- Email: [email protected]
The influence of experimental muscle pain on the human soleus stretch reflex during sitting and walking
Clinical Neurophysiology 110 (1999) 2033±2043 www.elsevier.com/locate/clinph
The in¯uence of experimental muscle pain on the human soleus stretch re¯ex during sitting and walking Dag®nn A. Matre a,b,*, Thomas Sinkjñr a, Stein Knardahl b, Jacob B. Andersen a, Lars Arendt-Nielsen a b
a Center for Sensory-Motor Interaction, Aalborg University, DK-9220 Aalborg E, Denmark Department of Physiology, National Institute of Occupational Health, N-0033 Oslo, Norway
Accepted 13 August 1999
Abstract Objectives: The stretch re¯ex is functionally important during human locomotion. Muscle pain has been found to increase the stretch re¯ex amplitude during sitting, possibly due to an altered fusimotor drive. To further study the importance of altered fusimotor activity due to muscle pain we investigated the combined effect of muscle pain and motor task on the soleus stretch re¯ex. Methods: Stretch re¯exes were elicited before, during and after experimentally induced muscle pain in soleus (i.m. infusion of 6% saline) in 3 experiments: (1) in the relaxed soleus muscle and before, during and after an isometric ramp contraction (500 ms, 0±10 Nm), (2) at 3 different time periods during walking, and (3) at matched pain intensity and soleus activity during sitting and walking. Results: Infusion of hypertonic saline into the soleus muscle caused a signi®cant facilitated stretch re¯ex in the relaxed muscle (P , 0:01), but not during walking or during sitting and walking at matched soleus EMG and matched pain levels. The infusion of isotonic saline (nonpainful) did not cause any changes (P 0:75). Conclusions: The main ®ndings of the present study were that experimental muscle pain facilitated the stretch re¯ex during pain in the relaxed muscle, but caused no changes in stretch re¯ex amplitude during sitting and walking at higher ``functional'' background EMG levels. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Experimental muscle pain; Stretch re¯ex; Motor control; Walking; Soleus muscle
1. Introduction The knowledge of the neural mechanisms leading to, and maintaining chronic musculoskeletal pain and occupational muscle pain is limited. Animal studies have shown that injection of various pain-producing substances into the muscle may activate the g -system (Schmidt et al., 1981; Ellaway et al., 1982; Appelberg et al., 1983; Jovanovic et al., 1990; Mense and Skeppar, 1991; Pedersen et al., 1997). There is an excitatory connection between muscle spindle afferents and homonymous motoneurones (Schomburg, 1990), and it has been suggested that an elevated g -activity, caused by muscle pain, facilitates this excitatory connection producing a higher electromyographic activity in painful muscles (Schmidt et al., 1981; Johansson and Sojka, 1991). Clinical and experimental studies of musculoskeletal pain and electromyographic activity (EMG) have not found any support for this hypothesis, neither during relaxed, static * Corresponding author. Tel.: 147-23-195-267; fax: 147-23-195-204. E-mail address: dag®[email protected] (D.A. Matre)
nor dynamic conditions (Lund et al., 1991; Arendt-Nielsen et al., 1996; Svensson et al., 1997; Graven-Nielsen et al., 1997c; Svensson et al., 1998b). However, in a recent human study we found that the soleus stretch re¯ex, but not the H re¯ex, was facilitated by experimentally induced muscle pain (Matre et al., 1998). This gives indirect support to a link between muscle pain and increased spindle sensitivity. Since the stretch re¯ex is functionally important in the control of joint stiffness (Sinkjñr, 1997) we wanted to investigate the possible effect of muscle pain on the re¯ex amplitude during different motor tasks. In the present study we compared the soleus stretch re¯ex in the relaxed muscle, in the active muscle during isometric ramp contractions, during walking, and during sitting and walking with matched pain intensity and matched ``motoneurone excitability'' as it can be measured from soleus background EMG (Capaday and Stein, 1986). The short latency stretch re¯ex amplitude was compared in each of the four situations before, during and after inducing experimental muscle pain, produced by intramuscular infusion of hypertonic saline into the soleus muscle. The following
1388-2457/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 1388-245 7(99)00211-4
CLINPH 99532
2034
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
questions were asked: is the stretch re¯ex modulation during ramp contractions, static contractions and walking during pain different from the relaxed state? If so, can the modulation be related to changes in g -activity? In the cat the g activity has been found to be dependent on motor task (Prochazka et al., 1988). 2. Materials and methods 2.1. Subjects and experimental conditions The investigation was divided into three experiments, performed on three different groups of healthy volunteers. They were recruited from university students after announcements in an internal newspaper. All subjects signed an informed consent form and the study was approved by the Local Ethical Committee and was conducted in accordance with the Helsinki Declaration. The subjects abstained from coffee and exercise prior to the investigation, and no subjects had any history of chronic musculoskeletal disorders. 2.1.1. Ramp experiment Ten healthy male subjects (age range 20±24 years) participated in an experiment where stretch re¯exes were elicited both in the relaxed soleus and at 6 regular periods before, during and after an isometric ramp contraction (500 ms). The ramp was similar in length to the rising phase of the soleus EMG in the stance phase during walking. Recordings were made before, during and after experimental muscle pain was produced in soleus by intramuscular infusion of hypertonic saline. Isotonic saline was used as control in a separate experiment separated by at least 5 days. 2.1.2. Walking experiment Fourteen healthy male subjects (age range 20±30 years) participated in an experiment where stretch re¯exes were elicited in soleus at three time periods of the stride while the subject was walking on a treadmill (walking speed 3.5± 4.5 km/h). The time periods corresponded to early stance phase, late stance phase and mid swing phase. Recordings were made, before, during and after experimental muscle pain was produced in soleus by intramuscular injections of hypertonic saline. 2.1.3. Walking and sitting with matched EMG One female (age 25) and seven male subjects (age range 18±25 years) participated in the experiment where stretch re¯exes were elicited in soleus in early stance phase (100± 200 ms after heel contact), and while sitting with matched background EMG in soleus. Recordings were made before, during and after injections of hypertonic saline. Four subsequent saline injections were made, and after each injection walking and sitting recordings were done in random order.
2.2. Pain stimulus Sterile hypertonic saline (6%) was infused to produce acute deep pain in the soleus muscle. In the relaxed and ramp conditions, the infusion was performed by a computer-controlled syringe pump (IVAC model 770, UK) with a 10 ml plastic syringe. A tube (IVAC G30303) was connected from the syringe to a catheter (VENFLON, 22G, 25 mm) (Graven-Nielsen et al., 1997d). The catheter was inserted into the soleus muscle, on the medial side, before the pre-infusion recording and remained in the muscle for the rest of the recordings. Using a computer controlled paradigm, 7.1 ml of saline was infused over a period of 15 min. The infusion rate was adjusted during the period to produce as constant a pain sensation as possible (90 ml/h for 20 s, 18 ml/h for 440 s, and 36 ml/h for 440 s) (Graven-Nielsen et al., 1997d; Matre et al., 1998). The subject rated the pain intensity on a 10 cm electronic visual analogue scale (VAS) and the ratings were stored on computer. The lower endpoint of the scale was marked ``no pain'' and rated `0', and the upper endpoint was marked ``most pain imaginable'' and rated `10'. Isotonic saline (0.9%, non-painful) was used as a control infusion in the ramp experiment. In the walking experiment three subsequent bolus injections of hypertonic saline (6%) were made. After each injection recordings were made for one time period. The injected volumes were 0.5, 0.6 and 0.7 ml (injected over 20±25 s) in the three time periods respectively. The locations of injection were separated by 2 cm (Graven-Nielsen et al., 1997b). Each pain stimulus lasted for approximately 5 min, and the pain intensity rating was given orally by the subject every 30 s as a number between 0 and 10. In the combined walking and matched EMG experiment four subsequent bolus injections of hypertonic saline (6%) were made. The injected volumes were 0.5, 0.6, 0.7 and 0.8 ml (injected over 20±30 s) respectively, and separated by 2 cm in location. The pain lasted for approximately 5 min, and the pain intensity was rated on an electronic visual analogue scale (VAS) and the ratings were stored on computer. 2.3. Experimental set-up 2.3.1. Relaxed and ramp experiment The subject was seated in a chair with the left foot strapped to a platform. Knee and ankle joints were extended to an angle of approximately 1008. Stretch-re¯exes were elicited in soleus by applying a 48 dorsi¯exion of the foot. The stretch had a rise time of 50 ms. The dorsi¯exion was made by rotating the platform with a motor (CEM, model 26) where the axis of rotation of the ankle joint was aligned with the axis of rotation of the platform. The motor was driven by a DC power ampli®er (BruÈel and Kjñr, model 2708). For further details, see Sinkjñr et al. (1988). The subject was trained to do isometric ramp contractions with
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
the ankle extensor muscles. Recti®ed and low-pass ®ltered soleus EMG was monitored on an oscilloscope and used as feedback for the subject. The ramp rose linearly (rise time 500 ms) to a soleus EMG level corresponding to approximately 10 Nm ankle torque. The ramp was drawn on the screen of an oscilloscope and the subject was trained to match the ramp as the monitored EMG swept across the screen. Stretches were elicited at 6 time points relative to the ramp start: 150 (before ramp onset), 0, 150, 300, 500 and 600 (after ramp) ms. In addition, stretch re¯exes were elicited in the relaxed muscle. The timing was triggered by the onset of the voluntary EMG. The trials were randomly alternated between trials with perturbation (ankle dorsi¯exion) and trials without perturbation. Recordings were made before infusion, during infusion, 20 min after and 40 min after the pain had disappeared. All subjects participated twice, receiving hypertonic and isotonic saline (control) in random order. The two sessions were separated by at least 1 week.
2035
subject rested in 5 min before the procedure was repeated with another time period, selected at random. 2.3.3. Walking and sitting with matched EMG The subjects were walking on a treadmill, and re¯exes were elicited as described above. The soleus EMG level 30 ms prior to the perturbation was noted. While the subject was seated re¯exes was elicited when an isometric contraction at the same strength (matched EMG) was held. The setup enabled equal stretch velocity, amplitude, ankle joint position, and pain intensity during walking and sitting, in addition to the matched EMG level. One series of walking and sitting measurements was made before pain, four series during pain, and one series after pain. Half of the subjects started with walking, the other half started with sitting. In each series 15 measurements were done walking, and 15 measurements were done sitting. 2.4. Data collection
2.3.2. Walking experiment The subjects were walking on a treadmill (Powerjog EG30). Stretch re¯exes were elicited in soleus by a semiportable stretch device attached to the subject's left foot performing a 88 dorsi¯exion of the foot with a velocity of approximately 3008/s (Andersen and Sinkjñr, 1995). The device, constructed as a functional joint, followed the ankle rotation during locomotion and was connected to a motor by means of bowed wires. The motor was regulated in such a way that it followed the movement of the ankle joint without in¯uencing the pattern of gait. During any time of the gait cycle it was possible to evoke a muscle stretch of the ankle extensors by rotating the ankle joint. The device was attached to a carbon ®bre shell which was strapped to the lower leg. For further details about the stretch device see Andersen and Sinkjñr (1995). The subject was walking for a few minutes prior to the recordings to adapt to the set-up and was asked to walk with a comfortable speed, usually between 3.5 and 4.5 km/h. The same speed was used in the rest of the experiment. During the experiment stretches were evoked for every 4±5 strides. For every stretch response recorded, the preceding stride was also recorded. The stride duration was measured from one heel contact to the succeeding. Each stride was divided into three time periods: (1) between heel down and heel off, (2) between heel off and toe off, and (3) between toe off and heel down. Stretches were elicited within the respective periods with the following delays from heel down: time period 1, 100± 200 ms; time period 2, 450±550 ms; time period 3, 900± 1100 ms. Thirty stretches were evoked in each time period. In the pre-pain and post-pain sessions stretches were evoked in all three segments successively. For the recordings during pain, the subject started walking at pain onset and stretches were evoked as long as the pain lasted, or for maximum 30 recordings. After one recording session with pain, the
EMG signals were recorded from the soleus and tibialis anterior muscles by bipolar Ag±AgCl surface electrodes (Medicotest; 2 cm apart). The skin was abraded, lightly rubbed with sand paper and cleaned with isopropanol before attaching the electrodes. The EMG signals were ampli®ed (DISA, model 15C01), ®ltered with a band pass ®lter (20 Hz to 2 kHz), sampled (4 kHz) and stored on computer. In the walking experiment optical encoders measured the actual position of the ankle joint. The position was measured, sampled and stored on a computer. In the ramp experiment a potentiometer measured the ankle position. For further details regarding the data collection of the walking experiment see Sinkjñr et al. (1996). 2.5. EMG analysis The perturbation resulted in a short latency stretch re¯ex in all subjects. For each series of recordings background EMG and stretch re¯ex peaks were averaged. In the ramp experiment nine re¯exes were averaged for the relaxed condition and for each time slot. In the walking experiment the inclusion criteria (a stretch velocity between 275 and 3008/s) reduced the number of valid re¯exes to between 10 and 20 for each segment. In the combined walking and matched EMG experiment the inclusion criteria (matched background EMG) reduced the number of valid re¯exes from 15 to between 9 and 10. The stretch re¯ex amplitudes during and after infusion was normalised to the recordings before infusion. Average background EMG 100 ms (static experiments) or 30 ms (walking experiments) prior to stretch onset was subtracted from the peak amplitudes. Due to a more variable EMG the average background EMG was calculated in a shorter time period in the walking experiments than in the static experiments.
2036
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
2.6. Statistics In the ramp experiment the data were analysed using a repeated measures 3 way analysis of variance (ANOVA) where time points (relaxed, 2150, 0, 150, 300, 500 and 600), experimental conditions (before, during and after infusion) and saline concentrations (hypertonic and isotonic) were used as factors. In the walking experiment the data were analysed using a repeated measures 2 way ANOVA with time periods (1, 2 and 3) and experimental conditions (before, during and after injections) as factors. In the combined walking and matched EMG experiment the data were analysed using a repeated measures 2 way ANOVA with experimental conditions (before, during and after injections) and motor task (walking and sitting) as factors. A signi®cance level of P , 0:05 was considered statistically signi®cant in both experiments. If a main effect was found for one of the factors or between two factors a post-hoc test was performed using Student±Newman±Keul's method, thus identifying the factor causing signi®cance.
3. Results 3.1. Stretch re¯ex modulation in the relaxed muscle and during the ramp contraction Fig. 1 shows the stretch response after a perturbation (48 dorsi¯exion) given at ramp onset (dashed line).Fig. 2 shows the background EMG and stretch re¯ex modulation during the ramp before saline infusion. The average stretch re¯ex increased with .100% from the relaxed state to 150 ms before ramp onset. The hypertonic saline infusion produced a stable deep pain sensation distributed on the posterior side of the lower leg. The mean pain intensity was signi®cantly higher (Wilcoxon: Z 2:8, P 0:005) during the hypertonic infusion (5:5 ^ 2:4 cm on the VAS) than during the isotonic infusion (0:4 ^ 0:6 cm) (Fig. 3C). For all subjects the EMG included a short latency response with onset and peak latencies 44:4 ^ 2:4 ms and 64:2 ^ 6:4 ms, respectively (mean ^ SD; relaxed muscle). No subjects exhibited the medium latency stretch re¯ex responses which is seen in the active muscle during similar experiments (Toft et al., 1991). The normalised stretch re¯ex amplitude (Fig. 3B) was higher during compared with before infusion of hypertonic saline (3 way ANOVA and SNK: P 0:012). A signi®cant interaction was found between saline concentration, experimental condition and time points (ANOVA: d:f: 12, F 2:01, P 0:044). Post-hoc analysis showed that the re¯ex facilitation was signi®cantly higher in the relaxed muscle (SNK: P , 0:005; indicated with *), but not during the ramp. During the isotonic control infusion the normalised stretch re¯ex amplitude was lower than during the hypertonic infusion (3 way ANOVA and SNK:
Fig. 1. Example of recorded averaged data from control contraction (heavy lines) and contractions with a perturbation at ramp onset (thin lines). Average of 9 re¯ex responses from subject JP. (A) Ankle angle positions, (B) recti®ed and ®ltered soleus muscle EMG. Time 0 (dashed line): ramp start. Time 500: ramp end.
P 0:008). The stretch re¯ex was not different during compared with before the the isotonic infusion (SNK: P 0:75). The re¯ex amplitude was not signi®cantly different after the infusions, compared with before the infusions. No change was found in soleus background EMG (Fig. 3A) when comparing the painful condition with the pre-pain condition (ANOVA: d:f: 3, F 2:25, P 0:18) or when comparing the saline concentrations (ANOVA: d:f: 1, F 0:15, P 0:74). 3.2. Stretch re¯ex modulation during walking Stretch re¯exes were measured in two experiments during walking. In the ®rst walking experiment re¯exes were elicited in early stance phase, late stance phase and in the
Fig. 2. Stretch re¯ex modulation during ramp contraction before saline infusion (X) and soleus background EMG (B). Each point is an average of 90 stretches (10 persons £ 9 stretches; mean ^ SE).
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
2037
Fig. 3. (A) Stretch re¯ex modulation (mean ^ SE; n 10) in the relaxed muscle, before and throughout an isometric ramp contraction before, during, 20 min after, and 40 min after infusion of hypertonic (6%) saline (X) and isotonic (0.9%) saline (W). Soleus background EMG (®lled squares) in the relaxed muscle and throughout the ramp contraction from 10 subjects (mean ^ SE) before, during, 20 min after, and 40 min after infusion. (B) Stretch re¯ex amplitude normalised to values before infusion (0%). The amplitude was signi®cantly higher during the hypertonic infusion in the relaxed muscle (*P , 0:005). (C) Pain intensity (mean ^ SE; n 10) during infusion of hypertonic (6%) saline (X) and isotonic (0.9%) saline (W). Note that the abscissa have a different time scale compared with the abscissa in (A) and (B). The stretch re¯ex was elicited at the different time points in random order during the 900-s infusion period.
swing phase. An example of average and recti®ed soleus EMG is shown in Fig. 4 with control steps (heavy lines) and steps with a perturbation (thin lines) in early stance phase (dashed line). The 88 dorsi¯exion of the foot was re¯ected in the soleus EMG (Fig. 4B, thin line). One subject was excluded from the study because only a few stretch re¯ex responses were different from the background EMG. In time period 1 (early stance phase) the short latency stretch re¯ex was present in all 13 remaining subjects, in time period 2 (late stance phase) it was present in 12 of 13 subjects, and in time period 3 (mid swing phase) in 6 of 13 subjects. The medium latency stretch re¯ex was present in time period 1 (10 of 13 subjects) and in time period 2 (6 of 13 subjects).
Fig. 5A shows the stretch re¯ex modulation (X) and soleus background EMG (W) in early stance phase, late stance phase and mid swing phase before, during and after injection of hypertonic saline. The pain intensities after the hypertonic saline injections in periods 1±3 are shown in Fig. 5B (mean of 13 subjects). The experimentally induced muscle pain did not in¯uence the short latency stretch re¯ex in either time period throughout the step cycle (ANOVA: d:f: 2, F 0:57, P 0:59). Neither did the medium latency stretch re¯ex change with pain (ANOVA: d:f: 2, F 2:5, P 0:1). Furthermore, average background EMG during the active periods of the soleus (stance phase) and tibialis anterior muscle (swing phase) did not change during injection of hypertonic saline (ANOVA-
2038
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
Fig. 4. Example of recorded averaged data from control steps (heavy lines) and steps with a perturbation (thin lines). Superimposition of data from time periods 1±3 (control steps) and time period 1 (steps with a perturbation). Average of 30 step cycles from subject KA. The time shown represents a full step cycle. (A) Ankle angle positions, (B) recti®ed and ®ltered soleus EMG with short and medium latency stretch re¯ex response. Dashed line: start of perturbation.
soleus: d:f: 2, F 0:85, P 0:44; ANOVA-tibialis anterior: d:f: 2, F 2:83, P 0:09). Onset and peak latencies (41:6 ^ 5:0 ms and 53:9 ^ 5:6 ms, respectively, in early stance phase) were unchanged during the pain condition (ANOVA: d:f: 2, F 2:73, P 0:09), and neither did the peak latencies of the medium latency stretch re¯ex depend on the pain condition (ANOVA: d:f: 2, F 2:5, P 0:11). The stride time was signi®cantly reduced from 1:21 ^ 0:07 s (pre-pain) to 1:19 ^ 0:08 s (during pain) (ttest dependent samples: d:f: 13, t 2:45, P 0:03). In the second walking experiment re¯exes were elicited alternately in early stance phase during walking and during sitting. Soleus background EMG and pain intensity were matched in the two conditions. Fig. 6A shows the stretch re¯ex amplitudes during and after injections of hypertonic saline normalised to the measurements before injection for each subject and averaged. The re¯exes were elicited at matched background EMG (Fig. 6B; ANOVA: d:f: 1, F 0:35, P 0:57), ankle angle, and pain intensity (Fig. 6C) while the subject was seated (W) keeping a static isometric contraction, and walking (X). The pain intensity (mean of 8 subjects over 4 infusions) was 4:0 ^ 1:9 cm during walking which was not different from 4.2 ^ 2.0
Fig. 5. (A) Stretch re¯ex amplitude (X) and soleus background EMG (W) during walking from 13 subjects (mean ^ SE) before, during, and 20 min after infusion of hypertonic (6%) saline. Between 20 and 30 measurements are averaged for each subject in each time period. No change was found during the hypertonic saline infusion. (B) Pain intensity (mean ^ SE; n 13) during periods 1±3. Injections indicated with arrows.
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
2039
3.3. Stretch re¯ex increase vs. background EMG When pooling data from the present experiments and from a previous study of similar nature (Matre et al., 1998) there is a signi®cant negative relation between background EMG and stretch re¯ex increase during the hypertonic saline infusion (r 20:72, P , 0:001). Fig. 7 shows the average change in the short latency stretch re¯ex amplitude during pain compared with before pain, versus soleus background EMG. The data are from the three experiments in the present study, and from a previous study with isometric contractions (Matre et al., 1998). The most prominent stretch re¯ex increase during pain is seen in the relaxed muscle. With increasing background EMG throughout the ramp contraction the pain-related increase becomes smaller. 4. Discussion The main ®ndings of the present study was that experimental muscle pain facilitated the stretch re¯ex during pain in the relaxed muscle, but caused no changes in stretch re¯ex amplitude during sitting and walking at higher ``functional'' background EMG levels. That is, the stretch re¯ex changes were depending on motoneurone excitability rather than on motor task. 4.1. Stretch re¯ex modulation during ramp contraction and during walking Fig. 6. (A) Normalised stretch re¯ex amplitude from 8 subjects (mean ^ SE) during walking (X) and sitting (W) before, during and after experimentally induced muscle pain. (B) Average soleus background EMG calculated 30 ms before stretch onset during walking (X) and 100 ms before stretch onset during sitting (W). (C) Pain intensity (mean ^ SD) during the four subsequent infusions indicated with arrows.
cm during sitting (ANOVA: d:f: 1, F 0:74, P 0:42). Neither were there any differences between the four subsequent infusions which resulted in pain intensities 4.5 ^ 1.9 cm, 4:2 ^ 2:3 cm, 3:8 ^ 2:0 cm and 4.0 ^ 1.7 cm, respectively (mean of 8 subjects during walking and sitting; ANOVA: d:f: 3, F 0:38, P 0:77). There was no difference in re¯ex amplitude between walking and sitting before the hypertonic saline injections (t-test dependent samples: d:f: 7, t 1:45, P 0:19). For both the walking and sitting measurements the experimental pain did not cause any signi®cant re¯ex changes when compared with before pain (ANOVA: d:f: 4, F 0:67, P 0:6). Neither was there any difference in re¯ex amplitude between sitting and walking during the experimental pain (ANOVA: d:f: 1, F 0:51, P 0:5). For all subjects soleus background EMG during stance phase did not change with pain (ANOVA: d:f: 5, F 0:36, P 0:87), neither did stride time (ANOVA: d:f: 5, F 1:81, P 0:14).
Stretch re¯ex modulation during the ramp was similar to the modulation of monosynaptic H-re¯exes (Wilmink et al., 1996). Fewer measuring points throughout the ramp in the present study makes it dif®cult to compare, but a similar ®nding is a facilitation of the re¯ex just before ramp onset. The stretch re¯ex was facilitated 150 ms before ramp onset, which is comparable to the H-re¯ex which was facilitated even 300 ms before ramp onset (Wilmink et al., 1996). Similar re¯ex facilitation in preparation to a movement has also been reported before (Bonnet et al., 1991), and could be attributed to changes in presynaptic Ia-inhibition (Meunier and Pierrot-Deseilligny, 1989) and produce an altered postsynaptic drive from the motoneurone (MN) pool. The soleus stretch re¯ex is modulated throughout the step cycle during human walking (Sinkjñr et al., 1996). In the present walking experiment it is hard to speak about an actual modulation throughout the step cycle with only three measuring points. However, the amplitudes in time periods 1 (100±200 ms), 2 (450±550 ms) and 3 (900± 1100 ms) in the present study are similar to the amplitudes in measuring points 2 (approx. 180 ms), 4 (approx. 550 ms) and 7 (approx. 1100 ms) respectively in Fig. 4A in Sinkjñr et al. (1996). The unchanged stretch re¯ex amplitude during walking and sitting with matched soleus background EMG is in accordance with previous studies (Sinkjñr et al., 1996). The onset and peak latencies are also comparable with Sink-
2040
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
Fig. 7. Stretch re¯ex increase during pain vs. soleus background EMG. Each data point is the average increase in stretch re¯ex amplitude in percent of the measurement before pain. Data from the ramp experiment (B), isometric contractions at matched background EMG (O), walking experiment (W), walking in combined experiment (A) and data from sitting isometric contractions in Matre et al. (1998) (X).
jñr et al. (1996), and indicate that the group Ia muscle spindle afferents are responsible for the short latency stretch re¯ex (Taylor et al., 1985; Stein et al., 1991; Nielsen et al., 1998). 4.2. Pain-related changes in re¯ex amplitude The experimentally induced muscle pain facilitated the stretch re¯ex amplitude, but only in the relaxed muscle. With increasing voluntary activity in the agonist muscle (during the ramp), the effect became smaller (Fig. 3B), and with even higher soleus activity (during stance phase and at matched background activity during sitting) there was no pain-related effect on the stretch re¯ex. Thus, two phenomena call for an explanation: (1) the pain-related re¯ex facilitation in the relaxed muscle, and (2) the attenuation of this effect at higher background EMG and during voluntary coordinated movements. (1) Several pathways, both re¯ex mediated and descending, may change the stretch re¯ex amplitude (Baldissera et al., 1981). In the relaxed muscle the motoneurone membrane potential is generally far from threshold, partly because of tonic inhibitory activity both from descending and peripheral inputs. A possibility exists that muscle nociceptive activity produces a decrease in the descending inhibitory activity, resulting in increased MN excitability and an increased stretch re¯ex. Although the transcortical effect of muscle nociceptive activity still is unclear (Mense, 1993), cortical activity has been shown to remove presynaptic Ia inhibition and increase a -MN excitability (Lundberg and
VyklickuÂ, 1963; Lund et al., 1965). Nociceptive muscle afferents have also been reported to be able to change the activity of interneurones mediating presynaptic Ia-inhibition and group I non-reciprocal inhibition (Rossi et al., 1999). An argument against changes in MN excitability is the ®nding that the H-re¯ex does not change during salineinduced muscle pain whether in the soleus muscle (Matre et al., 1998), or in the temporalis muscle (Svensson et al., 1998a). There are, however, methodological differences between the stretch and H-re¯ex since both the time course and amplitude of the Ia afferent volley are different between the two stimuli (Burke et al., 1984; Morita et al., 1998). Other factors that may change the transmission of the monosynaptic Ia excitation from the muscle spindle afferents are changes in post-activation depression, and presynaptic inhibition from interneurones mediating presynaptic inhibition of Ia terminals (Hultborn et al., 1996). Changes in reciprocal Ia inhibition are not a likely explanation since the stretch re¯ex is facilitated also with hypertonic saline infusion in the antagonist muscle (Matre et al., 1998). MN responsiveness may also be altered through modulation of intrinsic properties of the a -motoneurone (Windhorst et al., 1997a). Windhorst and co-workers showed that group III± IV afferent activity may affect MN-®ring via changes in membrane potential, changes in afterhyperpolarisation or increased synaptic noise. Furthermore, chemically activated group III±IV afferents may cause disinhibition of the a -MN pool by inhibiting Renshaw cells (Windhorst et al., 1997b). One might speculate whether such a disinhibition of a -MNs by group III±IV activity could contribute to stretch re¯ex facilitation. Another explanation for the increased re¯ex amplitude during experimental muscle pain might be elevation of muscle spindle sensitivity in the relaxed state. Such an increased sensitivity could facilitate the stretch re¯ex by giving a higher Ia afferent burst for a given muscle stretch. A linkage between muscle nociceptive input and g -activity is supported in several animal studies which have shown that injection of various algogenic substances into the muscle increases g -activity (Schmidt et al., 1981; Jovanovic et al., 1990; Johansson et al., 1993; DjupsjoÈbacka et al., 1995; Pedersen et al., 1997), which in turn could increase muscle spindle sensitivity. However, studies in unanaesthetized models are needed to con®rm this. (2) With higher background EMG (both during the ramp contraction and during walking) the stretch re¯ex was not facilitated with pain (Fig. 7). This is in contrast to the normal situation (non-painful muscle) where stretch re¯ex often is higher during contraction compared with rest (Toft et al., 1991; Matre et al., 1998). Several mechanisms contribute to a higher stretch re¯ex during contraction in the nonpainful muscle: the motoneurones are at or above their thresholds, muscle spindle sensitivity is probably increased, and presynaptic inhibition and post-activation depression are both decreased (Hultborn et al., 1987). Altered descending commands converging with afferent input onto common interneurones projecting to motoneurones could also contri-
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
bute to the reduced re¯ex facilitation with pain. It is not clear which of these mechanisms that are reducing the pain-related stretch re¯ex increase with increasing MNexcitability. The increased stretch re¯ex during muscle pain could be explained by a change in each of the above mechanisms when the muscle goes from an active to a relaxed condition. For example, if the increased stretch re¯ex in the relaxed muscle is caused by an increased g -drive that increases the Ia ®ring to a muscle stretch, then, since the increase will take place anyway when the muscle becomes active, the effect of muscle pain on the stretch re¯ex will dim the more the muscle is contracted. Most probably not one single neural mechanism is producing the stretch re¯ex changes seen in the present study, but a combination of effects. Both re¯exmediated pathways and descending pathways in¯uence the stretch re¯ex amplitude. Convergence of these pathways on interneurones projecting to either g - or a -motoneurones may affect the stretch re¯ex amplitude, and the spinal integration between these pathways still needs investigation. 4.3. Post infusion changes Although the analysis failed to ®nd any signi®cant changes in the post-infusion recordings, it is evident that the stretch re¯ex amplitude in the ramp experiment tended to be reduced both 20 and 40 min after infusion compared with before infusion. Since this effect was present both after isotonic and hypertonic saline infusion it could indicate that the stimulus had a non-nociceptive component, possibly increased intramuscular pressure. An argument for this is that the re¯ex amplitude is reduced not only after, but also during the isotonic infusion. Similar reductions in stretch re¯ex amplitude have been found after fatigue (Balestra et al., 1992). None of the subjects in the present ramp experiment reported tiredness after the contractions. However, it cannot be ruled out that the saline infusions produced fatigue-like effects in the absence of any subjective sensation of fatigue. If increased intramuscular pressure is the cause, a possible mediator is group III afferents responding to lowthreshold mechanical stimuli (Ellaway et al., 1982), since group III afferents have been found to inhibit a -motoneurones according to the ¯exor re¯ex pattern (Kniffki et al., 1981). 4.4. Methodological considerations Infusion of hypertonic saline in muscles is a reliable method to induce muscle pain (Kellgren, 1938; Stohler and Lund, 1994; Svensson et al., 1995; Graven-Nielsen et al., 1997a; Matre et al., 1998). The continuous infusion used in the ramp experiment and the subsequent bolus injections used in the other experiments all produced a deep pain sensation in the soleus muscle. The pain sensed by the subject (and indicated on the visual analogue scale) may not linearly re¯ect the true afferent discharge of the muscle nociceptors. The higher pain intensity during the continuous
2041
infusion could be due to the higher volume of infused saline (Graven-Nielsen et al., 1997a). A second explanation could be that pain perception is modulated by the attention of the subject. Experimental studies have shown that attention towards other stimuli alters pain thresholds (Janssen and Arntz, 1996). Walking on a treadmill is a complex task that might take the subject's attention away from the pain. In order to control for the lower pain intensity in the ®rst walking experiment compared with the ramp experiment, a third experiment was designed where both walking and sitting measurements were done after each hypertonic saline injection (Fig. 6C). Half of the subjects started with walking after the ®rst injection, the other half started with sitting. There were no differences in pain intensity between the two situations. In several experimental pain studies infusion of isotonic saline has been used as a control stimulus to hypertonic saline (Ashton-Miller et al., 1990; Vecchiet et al., 1993; Svensson et al., 1995; Graven-Nielsen et al., 1997a; Matre et al., 1998). Isotonic saline produces no or only little pain, and is thus believed to have primarily non-nociceptive (mechanical) effects by increasing the intramuscular pressure (Graven-Nielsen et al., 1997c). The nociceptive component of the afferent input is therefore believed to be the main cause for the increased stretch re¯ex in the relaxed muscle, although it cannot be ruled out that hypertonic saline may produce unknown pain-independent effects. In addition the combined effects of the discharge of nociceptive and non-nociceptive input may be more complex than the sum of the effects caused by the separate activation of these afferents. Despite equal pain intensity and background EMG, the stretch re¯ex was quite variable after the four subsequent bolus injections of hypertonic saline in the combined walking and sitting experiment. Some subjects responded to the injection with facilitation and some with inhibition. On average there was an alternation between inhibition and facilitation during walking, and a reduced facilitation during the second and fourth injection during sitting (Fig. 6A). Most probably infusion of hypertonic saline activates both nociceptive and non-nociceptive lowthreshold mechanosensitive (group II and group III) afferents (Mense, 1993). It has been suggested that the inputs from nociceptive and non-nociceptive afferents have opposite actions on homonymous g -motoneurones (Mense and Skeppar, 1991). Thus, if a different combination of nociceptive and non-nociceptive afferents is stimulated after each injection, this could explain that some subjects exhibited facilitation and others inhibition after the injections. The present methods to elicit stretch re¯exes in the human soleus muscle during sitting (Sinkjñr et al., 1988) and walking (Andersen and Sinkjñr, 1995) have been used in several studies of ankle joint stiffness in healthy and spastic man (Toft et al., 1991; Toft and Sinkjñr, 1993; Sinkjñr et al., 1993; Toft, 1995; Sinkjñr et al., 1996). Both during sitting and walking stretches with different amplitude and velocity may be applied in both directions.
2042
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
In the combined walking and matched EMG sitting experiment in the present study stretches were applied with equal amplitude, velocity and ankle joint angle at the same level of soleus activity during both walking and sitting. This ensured equal muscle length, and muscle stretch, in the two conditions. 4.5. Summary The mechanically evoked stretch re¯ex during different motor tasks and contraction levels was studied before, during and after infusion of hypertonic saline. The hypertonic saline infusion produced a deep pain sensation in the soleus muscle. At rest the stretch re¯ex was facilitated during the experimentally induced muscle pain. In the active muscle the stretch re¯ex was unaffected by pain. The painrelated changes in stretch re¯ex amplitude seemed to be more related to the excitability of the MN pool than to motor task. One possible explanation for the increased stretch re¯ex amplitude during hypertonic saline infusions in the relaxed muscle, is the linkage between muscle nociceptive afferent activity and an increased g -activity. Other explanations could be reduced presynaptic inhibition of Ia terminals, modulation of intrinsic properties of the a -motoneurones, or reduced Renshaw inhibition. The results show that the re¯ex changes which take place in the relaxed muscle due to the pain stimuli are of much less importance when the muscle is functionally active. Acknowledgements We kindly acknowledge Knud Larsen and Thomas Graven-Nielsen, Aalborg University, Denmark, for their technical assistance. The Danish National Research Foundation are kindly acknowledged for ®nancial support. References Andersen JB, Sinkjñr T. An actuator system for investigating electrophysiological and biomechanical features around the human ankle joint during gait. Trans Rehab Eng 1995;3(4):299±306. Appelberg B, Hulliger M, Johansson H, Sojka P. Actions on gamma-motoneurones elicited by electrical stimulation of group I muscle afferent ®bres in the hind limb of the cat. J Physiol 1983;335:237±253. Arendt-Nielsen L, Graven-Nielsen T, Svarrer H, Svensson P. The in¯uence of low back pain on muscle activity and coordination during gait Ð A clinical and experimental study. Pain 1996;64:231±240. Ashton-Miller JA, McGlashen KM, Herzenberg JE, Stohler CS. Cervical muscle myoelectric response to acute experimental sternocleidomastoid Pain. Spine 1990;15(10):1006±1012. Baldissera F, Hultborn H, Illert M. Integration in spinal neuronal systems. In: Brooks VB, editor. Handbook of physiology (a critical, comprehensive presentation of physilogical knowledge and concepts). Section 1: The nervous system. Vol. II: Motor control, part 1, Bethesda, MD: American Physiological Society. 1981. pp. 509±595. Balestra C, Duchateau J, Hainaut K. Effects of fatigue on the stretch re¯ex in a human muscle. Electroenceph clin Neurophysiol 1992;85:46±52. Bonnet M, Requin J, Stelmach GE. Changes in electromyographic
responses to muscle stretch, related to the programming of movement parameters. Electroenceph clin Neurophysiol 1991;81:135±151. Burke D, Gandevia SC, McKeon B. Monosynaptic and ogliosynaptic contributions to human ankle jerk and h-re¯ex. J Neurophysiol 1984;52(3):435±448. Capaday C, Stein RB. Amplitude modulation of the soleus H-re¯ex in the human during walking and standing. J Neurosci 1986;6(5):1308±1313. DjupsjoÈbacka M, Johansson H, Bergenheim M, SjoÈlander P. In¯uences on the gamma-muscle-spindle system from contralateral muscle afferents stimulated by KCl and lactic acid. Neurosci Res 1995;21:301±309. Ellaway PH, Murphy PR, Tripathi A. Closely coupled excitation of gammamotoneurones by group III muscle afferents with low mechanical threshold in the cat. J Physiol 1982;331:481±498. Graven-Nielsen T, Arendt-Nielsen L, Svensson P, Jensen TS. Experimental muscle pain: A quantitative study of local and referred pain in humans following injection of hypertonic saline. J Musculoskel Pain 1997a;5:49±69. Graven-Nielsen T, Arendt-Nielsen L, Svensson P, Jensen TS. Quanti®cation of local and referred muscle pain in humans after sequential i.m. injections of hypertonic saline. Pain 1997b;69:111±117. Graven-Nielsen T, McArdle A, Phoenix J, Arendt-Nielsen L, Jensen TS, Jackson MJ, Edwards RHT. In vivo model of muscle pain: Quanti®cation of intramuscular chemical, electrical, and pressure changes associated with saline-induced muscle pain in humans. Pain 1997c;69 (1-2):137±143. Graven-Nielsen T, Svensson P, Arendt-Nielsen L. Effects of experimental muscle pain on muscle activity and co-ordination during static and dynamic motor function. Electroenceph clin Neurophysiol 1997d; 105(2):156±164. Hultborn H, Meunier S, Pierrot-Dsesilligny E, Shindo M. Changes in presynaptic inhibition of Ia ®bres at the onset of voluntary contraction in man. J Physiol 1987;389:757±772. Hultborn H, Illert M, Nielsen J, Paul A, Ballegaard M, Wiese H. On the mechanism of post-activation depression of the H-re¯ex in human subjects. Exp Brain Res 1996;108(3):450±462. Janssen SA, Arntz A. Anxiety and pain: attentional and endorphinergic in¯uences. Pain 1996;66:145±150. Johansson H, Sojka P. Pathophysiological mechanisms involved in genesis and spread of muscular tension in occupational muscle pain and in chronic musculoskeletal pain syndromes: A hypothesis. Med Hypotheses 1991;35:196±203. Johansson H, DjupsjoÈbacka M, SjoÈlander P. In¯uences on the gammamuscle-spindle system from muscle afferents stimulated by KCl and lactic acid. Neurosci Res 1993;16:49±57. Jovanovic K, Anastasijevic R, Vuco J. Re¯ex effects on gamma-fusimotor neurones of chemically induced discharges in small-diameter muscle afferents in decerebrate cats. Brain Res 1990;521:89±94. Kellgren JH. Observations on referred pain arising from muscle. Clin Sci 1938;3:175±190. Kniffki K-D, Schomburg ED, Steffens H. Synaptic effects from chemically activated ®ne muscle afferents upon a-motoneurones in decerebrate and spinal cats. Brain Res 1981;206:361±370. Lund S, Lundberg A, Vyklicku L. Inhibitory action from the ¯exor re¯ex afferents on transmission to Ia afferents. Acta Physiol Scand 1965;64: 345±355. Lund JP, Donga R, Widmer CG, Stohler CS. The pain-adaptation model: a discussion of the relationship between chronic musculoskeletal pain and motor activity. Can J Physiol Pharmacol 1991;69:683±694. Lundberg A, Vyklicku L. Inhibitory interaction between spinal re¯exes to primary afferents. Experientia 1963;19:247±248. Matre DA, Sinkjñr T, Svensson P, Arendt-Nielsen L. Experimental muscle pain increases the human stretch re¯ex. Pain 1998;75(2±3):331±339. Mense S. Review article: Nociception from skeletal muscle in relation to clinical muscle pain. Pain 1993;54:241±289. Mense S, Skeppar P. Discharge behaviour of feline gamma-motoneurones following induction of an arti®cial myositis. Pain 1991;46:201±210. Meunier S, Pierrot-Deseilligny E. Gating of the afferent volley of the
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043 monosynaptic stretch re¯ex during movement in man. J Physiol 1989;419:753±763. Morita H, Petersen N, Christensen LOD, Sinkjñr T, Nielsen J. Sensitivity of H-re¯exes and stretch re¯exes to presynaptic inhibition in humans. J Neurophysiol 1998;80(2):610±620. Nielsen J, Sinkjñr T, Baumgarten J, Andersen JB, Toft E, Christensen LOD, Ladoceur M, Morita H. Modulation of the soleus stretch re¯ex and H-re¯ex during human walking after block of peripheral feedback. Neurosci Abstr 1998;24:2103. Pedersen J, SjoÈlander P, Wenngren BI, Johansson H. Increased intramuscular concentrations of bradykinin increases the static fusimotor drive to muscle spindles in neck muscles of the cat. Pain 1997;70(1):83±91. Prochazka A, Hulliger M, Trend P, DuÈrmuÈller N. Dynamic and static fusimotor set in various behavioural contexts. In: Hnik P, Soukup T, Vejsada R, Zelena J, editors. Mechanoreceptors. New York: Plenum, 1988. pp. 417±430. Rossi A, Decchi B, Ginanneschi F. Presynaptic excitability changes of group Ia ®bres to muscle nociceptive stimulation in humans. Brain Res 1999;818:12±22. Schmidt RF, Kniffki K-D, Schomburg ED. Der Ein¯uss Klinkalibriger Muskelafferenzen auf den Muskeltonus. In: Bauer H, Koella WP, Struppler H, editors. Therapie der Spastik, Munchen: Verlag fur angewandte Wissenschaft. 1981. pp. 71±86. Schomburg ED. Review article: Spinal sensorimotor systems and their supraspinal control. Neurosci Res 1990;7:265±340. Sinkjñr T. Muscle, re¯ex and central components in the control of the ankle joint in healthy and spastic man. Acta Neurol Scand 1997;96(Suppl 170):1. Sinkjñr T, Toft E, Andreassen S, Hornemann BC. Muscle stiffness in human ankle dorsi¯exors: intrinsic and re¯ex components. J Neurophysiol 1988;60:1110±1121. Sinkjñr T, Toft E, Larsen K, Andreassen S, Hansen HJ. Non-re¯ex and re¯ex mediated ankle joint stiffness in multiple sclerosis patients with spasticity. Muscle Nerve 1993;16:69±76. Sinkjñr T, Andersen JB, Larsen B. Soleus stretch re¯ex modulation during gait in humans. J Neurophysiol 1996;76(2):1112±1120. Stein RB, Yang J, Edamura M, Capaday C. Re¯ex modulation during normal and pathological human walking. In: Shimamura M, editor. Neurobiological basis of human locomotion, Tokyo: Jpn Sci Soc, 1991. pp. 335±346. Stohler CS, Lund JP. Effects of noxious stimulation of the jaw muscles on
2043
the sensory experinece of volunteer human subjects. In: Stohler CS, Carlson DS, editors. Biological & physiological aspects of orofacial pain. Craniofacial growth Series 29, Center for Human Growth & Development. The University of Michigan, 1994. pp. 55±73. Svensson P, Arendt-Nielsen L, Nielsen H, Larsen JK. Effect of chronic and experimental jaw muscle pain on pain-pressure thresholds and stimulusresponse curves. J Orofacial Pain 1995;9:347±356. Svensson P, Houe L, Arendt-Nielsen L. Bilateral experimental muscle pain changes electromyographic activity of human jaw-closing muscles during mastication. Exp Brain Res 1997;116:182±185. Svensson P, DeLaat A, Graven-Nielsen T, Arendt-Nielsen L. Experimental jaw-muscle pain does not change heteronymous H-re¯exes in the human temporalis muscle. Exp Brain Res 1998a;121(3):311±318. Svensson P, Graven-Nielsen T, Matre DA, Arendt-Nielsen L. Experimental muscle pain does not cause long-lasting increases in resting EMG activity. Muscle Nerve 1998b;21:1382±1389. Taylor J, Stein RB, Murphy PR. Impulse rates and sensitivity to stretch of soleus muscle spindle afferent ®bers during locomotion in premammilary cats. J Neurophysiol 1985;53(2):341±360. Toft E. Mechanical and electromyographical stretch responses in spastic and healthy subjects. Acta Neurol Scand 1995;92(Suppl 163):1. Toft E, Sinkjñr T. H-re¯ex changes during contractions of the ankle extensors in spastic patients. Acta Neurol Scand 1993;88:327±333. Toft E, Sinkjñr T, Andreassen S, Larsen K. Mechanical and electromyographic responses to stretch of the human ankle extensors. J Neurophysiol 1991;65(6):1402±1410. Vecchiet L, Dragani L, Bigontina P, Obletter G, Giamberardino MA. Experimental referred pain and hyperalgesia from muscles in humans. In: Vecchiet L, Albe-Fessard D, Lindblom U, editors. New trends in referred pain and hyperalgesia, Amsterdam: Elsevier, 1993. pp. 239± 249. Wilmink RJH, Slot PJ, Sinkjñr T. Modelling of the H-re¯ex facilitation during ramp and hold contractions. J Comput Neurosci 1996;3(4):337± 346. Windhorst U, Kirmayer D, Soibelman F, Misri A, Rose R. Effects of neurochemically excited group III-IV muscle afferents on motoneuron afterhyperpolarization. Neuroscience 1997a;76(3):915±929. Windhorst U, Meyer-Lohmann J, Kirmayer D, Zochodne D. Renshaw cell responses to intra-arterial injection of muscle metabolites into cat calf muscles. Neurosci Res 1997b;27(3):235±247.
The in¯uence of experimental muscle pain on the human soleus stretch re¯ex during sitting and walking Dag®nn A. Matre a,b,*, Thomas Sinkjñr a, Stein Knardahl b, Jacob B. Andersen a, Lars Arendt-Nielsen a b
a Center for Sensory-Motor Interaction, Aalborg University, DK-9220 Aalborg E, Denmark Department of Physiology, National Institute of Occupational Health, N-0033 Oslo, Norway
Accepted 13 August 1999
Abstract Objectives: The stretch re¯ex is functionally important during human locomotion. Muscle pain has been found to increase the stretch re¯ex amplitude during sitting, possibly due to an altered fusimotor drive. To further study the importance of altered fusimotor activity due to muscle pain we investigated the combined effect of muscle pain and motor task on the soleus stretch re¯ex. Methods: Stretch re¯exes were elicited before, during and after experimentally induced muscle pain in soleus (i.m. infusion of 6% saline) in 3 experiments: (1) in the relaxed soleus muscle and before, during and after an isometric ramp contraction (500 ms, 0±10 Nm), (2) at 3 different time periods during walking, and (3) at matched pain intensity and soleus activity during sitting and walking. Results: Infusion of hypertonic saline into the soleus muscle caused a signi®cant facilitated stretch re¯ex in the relaxed muscle (P , 0:01), but not during walking or during sitting and walking at matched soleus EMG and matched pain levels. The infusion of isotonic saline (nonpainful) did not cause any changes (P 0:75). Conclusions: The main ®ndings of the present study were that experimental muscle pain facilitated the stretch re¯ex during pain in the relaxed muscle, but caused no changes in stretch re¯ex amplitude during sitting and walking at higher ``functional'' background EMG levels. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Experimental muscle pain; Stretch re¯ex; Motor control; Walking; Soleus muscle
1. Introduction The knowledge of the neural mechanisms leading to, and maintaining chronic musculoskeletal pain and occupational muscle pain is limited. Animal studies have shown that injection of various pain-producing substances into the muscle may activate the g -system (Schmidt et al., 1981; Ellaway et al., 1982; Appelberg et al., 1983; Jovanovic et al., 1990; Mense and Skeppar, 1991; Pedersen et al., 1997). There is an excitatory connection between muscle spindle afferents and homonymous motoneurones (Schomburg, 1990), and it has been suggested that an elevated g -activity, caused by muscle pain, facilitates this excitatory connection producing a higher electromyographic activity in painful muscles (Schmidt et al., 1981; Johansson and Sojka, 1991). Clinical and experimental studies of musculoskeletal pain and electromyographic activity (EMG) have not found any support for this hypothesis, neither during relaxed, static * Corresponding author. Tel.: 147-23-195-267; fax: 147-23-195-204. E-mail address: dag®[email protected] (D.A. Matre)
nor dynamic conditions (Lund et al., 1991; Arendt-Nielsen et al., 1996; Svensson et al., 1997; Graven-Nielsen et al., 1997c; Svensson et al., 1998b). However, in a recent human study we found that the soleus stretch re¯ex, but not the H re¯ex, was facilitated by experimentally induced muscle pain (Matre et al., 1998). This gives indirect support to a link between muscle pain and increased spindle sensitivity. Since the stretch re¯ex is functionally important in the control of joint stiffness (Sinkjñr, 1997) we wanted to investigate the possible effect of muscle pain on the re¯ex amplitude during different motor tasks. In the present study we compared the soleus stretch re¯ex in the relaxed muscle, in the active muscle during isometric ramp contractions, during walking, and during sitting and walking with matched pain intensity and matched ``motoneurone excitability'' as it can be measured from soleus background EMG (Capaday and Stein, 1986). The short latency stretch re¯ex amplitude was compared in each of the four situations before, during and after inducing experimental muscle pain, produced by intramuscular infusion of hypertonic saline into the soleus muscle. The following
1388-2457/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 1388-245 7(99)00211-4
CLINPH 99532
2034
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
questions were asked: is the stretch re¯ex modulation during ramp contractions, static contractions and walking during pain different from the relaxed state? If so, can the modulation be related to changes in g -activity? In the cat the g activity has been found to be dependent on motor task (Prochazka et al., 1988). 2. Materials and methods 2.1. Subjects and experimental conditions The investigation was divided into three experiments, performed on three different groups of healthy volunteers. They were recruited from university students after announcements in an internal newspaper. All subjects signed an informed consent form and the study was approved by the Local Ethical Committee and was conducted in accordance with the Helsinki Declaration. The subjects abstained from coffee and exercise prior to the investigation, and no subjects had any history of chronic musculoskeletal disorders. 2.1.1. Ramp experiment Ten healthy male subjects (age range 20±24 years) participated in an experiment where stretch re¯exes were elicited both in the relaxed soleus and at 6 regular periods before, during and after an isometric ramp contraction (500 ms). The ramp was similar in length to the rising phase of the soleus EMG in the stance phase during walking. Recordings were made before, during and after experimental muscle pain was produced in soleus by intramuscular infusion of hypertonic saline. Isotonic saline was used as control in a separate experiment separated by at least 5 days. 2.1.2. Walking experiment Fourteen healthy male subjects (age range 20±30 years) participated in an experiment where stretch re¯exes were elicited in soleus at three time periods of the stride while the subject was walking on a treadmill (walking speed 3.5± 4.5 km/h). The time periods corresponded to early stance phase, late stance phase and mid swing phase. Recordings were made, before, during and after experimental muscle pain was produced in soleus by intramuscular injections of hypertonic saline. 2.1.3. Walking and sitting with matched EMG One female (age 25) and seven male subjects (age range 18±25 years) participated in the experiment where stretch re¯exes were elicited in soleus in early stance phase (100± 200 ms after heel contact), and while sitting with matched background EMG in soleus. Recordings were made before, during and after injections of hypertonic saline. Four subsequent saline injections were made, and after each injection walking and sitting recordings were done in random order.
2.2. Pain stimulus Sterile hypertonic saline (6%) was infused to produce acute deep pain in the soleus muscle. In the relaxed and ramp conditions, the infusion was performed by a computer-controlled syringe pump (IVAC model 770, UK) with a 10 ml plastic syringe. A tube (IVAC G30303) was connected from the syringe to a catheter (VENFLON, 22G, 25 mm) (Graven-Nielsen et al., 1997d). The catheter was inserted into the soleus muscle, on the medial side, before the pre-infusion recording and remained in the muscle for the rest of the recordings. Using a computer controlled paradigm, 7.1 ml of saline was infused over a period of 15 min. The infusion rate was adjusted during the period to produce as constant a pain sensation as possible (90 ml/h for 20 s, 18 ml/h for 440 s, and 36 ml/h for 440 s) (Graven-Nielsen et al., 1997d; Matre et al., 1998). The subject rated the pain intensity on a 10 cm electronic visual analogue scale (VAS) and the ratings were stored on computer. The lower endpoint of the scale was marked ``no pain'' and rated `0', and the upper endpoint was marked ``most pain imaginable'' and rated `10'. Isotonic saline (0.9%, non-painful) was used as a control infusion in the ramp experiment. In the walking experiment three subsequent bolus injections of hypertonic saline (6%) were made. After each injection recordings were made for one time period. The injected volumes were 0.5, 0.6 and 0.7 ml (injected over 20±25 s) in the three time periods respectively. The locations of injection were separated by 2 cm (Graven-Nielsen et al., 1997b). Each pain stimulus lasted for approximately 5 min, and the pain intensity rating was given orally by the subject every 30 s as a number between 0 and 10. In the combined walking and matched EMG experiment four subsequent bolus injections of hypertonic saline (6%) were made. The injected volumes were 0.5, 0.6, 0.7 and 0.8 ml (injected over 20±30 s) respectively, and separated by 2 cm in location. The pain lasted for approximately 5 min, and the pain intensity was rated on an electronic visual analogue scale (VAS) and the ratings were stored on computer. 2.3. Experimental set-up 2.3.1. Relaxed and ramp experiment The subject was seated in a chair with the left foot strapped to a platform. Knee and ankle joints were extended to an angle of approximately 1008. Stretch-re¯exes were elicited in soleus by applying a 48 dorsi¯exion of the foot. The stretch had a rise time of 50 ms. The dorsi¯exion was made by rotating the platform with a motor (CEM, model 26) where the axis of rotation of the ankle joint was aligned with the axis of rotation of the platform. The motor was driven by a DC power ampli®er (BruÈel and Kjñr, model 2708). For further details, see Sinkjñr et al. (1988). The subject was trained to do isometric ramp contractions with
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
the ankle extensor muscles. Recti®ed and low-pass ®ltered soleus EMG was monitored on an oscilloscope and used as feedback for the subject. The ramp rose linearly (rise time 500 ms) to a soleus EMG level corresponding to approximately 10 Nm ankle torque. The ramp was drawn on the screen of an oscilloscope and the subject was trained to match the ramp as the monitored EMG swept across the screen. Stretches were elicited at 6 time points relative to the ramp start: 150 (before ramp onset), 0, 150, 300, 500 and 600 (after ramp) ms. In addition, stretch re¯exes were elicited in the relaxed muscle. The timing was triggered by the onset of the voluntary EMG. The trials were randomly alternated between trials with perturbation (ankle dorsi¯exion) and trials without perturbation. Recordings were made before infusion, during infusion, 20 min after and 40 min after the pain had disappeared. All subjects participated twice, receiving hypertonic and isotonic saline (control) in random order. The two sessions were separated by at least 1 week.
2035
subject rested in 5 min before the procedure was repeated with another time period, selected at random. 2.3.3. Walking and sitting with matched EMG The subjects were walking on a treadmill, and re¯exes were elicited as described above. The soleus EMG level 30 ms prior to the perturbation was noted. While the subject was seated re¯exes was elicited when an isometric contraction at the same strength (matched EMG) was held. The setup enabled equal stretch velocity, amplitude, ankle joint position, and pain intensity during walking and sitting, in addition to the matched EMG level. One series of walking and sitting measurements was made before pain, four series during pain, and one series after pain. Half of the subjects started with walking, the other half started with sitting. In each series 15 measurements were done walking, and 15 measurements were done sitting. 2.4. Data collection
2.3.2. Walking experiment The subjects were walking on a treadmill (Powerjog EG30). Stretch re¯exes were elicited in soleus by a semiportable stretch device attached to the subject's left foot performing a 88 dorsi¯exion of the foot with a velocity of approximately 3008/s (Andersen and Sinkjñr, 1995). The device, constructed as a functional joint, followed the ankle rotation during locomotion and was connected to a motor by means of bowed wires. The motor was regulated in such a way that it followed the movement of the ankle joint without in¯uencing the pattern of gait. During any time of the gait cycle it was possible to evoke a muscle stretch of the ankle extensors by rotating the ankle joint. The device was attached to a carbon ®bre shell which was strapped to the lower leg. For further details about the stretch device see Andersen and Sinkjñr (1995). The subject was walking for a few minutes prior to the recordings to adapt to the set-up and was asked to walk with a comfortable speed, usually between 3.5 and 4.5 km/h. The same speed was used in the rest of the experiment. During the experiment stretches were evoked for every 4±5 strides. For every stretch response recorded, the preceding stride was also recorded. The stride duration was measured from one heel contact to the succeeding. Each stride was divided into three time periods: (1) between heel down and heel off, (2) between heel off and toe off, and (3) between toe off and heel down. Stretches were elicited within the respective periods with the following delays from heel down: time period 1, 100± 200 ms; time period 2, 450±550 ms; time period 3, 900± 1100 ms. Thirty stretches were evoked in each time period. In the pre-pain and post-pain sessions stretches were evoked in all three segments successively. For the recordings during pain, the subject started walking at pain onset and stretches were evoked as long as the pain lasted, or for maximum 30 recordings. After one recording session with pain, the
EMG signals were recorded from the soleus and tibialis anterior muscles by bipolar Ag±AgCl surface electrodes (Medicotest; 2 cm apart). The skin was abraded, lightly rubbed with sand paper and cleaned with isopropanol before attaching the electrodes. The EMG signals were ampli®ed (DISA, model 15C01), ®ltered with a band pass ®lter (20 Hz to 2 kHz), sampled (4 kHz) and stored on computer. In the walking experiment optical encoders measured the actual position of the ankle joint. The position was measured, sampled and stored on a computer. In the ramp experiment a potentiometer measured the ankle position. For further details regarding the data collection of the walking experiment see Sinkjñr et al. (1996). 2.5. EMG analysis The perturbation resulted in a short latency stretch re¯ex in all subjects. For each series of recordings background EMG and stretch re¯ex peaks were averaged. In the ramp experiment nine re¯exes were averaged for the relaxed condition and for each time slot. In the walking experiment the inclusion criteria (a stretch velocity between 275 and 3008/s) reduced the number of valid re¯exes to between 10 and 20 for each segment. In the combined walking and matched EMG experiment the inclusion criteria (matched background EMG) reduced the number of valid re¯exes from 15 to between 9 and 10. The stretch re¯ex amplitudes during and after infusion was normalised to the recordings before infusion. Average background EMG 100 ms (static experiments) or 30 ms (walking experiments) prior to stretch onset was subtracted from the peak amplitudes. Due to a more variable EMG the average background EMG was calculated in a shorter time period in the walking experiments than in the static experiments.
2036
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
2.6. Statistics In the ramp experiment the data were analysed using a repeated measures 3 way analysis of variance (ANOVA) where time points (relaxed, 2150, 0, 150, 300, 500 and 600), experimental conditions (before, during and after infusion) and saline concentrations (hypertonic and isotonic) were used as factors. In the walking experiment the data were analysed using a repeated measures 2 way ANOVA with time periods (1, 2 and 3) and experimental conditions (before, during and after injections) as factors. In the combined walking and matched EMG experiment the data were analysed using a repeated measures 2 way ANOVA with experimental conditions (before, during and after injections) and motor task (walking and sitting) as factors. A signi®cance level of P , 0:05 was considered statistically signi®cant in both experiments. If a main effect was found for one of the factors or between two factors a post-hoc test was performed using Student±Newman±Keul's method, thus identifying the factor causing signi®cance.
3. Results 3.1. Stretch re¯ex modulation in the relaxed muscle and during the ramp contraction Fig. 1 shows the stretch response after a perturbation (48 dorsi¯exion) given at ramp onset (dashed line).Fig. 2 shows the background EMG and stretch re¯ex modulation during the ramp before saline infusion. The average stretch re¯ex increased with .100% from the relaxed state to 150 ms before ramp onset. The hypertonic saline infusion produced a stable deep pain sensation distributed on the posterior side of the lower leg. The mean pain intensity was signi®cantly higher (Wilcoxon: Z 2:8, P 0:005) during the hypertonic infusion (5:5 ^ 2:4 cm on the VAS) than during the isotonic infusion (0:4 ^ 0:6 cm) (Fig. 3C). For all subjects the EMG included a short latency response with onset and peak latencies 44:4 ^ 2:4 ms and 64:2 ^ 6:4 ms, respectively (mean ^ SD; relaxed muscle). No subjects exhibited the medium latency stretch re¯ex responses which is seen in the active muscle during similar experiments (Toft et al., 1991). The normalised stretch re¯ex amplitude (Fig. 3B) was higher during compared with before infusion of hypertonic saline (3 way ANOVA and SNK: P 0:012). A signi®cant interaction was found between saline concentration, experimental condition and time points (ANOVA: d:f: 12, F 2:01, P 0:044). Post-hoc analysis showed that the re¯ex facilitation was signi®cantly higher in the relaxed muscle (SNK: P , 0:005; indicated with *), but not during the ramp. During the isotonic control infusion the normalised stretch re¯ex amplitude was lower than during the hypertonic infusion (3 way ANOVA and SNK:
Fig. 1. Example of recorded averaged data from control contraction (heavy lines) and contractions with a perturbation at ramp onset (thin lines). Average of 9 re¯ex responses from subject JP. (A) Ankle angle positions, (B) recti®ed and ®ltered soleus muscle EMG. Time 0 (dashed line): ramp start. Time 500: ramp end.
P 0:008). The stretch re¯ex was not different during compared with before the the isotonic infusion (SNK: P 0:75). The re¯ex amplitude was not signi®cantly different after the infusions, compared with before the infusions. No change was found in soleus background EMG (Fig. 3A) when comparing the painful condition with the pre-pain condition (ANOVA: d:f: 3, F 2:25, P 0:18) or when comparing the saline concentrations (ANOVA: d:f: 1, F 0:15, P 0:74). 3.2. Stretch re¯ex modulation during walking Stretch re¯exes were measured in two experiments during walking. In the ®rst walking experiment re¯exes were elicited in early stance phase, late stance phase and in the
Fig. 2. Stretch re¯ex modulation during ramp contraction before saline infusion (X) and soleus background EMG (B). Each point is an average of 90 stretches (10 persons £ 9 stretches; mean ^ SE).
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
2037
Fig. 3. (A) Stretch re¯ex modulation (mean ^ SE; n 10) in the relaxed muscle, before and throughout an isometric ramp contraction before, during, 20 min after, and 40 min after infusion of hypertonic (6%) saline (X) and isotonic (0.9%) saline (W). Soleus background EMG (®lled squares) in the relaxed muscle and throughout the ramp contraction from 10 subjects (mean ^ SE) before, during, 20 min after, and 40 min after infusion. (B) Stretch re¯ex amplitude normalised to values before infusion (0%). The amplitude was signi®cantly higher during the hypertonic infusion in the relaxed muscle (*P , 0:005). (C) Pain intensity (mean ^ SE; n 10) during infusion of hypertonic (6%) saline (X) and isotonic (0.9%) saline (W). Note that the abscissa have a different time scale compared with the abscissa in (A) and (B). The stretch re¯ex was elicited at the different time points in random order during the 900-s infusion period.
swing phase. An example of average and recti®ed soleus EMG is shown in Fig. 4 with control steps (heavy lines) and steps with a perturbation (thin lines) in early stance phase (dashed line). The 88 dorsi¯exion of the foot was re¯ected in the soleus EMG (Fig. 4B, thin line). One subject was excluded from the study because only a few stretch re¯ex responses were different from the background EMG. In time period 1 (early stance phase) the short latency stretch re¯ex was present in all 13 remaining subjects, in time period 2 (late stance phase) it was present in 12 of 13 subjects, and in time period 3 (mid swing phase) in 6 of 13 subjects. The medium latency stretch re¯ex was present in time period 1 (10 of 13 subjects) and in time period 2 (6 of 13 subjects).
Fig. 5A shows the stretch re¯ex modulation (X) and soleus background EMG (W) in early stance phase, late stance phase and mid swing phase before, during and after injection of hypertonic saline. The pain intensities after the hypertonic saline injections in periods 1±3 are shown in Fig. 5B (mean of 13 subjects). The experimentally induced muscle pain did not in¯uence the short latency stretch re¯ex in either time period throughout the step cycle (ANOVA: d:f: 2, F 0:57, P 0:59). Neither did the medium latency stretch re¯ex change with pain (ANOVA: d:f: 2, F 2:5, P 0:1). Furthermore, average background EMG during the active periods of the soleus (stance phase) and tibialis anterior muscle (swing phase) did not change during injection of hypertonic saline (ANOVA-
2038
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
Fig. 4. Example of recorded averaged data from control steps (heavy lines) and steps with a perturbation (thin lines). Superimposition of data from time periods 1±3 (control steps) and time period 1 (steps with a perturbation). Average of 30 step cycles from subject KA. The time shown represents a full step cycle. (A) Ankle angle positions, (B) recti®ed and ®ltered soleus EMG with short and medium latency stretch re¯ex response. Dashed line: start of perturbation.
soleus: d:f: 2, F 0:85, P 0:44; ANOVA-tibialis anterior: d:f: 2, F 2:83, P 0:09). Onset and peak latencies (41:6 ^ 5:0 ms and 53:9 ^ 5:6 ms, respectively, in early stance phase) were unchanged during the pain condition (ANOVA: d:f: 2, F 2:73, P 0:09), and neither did the peak latencies of the medium latency stretch re¯ex depend on the pain condition (ANOVA: d:f: 2, F 2:5, P 0:11). The stride time was signi®cantly reduced from 1:21 ^ 0:07 s (pre-pain) to 1:19 ^ 0:08 s (during pain) (ttest dependent samples: d:f: 13, t 2:45, P 0:03). In the second walking experiment re¯exes were elicited alternately in early stance phase during walking and during sitting. Soleus background EMG and pain intensity were matched in the two conditions. Fig. 6A shows the stretch re¯ex amplitudes during and after injections of hypertonic saline normalised to the measurements before injection for each subject and averaged. The re¯exes were elicited at matched background EMG (Fig. 6B; ANOVA: d:f: 1, F 0:35, P 0:57), ankle angle, and pain intensity (Fig. 6C) while the subject was seated (W) keeping a static isometric contraction, and walking (X). The pain intensity (mean of 8 subjects over 4 infusions) was 4:0 ^ 1:9 cm during walking which was not different from 4.2 ^ 2.0
Fig. 5. (A) Stretch re¯ex amplitude (X) and soleus background EMG (W) during walking from 13 subjects (mean ^ SE) before, during, and 20 min after infusion of hypertonic (6%) saline. Between 20 and 30 measurements are averaged for each subject in each time period. No change was found during the hypertonic saline infusion. (B) Pain intensity (mean ^ SE; n 13) during periods 1±3. Injections indicated with arrows.
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
2039
3.3. Stretch re¯ex increase vs. background EMG When pooling data from the present experiments and from a previous study of similar nature (Matre et al., 1998) there is a signi®cant negative relation between background EMG and stretch re¯ex increase during the hypertonic saline infusion (r 20:72, P , 0:001). Fig. 7 shows the average change in the short latency stretch re¯ex amplitude during pain compared with before pain, versus soleus background EMG. The data are from the three experiments in the present study, and from a previous study with isometric contractions (Matre et al., 1998). The most prominent stretch re¯ex increase during pain is seen in the relaxed muscle. With increasing background EMG throughout the ramp contraction the pain-related increase becomes smaller. 4. Discussion The main ®ndings of the present study was that experimental muscle pain facilitated the stretch re¯ex during pain in the relaxed muscle, but caused no changes in stretch re¯ex amplitude during sitting and walking at higher ``functional'' background EMG levels. That is, the stretch re¯ex changes were depending on motoneurone excitability rather than on motor task. 4.1. Stretch re¯ex modulation during ramp contraction and during walking Fig. 6. (A) Normalised stretch re¯ex amplitude from 8 subjects (mean ^ SE) during walking (X) and sitting (W) before, during and after experimentally induced muscle pain. (B) Average soleus background EMG calculated 30 ms before stretch onset during walking (X) and 100 ms before stretch onset during sitting (W). (C) Pain intensity (mean ^ SD) during the four subsequent infusions indicated with arrows.
cm during sitting (ANOVA: d:f: 1, F 0:74, P 0:42). Neither were there any differences between the four subsequent infusions which resulted in pain intensities 4.5 ^ 1.9 cm, 4:2 ^ 2:3 cm, 3:8 ^ 2:0 cm and 4.0 ^ 1.7 cm, respectively (mean of 8 subjects during walking and sitting; ANOVA: d:f: 3, F 0:38, P 0:77). There was no difference in re¯ex amplitude between walking and sitting before the hypertonic saline injections (t-test dependent samples: d:f: 7, t 1:45, P 0:19). For both the walking and sitting measurements the experimental pain did not cause any signi®cant re¯ex changes when compared with before pain (ANOVA: d:f: 4, F 0:67, P 0:6). Neither was there any difference in re¯ex amplitude between sitting and walking during the experimental pain (ANOVA: d:f: 1, F 0:51, P 0:5). For all subjects soleus background EMG during stance phase did not change with pain (ANOVA: d:f: 5, F 0:36, P 0:87), neither did stride time (ANOVA: d:f: 5, F 1:81, P 0:14).
Stretch re¯ex modulation during the ramp was similar to the modulation of monosynaptic H-re¯exes (Wilmink et al., 1996). Fewer measuring points throughout the ramp in the present study makes it dif®cult to compare, but a similar ®nding is a facilitation of the re¯ex just before ramp onset. The stretch re¯ex was facilitated 150 ms before ramp onset, which is comparable to the H-re¯ex which was facilitated even 300 ms before ramp onset (Wilmink et al., 1996). Similar re¯ex facilitation in preparation to a movement has also been reported before (Bonnet et al., 1991), and could be attributed to changes in presynaptic Ia-inhibition (Meunier and Pierrot-Deseilligny, 1989) and produce an altered postsynaptic drive from the motoneurone (MN) pool. The soleus stretch re¯ex is modulated throughout the step cycle during human walking (Sinkjñr et al., 1996). In the present walking experiment it is hard to speak about an actual modulation throughout the step cycle with only three measuring points. However, the amplitudes in time periods 1 (100±200 ms), 2 (450±550 ms) and 3 (900± 1100 ms) in the present study are similar to the amplitudes in measuring points 2 (approx. 180 ms), 4 (approx. 550 ms) and 7 (approx. 1100 ms) respectively in Fig. 4A in Sinkjñr et al. (1996). The unchanged stretch re¯ex amplitude during walking and sitting with matched soleus background EMG is in accordance with previous studies (Sinkjñr et al., 1996). The onset and peak latencies are also comparable with Sink-
2040
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
Fig. 7. Stretch re¯ex increase during pain vs. soleus background EMG. Each data point is the average increase in stretch re¯ex amplitude in percent of the measurement before pain. Data from the ramp experiment (B), isometric contractions at matched background EMG (O), walking experiment (W), walking in combined experiment (A) and data from sitting isometric contractions in Matre et al. (1998) (X).
jñr et al. (1996), and indicate that the group Ia muscle spindle afferents are responsible for the short latency stretch re¯ex (Taylor et al., 1985; Stein et al., 1991; Nielsen et al., 1998). 4.2. Pain-related changes in re¯ex amplitude The experimentally induced muscle pain facilitated the stretch re¯ex amplitude, but only in the relaxed muscle. With increasing voluntary activity in the agonist muscle (during the ramp), the effect became smaller (Fig. 3B), and with even higher soleus activity (during stance phase and at matched background activity during sitting) there was no pain-related effect on the stretch re¯ex. Thus, two phenomena call for an explanation: (1) the pain-related re¯ex facilitation in the relaxed muscle, and (2) the attenuation of this effect at higher background EMG and during voluntary coordinated movements. (1) Several pathways, both re¯ex mediated and descending, may change the stretch re¯ex amplitude (Baldissera et al., 1981). In the relaxed muscle the motoneurone membrane potential is generally far from threshold, partly because of tonic inhibitory activity both from descending and peripheral inputs. A possibility exists that muscle nociceptive activity produces a decrease in the descending inhibitory activity, resulting in increased MN excitability and an increased stretch re¯ex. Although the transcortical effect of muscle nociceptive activity still is unclear (Mense, 1993), cortical activity has been shown to remove presynaptic Ia inhibition and increase a -MN excitability (Lundberg and
VyklickuÂ, 1963; Lund et al., 1965). Nociceptive muscle afferents have also been reported to be able to change the activity of interneurones mediating presynaptic Ia-inhibition and group I non-reciprocal inhibition (Rossi et al., 1999). An argument against changes in MN excitability is the ®nding that the H-re¯ex does not change during salineinduced muscle pain whether in the soleus muscle (Matre et al., 1998), or in the temporalis muscle (Svensson et al., 1998a). There are, however, methodological differences between the stretch and H-re¯ex since both the time course and amplitude of the Ia afferent volley are different between the two stimuli (Burke et al., 1984; Morita et al., 1998). Other factors that may change the transmission of the monosynaptic Ia excitation from the muscle spindle afferents are changes in post-activation depression, and presynaptic inhibition from interneurones mediating presynaptic inhibition of Ia terminals (Hultborn et al., 1996). Changes in reciprocal Ia inhibition are not a likely explanation since the stretch re¯ex is facilitated also with hypertonic saline infusion in the antagonist muscle (Matre et al., 1998). MN responsiveness may also be altered through modulation of intrinsic properties of the a -motoneurone (Windhorst et al., 1997a). Windhorst and co-workers showed that group III± IV afferent activity may affect MN-®ring via changes in membrane potential, changes in afterhyperpolarisation or increased synaptic noise. Furthermore, chemically activated group III±IV afferents may cause disinhibition of the a -MN pool by inhibiting Renshaw cells (Windhorst et al., 1997b). One might speculate whether such a disinhibition of a -MNs by group III±IV activity could contribute to stretch re¯ex facilitation. Another explanation for the increased re¯ex amplitude during experimental muscle pain might be elevation of muscle spindle sensitivity in the relaxed state. Such an increased sensitivity could facilitate the stretch re¯ex by giving a higher Ia afferent burst for a given muscle stretch. A linkage between muscle nociceptive input and g -activity is supported in several animal studies which have shown that injection of various algogenic substances into the muscle increases g -activity (Schmidt et al., 1981; Jovanovic et al., 1990; Johansson et al., 1993; DjupsjoÈbacka et al., 1995; Pedersen et al., 1997), which in turn could increase muscle spindle sensitivity. However, studies in unanaesthetized models are needed to con®rm this. (2) With higher background EMG (both during the ramp contraction and during walking) the stretch re¯ex was not facilitated with pain (Fig. 7). This is in contrast to the normal situation (non-painful muscle) where stretch re¯ex often is higher during contraction compared with rest (Toft et al., 1991; Matre et al., 1998). Several mechanisms contribute to a higher stretch re¯ex during contraction in the nonpainful muscle: the motoneurones are at or above their thresholds, muscle spindle sensitivity is probably increased, and presynaptic inhibition and post-activation depression are both decreased (Hultborn et al., 1987). Altered descending commands converging with afferent input onto common interneurones projecting to motoneurones could also contri-
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
bute to the reduced re¯ex facilitation with pain. It is not clear which of these mechanisms that are reducing the pain-related stretch re¯ex increase with increasing MNexcitability. The increased stretch re¯ex during muscle pain could be explained by a change in each of the above mechanisms when the muscle goes from an active to a relaxed condition. For example, if the increased stretch re¯ex in the relaxed muscle is caused by an increased g -drive that increases the Ia ®ring to a muscle stretch, then, since the increase will take place anyway when the muscle becomes active, the effect of muscle pain on the stretch re¯ex will dim the more the muscle is contracted. Most probably not one single neural mechanism is producing the stretch re¯ex changes seen in the present study, but a combination of effects. Both re¯exmediated pathways and descending pathways in¯uence the stretch re¯ex amplitude. Convergence of these pathways on interneurones projecting to either g - or a -motoneurones may affect the stretch re¯ex amplitude, and the spinal integration between these pathways still needs investigation. 4.3. Post infusion changes Although the analysis failed to ®nd any signi®cant changes in the post-infusion recordings, it is evident that the stretch re¯ex amplitude in the ramp experiment tended to be reduced both 20 and 40 min after infusion compared with before infusion. Since this effect was present both after isotonic and hypertonic saline infusion it could indicate that the stimulus had a non-nociceptive component, possibly increased intramuscular pressure. An argument for this is that the re¯ex amplitude is reduced not only after, but also during the isotonic infusion. Similar reductions in stretch re¯ex amplitude have been found after fatigue (Balestra et al., 1992). None of the subjects in the present ramp experiment reported tiredness after the contractions. However, it cannot be ruled out that the saline infusions produced fatigue-like effects in the absence of any subjective sensation of fatigue. If increased intramuscular pressure is the cause, a possible mediator is group III afferents responding to lowthreshold mechanical stimuli (Ellaway et al., 1982), since group III afferents have been found to inhibit a -motoneurones according to the ¯exor re¯ex pattern (Kniffki et al., 1981). 4.4. Methodological considerations Infusion of hypertonic saline in muscles is a reliable method to induce muscle pain (Kellgren, 1938; Stohler and Lund, 1994; Svensson et al., 1995; Graven-Nielsen et al., 1997a; Matre et al., 1998). The continuous infusion used in the ramp experiment and the subsequent bolus injections used in the other experiments all produced a deep pain sensation in the soleus muscle. The pain sensed by the subject (and indicated on the visual analogue scale) may not linearly re¯ect the true afferent discharge of the muscle nociceptors. The higher pain intensity during the continuous
2041
infusion could be due to the higher volume of infused saline (Graven-Nielsen et al., 1997a). A second explanation could be that pain perception is modulated by the attention of the subject. Experimental studies have shown that attention towards other stimuli alters pain thresholds (Janssen and Arntz, 1996). Walking on a treadmill is a complex task that might take the subject's attention away from the pain. In order to control for the lower pain intensity in the ®rst walking experiment compared with the ramp experiment, a third experiment was designed where both walking and sitting measurements were done after each hypertonic saline injection (Fig. 6C). Half of the subjects started with walking after the ®rst injection, the other half started with sitting. There were no differences in pain intensity between the two situations. In several experimental pain studies infusion of isotonic saline has been used as a control stimulus to hypertonic saline (Ashton-Miller et al., 1990; Vecchiet et al., 1993; Svensson et al., 1995; Graven-Nielsen et al., 1997a; Matre et al., 1998). Isotonic saline produces no or only little pain, and is thus believed to have primarily non-nociceptive (mechanical) effects by increasing the intramuscular pressure (Graven-Nielsen et al., 1997c). The nociceptive component of the afferent input is therefore believed to be the main cause for the increased stretch re¯ex in the relaxed muscle, although it cannot be ruled out that hypertonic saline may produce unknown pain-independent effects. In addition the combined effects of the discharge of nociceptive and non-nociceptive input may be more complex than the sum of the effects caused by the separate activation of these afferents. Despite equal pain intensity and background EMG, the stretch re¯ex was quite variable after the four subsequent bolus injections of hypertonic saline in the combined walking and sitting experiment. Some subjects responded to the injection with facilitation and some with inhibition. On average there was an alternation between inhibition and facilitation during walking, and a reduced facilitation during the second and fourth injection during sitting (Fig. 6A). Most probably infusion of hypertonic saline activates both nociceptive and non-nociceptive lowthreshold mechanosensitive (group II and group III) afferents (Mense, 1993). It has been suggested that the inputs from nociceptive and non-nociceptive afferents have opposite actions on homonymous g -motoneurones (Mense and Skeppar, 1991). Thus, if a different combination of nociceptive and non-nociceptive afferents is stimulated after each injection, this could explain that some subjects exhibited facilitation and others inhibition after the injections. The present methods to elicit stretch re¯exes in the human soleus muscle during sitting (Sinkjñr et al., 1988) and walking (Andersen and Sinkjñr, 1995) have been used in several studies of ankle joint stiffness in healthy and spastic man (Toft et al., 1991; Toft and Sinkjñr, 1993; Sinkjñr et al., 1993; Toft, 1995; Sinkjñr et al., 1996). Both during sitting and walking stretches with different amplitude and velocity may be applied in both directions.
2042
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043
In the combined walking and matched EMG sitting experiment in the present study stretches were applied with equal amplitude, velocity and ankle joint angle at the same level of soleus activity during both walking and sitting. This ensured equal muscle length, and muscle stretch, in the two conditions. 4.5. Summary The mechanically evoked stretch re¯ex during different motor tasks and contraction levels was studied before, during and after infusion of hypertonic saline. The hypertonic saline infusion produced a deep pain sensation in the soleus muscle. At rest the stretch re¯ex was facilitated during the experimentally induced muscle pain. In the active muscle the stretch re¯ex was unaffected by pain. The painrelated changes in stretch re¯ex amplitude seemed to be more related to the excitability of the MN pool than to motor task. One possible explanation for the increased stretch re¯ex amplitude during hypertonic saline infusions in the relaxed muscle, is the linkage between muscle nociceptive afferent activity and an increased g -activity. Other explanations could be reduced presynaptic inhibition of Ia terminals, modulation of intrinsic properties of the a -motoneurones, or reduced Renshaw inhibition. The results show that the re¯ex changes which take place in the relaxed muscle due to the pain stimuli are of much less importance when the muscle is functionally active. Acknowledgements We kindly acknowledge Knud Larsen and Thomas Graven-Nielsen, Aalborg University, Denmark, for their technical assistance. The Danish National Research Foundation are kindly acknowledged for ®nancial support. References Andersen JB, Sinkjñr T. An actuator system for investigating electrophysiological and biomechanical features around the human ankle joint during gait. Trans Rehab Eng 1995;3(4):299±306. Appelberg B, Hulliger M, Johansson H, Sojka P. Actions on gamma-motoneurones elicited by electrical stimulation of group I muscle afferent ®bres in the hind limb of the cat. J Physiol 1983;335:237±253. Arendt-Nielsen L, Graven-Nielsen T, Svarrer H, Svensson P. The in¯uence of low back pain on muscle activity and coordination during gait Ð A clinical and experimental study. Pain 1996;64:231±240. Ashton-Miller JA, McGlashen KM, Herzenberg JE, Stohler CS. Cervical muscle myoelectric response to acute experimental sternocleidomastoid Pain. Spine 1990;15(10):1006±1012. Baldissera F, Hultborn H, Illert M. Integration in spinal neuronal systems. In: Brooks VB, editor. Handbook of physiology (a critical, comprehensive presentation of physilogical knowledge and concepts). Section 1: The nervous system. Vol. II: Motor control, part 1, Bethesda, MD: American Physiological Society. 1981. pp. 509±595. Balestra C, Duchateau J, Hainaut K. Effects of fatigue on the stretch re¯ex in a human muscle. Electroenceph clin Neurophysiol 1992;85:46±52. Bonnet M, Requin J, Stelmach GE. Changes in electromyographic
responses to muscle stretch, related to the programming of movement parameters. Electroenceph clin Neurophysiol 1991;81:135±151. Burke D, Gandevia SC, McKeon B. Monosynaptic and ogliosynaptic contributions to human ankle jerk and h-re¯ex. J Neurophysiol 1984;52(3):435±448. Capaday C, Stein RB. Amplitude modulation of the soleus H-re¯ex in the human during walking and standing. J Neurosci 1986;6(5):1308±1313. DjupsjoÈbacka M, Johansson H, Bergenheim M, SjoÈlander P. In¯uences on the gamma-muscle-spindle system from contralateral muscle afferents stimulated by KCl and lactic acid. Neurosci Res 1995;21:301±309. Ellaway PH, Murphy PR, Tripathi A. Closely coupled excitation of gammamotoneurones by group III muscle afferents with low mechanical threshold in the cat. J Physiol 1982;331:481±498. Graven-Nielsen T, Arendt-Nielsen L, Svensson P, Jensen TS. Experimental muscle pain: A quantitative study of local and referred pain in humans following injection of hypertonic saline. J Musculoskel Pain 1997a;5:49±69. Graven-Nielsen T, Arendt-Nielsen L, Svensson P, Jensen TS. Quanti®cation of local and referred muscle pain in humans after sequential i.m. injections of hypertonic saline. Pain 1997b;69:111±117. Graven-Nielsen T, McArdle A, Phoenix J, Arendt-Nielsen L, Jensen TS, Jackson MJ, Edwards RHT. In vivo model of muscle pain: Quanti®cation of intramuscular chemical, electrical, and pressure changes associated with saline-induced muscle pain in humans. Pain 1997c;69 (1-2):137±143. Graven-Nielsen T, Svensson P, Arendt-Nielsen L. Effects of experimental muscle pain on muscle activity and co-ordination during static and dynamic motor function. Electroenceph clin Neurophysiol 1997d; 105(2):156±164. Hultborn H, Meunier S, Pierrot-Dsesilligny E, Shindo M. Changes in presynaptic inhibition of Ia ®bres at the onset of voluntary contraction in man. J Physiol 1987;389:757±772. Hultborn H, Illert M, Nielsen J, Paul A, Ballegaard M, Wiese H. On the mechanism of post-activation depression of the H-re¯ex in human subjects. Exp Brain Res 1996;108(3):450±462. Janssen SA, Arntz A. Anxiety and pain: attentional and endorphinergic in¯uences. Pain 1996;66:145±150. Johansson H, Sojka P. Pathophysiological mechanisms involved in genesis and spread of muscular tension in occupational muscle pain and in chronic musculoskeletal pain syndromes: A hypothesis. Med Hypotheses 1991;35:196±203. Johansson H, DjupsjoÈbacka M, SjoÈlander P. In¯uences on the gammamuscle-spindle system from muscle afferents stimulated by KCl and lactic acid. Neurosci Res 1993;16:49±57. Jovanovic K, Anastasijevic R, Vuco J. Re¯ex effects on gamma-fusimotor neurones of chemically induced discharges in small-diameter muscle afferents in decerebrate cats. Brain Res 1990;521:89±94. Kellgren JH. Observations on referred pain arising from muscle. Clin Sci 1938;3:175±190. Kniffki K-D, Schomburg ED, Steffens H. Synaptic effects from chemically activated ®ne muscle afferents upon a-motoneurones in decerebrate and spinal cats. Brain Res 1981;206:361±370. Lund S, Lundberg A, Vyklicku L. Inhibitory action from the ¯exor re¯ex afferents on transmission to Ia afferents. Acta Physiol Scand 1965;64: 345±355. Lund JP, Donga R, Widmer CG, Stohler CS. The pain-adaptation model: a discussion of the relationship between chronic musculoskeletal pain and motor activity. Can J Physiol Pharmacol 1991;69:683±694. Lundberg A, Vyklicku L. Inhibitory interaction between spinal re¯exes to primary afferents. Experientia 1963;19:247±248. Matre DA, Sinkjñr T, Svensson P, Arendt-Nielsen L. Experimental muscle pain increases the human stretch re¯ex. Pain 1998;75(2±3):331±339. Mense S. Review article: Nociception from skeletal muscle in relation to clinical muscle pain. Pain 1993;54:241±289. Mense S, Skeppar P. Discharge behaviour of feline gamma-motoneurones following induction of an arti®cial myositis. Pain 1991;46:201±210. Meunier S, Pierrot-Deseilligny E. Gating of the afferent volley of the
D.A. Matre et al. / Clinical Neurophysiology 110 (1999) 2033±2043 monosynaptic stretch re¯ex during movement in man. J Physiol 1989;419:753±763. Morita H, Petersen N, Christensen LOD, Sinkjñr T, Nielsen J. Sensitivity of H-re¯exes and stretch re¯exes to presynaptic inhibition in humans. J Neurophysiol 1998;80(2):610±620. Nielsen J, Sinkjñr T, Baumgarten J, Andersen JB, Toft E, Christensen LOD, Ladoceur M, Morita H. Modulation of the soleus stretch re¯ex and H-re¯ex during human walking after block of peripheral feedback. Neurosci Abstr 1998;24:2103. Pedersen J, SjoÈlander P, Wenngren BI, Johansson H. Increased intramuscular concentrations of bradykinin increases the static fusimotor drive to muscle spindles in neck muscles of the cat. Pain 1997;70(1):83±91. Prochazka A, Hulliger M, Trend P, DuÈrmuÈller N. Dynamic and static fusimotor set in various behavioural contexts. In: Hnik P, Soukup T, Vejsada R, Zelena J, editors. Mechanoreceptors. New York: Plenum, 1988. pp. 417±430. Rossi A, Decchi B, Ginanneschi F. Presynaptic excitability changes of group Ia ®bres to muscle nociceptive stimulation in humans. Brain Res 1999;818:12±22. Schmidt RF, Kniffki K-D, Schomburg ED. Der Ein¯uss Klinkalibriger Muskelafferenzen auf den Muskeltonus. In: Bauer H, Koella WP, Struppler H, editors. Therapie der Spastik, Munchen: Verlag fur angewandte Wissenschaft. 1981. pp. 71±86. Schomburg ED. Review article: Spinal sensorimotor systems and their supraspinal control. Neurosci Res 1990;7:265±340. Sinkjñr T. Muscle, re¯ex and central components in the control of the ankle joint in healthy and spastic man. Acta Neurol Scand 1997;96(Suppl 170):1. Sinkjñr T, Toft E, Andreassen S, Hornemann BC. Muscle stiffness in human ankle dorsi¯exors: intrinsic and re¯ex components. J Neurophysiol 1988;60:1110±1121. Sinkjñr T, Toft E, Larsen K, Andreassen S, Hansen HJ. Non-re¯ex and re¯ex mediated ankle joint stiffness in multiple sclerosis patients with spasticity. Muscle Nerve 1993;16:69±76. Sinkjñr T, Andersen JB, Larsen B. Soleus stretch re¯ex modulation during gait in humans. J Neurophysiol 1996;76(2):1112±1120. Stein RB, Yang J, Edamura M, Capaday C. Re¯ex modulation during normal and pathological human walking. In: Shimamura M, editor. Neurobiological basis of human locomotion, Tokyo: Jpn Sci Soc, 1991. pp. 335±346. Stohler CS, Lund JP. Effects of noxious stimulation of the jaw muscles on
2043
the sensory experinece of volunteer human subjects. In: Stohler CS, Carlson DS, editors. Biological & physiological aspects of orofacial pain. Craniofacial growth Series 29, Center for Human Growth & Development. The University of Michigan, 1994. pp. 55±73. Svensson P, Arendt-Nielsen L, Nielsen H, Larsen JK. Effect of chronic and experimental jaw muscle pain on pain-pressure thresholds and stimulusresponse curves. J Orofacial Pain 1995;9:347±356. Svensson P, Houe L, Arendt-Nielsen L. Bilateral experimental muscle pain changes electromyographic activity of human jaw-closing muscles during mastication. Exp Brain Res 1997;116:182±185. Svensson P, DeLaat A, Graven-Nielsen T, Arendt-Nielsen L. Experimental jaw-muscle pain does not change heteronymous H-re¯exes in the human temporalis muscle. Exp Brain Res 1998a;121(3):311±318. Svensson P, Graven-Nielsen T, Matre DA, Arendt-Nielsen L. Experimental muscle pain does not cause long-lasting increases in resting EMG activity. Muscle Nerve 1998b;21:1382±1389. Taylor J, Stein RB, Murphy PR. Impulse rates and sensitivity to stretch of soleus muscle spindle afferent ®bers during locomotion in premammilary cats. J Neurophysiol 1985;53(2):341±360. Toft E. Mechanical and electromyographical stretch responses in spastic and healthy subjects. Acta Neurol Scand 1995;92(Suppl 163):1. Toft E, Sinkjñr T. H-re¯ex changes during contractions of the ankle extensors in spastic patients. Acta Neurol Scand 1993;88:327±333. Toft E, Sinkjñr T, Andreassen S, Larsen K. Mechanical and electromyographic responses to stretch of the human ankle extensors. J Neurophysiol 1991;65(6):1402±1410. Vecchiet L, Dragani L, Bigontina P, Obletter G, Giamberardino MA. Experimental referred pain and hyperalgesia from muscles in humans. In: Vecchiet L, Albe-Fessard D, Lindblom U, editors. New trends in referred pain and hyperalgesia, Amsterdam: Elsevier, 1993. pp. 239± 249. Wilmink RJH, Slot PJ, Sinkjñr T. Modelling of the H-re¯ex facilitation during ramp and hold contractions. J Comput Neurosci 1996;3(4):337± 346. Windhorst U, Kirmayer D, Soibelman F, Misri A, Rose R. Effects of neurochemically excited group III-IV muscle afferents on motoneuron afterhyperpolarization. Neuroscience 1997a;76(3):915±929. Windhorst U, Meyer-Lohmann J, Kirmayer D, Zochodne D. Renshaw cell responses to intra-arterial injection of muscle metabolites into cat calf muscles. Neurosci Res 1997b;27(3):235±247.