Diffuse Optical Tomography Activation in the Somatosensory Cortex: Specific Activation by Painful vs. Non-Painful Thermal Stimuli

Background

Pain is difficult to assess due to the subjective nature of self-reporting. The lack of objective measures of pain has hampered the development of new treatments as well as the evaluation of current ones. Functional MRI studies of pain have begun to delineate potential brain response signatures that could be used as objective read-outs of pain. Using Diffuse Optical Tomography (DOT), we have shown in the past a distinct DOT signal over the somatosensory cortex to a noxious heat stimulus that could be distinguished from the signal elicited by innocuous mechanical stimuli. Here we further our findings by studying the response to thermal innocuous and noxious stimuli.

Methodology/Principal Findings

Innocuous and noxious thermal stimuli were applied to the skin of the face of the first division (ophthalmic) of the trigeminal nerve in healthy volunteers (N = 6). Stimuli temperatures were adjusted for each subject to evoke warm (equivalent to a 3/10) and painful hot (7/10) sensations in a verbal rating scale (0/10 = no/max pain). A set of 26 stimuli (5 sec each) was applied for each temperature with inter-stimulus intervals varied between 8 and 15 sec using a Peltier thermode. A DOT system was used to capture cortical responses on both sides of the head over the primary somatosensory cortical region (S1). For the innocuous stimuli, group results indicated mainly activation on the contralateral side with a weak ipsilateral response. For the noxious stimuli, bilateral activation was observed with comparable amplitudes on both sides. Furthermore, noxious stimuli produced a temporal biphasic response while innocuous stimuli produced a monophasic response.

Conclusions/Significance

These results are in accordance with fMRI and our other DOT studies of innocuous mechanical and noxious heat stimuli. The data indicate the differentiation of DOT cortical responses for pain vs. innocuous stimuli that may be useful in assessing objectively acute pain.     

Mechanisms of pain arising from articular tissues

This paper reviews the peripheral and central neural mechanisms underlying pain from articular tissues innervated by spinal and trigeminal afferents. The paper especially addresses trigeminal mechanisms related to pain from the temporomandibular joint and its associated craniofacial musculature. Recent studies have shown the existence of articular nociceptive primary afferents that project to the spinal cord dorsal horn and trigeminal brainstem complex. A particular feature of most neurones receiving these deep nociceptive afferent inputs is their responsivity also to cutaneous nociceptive afferent inputs. This suggests the involvement of these neurones not only in the detection of acute articular pain, but also in the hyperalgesia and poor localization, spread, and referral of pain that characterize many painful conditions of joints and other deep structures. While only limited information is available on related higher brain centre mechanisms, convergence and interaction between cutaneous and deep afferent inputs also seem to be a characteristic of somatosensory neurones in the thalamus and somatosensory cerebral cortex. Muscle and autonomic reflexes may be induced by such deep noxious stimuli, but the functional significance of some of these effects (e.g., in relation to clinical concepts of myofascial dysfunction) requires further study in more appropriate functional settings.     

Population Coding of Forelimb Joint Kinematics by Peripheral Afferents in Monkeys

Various peripheral receptors provide information concerning position and movement to the central nervous system to achieve complex and dexterous movements of forelimbs in primates. The response properties of single afferent receptors to movements at a single joint have been examined in detail, but the population coding of peripheral afferents remains poorly defined. In this study, we obtained multichannel recordings from dorsal root ganglion (DRG) neurons in cervical segments of monkeys. We applied the sparse linear regression (SLiR) algorithm to the recordings, which selects useful input signals to reconstruct movement kinematics. Multichannel recordings of peripheral afferents were performed by inserting multi-electrode arrays into the DRGs of lower cervical segments in two anesthetized monkeys. A total of 112 and 92 units were responsive to the passive joint movements or the skin stimulation with a painting brush in Monkey 1 and Monkey 2, respectively. Using the SLiR algorithm, we reconstructed the temporal changes of joint angle, angular velocity, and acceleration at the elbow, wrist, and finger joints from temporal firing patterns of the DRG neurons. By automatically selecting a subset of recorded units, the SLiR achieved superior generalization performance compared with a regularized linear regression algorithm. The SLiR selected not only putative muscle units that were responsive to only the passive movements, but also a number of putative cutaneous units responsive to the skin stimulation. These results suggested that an ensemble of peripheral primary afferents that contains both putative muscle and cutaneous units encode forelimb joint kinematics of non-human primates.     

A new combination of methods for the localization,identification, and three-dimensional reconstruction of the sensory endings of articular afferents characterized by electrophysiology

A combination of methods is described to identify and reconstruct corpuscular and non-corpuscular sensory endings of group II and group III nerve fibers following functional examination by electrophysiology. Afferent units activated by electrical stimulation of the medial articular nerve of the cat's knee were analyzed by single fiber recordings and characterized by their responsiveness to mechanical stimuli. The receptive fields of the units were closely demarcated by fine needles when the responses elicited by insertion of the needles were being recorded. After fixation, the tissue around the demarcated field was dissected and histologically processed. Series of semithin sections were cut from the embedded tissue blocks containing the receptive fields. Corpuscular endings of group II fibers and peripheral myelinated group III nerve fibers, presumably corresponding to the characterized units, were identified by light microscopy of semithin sections and localized within the demarcated area. Non-corpuscular endings were identified by electron microscopy of ultrathin sections cut in alternation with, or after re-embedding of, semithin sections. Morphometric analysis of ultrathin section series allowed the measurement of parameters such as the mean axon diameter and the organelle content of the sensory endings. The methods described are appropriate for collecting data that correlate the structural and functional characteristics of sensory endings in deep tissues.     

Role of knee ligaments in proprioception and regulation of muscle stiffness

Physiologic evidence for the sensory role of the knee joint ligaments are reviewed. The cruciate and collateral ligaments accomodate morphologically different sensory nerve endings with different capabilities of providing the central nervous system (CNS) with information not only about noxious and chemical stimuli but also about mechanical events, e.g., movement- and position-related stretches of the ligaments. Available data show that low-threshold joint/ligament receptor (i.e., mechanoreceptor) afferents evoke only weak and rare effects in skeletomotor neurons (α-motoneurons), whereas they frequently and powerfully influence fusimotor neurons (γ-motoneurons). The effects on the γ-muscle-spindle system in the muscles around the knee are so potent that even stretches of the cruciate ligaments at relatively moderate loads (not noxious) may induce major changes in responses of the muscle spindle afferents. As the activity in the primary muscle spindle afferents modifies stiffness in the muscles, the cruciate ligament receptors may, through the γ-muscle-spindle system, participate in regulation and preparatory adjustment of the stiffness of the muscles around the knee joint and thereby of knee joint stiffness. Thus, the sensory system of the cruciate ligaments is able to contribute significantly to the functional stability of the knee joint. The possible role of (ligamentous) joint receptors in genesis and spread of muscular tension in occupational muscle pain and in chronic musculoskeletal pain syndromes is also discussed.     

Afferent signals from pigeon extraocular muscles modify the vestibular responses of units in the abducens nucleus.

Although the extraocular muscles contain stretch receptors it is generally believed that their afferents exert no influence on the control of eye movement. However, we have shown previously that these afferent signals reach various brainstem centres concerned with eye movement, notably the vestibular nuclei, and that the decerebrate pigeon is a favourable preparation in which to study their effects. If the extraocular muscle afferents do influence oculomotor control from moment-to-moment they should exert a demonstrable effect on the oculomotor nuclei. We now present evidence that extraocular muscle afferent signals do, indeed, alter the responses of units in an oculomotor nucleus (the abducens, VI nerve nucleus, which supplies the lateral rectus muscle) to horizontal, vestibular stimulation induced by sinusoidal oscillation of the bird. Such stimuli evoke a vestibulo-ocular reflex in the intact bird. The extraocular stretch receptors were activated by passive eye movement within the pigeon's saccadic range; such movements modified the vestibular responses of all 19 units studied which were all, histologically, in the abducens nucleus. The magnitude of the effects, purely inhibitory in 15 units, depended both on the amplitude and the velocity of the eye movement and most units showed selectivity for particular combinations of plane (e.g. horizontal versus vertical) and direction (e.g. rostral versus caudal) of eye movement. The results show that an afferent signal from the extraocular muscles influences vestibularly driven activity in the abducens nucleus to which it carries information related to amplitude, velocity, plane and direction of eye movement in the saccadic range. They thus strongly support the view that extraocular afferent signals are involved in the control of eye movement.     

猕猴执行痛,热延缓辨别作业时额叶神经元及肌电...

To study the functional implication of neuronal responses in frontal cortex to noxious and innocuous heat stimuli, experiments were carried out on two monkeys (Macaca mulatta) during performing a delayed discrimination GO/NO-GO task. The animals were trained to hold a lever for at least 3 s in the response period when an innocuous heat stimulus had been applied to the forearm in the cue period, or release it within 1 s with a noxious one. After a criterion of 90% of correct responses in 3 successive days was reached, single neuronal activity was recorded from the frontal cortex along with EMGs from six muscles in both arms. Of 142 task-related neurons recorded, 87 (66.4%) were related to heat, 34 to visual, and 21 to both visual and heat stimuli. Among the heat-related neurons, 22 were responsive to noxious heat, 18 to innocuous heat, and 47 to both. Of the neurons responding to both noxious and innocuous heat stimuli, 4 neurons showed changes in discharge rates, depending on the type of stimuli. In most cases, muscular activities just appeared at the moment of lever pressing and/or releasing in m. extensor digitorium communis, m. flexor carpi ulnaris and m. flexor carpi radialis of the performing arm. No regular muscular activities appeared in other muscles and in other periods of the task. The result showed that the neuronal responses in frontal cortex elicited by heat stimuli in the cue period were not related directly to the initiation and modulation of behaviours.(ABSTRACT TRUNCATED AT 250 WORDS)     

Responses of neurons in the rat nucleus submedius to noxious and innocuous mechanical cutaneous stimulation

Extracellular recordings were used to characterize responses to cutaneous mechanical stimulation of 78 neurons in the rat nucleus submedius (SM). Thirty-nine of these units were activated by some type of cutaneous mechanical stimulation. Eighteen cells were activated exclusively by noxious stimuli. In 13 of these cells, responses were of swift onset and relatively rapid termination following stimulus application. In contrast, in three neurons responses were delayed both in onset and termination, and in two the response was immediate, but the markedly increased evoked activity outlasted stimulus application by 13 min. Receptive fields (RFs) of these nociceptive neurons were generally large, although none were bilateral. Four SM neurons were activated by innocuous stimuli, but their maximal response was obtained only after noxious stimulation. Responses of all of these neurons were of immediate onset and recovery, and their RFs were large (two were bilateral). Twelve SM neurons were activated maximally by innocuous stimuli. Responses of seven of these cells were immediate in onset and termination, while that of three were delayed in both onset and termination. Two of the 12 innocuous-only neurons quickly became unresponsive to repeated stimulus applications, and could be reactivated only after a rest period during which no stimuli were applied. RFs of these units were also generally large, and in three cases were bilateral. Five SM neurons responded by decreasing, or completely ceasing, their firing subsequent to noxious-only (n = 2), or innocuous-only (n = 3) stimulation. Four of these units had large RFs (two were bilateral). The remaining 39 SM neurons could not be activated by any type of mechanical cutaneous stimulation we tried. Electrical stimulation of the ventrolateral orbital cortex (VLO) was employed to examine frontal cortical projections of 21 SM neurons. Ten of these units were activated, although all of them synaptically rather than antidromically, and two were inhibited. There was no clear-cut relationship between neuronal location, physiological type, RF site, or VLO stimulation effects among the 39 SM neurons. These results provide further support for the involvement of SM neurons in nociceptive information signaling, and suggest additionally that the role of the nucleus is not limited to nociception but encompasses a wider range of cutaneous sensations.     

Responses of Neurons in the Rat Nucleus Submedius to Noxious and Innocuous Mechanical Cutaneous Stimulation

Extracellular recordings were used to characterize responses to cutaneous mechanical stimulation of 78 neurons in the rat nucleus submedius (SM). Thirty-nine of these units were activated by some type of cutaneous mechanical stimulation. Eighteen cells were activated exclusively by noxious stimuli. In 13 of these cells, responses were of swift onset and relatively rapid termination following stimulus application. In contrast, in three neurons responses were delayed both in onset and termination, and in two the response was immediate, but the markedly increased evoked activity outlasted stimulus application by 13 min. Receptive fields (RFs) of these nociceptive neurons were generally large, although none were bilateral. Four SM neurons were activated by innocuous stimuli, but their maximal response was obtained only after noxious stimulation. Responses of all of these neurons were of immediate onset and recovery, and their RFs were large (two were bilateral). Twelve SM neurons were activated maximally by innocuous stimuli. Responses of seven of these cells were immediate in onset and termination, while that of three were delayed in both onset and termination. Two of the 12 innocuous-only neurons quickly became unresponsive to repeated stimulus applications, and could be reactivated only after a rest period during which no stimuli were applied. RFs of these units were also generally large, and in three cases were bilateral. Five SM neurons responded by decreasing, or completely ceasing, their firing subsequent to noxious-only (n = 2), or innocuous-only (n = 3) stimulation. Four of these units had large RFs (two were bilateral). The remaining 39 SM neurons could not be activated by any type of mechanical cutaneous stimulation we tried.

Electrical stimulation of the ventrolateral orbital cortex (VLO) was employed to examine frontal cortical projections of 21 SM neurons. Ten of these units were activated, although all of them synaptically rather than antidromically, and two were inhibited. There was no clear-cut relationship between neuronal location, physiological type, RF site, or VLO stimulation effects among the 39 SM neurons.

These results provide further support for the involvement of SM neurons in nociceptive information signaling, and suggest additionally that the role of the nucleus is not limited to nociception but encompasses a wider range of cutaneous sensations.     

猕猴执行痛、热延缓辨别作业时额叶神经元及肌电活动的观察

为研究额叶神经元对躯体痛、热刺激出现反应的机能意义,设计了痛、热延缓辨别作业对猕猴进行实验。痛和热仅在暗示期给予,要求动物对此不立即作出行为反应。而是暂时记住这个信号,等到行动期再作出反应。待作业正确率连续三天达90%以上,记录额叶神经元和两侧上肢肌肉的电活动.以观察神经元活动与信号刺激以及肌肉活动之间的关系。在记录的142个作业相关神经元中,与痛、热刺激相关者87个(66.4%)。其中22个仅对痛刺激、18个仅对热刺激起反应,47个对痛、热刺激都起反应(其中4个反应方向相反)。此外,有21个对痛、热、视都出现反应。其余仅对视觉刺激起反应。这些神经元主要位于额叶弓状沟上支内侧,包括前额叶和运动前区皮层。在两侧上肢12块肌肉中,除操作侧指总伸肌、尺侧腕屈肌和桡侧腕屈肌在压放杠杆时,可记录到短暂的肌电活动外,在作业的其它时期和其它肌肉均未记录到规律的肌电活动。这表明痛热刺激在暗示期引起的神经元反应与行为动作的发动没有直接关系,从而支持关于额叶皮层这一区域的神经元在对刺激物信号意义的辨别机制中,可能起着重要作用的假说。     

Unit activity in the superior colliculus of the cat following passive eye movements.

Unit response in the superior colliculus and underlying structures has been examined in the choralose-anaesthetized cat following passive movement of an occluded eye. One group of units was sensitive to small saccadic movements, responded regardless of the initial postion of the eye, and in most instances responded to movements in opposit directions. A second numerically smaller group also responded when they eye was moved at saccadic velocity but only when the eye passed a fixed point. Such units with fixed positional thresholds were found following movements in both nasal and temporal directions as well as to both upward and downward movement. Both types of unit response were found after transection of the optic nerve and were also recorded when individual extraocular muscles were subjected to controlled stretch. It is assumed that most unit activity seen after passive movement of the occluded eye is due to activity in extraocular muscle receptors. In the deep layers of the superior colliculus responses to small eye movements were found to be due to the activation of very low threshold receptors sensitive to vibration in the facial area.     

Activation of pontine and bulbar reticulo-spinal neurons in the cat by somatosensory stimuli of different modalities

The response pattern of reticulo-spinal (RS) neurons in two reticulo-spinal structures (n. reticularus pontis caudalis and n. reticularis gigantocellularis) to both electrical (somatic) nerve stimulation and natural mechanical innocuous (tapping with varying force) and noxious (pinch and prick) stimulation were investigated in chloralose-anesthetized cats. Bulbar and pontine neurons were found to vary considerably in their sensory characteristics: of the former 43% were activated only by high-threshold electrical nerve stimulation and noxious stimuli, while the remainder responded to innocuous stimuli as well. In the case of pontine neurons 81% produced a response to stimulation of low-threshold nerve fibers, and to innocuous as well as noxious stimuli. A relationship was found between the sensory characteristics of reticulo-spinal neurons and their axon conductance velocities. Various aspects and the likely functional significance of specialization in brainstem neurons of the pontine and bulbar reticular formation come under discussion.A. A. Bogomolets Institute of Physiology, Academy of Sciences of the Ukrainian SSR, Kiev. Translated from Neirofiziologiya, Vol. 18, No. 4, pp. 461–469, July–August, 1986.     

Central neurophysiological processing of joint pain on the basis of studies performed in normal animals and in models of experimental arthritis.

On the basis of anatomical and electrophysiological studies, this review summarizes first, the data dealing with the transmission of joint inputs in the central nervous system of normal animals at the spinal and supraspinal levels. It appears that in these conditions neuronal responses to mechanical noxious stimuli of the joints are relatively few and (or) weak. Second, in sharp contrast, the studies performed in polyarthritic rats have emphasized the profound changes in the activities (spontaneous firing and responsiveness) of the somatosensory neurones at various levels of the central nervous system (CNS), including the thalamus and primary somatosensory cortex; many were spontaneously active and a majority of them could be maximally activated by gentle mechanical stimuli applied to the inflamed joints. Although the change in the sensitivity of the peripheral mechanoreceptors has a major role in the modifications described in the CNS, additional observations have suggested a complex interaction between peripheral and central processes. On the basis of the recent data obtained in poly- and mono-arthritic animals; the following phenomena have been successively considered: the segmental and hetero-segmental "cross-talk" and their possible relationship with referred pain; the involvement of "new" neuronal populations as a possible basis of a selective system for joint pain; and the possible involvement of changes in the various control systems that normally modulate the nociceptive inputs at different levels of the CNS.     

Escape behavior in response to mechanical stimulation of hindwing in cricket, Gryllus bimaculatus

We studied behavioral responses of the cricket, Gryllus bimaculatus, to mechanical stimulation of the hindwing tip using three different kinds of stimuli: touching, bending and pinching. The most characteristic was a sequence of initial jump-like movements and subsequent running steps, that is referred to as escape behavior in this study. Touching stimulus elicited the escape behavior in 52% of resting animals tethered on a treadmill, whereas bending elicited the same behavior in 94% or 98% depending on the bend direction. Pinching was effective in all tested animals. The effectiveness of pinching stimulus in eliciting the escape behavior depends on the ongoing activity in the animal. Video and electromyographic recordings have revealed that, in the initial jump-like movements, forelegs and hindlegs move simultaneously on both sides while midlegs remain on the ground, followed by simultaneous movements of bilateral midlegs. The subsequent stepping was characterized by out-of-phase rhythmical activities of the leg muscles. Touching stimulus evoked tonic afferent responses of small amplitude in the second nerve root of the metathoracic ganglion. Bending stimulus evoked tonic responses of large units that showed rapid habituation and medium units that persisted during repeated stimulation. Pinching stimulus elicited only phasic responses of large and medium amplitude in the R2 afferents. The results suggest that touching, bending and pinching stimuli are transmitted to the metathoracic ganglion via different sensory systems having different effectiveness in activating the escape motor system.     

Afferent Modulation of Warmth Sensation and Heat Pain in the Human Hand

The hands of 14 normal humans were used to determine the somatotopic organization of the modulation of warmth sensation and heat pain by different forms of cutaneous stimuli. Test stimuli were 5-sec heat pulses ranging from 36° to 51°C, delivered to the fingerpads of digits 1, 2, 4, and 5 with a contact thermode. Conditioning stimuli (15 sec) bracketed the test stimuli and included vibration, noxious and innocuous heat, cold, and electrical pulses delivered to the fingerpads of digits that were adjacent or nonadjacent to the tested digits. Noxious (48° ± 1.3°C), but not innocuous (43°C), heat stimuli increased the perceived magnitude estimation of innocuous test stimuli (36–43°C) by 20–37% when delivered to adjacent, but not to nonadjacent, digits. No other conditioning stimuli had any effect on the intensity of warmth perception. In contrast, both noxious and innocuous heat or electrical conditioning reduced the magnitude estimation of noxious (50–5°C), but not innocuous, test pulses by 12–22% when delivered to adjacent digits. Conditioning of nonadjacent digits was significantly less effective. The analgesic effects of noxious and innocuous conditioning were approximately equal. Vibratory (120 Hz, 3.5 μm) and cold (15°C) conditioning stimuli were ineffective. The results are consistent with a dermatomal somatotopic organization of tactile and heat modulatory influences on warmth sensation and heat pain. The results further suggest that the neural mechanisms subserving warmth mediate a negative feedback influence on heat pain intensity.     

Cutaneous sensitivity to touch and low-frequency vibration in selachians

In 46 isolated and 4in situ preparations of perioral skin, afferent impulse patterns, leading from cutaneous mechanoreceptors and responding to orthogonal square as well as sine-wave stimuli, were investigated in selachians (Scyliorhinus canicula).

Use-dependent inhibition of C-axon multimodal units of cat skin by lidocaine and N-propyl-ajmaline]

Subcutaneous application of local anesthetic drug lidocaine and cardiac antiarrhythmic n-propyl-ajmaline produced the reversible use-dependent inhibition of feline polymodal mechano-heat C-fiber cutaneous sensory units (CMH-units) excited by moderate noxious mechanical stimulus. The discharge rate as well as the number of evoked spikes of polymodal sensory units treated with the drugs decreased below the values observed under noxious chemical excitation of CMH-units. The repeated mechano-stimulation with 5 to 30 sec interval between stimuli produced complete though a reversible block of the treated units. Quaternary amine n-propyl-ajmaline induced use-dependent inhibition of CMH-units in lower concentrations than tertiary amine lidocaine. The use-dependent inhibition of CMH-units is discussed in connection with nociception and local analgesia.     

Activity of precentral neurons during torque-triggered hand movements in awake primates.

In monkeys performing a handle-repositioning task involving primarily wrist flexion-extension (F-E) movements after a torque perturbation was delivered to the handle, single units were recorded extracellularly in the contralateral precentral cortex. Precentral neurons were identified by passive somatosensory stimulation, and were classified into five somatotopically organized populations. Based on electromyographic recordings, it was observed that flexors and extensors about the wrist joint were specifically involved in the repositioning of the handle, while many other muscles which act at the wrist and other forelimb joints were involved in the task in a supportive role. In precentral cortex, all neuronal responses observed were temporally correlated to both the sensory stimuli and the motor responses. Visual stimuli, presented simultaneously with torque perturbations, did not affect the early portion of cortical responses to such torque perturbations. In each of the five somatotopically organized neuronal populations, task-related neurons as well as task-unrelated ones were observed. A significantly larger proportion of wrist (F-E) neurons was related to the task, as compared with the other, nonwrist (F-E) populations. The above findings were discussed in the context of a hypothesis for the function of precentral cortex during voluntary limb movement in awake primates. This hypothesis incorporates a relationship between activities of populations of precentral neurons, defined with respect to their responses to peripheral events at or about single joints, and movements about the same joint.     

Response of Cat Cortical Neurons to Position and Movement of the Femur

The contribution of joint afferents to the response of cortical neurons in area 3a to mechanical stimulation of the contralateral hindlimb was evaluated in cats anesthetized with sodium pentobarbital and paralyzed with pancuronium bromide. The hindlimb projection to the pericruciate cortex was established by recording the evoked potentials to electrical stimulation of the sciatic nerve and some of its branches, the biceps-semitendinosus and the quadratus femoris

Out of 169 neurons, 63 responded exclusively to cutaneous stimuli (superficial), whereas the others could be activated by local pressure of hindlimb muscles and/or by joint rotation (deep). Deep neurons were classified as slowly adapting (SA) or rapidly adapting (RA) units. In the neurons responding exclusively to joint rotation, the site of the receptive field could not be identified with certainty. In 13 deep neurons, their firing was affected by rotation of multiple joints of the contralateral hindlimb

In an attempt to identify the source of activation of cortical neurons, partial denervations and muscle disconnections were performed in five animals to isolate and stimulate the hip capsule. In these preparations, in 14 of 15 cortical neurons the source of activation was localized in the periarticular muscles, with no response to mechanical stimulation of the joint capsule. Only one neuron (S A) could be selectively excited by punctate pressure on the hip capsule

Our results suggest that in neurons of area 3a of the cat, the information about the position of the femur relies mainly on muscle afferents