate buffer, stained with osmium, and dehydrated before embedding them in resin. Then specimens were cut transversely into serial 1 μm-thick sections for light microscopy. From these the innervation of the spindles could be reconstructed by following the afferent and efferent axons to their endings on the intrafusal fibres. In this way one could establish what proportion of the motor innervation had been identified functionally in the experiments.

Analysis

Spike trains were digitised with a 1401plus interface (Cambridge Electronic Design, Cambridge, UK) and analysed with SPIKE 2 software (version 4; Cambridge Electronic Design). The recordings were examined before analysis and sources of spurious interspike intervals eliminated. Data were exported to Microsoft Excel for further analysis.

When recording conditions were optimal in the anaesthetised preparation, afferent spikes in two or three fibres and efferent spikes in up to five fibres could be distinguished in the nerve supply of the spindle. Action potentials were first identified electronically by the SPIKE 2 wavemark facility, and then allocated to individual channels. Cross-correlograms between spike trains, and auto-correlograms were constructed with bin widths of 0.5 or 0.1 ms. The efficacy with which a spike train induced peaks in a cross-correlogram was calculated as described by Swadlow & Gusev (2001). The number of spikes was counted in a window 2.5 ms on either side of a peak in the cross-correlogram. The number of spikes that would occur by chance during this period was subtracted. This was estimated from the mean number of spikes in bins between 100 and 50 ms before the peak. The remainder was divided by the number of spikes used to construct the cross-correlogram. Joint interval histograms were constructed as described by Rodiek et al. (1962).

Results

The results were derived from 13 experiments on anaesthetised cats and 8 experiments on decerebrate cats. During experiments on single muscle spindles in the anaesthetised preparation between one and five static γ-motoneurones (usually three) became active, phasically or tonically, and evoked contractions in chain fibres and/or bag2 fibres at some stage. There was no obvious pattern to the recruitment of these fibre types in the proximal or distal poles. The total number of γ-motoneurones supplying the tenuissimus muscle is between 5 and 15 (Lev-Tov et al. 1988), so that the number thus studied was a significant proportion of the whole population. Spontaneous activity of a dynamic γ-motoneurone was detected in only one of the 13 experiments.

Static γ-motoneurones driving the output of the secondary sensory ending

At the start of recordings from the muscle spindle illustrated in Fig. 2 only one γ-motoneurone was spontaneously active. Additional γ-motoneurones were activated by pulling on other hindlimb muscles (caudo-femoralis, gastrocnemius and soleus muscles). As each γ-motoneurone was recruited the appearance of its characteristically shaped spike in the neurogram was correlated with the intrafusal contractions it elicited. At first two additional γ-motoneurones were activated phasically, and then one of these became tonically active (after upward-pointing arrow in Fig. 2A).

Figure 2. Spikes from a primary and a secondary ending coupled by fusimotor activity.

Figure 2

Open in a new tab

A, discharge frequencies of a primary (P) and secondary (S) ending in a single muscle spindle of the tenuissimus muscle, and of three static γ-motoneurones (γ1-3) that innervated it. Initially one static γ-motoneurone (γ2) was tonically active. Two other static γ-motoneurones (γ1 and γ3) were recruited phasically by reflex activation – repeated small stretches of the caudo-femoralis and gastrocnemius muscles (5-90 s and 115–130 s respectively). After stretching the soleus muscle, γ1 became tonically active (from upward-pointing arrow to 250 s). B, upper row, cross-correlograms between γ1 and P and S discharges after γ1 became tonically active show S spikes time-locked to γ1 spikes more securely and at shorter latency than P spikes (columns 1 and 2). Thus S and P spikes were coupled, with S spikes leading (column 3). Lower row, cross-correlograms between γ2 and P and S discharges for the whole period of this record show no relation (columns 1 and 2). When γ2 was active alone (between the downward-pointing arrows in A) the afferents were not coupled (column 3). All correlograms were normalised for spike rate with bin widths of 0.5 ms.

Cross-correlograms were constructed between the fusimotor spikes and spikes from the two sensory endings, and also between spikes from the secondary and primary endings. When the single γ-motoneurone was active neither ending was driven, and there was no coupling between them (Fig. 2B, lower row). However, after the second γ-motoneurone became spontaneously active it drove the secondary ending (Fig. 2B, centre of upper row). It also drove the primary ending, but less securely (compare the sharpness of the peaks in the first and second columns of Fig. 2B, lower row), and with a longer mean latency – 21.5 ms compared with 17.5 ms. Each spike from the secondary ending was therefore followed at a variable latency by a spike from the primary ending (see Fig. 9D).

Figure 9. Facilitation in spinal pathways from muscle spindle afferents by coupled output from secondary and primary endings.

Figure 9

Open in a new tab

A, diagram of branching of an axon from a secondary ending S. After leaving the dorsal root (dr) and subdividing into an ascending and descending branch it gives off several deeper projecting axon collaterals. Arrows indicate direct connections with α- and γ-motoneurones, and with two subpopulations of interneurones in the intermediate zone (Ii) and the dorsal horn (Di). The interneurones are interposed in disynaptic (Ii) and trisynaptic (Di and Ii) pathways between group II afferents and α-motoneurones. Both α-motoneurones and Ii interneurones are co-excited by the secondary (S) and primary (P) endings. Coupling between these endings will facilitate their synaptic actions at sites indicated by the coloured arrows provided that they are properly timed, i.e. the slower conducted S spikes leading at the level of the muscle spindle. B, diagram showing the estimated period of effective summation between the EPSPs: when they coincide and when the P EPSP precedes the S EPSP, or vice versa, by less than their half-width. C, diagram showing that in α-motoneurones in which EPSPs are evoked by all three pathways the window for effective summation (green line) is increased (to about 11 ms). D, afferent couplets leaving the spindle after four successive spikes of the coupling γ-motoneurone; as the green line indicates, all the larger P spikes occur within the time window for effective summation shown in C. E, cross-correlograms between 100 S and associated P spikes. The blue arrow indicates the position in the correlograms of a P spike that would arrive simultaneously at an α-motoneurone with an S spike, taking into account the faster peripheral and central conduction times of the P spike. The red and turquoise arrows show timing for spikes evoked through the di- and trisynaptic routes respectively. The total numbers of red spikes in each correlogram indicate the percentages of P spikes whose EPSPs would have summated on α-motoneurones with EPSPs from the interneurones Ii for the disynaptic pathway (red arrows in A). If EPSPs were evoked in the same α-motoneurone through the mono- and trisynaptic pathways in addition, the increased numbers summating are indicated in blue and turquoise respectively. See text for further explanation.

A similar entrainment of the output of the secondary ending was found in two other preparations. Coupling between the secondary and primary endings occurred during some periods of these recordings, with the spikes of the secondary ending leading those of the primary ending. In one case the secondary ending was also coupled with the primary ending of a neighbouring spindle which did not have a secondary ending.

Firing rate and the efficacy of coupling

To find out how the frequency of discharge of a γ-motoneurone affected the ability to drive the secondary ending and couple the afferent firing, γ-spikes were divided according to inter-spike intervals into three categories, < 10 Hz, 10–15 Hz and > 15 Hz, and cross-correlograms were constructed between spikes in each group and the afferents. In the case of the γ-motoneurone that could induce coupling illustrated in Fig. 2, both endings were entrained even when the γ-motoneurone was active at frequencies below 10 Hz (Fig. 3A) but the entrainment increased at higher frequencies. Above 10 Hz the primary ending tended to fire in doublets, giving two peaks in the cross-correlograms (Fig. 3A, upper row), and as the γ-frequency increased, the latencies for both the first and second peaks became shorter. The efficacy of coupling between the γ- and afferent spikes at each frequency was calculated. The efficacy was the ratio of the number of afferent spikes in a peak (minus the number that would occur in the same time window by chance) and the number of γ-spikes used to construct the correlogram. Figure 3B shows that the efficacy was inversely related to the latencies for the peaks in the cross-correlograms between the γ- and afferent spikes (first peak only for the primary ending). The outcome is that as the frequency of the γ-motoneurone that couples the afferents rises, more spikes from the primary ending will follow spikes from the secondary ending at latencies appropriate for the summation of EPSPs in common target cells in the spinal cord (see later).

Figure 3. γ-Motoneurone frequency and the security of coupling.

Figure 3

Open in a new tab

A, inter-spike intervals of the γ-motoneurone which coupled the afferents in Fig. 2A (γ1) were divided into three categories corresponding to frequencies < 10 Hz, 10–15 Hz and < 15 Hz. Cross-correlograms are from the spikes of this γ-motoneurone to P and S spikes (upper and lower rows respectively). Note that S spikes were almost optimally time-locked when the γ-motoneurone frequencies were between 10 and 15 Hz. B, the latencies of the peaks for the three frequency ranges shown in A were inversely related to the efficacy of the γ-motoneurone to afferent coupling (open squares: < 10 Hz; open diamonds: 10–15 Hz; filled circles: > 15 Hz). C, location and type of intrafusal fibres innervated by γ1 and γ2 whose frequencies are displayed in Figure 2; b1, b2 and ch: bag1, bag2 and chain fibres respectively; P and S: primary and secondary endings. The third γ-motoneurone (γ3 in Fig. 2) innervated a single chain fibre in the same pole as γ1. For simplicity the sensory endings are not indicated: primary sensory endings have terminals on all fibres, whereas the distribution of secondary endings to b1 and b2 fibres varies (see Banks et al. 1982).

The mechanism of coupling at constant length

Secure coupling of the afferent firing appeared to require contractions of the bag2 fibre in the opposite pole to the secondary ending. Contractions of chain fibres alone did not induce any coupling of afferent firing (9 cases), whether the contractions were in the same or in the opposite pole to the secondary ending. Sometimes there was some weak correlation between the γ-spikes and the firing of a secondary ending, but this occurred only when most or all of the chain fibres in one pole were innervated, and the γ-motoneurone that supplied them was firing continuously (mean firing rate 13 Hz).

It is very surprising that contraction of bag2 fibres could have a strong effect on secondary endings, and at low frequencies. As mentioned in the Introduction, when bag2 fibres are activated by stimulating single γ-axons that innervate them at high frequencies (>20 Hz) the contraction has little effect on secondary endings. It is even more surprising that contractions of a bag2 fibre on the other side of the primary ending should excite the secondary ending more effectively than the primary ending itself (see Discussion). Part of the explanation appears to be that a tensing of the intrafusal bundle set up by steady discharges of other γ-motoneurones created favourable mechanical conditions to enhance the action of bag2 fibres on the secondary endings. For example, in the muscle spindle illustrated in Fig. 2, weak but steady contractions of the bag2 and chain fibres were observed in the same pole as the secondary ending (see Fig. 3C), set up by the γ-motoneurone that did not couple the afferent discharges (γ2 in Fig. 2 and Fig. 3). It may be significant that the putative tensing action occurred in the opposite pole to the one that was innervated by the γ-motoneurone which coupled the afferent firing. When intrafusal fibres contract in one pole of a muscle spindle, the non-contracting portions of the fibres in the other pole will extend passively. The contracting portions of the fibres will shorten, and the contraction will become weaker because of the length/tension relation. The problem will be worst in full flexion because the active sarcomeres are at their shortest physiological length. Thus the effectiveness of an intrafusal contraction should be increased by intrafusal contractions that stop fibres in the opposite pole from extending passively, or by extending the entire muscle spindle. It was impossible to test this idea in the anaesthetised preparation because the firing rates and recruitment of the γ-motoneurones could not be controlled directly. Instead, tests were applied to a muscle spindle in the decerebrate preparation.

Tests to clarify the mechanism of coupling at constant length

The innervation of the muscle spindle used for these tests is shown in Fig. 4A. Two static γ-axons were stimulated. One innervated the bag2 and chain fibres in the opposite pole to the secondary ending, and thus should have entrained its discharge in favourable mechanical conditions. The other caused strong contractions of the bag2 and chain fibres in the same pole as the secondary ending, and weak contraction of the bag2 in the opposite pole. It was used to tense the intrafusal fibres. When stimulated at constant frequencies this static γ-axon had the stronger effect (Fig. 4B).

Figure 4. Test of the idea that static γ-motoneurones coupling the spindle afferents rely on mechanical conditions set by other static γ-motoneurones.

Figure 4

Open in a new tab

An analysis was made to test the idea that the action of a γ-motoneurone which could couple spindle afferent discharges (as in Fig. 2) was enhanced by tensing the intrafusal bundle, either by stretching the spindle, or by tonic intrafusal contraction evoked by a second γ-motoneurone. A, diagram of the innervation of the test muscle spindle. By comparison with the spindle illustrated in Fig. 3C, γ1 should have a coupling action, and γ2 could be used to tense the intrafusal fibres. B, effects of stimulating γ1 (open squares) and γ2 (filled diamonds) with the constant frequencies indicated; under these conditions the secondary ending was not entrained by stimulating either axon. C, cross-correlograms of effects during random stimulation; γ1 could weakly entrain both afferents and couple them, better when the muscle was extended (cf. columns 1 and 3), whereas γ2 could not at either length (cf. columns 2 and 4); simultaneous tonic stimulation of γ2 at 40 Hz (column 6) much improved the ability of γ1 to drive P but not S, presumably due to occlusion by direct excitation from γ2. Simultaneous random stimulation of γ2 was not effective (column 5). For further description see text. The values below each correlogram are for efficacy. D, joint interval histograms. Plots of interspike intervals (τ1) against immediately preceding interspike intervals (τ2) show that S tended to fire in bursts when γ1 was stimulated randomly (graphs 1 and 2). Neuromuscular block was an unlikely cause because bursting occurred irrespective of muscle length. Bursting of the secondary discharge did not occur without γ-stimulation (graph 4, S rest), or when γ2 was stimulated randomly (graph 3), or when random stimulation of γ1 was combined with tonic stimulation of γ2 at 40 Hz (graph 4).

When the γ-axon that innervated fibres in the opposite pole to the secondary ending (γ1 in Fig. 4) was stimulated with constant low frequency pulses, in the output range of γ-motoneurones that can couple the afferent discharge (10 or 20 Hz) for periods of 2 s, it did not entrain the secondary ending. However, it did weakly drive the secondary ending, and couple both endings when stimulated for longer periods with randomly varying frequencies (Fig. 4C). The advantage of random stimulation is that a wide range of frequencies can be tested in a very much shorter time than if the same range of constant frequencies is tested separately. The coupling was more effective when the muscle was in full physiological extension. Interestingly, the second γ-axon that caused contractions at the same pole as the secondary ending did not entrain the secondary ending or cause coupling, even though it had a stronger action on both endings. This confirms the observations of the actions of γ-motoneurones in the exteriorised muscle spindles described above – that strong contractions of bag2 fibres in the opposite pole to the secondary ending are necessary for the afferent discharge to be coupled.

In the test case, the idea that a γ-motoneurone can extend the sensory endings more effectively if fibres in the other pole are tonically contracting was tested by stimulating the two γ-axons innervating the test spindle together. When they were both stimulated at random frequencies, the secondary ending was driven less effectively, even if the muscle was fully extended (compare Fig. 4C, columns 1 and 5). This suggests that fluctuating contractions in the opposite pole do not improve mechanical conditions for entraining the secondary ending. When the second γ-axon was stimulated at a constant frequency of 40 Hz, and the γ-axon that could cause coupling was randomly stimulated, the primary ending was driven very effectively, even when the muscle was in flexion (compare Fig. 4C, columns 1, 3 and 6, row 1; note the marked improvement in the efficacy value). Interestingly, however, the entrainment of the secondary ending was not so markedly improved (compare Fig. 4C, columns 3 and 6, row 2) by co-stimulating the second γ-axon at a constant frequency of 40 Hz. This frequency was too high, because it dominated the output of the secondary ending – the discharge became highly regular (compare the auto-correlograms in Fig. 4C, row 4, column 6 with those in columns 1-5), and the number of spikes increased by a factor of two. This supports the idea that a γ-motoneurone can excite sensory endings much more effectively if fibres in the opposite pole are tensed by another γ-motoneurone firing, and suggests why a constant low frequency would be most effective.

An interesting feature of the response of the secondary ending to stimulating the γ-axon that could weakly drive it at random frequencies was a tendency to burst. This was investigated by plotting each interspike interval against the preceding interspike interval. Bursting is indicated in these joint interval histograms (Fig. 4D) by points aligned with the axes, giving the display an arrow-head appearance in contrast to the compact circular displays for resting discharge of the secondary ending. There was no bursting under resting conditions, or during random stimulation of the second γ-axon. The bursting might have been caused by neuromuscular block, which can occur at intrafusal neuromuscular junctions (see Bessou & Pagès, 1972). However, stretching muscles improves transmission at partially blocked neuromuscular junctions (Hutter & Trautwein, 1956). The joint interval histograms show that the bursting of the secondary ending was present in both flexion and extension (Fig. 4D, graphs 1 and 2). It was therefore unlikely to be due to neuromuscular block. It might have been caused by the giving way of resting bonds in the sarcomeres of the intrafusal fibres in the opposite, non-contracting pole. In support of this suggestion, there was no bursting when those parts of the fibres were made to contract by combining stimulation of the γ-axon that caused coupling at random frequencies with tonic stimulation of the second γ-axon (Fig. 4D, graph 4).

Coupled Ia and II output during random length changes

Coupling between the primary and secondary endings also occurred during muscle stretches in the decerebrate preparation, and under these conditions also spikes from the secondary ending left the muscle spindle ahead of those from the primary ending. The length signal was imposed after a period of stimulating nociceptors to excite the fusimotor system. This period is shown in Fig. 5. The stimulation mainly affected the primary ending. At the beginning of recording illustrated in Fig. 5A there was no coupling between the afferents (Fig. 5Ba). However, after 50 s there was a precise relation between the spikes from the two endings, indicating that the fusimotor activity had coupled the afferent output (Fig. 5B, compare b with d, the autocorrelogram of the secondary ending). The pattern is striking – there are exactly three peaks of primary spikes to each interspike interval of the secondary ending. After this the frequency of the secondary ending fell even though the output of the primary ending continued to rise, and there was now no coupling (Fig. 5Bc). Presumably the γ-activity had unloaded the secondary ending slightly.

Figure 5. Afferents coupled by natural fusimotor excitation in the decerebrate preparation.

Figure 5

Open in a new tab

A, instantaneous frequencies of spikes from primary (P) and secondary (S) endings of a spindle during increasing excitation of γ-motoneurones elicited by stimulating nociceptors. Periods marked a, b and c were used for construction of the cross-correlograms in B. Note the different scaling for P and S. B, a-c, cross-correlograms between S and P spikes (normalised, bin widths all 0.5 s). d, S auto-correlogram. Three groups of P spikes within the S interspike interval are seen clearly in b but there is no obvious relation in a or c.

After this period of increasing fusimotor activity at constant length, a random length signal was imposed; it may be assumed that some fusimotor excitation continued because the afferent responses to stretching were unchanged except for a slight increase in the maximum frequency of the primary ending (Fig. 6A). A cross-correlogram constructed by triggering off spikes from the secondary ending shows strong coupling, with the majority of spikes from the primary ending following those from the secondary ending (Fig. 6B). This can be confirmed by simple visual inspection of a representative section taken from the middle of the recording (Fig. 6C). Out of the six muscle stretches illustrated, the secondary ending responded before the primary ending to all but the fifth stretch. It is well documented that during muscle stretches both sensory endings respond in advance of the length stimulus (see Matthews & Stein, 1969; Cussons et al. 1977). However, in this instance the phase advance of the primary ending was reduced: compare the cross-correlograms between the peaks of the stretches and the two endings (Fig. 6D), with Fig. 6E, similar cross-correlograms for the test spindle without any fusimotor stimulation. The phase advance cannot be properly compared for each ending from records taken at the level of the dorsal roots because of the slower conduction velocity of group II axons. Histograms of the estimated times that the first spike from each ending would leave the spindle in relation to the peaks of 1000 stretches are shown in Fig. 6F. The ratio of the number of these first spikes leaving the primary ending before and after the peak of the stretch was 1.3:1; in contrast it was 3:1 for the secondary ending. The transmission of stretch to the primary ending appears to have been delayed. On the other hand, the sensitivity to stretching of the primary ending was higher; it failed to respond to only one of the 1000 stretches, whereas the secondary ending failed in 4.3 % of stretches. Also, the average number of spikes per stretch was about 3 for the primary ending, and 1.5 for the secondary ending. Without fusimotor influences, the sensitivity of primary endings to stretching is around 6–30 times greater than that of the secondary endings, but stimulation of single static γ-axons reduces their sensitivity so their responsiveness is similar to that of secondary endings (Cussons et al. 1977). The importance of the results illustrated in Fig. 6 is that the majority of spikes from the secondary ending would reach the spinal cord ahead of the spikes from the primary ending (see Discussion and Fig. 9B).

Effects of length change in the test spindle

In Fig. 4 it was shown that the coupling and tensing actions of naturally active static γ-motoneurones could be partially simulated by stimulating two γ-axons with appropriate intrafusal connections in a test spindle (γ1 and γ2 in Fig. 4). It was therefore interesting to analyse the effects of stimulating these two γ-axons during length change. This is illustrated in Fig. 7. Random length changes were applied with the hindlimb in full extension, so that the effectiveness of any γ-actions should have been optimal.

Length change alone coupled the afferent output (Fig. 7A, column 2), but cross-correlograms between the afferent spikes show that most spikes from the primary ending preceded the spikes from the secondary ending (Fig. 7A, column 2). This is in marked contrast to the situation illustrated in Fig. 6. Random stimulation of γ1 and of γ2 applied during the length change reduced this coupling, but γ2 had the more profound effect (Fig. 7A, columns 3 and 4). Cross-correlograms constructed between the peaks of the stretches and the afferents individually while the static γ-axons were stimulated (Fig. 7B, columns 2 and 3) show that the fusimotor stimulation reduced the response to length of the secondary ending. Jami et al. (1980) showed that strong participation of chain fibres enhances the length sensitivities of secondary endings but not primary endings. This suggests that in the case of the test spindle the contractions of the bag2 fibre may have dominated (both γ-axons innervated chain and bag2 fibres). The auto-correlograms for the secondary ending (Fig. 7A, lower row) show that its output was still rhythmic despite the length change (column 2), and that γ2 increased the frequency of the secondary ending without destroying its rhythmicity (column 4). On the other hand, γ1 abolished the rhythmicity (Fig. 7A column 3, lower row) and could still entrain the secondary ending despite the changing length (Fig. 7B, column 4, lower row). In contrast, γ1 could still very weakly entrain the primary ending (Fig. 7B, column 5, upper row). These results are surprising considering that both γ-axons innervated the same type of fibre (both γ1 and γ2 innervated bag2 and chain fibres). As with the results for the test spindle at constant length this demonstrates that a γ-axon innervating fibres solely in the opposite pole to the secondary ending can have better access to it than one innervating the same pole.

Natural fusimotor action unloading the secondary ending

A function of intrafusal contraction is to prevent the unloading of muscle spindles by extrafusal contraction, and in the present experiments examples of this were observed. However, contractions of bag2 fibres have been shown to unload chain fibres (see, for example, Boyd et al. 1977), and this has been put forward as an explanation for the slowing of secondary endings during stimulation of a small minority of static γ-axons (Gioux et al. 1990). It could also explain the slight slowing of the secondary ending as fusimotor activity increased in the spindle illustrated in Fig. 5 and Fig. 6, just before the random length signal was applied. In another example a secondary ending appeared to be particularly reliant on the continued activity of one of three spontaneously active γ-motoneurones. The secondary ending actually stopped temporarily when the discharge of one γ-motoneurone slowed, but the primary ending was not affected.

Figure 8 illustrates a case in which there was prolonged unloading of the secondary ending. Responses were elicited from the primary ending of the muscle spindle of a decerebrate preparation with intact ventral roots illustrated in Fig. 8, by stimulating nociceptors in the forelimb. However, the secondary ending was completely unresponsive. In addition there was a marked discrepancy in the background discharges of the two endings (mean frequencies 17.1 spikes s−1 for the primary ending, and only 4.2 spikes s−1 for the secondary ending). After the ventral roots had been cut the spontaneous discharge of the secondary ending was much higher (mean frequency 15.9 spikes s−1). This suggested that the secondary ending had been tonically unloaded by γ-activity before the ventral roots were cut. The innervation of this spindle is shown in Fig. 8F. To attempt to mimic the effects recorded while the ventral roots were intact, all three γ-axons innervating the pole containing the secondary ending were stimulated singly. Fig. 8B and C shows that when the static γ-axon that innervated the bag2 fibre and one chain fibre (γs1) was stimulated tonically at 100 Hz, or with a ramp frequency to 50 Hz the primary ending reached the maximum frequencies achieved during the natural fusimotor activation. However, the secondary ending then had a distinct response which it did not have under natural conditions. Stimulating the dynamic γ-axon (γd) with a ramp (Fig. 2E) came nearest to mimicking the natural effects on the primary ending without provoking a response from the secondary ending. However, the maximum frequencies recorded from the primary ending were too low; they could not even be reached by stimulating the γd-axon supramaximally at 100 Hz (Fig. 8D). The remaining axon (γs2), which innervated chain fibres, excited both primary and secondary endings (not illustrated). Histological reconstruction of the secondary ending showed that it was quite extensive, and had several contacts with both bag fibres. Contractions set up by activity in the two axons found by histology to innervate the other pole (Fig. 8F, γsh and γdh), could not have unloaded the secondary ending – they would have extended it. Possibly low frequency (< 10 Hz) tonic activity of the static γ-motoneurone that innervated the bag2 fibre (γs2) unloaded this secondary ending. Neither ending would have been excited appreciably by such a low frequency, but, as has been argued above, such activity may be sufficient to change the mechanical status of the spindle.

Figure 8. Activation of the primary, but not the secondary ending of a muscle spindle in the decerebrate preparation.

Figure 8

Open in a new tab

A, only the primary (P) ending could be excited by pinching a forepaw. Note that after cutting the ventral roots the resting discharge of the secondary ending (S) was much higher (compare with frequency of S in B, C, D and E before and after the γ-axons were stimulated). B-E, failure to mimic natural effects by stimulating single γ-axons. Stimulation of a single γs-axon (γs1) at 100 Hz (B) or with a ramp frequency (C) elicited an increase in the frequency of P (B and C, upper records) similar to that in A, but co-excited S (B and C, lower records). Maximal stimulation of a dynamic γ-axon (γd) did not match the natural excitation of P (D and E, upper records), although its action on S was minimal (D and E, lower records). F, schematic representation of the innervation of this muscle spindle. End-plates of the three axons stimulated are shown filled; those of two additional axons (γdh and γsh) found during the reconstruction are shown empty. Labels as in Fig. 3C.

Estimation of strength of summation of EPSPs in spinal pathways from secondary endings

Both primary and secondary endings have projections to α-motoneurones and to interneurones in the intermediate zone (see Fig. 9A). EPSPs evoked in these neurones by spikes from primary and secondary endings (P EPSPs and S EPSPs respectively) will summate most effectively when P EPSPs occur during an S EPSP, and when the P EPSPs occur just before the S EPSP. Time windows during which EPSPs will summate effectively were estimated taking into account the half-widths (when the amplitude is more than half the maximal amplitude). If, as shown diagrammatically in Fig. 9B, the time window corresponds to the sum of the half-widths of the P and S EPSPs (3.9 ms and 4.4 ms respectively; Sypert et al. 1980), the summation of EPSPs evoked by spikes from P and S endings would be appreciable when they reach their target cells within about 8 ms of each other.

In order to estimate the percentages of P and S EPSPs summating within such time windows when evoked in α-motoneurones, the following procedure was used. Cross-correlograms were constructed between 100 successive spikes from secondary endings and the associated spikes from primary endings at the level of the muscle spindle in the tenuissimus muscle. The differences in conduction times for the P and S spikes were estimated taking into account the known peripheral delays, and central delays for tenuissimus afferents (Fu & Schomberg, 1974). A spike leaving the muscle spindle from the secondary ending at the time indicated by ‘sp’ in Fig. 9E, would thus reach an α-motoneurone simultaneously with spikes from the primary ending occurring in the cross-correlograms around 2.6 ms later, as indicated by the blue arrow. P spikes falling within an 8 ms period approximately centred on this point (to be more exact, 3.9 ms before and 4.4 ms after, see above) would summate effectively with the S EPSPs. Knowing the number of coupled S and P spikes at the level of the spindles, their percentages under different experimental conditions could thus be compared. In the case of the spindle illustrated in Fig. 2, it was calculated that when only one γ-motoneurone (γ2) was active, only 5 % of monosynaptically-evoked S EPSPs would summate with a P EPSP on α-motoneurones. However, when a second γ-motoneurone was recruited (γ1), 60 % would summate. A 12-fold increase would thus be a direct consequence of the coupling action of the second motoneurone. During muscle stretching in the decerebrate cat the percentage was 146 % (the spindle illustrated in Fig. 6) – that is, all S EPSPs would summate with one P EPSP, and 46 % would summate with two. When γ-axons were stimulated artificially with random frequencies of pulses (see Fig. 4) the summation calculated would have increased by a factor of only three from the resting conditions.

Di- and trisynaptically evoked EPSPs in α-motoneurones following the same spike potentials from P and S endings, and secondary to the activation of intermediate zone interneurones (Ii), or both dorsal horn interneurones (Di) and Ii, will be induced with additional central delays of just over 1 and 2 ms. Spikes in interneurones activated by these endings would thus reach α-motoneurones at the times indicated by the red and turquoise arrows respectively in Fig. 9E. The half-width for EPSPs evoked by the interneurones (Cavallari et al. 1987) is not known but it may be assumed that it is about the same as for EPSPs evoked by secondary endings. Adding the half-width of such EPSPs to the effective summation period for monosynaptically evoked EPSPs on α-motoneurones will extend the period during which coupled actions of P and S endings will be most effectively facilitated to about 11 ms. Estimates of the percentages of effective summation of S and P EPSPs are given in Table 1 for EPSPs evoked in α-motoneurones through the mono-, di- and trisynaptic pathways separately, and for all three pathways converging on the same α-motoneurone. The effects of P and S coupling at the level of the intermediate zone interneurones cannot yet be estimated but should add to these effects. Facilitation at the level of Ii will also increase the excitation of γ-motoneurones, including homonymous γ-motoneurones (Gladden et al. 1998). This should provide compensation for the unloading of muscle spindles when the homonymous muscle contracts.

Table 1.

Estimates of effective summation of EPSPs evoked in α-motoneurones by spikes from primary and secondary sensory endings

Monosynaptic group II pathway only Disynaptic group II pathway only Trisynaptic group II pathwayonly All Mean group II freq. (Hz)
One γ-motoneurone active, no coupling (γ2 in Fig. 2) 5 6 7 8 8
Two γ-motoneurones active, coupling present (γ1 + γ2 in Fig. 2) 60 61 64 79 14
γ-Activity + muscle stretching (Fig. 6) 146 160 182 217 42
Resting conditions (spindle of Fig. 4) 9 10 11 11 23
Stimulation of γI (Fig. 4) with random frequencies 27 26 17 34 26
Open in a new tab

Numbers are the mean numbers of P spikes associated with 100 consecutive S spikes.

Discussion

It has been demonstrated that γ-motoneurones activated by natural stimuli can couple the output of secondary and primary endings, with the spikes from the secondary ending consistently leaving the muscle spindle first. Stable coupling of this type was demonstrated both at constant length and during muscle stretching. At constant length the surprising finding was that bag2 fibres could excite secondary endings more easily than the primary ending when contracting on the opposite side of the primary ending. Three points can be offered in explanation. Firstly, when stretched passively, secondary endings stretch more than primary endings (Poppele & Quick, 1985). In effect, the primary ending is a stiffer spring. Secondly, under the primary ending the bag2 and chain fibres are bound together by the ending terminals and internal capsule. Forces set up by a bag2 fibre pulling on the primary ending may thus be transmitted more evenly to a secondary ending on the other side of it than by fibres contracting on the same side. Also, in the region of intrafusal fibres that secondary endings contact there are few central nuclei, mainly myofibrils. It is therefore possible that contraction of fibres in the same pole as the secondary ending could cause some shortening of the ending. The third point is that at the low frequencies at which the γ-motoneurones were active, bag2 fibres may be more effective than chain fibres in pulling on the primary ending because their contraction summates tetanically at frequencies less than 10 Hz (see Bessou & Pagès, 1975). Thus, as the tension exerted by the bag2 fibres rises following each γ-spike, the secondary ending on the other side of the primary ending may be extended earlier than the stiffer primary ending. As explained in the Results, the effectiveness with which the bag2 fibre can extend the secondary ending should be increased if the fibres on the opposite side of the primary ending are prevented from lengthening passively. Passive extension will be reduced when the muscle spindle is extended, and also if another γ-motoneurone induces small, steady tonic contractions. There was evidence to support this idea (see Fig. 4). Functionally this would imply that coupling becomes more effective as muscle length is increased, and with background excitation of the static γ-motoneurone pool. During muscle stretching, γ-activity delayed the application of stretch to the primary sensory ending, though the ending remained more sensitive to stretch than the secondary ending (see Fig. 6).

Functional implications

Coupling between the primary and secondary endings and changes in the timing of the arrival of spikes in group Ia and group II afferents to their spinal target neurones appeared to be of considerable functional consequence. It was calculated that the percentage of EPSPs evoked by these afferents that summate is increased several fold by this coupling. The directly evoked excitation of α-motoneurones from secondary endings may be rather weak (Lundberg et al. 1977). Secondary endings with low conduction velocities (< 52 m s−1, as had secondary endings in the present experiments) have been found to have functional connections with only 27 % of homonymous α-motoneurones (Munson et al. 1980), and EPSPs evoked by them are 4–6 times smaller than EPSPs evoked by the primary endings (Sypert et al. 1980). However, a great proportion of interneurones in di-synaptic excitatory pathways, and some inhibitory pathways from group II and group I afferents are co-excited by muscle spindle primaries and secondaries (Fig. 9A; see Harrison & Jankowska, 1985; Cavallari et al. 1987; Edgley & Jankowska, 1987; Riddell & Hadian, 2000). Potent effects of these interneurones upon α-motoneurones would thus be greatly facilitated by γ-mediated coupling between P and S endings. The leading output from the secondary endings would also be important for the proper timing between the indirect input from secondary endings to intermediate zone interneurones via dorsal horn interneurones (see Fig. 9A; Jankowska et al. 2002a,b).

For the combined actions of the mono-, di- and trisynaptic routes to α-motoneurones the facilitation was estimated to be particularly strong during muscle stretch in the decerebrate cat preparation (Fig. 9E). Around 1.6 P EPSPs were estimated to summate with every S EPSP or disynaptically evoked EPSP in an α-motoneurone, providing a potent mechanism for integrating the output of primary and secondary endings in the segmental stretch reflex. Interneuronal pathways from secondary endings are implicated in the human stretch reflex both under normal conditions (Corna et al. 1995; Marque et al. 1996) and in pathological states associated with spasticity (for references see Jankowska & Hammar, 2002).

Groups of static γ-motoneurones

It has been argued that there are different groups of static γ-motoneurones controlled separately from the midbrain (Gladden & McWilliam, 1977; Wand & Schwarz, 1985; Dickson & Gladden, 1992) and cortex (Asgari-Khozankalaei & Gladden, 1990). Boyd (1986) proposed that there are two groups, based on the type of intrafusal fibre that static γ-axons innervate in different muscle spindles. However, this was contested (see Dickson et al. 1993; Celichowski et al. 1994; Emonet-Dénand et al. 1998; Taylor et al. 1998; Petit et al. 1999). Recently Taylor et al. (2000) described two patterns of activation of static γ-motoneurones during locomotion. The present results support the idea that static γ-motoneurones have different functional roles, but suggest that it is important to take account of their effects on secondary endings.

Acknowledgments

The study was supported by a grant from the Wellcome Trust to M.H.G. and by the Boyd Memorial Fund. Our warmest thanks are due to Miss J. Wilson for expert technical assistance and to Miss B. Sanford for cutting histological sections. The authors would like to thank Professor Elzbieta Jankowska and Dr Léna Jami for helpful discussions.

References

  1. Asgari-Khozankalaei A, Gladden MH. Simultaneous excitation and inhibition of tenussimus static γ motoneurones during cortical stimulation in the anaesthetised cat. Journal of Physiology. 1990;429:9P. [Google Scholar]
  2. Banks RW, Barker D, Stacey MJ. Form and distribution of sensory terminals in cat hindlimb muscle spindles. Philosophical Transactions of the Royal Society B. 1982;299:329–364. doi: 10.1098/rstb.1982.0136. [DOI] [PubMed] [Google Scholar]
  3. Bessou B, Laporte Y. Responses from primary and secondary endings of the same neuromuscular spindle of the tenuissimus of the cat. In: Barker D, editor. Symposium on Muscle Receptors. Hong Kong: Hong-Kong University Press; 1962. pp. 105–119. [Google Scholar]
  4. Bessou B, Pagès B. Intracellular potentials from intrafusal muscle fibres evoked by stimulation of static and dynamic fusimotor axons in the cat. Journal of Physiology. 1972;227:709–727. doi: 10.1113/jphysiol.1972.sp010055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bessou B, Pagès B. Cinematographic analysis of contractile events produced in intrafusal muscle fibres by stimulation of static and dynamic fusimotor axons. Journal of Physiology. 1975;252:397–428. doi: 10.1113/jphysiol.1975.sp011150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Boyd IA. The action of the three types of intrafusal fibre in isolated cat muscle spindles on the dynamic and length sensitivities of primary and secondary sensory endings. In: Taylor A, Prochazka A, editors. Muscle Receptors and Movement. London: Macmillan; 1981. pp. 17–32. [Google Scholar]
  7. Boyd IA. Two types of static γ axons in cat muscle spindles. Experimental Physiology. 1986;71:307–327. doi: 10.1113/expphysiol.1986.sp002987. [DOI] [PubMed] [Google Scholar]
  8. Boyd IA, Gladden MH, McWilliam PN, Ward J. Control of dynamic and static nuclear bag fibres and nuclear chain fibres by gamma and beta axons in isolated cat muscle spindles. Journal of Physiology. 1977;265:133–162. doi: 10.1113/jphysiol.1977.sp011709. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cavallari P, Edgley SA, Jankowska E. Post-synaptic actions of midlumbar interneurones on motoneurones of hind-limb muscles in the cat. Journal of Physiology. 1987;389:675–689. doi: 10.1113/jphysiol.1987.sp016677. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Celichowski J, Emonet-dénand F, Laporte Y, Petit J. Distribution of static γ axons in cat peroneus tertius spindles determined by exclusively physiological criteria. Journal of Neurophysiology. 1994;71:722–732. doi: 10.1152/jn.1994.71.2.722. [DOI] [PubMed] [Google Scholar]
  11. Corna S, Grasso M, Nardone A, Schieppati M. Selective depression of medium latency leg and foot responses to stretch by an α2-agonist in humans. Journal of Physiology. 1995;484:803–809. doi: 10.1113/jphysiol.1995.sp020705. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cussons PD, Hulliger M, Matthews PBC. Effects of fusimotor stimulation on the responses of the secondary ending of the muscle spindle to sinusoidal stretching. Journal of Physiology. 1977;270:835–850. doi: 10.1113/jphysiol.1977.sp011984. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dickson M, Emonet-DÉNAND F, Gladden MH, Petit J, Ward J. Incidence of non-driving excitation of Ia afferents using ramp frequency stimulation of static γ-axons in cat hindlimbs. Journal of Physiology. 1993;460:657–673. doi: 10.1113/jphysiol.1993.sp019492. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Dickson M, Gladden MH. Central and reflex recruitment of γ motoneurones of individual muscle spindles of the tenuissimus muscle in anaesthetised cats. In: Jami L, Pierrot-Deseilligny E, Zytnicki D, editors. Muscle afferents and Spinal Control of Movement. Oxford: Pergamon; 1992. pp. 37–42. [Google Scholar]
  15. Duggan AW, Hall JG, Foong FW, Zhao ZQ. A differential effect of naloxone on transmission of impulses in primary afferents to ventral roots and ascending spinal tracts. Brain Research. 1985;344:316–321. doi: 10.1016/0006-8993(85)90809-1. [DOI] [PubMed] [Google Scholar]
  16. Edgley SA, Jankowska E. An interneuronal relay for group I and II muscle afferents in the midlumbar segments of the cat spinal cord. Journal of Physiology. 1987;389:647–674. doi: 10.1113/jphysiol.1987.sp016676. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Emonet-DÉnand F, Laporte Y, Petit J. Comparison of static fusimotor innervation in cat peroneus tertius and longus muscles. Journal of Neurophysiology. 1998;80:249–254. doi: 10.1152/jn.1998.80.1.249. [DOI] [PubMed] [Google Scholar]
  18. Fu T -C, Schomburg ED. Electrophysiological investigation of the projection of secondary muscle spindle afferents in the cat spinal cord. Acta Physiologica Scandinavica. 1974;91:314–329. doi: 10.1111/j.1748-1716.1974.tb05687.x. [DOI] [PubMed] [Google Scholar]
  19. Gioux M, Petit J, Proske U. Slowing of the discharge of secondary endings of cat muscle spindles during fusimotor stimulation. Experimental Brain Research. 1990;83:164–171. doi: 10.1007/BF00232205. [DOI] [PubMed] [Google Scholar]
  20. Gladden MH. Structural features relative to the function of intrafusal muscle fibres in the cat. Progress in Brain Research. 1976;44:51–59. doi: 10.1016/S0079-6123(08)60722-0. [DOI] [PubMed] [Google Scholar]
  21. Gladden MH, Jankowska E, Czarkowska-Bauch J. New observations on coupling between group II muscle afferents and feline gamma motoneurones. Journal of Physiology. 1998;512:507–520. doi: 10.1111/j.1469-7793.1998.507be.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Gladden MH, McWilliam PN. The activity of intrafusal muscle fibres during cortical stimulation in the cat. Journal of Physiology. 1977;273:28–29P. [PubMed] [Google Scholar]
  23. Gladden MH, Matsuzaki H. Co-operative actions of static γ-motoneurones couple group Ia and II afferents. Journal of Physiology. 2001;536.P:51–52P. doi: 10.1113/jphysiol.2001.013229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Gladden MH, Matsuzaki H. Fusimotor-generated coupling between group Ia and II afferents during muscle stretch. European Journal of Physiology. 2002;443:31–32P. [Google Scholar]
  25. Harrison PJ, Jankowska E. Sources of input to interneurones mediating group I non-reciprocal inhibition of motoneurones in the cat. Journal of Physiology. 1985;361:379–401. doi: 10.1113/jphysiol.1985.sp015651. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Hutter OF, Trautwein W. Neuromuscular facilitation by stretch of motor nerve endings. Journal of Physiology. 1956;133:610–625. doi: 10.1113/jphysiol.1956.sp005613. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jami L, Lan-Couton D, Petit J. Glycogen-depletion method of intrafusal distribution of γ-axons that increase sensitivity of spindle secondary endings. Journal of Neurophysiology. 1980;43:16–26. doi: 10.1152/jn.1980.43.1.16. [DOI] [PubMed] [Google Scholar]
  28. Jankowska E, Hammar I. Spinal interneurones; how can studies in animals contribute to the understanding of spinal interneuronal systems in man? Brain Research Reviews. 2002 doi: 10.1016/s0165-0173(02)00185-6. in the Press. [DOI] [PubMed] [Google Scholar]
  29. Jankowska E, Slawinska U, Hammar I. Differential presynaptic inhibition of actions of group II afferents in di- and polysynaptic pathways to feline motoneurones. Journal of Physiology. 2002a;542:287–299. doi: 10.1113/jphysiol.2001.014068. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Jankowska E, Slawinska U, Hammar I. On organization of a neuronal network in pathways from group II muscle afferents in feline lumbar spinal segments. Journal of Physiology. 2002b;542:301–314. doi: 10.1113/jphysiol.2001.014076. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Lev-Tov A, Pratt CE, Burke RE. The motor unit population of the cat tenuissimus muscle. Journal of Neurophysiology. 1988;59:1128–1142. doi: 10.1152/jn.1988.59.4.1128. [DOI] [PubMed] [Google Scholar]
  32. Lundberg A, Malmgren K, Schomburg ED. Comments on reflex actions evoked by electrical stimulation of group II muscle afferents. Brain Research. 1977;122:551–555. doi: 10.1016/0006-8993(77)90466-8. [DOI] [PubMed] [Google Scholar]
  33. Matthews PBC, Stein RB. The sensitivity of muscle spindle afferents to small sinusoidal changes of length. Journal of Physiology. 1969;200:723–743. doi: 10.1113/jphysiol.1969.sp008719. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Marque P, Pierrot Deseilligny E, Simonetta Moreau M. Evidence for excitation of the human lower limb motoneurones by group II muscle afferents. Experimental Brain Research. 1996;109:357–360. doi: 10.1007/BF00231793. [DOI] [PubMed] [Google Scholar]
  35. Munson JB, Fleshman JW, Sypert GW. Properties of single-fibre spindle group II EPSPs in triceps surae motoneurons. Journal of Neurophysiology. 1980;44:713–725. doi: 10.1152/jn.1980.44.4.713. [DOI] [PubMed] [Google Scholar]
  36. Petit J, Banks RW, Laporte Y. Testing the classification of static γ axons using different patterns of random stimulation. Journal of Neurophysiology. 1999;81:2823–2832. doi: 10.1152/jn.1999.81.6.2823. [DOI] [PubMed] [Google Scholar]
  37. Poppele RE, Quick DC. Effect of intrafusal muscle mechanics on mammalian muscle spindle sensitivity. Journal of Neuroscience. 1985;5:1881–1885. doi: 10.1523/JNEUROSCI.05-07-01881.1985. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Riddell JS, Hadian M. Interneurones in pathways from group II muscle afferents in the lower-lumbar segments of the feline spinal cord. Journal of Physiology. 2000;522:109–123. doi: 10.1111/j.1469-7793.2000.t01-2-00109.xm. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rodiek RW, Kiang NY-S, Gerstein GL. Some quantitative methods for the study of spontaneous activity of single neurons. Biophysical Journal. 1962;2:351–368. doi: 10.1016/s0006-3495(62)86860-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Ruffini A. On the minute anatomy of the neuromuscular spindles of the cat, and their physiological significance. Journal of Physiology. 1898;23:190–208. doi: 10.1113/jphysiol.1898.sp000723. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Swadlow HA, Gusev AG. The impact of ‘bursting’ thalamic impulses at a neocortical synapse. Nature Neuroscience. 2001;4:402–408. doi: 10.1038/86054. [DOI] [PubMed] [Google Scholar]
  42. Sypert GW, Fleshman JW, Munsen JB. Comparison of monosynaptic actions of medial gastrocnemius group Ia and group II muscle spindle afferents on triceps surae motoneurons. Journal of Neurophysiology. 1980;44:726–738. doi: 10.1152/jn.1980.44.4.726. [DOI] [PubMed] [Google Scholar]
  43. Taylor A, Ellaway PH, Durbaba R. Physiological signs of the activation of bag2 and chain intrafusal muscle fibres of gastrocnemius muscle spindles in the cat. Journal of Neurophysiology. 1998;80:130–142. doi: 10.1152/jn.1998.80.1.130. [DOI] [PubMed] [Google Scholar]
  44. Taylor A, Ellaway PH, Durbaba R, Rawlinson S. Distinctive patterns of static and dynamic gamma motor activity during locomotion in the decerebrate cat. Journal of Physiology. 2000;529:825–836. doi: 10.1111/j.1469-7793.2000.00825.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Wand P, Schwarz M. Two types of static fusimotor neurones under separate central control? Neuroscience Letters. 1985;58:145–149. doi: 10.1016/0304-3940(85)90344-1. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of Physiology are provided here courtesy of The Physiological Society

RESOURCES