| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
MS 8817 Received 7 October 1998; accepted after revision 7 June 1999.
 |
ABSTRACT |
|---|
 TopAbstract IntroductionMethodsResultsDiscussionReferences |
|---|
0·8) before declining as a function of frequency, with R values of ~0·6 at 300 Hz and <= 0·4 at 800 Hz. Both the qualitative and quantitative aspects of ECN responsiveness to the vibro-stretch disturbances were indistinguishable from those of the main cuneate neurones.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Inputs arising from a particular class of receptor may be directed to multiple targets within the central nervous system and utilized for a variety of processing tasks at the different sites. In the case of muscle receptors, the inputs are conveyed over segmental pathways at the spinal cord level for the reflex regulation of posture and movement. However, these muscle inputs are also directed over ascending pathways for processing at hierarchically higher levels of the nervous system, including the cerebral cortex, where they contribute to kinaesthetic sensation (McCloskey, 1978), and the cerebellum where they are presumably utilized for the regulation and control of voluntary movements (e.g. Cooke et al. 1971). Muscle inputs from the forelimb project to the cuneate and external cuneate nuclei which form parallel synaptic relays in these ascending pathways to the cerebral cortex and cerebellum, respectively (Cooke et al. 1971).
The input to these parallel relay nuclei from muscle spindle afferents is known to contain precise information about both static and dynamic aspects of muscle length changes. Furthermore, the responses, in particular of primary spindle afferent fibres, to controlled forms of muscle vibration reveal their capacity for signalling, with great precision, information about high-frequency, low-amplitude perturbations in muscle length (e.g. Bianconi & Van der Meulen, 1963; Brown et al. 1967; Mackie et al. 1998). However, it is uncertain whether equivalent information is extracted from the muscle afferent inputs by neurones within the main and external cuneate nuclei, arranged as they are, in parallel, for conveying information rostrally, predominantly for the purpose of kinaesthesia in the case of the main cuneate nucleus and for motor control in the case of the external cuneate nucleus. Some of the early studies on neurones of both nuclei left some doubt over their capacities to retain the precision of impulse patterning evident in the responses of their spindle afferent inputs. Neurones of both the cuneate and external cuneate nuclei of the macaque monkey displayed responses to brief trains (2-8 cycles) of muscle vibration that were phase locked at frequencies up to 50-100 Hz, although contributions to these responses from other mechanoreceptors, including Pacinian corpuscles and joint receptors, could not be excluded (Hummelsheim & Wiesendanger, 1985). Both cuneate and external cuneate neurones in the cat were shown by Rosén & Sjölund (1973a) to respond with high discharge rates to muscle vibration but no assessment was made of how precisely the pattern of discharge could reflect the temporal detail in the vibratory stretch disturbance. However, our recent quantitative analysis of cat main cuneate responses to muscle vibration revealed that these neurones signal, with great precision, the temporal features of vibro-stretch perturbations up to frequencies of 400-500 Hz and display an overall bandwidth of vibration sensitivity that extends to 
800 Hz (Mackie et al. 1998). These attributes enable the cuneate neurones to contribute accurate temporal information for kinaesthetic sensation in circumstances involving dynamic length changes in skeletal muscle.
In the present study, in part reported in preliminary abstract form (Mackie et al. 1994), we have examined quantitatively the capacity of external cuneate neurones in the cat to respond to the same forms of static and vibrational stretch of forearm extensor muscles. The aim was to determine whether the information signalled by these neurones to the cerebellum for the purposes of motor control provides evidence for differential processing of muscle spindle inputs at the parallel synaptic relays of the main cuneate and external cuneate nuclei.
|
METHODS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Data on the responses of neurones of the external cuneate nucleus (ECN) to spindle inputs from forearm extensor muscles were obtained from experiments in seven adult cats anaesthetized with pentobarbitone sodium (Nembutal, 40 mg kg-1, I.P.) and maintained on an I.V. infusion of the same agent (1·5-2·0 mg kg-1 h-1). Atropine sulphate was administered subcutaneously to minimize respiratory secretions (0·05-0·1 mg kg-1). The femoral artery and vein were cannulated, and a tracheal cannula inserted. Blood pressure and heart rate were monitored continuously, and pupillary aperture and responsiveness, together with withdrawal reflexes, were frequently examined to ensure deep levels of anaesthesia were maintained throughout the course of each experiment. Rectal temperature was maintained at 37-38·5°C. All experiments conformed with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. Animals were killed at the end of experiments by I.V. administered anaesthetic overdose.
Forearm preparation
Preparation of the forelimb and the application of mechanical stimuli to individual forearm extensor muscles was carried out as described previously (Mackie et al. 1998). In particular, the left forelimb was substantially denervated, apart from the deep radial nerve, by severing the median, ulnar, musculocutaneous and superficial radial nerves. A cautery device was also used to sever any branches of the dorsal interosseus wrist joint nerve which may have supplied non-muscle receptors in the interosseus membrane or wrist joint. The tendons of the following forearm extensor muscles were cut at the wrist and the muscles separated from each other over almost their entire length: extensor digitorum communis (EDC), extensor digitorum lateralis (EDL), extensor carpi ulnaris (ECU), indicis proprius (IPr), extensor pollicis longus (EPL), abductor pollicis longus (APL), extensor carpi radialis longus (ECRL), extensor carpi radialis brevis (ECRB), and usually, brachioradialis (Bra). A pool, filled with warm isotonic saline, was formed by the skin flaps of the forearm which were attached to a brass ring.
Stimulation of forearm extensor muscles
A feedback-controlled mechanical stimulator was attached to individual muscle tendons in order to apply controlled ramp-and-hold stretches of up to 2 mm in amplitude, and longitudinal vibration of up to 200 µm peak-to-peak amplitude at frequencies of 50-150 Hz, 100 µm at 500 Hz, and 75 µm at 800 Hz. A low level of background resting tension was applied to each muscle under observation, usually 0·1 N (10 g wt), which was sufficient to take up any slack in the muscle. This involved a 1-2 mm extension of the muscle from the length that the muscle assumed in the extended position of the wrist and digits prior to cutting the muscle tendon. One-second trains of longitudinal vibration were normally superimposed on a controlled steady stretch of the muscle (up to 2 mm and lasting 5 s) at stimulus intervals of 12 s or greater. In any analysis of responses to muscle vibration the initial movement in the vibratory displacement always occurred from the null position in a fixed direction. This was a crucial requirement for the analysis of phase locking in the responses to the vibro-stretch disturbances.
Recording and analysis procedures for the study of external cuneate neurones
Recordings from ECN neurones were also carried out as described previously for the study of main cuneate neurones (Mackie et al. 1998). Lacquer-coated tungsten microelectrodes with measured tip impedances of 0·5-2 M
were inserted into the nucleus at depths of up to 2500 µm below the brainstem surface and at the mediolateral and rostrocaudal locations indicated in the Results. In two experiments, a pair of low-impedance tungsten electrodes with a 4-5 mm spacing was inserted into the ipsilateral restiform body, based on the stereotaxic coordinates of Snider & Niemer (1961) and initial recording from the electrodes, to maximize responses to forearm extensor muscle inputs. These electrodes were fixed in place by means of a dental acrylic attachment to the occipital region of the skull and were used as stimulating electrodes for antidromic activation of ECN neurones isolated for study. However, the antidromic stimulation electrodes were not routinely used because of the proximity of their insertion path to the ECN itself and the risk of damage to the nucleus.
Signals from individual neurones were displayed on oscilloscopes and stored on magnetic tape and in computer records. Procedures for data analysis were described in detail by Mackie et al. (1998). Stimulus-response relations were constructed for individual ECN neurones to show the mean response (impulses s-1; ± S.D.), as a function of static stretch amplitude for stretch applied via the tendon to the individual forearm extensor muscles. The frequency bandwidth of responsiveness to muscle vibration was also plotted for individual neurones in terms of response level (impulses s-1) as a function of vibration frequency at fixed vibration amplitudes. The extent of entrainment or phase locking of impulse activity to the muscle vibration waveform was examined by constructing phase-scatter and cycle histogram distributions by means of a laboratory computer. The phase-scatter graphs (Fig. 3A-C) display the time of occurrence of individual impulses as a function of both the phase of the vibration cycle (ordinate) and the time within the overall 1 s train of muscle vibration (abscissa). The corresponding cycle histograms display the probability of an impulse occurring at different times throughout the period of the vibration cycle (Mackie et al. 1998).
A quantitative measure of phase locking for responses to the vibration was derived from these cyclic distributions. The resultant, R, which is an index of vector strength in a cyclic distribution (Mardia, 1972), is a measure of the phase coherence or synchronization in the impulse distribution and is inversely related to the dispersion in the impulses around the mean phase of the response (Mackie et al. 1998). It is calculated from the cyclic distributions according to the formula:
[(
where n is the total number of impulse occurrences, and xi(1
n) is the phase angle (in radians) of each spike occurrence time relative to the start of the vibration cycle (Zar, 1984).
The R value for each distribution ranges between 0 and 1, where 0 indicates an absence of phase locking and is associated with a flat distribution in the cycle histogram, and where 1 indicates perfect phase locking of the responses. This measure has been used in evaluating the entrainment of auditory responses to tonal stimuli (e.g. Lavine, 1971; Bledsoe et al. 1982) and in the somatosensory system for quantifying the entrainment of responses to vibrotactile stimuli (e.g. Greenstein et al. 1987; Vickery et al. 1994; Mackie et al. 1998). The sample size (n), which is the total number of impulse occurrences making up the cycle histogram, determines the significance level of R. Tabulated significance levels for various sized samples are available, based on the non-parametric Rayleigh z statistic (Zar, 1984; Table B32). Application of the Rayleigh test to a distribution containing 200 impulse occurrences demonstrates, at the 5 % significance level, that R values of 
0·12 are indicative of a uniform (i.e. random) distribution. As all calculations of R for the present data were based on sample sizes in excess of 200, it can be accepted that R values exceeding 0·12 were indicative of significant phase locking (Zar, 1984; Table B32). Details of the computation algorithm for R have been described previously (Mackie et al. 1998).
|
RESULTS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Location and identity of muscle-related ECN neurones
Neurones of the ECN that were studied in detail demonstrated a high sensitivity in their responses to both the static and dynamic components of muscle stretch stimuli, consistent with their dominant input being derived from muscle spindle afferent fibres. Forearm denervation procedures ensured that contamination from non-muscular sources, such as Pacinian corpuscle inputs originating in the interosseous membrance or wrist joint, was eliminated. One neurone appeared to derive its input from Pacinian or paciniform sources within the EDL muscle, based on the observation that this neurone had no background activity or sustained responses during the static components of muscle stretch stimuli, but was responsive to longitudinal vibration of the muscle, with peak sensitivity at a vibration frequency of 300 Hz. This unit was unresponsive to stimuli applied to other parts of the body, including other muscles, or to structures such as bone or the brass ring supporting the skin-flap pool, confirming that the location of the ending was almost certainly within the EDL muscle itself. This was an unusual finding, but was by no means anomalous as Pacinian and paciniform endings have been reported within the muscle fascia (Barker, 1967). As this study was concerned with the response characteristics of neurones which derived their input from muscle spindle endings, the data from this neurone are not considered further.
The muscle-related ECN neurones were isolated between 2·75-3·5 mm rostral, and 3·5-4·5 mm lateral to the obex which conforms to the rostrocaudal and mediolateral location of the ECN in the stereotaxic atlas of Berman (1968). The midline and obex provided visible reference locations on the exposed dorsal surface of the brainstem which enabled these recording locations to be identified accurately and permitted confident identification of the ECN as the recording locus. The depths below the brainstem surface at which neurones were isolated were not determined precisely in all cases, as many penetrations were made through the posterior region of the cerebellum. However, where neurone depths were reliably measured, they were between 1 and 2 mm below the surface. These distributions, though slightly more rostral and lateral, are consistent with those reported by Rosén & Sjölund (1973a) for ECN neurones which were located in penetrations between 1·3 and 3·2 mm rostral, and 3·0 to 4·2 mm lateral to the obex. Two out of five ECN neurones tested were antidromically activated by means of stimulating electrodes inserted into the ipsilateral restiform body (see Methods), confirming their identity as cuneo-cerebellar neurones. The 4-5 mm tip spacing of the electrodes may not have permitted sufficient current density to activate the remaining three units. However, it is known that almost all neurones within the cat ECN project to the cerebellum (Cooke et al. 1971).
Quantitative stimulus-response characteristics for twelve ECN neurones that were activated by spindle afferent sources were examined in detail. Four neurones were activated by stretch of the EDL, three by EDC, two by APL, and one by each of ECRB, EPL and IPr. None showed evidence of activation by more than one muscle, a finding consistent with the observation reported by Rosén & Sjölund (1973b) that almost 90 % of their sample was activated by stretch of only one forelimb muscle. All muscle-related ECN neurones examined in this study are believed to derive their dominant input from large afferent nerve fibres which form primary annulospiral endings on either or both nuclear bag or nuclear chain intrafusal fibres. The fact that all responded with high sensitivity, and a maintained discharge to both muscle stretch and longitudinal muscle vibration, precludes their drive originating from Golgi tendon organ receptors (Lundberg & Winsbury, 1960). The dynamic response properties of all ECN units described below indicated that the input was unlikely to have come from secondary spindle endings (Cooper, 1961; Brown et al. 1967).
We believe that the ECN responses recorded in this study should not have been substantially affected by the anaesthetic agent, (i)because of the evidence of high transmission fidelity revealed by the responses, and (ii) because of their similarity to the muscle-related responses recorded (Mackie et al. 1998) in the main cuneate nucleus (MCN) where our earlier observations have demonstrated that remarkably secure transmission occurs across the cuneate nucleus for inputs from single Pacinian-related sensory fibres in both anaesthetized and decerebrate cats. Although some studies have indicated that cerebellar Purkinje cell responses may be depressed by anaesthetic agents (Gordon et al. 1972; Perciavalle et al. 1995) this probably reflects a greater susceptibility to anaesthetic disruption of the cerebellar cortical synapses than the earlier relays in the pathway, through structures such as the ECN.
Responses of ECN neurones to static muscle stretch
All ECN neurones activated by forearm extensor muscles (with the exception of the Pacinian corpuscle-related neurone referred to above) demonstrated the capacity to respond in a maintained manner to static muscle stretch, as shown for one neurone in Fig. 1A. The response features for this neurone, reflected in the impulse record, peristimulus time histogram and instantaneous frequency plot, are consistent with the behaviour of group Ia spindle afferent fibres to ramp-and-hold stretches. These features include the increase in discharge rate during the application of the muscle stretch, an abrupt fall in the impulse rate at the end of the onset ramp (seen most clearly in the histogram), the slow adaptation of the response at the new steady muscle length, and a brief period of 'silence' in the impulse trace as the muscle was returned to its shorter length. These properties were also reported for neurones of the main cuneate nucleus (Mackie et al. 1998).
 |
View larger version [in this window] [in a new window] |
|
|
A, 2 mm ramp-and-hold stretch (0·5 s onset and offset ramps; waveform at top) applied to the IPr muscle generated a maintained response in an ECN neurone, shown by the impulse record (top trace), the peristimulus time histogram (constructed from 6 successive responses; bin width 80 ms) (middle), and the instantaneous frequency plot (bottom). B, stimulus-response relations for 7 ECN neurones tested over the 2 mm range of stretch applied to forearm extensor muscles (indicated next to each curve). Each point shows the mean impulse counts (± S.D.) evoked in response to 6 repetitions of the 5 s static component of muscle stretch (of the same form as that represented in A). S.D. fell within the bounds of some symbols here and in other figures. | ||
The maintained, tonic responses of ECN neurones (e.g. Fig. 1A) displayed a greater degree of irregularity in inter-impulse intervals than is shown by primary spindle afferents from either the cat forelimb (Mackie et al. 1998) or hindlimb (Matthews & Stein, 1969). Calculations were made of the coefficient of variation of inter-impulse intervals (S.D./mean interval) in the final 2 s of static stretches lasting 5 s, during which the discharge rates of ECN neurones were 20-40 s-1. These were found to range between 0·32 and 0·69, somewhat higher than reported values of 0·17-0·30 for spindle afferent fibres (Matthews & Stein, 1969; Mackie et al. 1998), but were consistent with the values (0·35-0·65) reported for main cuneate neurones (Mackie et al. 1998) and for dorsal spino-cerebellar tract (DSCT) neurones (Jansen et al. 1966) activated by muscle spindle inputs.
Graded, approximately linear stimulus-response relations were obtained for ECN neurones when different magnitudes of muscle stretch (up to 2 mm) were applied to the individual forearm muscles, as shown in Fig. 1B. These relations show the mean response (± S.D.) for seven different ECN neurones as a function of the magnitude of the 5 s steady or hold component of muscle stretches (of the form shown at the top of Fig. 1A). Although there was some variability in the slope of these relations and in the absolute response levels at given amplitudes of stretch, there were no striking differences depending upon the forearm extensor muscle studied. For example, the two neurones activated by EDL stretch displayed the least steep and the steepest slopes observed among the set of stimulus-response relations in Fig. 1B. The maximum stretch imposed (consisting of the 'background' of 1-2 mm (see Methods), plus the further 2 mm of static stretch stimulus) for constructing the Fig. 1B stimulus-response relations corresponded to approximately half of the physiological range of stretch for these forearm muscles, and therefore approximately half the joint rotation range controlled by these muscles. The 2 mm imposed static stretch therefore covered approximately the middle one-fifth to one-third of the physiological range of 6-10 mm for these muscles. These estimates were based, in each experiment, on an initial observation of the length changes produced by the flexion-extension movements of the associated joints prior to tendon detachment.
The low variability in response level at each amplitude of static stretch, indicated by the S.D. bars, provides a guide to the reliability of the response as a signal of static muscle length. No systematic change in response variability was apparent with changes in response magnitude. Background impulse rates (at the zero stretch amplitude in Fig. 1B) averaged about 10 impulse s-1, which was similar to values for muscle-related MCN neurones (Mackie et al. 1998) and for many tactile cuneate neurones (Douglas et al. 1978; Ferrington et al. 1987a,b) and may be largely attributable to spontaneous activity in the central neurones rather than a background muscle input.
Each stretch-sensitive neurone showed a marked sensitivity to the dynamic components of muscle stretch stimuli, illustrated by the sharp increase in spike discharge rates during the onset ramp component of stretch and the sharp decrease, or cessation in response during the ramp-off components of muscle stretches (e.g. Fig. 1A). However, the use of controlled, sinusoidal vibration applied longitudinally to muscles provided a more demanding test of dynamic sensitivity.
Responses to muscle vibration
The sensitivity of ECN neurones to the dynamic components of muscle stretch was examined by applying 1 s trains of sinusoidal stretch displacements to the muscle tendon at frequencies up to 800 Hz. These controlled vibratory perturbations were superimposed on background static stretches lasting 5 s (of the form shown in Fig. 1), and commenced 3 s after the start of the 'hold' component of the stretch to allow adaptation of the response to a new stable level. Figure 2A shows the responses of a representative ECN neurone to 1 s vibration trains at a range of frequencies (50-800 Hz), all at a peak-to-peak amplitude of 10 µm, applied to the EPL muscle. The response traces and the 'bandwidth curves' in Fig. 2B show (i) the marked elevation in activity above that elicited by the background static stretch where the mean response (
in Fig. 2B) was 25 impulses s-1, (ii) that spike discharge rates were graded over a range of amplitudes for a given frequency, and (iii) that the highest response levels were evoked over a broad range of frequencies from 150-500 Hz but that some capacity to respond to even higher frequency (800 Hz) vibro-stretch disturbances existed. These attributes of the muscle-activated ECN neurones were all consistent with the input being derived predominantly from primary spindle afferents as neither the secondary spindle endings nor Golgi tendon organ receptors display such sensitivity to dynamic components of muscle stretch (Cooper, 1961) or vibration (Brown et al. 1967; and see Jack, 1978, for review).
 |
View larger version [in this window] [in a new window] |
|
|
A, impulse traces recorded from an ECN neurone in response to 1 s trains of sinusoidal vibration applied to the EPL muscle, at 10 µm amplitude and frequencies of 50-800 Hz (as indicated), all superimposed on a background 1 mm steady stretch. Bottom trace shows 150 Hz vibration stimulus train. B, mean response (impulses s-1, ± S.D., from 6 responses to 1 s trains of vibration) is plotted as a function of vibration frequency to show the bandwidth of vibration sensitivity at a series of different amplitudes (5, 10, 20, 50 µm peak-to-peak). Impulse counts for the same neurone for which response traces are shown in A. | ||
Phase locking of ECN neurone responses to muscle vibration
In a previous study (Mackie et al. 1998), we quantified the precision of phase locking in the responses of primary spindle afferent fibres and MCN neurones to muscle vibration and, in the present study, have applied the same analytical procedures to the responses of ECN neurones, in order to evaluate the fidelity with which the temporal features of afferent fibre responses to muscle vibration were relayed across the first synaptic junction in the cuneo-cerebellar pathway. This permitted direct quantitative comparison of responses of individual ECN neurones with those of the MCN (Mackie et al. 1998) for their capacities to relay precise temporal information about dynamic muscle length changes to their output targets, principally in the cerebellum and the thalamus.
The effect of vibration amplitude and frequency on the extent to which ECN responses were phase locked to the muscle vibration was evaluated by constructing phase-scatter plots and cycle histograms (e.g. Fig. 3A-C), from which quantitative measures, in particular, the resultant (R), were derived (see Methods). The phase-scatter graphs on the left-hand side of Fig. 3A-C, on the ordinate, show the time (or phase) of impulse occurrence within the vibration cycle period (denoted by the sine wave on the left, which lasts for 6-7 ms for the uppermost graph in Fig. 3A where the vibration frequency was 150 Hz) as a function of time from the onset of the 1 s long vibration train. The phase-scatter graphs were based on the accumulated impulse occurrences from the responses of a representative ECN neurone to five applications of the 1 s train of longitudinal muscle vibration, applied at frequencies of 150 Hz (A), 300 Hz (B) and 500 Hz (C), each at a peak-to-peak amplitude of 20 µm.
 |
View larger version [in this window] [in a new window] |
|
|
A-C, phase-scatter plots and cycle histograms constructed from the responses of an ECN neurone to 5 applications of 1 s trains of vibration (20 µm amplitude) applied at frequencies of 150, 300 and 500 Hz, respectively. Phase-scatter plots show the time of impulse occurrences (single dots) during each cycle period (ordinate) as a function of time after the onset of the 1 s vibration train (abscissa). Cycle histograms share the same ordinate as scatter plots and display the cumulative distribution of impulses within the vibration stimulus cycle (abscissa: number of impulses in each of the 50 bins into which each ordinate scale is divided). R (resultant) value accompanying each distribution shows the level of phase locking in response (see text). D, tightness of phase locking indicated by R (resultant) values, plotted as a function of amplitude for vibration frequencies in the range 50-800 Hz (for the same neurone whose data are presented in A-C). E, representative points of the same data shown in D, re-plotted to show phase locking as a function of frequency. The 0 Hz value at the origin of the abscissa represents values of R obtained in the absence of vibration. | ||
The cycle histograms on the right-hand side in A, B and C were constructed from the same sets of responses, and display the cumulative distribution of impulse events elicited by the five stimulus applications at each frequency of muscle vibration. The two plots in A show that almost all impulses generated in response to the 150 Hz muscle vibration were confined within a narrow segment (20 %) of the vibration cycle period, reflected in the very high vector strength or resultant (R = 0·97) for the distributions. Furthermore, as the phase-scatter graph reveals any trend in the phase of response over the 1 s duration of the vibration train, it is apparent that the phase locking of responses at 150 Hz is even tighter and more stable once the first 100 ms have elapsed. In that initial segment of response there is a brief phase drift and a few outlying impulse occurrences, reflecting the presence of pairs of impulses on some of the early cycles of vibration. A brief initial phase drift is also apparent in responses to the 300 and 500 Hz vibration stimuli, but thereafter a stability of response phase is apparent, as it was for the 150 Hz response. Although the overall scatter of impulse occurrences is greater at the higher vibration frequencies, reflected in the fall in the value of R to 0·85 (at 300 Hz) and 0·72 (at 500 Hz), there is still very substantial phase locking, even at the highest frequency of 500 Hz.
Systematic quantitative assessment of the effect of amplitude and frequency of muscle vibration on the precision of phase locking in ECN responses was obtained from data such as that in Fig. 3A-C and plotted as a function of the amplitude and frequency parameters of the vibro-stretch stimulus (e.g. Figs 3D and E and 4). Figure 3D shows that at vibration frequencies of 50 and 150 Hz the phase locking in the ECN responses was very tight (R > 0·8) at vibration amplitudes as low as 2-5 µm and remained essentially independent of amplitude over the broad amplitude range up to 200 µm. At higher frequencies of vibration (300-500 Hz) the phase locking was poorer, but nevertheless, over a broad range of vibration amplitudes (>100-150 µm) at these frequencies, the responses displayed R values of 0·6, which are considered to represent moderate levels of phase locking (see Lavine, 1971). Furthermore, it is quite striking that even at 800 Hz there is some evidence of phase locking in the responses of this ECN neurone to muscle vibration (R > 0·4 at amplitudes of 20-40 µm). The frequency bandwidth over which phase-locked responses could be obtained for the neurone is represented in Fig. 3E by the plot of R values (for four vibration amplitudes) as a function of vibration frequency. Despite there being a progressive decline in R values at vibration frequencies above 150 Hz, phase locking was retained at substantial levels (R 0·5) to vibration frequencies up to at least 500 Hz.
Although considerable variation was found (Fig. 4A-C) from neurone-to-neurone in the indices of phase locking, the same fundamental trend in the tightness of phase locking as a function of vibration frequency was apparent for the eight ECN neurones examined in this way for their capacity to signal temporal information about vibro-stretch disturbances in the forearm extensor muscles. The three graphs in Fig. 4A-C plot, for these eight ECN neurones, the resultant (R) as an index of phase locking in response to vibration frequencies in the range 50-800 Hz. At each amplitude (20, 50 and 100 µm) in Fig. 4A-C, significant levels of phase locking were observed for all points except for the seven points designated N, where values of R were not significantly different (P > 0·05) from values obtained in random distributions, based on Rayleigh's z statistic (see Methods), or from values indicated at the origin of the abscissa at 0 Hz in Fig. 4A-C, obtained as control values in the absence of vibration.
 |
View larger version [in this window] [in a new window] |
|
|
A-C, R values for up to 8 ECN neurones plotted against muscle vibration frequency at amplitudes of 20, 50 and 100 µm, respectively. The particular forelimb extensor muscle providing input to each neurone is indicated in the legend. Points marked N, including those circled in A were not significantly different from random distributions based on Raleigh's z statistic (see Methods). The 0 Hz value at the origin of the abscissa represents values of R obtained in the absence of vibration. D-F, pooled R values (means ± S.D.) for up to 8 ECN neurones and up to 6 MCN neurones (data from Mackie et al. 1998; see text) as a function of frequency at amplitudes of 20, 50 and 100 µm, respectively. In most cases muscle vibration was superimposed on a 2 mm stretch of muscle (in a few cases muscle stretch was 1 mm). The two groups of neurones (MCN and ECN) were not significantly different at each of the three vibration amplitudes (P =~ 0·18, 0·15, 0·08 for 20, 50 and 100 µm, respectively; Student's paired t test). | ||
The mean values (± S.D.) for the pooled eight ECN neurones have been plotted as a function of vibration frequency in the lower graphs (Fig. 4D -F), together with mean values calculated from the equivalent measures of R for six MCN neurones whose individual data were plotted in Fig. 11 of Mackie et al. (1998). The same trend with vibration frequency is apparent for both ECN and MCN neurones and no systematic difference is apparent between the two samples. Student's paired t test confirmed the absence of a significant difference at the 5 % level between the two groups.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Signalling of information about static muscle stretch by ECN neurones
This report provides the first quantitative evaluation of how effectively forelimb muscle spindle signals of static muscle stretch are transmitted across the ECN synaptic relay in the path predominantly directed to the cerebellum. The ECN neurones displayed approximately linear stimulus-response relations and little evidence of a plateau in these relations over the 2 mm range of static stretch employed. They also show a high level of consistency (low S.D. values in Fig. 1B) in their responses at fixed amplitudes of static stretch. In all these respects they therefore resemble muscle-related neurones of the MCN (Mackie et al. 1998) and appear to display a similar capacity to that of MCN neurones for contributing discriminative information about the degree of static muscle stretch. Earlier studies have provided only fragmentary observations on ECN neurone responses to static muscle stretch, although previous observations on the 'equivalent' cerebellar relay neurones for hindlimb spindle inputs in the DSCT had provided some evidence for similar linear stimulus-response relations (Jansen & Rudjord, 1965; Jansen et al. 1966). As the DSCT counterparts commonly receive convergent group Ia and II spindle afferent input (for reviews, see Mann, 1973; Walmsley, 1991), this may also be the case for ECN neurones. However, we have not specifically addressed this issue, nor whether group Ib inputs are directed to a separate subset of neurones, as appears to be the case for hindlimb muscle inputs to DSCT neurones (Mann, 1973).
Although Rosén & Sjölund (1973a) reported that their sample of muscle-activated MCN and ECN neurones responded to static stretch, no stimulus-response data were presented. Hummelsheim & Wiesendanger (1985), in studying muscle-related MCN and ECN neurones in the macaque monkey, used only brief (30-40 ms) 'ramp and hold' stimuli of low amplitude ( 80 µm). As the responses they observed lacked any maintained activity, it is possible that the input may have arisen from phasically sensitive Pacinian corpuscles which the authors allowed may have contaminated the input generated by these abrupt stretch stimuli. In the present study, the extensive forearm deafferentation ensured that the dynamic components of the ECN responses could not have arisen from extramuscular Pacinian corpuscles. However, we cannot exclude unequivocally some intramuscular Pacinian contribution (Barker, 1967; Jack, 1978), although Barker (1967) has commented that Paciniform corpuscles are scarce in muscle, and that the larger Pacinian corpuscles are seldom encountered. Furthermore, we did not encounter any Pacinian-related afferent fibres among single fibres recorded from various forearm extensor muscle nerves (Mackie & Rowe, 1997; Mackie et al. 1998). In view of these considerations, we believe that the combination of the maintained responses of ECN neurones to static muscle stretch, and their sensitivity to low-amplitude muscle vibration, is most probably attributable to primary spindle input.
Responsiveness of ECN neurones to forelimb muscle vibration
In response to sinusoidal stretch superimposed upon a background static stretch, the ECN neurones displayed an augmented response (Fig. 2A) resembling that seen in both the primary spindle afferents and the MCN neurones (Mackie et al. 1998). Furthermore, as this response was clear for some neurones, even at vibration amplitudes of 5 µm (Fig. 2B), and was apparent over a broad frequency range, from 50 to 800 Hz (Figs 2B, 3 and 4), it appears that ECN neurones display a sensitivity and bandwidth to vibro-stretch perturbations in skeletal muscle that at least matches that for the MCN neurones (cf. Figs 4 and 6 in Mackie et al. 1998). Furthermore, as these response measures approximate those of the primary spindle afferent fibres (cf. Figs 4 and 5 in Mackie et al. 1998) there appears to be no substantial limitation imposed at the ECN synaptic junction on (i) the threshold for detection of dynamic stretch events in forearm muscles, and (ii) the frequency range of the vibrational stretch disturbances that can be detected.
Capacity of ECN neurones to retain temporal precision in the signalling of vibro-stretch information from skeletal muscle
In their responses to muscle vibration ECN neurones display high levels of phase locking over a broad range of frequencies and over a broad range of amplitudes at a given frequency (Figs 3 and 4). Tightest phase locking was observed at vibrational stretch frequencies of 50-150 Hz, as was also found for MCN neurones (Mackie et al. 1998). The quantitative indices of phase locking calculated as the resultant, R, at these 'best' frequencies were 0·8-0·9, indicative of very tight phase locking in the ECN responses (Lavine, 1971; Bledsoe et al. 1982; Greenstein et al. 1987) and were very similar to those of MCN responses at these stretch frequencies (Fig. 4). Although there was a tendency for phase locking to improve at higher amplitudes of muscle vibration (>10 µm; Fig. 3E), as was observed in MCN neurones (Figs 7, 8 and 11 in Mackie et al. 1998), there were individual ECN neurones (Fig. 3D) that displayed consistently high R values (R > 0·8) over the complete range of vibration amplitudes tested, from 2 to 200 µm.
At vibration frequencies above the 50-150 Hz range, the phase locking in ECN responses declined gradually as a function of frequency (Figs 3D and 4), as was found previously for their MCN counterparts (Mackie et al. 1998). However, as was found for the MCN neurones, the mean value for the resultant was 0·6 at 300 Hz and 0·4-0·5 at 500 Hz, indicative of moderate levels of phase locking in the responses of both groups of central neurones at muscle vibration frequencies 500 Hz. These quantitative comparisons provided no evidence for any difference in the capacities of individual neurones from the ECN and MCN groups to signal temporal information about the vibro-stretch disturbances to their respective targets.
Both central groups display greater phase scatter in their responses to vibration than do the muscle afferent fibres whose activity is locked with great precision to a fixed phase of the muscle vibration waveform, as reflected in vector strength values of 1·0 in their responses to vibration frequencies up to 200 Hz, and > 0·8 in their responses to vibration frequencies of up to 800 Hz (see Figs 11 and 12 in Mackie et al. 1998), confirming the capacity of the spindle afferent impulse trains to provide a precise metronome-like reflection of the periodicity inherent in the vibrational stimulus. The phase-scatter graphs (e.g. Fig. 3) show that the poorer values of R for the central neurones are not, to any substantial extent, attributable to a systematic drift in the phase of the response over the 1 s duration of the vibration stimulus train. Presumably, the similarly lower values of R in both ECN and MCN responses (compared with spindle afferent responses) demonstrate that synaptic transmission within each of these parallel relays introduces an equivalent deterioration in the reliability of temporal signalling. Part of this deterioration in the capacity of ECN and MCN neurones may be attributable to the convergence on individual central neurones (Gordon & Jukes, 1964; Andersen et al. 1964; Tracey, 1980; Ferrington et al. 1987b) from input fibres, each of which has tightly phase-locked patterns of response to high-frequency vibration ( 800 Hz; Ferrington et al. 1984, 1987a; Mackie et al. 1998), but whose ensemble activity, at least for vibrotactile inputs, is not synchronized at vibration frequencies above 100 Hz (Greenstein et al. 1987; Ferrington et al. 1987b; Rowe, 1990). However, convergence of inputs with asynchronous activity is unlikely to be the entire explanation, as elsewhere we have shown that central neurones whose input comes selectively from a single afferent fibre also show poorer phase locking in response to oscillatory stimuli than does the single input fibre (Ferrington et al. 1987a,b; Rowe, 1990; Vickery et al. 1994; Gynther et al. 1995). It appears therefore that the similar deterioration in fidelity of temporal signalling in both ECN and MCN responses to muscle vibration reflects a similar uncertainty or 'noise' in the synaptic transmission process within these parallel relays, an effect that introduces temporal variability in the central impulse activity (Mackie et al. 1998). Nevertheless, it remains striking that ECN and MCN responses retain tight phase locking to muscle or skin vibration up to 500 Hz, a capacity that may be attributable to a functional domination of the neurones, in particular, their phase of response to vibration stimuli, by the powerful synaptic actions of just one or a few of the convergent afferents (Ferrington et al. 1987b; Rowe, 1990).
Signalling of spindle afferent information to the cerebellum by individual ECN neurones
The results demonstrate that muscle-related ECN neurones in the cuneo-cerebellar pathway for motor control and regulation, match the signalling capacities of their MCN counterparts that are involved principally in transmitting muscle afferent signals to higher centres for kinaesthetic sensation. This was surprising as studies on muscle inputs to cerebellar Purkinje cells have shown that group Ia spindle inputs have weaker actions than those exerted by tactile inputs or by group Ib or II muscle inputs (Eccles et al. 1971a,b; Ishikawa et al. 1972a,b; Iosif et al. 1972). In particular, few Purkinje cells (< 5 %) responded to muscle vibration at amplitudes consistent with a primary spindle input. Furthermore, they were unable to respond in a sustained way to muscle vibration at strengths selective for primary spindle input (Ishikawa et al. 1972a), in contrast to the ECN neurones. The earlier observations on the weak actions of primary spindle inputs on Purkinje cells had themselves been surprising because of some expectation at the time that muscle spindle inputs might be the pre-eminent input for the cerebellar regulation of motor function. Furthermore, these earlier observations were, in part, responsible for the hypothesis behind the present study, that ECN transmission characteristics for processing muscle inputs may be very different from the high-security, temporally precise linkage established recently for muscle-related MCN neurones (Mackie et al. 1998). However, as no such differential capacity was revealed for transmission characteristics across these two parallel synaptic relay sites, it appears that the limitations in the capacities of cerebellar Purkinje cells to respond to primary spindle inputs must be imposed at synapses within the cerebellum, either at the mossy fibre-granule cell relay formed by the cuneo-cerebellar output from ECN, or at the parallel fibre-Purkinje cell synapse itself.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
| Andersen, P., Eccles, J. C., Oshima, T. & Schmidt, R. F. (1964). Mechanisms of synaptic transmission in the cuneate nucleus. Journal of Neurophysiology 27, 1096-1116. | |
| Barker, D. (1967). The innervation of mammalian skeletal muscle. In Myotatic, Kinesthetic and Vestibular Mechanisms, ed. de Reuck, A. V. S. & Knight, J., Ciba Foundation Symposium, pp. 3-15. Churchill, London. | |
| Berman, A. L. (1968). The Brain Stem of the Cat; A Cytoarchitectonic Atlas with Stereotaxic Coordinates. University of Wisconsin Press, Madison. | |
| Bianconi, R. & Van der Meulen, J. P. (1963). The response to vibration of the end organs of the mammalian muscle spindles. Journal of Neurophysiology 26, 177-190. | |
| Bledsoe, S. C., Rupert, A. L. & Moushegian, G. (1982). Response characteristics of cochlear nucleus neurons to 500 Hz tones and noise: findings relating to frequency-following potentials. Journal of Neurophysiology 47, 113-127 | [Medline] |
| Brown, M. C., Engberg, I. & Matthews, P. B. C. (1967). The relative sensitivity to vibration of muscle receptors of the cat. The Journal of Physiology 192, 773-800 | [Medline] |
| Cooke, J. D., Larson, B., Oscarsson, O. & Sjölund, B. (1971). Origin and termination of cuneo-cerebellar tract. Experimental Brain Research 13, 339-358 | [Medline] |
| Cooper, S. (1961). The responses of the primary and secondary endings of muscle spindles with intact motor innervation during applied stretch. Quarterly Journal of Experimental Physiology 46, 389-398. | |
| Douglas, P. R., Ferrington, D. G. & Rowe, M. J. (1978). Coding of information about tactile stimuli by neurones of the cuneate nucleus. The Journal of Physiology 285, 493-513 | [Abstract] |
| Eccles, J. C. Faber, D. S., Murphy, J. T., Sabah, N. H. & Táboriková, H. (1971a). Afferent volleys in limb nerves influencing impulse discharges in cerebellar cortex. II. In Purkyne cells. Experimental Brain Research 13, 36-53 | |
| Eccles, J. C., Sabah, N. H., Schmidt, R. F. & Táboriková, H. (1971b). Cerebellar Purkyne cell responses to inputs from cutaneous mechanoreceptors. Brain Research 30, 419-424 | [Medline] |
| Ferrington, D. G., Hora, M. O. H. & Rowe, M. J. (1984). Functional maturation of tactile sensory fibers in the kitten. Journal of Neurophysiology 52, 74-85 | [Medline] |
| Ferrington, D. G., Rowe, M. J. & Tarvin, R. P. C. (1987a). Action of single sensory fibres on cat dorsal column nuclei neurones: vibratory signalling in a one-to-one linkage. The Journal of Physiology 386, 293-309 | [Abstract] |
| Ferrington, D. G., Rowe, M. J. & Tarvin, R. P. C. (1987b). Integrative processing of vibratory information in cat dorsal column nuclei neurones driven by identified sensory fibres. The Journal of Physiology 386, 311-331 | [Abstract] |
| Gordon, G. & Jukes, M. G. M. (1964). Dual organization of the exteroceptive components of the cat's gracile nucleus. The Journal of Physiology 173, 263-290. | |
| Gordon, M., Rubia, F. J. & Strata, P. (1972). The effect of barbiturate anaesthesia on the transmission to the cerebellar cortex. Brain Research 43, 677-679 | [Medline] |
| Greenstein, J., Kavanagh, P. & Rowe, M. J. (1987). Phase coherence in vibration-induced responses of tactile fibres associated with Pacinian corpuscle receptors in the cat. The Journal of Physiology 386, 263-275 | [Abstract] |
| Gynther, B. D., Vickery, R. M. & Rowe, M. J. (1995). Transmission characteristics for the 1:1 linkage between slowly adapting type II fibers and their cuneate target neurons in cat. Experimental Brain Research 105, 67-75 | [Medline] |
| Hummelsheim, H. & Wiesendanger, M. (1985). Neuronal responses of medullary relay cells to controlled stretches of forearm muscles in the monkey. Neuroscience 16, 989-996 | [Medline] |
| Iosif, G., Pompeiano, O., Strata, P. & Thoden, U. (1972). The effect of stimulation of spindle receptors and Golgi tendon organs on the cerebellar anterior lobe II. Responses of Purkinje cells to sinusoidal stretch or contraction of hindlimb extensor muscles. Archives Italiennes de Biologie 110, 502-542. | |
| Ishikawa, K., Kawaguchi, S. & Rowe, M. J. (1972a). Actions of afferent impulses from muscle receptors on cerebellar Purkynè cells. I. Responses to muscle vibration. Experimental Brain Research 15, 177-193 | [Medline] |
| Ishikawa, K., Kawaguchi, S. & Rowe, M. J. (1972b). Actions of afferent impulses from muscle receptors on cerebellar Purkynè cells. II. Responses to muscle contraction: effects mediated via the climbing fibre pathway. Experimental Brain Research 16, 104-114 | [Medline] |
| Jack, J. J. B. (1978). Some methods for selective activation of muscle afferent fibres. In Studies in Neurophysiology, ed. Porter, R., pp. 155-176. Cambridge University Press, Cambridge. | |
| Jansen, J. K. S., Nicolaysen, K. & Rudjord, T. (1966). Activity in the dorsal spinocerebellar tract induced by muscle stretch. In Control and Innervation of Skeletal Muscle, ed. Andrew, B. L., pp. 119-124. Livingstone, Edinburgh. | |
| Jansen, J. K. S. & Rudjord, T. (1965). Dorsal spinocerebellar tract: Response pattern of nerve fibers to muscle stretch. Science 149, 1109-1111 | [Medline] |
| Lavine, R. A. (1971). Phase-locking in response of single neurons in cochlear nuclear complex of the cat to low-frequency tonal stimuli. Journal of Neurophysiology 34, 467-483 | [Medline] |
| Lundberg, A. & Winsbury, G. (1960). Selective adequate activation of large afferents from muscle spindles and Golgi tendon organs. Acta Physiologica Scandinavica 49, 155-164. | |
| McCloskey, D. I. (1978). Kinesthetic sensibility. Physiological Reviews 58, 763-820 | [Medline] |
| Mackie, P. D., Morley, J. W. & Rowe, M. J. (1994). Responses of cells in the external cuneate nucleus of the cat to forelimb muscle vibration. Proceedings of the Australian Neuroscience Society 5, 66. | |
| Mackie, P. D., Morley, J. W., Zhang, H. Q., Murray, G. M. & Rowe, M. J. (1998). Signalling of static and dynamic features of muscle spindle input by cuneate neurones in the cat. The Journal of Physiology 510, 923-939 | [Abstract/Full Text] |
| Mackie, P. D. & Rowe, M. J. (1997). An intact peripheral nerve preparation for monitoring inputs from single muscle afferent fibres. Experimental Brain Research 113, 186-188 | [Medline] |
| Mann, M. D. (1973). Clarke's column and the dorsal spinocerebellar tract. Brain Behaviour and Evolution 7, 34-83. | |
| Mardia, K. V. (1972). Statistics of Directional Data. Academic Press, London. | |
| Matthews, P. B. C. & Stein, R. B. (1969). The regularity of primary and secondary muscle spindle afferent discharges. The Journal of Physiology 202, 59-82 | [Medline] |
| Perciavalle, V., Bosco, G. & Poppele, R. (1995). Correlated activity in the spinocerebellum is related to spinal timing generators. Brain Research 695, 293-297 | [Medline] |
| Rosén, I. & Sjölund, B. (1973a). Organisation of group I activated cells in the main and external cuneate nuclei of the cat: Identification of muscle receptors. Experimental Brain Research 16, 221-237 | [Medline] |
| Rosén, I. & Sjölund, B. (1973b). Organisation of group I activated cells in the main and external cuneate nuclei of the cat: Convergence patterns demonstrated by natural stimulation. Experimental Brain Research 16, 238-246 | [Medline] |
| Rowe, M. J. (1990). Impulse patterning in central neurons for vibrotactile coding. In Information Processing in Mammalian Auditory and Tactile Systems, ed. Rowe, M. J. & Aitkin, L. M., pp. 111-125. Wiley-Liss, New York. | |
| Snider, R. S. & Niemer, W. T. (1961). A Stereotaxic Atlas of the Cat Brain. The University of Chicago Press, Chicago. | |
| Tracey, D. J. (1980). The projection of joint receptors to the cuneate nucleus in the cat. The Journal of Physiology 305, 433-449 | [Abstract] |
| Vickery, R. M., Gynther, B. D. & Rowe, M. J. (1994). Synaptic transmission between single slowly adapting type I fibres and their cuneate target neurones in cat. The Journal of Physiology 474, 379-392 | [Abstract] |
| Walmsley, B. (1991). Central synaptic transmission: studies at the connection between primary afferent fibres and dorsal spinocerebellar tract (DSCT) neurones in Clarke's column of the spinal cord. Progress in Neurobiology 36, 391-423 | [Medline] |
| Zar, J. H. (1984) Biostatistical Analysis, 2nd edn. Prentice-Hall, Englewood Cliffs, NJ, USA. | |
The authors acknowledge the technical assistance provided by H. Bahramali, C. Riordan, F. Spicer and P. Farrell. The work was supported by the National Health and Medical Research Council of Australia and the Australian Research Council.
Corresponding author
M. J. Rowe: School of Physiology & Pharmacology, University of New South Wales, Sydney NSW 2052, Australia.
Email: m.rowe{at}unsw.edu.au
This article has been cited by other articles:
|
|
 |
|
  M. K. Zachariah, G. T. Coleman, D. A. Mahns, H. Q. Zhang, and M. J. Rowe Transmission Security for Single, Hair Follicle-Related Tactile Afferent Fibers and Their Target Cuneate Neurons in Cat J Neurophysiol, August 1, 2001; 86(2): 900 - 911. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
