J Physiol Volume 510, Number 3, 923-939, August 1, 1998
The Journal of Physiology (1998), 510.3, pp. 923-939
© Copyright 1998 The Physiological Society
Signalling of static and dynamic features of muscle spindle input by cuneate neurones in the cat
P. D. Mackie, J. W. Morley, H. Q. Zhang, G. M. Murray and M. J. Rowe
School of Physiology and Pharmacology, The University of New South Wales, Sydney 2052, Australia
Received 20 November 1997; accepted after revision 16 April 1998.
 |
ABSTRACT |
- The capacity of cuneate neurones to signal information derived from muscle spindle afferent fibres about static stretch or vibration of forearm extensor muscles was examined electrophysiologically in anaesthetized cats.
- Static stretch (<= 2 mm in amplitude) and sinusoidal vibration (at frequencies of 50-800 Hz) were applied longitudinally to individual muscle tendons by means of a feedback controlled mechanical stimulator, and responses were recorded from individual cuneate neurones and from individual spindle afferent fibres.
- Cuneate neurones sampled were located caudal to the obex and displayed a sensitivity to both vibration and static stretch of forearm muscles that was consistent with their input arising from primary spindle endings. In response to static muscle stretch, they displayed graded and approximately linear stimulus-response relations, and a stability of response level at fixed lengths that was consistent with these neurones contributing discriminative information about static muscle stretch.
- In response to sinusoidal muscle vibration the cuneate neurones also showed graded stimulus-response relations (in contrast to spindle afferents which at low vibration amplitudes attain a plateau response level corresponding to a discharge of 1 impulse on each vibration cycle). Lowest thresholds were at 100-300 Hz and bandwidths of vibration sensitivity extended up to ~800 Hz.
- Temporal precision in cuneate responses to muscle vibration was assessed by constructing phase scatter and cycle histograms from which measures of vector strength could be calculated. Cuneate responses displayed somewhat poorer phase locking (and lower vector strengths) than spindle afferent responses to vibration (a reflection of uncertainties associated with synaptic transmission). Nevertheless, the remarkable feature of cuneate responses to muscle vibration is the preservation of tight phase locking at frequencies up to 400-500 Hz, which presumably enables these central neurones to contribute accurate temporal information for the kinaesthetic sense in a variety of circumstances involving dynamic perturbations to skeletal muscle.
|
INTRODUCTION |
Signals about both static and dynamic components of muscle stretch are conveyed to the central nervous system by muscle spindle stretch receptors. However, the information about dynamic features of the stretch is derived principally from the primary spindle endings (Matthews, 1933; Lundberg & Winsbury, 1960; Cooper, 1961; Bessou & Laporte, 1962; Bianconi & Van der Meulen, 1963; Brown, Engberg & Matthews, 1967; Stuart, Mosher, Gerlach & Reinking, 1970). These primary endings may be activated selectively by longitudinal vibration of the muscle over a broad range of frequencies. In response to this controlled, dynamic form of muscle stretch their impulse activity is phase locked to the vibration waveform, reflecting with great precision the temporal features of the sinusoidal stretch perturbation up to frequencies of 500 Hz (Bianconi & Van der Meulen, 1963; Brown et al. 1967).
Limited data are available on the extent to which this temporal precision in impulse patterning is retained across central synapses in the response behaviour of neurones in the ascending kinaesthetic pathways. Central responses to vibratory muscle stretch observed in earlier studies have been confined largely to brief trains of vibration and in some cases may have been contaminated by other vibration-sensitive inputs, in particular from Pacinian corpuscle receptors. For example, cuneate neurones of the macaque monkey display some entrainment of responses to brief trains (usually 2-8 cycles) of forelimb muscle vibration at frequencies up to 50-100 Hz, but as all forearm nerves remained intact the authors acknowledged that other receptors such as Pacinian corpuscles, joint receptors or muscle spindle secondary endings may have contributed to the responses (Hummelsheim & Wiesendanger, 1985; Wiesendanger & Hummelsheim, 1985). In the cat, nucleus Z neurones responded to primary and secondary spindle inputs generated by brief trains of vibration (usually < 5 cycles) applied to hindlimb muscles but showed little evidence of phase locking to the stimulus (Magherini, Pompeiano & Seguin, 1975). In contrast, neurones of both the main and external cuneate nucleus of the cat display powerful responses to muscle vibration with some evidence for temporal patterning of responses to vibration frequencies of up to 100-200 Hz (Rosén & Sjölund, 1973a). However, none of these studies has attempted to quantify the temporal precision of central responses to vibro-stretch stimulation of skeletal muscle.
In the present study, conducted in the cat and reported in preliminary abstract form (Mackie, Morley, Murray, Zhang, Bahramali & Rowe, 1992; Mackie, Morley, Bahramali, Zhang, Murray & Rowe, 1993), we have examined quantitatively the stimulus-response characteristics of cuneate neurones activated by static stretch and sinusoidal vibration of forearm extensor muscles. In addition, we have quantified the extent of phase locking in responses to muscle vibration in order to compare the capacities of cuneate neurones to signal the temporal features of these dynamic stretch disturbances with those of spindle primary afferent fibres.
|
METHODS |
Data were obtained from experiments in eighteen adult cats, fourteen 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), and four anaesthetized with alphaxolone/alphadolone (Saffan, Glaxo; 18 mg kg-1, I.M.) and maintained by means of an I.V. infusion of the same agent (12-48 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 that deep levels of anaesthesia were maintained throughout the course of the experiments. Rectal temperature was maintained at 37-38·5°C. All the experimental work 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.
Forelimb preparation
In all experiments the left forearm was denervated, apart from the deep radial nerve that supplied the forearm extensor muscles, by sectioning of the median, ulnar, musculocutaneous and superficial radial nerves. Furthermore in some experiments the denervation was extended to include the axillary, suprascapular, subscapular, dorsal and lateral thoracic, and the fifth cervical nerve, as well as branches of the deep radial nerve supplying the triceps and epitrochlearis muscles. A cautery device was also used to disrupt any branches of the dorsal interosseus wrist joint nerve which may have supplied non-muscle receptors in the interosseus membrane or wrist joint.
The tendon of each of the following forearm extensor muscles was cut at the wrist and the individual muscles separated from each other over almost their entire length, without compromising nerve and blood supplies: 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). For experiments in which recordings were made from muscle spindle afferent fibres supplying IPr, the overlying EDC, EDL and ECU muscles were removed. A pool formed by the skin flaps attached to a brass ring was filled with isotonic saline when recordings were made only from central neurones, and filled with liquid paraffin when recordings were made from IPr muscle spindle afferents in the forearm. In five experiments, the radius bone was fixed rigidly to the brass ring by means of a dental acrylic 'bridge', built around a stiff wire strand which was fixed to the brass ring and a small self-tapping screw implanted into the bone. This fixation was an added precaution to ensure that longitudinal stretching of the extensor muscles took place without any movement of the upper part of the limb due to extension at the elbow joint.
Muscle afferent fibre recordings
Segments of the IPr branch of the deep radial nerve, 2-4 cm in length, were dissected free from surrounding muscle tissue, and left in continuity with both the IPr muscle and the parent deep radial nerve (for details see Mackie & Rowe, 1997). Any distal branches of the nerve supplying structures other than the IPr muscle (see Mackie & Rowe, 1998) were cut. A silver hook electrode was placed under the nerve and a thin sheet of plastic film inserted beneath the dissected segment of the nerve to ensure electrical isolation. The impulse activity of muscle spindle afferent fibres was recorded conventionally and stored on magnetic tape and on computer.
Recordings from the main cuneate nucleus
Extracellular recordings were made by means of tungsten microelectrodes (tip impedances of 0·5-2 M
), from single neurones of the main cuneate nucleus which was made accessible by exposure of the dorsal surface of the brainstem in the region of the obex. The brainstem was covered with agar (4 % in saline), which prevented drying of the exposed tissue and minimized brainstem movement due to respiration and arterial pulsations.
As both the mid-line and obex provided visible reference locations on the exposed dorsal surface of the brainstem, it was possible from these co-ordinates to reliably specify recording positions (see Results) and be confident that these were within the main cuneate nucleus. Furthermore, antidromic activation from the contralateral medial lemniscus confirmed the cuneo-thalamic identity of many of the sampled neurones. This was achieved by stimulation with a low impedance tungsten electrode (
0·5 M) that had been inserted into the caudal region of the ventrobasal thalamus and secured by cementing to the overlying skull with dental acrylic. Impulses evoked by lemniscal stimulation were always all-or-none in nature, invariant in both stimulation threshold and response latency, and successfully collided with orthodromically evoked impulses (Darian-Smith, Phillips & Ryan, 1963). The ability to follow high frequencies of stimulation was not considered a suitable criterion for antidromic activation of cuneate cells (and therefore was not employed in this study) as we have shown, in an earlier study, that neurones in this nucleus are capable of responses exceeding many hundreds of impulses per second with orthodromic stimulation (e.g. Douglas, Ferrington & Rowe, 1978).
Mechanical stimulation of forelimb extensor muscles
A feedback-controlled mechanical stimulator was attached securely to individual muscle tendons in order to apply controlled ramp-and-hold stretches of up 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 resting tension, just enough to take up any slack in the muscle under observation, was applied and measured 
5-10 g wt based on a calibrated, DC-amplified feedback signal generated by the stimulator. 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, allowing enough time in each instance for full recovery of both muscle and neurone properties to obtain consistently reproducible activity at any given set of stimulus conditions. 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. When certain stimulus parameters were altered so as to obtain a graded range of responses, this was done so in a pseudo-random order to prevent adaptation or habituation effects accounting for any observed trends.
Quantitative analysis of spindle afferent and cuneate neurone responses to muscle stretch and vibration
Stimulus-response relations were constructed in order to quantify the sensitivity and responsiveness of peripheral fibres and cuneate neurones in their responses to muscle stimulation. These were obtained by plotting, as a function of static or vibratory stretch amplitude, the mean response (impulses s-1 ± S.D.) for five to ten repetitions of a fixed static stretch or vibratory stimulus applied via the tendon to the individual forearm extensor muscles. The frequency bandwidth of responsiveness to muscle vibration was also plotted for individual units in terms of response level (impulses s-1) as a function of vibration frequency at fixed vibration amplitudes.
The overall distributions of impulse activity in relation to the controlled muscle stimuli were displayed in the form of peristimulus time histograms, and 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 computer-constructed phase scatter graphs (Figs 7 and 9) 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 cycle histograms (Figs 7 and 9) use a pulse associated with the onset of each successive cycle in the vibration train as the stimulus marker and display the probability of an impulse occurring at different times throughout the period of the vibration cycle. The analysis period corresponds to the cycle period of the vibration and was divided into a number of time segments (usually 50). The cycle histograms were constructed from the impulse activity recorded in response to 900-3000 cycles of muscle vibration. As these distributions are used to quantify the extent of phase locking in a cyclic event it is most appropriate to use directional or angular statistics (Mardia, 1972). The measure obtained is the resultant, R, which is a measure of vector strength in a cyclic distribution. The resultant measures 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. This measure has been used in evaluating the entrainment of auditory responses to tonal stimuli (e.g. Lavine, 1971; Bledsoe, Rupert & Moushegian, 1982) and in the somatosensory system for quantifying the entrainment of responses to vibrotactile stimuli (Greenstein, Kavanagh & Rowe, 1987; Vickery, Gynther & Rowe, 1994). The resultant, R, was calculated from each of these 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).
|
RESULTS |
Muscle-related cuneate neurones
Cuneate neurones selected for study were highly sensitive to both steady stretch and low amplitude, longitudinal vibration of individual forearm extensor muscles. Their sensitivity to static muscle stretch, manifest as a maintained response, was consistent with the input coming from spindle afferent fibres rather than, for example, Pacinian receptors. Although any inputs from the latter should have been eliminated by the comprehensive forearm denervation (see Methods) they may sometimes be found in the fascia surrounding the skeletal muscle itself (Barker, 1967). Care was also taken to confirm that neurones selected for study were activated only by muscle stimulation and not from any extra-muscular forearm sources such as skin, or structures in the upper part of the limb.
Location and identity of muscle-related cuneate neurones
All eighteen neurones studied were located between 0 and 3·5 mm caudal to the obex (with all but one within 2 mm of the obex) and all but one between 1·0 and 2·5 mm lateral to the mid-line. Their depths below the brainstem surface ranged from 800 to 2200 µm. The locations were entirely consistent with the neurones being within the cuneate nucleus and, in rostro-caudal position, were consistent with Rosén's report (1969) that the largest 'focal potentials' evoked in the cuneate nucleus by electrical stimulation of the deep radial nerve at group I strength were usually recorded immediately caudal to the obex. Furthermore, as seven of eight neurones studied for antidromic activation from the contralateral medial lemniscus (see Methods) were identified as thalamic projecting neurones it is probable that most of the muscle-related neurones sampled were cuneothalamic neurones located caudal to the obex extending into the middle zone of the nucleus (Hand & Van Winkle, 1977).
Eighteen cuneate neurones that met the above requirements were examined in detail for their responsiveness to controlled static stretch or vibro-stretch stimulation of individual forearm muscles. The eighteen were made up of six activated by stretch of APL, four by EDC, three by ECRB, two by EPL, and one by each of ECU, EDL and IPr. In order to make reliable and systematic comparison between cuneate response behaviour and that of primary spindle afferent fibres under similar experimental conditions, a sample of six spindle afferents was also examined in detail with the same controlled static and vibrational stretch of forearm muscles.
Extent of muscle afferent convergence onto cuneate neurones
Although all eighteen neurones studied were examined for their responsiveness to stretch of each of the dissected forearm extensor muscles, all but one were activated selectively by just one muscle. This low incidence of suprathreshold convergence (6 % of neurones) was consistent with the very limited convergence reported for muscle-related neurones of the macaque cuneate nucleus (Hummelsheim & Wiesendanger, 1985) but lower than the 28 % incidence of convergence reported by Rosén & Sjölund (1973a,b) for cuneate neurones activated by forearm extensor muscles in the cat. However, in the latter study, individual muscles may not have been as extensively separated from one another as was the case in the present study (see Methods). If just the individual tendons are freed as appears to have been the case in the study of Rosén & Sjölund (1973a,b), there is greater likelihood of mechanical spread of the imposed stretch from one muscle to another.
Cuneate neurone responses to static muscle stretch
The capacity of cuneate neurones to respond sensitively and in a maintained manner to static muscle stretch is shown in Fig. 1 for a cuneothalamic neurone, associated with the large spike in the impulse trace, activated by a 2 mm static stretch of EDL lasting 5 s (with 0·5 s onset and offset ramps). The response features of the neurone, as reflected in the peristimulus time histogram, the instantaneous frequency plot, and the impulse record, are consistent with the behaviour of group Ia spindle afferent fibres in their responses to ramp-and-hold stretches (see, for example, Matthews, 1972). In particular, these features are the substantial increase in response rate during the application of the stretch, an abrupt fall in discharge frequency at the end of the onset ramp (most clearly seen in the histogram), the slow adaptation of the response at the new muscle length, and the brief period of impulse silence at the moment the muscle is returned to its shorter length. These same properties are also evident in the responses illustrated for another cuneate neurone (Fig. 3C and D), as well as for a muscle spindle afferent fibre (Fig. 3A and B), and were present in all cuneate neurones and afferent fibres studied. The cuneate response characteristics confirm unequivocally the strong maintained response to the static muscle disturbance, which is in contrast to the reported failure of nucleus Z neurones to signal information about static muscle length (Magherini et al. 1975) (see also Discussion).
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 1. Capacity of cuneate neurones to respond to static muscle stretch
A 2 mm ramp-and-hold stretch lasting 5 s (0·5 s onset and offset ramps; waveform at top) generated a maintained response reflected, first, in the impulse record (large spike, ~1 mV peak-to-peak amplitude; negativity upwards here and in other response traces), second, in the peristimulus time histogram (constructed from 6 successive responses; address width, 80 ms), and third, in the instantaneous frequency plot (lowest panel). The neurone was identified as a cuneothalamic projecting neurone according to criteria described in Methods.
|
The maintained response of the cuneate neurone to static stretch in Fig. 1, reflected in both the spike trace and the instantaneous frequency plot, displays occasional pairs of spikes at high frequency, and a greater degree of irregularity in interimpulse intervals than that seen in primary spindle afferents. For example, the coefficients of variation for interimpulse intervals (S.D./mean interval) were assessed over the last 2 s of the static 5 s stretch (during which the mean impulse rates were 20-40 s-1), and were found to be 0·5-0·65 for two representative cuneate neurones. In contrast, the coefficients of variation assessed in the same way for three representative spindle afferent fibres were found to be 0·17-0·3. Furthermore, this degree of variation in the afferent fibre was found whether or not the fibre was 'de-efferented' (i.e. whether the nerve was cut proximally to the peripheral recording electrode, or left intact), which also indicates that under the present experimental conditions for analysing cuneate responses (e.g. the prevailing depth of anaesthesia, and with 
-innervation of spindles intact) there was little or no influence of fusimotor activity upon the responses observed. The coefficients of variation in the responses of the cuneate neurones were consistent with those reported for Ia-activated dorsal spinocerebellar tract (DSCT) neurones (0·35-0·6, Jansen, Nicolaysen & Rudjord, 1966). Although the variability in the spindle afferent fibre responses was somewhat higher than that reported previously for de-efferented primary endings supplying the hindlimb soleus muscle (0·06, Matthews & Stein, 1969), the difference is probably related to the times at which the assessments were made; in our study, for 2 s starting 3 s after the onset of the static component of the muscle stretch and, in the study of Matthews & Stein, a delay of 5-10 s elapsed after applying a stretch (perhaps achieving greater stationarity in responses) before analysing the interpulse intervals.
When different magnitudes of static stretch were applied to individual forearm muscles the cuneate neurones displayed graded, approximately linear stimulus-response relations over the 2 mm range of static stretch that could be delivered with the feedback-controlled mechanical stimulator (Fig. 2). The stimulus-response relations in Fig. 2 plot the mean impulse counts (± S.D., n = 6) during the 5 s period of static stretch as a function of different stretch amplitudes for five cuneate neurones. Only one of the five relations appears to attain a plateau level of response within this 2 mm stretch range (which was superimposed on a small pre-existing stretch, see Methods). There appeared to be no striking muscle-to-muscle differences among the forearm extensors studied, as both the steepest and least-steep relations in Fig. 2 were obtained for cuneate neurones with EDC input. The maximum stretch imposed (background plus the 2 mm stretch) for the Fig. 2 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 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. Although the variability is slightly greater at some of the largest stretch amplitudes, there is little evidence for a systematic change in response variability with response magnitude.
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 2. Stimulus-response relations for five cuneate neurones responsive to static stretch of forearm extensor muscles
The neurones exhibited a graded response as a function of static stretch amplitude over the 2 mm range tested. Each point represents the mean (± S.D.) impulse count in six consecutive responses to the 5 s static stretch of different extensor muscles. (S.D. fell within the bounds of some symbols here and in other figures.) The top four labelled curves were from neurones tested and confirmed as cuneothalamic projecting neurones.
|
Responses of cuneate neurones to vibratory stretch of skeletal muscle
The capacity of cuneate neurones to respond to the dynamic features of muscle stretch was examined by applying 1 s long, sinusoidal stretch displacements to the forearm extensor muscles via their tendons. These vibro-stretch perturbations were superimposed on a background static stretch of the muscle lasting 5 s, applied after a delay of 3 s to ensure that the neurone had adapted to an approximately stable level of response to the static component of stretch (Fig. 3). The impulse traces and peristimulus time histograms of Fig. 3 are consistent with the known effectiveness of the superimposed dynamic perturbation, in this case at 150 Hz, for activating primary spindle afferent fibres (Fig. 3A and B; and Lundberg & Winsbury, 1960; Bianconi & Van der Meulen, 1963; Brown et al. 1967; Stuart et al. 1970), but also demonstrate that cuneate neurones display a clear and maintained increase in activity above that elicited by static stretch (Fig. 3C and D). Furthermore, the response level in cuneate neurones may be sensitively graded as a function of the intensity of the vibro-stretch stimulus (Fig. 4B) in contrast to the individual primary spindle afferents which at low amplitudes of vibration attain a 1 : 1 plateau level of response corresponding to the discharge of one impulse per cycle of vibration (Fig. 4A). In Fig. 4A the spindle afferent retains its 1 : 1 plateau level over a broad range of amplitudes, from 10 to >150 µm, in response to the 150 Hz vibration of the muscle, whereas the cuneate response is graded over the whole 200 µm intensity range. If one were to assess the responsiveness of the cuneate neurone over very brief time spans, <100 ms, the range of amplitudes over which graded response levels are displayed would be narrower (Fig. 4B). However, even over the first 100 ms segment of response in Fig. 4B, the cuneate response is graded over an amplitude range of 100 µm, many times (5-10 times) the range of intensities over which the spindle afferent response is graded (Fig. 4A).
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 3. The capacity of cuneate neurones and spindle afferent fibres to respond to muscle vibration superimposed upon static stretch of the muscle
Impulse records and peristimulus time histograms (constructed from 6 consecutive responses; address width, 80 ms) for a representative spindle afferent fibre (A and B) and cuneothalamic neurone (C and D) activated by both static muscle stretch (1 mm amplitude, 5 s duration) and sinusoidal vibration (150 Hz, 1 s duration starting 3 s after onset of stretch; amplitude, 20 µm for A and B, 25 µm for C and D) applied to the muscle tendon (vibration amplitude is exaggerated on stimulus waveform in upper trace for purposes of illustration). Spike height in A, 0·5 mV; in B, 2·6 mV.
|
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 4. Amplitude dependence of responses to muscle vibration in representative spindle afferent fibre (A) and cuneothalamic neurone (B)
Vibration at 150 Hz (1 s duration) was applied longitudinally to the muscle tendon. The primary afferent attained a plateau 1 : 1 response level at ~10 µm whereas the cuneate neurone response was graded up to 150-200 µm. Vibration was superimposed on a 2 mm stretch of each muscle (cf. Fig. 3) applied from a low background tension of ~5-10 g wt. Spike heights, ~0·5 mV (A) and 1 mV (B).
|
Quantification of the response (in impulses s-1) as a function of the amplitude of the vibro-stretch displacement permitted construction (in Figs 5 and 6) of stimulus-response relations for the representative primary spindle afferent and cuneate neurone of Fig. 4. For the peripheral fibre, the relations in Fig. 5, plotted for five frequencies of muscle vibration from 50 to 800 Hz, show lowest absolute and 1 : 1 thresholds at frequencies of 150 and 300 Hz. At lower (50 Hz) and higher (500 and 800 Hz) frequencies there was no distinct 1 : 1 plateau, although at 50 Hz, a 2 : 1 plateau level of response was attained over the amplitude range 50-100 µm. At 500 Hz, a 1 : 1 response level was just attained at the highest vibration amplitude of 100 µm, whereas at 800 Hz the fibre was less sensitive and had a maximum discharge rate of 350 impulses s-1 over the 75 µm amplitude range available from the stimulator at this frequency. The frequency bandwidth for vibro-stretch responsiveness for this spindle afferent fibre is plotted in Fig. 5B at a series of different amplitudes (5-100 µm). At low amplitudes of the muscle vibration (10 µm), responsiveness is greatest at vibration frequencies below 250 Hz but extends into the higher frequencies with larger vibro-stretch amplitudes.
The equivalent stimulus-response relations and responsiveness profiles constructed in Fig. 6 from the responses of the representative cuneate neurone of Fig. 4 to the same forms of vibrational muscle stretch show that these low amplitude (l&t; 10-20 µm), sinusoidal stretches can also be best detected in the behaviour of the cuneate neurone in the frequency range 150-300 Hz, reflecting the sensitivity and frequency bandwidth of spindle afferent fibres. Although some cuneate neurone thresholds appeared a little higher than those of spindle afferents (Figs 5 and 6), this was almost certainly related to the fact that the afferent fibre thresholds were all obtained from the small indicis proprius muscle (Fig. 5) whereas many of the cuneate neurone thresholds were based on inputs from larger forearm muscles such as the APL and EDL.
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 5. Stimulus-response relation (A) and bandwidth of vibration sensitivity (B) for a representative muscle spindle afferent fibre
A, mean response (impulses s-1 ± S.D.; n = 6) as a function of vibration amplitude for five vibration frequencies (50-800 Hz) applied in 1 s trains to the muscle tendon. In B the mean response is replotted as a function of vibration frequency to show the bandwidth of vibration sensitivity at five different vibration amplitudes.
|
The stimulus-response relations for cuneate neurones (upper graph, Fig. 6) lack the discontinuities (in the form of the broad 1 : 1 response plateau) displayed in the relations of spindle afferents in Fig. 5 but show a continuously graded response, at least up to amplitudes of 100-150 µm. Although the maximum discharge rates of the cuneate neurones (125 s-1 in Fig. 6) were generally lower than those of the peripheral fibres, the graded stimulus-response relations should ensure that individual cuneate neurones can contribute a sensitive signal of the magnitude of these dynamic stretch perturbations.
The response profiles in the lower graph of Fig. 6 show the weaker effects on cuneate responsiveness of both low frequencies (100 Hz) and high frequencies (>300 Hz) of muscle vibration which is consistent with the input to these cuneate neurones arising from the primary spindle endings whose vibration sensitivity appears greatest at frequencies of up to 150-300 Hz (Fig. 5; Brown et al. 1967).
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 6. Stimulus-response relation (A) and bandwidth of vibration sensitivity (B) for a representative cuneothalamic neurone
A, mean response (impulses s-1 ± S.D.; n = 6) as a function of vibration amplitude for seven vibration frequencies (50-800 Hz) applied in 1 s trains to the muscle tendon. B, mean response is replotted as a function of vibration frequency to show the bandwidth of vibration sensitivity at a series of different amplitudes.
|
Entrainment and patterning of cuneate neurone responses to muscle vibration
Previous studies have demonstrated that primary spindle afferent fibres respond with a tightly phase locked pattern of activity to imposed muscle vibration (Brown et al. 1967). However, there has been no quantification of the precision of phase locking in these responses, nor comparison of entrainment measures in the responses of spindle afferents and central neurones in order to evaluate the fidelity with which these temporal features of the afferent responses are conveyed across central synaptic junctions in the pathway for kinaesthetic sensations. The effect of vibration amplitude and frequency on the extent to which impulse activity in both primary and cuneate units was phase locked to the muscle vibration was evaluated by constructing phase scatter graphs and cycle histograms (Figs 7 and 9) from which quantitative measures of phase locking could be derived (Figs 8 and 10). The effect of vibration amplitude on phase locking in a representative spindle afferent and cuneate neurone is illustrated in Fig. 7 for the responses at the approximate 'best' frequency for vibrostretch sensitivity, 150 Hz (Figs 4 and 5; Brown et al. 1967). The phase scatter graphs on the left-hand side in Fig. 7A-F plot, on the ordinate, the time (or phase) of impulse occurrence within the vibration cycle period (6·7 ms for the 150 Hz muscle vibration) against the time from the onset of the 1 s vibration train. For the muscle afferent fibre (Fig. 7A-C), considerable change occurs in the tightness of phase locking as a function of vibration amplitude, as the phase scatter of the impulses occurs over a substantial fraction (around one-half) of the vibration cycle period at the 1 µm amplitude, where the response level was 40 impulses s-1, but tightens to about one-fifth of the cycle period at 10 µm where the response rate was 120 impulses s-1, and becomes very tightly entrained at 50 µm where the fibre was responding at a 1 : 1 level of 150 impulses s-1.
The cycle histograms (see legend) for the fibre, on the right hand side in Fig. 7A-C, plot the cumulative distribution of the impulse events displayed in the corresponding phase scatter graph and show a preferential grouping of impulse occurrences even at the lowest vibration amplitude of 1 µm. From these distributions we derived a quantitative measure of phase locking, the resultant, R (see Methods), which ranges in value from 0 (absence of phase locking) to 1 (maximum phase locking). Even at the lowest amplitude of 1 µm, R for the afferent fibre's response is 0·77 indicating a high degree of phase locking (Lavine, 1971; Bledsoe et al. 1982; Greenstein et al. 1987). However, at 10 and 50 µm the phase locking is extremely tight with R of 0·97 and 1·0, respectively.
The responses to muscle vibration in representative cuneate neurones displayed greater phase scatter (Fig. 7D -F) than was seen in the peripheral fibres. Nevertheless, phase locking was retained in responses to the 150 Hz frequency even at the lowest amplitude (20 µm) shown in Fig. 7D, where R was 0·61, indicative of moderately tight phase locking (Lavine, 1971). As was observed in the peripheral fibres, phase locking became much tighter at higher amplitudes with R of 0·90 and 0·92 at 50 and 200 µm, respectively, in Fig. 7E and F. The phase scatter graphs of Fig. 7D -F indicate that the somewhat lower R in the cuneate responses compared with those for afferent fibres are not, to any substantial extent, attributable to a systematic shift in the phase of response over the course of the 1 s period of the vibration train. The major factor in the lower vector strength (R) in the cuneate responses is presumably related to uncertainty or 'noise' in the synaptic transmission process which introduces a component of variability or 'jitter' into the time of occurrence of the central neurone's responses. A further factor that contributes to some degradation in R for the central neurone, compared with the primary fibres, is the occurrence of pairs of spikes in response to some cycles of the vibration. This is reflected in the spike occurrences lying outside the main band of impulse events in the phase scatter graphs at 50 and 200 µm.
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 7. Effect of vibration amplitude on phase locking of spindle afferent (A-C) and cuneothalamic neurone (D-F) responses to muscle vibration
Phase scatter graphs (left side) and cycle histograms (right side) constructed from responses of a spindle afferent fibre (A-C) and cuneothalamic neurone (D-F) to six repetitions of a 1 s train of vibration applied to the muscle tendon (IPr muscle in A-C; EDL in D-F; 150 Hz vibration, amplitudes indicated in each cycle histogram). Phase scatter graphs display the time of each impulse occurrence (single dot) 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 the phase scatter graphs and display the cumulative distribution of impulse occurrences within the vibration stimulus cycle (abscissa is number of impulse counts in each of the 50 time addresses into which the 6·7 ms ordinate scale was divided).
|
Quantitative comparison of the capacities of spindle afferents and cuneate neurones for entrained, phase locked patterns of response to muscle vibration is illustrated as a function of vibration amplitude in Fig. 8. The data in Fig. 8A for a spindle fibre reveal very tight phase locking (R > 0·8) which, over a range of frequencies, is essentially independent of the vibration amplitude, at least over the range from 10 to 200 µm. In contrast, the representative cuneate neurone (Fig. 8B) attains high levels of phase locking (R > 0·7) only when vibration amplitudes exceed 20 µm and then only at low frequencies of vibration ( 150 Hz).
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 8. Quantitative comparison of the capacities of spindle afferents and cuneate neurones to display phase locked responses to muscle vibration
A, resultant values for responses of a muscle afferent fibre to muscle vibration at frequencies in the range 50-800 Hz and at amplitudes up to 150 µm. B, resultant values for responses of a cuneothalamic neurone to muscle vibration frequencies in the range 50-800 Hz and at amplitudes up to 200 µm.
|
The effect of frequency on the capacity of primary fibres and cuneate neurones to display entrained responses to muscle vibration
The capacity of spindle afferent fibres to retain tight phase locking over a broad bandwidth of vibration frequencies (150-500 Hz) is illustrated by the phase scatter graphs and cycle histograms of Fig. 9A-C. The phase locking is very tight at all three frequencies, as reflected in the vector strengths of almost 1·0, and is just as tight at 500 Hz as at 150 and 300 Hz. In the three examples illustrated where a fixed vibration amplitude of 10 µm was used, the response level was below the 1 : 1 level (50 s-1 at 150 Hz, 100 s-1 at 300 Hz, and 110 s-1 at 500 Hz). At higher amplitudes of muscle vibration the entrainment was even tighter in the primary fibres as reflected in the plot of R in Fig. 10A.
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 9. Effect of vibration frequency on phase locking of spindle afferent (A-C) and cuneate neurone (D-F) responses to muscle vibration
Phase scatter graphs and cycle histograms constructed from responses of representative spindle afferent fibre (A-C) and a cuneothalamic neurone (D-F) to six repetitions of 1 s trains of vibration (10 µm amplitude, A-C; 50 µm, D-F) applied at three frequencies (150 Hz in A and D, 300 Hz in B and E, 500 Hz in C and F). Format as in Fig. 7, except ordinates in A and D, B and E, and C and F, correspond to vibration cycle periods of 6·7 ms, 3·3 ms and 2·0 ms, respectively.
|
In contrast to the capacity of primary spindle afferent fibres to sustain tightly phase locked responses to muscle vibration at frequencies up to and beyond 500 Hz, the cuneate neurones display a marked phase dispersion in their responses as the vibration frequency applied to the muscle increases from 150 Hz to 300 and then 500 Hz (Fig. 9D-F, left-hand graphs). The vector strength fell from a value of R of 0·9 at 150 Hz to only moderate levels of phase locking (Lavine, 1971), with values of R of 0·60 at 300 Hz and 0·51 at 500 Hz. Furthermore, the phase scatter graphs in Fig. 9D-F plotted over the 1 s segment of muscle vibration show that the decline in phase locking at the higher frequencies is not attributable to a progressive phase drift in the response over the 1 s of vibration. The example in Fig. 9D-F also reflects the general finding that cuneate response levels declined at higher frequencies of muscle vibration, as the mean response (obtained by dividing the total counts in the histogram by 6, the number of repetitions of the vibration stimulus) fell from 77 impulses s-1 at 150 Hz to 50 impulses s-1 at 300 Hz and 45 impulses s-1 at 500 Hz.
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 10. Quantitative comparison of the frequency bandwidths over which spindle afferents and cuneate neurones display phase locked responses to muscle vibration
Resultant values plotted as a quantitative measure of phase locking in the responses to muscle vibration of a spindle afferent fibre (A) and cuneothalamic neurone (B) as a function of vibration frequency for a range of different amplitudes.
|
When the vector strengths are plotted as a function of vibration frequency (Fig. 10), the distinction between the spindle afferent fibres and cuneate neurones in their capacity to sustain phase locked responses, in particular at high frequencies (>200 Hz) of muscle vibration, becomes very clear and demonstrates that a marked restriction is imposed at the first synaptic junction on the fidelity and precision of temporal signalling. The decline in the tightness of phase locking for the cuneate neurone in Fig. 10B was representative of the behaviour of the neurones sampled as shown by Fig. 11, which plots, at three vibration amplitudes, the vector strength values as a function of the frequency of muscle vibration for all six cuneate neurones studied in detail over a broad range of frequencies.
|
 |
View larger version
[in this window]
[in a new window] |
|
|
Figure 11. Frequency bandwidths over which cuneate neurones display phase locked responses to muscle vibration
The three graphs plot the values of the resultant for six cuneate neurones as a function of vibration frequency at three different amplitudes of the muscle vibration (20, 50 and 100 µm). The bottom four of the six labelled curves (APL, EDL, ECRB and EDC) were obtained from neurones tested and identified as cuneothalamic projecting neurones.
|
|
DISCUSSION |
The capacity of cuneate neurones to signal static stretch in skeletal muscle
The present results provide the first quantitative evaluation of how effectively the muscle spindle signals about static muscle stretch are transmitted to central neurones in the principal kinaesthetic pathway through the cuneate relay of the dorsal column nuclei. These muscle-related cuneate neurones displayed graded stimulus-response relations, of an approximately linear form, over the range of static stretch examined which covered approximately the middle third of the physiological range (see Results). As most relations (Fig. 2) did not attain a plateau response level over this extent of stretches, the dynamic range of signalling for individual neurones, defined as the range of static muscle lengths over which the neurone displayed graded responsiveness, was almost certainly broader than the 2 mm extent of static stretch used. Furthermore, we may conclude, from the graded stimulus-response relations and the consistency in the response levels at fixed amplitudes of stretch (indicated by the low S.D. values in Fig. 2) that these individual cuneate neurones can contribute discriminative information about static muscle stretch. However, in the central interpretation of neural information about this parameter of skeletal muscle function it is possible that the more important neural signal of muscle length is the total level of impulse traffic in the population of responding neurones as appears to be the case for intensity coding in other sensory continua (e.g. Johnson, 1974).
In earlier studies, there have been only fragmentary observations on cuneate neurone responses to static muscle stretch. For example, although Rosén & Sjölund (1973a) commented that their sample of muscle-activated cuneate and external cuneate neurones responded to static stretch, no stimulus-response data were presented. Furthermore, while there have also been reports of cuneate neurone responses to static inputs associated with limb movement (Amassian, Macy, Waller, Leader & Swift, 1962), there was no attempt made, or grounds available, to identify reliably the receptor source. Hummelsheim & Wiesendanger (1985), in studying muscle-related cuneate neurones in the macaque monkey, used only brief (30-40 ms) 'ramp and hold' stimuli of very low amplitude (80 µm). As the responses they observed lacked any maintained activity, consisting of just a transient response to the onset ramp (Fig. 3 in Hummelsheim & Wiesendanger, 1985), it is possible that the responsible input may have arisen from phasically sensitive Pacinian corpuscles which the authors allowed may have contaminated the input generated by these rather abrupt muscle stretch stimuli. Because of their chosen stimulus parameters it would be inappropriate to conclude from their data that primate cuneate neurones were unable to signal the static components of muscle stretch, in particular as Maendly, Rüegg, Wiesendanger, Wiesendanger, Lagowska & Hess (1981) have observed maintained responses to static stretch in primate thalamic neurones whose inputs are probably mediated via the cuneate nucleus.
The response characteristics displayed by the cuneate neurones in our study, that is, maintained responses to static muscle stretch and high sensitivity to low amplitude muscle vibration, provide strong evidence that the input was derived from primary spindle afferents. Furthermore, as the forearm deafferentation was complete, apart from the nerves to the particular muscles under study, we can be certain that the static responses we observed, lasting 5 s, could not arise from purely dynamically sensitive receptors such as Pacinian corpuscles even if there were some corpuscles associated with the fascia of the muscle under study (Barker, 1967).
In the principal brainstem relay for hindlimb muscle kinaesthetic information, nucleus Z, there were no static responses to maintained stretch of the gastrocnemius-soleus muscle, even with up to 8 mm of extension (Fig. 12 in Magherini et al. 1975). The observations in the cat of Magherini et al. suggest that there may be other pathways over which signals of static muscle length from the hindlimb are conveyed to higher centres for these aspects of hindlimb kinaesthetic sensation. However, at higher levels of the pathway, Hore, Preston, Durkovic & Cheney (1976) have shown, in the baboon, that group I-driven cells in cortical area 3a, where nucleus Z is known to project, do display tonic responses to stretch of single hindlimb muscles.
Responsiveness of cuneate neurones to dynamic, vibro-stretch displacements in skeletal muscle
Sinusoidal vibration was used as a controlled form of dynamic muscle stretch whose parameters (frequency and amplitude) could be precisely quantified. This was chosen in order to examine the extent to which information about dynamic aspects of muscle perturbations are reliably conveyed across the cuneate synaptic junction in the kinaesthetic pathway to the cortex. More complex, and less easily quantified forms of dynamic muscle stretch may arise in many manual tasks carried out by the human hand and forearm.
In the present study, the occurrence of sinusoidal muscle stretch superimposed on a background stretch of the forearm extensor muscles was signalled by an augmented response level in cuneate neurones (Figs 3 and 4) demonstrating that the spindle afferent information about this dynamic perturbation is reliably conveyed across the cuneate synapse in the pathway to cortex. Furthermore, as the sensitivity and bandwidth of vibratory stretch responses in cuneate neurones approximated that of the primary spindle afferent fibres (Figs 5 and 6), there appears to be no substantial limitation imposed at the cuneate synapse on, first, the threshold for detection of dynamic stretch disturbances in forearm muscles, and second, the frequency range of the vibrational stretch disturbances that can be detected. In addition, the graded stimulus-response relations of the cuneate neurones (Fig. 6) indicate that the magnitude of the dynamic stretch changes can be signalled sensitively in the response rate of individual central neurones. Presumably the presence of these graded relations in the cuneate neurones, in contrast to the discontinuities associated with the 1 : 1 response plateau in the stimulus-response relations of spindle afferents, reflects the convergence upon individual cuneate neurones of a number of muscle spindle afferent fibres that vary somewhat in the vibration stretch amplitude required for a 1 : 1 discharge level. Even if there were few converging fibres, there need not be a series of response steps or plateaux in the stimulus-response relations of cuneate neurones, as first, the converging afferents may not be recruited in any strict sequence; second, the converging afferents may vary substantially in the potency of their actions on the central neurone; and third, there may be afferent inhibitory actions, generated with the recruitment of fibres, that become more potent at higher levels of afferent drive leading to a reduction in response gain and expansion of dynamic range for these muscle-related cuneate neurones.
Although the spindle afferent data were obtained from one forearm extensor, the IPr muscle, and cuneate responses came from a variety of different forearm extensors, the observed differences between the afferent fibres and cuneate neurones in their stimulus-response relations are almost certainly attributable to the consequences of central synaptic processing rather than to differences in the identity of the muscle input. It should be emphasized that a similar transformation of stimulus-response relations is apparent across the first central synapse in the tactile pathway if one compares primary vibrotactile afferent fibres, such as those associated with Pacinian corpuscles, with their central target neurones of the cuneate nucleus (Douglas et al. 1978).
Whether the present sample of cuneate neurones activated by primary spindle inputs also received input from secondary spindle endings or Golgi tendon organs was not investigated. However, the responses observed appear to be associated with primary spindle inputs as they displayed high sensitivity to static and dynamic components of muscle stretch and to muscle vibration (Matthews, 1972). The vibration stimuli that were delivered longitudinally to the extensor muscle tendons at amplitudes below 200 µm and at frequencies of 50-800 Hz have little or no effect on the secondary endings or tendon organ receptors and are largely selective for spindle primary endings (Lundberg & Winsbury, 1960; Bianconi & Van der Meulen, 1963; Brown et al. 1967; Stuart et al. 1970).
Fidelity of temporal signalling about muscle vibration in the input to cuneate neurones
The phase scatter and cycle histogram plots for primary spindle afferent fibres in Figs 7 and 9 show that impulse activity in the input to cuneate neurones is locked with great precision to a fixed phase of the muscle vibration waveform. In particular, this is the case once the discharge rate attains the regular 1 : 1 pattern of response that reflects on a cycle-by-cycle basis the periodicity inherent in the vibratory disturbance. In this aspect, the behaviour of the spindle primary afferents resembles that of vibrotactile afferent fibres, in particular those of the rapidly adapting (RA) and Pacinian corpuscle (PC)-related tactile fibre classes (Talbot, Darian-Smith, Kornhuber & Mountcastle, 1968; Ferrington & Rowe, 1980; Ferrington, Hora & Rowe, 1984; Ferrington, Rowe & Tarvin, 1987a). The improvement observed in the tightness of phase locking for both spindle and tactile afferent fibres as the amplitude of the vibratory disturbance increases is presumably in part due to the effective acceleration or velocity component of the stimulus occurring more abruptly at higher amplitudes which will serve to synchronize better the discharge of the fibre. However, another important contribution, once the 1 : 1 response level is reached, must be the stabilization of post-excitatory factors influencing the responsiveness of the fibre. That is, after any spike occurrence, excitability will vary as a function of time because of factors such as relative refractoriness and after-hyperpolarization. The influence of these factors on the phase of spike discharge will vary depending on whether the spikes occur at a relatively constant interval, as takes place in the 1 : 1 pattern of response when intervals approximate the cycle period, or when spikes occur at different intervals, corresponding roughly to multiples of the vibration cycle period; for example, at 1, 2, 3, etc. times the cycle period, as occurs when the fibre discharges at rates below the 1 : 1 level of response.
It is presumably as a consequence of these factors that the phase locking, quantified in terms of the vector strength, R, is very high (1·0) once the afferent fibre attains a 1 : 1 level of response. At this level of response there are therefore two factors that confer great reliability and precision on the afferent fibre signal about the periodicity of the vibrostretch stimulus or, in the case of the tactile system, vibrotactile stimuli. First, the phase of impulse occurrence on individual vibration cycles is 'fixed' (R 
1·0) and second, once the 1 : 1 response level is attained, the impulse sequence provides a precise, metronome-like reflection of the periodicity inherent in the vibratory disturbance.
The capacity of cuneate neurones to retain temporal precision in their signalling of vibrostretch information from skeletal muscle
Phase scatter and cycle histogram distributions for the cuneate neurones (Figs 8 and 10) demonstrate that the tightness of phase locking, at a given vibration frequency, improves as a function of the amplitude of muscle vibration. However, as happened with response levels in the stimulus- response relations for the cuneate neurones (Fig. 6), the increase occurs much more gradually than is the case in the input fibres (Figs 7 and 9). Nevertheless, even when the cuneate neurones reach their maximum level of phase locking (Figs 8B, 10B and 11) the values of R remain lower than those of the spindle afferent fibres and demonstrate that the process of synaptic transmission introduces some temporal degradation in the reliability of neural signalling of the frequency parameter of the vibrostretch muscle stimulus. For example, at lower vibration frequencies, 200 Hz, R for cuneate neurones (Figs 7-11) ranges from moderate to high levels of phase locking (0·5-0·9) compared with the very high values of 1·0 in the spindle afferents (Figs 7 and 9). At high vibration frequencies, 
300 Hz, the discrepancies become much more marked with R in the cuneate neurones declining steeply as a function of vibration frequency (Figs 10 and 11) despite the values for the afferent fibre responses remaining high (> 0·8, even up to vibration frequencies of 800 Hz; Figs 8 and 10). Some of this decline for cuneate neurones, which reflects greater dispersion in the preferred point of discharge on each vibration cycle, may, in part, reflect the convergence on individual neurones of a number of spindle afferent fibres that vary both in their conduction velocity and in their preferred points of discharge to the vibration waveform. However, this may not be the entire explanation because, in the tactile modality, when responses of vibrotactile neurones of dorsal column nuclei are examined in the absence of convergence (that is, when their input is derived selectively from a single Pacinian afferent fibre activated by focal vibration of the associated Pacinian corpuscle) the central neurone also shows poorer phase locking than does the single input fibre and, once vibration frequencies reach 400-500 Hz, there is a virtual disappearance of phase locking in the central neurone (Ferrington et al. 1987a,b; Rowe, 1990).
Nevertheless, it should be emphasized that the remarkable feature of cuneate responses to vibrational stimuli, whether to muscle inputs (present observations) or Pacinian inputs (Douglas et al. 1978; Ferrington et al. 1987a,b; Rowe, 1990), is not the decline and disappearance above 400-500 Hz of phase locking in the central neurones' responses, but rather, the survival of relatively tight phase locking up to these vibration frequencies of 400-500 Hz. This is remarkable because of the following considerations. First, there is considerable convergence of input fibres upon individual dorsal column nuclei neurones (e.g. Gordon & Jukes, 1964; Andersen, Eccles, Oshima & Schmidt, 1964; Tracey, 1980; Ferrington et al. 1987a,b) and second, although the individual tactile and muscle afferent fibres retain precise phase locking in response to vibration frequencies as high as 800-1000 Hz (Figs 8 and 10; and Talbot et al. 1968; Ferrington & Rowe, 1980; Ferrington et al. 1984, 1987a), there is little or no phase coherence at a population response level, at least among vibrotactile Pacinian afferent fibres, at vibration frequencies of 100 Hz (Greenstein et al. 1987; Rowe, 1990). Although the conduction velocities of primary spindle afferent fibres are slightly faster than those of Pacinian afferents it is also improbable that any population coherence exists in the responses of these input fibres to muscle vibration at frequencies 100 Hz. It appears that the behaviour of central tactile neurones, in particular their phase of response to vibration stimuli, is functionally dominated by the powerful synaptic actions of just one or a few of its convergent Pacinian input fibres (Greenstein et al. 1987; Ferrington et al. 1987a,b; Rowe, 1990). The present observation that muscle-related cuneate neurones also retain precisely phase locked responses to vibrational stimuli up to 400-500 Hz (Figs 10 and 11) suggests that these muscle-related neurones share the integrative processing properties of their vibrotactile counterparts of the dorsal column nuclei, and that their responses to muscle vibration may also be functionally dominated by the input from just one or a few of their convergent spindle afferent fibres.
Importance of temporal precision in the transmission and coding of kinaesthetic information
Although there may be no direct kinaesthetic counterpart to the tactile vibratory sense, the use of sinusoidal muscle vibration provides a well-controlled experimental form of dynamic muscle stimulation that allows the signalling and transmission characteristics of the kinaesthetic pathway to be quantified reliably. However, the precision of impulse entrainment and patterning may nevertheless be of physiological importance in kinaesthesia for the subjective recognition of dynamic perturbations of length within skeletal muscle in a variety of circumstances. These may include the abrupt perturbations of joint angle (and therefore muscle length) encountered in the day-to-day movements of all animals, for example, in climbing or locomotion, in particular over uneven or unstable surfaces. Other complex perceptual events may depend upon a concatenation of tactile and kinaesthetic inputs generated by dynamic perturbations affecting both skin receptors and deep receptors in muscles and joints. In order to respond appropriately under these conditions of input, there is a need for accurate signalling of the temporal features of the muscle perturbations, first in the spindle afferent fibres and then in the central target neurones of these afferent fibres. Of course, for these temporal details of the muscle performance to be utilized for kinaesthetic experience requires some preservation of this information, despite the additional temporal dispersion of signals that must arise at thalamic and cortical synaptic junctions in the pathway. It should also be emphasized that in a conscious, behaving animal the response characteristics of central neurones at each level of the system may be modified during normal active movements by descending influences which may serve to gate or otherwise modify response features according to the demands of the movement task. It should also be emphasized that, although we identified only a limited convergence from different muscles onto individual cuneate neurones, there may be more widespread subthreshold convergence which, once again, may be important in the conscious, behaving animal.
Nevertheless, the present observations on muscle-related neurones of the cuneate nucleus show that these neurones have the capacity to contribute to the kinaesthetic acuity needed to ensure fine motor control necessary for complex motor tasks. The cuneate neurones have the capacity to code in a graded and sensitive way the magnitude of static and dynamic changes in muscle length. Furthermore, although these central neurones cannot maintain the same precision and fidelity as the spindle afferent fibres for signalling information about temporal features of dynamic stretch perturbations, they can retain in their entrained, phase locked responses an accurate signal of these vibratory stretch disturbances up to frequencies of 400-500 Hz.
|
REFERENCES |
| Amassian, V. E., Macy, J. Jr, Waller, H. J., Leader, H. S. & Swift, M. (1962). Transformation of afferent activity at the cuneate nucleus. In Information Processing in the Nervous System, ed. Gerard, R. W. & Duyff, J. W., pp. 235-254. Excerpta Medica Foundation, Amsterdam. |
|
| 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, Ciba Foundation Symposium, ed. de Reuck, A. V. S. & Knight, J., pp. 3-15. Churchill, London. |
|
| Bessou, P. & Laporte, Y. (1962). Responses from primary and secondary endings of the same neuromuscular spindle of the tenuissimus muscle of the cat. In Symposium on Muscle Receptors, ed. Barker, D., pp. 105-119. Hong Kong University Press, Hong Kong. |
|
| 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] |
| 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. |
|
| Darian-Smith, I., Phillips, G. & Ryan, R. D. (1963). Functional organization in the trigeminal main sensory and rostral spinal nuclei of the cat. The Journal of Physiology 168, 129-146. |
|
| 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] |
| 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. (1980). Functional capacities of tactile afferent fibres in neonatal kittens. The Journal of Physiology 307, 335-353 |
[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. |
|
| 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] |
| Hand, P. J. & Van Winkle, T. (1977). The efferent connections of the feline nucleus cuneatus. Journal of Comparative Neurology 171, 83-110 |
[Medline] |
| Hore, J., Preston, J. B., Durkovic, R. G. & Cheney, P. D. (1976). Responses of cortical neurons (areas 3a and 4) to ramp stretch of hindlimb muscles in the baboon. Journal of Neurophysiology 39, 484-500 |
[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] |
| 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. E. S. Livingstone Ltd, Edinburgh. |
|
| Johnson, K. O. (1974). Reconstruction of population response to a vibratory stimulus in quickly adapting mechanoreceptive afferent fiber population innervating glabrous skin of the monkey. Journal of Neurophysiology 37, 48-72 |
[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. |
|
| Mackie, P. D., Morley, J. W., Bahramali, H., Zhang, H. Q., Murray, G. M. & Rowe, M. J. (1993). Phase relations of muscle afferent and cuneate relay cell responses to muscle vibration in the cat. Proceedings of the Australian Neuroscience Society 4, 82. |
|
| Mackie, P. D., Morley, J. W., Murray, G. M., Zhang, H. Q., Bahramali, H. & Rowe, M. J. (1992). Response characteristics of group I muscle afferent relay cells in the main cuneate nucleus of the cat. Proceedings of the Australian Physiological and Pharmacological Society 23, 137P. |
|
| 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] |
| Mackie, P. D. & Rowe, M. J. (1998). Central projection of proprioceptive information from the wrist joint via a forearm 'muscle' nerve in the cat. The Journal of Physiology 510, 261-267 |
[Abstract/Full Text] |
| Maendly, R., Rüegg, D. G., Wiesendanger, M., Wiesendanger, R., Lagowska, J. & Hess, B. (1981). Thalamic relay for group I muscle afferents of forelimb nerves in the monkey. Journal of Neurophysiology 46, 901-917 |
[Medline] |
| Magherini, P. C., Pompeiano, O. & Seguin, J. J. (1975). Responses of nucleus Z neurons to vibration of hindlimb extensor muscles in the decerebrate cat. Archives Italiennes de Biologie 113, 150-187 |
[Medline] |
| Mardia, K. V. (1972). Statistics of Directional Data. Academic Press, London. |
|
| Matthews, B. H. C. (1933). Nerve endings in mammalian muscle. The Journal of Physiology 78, 1-53. |
|
| Matthews, P. B. C. (1972). Mammalian Muscle Receptors and their Central Actions. Monographs of the Physiological Society No. 23, pp. 630. Edward Arnold, 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] |
| Rosén, I. (1969). Localization in caudal brain stem and cervical spinal cord of neurones activated from forelimb group I afferents in the cat. Brain Research 16, 55-71 |
[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. |
|
| Stuart, D. G., Mosher, C. G., Gerlach, R. L. & Reinking, R. M. (1970). Selective activation of Ia afferents by transient muscle stretch. Experimental Brain Research 10, 477-487 |
[Medline] |
| Talbot, W. H., Darian-Smith, I., Kornhuber, H. H. & Mountcastle, V. B. (1968). The sense of flutter-vibration: comparison of the human capacity with response patterns of mechano-receptive afferents from the monkey hand. Journal of Neurophysiology 31, 301-334 |
[Medline] |
| 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] |
| Wiesendanger, M. & Hummelsheim, H. (1985). The medullary relay of activity from low-threshold proprioceptive afferents of forelimbs in the monkey. In Development, Organization and Processing in Somatosensory Pathways, ed. Rowe, M. J. & Willis, W. D., pp. 175-181. Liss, New York. |
|
| Zar, J. H. (1984). Biostatistical Analysis, 2nd edn. Prentice-Hall, Englewood Cliffs, NJ, USA. |
|
Acknowledgements
The authors acknowledge the technical assistance of H. Bahramali, C. Riordan, F. Spicer, P. Farrell and T. Ingall. 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 and 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:
|
|
|
|
 
V. Sahai, D. A. Mahns, N. M. Perkins, L. Robinson, and M. J. Rowe
Vibrotactile Coding Capacities of Spinocervical Tract Neurons in the Cat
J Neurophysiol,
March 1, 2006;
95(3):
1465 - 1477.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
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]
|
|
|
|
|
|
|
H. Q. Zhang, G. M. Murray, G. T. Coleman, A. B. Turman, S. P. Zhang, and M. J. Rowe
Functional Characteristics of the Parallel SI- and SII-Projecting Neurons of the Thalamic Ventral Posterior Nucleus in the Marmoset
J Neurophysiol,
May 1, 2001;
85(5):
1805 - 1822.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Introduction
