Kinaestheticsignalsandmusclecontradion S. C. G a n d e v i a , D. I. M c C I o s k e y and D. Burke
S. C Gandevia, D. I. McC/oskeyand D. Burkeare at the Prince of Wales Medical Research Institute and Faculty of Medicine, University of New 5outh Wales, Sydney 2033, Australia.
Signals generated both peripherally and centrally contribute to the group of sensations termed kinaesthesia. Many experiments report sensations of position and movement under passive relaxed conditions without muscle contraction. However, kinaesthetic acuity is probably of greater functional value when subjects are active rather than passive and, accordingly, movement detection is markedly improved by muscular contraction. One mechanism contributing to this enhancement is likely to involve muscle spindle volleys. When identical microstimulation techniques are applied to skin, joint and muscle spindle endings innervating the hand, some cutaneous afferents and some joint afferents elicit a sensation, but activation of certain other cutaneous afferents and muscle spindle afferents rarely does. Activity in more than one muscle spindle afferent may be required for kinaesthetic sensations, whereas some single cutaneous and joint afferents may have a more 'secure' central projection. Kinaesthesia encompasses three main sensations: the sensations of position and movement of joints; the sensations of force, effort and heaviness associated with muscular contractions; and the sensations of the perceived timing of muscular contractions. Theoretically, each of these types of sensation could rely solely upon the discharge of receptors in the skin, joint and muscles. However, there is evidence that signals related to centrally generated motor commands play a role in each sensation group (for reviews see Refs 1-4). This is not surprising given that these sensations are rarely generated in the absence of muscular contraction. Thus, it is anomalous that the sensation of joint rotation is routinely assessed in the clinic when joints are moved passively. Despite this anomaly, this brief review describes the mechanisms involved in detection of passively applied movements, and then goes on to address some unresolved questions about the effect of muscle contraction on kinaesthesia, and, based on results from intrafascicular microstimulation, describes the ability of different afferent classes to signal that a joint has moved. Data from several psychophysical approaches have pointed to a major role for intramuscular receptors, particularly muscle spindle endings, in generating the sensation of passive joint movement L3. (1) Transverse vibration of muscles and tendons using stimulus parameters that favour activation of muscle spindle endings (e.g. see Ref. 5) produces illusions in which joints rotate and occupy abnormal positions. Such illusions, consistent with perceived muscle lengthening, occur when the vibration frequency is above about 20 Hz, and even when the amplitude of vibration is as low as 20 ~m applied longitudinally to a surgically exposed tendon s. Lowamplitude vibration preferentially excites primary muscle spindle afferents. Electrical stimulation of the ulnar nerve at intensities below threshold for activation of motor axons also produces similar illusions of joint movement, as well as short-latency cortical
62
potentials from the contralateral somatosensory cortex 7. Illusions elicited below motor threshold probably involve the large diameter primary muscle spindle af[erents. (2) The ability to detect passive movements is not abolished by anaesthesia of skin and joint afferents. Furthermore, detection of passive movements is impaired when the joint is positioned so that the muscles normally acting on it cannot produce joint movement s. For the distal interphalangeal joint of the finger, this deficit is greater when more than one muscle group (i.e. agonist and antagonist) is eliminated 9. The residual ability to detect the applied movements in the absence of a muscle afferent contribution is partly due to a contribution from joint receptors (Refs 10, 11; but see Ref. 12). (3) The detection thresholds for passive movements applied at various velocities to different joints of the upper limb can be compared in terms of the relative change in muscle length, linear distance moved at the extremity, angular velocity about the moved joint, or the estimated distortion of the joint. The best description of proprioceptive acuity across several joints occurs when data are expressed as the relative change in muscle length 13. This suggests that signals related to muscle length and velocity are important for detection of passive movements applied at velocities commonly encountered during deliberate accurate movements. During a contraction (i.e. a shortening) of an agonist, the spindle-derived signals about length coming from the (passively) lengthened antagonist will continue, irrespective of complex changes in signals from the agonlst muscle receptors due to fusimotor activity and mechanical unloading of the spindles. It is possible to detect changes in position made at velocities below those encountered during usual voluntary movement - such velocities, usually less than one degree per minute (° per min), produce no sensation that the joint is moving. Once the rotation has exceeded 2-5 °, it is detected as a change in position about the joint; such detection has been used by Clark and colleagues to test 'static' or absolute position sense 14. This sensation is certainly present for a variety of joints: the ankle, knee and the joints within the hand. An initial report that it was absent for the proximal interphalangeal joint l~ has not been confirmed 16. Even rotation of the neck and trunk has revealed a capacity to detect changes in position without a sense of movement 17'18. For joints in the leg, detection of extremely slow displacements appears to be dependent on input from muscle afferents 14. By contrast, in the distal joint of the finger, detection of the static position is not impaired when the agonist and antagonist muscles are effectively disengaged from the jointlS: this suggests that for the finger joint, adequate position information is available without recourse to inputs from muscle receptors, while this is not true for large joints in the lower limb. However, when more rapid movements are applied (within the range of natural movements),
© 1992,ElsevierSciencePublishersLtd.(UK) 0166-2236/92/$05.00
TINS, Vol. 15, No. 2, 1992
muscle disengagement impairs the sensation of movement of the finger8 - suggesting that receptors within muscle are especially important in providing signals of joint movement.
Movement detection during muscular contraction How acute are sensations of limb position and movement when tested under more natural conditions; that is, when muscles are contracting? Anaesthetic block of the digital nerves impairs the ability to detect movements applied to the big toe 19, but it was noted that muscle contraction restored kinaesthetic acuity. This observation has been confirmed for the distal interphalangeal joint (Fig. 1A). The digital nerves of the finger were blocked by local anaesthesia and the hand positioned so that only contraction of one muscle could exert torque at the joint. The ability to detect the direction of applied movements of 10° towards either flexion or extension was dramatically enhanced when the movements were applied while the subject continuously exerted a weak voluntary flexion. Corroborative data have been obtained for the elbow joint (Fig. 1B): angular movements o f - 0 . 2 ° were detected at velocities of 0.1 ° per s during contraction of the elbow flexor muscles 2°, whereas the threshold for detection of passive movement at this joint was ten times greater la than when the muscles were active. A recent study confirmed this result for a broader range of angular velocities (Taylor, J. L. and McCloskey, D. I., unpublished observations) and showed that thresholds for detection during contraction diminish further at higher angular velocities. With contraction of elbow flexors, the reduction in threshold occurred for both flexion and extension movements - although thresholds were significantly lower for extension (i.e. stretch of the contracting muscles) than for flexion movements. The low detection threshold during active tasks is highlighted by the observation that distances of 75 Izm between the thumb and index finger can be detected reliably in tests of 'thickness' discrimination2~: indirect evidence points to a role for muscle afferents in this detection, but remote cutaneous afferents within the territory of the superficial radial nerve might also contribute 22. Several mechanisms could underlie the reduction in detection thresholds during active movement. The signals from joint and cutaneous afferents may be altered, but it is unlikely that this is the full explanation. Some joint receptors discharge upon movement through the mid-range of joint motion23-25; however, increases in muscle force do not enhance the mid-range discharge of joint receptors at the cat elbow ~6. Slowly and rapidly adapting cutaneous afferents discharge in response to active and passive movements 22'23'27, but their discharge is unlikely to encode active movements more sensitively than passive ones. Furthermore, the enhancement of movement detection of the distal joint of the finger occurs even when the input from cutaneous and joint afferents has been eliminated (Fig. 1A). By exclusion, the responsible mechanism involves intramuscular receptors and their central projections. The discharge of muscle spindle endings increases during voluntary isometric contraction: the number of spindle endings with a background discharge inTINS, Vol. 15, No. 2, 1992
A I00-
•
•
•
•
• "O
•
•
[]
•
80-
O
O
0
cO
¢-
60-
O
._o "~ 40E E ~. e00
O
withtension: l relaxed
0- O0 0
0
O0 I
5
I
0
I
I
I
10 15 20 Angular velocity (degrees per s)
25
B
2.5-
E
2.0
t-
"6 1.5 0 t~
relaxed O
~ 1.0
[with tension •
I
I
"O
E 0.5 E •
0.0 0.01
o
o11
1'0 Angular velocity (degrees per s)
0
0
100
Fig. 1. (A)
The ability of one subject to detect the direction of 10° displacements of the distal interphalangeal joint of the middle finger at various angular velocities. The test finger was anaesthetized by digital nerve block, thus removing joint and cutaneous contributions to detection. In one series of experiments, the subject exerted no tension with the anaesthetized finger (circles), but in the other, active tension was exerted by contraction of the long flexor muscles (squares). When the relaxed long flexor muscle provided the only source of kinaesthetic information, performance was poor; only 80% of movement directions were detected at 20 ° per s. When the flexors voluntarily exerted tension, performance improved: the subject detected the direction of 90-100% of all movements, even at a velocity of 1° per s, and the performance was within the range observed when joint and cutaneous afferents were also active. (B) Detection thresholds for movements (in degrees) of the elbow over a range of angular velocities. Thresholds for the relaxed arm were set at a movement detection level of 70% (Ref. 13) and are shown as open circles (mean +_ SEM). A threshold for detection with active flexion of the e l b o w 2° is marked by a filled square. [Part (A) taken, with permission, from Ref. 8. Part (B) taken, with permission, from Ref. 13.]
creases and the discharge rate of most endings increases in proportion to muscle force 28-3°. This reflects an increase in fusimotor (or skeletofusimotor) outflow. If this discharge was due simply to static 63
thumb
Elicited sensations 2 out of 18 SA II afferents , , q ~
8 out of 11 joint afferents
1 out of 16 muscle spindle afferents
record
stimulate
microelectrode intrafascicular recording site
Fig. 2. Experimental arrangement for recording and stimulating intrafascicular sites that are associated with single afferents originating in the skin, joint or muscles of the hand. As indicated in the expanded diagrammatic view of a cutaneous fascicle of the median nerve, the tip of the microelectrode is manipulated until it is near a single axon. The behaviour of the afferent during natural stimulation of its receptive field is recorded before and after microstimulation at the recording site. Subjects receive single stimufi and trains of stimuli at different intensities and are required to indicate their response using a questionnaire expanded from that used by Vallbo and colleagues 41. Particular attention was paid to sensations of joint displacement or movement. Recordings were made in the median and ulnar nerves. The frequency with which stimulus trains elicited sensations are shown. (Modified from Ref. 42.)
fusimotor drive, then the contraction-induced improvement in movement detection would suggest that the CNS is influenced by the number of active spindle endings, since static fusimotor drive reduces the sensitivity of primary spindle afferents to small stretches al. For muscles with long tendons, muscle contraction is important in improving the mechanical linkage between the joint and muscle fibres32. The discharge from Golgi tendon organs also increases during voluntary contraction. Given the stiffness of actively contracting muscle and the high sensitivity of tendon organs to such a forcea3, the discharge from these organs will follow changes in muscle length34 and could thus contribute to detection of applied displacements. The history of previous muscle con64
traction is important in determining the sensitivity of muscle receptors - muscle spindle endings in a relaxed muscle discharge more in response to a standard lengthening when the muscle has just previously contracted to a short length. This induces predictable errors in position sense 35. During voluntary contraction, changes in central transmission of afferent input may modulate the detection of kinaesthetic signals. However, many studies have confirmed the seemingly paradoxical decrease in detectability and perceived intensity of cutaneous stimuli during voluntary contraction, particularly under non-isometric conditions (e.g. Refs 36, 37). The arguments above suggest that muscle afferent input during contraction is augmented. It is also possible that there is enhanced central transmission of muscle afferent volleys. If this is so, then muscle afferent signals are being processed differently to signals from specialized cutaneous afferents. Irrespective of the mechanisms involved, perception of cutaneous inputs but not that of muscle afferent inputs appears to be functionally gated during voluntary contractions.
Microstimulation of single afferents Not only has microneurography facilitated the recording of the behaviour of single afferents innervating skin, joint and muscle in humans, but it can also be used to stimulate axons so that their central projections and the perceptual effects of these projections can be studied. Experiments involving the stimulation of slowly and rapidly adapting cutaneous afferents innervating the human hand have resulted in subjective reports that are 'remarkably congruent in studies performed independently by different research groups in different laboratories' (Ref. 38; see also Refs 39--42), while criticisms of the technique have been largely refuted38. Specialized cutaneous innervation of the glabrous skin is divided into rapidly adapting receptors in Meissner's corpuscles (RA receptors), in Pacinian/ paciniform corpuscles (PC receptors), in slowly adapting receptors in Merkel cell neurite complexes (SA I receptors), and in Ruffmi endings (SA II receptors). Compared with SA I afferents, the SA II afferents have larger receptive fields, greater sensitivity to lateral skin stretch and often a background discharge. Stimuli are delivered through the insulated microelectrode at the intrafascicular site, from which the discharge of a single afferent is recorded. The uniform findings are as follows. (1) Low-level voltage or current pulses elicit a focal sensation of superficial flutter or vibration when applied at the site of recording of rapidly adapting afferents, and a sensation of sustained pressure when applied to SA I sites that include the distal region of the digits. (2) Microstimulation of S A I I afferents innervating the glabrous skin of the hand usually elicits no sensation. The technique has subsequently been applied to assess whether single muscle and joint afferents convey information about joint movement (Fig. 2)42. The study corroborated findings previously reported for cutaneous afferents, but, in addition, microstimulation with controlled electrical stimuli elicited sensations from 8 out of 11 intrafascicular sites from which the discharge of single slowly adapting joint TINS, Vol. 15, No. 2, 1992
afferents was recorded. Half the sensations elicited by single stimuli or trains of stimuli were of focal pressure at or very close to the receptive field. The remainder were sensations of joint movement - small illusory displacements (~10 °) that increased with stimulus frequency and that were not necessarily perceived to occur about the extremes of joint rotation where the afferents responded best. Sometimes the elicited illusions corresponded exactly with the complex multiaxial responsiveness of the afferent. In the same experimental sessions, microstimulation was performed (at intensities identical to those for cutaneous and joint afferents) at intrafascicular sites from which the discharge of 16 identified muscle spindle endings in intrinsic muscles of the hand was recorded. At only one spindle site did microstimulation yield a sensation when the stimulus intensity was below that necessary to activate nearby motor axons and so produce a twitch of the innervated muscle. The one positive response was a sensation normally associated with muscle lengthening. In agreement with other results (e.g. see Refs 38, 40, 41), stimulation of most S A I I afferent sites (15 out of 18) elicited no consistent sensations. Exceptions were two SA II afferents innervating the skin near the nallbed, which, when stimulated, elicited perceived flexion at the distal interphalangeal joint - an illusion consistent with their responses to passive movement.
Implications Although digital joint afferents have a limited capacity to encode movement and position 23, some must have a 'secure' projection to consciousness. Also, the discharge of single muscle spindle afferents appears to be insufficient to elicit the sensations attributed to the discharge of a population of them. Reliance on the population input from muscle spindle afferents would prevent the sensation being 'distorted' by any 'irregularities' introduced as a result of individual spindles responding to local disturbances within the parent muscle. Indeed, psychophysical studies suggest that the number and discharge frequency of spindle afferents influence the perceived velocity of illusory movements (e. g. Refs 5, 7). Thus, 'spatial coding' differs for receptors involved in control of the fingers: single cutaneous afferents (RA and SA I) can provide focal information about the skin surface, but the processing of movement sensation in the CNS relies on the population input from muscle spindle endings. Single stimuli applied to joint and some cutaneous afferents (RA and SA I) can elicit sensations (a finding that applies to muscle spindle afferents), but only when a sufficient number of them are activated (see also Ref. 7). S A I I afferents can encode local and remote joint movement 22, but individual afferents usually fail to arouse perception. It is concluded that: (1) a coherent discharge in any mechanoreceptor channel (cutaneous, muscle or joint) will not be entirely ignored by the CNS when computing limb position and movement sensations; (2) all afferent species can encode specific aspects of movement and each can elicit a movement percept (although relatively greater spatial summation may be required to elicit a sensation for muscle spindle afferents and S A I I afferents than for RA and SA I TINS, Vol. 15, No. 2, 1992
afferents); (3) movement enhances kinaesthetic acuity Acknowledgements via mechanisms that have not yet been fully defined in We are grateful to the humans; and (4) the contraction-associated improve- National Health and ment in proprioceptive acuity reinforces the major Medical Research Council of Australia role of muscle receptors in kinaesthesia. Selected references 1 McCIoskey, D. I. (1978) Physiol. Rev. 58, 763-820 2 McCIoskey, D. I. (1981) in Handbook of Physiology (The Nervous System, Vol. I1: Motor Control) (Brooks, V. B., ed.), pp. 1415-1447, American Physiological Society 3 Matthews, P. B. C. (1988) Can. J. Physiol. Pharmacol. 66, 430-438 4 Gandevia, S. C. (1987) Trends NeuroscL 10, 81-85 5 Roll, J. P. and Vedel, J. P. (1982) Exp. Brain Res. 47, 177-190 6 McCIoskey, D. I., Cross, M. J., Honner, R. and Potter, E. K. (1983) Brain 106, 21-37 7 Gandevia, S. C. (1985) Brain 108, 965-981 8 Gandevia, S. C. and McCIoskey, D. I. (1976) J. Physiol. 260, 387-407 9 Gandevia, S. C., Hall, L. A., McCIoskey, D. I. and Potter, E. K. (1983) J. Physiol. 335, 507-517 10 Ferrell, W. R., Gandevia, S. C. and McCIoskey, D. I. (1987) J. Physiol. 386, 63-71 11 Ferrell, W. R. and Smith, A. (1988) J. Physiol. 399, 49-61 12 Clark, F. J., Grigg, P. and Chapin, J. W. (1989) J. Neurophysiol. 61, 186-193 13 Hall, L. A. and McCIoskey, D. I. (1983) J. Physiol. 335, 519-533 14 Clark, F. J., Burgess, R. C., Chapin, J. W. and Lipscomb, W. T. (1985) J. Neurophysiol. 54, 1529-1540 15 Clark, F. J., Burgess, R. C. and Chapin, J. W. (1986) Brain 109, 1195-1208 16 Taylor, J. L. and McCIoskey, D. I. (1990) Brain 113,157-166 17 Taylor, J. L. and McCIoskey, D. I. (1988) Brain Res. 70, 351-360 18 Taylor J. L. and McCIoskey, D. I. (1990) Exp. Brain Res. 81, 413-416 19 Brown, K., Lee, J. and Ring, P. A. (1954) J. Physiol. 126, 448-458 20 Colebatch, J. G. and McCIoskey, D. I. (1987) J. Physiol. 386, 247-261 21 John, K. T., Goodwin, A. W. and Darian-Srnith, I. (1989) Exp. Brain Res. 78, 62-68 22 Edin, B. B. and Abbs, J. H. (1991) J. NeurophysioL 65, 657-670 23 Burke, D., Gandevia, S. C. and Macefield, G. (1988) J. Physiol. 402, 347-361 24 Ferrell, W. R. (1980) J. Physiol. 299, 85-99 25 Proske, U., Schaible, H-G. and Schmidt, R. F. (1988) Exp. Brain Res. 72, 219-224 26 Baxendale, R. H. and Ferrell, W. R. (1983) Brain Res. 261, 195-203 27 Hulliger, M., Nordh, E., Thelin, A-E. and Vallbo, A. B. (1979) J. Physiol. 291,233-249 28 Burke, D. (1981) Int. Rev. PhysioL 25, 91-126 29 Edin, B. B. and Vallbo, /k. B. (1990) J. Neurophysiol. 63, 1307-1313 30 Vallbo, A. B. (1974) Acta PhysioL Scand. 90, 319-336 31 Hulliger, M., Matthews, P. B. C. and Noth, J. (1977) J. PhysioL 267, 811-838 32 Rack, P. M. H. and Ross, H. F. (1984) J. Physiol. 351,99-110 33 Proske, U. (1981) Int. Rev. Physiol. 25, 127-171 34 Stauffer, E. K. and Stephens, J. A. (1977) J. Neurophysiol. 40, 681-691 35 Gregory, J. E., Morgan, D. L, and Proske, U. (1988) J. Neurophysiol. 59, 1220-1230 36 Angel, R. W. and Malenka, R. C. (1982) Exp. Neurol. 77, 266-274 37 Milne, R. J., Aniss, A. M., Kay, N. E. and Gandevia, S. C. (1988) Exp. Brain Res. 70, 569-576 38 TorebjSrk, E., Vallbo, ~,. B. and Ochoa, J. L. (1987) Brain 110, 1509-1529 39 Konietzny, F., Perl, E. R., Trevino, D., Light, A. and Hensel, H. (1981) Exp. Brain Res. 42, 219-222 40 Ochoa, J. and TorebjSrk, E. (1983) J. PhysioL 342, 633-654 41 Vallbo, A. B., Olsson, K. ]k., Westberg, K-G. and Clark, F. J. (1984) Brain 107, 727-749 42 Macefield, G., Gandevia, S. C. and Burke, D. (1990) J. Physiol. 429, 113-129
for continuing financialsupport. We acknowledge the contribution of our collaborators to the work cited here.
65
S. C Gandevia, D. I. McC/oskeyand D. Burkeare at the Prince of Wales Medical Research Institute and Faculty of Medicine, University of New 5outh Wales, Sydney 2033, Australia.
Signals generated both peripherally and centrally contribute to the group of sensations termed kinaesthesia. Many experiments report sensations of position and movement under passive relaxed conditions without muscle contraction. However, kinaesthetic acuity is probably of greater functional value when subjects are active rather than passive and, accordingly, movement detection is markedly improved by muscular contraction. One mechanism contributing to this enhancement is likely to involve muscle spindle volleys. When identical microstimulation techniques are applied to skin, joint and muscle spindle endings innervating the hand, some cutaneous afferents and some joint afferents elicit a sensation, but activation of certain other cutaneous afferents and muscle spindle afferents rarely does. Activity in more than one muscle spindle afferent may be required for kinaesthetic sensations, whereas some single cutaneous and joint afferents may have a more 'secure' central projection. Kinaesthesia encompasses three main sensations: the sensations of position and movement of joints; the sensations of force, effort and heaviness associated with muscular contractions; and the sensations of the perceived timing of muscular contractions. Theoretically, each of these types of sensation could rely solely upon the discharge of receptors in the skin, joint and muscles. However, there is evidence that signals related to centrally generated motor commands play a role in each sensation group (for reviews see Refs 1-4). This is not surprising given that these sensations are rarely generated in the absence of muscular contraction. Thus, it is anomalous that the sensation of joint rotation is routinely assessed in the clinic when joints are moved passively. Despite this anomaly, this brief review describes the mechanisms involved in detection of passively applied movements, and then goes on to address some unresolved questions about the effect of muscle contraction on kinaesthesia, and, based on results from intrafascicular microstimulation, describes the ability of different afferent classes to signal that a joint has moved. Data from several psychophysical approaches have pointed to a major role for intramuscular receptors, particularly muscle spindle endings, in generating the sensation of passive joint movement L3. (1) Transverse vibration of muscles and tendons using stimulus parameters that favour activation of muscle spindle endings (e.g. see Ref. 5) produces illusions in which joints rotate and occupy abnormal positions. Such illusions, consistent with perceived muscle lengthening, occur when the vibration frequency is above about 20 Hz, and even when the amplitude of vibration is as low as 20 ~m applied longitudinally to a surgically exposed tendon s. Lowamplitude vibration preferentially excites primary muscle spindle afferents. Electrical stimulation of the ulnar nerve at intensities below threshold for activation of motor axons also produces similar illusions of joint movement, as well as short-latency cortical
62
potentials from the contralateral somatosensory cortex 7. Illusions elicited below motor threshold probably involve the large diameter primary muscle spindle af[erents. (2) The ability to detect passive movements is not abolished by anaesthesia of skin and joint afferents. Furthermore, detection of passive movements is impaired when the joint is positioned so that the muscles normally acting on it cannot produce joint movement s. For the distal interphalangeal joint of the finger, this deficit is greater when more than one muscle group (i.e. agonist and antagonist) is eliminated 9. The residual ability to detect the applied movements in the absence of a muscle afferent contribution is partly due to a contribution from joint receptors (Refs 10, 11; but see Ref. 12). (3) The detection thresholds for passive movements applied at various velocities to different joints of the upper limb can be compared in terms of the relative change in muscle length, linear distance moved at the extremity, angular velocity about the moved joint, or the estimated distortion of the joint. The best description of proprioceptive acuity across several joints occurs when data are expressed as the relative change in muscle length 13. This suggests that signals related to muscle length and velocity are important for detection of passive movements applied at velocities commonly encountered during deliberate accurate movements. During a contraction (i.e. a shortening) of an agonist, the spindle-derived signals about length coming from the (passively) lengthened antagonist will continue, irrespective of complex changes in signals from the agonlst muscle receptors due to fusimotor activity and mechanical unloading of the spindles. It is possible to detect changes in position made at velocities below those encountered during usual voluntary movement - such velocities, usually less than one degree per minute (° per min), produce no sensation that the joint is moving. Once the rotation has exceeded 2-5 °, it is detected as a change in position about the joint; such detection has been used by Clark and colleagues to test 'static' or absolute position sense 14. This sensation is certainly present for a variety of joints: the ankle, knee and the joints within the hand. An initial report that it was absent for the proximal interphalangeal joint l~ has not been confirmed 16. Even rotation of the neck and trunk has revealed a capacity to detect changes in position without a sense of movement 17'18. For joints in the leg, detection of extremely slow displacements appears to be dependent on input from muscle afferents 14. By contrast, in the distal joint of the finger, detection of the static position is not impaired when the agonist and antagonist muscles are effectively disengaged from the jointlS: this suggests that for the finger joint, adequate position information is available without recourse to inputs from muscle receptors, while this is not true for large joints in the lower limb. However, when more rapid movements are applied (within the range of natural movements),
© 1992,ElsevierSciencePublishersLtd.(UK) 0166-2236/92/$05.00
TINS, Vol. 15, No. 2, 1992
muscle disengagement impairs the sensation of movement of the finger8 - suggesting that receptors within muscle are especially important in providing signals of joint movement.
Movement detection during muscular contraction How acute are sensations of limb position and movement when tested under more natural conditions; that is, when muscles are contracting? Anaesthetic block of the digital nerves impairs the ability to detect movements applied to the big toe 19, but it was noted that muscle contraction restored kinaesthetic acuity. This observation has been confirmed for the distal interphalangeal joint (Fig. 1A). The digital nerves of the finger were blocked by local anaesthesia and the hand positioned so that only contraction of one muscle could exert torque at the joint. The ability to detect the direction of applied movements of 10° towards either flexion or extension was dramatically enhanced when the movements were applied while the subject continuously exerted a weak voluntary flexion. Corroborative data have been obtained for the elbow joint (Fig. 1B): angular movements o f - 0 . 2 ° were detected at velocities of 0.1 ° per s during contraction of the elbow flexor muscles 2°, whereas the threshold for detection of passive movement at this joint was ten times greater la than when the muscles were active. A recent study confirmed this result for a broader range of angular velocities (Taylor, J. L. and McCloskey, D. I., unpublished observations) and showed that thresholds for detection during contraction diminish further at higher angular velocities. With contraction of elbow flexors, the reduction in threshold occurred for both flexion and extension movements - although thresholds were significantly lower for extension (i.e. stretch of the contracting muscles) than for flexion movements. The low detection threshold during active tasks is highlighted by the observation that distances of 75 Izm between the thumb and index finger can be detected reliably in tests of 'thickness' discrimination2~: indirect evidence points to a role for muscle afferents in this detection, but remote cutaneous afferents within the territory of the superficial radial nerve might also contribute 22. Several mechanisms could underlie the reduction in detection thresholds during active movement. The signals from joint and cutaneous afferents may be altered, but it is unlikely that this is the full explanation. Some joint receptors discharge upon movement through the mid-range of joint motion23-25; however, increases in muscle force do not enhance the mid-range discharge of joint receptors at the cat elbow ~6. Slowly and rapidly adapting cutaneous afferents discharge in response to active and passive movements 22'23'27, but their discharge is unlikely to encode active movements more sensitively than passive ones. Furthermore, the enhancement of movement detection of the distal joint of the finger occurs even when the input from cutaneous and joint afferents has been eliminated (Fig. 1A). By exclusion, the responsible mechanism involves intramuscular receptors and their central projections. The discharge of muscle spindle endings increases during voluntary isometric contraction: the number of spindle endings with a background discharge inTINS, Vol. 15, No. 2, 1992
A I00-
•
•
•
•
• "O
•
•
[]
•
80-
O
O
0
cO
¢-
60-
O
._o "~ 40E E ~. e00
O
withtension: l relaxed
0- O0 0
0
O0 I
5
I
0
I
I
I
10 15 20 Angular velocity (degrees per s)
25
B
2.5-
E
2.0
t-
"6 1.5 0 t~
relaxed O
~ 1.0
[with tension •
I
I
"O
E 0.5 E •
0.0 0.01
o
o11
1'0 Angular velocity (degrees per s)
0
0
100
Fig. 1. (A)
The ability of one subject to detect the direction of 10° displacements of the distal interphalangeal joint of the middle finger at various angular velocities. The test finger was anaesthetized by digital nerve block, thus removing joint and cutaneous contributions to detection. In one series of experiments, the subject exerted no tension with the anaesthetized finger (circles), but in the other, active tension was exerted by contraction of the long flexor muscles (squares). When the relaxed long flexor muscle provided the only source of kinaesthetic information, performance was poor; only 80% of movement directions were detected at 20 ° per s. When the flexors voluntarily exerted tension, performance improved: the subject detected the direction of 90-100% of all movements, even at a velocity of 1° per s, and the performance was within the range observed when joint and cutaneous afferents were also active. (B) Detection thresholds for movements (in degrees) of the elbow over a range of angular velocities. Thresholds for the relaxed arm were set at a movement detection level of 70% (Ref. 13) and are shown as open circles (mean +_ SEM). A threshold for detection with active flexion of the e l b o w 2° is marked by a filled square. [Part (A) taken, with permission, from Ref. 8. Part (B) taken, with permission, from Ref. 13.]
creases and the discharge rate of most endings increases in proportion to muscle force 28-3°. This reflects an increase in fusimotor (or skeletofusimotor) outflow. If this discharge was due simply to static 63
thumb
Elicited sensations 2 out of 18 SA II afferents , , q ~
8 out of 11 joint afferents
1 out of 16 muscle spindle afferents
record
stimulate
microelectrode intrafascicular recording site
Fig. 2. Experimental arrangement for recording and stimulating intrafascicular sites that are associated with single afferents originating in the skin, joint or muscles of the hand. As indicated in the expanded diagrammatic view of a cutaneous fascicle of the median nerve, the tip of the microelectrode is manipulated until it is near a single axon. The behaviour of the afferent during natural stimulation of its receptive field is recorded before and after microstimulation at the recording site. Subjects receive single stimufi and trains of stimuli at different intensities and are required to indicate their response using a questionnaire expanded from that used by Vallbo and colleagues 41. Particular attention was paid to sensations of joint displacement or movement. Recordings were made in the median and ulnar nerves. The frequency with which stimulus trains elicited sensations are shown. (Modified from Ref. 42.)
fusimotor drive, then the contraction-induced improvement in movement detection would suggest that the CNS is influenced by the number of active spindle endings, since static fusimotor drive reduces the sensitivity of primary spindle afferents to small stretches al. For muscles with long tendons, muscle contraction is important in improving the mechanical linkage between the joint and muscle fibres32. The discharge from Golgi tendon organs also increases during voluntary contraction. Given the stiffness of actively contracting muscle and the high sensitivity of tendon organs to such a forcea3, the discharge from these organs will follow changes in muscle length34 and could thus contribute to detection of applied displacements. The history of previous muscle con64
traction is important in determining the sensitivity of muscle receptors - muscle spindle endings in a relaxed muscle discharge more in response to a standard lengthening when the muscle has just previously contracted to a short length. This induces predictable errors in position sense 35. During voluntary contraction, changes in central transmission of afferent input may modulate the detection of kinaesthetic signals. However, many studies have confirmed the seemingly paradoxical decrease in detectability and perceived intensity of cutaneous stimuli during voluntary contraction, particularly under non-isometric conditions (e.g. Refs 36, 37). The arguments above suggest that muscle afferent input during contraction is augmented. It is also possible that there is enhanced central transmission of muscle afferent volleys. If this is so, then muscle afferent signals are being processed differently to signals from specialized cutaneous afferents. Irrespective of the mechanisms involved, perception of cutaneous inputs but not that of muscle afferent inputs appears to be functionally gated during voluntary contractions.
Microstimulation of single afferents Not only has microneurography facilitated the recording of the behaviour of single afferents innervating skin, joint and muscle in humans, but it can also be used to stimulate axons so that their central projections and the perceptual effects of these projections can be studied. Experiments involving the stimulation of slowly and rapidly adapting cutaneous afferents innervating the human hand have resulted in subjective reports that are 'remarkably congruent in studies performed independently by different research groups in different laboratories' (Ref. 38; see also Refs 39--42), while criticisms of the technique have been largely refuted38. Specialized cutaneous innervation of the glabrous skin is divided into rapidly adapting receptors in Meissner's corpuscles (RA receptors), in Pacinian/ paciniform corpuscles (PC receptors), in slowly adapting receptors in Merkel cell neurite complexes (SA I receptors), and in Ruffmi endings (SA II receptors). Compared with SA I afferents, the SA II afferents have larger receptive fields, greater sensitivity to lateral skin stretch and often a background discharge. Stimuli are delivered through the insulated microelectrode at the intrafascicular site, from which the discharge of a single afferent is recorded. The uniform findings are as follows. (1) Low-level voltage or current pulses elicit a focal sensation of superficial flutter or vibration when applied at the site of recording of rapidly adapting afferents, and a sensation of sustained pressure when applied to SA I sites that include the distal region of the digits. (2) Microstimulation of S A I I afferents innervating the glabrous skin of the hand usually elicits no sensation. The technique has subsequently been applied to assess whether single muscle and joint afferents convey information about joint movement (Fig. 2)42. The study corroborated findings previously reported for cutaneous afferents, but, in addition, microstimulation with controlled electrical stimuli elicited sensations from 8 out of 11 intrafascicular sites from which the discharge of single slowly adapting joint TINS, Vol. 15, No. 2, 1992
afferents was recorded. Half the sensations elicited by single stimuli or trains of stimuli were of focal pressure at or very close to the receptive field. The remainder were sensations of joint movement - small illusory displacements (~10 °) that increased with stimulus frequency and that were not necessarily perceived to occur about the extremes of joint rotation where the afferents responded best. Sometimes the elicited illusions corresponded exactly with the complex multiaxial responsiveness of the afferent. In the same experimental sessions, microstimulation was performed (at intensities identical to those for cutaneous and joint afferents) at intrafascicular sites from which the discharge of 16 identified muscle spindle endings in intrinsic muscles of the hand was recorded. At only one spindle site did microstimulation yield a sensation when the stimulus intensity was below that necessary to activate nearby motor axons and so produce a twitch of the innervated muscle. The one positive response was a sensation normally associated with muscle lengthening. In agreement with other results (e.g. see Refs 38, 40, 41), stimulation of most S A I I afferent sites (15 out of 18) elicited no consistent sensations. Exceptions were two SA II afferents innervating the skin near the nallbed, which, when stimulated, elicited perceived flexion at the distal interphalangeal joint - an illusion consistent with their responses to passive movement.
Implications Although digital joint afferents have a limited capacity to encode movement and position 23, some must have a 'secure' projection to consciousness. Also, the discharge of single muscle spindle afferents appears to be insufficient to elicit the sensations attributed to the discharge of a population of them. Reliance on the population input from muscle spindle afferents would prevent the sensation being 'distorted' by any 'irregularities' introduced as a result of individual spindles responding to local disturbances within the parent muscle. Indeed, psychophysical studies suggest that the number and discharge frequency of spindle afferents influence the perceived velocity of illusory movements (e. g. Refs 5, 7). Thus, 'spatial coding' differs for receptors involved in control of the fingers: single cutaneous afferents (RA and SA I) can provide focal information about the skin surface, but the processing of movement sensation in the CNS relies on the population input from muscle spindle endings. Single stimuli applied to joint and some cutaneous afferents (RA and SA I) can elicit sensations (a finding that applies to muscle spindle afferents), but only when a sufficient number of them are activated (see also Ref. 7). S A I I afferents can encode local and remote joint movement 22, but individual afferents usually fail to arouse perception. It is concluded that: (1) a coherent discharge in any mechanoreceptor channel (cutaneous, muscle or joint) will not be entirely ignored by the CNS when computing limb position and movement sensations; (2) all afferent species can encode specific aspects of movement and each can elicit a movement percept (although relatively greater spatial summation may be required to elicit a sensation for muscle spindle afferents and S A I I afferents than for RA and SA I TINS, Vol. 15, No. 2, 1992
afferents); (3) movement enhances kinaesthetic acuity Acknowledgements via mechanisms that have not yet been fully defined in We are grateful to the humans; and (4) the contraction-associated improve- National Health and ment in proprioceptive acuity reinforces the major Medical Research Council of Australia role of muscle receptors in kinaesthesia. Selected references 1 McCIoskey, D. I. (1978) Physiol. Rev. 58, 763-820 2 McCIoskey, D. I. (1981) in Handbook of Physiology (The Nervous System, Vol. I1: Motor Control) (Brooks, V. B., ed.), pp. 1415-1447, American Physiological Society 3 Matthews, P. B. C. (1988) Can. J. Physiol. Pharmacol. 66, 430-438 4 Gandevia, S. C. (1987) Trends NeuroscL 10, 81-85 5 Roll, J. P. and Vedel, J. P. (1982) Exp. Brain Res. 47, 177-190 6 McCIoskey, D. I., Cross, M. J., Honner, R. and Potter, E. K. (1983) Brain 106, 21-37 7 Gandevia, S. C. (1985) Brain 108, 965-981 8 Gandevia, S. C. and McCIoskey, D. I. (1976) J. Physiol. 260, 387-407 9 Gandevia, S. C., Hall, L. A., McCIoskey, D. I. and Potter, E. K. (1983) J. Physiol. 335, 507-517 10 Ferrell, W. R., Gandevia, S. C. and McCIoskey, D. I. (1987) J. Physiol. 386, 63-71 11 Ferrell, W. R. and Smith, A. (1988) J. Physiol. 399, 49-61 12 Clark, F. J., Grigg, P. and Chapin, J. W. (1989) J. Neurophysiol. 61, 186-193 13 Hall, L. A. and McCIoskey, D. I. (1983) J. Physiol. 335, 519-533 14 Clark, F. J., Burgess, R. C., Chapin, J. W. and Lipscomb, W. T. (1985) J. Neurophysiol. 54, 1529-1540 15 Clark, F. J., Burgess, R. C. and Chapin, J. W. (1986) Brain 109, 1195-1208 16 Taylor, J. L. and McCIoskey, D. I. (1990) Brain 113,157-166 17 Taylor, J. L. and McCIoskey, D. I. (1988) Brain Res. 70, 351-360 18 Taylor J. L. and McCIoskey, D. I. (1990) Exp. Brain Res. 81, 413-416 19 Brown, K., Lee, J. and Ring, P. A. (1954) J. Physiol. 126, 448-458 20 Colebatch, J. G. and McCIoskey, D. I. (1987) J. Physiol. 386, 247-261 21 John, K. T., Goodwin, A. W. and Darian-Srnith, I. (1989) Exp. Brain Res. 78, 62-68 22 Edin, B. B. and Abbs, J. H. (1991) J. NeurophysioL 65, 657-670 23 Burke, D., Gandevia, S. C. and Macefield, G. (1988) J. Physiol. 402, 347-361 24 Ferrell, W. R. (1980) J. Physiol. 299, 85-99 25 Proske, U., Schaible, H-G. and Schmidt, R. F. (1988) Exp. Brain Res. 72, 219-224 26 Baxendale, R. H. and Ferrell, W. R. (1983) Brain Res. 261, 195-203 27 Hulliger, M., Nordh, E., Thelin, A-E. and Vallbo, A. B. (1979) J. Physiol. 291,233-249 28 Burke, D. (1981) Int. Rev. PhysioL 25, 91-126 29 Edin, B. B. and Vallbo, /k. B. (1990) J. Neurophysiol. 63, 1307-1313 30 Vallbo, A. B. (1974) Acta PhysioL Scand. 90, 319-336 31 Hulliger, M., Matthews, P. B. C. and Noth, J. (1977) J. PhysioL 267, 811-838 32 Rack, P. M. H. and Ross, H. F. (1984) J. Physiol. 351,99-110 33 Proske, U. (1981) Int. Rev. Physiol. 25, 127-171 34 Stauffer, E. K. and Stephens, J. A. (1977) J. Neurophysiol. 40, 681-691 35 Gregory, J. E., Morgan, D. L, and Proske, U. (1988) J. Neurophysiol. 59, 1220-1230 36 Angel, R. W. and Malenka, R. C. (1982) Exp. Neurol. 77, 266-274 37 Milne, R. J., Aniss, A. M., Kay, N. E. and Gandevia, S. C. (1988) Exp. Brain Res. 70, 569-576 38 TorebjSrk, E., Vallbo, ~,. B. and Ochoa, J. L. (1987) Brain 110, 1509-1529 39 Konietzny, F., Perl, E. R., Trevino, D., Light, A. and Hensel, H. (1981) Exp. Brain Res. 42, 219-222 40 Ochoa, J. and TorebjSrk, E. (1983) J. PhysioL 342, 633-654 41 Vallbo, A. B., Olsson, K. ]k., Westberg, K-G. and Clark, F. J. (1984) Brain 107, 727-749 42 Macefield, G., Gandevia, S. C. and Burke, D. (1990) J. Physiol. 429, 113-129
for continuing financialsupport. We acknowledge the contribution of our collaborators to the work cited here.
65
