Group III and IV muscle afferents differentially affect the motor cortex and motoneurones in humans

Authors


Corresponding author J. L. Taylor: Prince of Wales Medical Research Institute, University of New South Wales, Barker Street, Randwick, Sydney, NSW 2031, Australia. Email: jl.taylor@unsw.edu.au

Abstract

The influence of group III and IV muscle afferents on human motor pathways is poorly understood. We used experimental muscle pain to investigate their effects at cortical and spinal levels. In two studies, electromyographic (EMG) responses in elbow flexors and extensors to stimulation of the motor cortex (MEPs) and corticospinal tract (CMEPs) were evoked before, during, and after infusion of hypertonic saline into biceps brachii to evoke deep pain. In study 1, MEPs and CMEPs were evoked in relaxed muscles and during contractions to a constant elbow flexion force. In study 2, responses were evoked during elbow flexion and extension to a constant level of biceps or triceps brachii EMG, respectively. During pain, the size of CMEPs in relaxed biceps and triceps increased (by ∼47% and ∼56%, respectively; P < 0.05). MEPs did not change with pain, but relative to CMEPs, they decreased in biceps (by ∼34%) and triceps (by ∼43%; P < 0.05). During flexion with constant force, ongoing background EMG and MEPs decreased for biceps during pain (by ∼14% and 15%; P < 0.05). During flexion with a constant EMG level, CMEPs in biceps and triceps increased during pain (by ∼30% and ∼26%, respectively; P < 0.05) and relative to CMEPs, MEPs decreased for both muscles (by ∼20% and ∼17%; P < 0.05). For extension, CMEPs in triceps increased during pain (by ∼22%) whereas MEPs decreased (by ∼15%; P < 0.05). Activity in group III and IV muscle afferents produced by hypertonic saline facilitates motoneurones innervating elbow flexor and extensor muscles but depresses motor cortical cells projecting to these muscles.

The central actions of group III and IV muscle afferents are poorly understood. They respond to a range of local chemical and mechanical changes (e.g. Mense, 1977; Kniffki et al. 1978; Kindig et al. 2007), including fatiguing contractions and the metabolic products of fatigue (e.g. Kaufman et al. 1983). Many of them are nociceptive and their activity can be increased by intramuscular injection of hypertonic saline (e.g. Paintal, 1960; Kumazawa & Mizumura, 1977). Activity in these afferents may produce changes at motoneurones or higher in the motor pathways, possibly at the motor cortex (for reviews see Garland & Kaufman, 1995; Gandevia, 2001).

In the cat, group III and IV muscle afferents have different actions on flexor and extensor motoneurone pools (e.g. Kniffki et al. 1979, 1981; Schomburg et al. 1999; Schomburg & Steffens, 2002). During chemical activation of nociceptive afferents, many motoneurones innervating extensor muscles are hyperpolarized, whereas those innervating flexor muscles are depolarized. Excitation of flexor muscles along with inhibition of extensors is consistent with the predominant actions of several classes of afferents collectively referred to as flexor reflex afferents (FRA) (e.g. Eccles & Lundberg, 1959). We recently showed that, as for cats, human motoneurone pools are not uniformly affected by inputs from these afferents (Martin et al. 2006b). Responses of motoneurones to direct stimulation of the corticospinal tract are facilitated in elbow flexor muscles and inhibited in extensor muscles when activity in group III and IV muscle afferents is maintained by ischaemia after a fatiguing contraction. H-reflexes evoked in the wrist flexors or ankle plantarflexors tend to be reduced by activity in group III and IV muscle afferents (e.g. Garland & McComas, 1990; Le Pera et al. 2001) but this could be due to their presynaptic effect on Ia afferents (e.g. Rossi et al. 1999; Kalezic et al. 2004).

Responses to magnetic stimulation of the motor cortex are variously reported to increase, decrease or remain unaffected by inputs from group III and IV muscle afferents under different conditions (Taylor et al. 2000; Le Pera et al. 2001; Del Santo et al. 2007). Some of this uncertainty is because muscle responses to cortical stimulation depend on excitability at the motoneuronal as well as the cortical level. Thus it is critical to determine the effects of these afferents at both sites.

In the present studies, activation of group III and IV muscle afferents by intramuscular infusion of hypertonic saline was used to investigate their actions on the human motor pathways. To detect changes at different levels, we stimulated the motor cortex as well as the corticospinal tract to elicit responses in elbow flexor and extensor muscles. Corticospinal synapses onto motoneurones lack classical presynaptic inhibition (e.g. Nielsen & Petersen, 1994; Jackson et al. 2006) and thus responses to corticospinal stimulation can be used to assess motoneurone excitability. Because activity of these muscle afferents after a fatiguing contraction does not produce uniform effects for homonymous and antagonist muscles (Martin et al. 2006b), we investigated the behaviour of both elbow flexor and extensor muscles.

Methods

In two studies, stimulation of the motor cortex and corticospinal tract was used to assess the effects of activity of group III and IV muscle afferents, evoked by an infusion of hypertonic saline, on the excitability of the motor pathways. An additional control study was also performed. Experiments were separated by at least three weeks. Nine subjects (2 females; 37 ± 15 years, mean age ±s.d.) took part in one or more of the studies. The procedures were approved by the local ethics committee and the study was conducted according to the Declaration of Helsinki. All subjects gave their written informed consent to participate.

Force and EMG recordings

Subjects were seated with the right shoulder flexed at 90 deg and the elbow flexed at 120 deg. The arm was held in an isometric myograph with a strap at the wrist. EMG was recorded from biceps and triceps brachii by surface electrodes secured to the skin over the belly of each muscle (Ag–AgCl, 10 mm diameter). The signals were amplified, filtered (16–1000 Hz), and collected at 2 kHz for off-line analysis using customized software (CED 1401 with Spike and Signal software; Cambridge Electronic Design, Cambridge, UK).

Stimulation

Recordings were made of the motor responses in the elbow flexor and extensor muscles to stimulation at the brachial plexus, stimulation between the mastoids, and magnetic stimulation over the motor cortex.

Brachial plexus stimulation Single electrical stimuli were delivered with a cathode in the supraclavicular fossa at Erb's point and an anode on the acromion (100 μs duration; Digitimer DS7AH constant-current stimulator; Digitimer, Welwyn Garden City, UK). For each subject, the intensity of stimulation was gradually increased until the biceps and triceps compound muscle action potentials were maximal (Mmax). The sizes of Mmax were then used to determine the appropriate sizes of responses to motor cortex and corticospinal stimulation (see below).

Transcranial magnetic stimulation A circular coil (13.5 cm outside diameter) positioned over the vertex elicited motor evoked potentials (MEPs) recorded from biceps and triceps (Magstim 200; Magstim, Dyfed, UK). The direction of current flow in the coil preferentially activated the left motor cortex. In study 1, stimulus intensities were set to evoke responses in biceps of ∼10–15%Mmax during relaxation (46–90% of maximal stimulator output) and during elbow flexion (33–42%). In study 2, stimulus intensities were set to evoke responses of ∼10–15%Mmax in biceps during elbow flexion (30–41%) and in triceps during elbow extension (38–50% stimulator output). In study 3, responses of ∼10–15%Mmax were evoked in biceps during flexion and triceps during extension by stimulus intensities of 30–38% and 35–50% of stimulator output, respectively.

Corticospinal tract stimulation The corticospinal tract was stimulated by passing an electrical current (100 μs duration; Digitimer DS7AH) between electrodes placed over the mastoids (1–2 cm posterior and superior to the tip of the mastoid processes with the cathode on the left side) (e.g. Ugawa et al. 1991; Gandevia et al. 1999). Activation occurs at the cervicomedullary junction and evokes large, short-latency responses (cervicomedullary motor evoked potentials (CMEPs)) in arm muscles. The stimulus activates many of the same axons as motor cortical stimulation because the single volley evoked by transmastoid stimulation can largely occlude the response to cortical stimulation when the interstimulus interval is appropriate (Taylor et al. 2002). A significant proportion of the motoneuronal response to cervicomedullary stimulation is monosynaptic for biceps (Petersen et al. 2002). The latency of responses was monitored to ensure that high stimulation intensities did not activate the motor axons at or near the ventral roots. A jump in latency of ∼2 ms occurs when the site of stimulation spreads from descending tracts to the ventral roots (Taylor & Gandevia, 2004). In study 1, stimulus intensities were set to evoke responses in biceps of ∼10–15%Mmax during relaxation (240–320 mA) and during elbow flexion (190–290 mA). In study 2, stimulus intensities were set to evoke responses of ∼10–15%Mmax in biceps during elbow flexion (165–275 mA) and in triceps during elbow extension (210–320 mA). In study 3, stimulus intensities evoked responses of ∼10–15%Mmax in biceps (165–255 mA) and triceps (190–320 mA) during flexion and extension, respectively. For one subject, the high stimulus intensities needed to evoke triceps CMEPs during elbow extension activated the ventral roots projecting to biceps and hence, data for this subject for biceps were not used.

Experimental muscle pain

Experimental muscle pain was induced by infusion of sterile hypertonic saline (5%) into biceps with a computer-controlled syringe pump (Graseby Syringe Pump 3100, Graseby, Watford, Herts) and a 24-gauge catheter inserted into the proximal part of the muscle belly. Prior to the infusion, a 0.5 ml bolus of hypertonic saline was delivered. The infusion rate was adjusted to maintain the pain level at ∼3–5 level on a rating scale (150–250 μl min−1; see below). In study 3, the same procedures and infusion rates were used for injection of 0.9% isotonic saline, which generally does not induce pain.

Participants were asked to rate verbally their perceived pain on a modified Borg scale consisting of numbers from 0 to 10 and corresponding descriptors of pain: 0, infinitely small; 0.5, just noticeable; 1, slight; 2, mild; 3, moderate; 4, considerable; 5, large; 7, very large; 10, extremely large (maximal). For relaxed sets, pain ratings were made before and after each stage. For contracting sets, pain ratings were made during each contraction and during the rest intervals between contractions. Since ratings during and after contraction were similar (P > 0.05), only ratings made during contractions are reported.

Protocol

Figure 1A illustrates the general experimental design for all studies. Data for the elbow flexors and extensors were collected in three stages; before (baseline), during (pain), and after (recovery) an infusion. Stages were separated by 15 min rest intervals and within a study each stage was identical. Representative stages from study 1 and study 2 are illustrated in Fig. 1B and C, respectively.

Figure 1.

Experimental protocol
A, general experimental design. In each study data were collected in 3 stages; before (baseline), during (pain), and after (recovery) an infusion into biceps muscle (timing of infusion indicated by large vertical arrows). Data collection in each stage lasted ∼10 min and stages were separated by 15 min. Within each study, the protocol in each stage was identical. B, expanded time scale of a stage in study 1. First, motor cortex and corticospinal tract stimuli were delivered alternately (5 s interstimulus interval) with the muscles at rest (n= 8 subjects). Fifteen of each stimuli were applied. Second, subjects (n= 7) performed 6 contractions of the elbow flexors to 20% of the force of a maximal voluntary contraction (MVC). Motor cortical and corticospinal stimuli were delivered alternately (5 s interstimulus interval) during these contractions. C, expanded time scale of a stage in study 2. First, subjects (n= 6) performed 6 elbow flexion contractions to a target corresponding to the amount of biceps brachii EMG produced in a 20% MVC of the elbow flexors. Second, subjects performed 6 elbow extension contractions to a target corresponding to the amount of triceps brachii EMG produced in a 20% MVC of the elbow extensors. Motor cortical and corticospinal stimulation were delivered alternately (5 s interstimulus interval) during flexion and extension contractions. Study 3 (n= 6) was identical to study 2, except that the infusion of hypertonic saline was replaced with an infusion of isotonic saline.

Study 1 The first study assessed MEPs and CMEPs at rest and during contractions of the elbow flexors to a force target before, during, and after pain produced by infusion of hypertonic saline. Each stage had two parts (Fig. 1B). First, motor cortex and corticospinal stimulation were delivered alternately (5 s interstimulus interval) with the muscles at rest. A total of 15 MEPs and 15 CMEPs were collected in each stage. Subjects (n= 7) were continuously reminded to maintain complete relaxation of the muscles of the arm throughout the experiment. However, subsequent analysis of EMG showed that three subjects were unable to maintain relaxation across all stages. Hence, these data were excluded from further analysis. One of these subjects repeated the relaxed sets on another day but with complete relaxation and an additional subject performed only the part of this study with relaxed muscles. Thus the final analysis included data for six subjects. Second, subjects (n= 7) performed contractions of the elbow flexors (Fig. 1B). Ongoing elbow flexion force and a target of 20% of maximal elbow flexion force were displayed on an oscilloscope. Subjects performed six contractions to this target each held for 20 s. During contractions motor cortical and corticospinal tract stimulation were delivered alternately (5 s interstimulus interval). Contractions were separated by 40 s rest to avoid fatigue. A total of 15 MEPs and 15 CMEPs during contractions were collected in each stage.

Study 2 The second study assessed MEPs and CMEPs in the elbow flexors and extensors during elbow flexion and extension performed to an EMG target before (baseline), during (pain), and after (recovery) pain produced by infusion of hypertonic saline. Each stage had two parts (Fig. 1C). First, subjects (n= 6) performed contractions of elbow flexors. Rectified integrated EMG activity (time constant of 100 ms) from biceps was displayed on an oscilloscope with a target corresponding to the EMG produced during a 20% maximal voluntary contraction (MVC) of the elbow flexors. Subjects performed six contractions to this target each held for 20 s. Motor cortical and corticospinal tract stimulation were delivered alternately during these contractions (5 s interstimulus interval). Contractions were separated by 40 s rest. A total of 15 MEPs and 15 CMEPs were collected during elbow flexion in each stage. Second, subjects performed contractions of the elbow extensors. EMG from triceps was displayed on the oscilloscope with a target corresponding to the amount of EMG produced during a 20% MVC of the elbow extensors. Subjects performed six contractions (40 s apart) during which stimuli were delivered. A total of 15 MEPs and 15 CMEPs were collected during elbow extension in each stage.

Study 3 A control study assessed MEPs and CMEPs before, during, and after infusion of isotonic saline. This study was identical to study 2 (and involved the same subjects; n= 6), except that the infusion of hypertonic saline was replaced with an infusion of isotonic saline.

Data analysis

The areas of MEPs and CMEPs are reported, although similar results occurred if the peak-to-peak amplitudes were used. Areas were measured between cursors appropriately positioned for each potential. The cursors encompassed a region from the initial deflection from the horizontal axis to the second crossing of the horizontal axis (Fig. 1D in Martin et al. 2006a; see also Martin et al. 2006b). For each subject, mean areas for each stage were calculated. MEPs and CMEPs were normalized to their mean value prior to the infusion (i.e. at baseline). MEPs are influenced by the excitability of both the motor cortex and motoneurones, whereas CMEPs are not influenced by cortical excitability (Taylor & Gandevia, 2004). Hence, to provide a measure of changes at the cortex, MEPs were also expressed relative to CMEPs (i.e. MEPs/CMEPs). Torque and root mean square (r.m.s.) EMG from each muscle were measured for a 50 ms window prior to each stimulus and a mean calculated for each subject for each stage.

Group data are presented as means ± standard deviation (s.d.) in the text and means ± standard error of the mean (s.e.m.) are shown in the figures (with n in the legends). All statistical analysis was performed on non-normalized values. One-way repeated-measures ANOVAs were performed (with Student–Newman–Keuls post hoc tests) to assess the differences between the three stages for the following variables: MEPs, CMEPs, MEP to CMEP ratios, torque, background EMG from each muscle, and pain intensity ratings. Student's t test for paired data compared the sizes of MEPs and CMEPs (normalized to Mmax) at baseline in each experiment. Statistical significance was set at P < 0.05.

Results

Study 1

Relaxed muscles Stimuli were delivered to the motor cortex and the corticospinal tract and EMG responses were recorded in relaxed muscles before (baseline), during (pain), and after (recovery) pain produced by infusion of hypertonic saline. At baseline, there were no significant differences in the sizes of MEPs compared to CMEPs for biceps (7 ± 4%Mmax and 13 ± 10%Mmax, respectively) or triceps (2 ± 2%Mmax and 2 ± 1%Mmax, P > 0.05). The infusion of hypertonic saline into biceps muscle increased the level of pain from zero at baseline to 4 ± 2 (‘considerable’, P < 0.001). Pain returned to zero before the recovery stage for all subjects. Figure 2A shows typical superimposed MEPs and CMEPs for one subject for biceps before, during, and after pain. For this subject, the area of MEPs decreased slightly during pain compared to baseline, but CMEPs increased dramatically in size during pain (Fig. 2B). For the group of subjects, the sizes of CMEPs increased by 47 ± 40% (P < 0.05) during pain compared to baseline and some of this increase persisted in recovery (by 16 ± 46%, P < 0.05; Fig. 3A). MEPs did not change among the three stages (baseline, pain, recovery, P > 0.05), but relative to CMEPs, they decreased by 34 ± 23% during pain (P < 0.05; Fig. 3A). A similar pattern of change occurred for triceps. CMEPs increased during pain (by 56 ± 57%, P < 0.05; Fig. 3B) and remained larger in recovery (by 30 ± 48%, P < 0.05). MEPs did not change among the three stages but relative to CMEPs they were reduced during pain (by 43 ± 15%, P < 0.05; Fig. 3B).

Figure 2.

EMG responses in relaxed biceps of a typical subject
Responses are shown before (baseline), during (pain), and after (recovery) activation of group III and IV muscle afferents produced by infusion of hypertonic saline into biceps. A, single subject data showing superimposed EMG responses in relaxed biceps to stimulation of the motor cortex (MEPs; upper traces) and the corticospinal tract (CMEPs; lower traces). Dashed horizontal lines indicate peak responses. B, corresponding data for the same subject plotted as areas of MEPs (○) and CMEPs (•). During the infusion of hypertonic saline, the average pain intensity rating for this subject was 4 out of 10 (‘moderate’ to ‘considerable’) and this rating returned to 0 before the recovery stage.

Figure 3.

Group data showing normalized EMG responses in relaxed biceps and triceps
Responses are shown before (baseline), during (pain), and after (recovery) activation of group III and IV muscle afferents produced by infusion of hypertonic saline in biceps. Mean area (±s.e.m.; n= 6 subjects) of responses to stimulation of the motor cortex (MEP, ○) and the corticospinal tract (CMEP, •), expressed as a proportion of the mean response prior to the infusion (i.e. at baseline). The ratio of the MEP to CMEP is also shown (MEP/CMEP; ▵). A, responses evoked in biceps. B, responses in triceps. An infusion of hypertonic saline produced a mean pain intensity rating of 4 out of 10 (‘considerable’) which returned to 0 before the recovery sets. #P < 0.05, CMEP different to baseline. ‡P < 0.05, MEP/CMEP different to baseline.

Constant-force flexor contractions Responses were evoked during elbow flexion contractions to a force target before, during, and after pain. Flexion force was stable and differed by < 1.3% in contractions performed across the three stages. During pain, the amount of EMG produced in biceps to maintain the target force decreased by 14 ± 25% (P < 0.05; Fig. 4A, dashed line).

Figure 4.

Group data showing normalized EMG responses in biceps and triceps during elbow flexion to a force target
Responses are shown before (baseline), during (pain) and after (recovery) activation of group III and IV muscle afferents produced by infusion of hypertonic saline in biceps. All responses were evoked during contractions performed to a force target of 20% MVC of the elbow flexors. Mean area (±s.e.m.; n= 7 subjects) of responses to stimulation of the motor cortex (MEP, ○) and the corticospinal tract (CMEP, •), expressed as a proportion of the mean response prior to the infusion (i.e. at baseline). A, responses evoked in biceps. The dashed line shows the mean r.m.s. EMG during contraction to the target force. B, responses in triceps. Infusion of hypertonic saline produced a mean pain intensity rating of 4 out of 10 (‘considerable’) which returned to 0 before the recovery sets. *P < 0.05, MEPs different to baseline. §P < 0.05, EMG different to baseline.

At baseline, there were no significant differences in the sizes of MEPs compared to CMEPs in biceps (14 ± 4%Mmaxversus 10 ± 3%Mmax, respectively) or triceps (2 ± 1%Mmaxversus 2 ± 1%Mmax, P > 0.05). Infusion of hypertonic saline into biceps increased pain from zero at baseline to 4 ± 1 (‘considerable’, P < 0.001). Pain returned to zero before the recovery stage for all but one subject in whom pain was ‘just noticeable’ (0.5) during some contractions. During pain, the sizes of MEPs decreased for biceps (by 15 ± 24%, P < 0.001; Fig. 4A) and triceps (by 15 ± 21%, P < 0.05; Fig. 4B) whereas CMEPs did not change across the stages for either muscle (P > 0.05; Fig. 4).

Study 2

Constant-EMG flexor contractions For flexor contractions to a constant force, biceps EMG decreased during pain. This could be due to the withdrawal of voluntary descending drive to motoneurones (i.e. disfacilitation) causing a change in net excitation to the motoneurone pool. Under those conditions it is difficult to assess the direct influence of group III and IV muscle afferents on motoneurones. For example, facilitation of motoneurones could be masked by the withdrawal of descending drive. Thus in the second study we investigated changes in MEPs and CMEPs evoked during elbow flexion contractions to a constant level of biceps EMG. Biceps EMG was stable and differed by < 1.4% in contractions performed across the three stages. Although not deliberately matched, triceps EMG was also similar across stages (P > 0.05).

At baseline, there were no significant differences in the sizes of MEPs compared to CMEPs in biceps (15 ± 5%Mmaxversus 14 ± 5%Mmax, respectively) or triceps (4 ± 2%Mmaxversus 3 ± 2%Mmax, P > 0.05). Infusion of hypertonic saline into biceps increased the pain from zero at baseline to 4 ± 2 (‘considerable’; P < 0.001). Pain returned to zero before the recovery stage for all subjects except for one subject who reported ‘just noticeable’ (0.5) pain during some contractions. Figure 5A shows typical superimposed MEPs and CMEPs for one subject for biceps during elbow flexion performed before, during, and after pain. For this subject, CMEPs but not MEPs increased during pain (Fig. 5C, circles). For the group, CMEPs increased in size in biceps (by 33 ± 30%, P < 0.05; Fig. 6A) and triceps (by 26 ± 25%, P < 0.001; Fig. 6B) during pain. MEPs did not change across the three stages for either muscle (P > 0.05); however, relative to CMEPs they were reduced for biceps (by 20 ± 36%, P < 0.05; Fig. 6A) and triceps during pain (by 17 ± 23%, P < 0.05; Fig. 6B).

Figure 5.

EMG responses in biceps and triceps of a typical subject during elbow flexion or extension to an EMG target
Responses are shown before (baseline), during (pain), and after (recovery) activation of group III and IV muscle afferents produced by infusion of hypertonic saline into biceps. All responses were evoked during contractions performed to EMG targets corresponding with the EMG produced during contractions of 20% MVC in either flexion or extension. A and B, single subject data showing superimposed EMG responses to stimulation of the motor cortex (MEP; upper traces) and the corticospinal tract (CMEP; lower traces). Dashed horizontal lines show peak responses. C and D, corresponding data for the same subject plotted as areas of MEPs (○) and CMEPs (•). □ and ▪ show MEP and CMEP areas for the same subject during an infusion of isotonic saline in a control experiment. A and C, MEPs and CMEPs in biceps during elbow flexion. B and D, responses in triceps during elbow extension. During the infusion of hypertonic saline, the mean pain intensity rating for this subject was 3 out of 10 (‘moderate’ pain) during both flexion and extension contractions and this returned to 0 before the recovery set. This subject reported no pain during infusion of isotonic saline.

Figure 6.

Group data showing normalized EMG responses in biceps and triceps during elbow flexion or elbow extension to an EMG target
Responses are shown before (baseline), during (pain) and after (recovery) activation of group III and IV muscle afferents produced by infusion of hypertonic saline. All responses were evoked during contractions performed to EMG targets corresponding with the EMG produced during contractions of 20% MVC in either flexion or extension for each subject. Mean area (±s.e.m.) of responses to stimulation of the motor cortex (MEP, ○) and the corticospinal tract (CMEP, •), expressed as a proportion of the mean response prior to the infusion (i.e. at baseline). The ratio of the MEP to the CMEP is also shown (▵). Upper and lower panels are responses during flexion and extension, respectively. A and B, MEPs and CMEPs in biceps and triceps during elbow flexion (n= 6). C and D, responses in biceps (n= 5) and triceps (n= 6) during elbow extension. An infusion of hypertonic saline (○, •, ▵ and continuous lines) produced a mean pain intensity rating of 4 out of 10 (‘moderate’ pain) which returned to 0 before the recovery sets. A control infusion of isotonic saline (n= 6; □, ▪ and dashed lines) produced no perceived pain and did not alter MEPs or CMEPs. *P < 0.05, MEPs different to baseline. #P < 0.05, CMEPs different to baseline. ‡P < 0.05, MEPs/CMEPs different to baseline.

Constant-EMG extensor contractions Activity in group III and IV muscle afferents maintained by ischaemia following exercise facilitates flexor motoneurones but inhibits extensor motoneurones (Martin et al. 2006b). Thus to determine whether the actions of afferents activated by hypertonic saline differed for extensors and flexors, responses were also evoked during elbow extension contractions to a constant level of triceps EMG. Triceps EMG was stable and differed by < 3.5% in contractions performed across the three stages. Although not deliberately matched, biceps EMG was also similar across stages (P > 0.05).

At baseline, there were no differences between the sizes of MEPs and CMEPs for triceps (13 ± 8%Mmaxversus 11 ± 9%Mmax, P > 0.05), but for biceps, MEPs were smaller than CMEPs (4 ± 4%Mmaxversus 34 ± 13%Mmax, P < 0.01). For the typical subject in Fig. 5 (B and D, circles), there was a progressive increase in the area of triceps CMEPs across the three stages whereas there was a small decrease in MEPs during pain. For the group, triceps CMEPs increased during pain (by 22 ± 32%, P < 0.05) compared to baseline and this persisted in recovery (P < 0.05; Fig. 6D). Biceps CMEPs also tended to be larger during pain, but because there were fewer subjects (see Methods) this increase did not reach significance (P= 0.181; Fig. 6C). MEPs in triceps were reduced in size during pain (by 15 ± 37%, P < 0.05; Fig. 6D) but no change occurred for biceps (P > 0.05; Fig. 6C).

Study 3

Control infusion of isotonic saline To ensure that differences among stages were due to muscle pain per se, study 2 was repeated but with an infusion of isotonic saline, which induces minimal pain. There were no differences in biceps or triceps EMG between the three stages for either flexion or extension contractions (P > 0.05). During initial baseline sets, there were no significant differences between the sizes of MEPs and CMEPs in either biceps or triceps for flexion or extension contractions (P > 0.05). In addition, baseline MEPs and CMEPs for both muscles were similar in size to those evoked at baseline in study 2 (P > 0.05). The infusion of isotonic saline evoked no pain, except for one subject who reported ‘just noticeable’ pain during some contractions. There were no changes in MEPs or CMEPs for either muscle across the three stages during flexion (Figs 5A and C, and 6A and B, squares and dashed lines) or extension contractions (P > 0.05; Figs 5B and D, and 6C and D, squares and dashed lines).

Discussion

These studies provide new data on the central actions of group III and IV muscle afferents in humans. Using stimulation of the corticospinal tract to test motoneurone excitability we have demonstrated that group III and IV muscle afferents activated by hypertonic saline produce a uniform effect on flexor and extensor motoneurone pools. The uniform facilitation of flexors and extensors differs from the pattern predicted for flexor reflex afferents in animals (FRAs) and from the pattern which we have previously demonstrated for group III and IV muscle afferents in humans (Martin et al. 2006b). Furthermore, we argue that facilitation of CMEPs induced by intramuscular hypertonic saline during contractions performed to a matched level of EMG is most likely due to preferential excitation of high-threshold motoneurones by group III and IV muscle afferents. Finally, based on comparison of responses to corticospinal and motor cortical stimulation we argue that motor cortical output cells are inhibited by these afferents.

Motoneuronal excitability was tested by stimulation of the corticospinal tract. Corticospinal synapses lack presynaptic inhibition (e.g. Nielsen & Petersen, 1994; Jackson et al. 2006) and, for biceps, a significant proportion of the response is monosynaptic (Petersen et al. 2002). Although corticospinal connections to triceps are weaker, they include a monosynaptic corticomotoneuronal connection (Maertens de Noordhout et al. 1999). It is unlikely that hypertonic saline caused changes in the periphery to alter the compound muscle action potential (e.g. Farina et al. 2005; Qerama et al. 2005). Hence, changes in responses to this stimulus probably reflect effects at a motoneuronal level. The increase in CMEPs for relaxed biceps and triceps with hypertonic saline is therefore consistent with increased excitability of motoneurones.

It is well established that hypertonic saline activates many group III and IV muscle afferents, whereas direct effects on muscle spindles are rare (e.g. Paintal, 1960; Kumazawa & Mizumura, 1977; Thunberg et al. 2002). However, several studies have shown that activity in small-diameter muscle afferents can increase or decrease fusimotor discharge. Thus, hypertonic saline may indirectly influence Ia afferent behaviour (e.g. Ro & Capra, 2001; Thunberg et al. 2002). Although, it is possible that excitatory input from muscle spindles during pain could increase motoneurone excitability, the specific pattern of changes we observed during pain in relaxation and contraction argues against this. Inputs from Ia afferents generate larger EPSPs in low-threshold than in high-threshold motoneurones, an effect which increases the range of recruitment thresholds within the pool beyond the typical 11-fold range caused by variations in intrinsic motoneurone properties (e.g. Burke & Rymer, 1976; Heckman & Binder, 1988, 1993). In contrast, our results indicate that the additional excitatory input to motoneurones during pain is probably greater for higher-threshold motoneurones and that this causes compression, rather than expansion of the range of thresholds within the motoneurone pool (see below). Therefore, the increase in motoneuronal excitability revealed by corticospinal stimulation most likely relates to an excitatory input from group III and IV muscle afferents.

Uniform effect on flexor and extensor motoneurone pools

Excitation of flexor motoneurones is consistent with the effect produced by activity in group III and IV muscle afferents maintained by ischaemia after exercise (Martin et al. 2006b) and with the major actions of FRAs (e.g. Eccles & Lundberg, 1959). However, the excitation of extensor motoneurones is unexpected. Afferents belonging to the FRA group may also use separate ‘private’ pathways (e.g. Holmqvist & Lundberg, 1961). For example, an excitatory nociceptive non-FRA pathway from the foot pad to extensors of the foot exists in the cat (Schmidt et al. 1991; Schomburg et al. 2001). Our results suggest that humans have both excitatory and inhibitory pathways from group III and IV muscle afferents to extensor motoneurones. In cats, excitatory and inhibitory effects on motoneurones can be mediated by the same afferents with transmission controlled by descending commands (Holmqvist & Lundberg, 1961; Schmidt et al. 1991; Schomburg & Steffens, 1998). This may also occur in humans. Alternatively, fatigue and hypertonic saline may activate different subsets of group III and IV muscle afferents which predominantly access the excitatory or inhibitory paths. For example, changes at motoneurones during activation of nociceptive afferents may be mediated by group III muscle afferents (Schomburg & Steffens, 2002), whereas group IV muscle afferents are probably responsible for motoneuronal effects during maintained ischaemia following exercise (Martin et al. 2006b; Kindig et al. 2007). Alternatively, the behaviour of motoneurones within the pool may differ. In our study of fatigue, we used large CMEPs (∼50%Mmax) during MVCs (Martin et al. 2006b) whereas here small CMEPs (< 15%Mmax) were elicited during relaxation and weak contractions and hence we have assessed different parts of the motoneurone pool.

The facilitation of CMEPs often persisted several minutes after the pain sensation dissipated. This concurs with previous reports of enduring central changes following nociceptive stimulation (e.g. Wall & Woolf, 1984; Le Pera et al. 2001; Svensson et al. 2003). Flexor and extensor motoneurones show long-lasting effects after chemical activation of group III and IV muscle afferents in decerebrate cats (Kniffki et al. 1981). For example, the duration of responses induced by a second injection of potassium chloride is longer compared to the first injection, and the response is enhanced. Similarly, activation of nociceptors by brief conditioning stimuli facilitates the flexor withdrawal reflex in spinal rats, lasting for up to 90 min (e.g. Wall & Woolf, 1984; Woolf & Wall, 1986). Moreover, inputs via muscle afferents are more effective than those via cutaneous afferents at inducing prolonged excitability changes (Wall & Woolf, 1984). These effects are not due to changes in afferent terminal or motoneurone excitability (Cook et al. 1986), and are likely to be due to changes in interneurones (Wall & Woolf, 1984; Woolf & Wall, 1986), probably through long-term synaptic potentiation (for review see Sandkühler, 2000). Therefore it is possible that the enduring facilitation of CMEPs is an indirect result of the operation of pain-induced long-term plasticity.

Non-uniform effect across the motoneurone pool

CMEPs were also evoked during contractions to constant levels of force or EMG. In constant-force contractions, voluntary EMG decreased and CMEPs were unchanged by hypertonic saline whereas CMEPs were facilitated in constant-EMG contractions. This suggests changes in the input–output properties of the motoneurone pool. The ability of a motoneurone to generate an action potential depends on both its intrinsic excitability and the sum of its synaptic inputs (for review see Hultborn et al. 2004). An additional excitatory input to the motoneurone pool should alter recruitment gain if its distribution across the motoneurones of different intrinsic excitability differs from the distribution of synapses that are already active (Kernell & Hultborn, 1990). Hence, delivery of more excitation to high- than low-threshold motoneurones would decrease differences in recruitment threshold within the pool and increase recruitment gain by threshold compression (Burke, 2004). This may occur for inputs to motoneurones from skin afferents or descending tracts (e.g. Burke et al. 1970; Garnett & Stephens, 1981; Heckman & Binder, 1993; Nielsen & Kagamihara, 1993). For example, the rubrospinal system generates ∼5 times as much synaptic current in high- as in low-threshold motoneurones (Powers et al. 1993), and can reduce the range of recruitment thresholds from ∼11-fold to < 2-fold (Heckman & Binder, 1993).

This fits our findings during pain. For contractions in which EMG output was constant, the total synaptic input to the pool probably remained unchanged. Thus, since group III and IV muscle afferents provide an additional excitatory input in the painful condition (indicated by the facilitation of motoneurones at rest), other excitatory input, presumably from descending pathways, must be reduced to maintain constant output. As total excitation to the motoneurone pool was adjusted voluntarily to match the required EMG, output evoked by the corticospinal stimulus should also be similar in painful and non-painful conditions. However, despite the adjustment, CMEPs were larger during tonic pain. This implies that the additional excitatory input from group III and IV muscle afferents has a different distribution to the pool than the descending drive that is withdrawn, and that this excitatory input is more pronounced for the extra, higher-threshold motoneurones recruited into the CMEP than those activated voluntarily during contraction. Findings for contractions in which force was kept constant are consistent with this. CMEPs were not reduced during pain despite reduced EMG. If the distribution of excitation from group III and IV muscle afferents were similar to the synaptic influences already acting on the motoneurones at baseline, the net reduction in synaptic inputs should also have reduced CMEPs.

A compression of thresholds across the motoneurone pool suggests that similar EMG output in painful conditions is produced by different firing of motoneurones. Additional motoneurones should be recruited and motoneurones should fire more slowly. Reduced firing rates have been recorded in painful contractions (e.g. Sohn et al. 2000; Farina et al. 2004). Changes in firing rates will in turn influence the responsiveness of the motoneurone pool to corticospinal inputs. Because recovery of motoneuronal excitability after an action potential slows as firing rates decrease, the probability that a stimulus elicits an action potential is greater if the membrane potential approaches threshold slowly (Matthews, 1999; Martin et al. 2006a). Hence, any slowing of firing rates would contribute to increase CMEPs during contraction.

Differential effects at the motoneurones and the motor cortex

Whereas group III and IV afferents facilitated motoneurones, they inhibited the motor cortex. The size of MEPs depends on the excitability of cortical and spinal circuits. Hence, while MEPs generally did not change upon activation of group III and IV muscle afferents produced by hypertonic saline, facilitation at the spinal level suggests that the excitability of cortical output cells must have decreased substantially to produce no change in MEPs. To account for changes at the spinal level we compared the ratios of MEPs to CMEPs. MEPs were reduced (relative to CMEPs) for biceps and triceps in relaxation or during flexion or extension. For contractions to a constant EMG, the increase in motoneuronal excitability suggests that fewer motor cortical cells would be activated by volition to obtain the EMG and this would contribute to reduced cortical excitability. However, the reduction in the ratio of MEPs to CMEPs in relaxation confirms that group III and IV muscle afferents must also inhibit cortical output cells.

Previous studies have shown that hypertonic saline-induced firing of group III and IV afferents decreases MEPs in relaxation (Le Pera et al. 2001) but increases them during contraction (Del Santo et al. 2007). Although this suggests that these afferents influence motor cortical excitability, the changes are impossible to interpret without understanding the behaviour of motoneurones under these conditions. In one study, changes in MEPs were compared to changes in H-reflexes. Suppression of MEPs was accompanied by reduced H-reflexes and hence it was concluded that at least some of the reduction in MEPs was due to decreased motoneuronal excitability (Le Pera et al. 2001). However, as the H-reflex change may reflect altered presynaptic inhibition (e.g. Rossi et al. 1999), it was difficult to determine the changes that occurred at the motoneurones and hence, at the cortex.

The current studies show that group III and IV muscle afferents activated by hypertonic saline produce opposite effects on motor cortex and motoneuronal outputs. Furthermore, the uniform excitatory effect on different motoneurone pools differs from the flexor excitation and extensor inhibition that characterizes the group III and IV muscle afferent input maintained by ischaemia after exercise. It also differs from the pattern predicted for flexor reflex afferents. Finally, the excitation of motoneurones by these inputs appears differentially directed to higher threshold motoneurones.

Appendix

Acknowledgements

This work was supported by the National Health and Medical Research Council of Australia. We are grateful to Dr Peter Nickolls for assistance with the experiments.

Ancillary