METHODOLOGICAL CONSIDERATIONS AND THE EFFECT OF PAIN ON THE H-REFLEX AND MAXIMAL M-WAVE IN THE HUMAN TRICEPS SURAE DOCTOR OF PHILOSOPHY

Transcription

1 METHODOLOGICAL CONSIDERATIONS AND THE EFFECT OF PAIN ON THE H-REFLEX AND MAXIMAL M-WAVE IN THE HUMAN TRICEPS SURAE A thesis submitted for the degree of DOCTOR OF PHILOSOPHY by Kylie Jane Tucker BA, BSc Hons (Physiolgy) Research Centre for Human Movement Control Discipline of Physiology School of Molecular and Biomedical Science The University of Adelaide June

2 Table of Contents Table of Conents... I List of Figures... VII Abstract... IX Declaration... X Publications... XI Acknowledgements... XII Summary... XIII Table of Contents 1. Introduction Neuromuscular Physiology Motor units and fibre types Recording motor unit activity Modulation of motor units Central control of motor units Peripheral control of motor units Interneurones Hyperpolarisation of neurons Renshaw cells Fatigue Electrical Evoked Responses M-wave Antidromic action potentials and Renshaw cells The F-wave The H-reflex Modulation of the H-reflex Voluntary contraction After Hyperpolarisation and absolute refractory periods Postural variation Cortical influences Training Age Other methodological considerations I

3 Table of Contents 1.3. The Triceps Surae Structure of the muscles of the triceps surae Function of the muscles of the triceps surae Motor points Experimental Considerations in the Triceps Surae Common protocol Placement of the stimulating electrodes Stimulation duration, frequency and variability Use of constant current or voltage stimulator Recording with monopolar v bipolar surface electrodes Recording with surface v intramuscular electrodes Noise reduction Analysis methods Pain Painful stimuli and the perception of pain The effect of pain on muscle activity Pain and background muscle activity Pain and motor unit firing rate Pain and reflex strength Pain and EMG spectral analysis Pain and the recruitment order of muscles Conclusions and Directions Muscle spindle feedback differs between the soleus and gastrocnemius in humans Outline of study Introduction Materials and Methods Subjects Protocol Recording Feedback of voluntary contraction level Stimulus Analysis Statistics Results H-reflex M-wave II

4 Table of Contents Gastrocnemius versus soleus Background Muscle Activity Discussion Gastrocnemius and soleus motoneurons are recruited differently Primary afferent depolarisation Recurrent inhibition Hyperpolarisation Changes in M max during contraction studies Further discussion points and limitations Background muscle activity M-response threshold Voltage stimulator Stimulus rate Significance A new method to estimate signal cancellation in the human maximal M-wave Outline of study Introduction Materials and Methods Subjects Protocol Recording Feedback of voluntary contraction level Stimulus Analysis Statistics Results Maximal M-wave magnitude at different electrode placements Maximal M-wave magnitude during rest and contraction Area vs PTP amplitude analysis methods Discussion Monopolar vs bipolar recording and the signal cancellation error Signal cancellation and fibre type variability Monopolar vs bipolar recording and changes in the maximal M-wave magnitude Area vs PTP amplitude analysis: Dispersion of action potentials Methodological considerations Plantar flexion III

5 Table of Contents Electrode type and configuration Spatial filtering End plate potentials and tendon potentials Common mode rejection and cross talk Innervation zones Conclusions and significance Triceps surae stretch and voluntary contraction alters maximal M-wave magnitude Outline of study Introduction Materials and Methods Protocol Recording Feedback of voluntary contraction level Stimulus Analysis Statistics Results Maximal M-wave magnitude during different muscle conditions Maximal M-wave magnitude: bipolar vs monopolar recording Shape of maximal M-wave Discussion The M-wave and muscle condition: monopolar vs bipolar recording Changes in muscle geometry during muscle stretch and contraction Area vs PTP amplitude analysis: Dispersion of action potentials Methodological considerations Conclusions and significance Hypertonic saline induced pain does not affect agonist or antagonist H-reflex strength in the triceps surae Chapter Outline Introduction Materials and Methods Protocol Electrical Stimulus Recording IV

6 Table of Contents Analysis Statistics Results Pain scores Background muscle activity and H-reflex size Foot plate force record Discussion Pain levels throughout the trials Background muscle activity during symmetrical standing Background muscle activity during pain Weight distribution during pain and non pain states H-reflex modulation during pain Implications and conclusions Concluding remarks References Appendix V

7 List of Figures List of Figures Figure 1: Electrical stimulation of the mixed nerve Figure 2: Position of the popliteal fossa and innervation zones of the triceps surae Figure 3: Average trace, average rectified trace and CUSUM of M-wave and H-reflex Figure 4: Average of 3 raw data traces of SEMG for one subject throughout the stimulus range Figure 5: Normalised H/M max curves Figure 6: Maximal M-wave for each contraction level, normalised to overall maximal M-wave Figure 7: Mean SEM of H/M max, data pooled from 8 subjects Figure 8: Correlation between levels of normalised pre-stimulus background muscle activity in the gastrocnemius and soleus Figure 9: Bipolar and monopolar electrode arrangement Figure 10: Maximal M-wave magnitude at each electrode arrangement Figure 11: Magnitude of the maximal M-wave obtained during all muscle contraction trials Figure 12: Representation of the signal cancellation error Figure 13: Simulation and actual data of signal cancellation that occurs during bipolar recording Figure 14: Ankle positions during test conditions Figure 15: Mean and SD, of normalised maximal M-wave Figure 16: Maximal M-wave magnitude during monopolar vs bipolar recording Figure 17: Maximal M-wave: bipolar vs monopolar recording Figure 18: Raw and curve fitted H-reflex and M-wave curve Figure 19: Isotonic and hypertonic saline injection Figure 20: Pain scores throughout pain trials VI

8 List of Figures Figure 21: Area of pain during hypertonic saline (pain) trials Figure 22: Background muscle activity during control, sham and pain trials Figure 23: H/M max during control, sham and pain trials Figure 24: Weight distribution during control, sham and pain trials Figure 25: Modelling the H-reflex and M-wave VII

9 Abstract Abstract The control of skeletal muscle relies on a complex integration between descending central input and information that originates from receptors that lie within peripheral tissue. The following investigations were performed to contribute to our understanding of this control. Study 1 (Chapter 2) was designed to determine (using the H-reflex) if muscle spindle feedback is similar in the gastrocnemius and soleus. The strength of the H-reflex at rest and during contraction was compared between muscles. The results showed that the maximal H-reflex obtained at any level of contraction is larger in the soleus than in the gastrocnemius. We argue that along with the muscles having different structures and functions, the recruitment capabilities of their motoneurons are quite different. We also found that the maximal M-wave, which has for years been thought to be a consistent measure of maximal muscle activity, was quite variable within subjects during different conditions. Review of the maximal M-wave literature showed evidence that variability in this response did exist between conditions, but that the variability was rarely seen in pooled data, and was therefore not often reported. Study 2 (Chapter 3) was developed to determine if experimental recording techniques, or analysis methods, could affect the magnitude of the maximal M-wave within subjects. The first finding of this study showed that the two most commonly used analysis methods (peak-topeak amplitude and area) provided comparable results, and could not account for the differences seen in the maximal M-wave magnitude. The study did however suggest that the orientation of surface recording electrodes can significantly alter the recorded signal. We argue that although bipolar surface recording is considered superior to monopolar recording in its ability to record a clean signal, it has a large limiting factor, which we call signal cancellation. The third study (Chapter 4) focused on the variability in M-wave strength in the gastrocnemius and soleus during a variety of ankle orientations and voluntary contraction levels. This study supported our previous work, and showed that when monopolar recording is used, consistent and significant differences exist in the VIII

10 Abstract strength on the M-wave obtained during different conditions that were not seen in bipolar recordings. It was concluded that the difference in maximal M-wave strength obtained during different muscle conditions may be related to a change in the recording electrode to muscle bulk relationship. This finding is important as M-wave strength is consistently used as a normalisation factor in reflex studies, and therefore variability in this measure may seriously affect the results obtained during muscle reflex investigations. The final study (chapter 5) considered the size of the H-reflex, the level of background muscle activity, and the subjects weight distribution, during painful and non-painful conditions. We determined that these factors were not modified by pain induced in either agonist or antagonist muscles. The final chapter outlines the major findings from this work, highlights limitations to the research conducted using the H-reflex, and makes suggestions for future research in this area. IX

11 Declaration Declaration This work contains no material which has been accepted for the award of any other degree or diploma in any university or other tertiary institution, and to the best of my knowledge and belief, contains no material previously published or written by another person, except where due reference has been made in the text. I give consent to this copy of my thesis, when deposited in the University Library, to be available for loan and photocopying. I acknowledge that the copyright of the published articles included as an appendix in this thesis resides with the copyright holders of those works. Signed: Date: X

12 Publications Publications The following is a list of peer reviewed journal publications arising from this thesis to date. Each article is included in the appendix of this thesis. Chapter 1: Tucker KJ, Tuncer M, Türker KS. A review of the H-reflex and M-wave in the human triceps surae. Hum Mov Sci. 2005; 24(5-6): Chapter 2: Tucker KJ, Türker KS. Muscle spindle feedback differs between the soleus and gastrocnemius in humans. Somatosensory and Motor Research 2004; 21(3-4): Chapter 3: Tucker KJ, Türker KS. A new method to estimate signal cancellation in the human maximal M-wave. Journal of Neuroscience Methods. 2005; 149(1):31-41 Chapter 4: Tucker KJ, Türker KS. Triceps surae stretch and voluntary contraction alters maximal M-wave magnitude. Journal of Electromyography and Kinesiology 2006 (In Press) Further possible publications arising from this thesis include: Chapter 5: (Currently in submission) Tucker KJ, Brinkworth, RS, and Türker KS. Hypertonic saline induced pain does not affect agonist or antagonist H-reflex strength in the triceps surae. XI

13 Acknowledgements Acknowledgements The work contained in this thesis was conducted in the Oral Neurophysiology Laboratory in the University of Adelaide under the direction of Associate Professor Kemal Türker. The laboratory funding came from a grant supplied by the National Health and Medical Research Council of Australia (NH&NRC). Part of my salary was obtained from an Australia Postgraduate Award scholarship. It is very important at this point to acknowledge the wonderful team of collaborators, workmates and friends that have helped to make this possible. Firstly I must thank my supportive and caring supervisor, Associate Professor Kemal Türker. Kemal has not only provided me with the scientific background and research direction that I have needed to accomplish this goal, but also the support of a friend. I am grateful to have had the opportunity to work with Kemal in different roles over the last 8 years. Dr Russell Brinkworth, (soon to be Dr) Paul Sowman and I shared a very small space for the majority of the time spent on my PhD. With many people this would have been difficult; however I believe that we have all come out the other end better off for the experience. I would not have been able to complete this work without the technical, statistical, language and general research skills of these two brilliant men. I want to also mention the extended motor control research group at Adelaide University. Whether it be morning cake break, company at the conferences, general conversations, or sometimes even scientific discussions, Paul, Stan, Michelle, Sophie, Martin and Gabby have been inspirational friends and work mates. Finally, thanks goes to my family, who have always supported me with what ever decision I have made, and in particular to my sister Megan, who has washed more dishes, cooked more meals, done more shopping and turned on the washing machine many more times than me over the past few years. XII

14 Summary Summary The control of skeletal muscle relies on complex communication between the central and peripheral nervous system. Constant integration in the spinal cord of information that originates from receptors lying within many peripheral tissues and continual descending central input, contribute to ensure that the level on ongoing muscle activity is appropriate for each condition. A large amount of research has been conducted into this complex integrative system. With every answer however, more questions are posed. The following investigations were performed to add to this body of research. When these investigations began, the main aim of my PhD was to investigate how musculo-skeletal pain may affect the control of muscle activity. The first investigation was designed to determine if pain alters muscle activity at the spinal cord level. To do this the Hoffmann reflex (H-reflex described in chapter 1) was elicited in the triceps surae of asymptomatic human subjects, before, during, and following, the induction of short-term pain. The results of this preliminary investigation were quite varied, and this variability led to a change in direction of the research. Through discussions with my supervisor it was determined that some of the variability in the results may have been due to the different structure and function of the muscles that make up the triceps surae. Subsequently, the first investigation that is described in this thesis was proposed. Study 1 (Chapter 2), was designed to determine if the muscle spindle input to the gastrocnemius and soleus muscles were different. In this investigation the strength of the H-reflex during rest and different levels of contraction were compared between muscles. The results showed that the maximal H-reflex obtained at any level of contraction is larger in the soleus than in the gastrocnemius. The study also showed XIII

15 Summary that while the size of the maximal H-reflex increased steadily with increased contraction in the gastrocnemius to approximately 50% of the maximal muscle response (maximal M-wave), the maximal H-reflex in the soleus peaked, when it reached almost 100% of its maximum possible size, at just 40% maximal voluntary contraction. The findings of this study suggest that the along with the muscles having quite different structures and functions, the recruitment capabilities of the motoneurons are also quite different. The other main finding of this study was that the maximal M-wave, which has for years been thought to be a consistent measure of maximal muscle activity, was quite variable within subjects at different levels of voluntary contraction. Further review of the maximal M-wave literature showed that some variability in the response did exist between conditions within subjects, but that the variability was rarely seen in the pooled data. This intriguing finding led to the second study presented in this thesis. Study 2 (Chapter 3) was developed to determine if the experimental recording techniques, or analysis methods commonly used in human muscle response investigations, could further change the magnitude of the reported maximal M-wave within subjects. To investigate this, a variety of techniques were utilized to record and analyse the maximal M-wave as it was elicited in the gastrocnemius and soleus of human subjects. The first finding of this study showed that the two most commonly used analysis methods (peak-to-peak amplitude and area) provided comparable results, and could not account for the differences seen in the maximal M-wave magnitude between conditions. The study did however suggest that the orientation of surface recording electrodes could significantly alter the signal that is recorded in such studies. It was concluded from this study that although bipolar surface recording is XIV

16 Summary considered superior to monopolar recording in its ability to record a clean signal, it has a large limiting factor, which we called signal cancellation. Signal cancellation occurs when the second negative electrode is recording part of the muscle response, before the wave has completely passed the first positive electrode. This study showed that bipolar recording reduced the signal strength by up to 80% of its actual size, and that when this occurred no consistent variability existed between experimental conditions. In contrast, when monopolar recording electrodes were used, a consistent and significant difference in maximal M-wave strength was seen between the resting and voluntary contracting test muscles. This finding reinforces the importance of using the correct tools, in particular when recording large muscle responses during different conditions. The finding also led to the third study (Chapter 4), which focused on the variability in M-wave strength in the gastrocnemius and soleus muscles during a variety of different ankle orientations and voluntary contraction levels. This study supported our previous work, and showed that when monopolar recording is used (as opposed to bipolar recording electrodes), consistent and significant differences existed in the strength of the M-wave obtained during different muscle conditions. It was concluded that the difference in maximal M-wave strength obtained during different muscle conditions may be related to a change in the recording electrode to muscle bulk relationship. This finding is important as the M-wave strength is consistently used as a normalising factor in many reflex studies, and therefore variability in this measure may seriously affect the results obtained during many reflex investigations. Once the first three studies had been concluded, we decided to reassess the initial question of this thesis. Does pain affect muscle activity by influencing the excitability of the H-reflex circuit? From the results of the first study, it had been determined that XV

17 Summary the gastrocnemius and soleus muscles were quite different in their recruitment tendencies, so both muscles were to be recorded simultaneously and independently. Furthermore both monopolar and bipolar recording electrodes were used to record the H-reflex and M-wave to determine if the results would be affected by the signal cancellation error as seen in the M-wave during study 2 and 3. The final study considered both the size of the H-reflex, the level of background muscle activity, and the weight distribution of subjects during control, painful and sham conditions. Subjects stood comfortably throughout the duration of the investigation, which ensured some postural background muscle activity. The major findings from this study were that the use of hypertonic saline induced pain in either agonist or antagonist muscles in the leg did not alter the H-reflex size in the soleus or medial gastrocnemius. It also showed that this kind of pain did not significantly alter the level of background muscle activity in the soleus, medial gastrocnemius or tibialis anterior. Finally, the study also showed no change in the weight distribution of the subjects on the test leg between the different experimental conditions. This work supports some, but refutes other data published in the area of pain and the H-reflex. We suggest that the variability in the results seen by different authors may be specifically due to the degree of motor control required to complete the tasks required of the test muscle during the experimental conditions. We therefore suggest that greater changes in the H-reflex and background muscle activity may occur in the test muscle when it is completing more complex tasks during pain compared to nonpain conditions. Due to the conflicting results in this area, much further research is required before definite conclusions can be drawn as to the location and the extent of the influence that pain may have on the human H-reflex circuit. All of the work completed as part of this thesis helps to improve our understanding of the methodological considerations required to maintain a high standard of research XVI

18 Summary when studying the H-reflex and M-wave. It also adds to our understanding of the influence that both voluntary contraction and pain has on the human motoneuron pool, with particular focus on the triceps surae. XVII

19 H-reflex and M-wave in the Triceps Surae 1. Introduction Parts of this chapter are an edited version of the manuscript A review of the H-reflex and M-wave in the Triceps Surae by K.J Tucker, M. Tuncer and K.S. Türker, which has been published in Human Movement Science (2005); 24(5-6): pgs A copy of this article is included in the appendix of this thesis Neuromuscular Physiology The control of skeletal muscle relies on complex communication between the central and peripheral nervous system, including the integration of information from receptors that lay within many peripheral structures. The resulting activity that occurs in the destination muscle also depends on the functional characteristics of the muscle fibre types, and that of the surrounding tissues. To accurately introduce the topic of this thesis, a comprehensive understanding of the neurological components of the reflex loop and a basic understanding of different skeletal muscle fibre properties is required Motor units and fibre types A motor unit describes the smallest functional unit of the motor system. The single motor unit (SMU) comprises of one motoneuron and all of the muscle fibres that the motoneuron supplies (Calancie and Bower, 1990; Schmidt, 1985). The group of motoneurons that control an entire muscle is referred to as the motoneuron pool for that muscle. Studies on limb muscles have shown that all muscle fibres of a motor unit lie within a single muscle (Stuart and Enoka, 1983) and that there is a one-to-one relationship 1-18

20 H-reflex and M-wave in the Triceps Surae between the discharge of a motoneuron and all the muscle fibres it controls (Bigland- Ritchie et al., 1979). In addition, it is known that each muscle fibre is innervated by only one motoneuron (Burke, 1981) and each fibre within a motor unit has similar metabolic properties (Burke et al., 1971; Nemeth et al., 1986), although the muscle fibres of a given motor unit are widely scattered over a cross-section of the muscle (Nemeth et al.,1981). Skeletal muscles contain three distinct types of motor units, which allow varied contraction force outputs and durations, which is important given the diverse range of activities that skeletal muscles are required to achieve. The three distinct groups of muscle fibre types (and therefore motor unit types) are slow-twitch (S), fast fatiguing (FF) and fatigue resistant (FR) (Burke et al., 1973; Taylor, 1976). These muscle fibre types can be distinguished by force output, contraction duration and by colour in histochemical studies. The slow-twitch (S or type I) muscle fibres are reddish in colour, they produce relatively small contractile forces and are resistant to fatigue. These fibres are able to maintain a constant contractile force for very long periods. S fibres may be further categorised to those with greater fatigable (SF) or fatigue resistant (SFR) qualities (Shall and Goldberg, 1992). S fibres are oxygen dependant and as such are generally found in the deep parts of muscles where the blood supply is greatest. They are innervated by small motoneurons and are therefore the first to be recruited during voluntary contraction due to the large input resistance in the motoneuron (Henneman, 1985). The fast, fatigue-resistant (FR or type IIA) fibres can use both oxygen and glycogen for fuel (Buchthal and Schmalbruch, 1980). FR fibres are innervated by intermediately sized motoneurons and are recruited during moderately intensive 1-19

21 H-reflex and M-wave in the Triceps Surae tasks. They produce relatively fast and strong twitches and while they are fatigue resistant the contractile force does decrease over time. The fast, fatigable (FF or type IIB) fibres are glycolytic and are generally found on the superficial layer of skeletal muscles. FF fibres are innervated by the largest motoneurons and are only recruited during powerful voluntary contractions; they produce the fastest and strongest twitches (about 10 x the force produced by S fibres) but fatigue quickly (Burke, 1968; Brooke and Kaiser, 1970; Burke et al., 1973). The large motoneurons that innervate these muscle fibres, have a lower resistance than smaller motoneurons, and therefore have a faster action potential conduction time along the nerve axon. The diameter of a motor nerve (axon of the motoneuron) is directly related to the type of muscle fibres recruited (S<FR<FF; Burke, 1991) and number of muscle fibres it supplies (Henneman et al., 1965). Individual motor units can therefore vary significantly in size and functional capacities which relates directly to the functional requirements of the skeletal muscle and the input current required to stimulate these neurons (Wuerker et al., 1965; Burke et al., 1973; McPhedran et al., 1998). The size of the smallest motoneurons therefore also predict the stimulation threshold, the type of muscle fibers, and the number of the muscle fibers it innervates. This size principle dictates the order of recruitment of motoneurons as has been shown during investigations using both the stretch reflex and H-reflex (see below) (Henneman et al., 1965) Recording motor unit activity The activity of SMU s can be recorded in human experiments by using intramuscular electrodes. The ability to make a SMU fire is the best indication of motoneuron excitability in human experiments, because if the motoneuron fires an action potential 1-20

22 H-reflex and M-wave in the Triceps Surae along its axon, the SMU will always fire the same way. This is called the all or nothing twitch response. The size and shape of a recorded SMU twitch is determined by a small fraction of all fibres in the motor unit. The spike component represents the temporal summation of the action potentials of fibres closets to the electrode, mainly within a radius of 0.5 mm. The shape of the spike part is determined by the synchronicity of the action potentials of the most adjacent muscle fibres. Therefore, they give only limited information about individual muscle fibres and no information about number of muscle fibres of the entire motor unit (Stålberg, 1986). Although intramuscular recording is very useful in many experiments, its ability to only record from a small fraction of the test muscle, excludes it from being used in the investigations described in this thesis Modulation of motor units The output of a motoneuron is a result of the net effect of a complex process of integration (excitatory and inhibitory) of inputs from many other neurons. Therefore some exceptions to the size principle (in the order of the recruited motor untis) can occur due to variations in the mixture of excitatory and inhibitory impulses that synapse onto the motoneuron (Henneman et al., 1965; Burke et al., 1970; Calancie and Bawa, 1990; Burke, 1991; Semmler and Türker, 1994). Several studies have shown for example that cutaneous input (Kanda et al., 1977; Garnett and Stephens, 1981; Kanda and Desmedt, 1983) and pain can alter the recruitment order of motor units (Burke, 1973; Burke et al., 1981). Also, the recruitment of a new motor unit may have an inhibitory influence on the firing rates of previously activated motor units. For example, when large (FF) motor units are recruited, the smaller (S) units are inhibited by Renshaw inhibition (discussed in 1-21

23 H-reflex and M-wave in the Triceps Surae sections below). It is therefore understood that selective recruitment exists within a motor unit pool, which activates only the muscle units that have appropriate mechanical characteristics to suit the specific demand of a task. The total muscle output that results during specific tasks is a result of both the number of motor units recruited and the firing rates of those motor units, which are directly influenced by the all the factors discussed above. Such mechanisms are functionally useful in providing smooth control of the muscle output via peripheral circuitry, and lessening the amount of control required by the central nervous system over the α-motoneuron pool during normal movement (Broman et al., 1985) Central control of motor units The excitability of a motoneuron is reliant on the bombardment of excitatory and inhibitory input that it receives from both peripheral and central origins. In man a profound influence upon any motoneuronal pool is the tonic facilitatory influence which originates from the motor centres in the brain, the cortex, cerebellum and reticular formation (Greenwood and Sessle, 1976). An example of the large role that supra spinal presynaptic inhibition (the influence of supraspinal centers on spinalmediated presynaptic inhibition) plays in the regulation of normal muscle activity is particularly well described in studies detecting changes in muscle activity following spinal cord injury (Calancie et al., 1993). Other variable supra spinal influences on normal muscle activity are shown during different mental tasks. Studies have shown for example, that stress induced by the anticipation of pain, can markedly reduce an inhibitory reflex in the human jaw, while not altering the level of background muscle activity (Burke et al., 1980; Scott and Cadden, 1996), and that level of alertness or drowsiness can alter motoneuron 1-22

24 H-reflex and M-wave in the Triceps Surae excitability as observed by changes in the H-reflex (Táborìková and Sax, 1968) (Hreflex is discussed in 1.2.4). The process of fatigue can also be partially regulated by central mechanisms. The central mechanisms of fatigue are known to reduce the discharge rate of motoneurons (Gandevia, 1992), and by doing so, this central fatigue reduces the maximal force able to be generated by a muscle. (The mechanisms of fatigue are discussed further in section below). Central mechanisms are not effective controllers of skeletal muscle activity alone however, and subsequently proprioceptive input (eg muscle spindle activity) is constantly required to monitor kinaesthetic sensations, and update the centrally generated programme for movement (Burke, 1988) Peripheral control of motor units The muscle spindle is the primary sensory structure within a muscle. The spindle system provides approximately one-third of the total peripheral input to postural muscle motoneurons, and is especially important in maintaining tonus and static posture in these muscles throughout the day (Macefield et al., 1993; Eriksson et al., 1984). They receive motor innervation from the fusimotor system (gamma motor system), which keeps the spindle in an active state, and therefore ready to fire at any particular length or tension. The activity of a muscle spindle is a direct example of central mechanisms (the fusimotor system) being monitored and altered by a peripheral feedback control mechanism. Cutaneous receptors (Knikou and Rymer, 2002), tendon organs (Pierrot-Deseiligny et al., 1981; Burke et al., 1984), temperature and pain receptors (McMullan, 2004; Graven-Nielsen et al., 1997b, 1998) are also important peripheral modulators of 1-23

25 H-reflex and M-wave in the Triceps Surae motor unit activity. Information from both the central and peripheral influences converge at the spinal cord level, and it is the net effect of all the inhibitory and excitatory inputs to the motoneuron that determines its ability to fire Interneurones Interneurones in the spinal cord synapse on alpha and gamma motoneuron pathways, and primary afferent terminals (as well as autonomic efferent neurons). The interneurones have an important integrative function, and allow the combination of a great diversity of afferent and central information to produce a single response (Jankowska and Lundberg, 1981). Modulation can occur directly on the motoneuron (postsynaptic: see Funase and Miles, 1999), directly on the primary afferent (presynaptic - primary afferent depolarisation (PAD): see Pierrot-Deseilligny, 1997), or at the interneuronal level (postsynaptic and presynaptic) within the central nervous system. The type of interneuron (inhibitory or excitatory) determines if an excitatory postsynaptic potential (EPSP) or inhibitory postsynaptic potential (IPSP) will be created in the postsynaptic neuron. It is the combination of all the neuronal inputs that determine if a post synaptic neuron will fire. It is also important to note that primary afferent fibres (originating from muscle spindles) can affect motoneuron-firing frequency though oligosynaptic (many synapses) pathways (Pierrot-Deseilligny et al., 1981; Burke et al., 1984). This long latency effect may be responsible for both inhibitory and excitatory effects, determined by the type of interneurons that complete the oligosynaptic pathway Hyperpolarisation of neurons Hyperpolarisation of a neuron occurs immediately following an action potential. This hyperpolarisation is a direct result of the flow of ions over the nerve membrane. The 1-24

26 H-reflex and M-wave in the Triceps Surae hyperpolarisation occurs when the membrane potential is further from its firing threshold than the resting membrane potential. During this period it is more difficult for subsequent inputs to fire another action potential (see also below). In anaesthetised animal trials, the membrane potential, following an intracellular injection of current, increases linearly up to the threshold where an action potential is fired (Schwindt and Calvin, 1972). In man however, this trajectory is not linear because of the constant central and peripheral inputs to a motoneuron. This fundamental difference between the anaesthetised animal and normal human models, limits the ability for many animal findings to be automatically implied in humans Renshaw cells Total muscle output or muscle force is a result of both the number of motor units recruited and the firing rates of those motor units. When large forces are required, larger motor units are recruited. The recruitment of these larger units is known to inhibit the recruitment of smaller motor units Friedmann et al., (1981). This process, known as Renshaw inhibition, ensures that only those units required for the activity are active, and those motor units whose effect would be redundant during certain activities are not firing unnecessarily Fatigue Fatigue is generally noted as a reduction in the force output from a motor unit, as the result of repeated activation (Burke, 1981). Fatigue can be the result of the failure of neuromuscular transmission, the muscle fibre action potential, excitation-contraction coupling or the contractile machinery itself (Gandevia, 1991). 1-25

27 H-reflex and M-wave in the Triceps Surae The two types of fatigue are peripheral and central. Central fatigue is a progressive reduction in the voluntary activation of a muscle during exercise and is caused by a reduction in the frequency of motoneuron firing, or a reduction in the number of motor units recruited. Peripheral fatigue is produced by changes at, or distal to, the neuromuscular junction, and follows changes in the local ph, an increase in lactic acid concentration and t-tubule failure (Gandevia, 2001). Given that fatigue will reduce the muscles potential output, it is important to ensure that the test muscle does not fatigue throughout the duration of an experiment (if fatigue is not being tested) Electrical Evoked Responses The Hoffmann (H) reflex and Motor (M) wave are initiated by passing an electrical current through a mixed peripheral nerve that includes muscle spindle afferents and motor efferents of the test muscle. To elicit these responses in the human triceps surae the electrical stimulus (of approximately ma) is passed through the tibial nerve in the popliteal fossa of the same leg (see Figure 1: page 1-31 and Figure 2: page 1-42). The direct M-wave generally has a higher activation threshold than the H-reflex because of the relatively thinner size of the motor axons compared to the muscle spindle Ia afferents. The M-wave occurs at a shorter latency (approximately 5-8 ms in the soleus) to the H-reflex (approximately ms in the soleus) (Kudina and Alexeeva, 1992; Ali and Sabbahi, 2001). This is because the M-wave volley only has to travel along the motor axon while the H-reflex volley begins in the afferent fibre, synapses onto the motoneuron in the spinal cord, then travels along the efferent nerve fibre before eliciting a response in the target muscle (see Figure 1: page1-31). 1-26

28 H-reflex and M-wave in the Triceps Surae M-wave For many years, it was thought that the maximal M-wave, which is obtained by increasing stimulation intensity until no further increases in the M-wave strength is observed, is a steady and accurate measurement of total muscle activation (Brunia, 1973; Hwang, 2002). It is therefore often used as a normalisation factor in reflex studies, with the strength of the H-reflex being displayed as a fraction of the maximal M-wave strength (H/M-max: reviewed in Pierrot-Deseilligny and Mazevet, 2000; Zehr, 2002). The stability of M-wave strength during muscle contraction or limb movement is also used as an indicator of nerve to stimulating electrode and muscle to recording electrode relationships across an experimental sessions (Simonsen and Dyhre- Poulsen, 1999; reviewed in Zehr, 2002). It has therefore been thought that by maintaining the strength of a sub-maximal M-wave throughout a study, the level of electrical stimulation to the mixed nerve and the contribution of the muscle s electrical activity to the recoding electrode remains constant, and that therefore any changes in the H-reflex size is due to changes in other test conditions. Recently however, it has been shown that the size of the maximal M-wave may change considerably throughout an experiment at different limb positions (Simonsen, et al., 1995; Simonsen and Dyhre-Poulsen, 1999), or during different levels of muscle contraction (Nagata and Christianson, 1995). The variability has been attributed to changes in the angle of the underlying muscle fibres (Gerilovsky et al., 1989), or changes in the effectiveness of the electrical stimulation (Simonsen and Dyhre- Poulsen, 1999; Fumoto et al., 2002). The strength of the maximal M-wave can also significantly reduce over the period of an experiment while the subject is at rest (Crone et al., 1999), or from morning to evening experimental session (Castaingts et al., 2004) which cannot be explained by changes in the angle of underlying muscle 1-27

29 H-reflex and M-wave in the Triceps Surae fibres. Other factors including the temperature of the test muscles, the accumulation of interstitial fluid in the test limb (Castaingts et al., 2004), and depletion of the NA + /K + pump efficiency, which can occur with repeated activation of the motor units (Nielsen and Clausen, 2000) may all affect the total recorded muscle output. When comparing across different conditions, it can be useful to obtain the full H/Mwave curve of each of the conditions for comparison (see Figure 5: Normalised H/M max curves. Then the effect of the different conditions may be described at the stimulus intensity at which the same strength M-wave is obtained (Zehr, 2002). Although this may be considered a useful way to determine that the same stimulus strength is being applied to the motor nerves (contained within the mixed nerve), the relative strength of the H-reflex and M-wave can be altered by slight changes in the stimulating electrode to mixed nerve, or recording electrode to muscle relationship (Brunia, 1973). Furthermore, other complications exist when using M-wave strength as an indicator of stimulus strength, particularly when large M-waves are being used, given the potential of the F-wave to influence the recorded H-reflex strength (Trontelj, 1973) (see Figure 1; page 1-31). Further investigation is required to determine the best methods to control for potential changes in M-wave activity, and is discussed comprehensively in chapters 3 and 4 of this thesis Antidromic action potentials and Renshaw cells An electrical stimulus of a mixed nerve evokes action potentials that propagate both orthodromically and antidromically in non-refractory motor and sensory axons. The orthodromic volley from the motor nerves produces the M-wave, while the orthodromic volley from the Ia afferents, travels to the spinal cord, across the synapse to the motor nerves, and produces the H-reflex (Figure 1: page 1-31). The 1-28

30 H-reflex and M-wave in the Triceps Surae electrical stimulus simultaneously elicits an antidromic action potential volley in both fibre types, which propagates towards the motoneuron (soma) in the motor nerve, and towards the test muscle in the Ia afferents. While these propagating action potential volleys have no significant effect in the Ia afferents, the volley in the motor nerve can either collide with the orthodromic sensory (Ia) action potential in the axon, or cause hyperpolarisation of the motoneuron soma (Gottlieb and Agarwal, 1976). When collision occurs between the Ia afferent orthodromic volley, and the motor nerve antidromic volley, a significant reduction in the H-reflex strength will result (Brunia, 1973). This antidromic volley can also collide with voluntary produced action potentials and therefore reduce the output of the test muscle during the voluntary contraction (Brock, 1952). The antidromic volley may also propagate along Renshaw cells (see 1.1.8) evoking inhibitory postsynaptic potentials in other motoneurons. Renshaw (or recurrent) inhibition occurs as a result of larger motor neurons inhibiting the recruitment of smaller motor units, which is an effective process that ensures only those units required for the activity are active (Broman et al., 1985; reviewed in Windhorst, 1996). In the soleus, Renshaw inhibition becomes particularly important during voluntary contraction, and during dorsi flexion, reducing the stretch reflex, and allowing the normal walking cycle (Katz and Pierrot Deseilligny, 1984; Löscher et al., 1996) The F-wave The F-wave occurs at a similar latency to the H-reflex in the test muscle. It is a secondary response to the antidromic action potential volley that can occur in the activated motor nerves following an electrical stimulus (Trontelj, 1973; Dengler et al., 1992), normally at an intensity which has eliminated the H-reflex due to high 1-29

31 H-reflex and M-wave in the Triceps Surae antidromic motor nerve activity. Although the F-wave occurs at a similar latency, it is also distinguishable from the H-reflex due to its much larger stimulus threshold, and its insusceptibility to change in strength during voluntary contraction (Fujii, 1977) The H-reflex The monosynaptic Hoffmann (H) reflex was first described in the triceps surae by Paul Hoffmann in the early 20 th century. Approximately 60 years later, in 1973, Hugon published what has been regarded for many years as a definitive guideline to the parameters of eliciting the H-reflex (Hugon, 1973). In his article, Hugon describes the correct arrangement for eliciting the H-reflex in the calf and the expected responses seen in the triceps surae, as well as answering numerous other methodological questions. Since 1973, many other H-reflex review articles have been published. Some of these reviewers suggest changes to Hugon s protocols, while others describe the H-reflex parameters in a very similar light. Although the procedures are well documented, it is true that poor conclusions have been drawn from investigations following incorrect placement of stimulating and recording electrodes, inappropriate stimulus parameters, uncontrolled postural changes during an investigation, and by variation in geometry of the test muscles. 1-30

32 H-reflex and M-wave in the Triceps Surae Figure 1: Electrical stimulation of the mixed nerve Electrical stimulation of a mixed nerve is used to elicit a reflex response in the test muscle. Many Ia afferents and motor nerves lie within a mixed nerve. In the case of the tibial nerve, Ia afferents and motor nerves from the soleus and medial and lateral gastrocnemius all lie within the mixed nerve. During experimental conditions, when the stimulus intensity is sufficient, the electrical current elicits action potentials in the individual nerve fibres lying within the mixed nerve. The stimulus intensity required to elicit an action potential within the individual nerve fibres is dependant on the size of the fibres. It is common for the diameter of the largest Ia afferents to be bigger than the largest motor nerves. Because of this, the stimulus intensity required to elicit a small Hoffmann (H) reflex, which is a result of an electrical response elicited in the Ia afferents, occurs at a lower stimulus intensity than that required to elicit a direct motor response (M-wave). Furthermore, the latency of the H-reflex is greater than that of the direct M-wave given the distance that the action potentials must travel before reaching the target muscle. 1-31

33 H-reflex and M-wave in the Triceps Surae The H-reflex has been described as a monosynaptic reflex, and as an excellent tool in determining strength and distribution of spindle input to a motoneuronal pool (Hugon, 1973; Táboríková, 1973). It cannot however be described as a true direct measure of α-motoneuron excitability due to the effects of both pre-synaptic and postsynaptic effects on motor neuron excitability and therefore the strength of the reflex (for reviews see Capaday, 1997; Zehr, 2002). Furthermore, given the duration of the H-reflex, it is likely that the later proportion of the H-reflex waveform will have oligosynaptic contributions. In particular, there may be some inhibitory Ib effects from Golgi tendon organs (Pierrot-Deseilligny et al., 1981; Burke et al., 1984). It has also been demonstrated in human subjects that it is possible the electrical stimulus elicits a response from other proximal muscle and cutaneous afferents given that their activation thresholds are similar (Burke et al., 1983) Modulation of the H-reflex The strength of the H-reflex can be modulated by many central and peripheral factors. The first method of modulation is by primary afferent depolarisation (PAD) which reduces the amount of neurotransmitter released from the Ia afferent presynaptic terminal. PAD acts as an inhibitor to the circuit by reducing the potential for an action potential to be elicited in the target motoneurons (eg by skin stimulation: Sabbahi and De Luca, 1981). The majority of the changes in H-reflex strength are attributed to changes in the amount of presynaptic inhibition including but not solely PAD (for review see Schieppati, 1987; Zehr, 2002). The second way to modulate the H-reflex is via postsynaptic effects on the motoneuron pool. Postsynaptic modulation occurs by changing the potential for an incoming Ia volley to elicit an action potential in the motoneurons. The postsynaptic effects can be inhibitory or excitatory and act directly on the motoneuron. 1-32

34 H-reflex and M-wave in the Triceps Surae The strength of the maximal H-reflex is therefore dependant on both pre-synaptic and post-synaptic modulation of the monosynaptic loop and the amount of antidromic activity in the motor nerves. In the soleus, the maximal H-reflex obtained at rest is approximately 50%-60% of the maximal M-wave (Táborìková and Sax, 1968; Hugon, 1973). While in the gastrocnemius, some authors report that the H-reflex cannot be obtained before the M-wave (Hugon, 1973) Voluntary contraction It has been well established that increasing voluntary drive to the motoneuronal pool increases the possibility of an H-reflex to be initiated without the M-wave (Burke et al., 1989). Voluntary input both reduces the stimulus intensity required to initiate an H-reflex (Burke et al., 1989) and increases the strength of the H-reflex at a given stimulus intensity (Svensson et al., 1999). During voluntary muscle contraction the H/M max can increase to approximately 90% in the soleus muscle (Fumoto et al., 2002). This increase in H/M max size is due to voluntary drive bringing more motoneurons closer to the firing threshold. Once a motoneuron is close to firing threshold, previously sub threshold stimuli are more likely to recruit it and therefore make it fire (Burke et al., 1989). Furthermore, as weaker stimulus intensities are required to elicit an H-reflex during contraction, and these weaker stimuli are below the M-wave threshold, no antidromic action potential is produced in the motor nerves and therefore no antidromic collision can occur. Given the reduction in possible antidromic activity, it is possible to estimate more accurately the extent of the synaptic connections of spindle afferents to a motoneuron pool during high levels of voluntary contraction with low stimulus intensities. 1-33

35 H-reflex and M-wave in the Triceps Surae In the triceps surae, the H-reflex amplitude has been shown to increase linearly with background muscle activity in most subjects, but in others, to only increase up to approximately 50% maximal voluntary contraction (MVC) (Löscher et al., 1996). The size of the triceps surae H-reflex is also known to increase at the beginning of rapid plantar flexion, when the level of back ground activity is increasing, then tends to decrease towards the end of the motion (Simonsen and Dyhre-Poulsen, 1999; Fumoto et al., 2002). The amplitude of the M-wave and H-reflex are also known to change as a result of muscle length and architecture in leg muscles (Garland et al., 1994), as for example the strength of the soleus H-reflex decreases during dorsiflexion (Mark et al., 1968; Robinson et al., 1982) After Hyperpolarisation and absolute refractory periods The effect of after hyperpolarisation and refractory periods in both the afferent and efferent nerve fibres must be considered in reflex studies, particularly when the motor units from the test muscle are discharging at different rates. During the absolute refractory period, which occurs directly following the action potential, it is impossible to elicit a response in the affected nerve. In human nerves, the average absolute refractory period lasts for approximately 4-5 ms (Capaday, 1997). Given that the average firing frequency of units in the soleus is between 7-15 Hz (Capaday, 1997; Kuchinad et al., 2004), when the soleus is firing at its maximum frequency, a potential exists for an electrical stimulus to be unsuccessful in eliciting a response in the motor nerve for between ms every second (Capaday, 1997). This equation becomes more complicated however, because of the after hyperpolarisation period trajectory. Following an action potential in a nerve fibre, the membrane conductance to potassium is increased resulting in a decline in excitability of the neuron. This after hyperpolarisation period gradually recovers over a period 1-34

36 H-reflex and M-wave in the Triceps Surae that is dependant on the size of the motor neuron. In the soleus, the after hyperpolarisation period can last for ms (Kudina and Alexeeva, 1992). Many studies have now shown that the after hyperpolarisation period trajectory includes a longer slow depolarisation period when the motoneuron is firing at a slower rate (Miles et al., 1989; Türker, 1995; Matthews, 1999; Türker and Powers, 2002). This longer slow depolarisation period, allows greater chance for the same stimulus intensity to achieve a reflex in slower firing motor units, than in faster firing units, and may therefore result in larger reflex responses being obtained in muscles with more slow twitch fibres (Miles et al., 1989; Türker and Powers, 2002; also discussed in Todd et al., 2003) Postural variation It is well understood that the maximal H/M ratio achieved within a subject will change significantly between different postural tasks. For example, the maximal H/M ratio is significantly different between lying prone, standing (Abbruzzese et al., 1996) and at different supported body angles (Knikou and Rymer, 2003). The H-reflex is also modified at different hip joint angles (Knikou and Rymer, 2002) and during different phases of the walking and running cycle (Crenna and Frigo, 1987; Simonsen et al., 1995; Simonsen and Dyhre-Poulsen, 1999; for review see Schieppati, 1987; Zehr, 2002), and at different levels of stretch in the soleus (Gerilovsky et al., 1977; Gottlieb and Agarwal, 1978; Burke et al., 1983; Gerilovsky et al., 1989). Furthermore, the level of suppression or facilitation of the H-reflex has not always been shown to be relative to the level of background muscle activity, particularly during the gait cycle where changes may be related to the influence of a central pattern generator (Duysens et al., 1992). Other factors such as the influence of cutaneous and muscular mechanoreceptors must be considered during an experimental set up, as they are known to have either an inhibitory effect, or in some cases a facilitatory 1-35

37 H-reflex and M-wave in the Triceps Surae effect on the strength of the H-reflex (Robinson et al., 1982; Burke et al., 1983; Goldberg et al., 1992; Knikou and Rymer, 2002) Cortical influences The strength of the H-reflex has also been shown to vary with changes in alertness or awareness of the subject as it is known to decrease during light sleep, and be abolished during deep sleep (reviewed in Schieppati, 1987). It is possible that the level of the neurotransmitter serotonin is responsible for these changes as it effects both the sleep cycle and the level of alertness, and is known to change the excitability of the α-motoneuron pool (Heckman, 1994). It is therefore important that the level of alertness of the subjects described in the following trials is maintained throughout the investigations Training Specific adaptations in the motor control mechanisms behind the H-reflex and M- wave may occur as a result of training (running) or conditioned learning (pull or push tasks) (Wolpaw and Carp, 1990), with changes in the H/M ratio having been shown over both long term (4-8 week) and short term (1 day) training tasks (reviewed in Zehr, 2002) Age Aging has also been shown to affect the maximal H/M ratio, with a reduction in the maximal H/M ratio occurring in older subjects. This reduction in the strength of the maximal H-reflex has been attributed to an increase in presynaptic inhibition (reviewed in Zehr, 2002) or other degenerative phenomena (Scaglioni et al, 2002) with age. 1-36

38 H-reflex and M-wave in the Triceps Surae Other methodological considerations When looking for changes in muscle activity during an investigation, there are many other methodological issues that must be considered. For example it is known that highly significant inter-subject differences exist in almost all reflexes, and that therefore to attempt to overcome this variability, it is imperative that test and control groups are age and sex matched (Visser et al., 1983; Stetson et al., 1992; Falco et al., 1994). Recent consumption of drugs, such as caffeine, can influence experimental results (Kalmar and Cafarelli, 1999), and therefore should be avoided before any investigations. The influence of pain on the H-reflex is also important and is described in full in section and chapter 5 of this thesis The Triceps Surae The majority of work completed for this thesis utilised the electrically induced H-reflex and direct M-wave (also known as the M-response or motor wave) elicited in the human soleus and gastrocnemius muscles. This section reviews the fundamental differences between the structure and function of the human soleus and gastrocnemii muscles, including recent data published about the complex innervation zones within these muscles Structure of the muscles of the triceps surae The triceps surae consists of the medial and lateral heads of the gastrocnemius, and the soleus muscles. The gastrocnemius is the most superficial muscle of the calf; its two heads originate on the medial and lateral posterior condyle of the femur respectfully, while the soleus is a broad flat muscle that lies beneath the gastrocnemius with origins on both the upper posterior surface of the tibia and fibula. Both muscles insertion point is on the calcaneus via the Achilles tendon. 1-37

39 H-reflex and M-wave in the Triceps Surae The distribution of motor unit types is quite different between the muscles. There is an approximately equal distribution of slow (type I) and fast (type II) twitch fibres in the gastrocnemius that allows for a large force output range. While the soleus has about 70-90% slow and 10-30% fast twitch fibres in humans which restricts the muscle to more low force output roles (Burke, 1981; Johnson et al., 1973; Edgerton et al., 1975; Green et al., 1981). The number and density of muscle spindles is also far greater in the soleus (408 spindles; 0.94 muscle spindles / gram) than in the gastrocnemius (156 spindles; 0.4 muscle spindles / gram) (Voss, 1971), and the volumes of the soleus, lateral gastrocnemius and medial gastrocnemius in human adults are approximately 450, 145 and 260 cm 3 respectfully (Fukunaga et al., 1992; Elliott et al., 1997) Function of the muscles of the triceps surae The structure and constitution of the gastrocnemii and soleus support their different functional roles during standing and during movement of the lower limb. The gastrocnemii and soleus muscles have a postural function, in that their activity is required to maintain an upright stance, and they are used under load in normal conditions (Nardone and Schieppati, 1988b). Given the attachment to the Achilles tendon, all three muscles are plantar flexors of the foot, while the gastrocnemii are also a flexor of the knee. Other muscles involved with plantar flexion include the plantaris, tibialis posterior, flexor hallucis posterior, and flexor digitorum longus. While the gracilis, sartorius and popliteus muscles combine with the gastrocnemius to allow knee flexion. The close proximity of these muscles and their tendency to co-active during certain movements can result in cross talk particularly when recording by surface electromyography 1-38

40 H-reflex and M-wave in the Triceps Surae (Isaacs et al., 1968; Türker and Miles, 1990). The experimental considerations for such factors are discussed in section During symmetrical standing, the soleus consistently exhibits activity, while the gastrocnemii muscles are active intermediately in most cases (Joseph et al., 1955). The soleus also has higher levels of background muscle activity than the gastrocnemii during normal standing (Joseph et al., 1955; Monster et al., 1978). The soleus is most active when the foot is in dorsiflexion (lengthening contraction), while the gastrocnemii are most active when the ankle is plantar flexed, during strong contraction or rapid development of tension (Hollinshead, 1951; Herman, 1967). Animal studies have shown that activities such as walking and running that involve mechanical coupling of the ankle and knee joints can result in suppression of the soleus, and simultaneous facilitation of the gastrocnemii muscles (Hutchison et al., 1989; Nichols, 1999). This is supported by the suggestion that there is no heteronymous facilitation, but instead only inhibitory input from medial gastrocnemius Ia afferents to soleus α-motoneurons in humans (Mao et al., 1984). Mao and colleagues further suggest that although no heteronymous facilitatory projections exist from the human medial gastrocnemius to soleus motoneurons, however that the reverse is likely. They also showed heteronymous projections between the two heads of the gastrocnemius (Mao et al., 1984). This story is complicated further by the introduction of the Ib pathway, which is known to be important in the co-ordination of multi-muscle movements (Rossi and Decchi, 1995). Both inhibitory and excitatory pathways from gastrocnemius tendon organs (Ib) to soleus motoneurons have now been shown in both the cat (reviewed in Rossignol et al 2006) and in humans (Faist et al., 2006). Although not thoroughly examined to date in humans, it is currently understood that while the facilitatory 1-39

41 H-reflex and M-wave in the Triceps Surae (gastrocnemius Ib - to soleus α-motoneouron) pathway requires some form of locomotion, the inhibitory pathway, which has previously been studied in more detail, requires only load bearing (Faist et al., 2006). During walking, recordable muscle activity of the soleus starts earlier in the stance phase than in the gastrocnemius (Duysens et al., 1991), and similarly during this phase, the increase in reflex response is larger in soleus than in gastrocnemii (Duysens et al., 1991). The soleus also de-recruits differently to the gastrocnemius during both plantar flexion and dorsi flexion in humans (Nardone and Schieppati, 1988a). Combined, these findings indicate that the motor neuron pools of each muscle function quite separately Motor points The motor point of a muscle corresponds with the entrance point of the motor nerves into the muscle. Muscle fibres of a given unit, and even across units, tend to receive their innervation in this spatially centred region of their muscles belly. From this region, action potentials propagate symmetrically in either direction, to the end of each muscle fibre within a given motor unit. It is commonly thought that the innervation zone of most skeletal muscles, lie perpendicular to the direction of the fibres, in the mid belly of the muscle. Although this is seen in the soleus, recent publications have shown that the gastrocnemius innervation zones are more scattered and distributed in quite complex configurations (Rainoldi et al., 2004; Parratte et al., 2002) (see Figure 2: page 1-42). 1-40

42 H-reflex and M-wave in the Triceps Surae 1.4. Experimental Considerations in the Triceps Surae Much work has now been completed on the Hoffmann reflex (H-reflex) and direct motor response (M-wave) in the triceps surae, with greatest focus on the soleus because of the larger separation between the stimulus thresholds of its H-reflex and M-wave (Capaday, 1997). There are however limitations to extrapolating complete and comparative results from many of these investigations, due to the large variability that exists between the experimental protocols of different authors. 1-41

43 H-reflex and M-wave in the Triceps Surae Figure 2: Position of the popliteal fossa and innervation zones of the triceps surae (see next page for explanation) 1-42

44 H-reflex and M-wave in the Triceps Surae Figure 2: Position of the popliteal fossa and innervation zones of the triceps surae The popliteal fossa (popliteal cavity) is a diamond shaped region situated behind the knee. It is bound superiorly and laterally by the biceps femoris, inferiorly and laterally by the plantaris and lateral head of the gastrocnemius, superiorly and medially by the semitendinous and semimembranosus, and inferiorly and medially by the medial head of the gastrocnemius. The most densly innervation zones illustrated by (Rainoldi et al., 2004) are shown in the dense rectangular area: Medial gastrocnemius: Approximately half way between the medial side of the popliteus cavity and medial gastrocnemius-achillies tendon insertion point. Lateral gastrocnemius: Approximately 60% of the distance between the lateral gastrocnemius-achillies tendon insertion point and the lateral side of the popliteus cavity. Soleus: Approxiamtely 75% of the distance between the soleus-achillies tendon insertion point and the tibial tuberosity. The innervation areas for each muscle indicated in the ovals were demonstrated by (Parratte et al., 2002; Saitou et al., 2000). 1-43

45 H-reflex and M-wave in the Triceps Surae Common protocol One of the most fundamental and overlooked factors that will alter the strength of both background muscle and reflex activity is the influence of the vestibular system (Knikou and Rymer, 2002: for review see Schieppati, 1987). To overcome these effects the most commonly used position for subjects in H-reflex studies is the semi reclined position with the head and arms supported (Hugon 1973; Brunia et al., 1973). It is suggested that the foot should be supported in a slightly flexed position, with the knee angle of approximately This allows adequate stretch of the soleus and Achilles tendon, while the gastrocnemius is relaxed due to the degree of knee flexion. The relaxation reduces the potential of any depressive influence on the soleus that may originate from the gastrocnemius (Hugon, 1973). Subjects may also be supported in the prone or supine laying position, providing the leg and body are adequately supported, and the subjects comfort level is maintained throughout the experiment (Hugon, 1973). Although this set up is commonly used it is also known that the test-retest reliability of the H-reflex is greater when the subject is in a loaded position (Ali and Sabbahi, 2001), rather than relaxed as suggested above. Furthermore, although these are the ideal postures for eliciting reliable H-reflex strengths in the soleus, it is not always suitable for many of today s investigations, particularly when movement and postural changes are being examined Placement of the stimulating electrodes Variability also exists in the preferred placement of stimulating electrodes between some investigators. Bipolar stimulating electrodes are generally placed along the skin above the tibial nerve in the popliteal fossa (Burke et al., 1989), while some prefer to 1-44

46 H-reflex and M-wave in the Triceps Surae ensure that the stimulus passes through the tibial nerve, and therefore place the bipolar electrodes on either side on the leg (Burke et al., 1984). The electrical current used to elicit an H-reflex is however most commonly delivered via an AgCl cathode (2 cm 2 ), which is placed in the popliteal fossa, and stainless steel or brass anode (20cm 2 ) placed on or just above the patella. This arrangement reduces the artefact of the electrical stimulus, and allows more selective stimulation of the tibial nerve when compared to the longitudinal stimulating electrode arrangement (Hugon, 1973; Capaday, 1997). Movement of the stimulating electrode to nerve relationship, that may occur for example during limb movement, can alter the relative activation of the underlying Ia afferents and therefore lead to changes in the recorded H-reflex strength (Zehr, 2002). Therefore during experiments where the test limb will be moving (e.g. walking and running trials) it is suggested that the cathode be placed just above or below the popliteal fossa, to reduce the potential for changes in the stimulating electrode to nerve relationship (Capaday, 1997). The placement of the cathode is considered adequate when the stimulus threshold for eliciting the H-reflex is less than that required to induce an M-wave in the soleus muscle (Hugon, 1973). It is well understood that with increased stimulus intensity, an M-wave will be elicited in the soleus and then in the gastrocnemius (Hugon, 1973) Stimulation duration, frequency and variability The optimal stimulus duration for eliciting the H-reflex in the triceps surae has been debated for a number of years. Hugon suggested that it should be 1 ms, with shorter durations favouring the activation of α-motor axons (Hugon, 1973). While Capaday suggest that 0.5 ms duration is a good compromise for eliciting the response in the appropriate nerve fibres, and reducing unpleasant sensations. Capaday continues by 1-45

47 H-reflex and M-wave in the Triceps Surae suggesting that although shorter stimulus durations may be used, and would favour a response being elicited in the Ia afferents before the motor nerves, greater stimulus intensities would be required, and therefore a greater risk of stimulating nociceptors in the skin exist (Capaday, 1997). Originally it was suggested that the period between stimuli should be no less than 5 seconds (Hugon, 1973). However more recent studies indicate that stimuli given 3 seconds apart will avoid any post activation depression when the test muscle is at rest, and that stimuli may be given up to 4 Hz without loss of reflex amplitude if the test muscle is voluntarily contracting (Burke et al., 1989; Rossi-Durand et al., 1999; for review see Capaday, 1997; Zehr, 2002). To overcome trial-to-trial variability of the H-reflex, it has been suggested that the reflex response to 10 stimuli are averaged for each experimental condition (Hugon, 1973), however it is common for up to 20 (Knikou and Rymer, 2002) or as few as 3 stimuli to be repeated in more recent investigations (Zehr and Stein, 1999) Use of constant current or voltage stimulator It is commonly agreed that the use of a constant current stimulator will promote the use of a stable stimulus intensity throughout the study, as it will reduce the effect that changes in the skin-electrode impedance may have on the size of the responses obtained (Hugon, 1973; Zehr, 2002). However, since the objective of many of these studies is to maintain the M-wave strength, or to normalize the size of the reflex response to the maximal M-wave strength with little regard to the actual stimulus intensity required, the source of the current may be irrelevant (Capaday, 1997). 1-46

48 H-reflex and M-wave in the Triceps Surae Recording with monopolar v bipolar surface electrodes Most authors suggest similar electrode placement for recording surface muscle activity from the soleus. Hugon suggested that recording electrodes should be 1 cm 2, and placed 3 cm apart on the skin above the soleus muscle, along the mid-dorsal line of the leg, about 4 cm below the point where the two heads of the gastrocnemius join the Achilles tendon (Hugon, 1973). Similarly, Zehr suggested that they should be placed one half of the distance between the mid-popliteal crease and the medial malleolus, or just distal to the belly of the medial gastrocnemius, medial to the Achilles tendon (Zehr, 2002). Gerilovsky and colleagues suggest that the optimum position for recording electrodes depend on the type being used. They suggest that bipolar electrodes should be placed 2 and 4 cm below the insertion of the gastrocnemius, while the active monopolar electrode should be placed a further 3 cm distal (Gerilovsky et al., 1989). It has long been understood that the relationship between the muscle bulk and the skin (on which the surface electrodes are attached) can change during motion, and that monopolar SEMG recordings are more effective at measuring reflex activity during contraction as they reduce the effect of changes in the geometry of muscles has on the recorded output (Gerilovsky et al., 1989). Muscle architecture is also known to influence the shape of the potentials recorded by surface electrodes, and therefore placement of the electrodes becomes important when the underlying muscle is complex (Garland et al., 1994). While monopolar recordings of leg muscles have been preferred by some due to better reproducibility of the shape, and easier distinction between the M-wave and H-reflex (Garland et al., 1994), it is generally considered inferior to bipolar recording because of the tendency for increased cross talk and noise during monopolar recording (Türker, 1993). 1-47

49 H-reflex and M-wave in the Triceps Surae Cross-talk results when potentials from adjacent muscles reach the recording site through volume conduction, thus contributing to the EMG signal. There is strong evidence for example that the reflexes observed in the tibialis anterior could be caused by cross-talk from more active muscles, such as the soleus (De Luca and Merletti, 1988; Türker and Miles, 1990). In relation to the maximal M-wave, it is believed that recording from the soleus with 3 cm bipolar recording electrodes (rather than monopolar electrodes) eliminates cross talk from the gastrocnemius. It is not believed to be the case in reverse however, with muscle activity recorded in the gastrocnemius being influenced by the soleus activity due to its proximal location (Hugon, 1973). The benefits and pitfalls of using bipolar and monopolar surface recording electrodes are discussed further in chapter 3 of this thesis Recording with surface v intramuscular electrodes Electrical activity of a human muscle may be recorded using either surface (as discussed above in 1.4.5) or intramuscular electrodes (see 1.1.2). While the surface electrodes lay on the skin above the test muscle, intramuscular electrodes are inserted into the muscle. Surface electromyography (SEMG) recording is good for determining the effect an input stimulus has on a whole, superficial muscle, however it cannot show how the stimulus has affected individual motoneurons (Hugon, 1973; Türker, 1993). The electrical activity recorded during muscle investigations result from transient ionic potentials in the muscles motor units. The motor unit potential as recorded from surface electrodes is a result of the summation of many motor unit potentials. 1-48

50 H-reflex and M-wave in the Triceps Surae While there is no doubt that intra-muscular electromyogram (IM-EMG) recordings have some superior characteristics to surface recordings (Türker, 1993). IM electrodes can only record from proximal motor units, which is not desirable if the investigator is interested in the response of the whole muscle. IM electrodes are also more expensive and more invasive than surface electrodes, and IM-EMG recordings require faster and more expensive equipment due to higher frequency characteristics of the signal they are recording. The main benefit of IM-EMG is that it is not subject to as much cross-talk, stimulus and movement artefact as the SEMG, and the signals recorded from within the muscle are not as filtered (Türker et al., 1999) due to the closeness of the electrodes to the motor units. Reproducibility of results in different experimental sessions is known to be better with SEMG than IM-EMG (Komi and Buskirk, 1970). However, this may be due to an inability to ensure the exact position and depth of the IM-electrodes for the second recording session. Comparatively, in SEMG recordings, skin markers can be used to ensure correct placement of the surface electrodes in different experimental sessions. Furthermore, given that the SEMG electrodes record from a large area of muscle, greater chance exists for a similar response to be recorded during different experimental sessions. SEMG recording is used throughout the experiments described in this thesis given that each investigation focused on whole muscle responses. 1-49

51 H-reflex and M-wave in the Triceps Surae Noise reduction As well as cross talk interfering with the recorded signal, the impact of other influences such as the stimulus artefact, cable motion artefact and power line interference must be considered (Türker, 1993; Clancy et al., 2002). Many signal amplifiers are now fitted with stimulus artefact suppressors, which can temporarily stop recording for the period of the stimulation. Without this suppressor, the duration of the artefact can spread well into the latency of the M-wave. Today s amplifiers are also fitted with low frequency high pass filters. By eliminating the low frequency components, artefacts such as that produced by limb movement or electrode cable motion will be eliminated (Clancy et al., 2002). Further noise reduction occurs due to common mode rejection, which occurs when bipolar recording electrodes receive the same signal at the same time, and subsequently much of the common signal is removed from the recorded trace. Common mode rejection is important in reducing cross-talk signals, other electrical activity from within the body, and electrical noise external to the body (De Luca and Merletti, 1988; and reviewed in Türker, 1993; Clancy et al., 2002). The use of a low impedance lip clip (Türker et al., 1988), and 50Hz notch filters will further reduce the effect of this noise on the recorded signal (Clancy et al., 2002). Common mode rejection and cross talk are discussed in detail in chapters 3 and 4. To ensure that the signal to noise ratio remains high throughout the experimental session, the subjects skin should be prepared by removing any excess dead skin with an abrasive paper, and using silver disc surface electrodes filled with an electrode jelly (Brunia et al., 1973; Clancy et al., 2002). The skin impedance should be kept around or below 3kΩ (Basmajian, 1973). 1-50

52 H-reflex and M-wave in the Triceps Surae Analysis methods It is well established that single EMG traces are not reliable for measuring reflex size given the large variability that exists between individual traces (Lavigne et al., 1983). Because of this, many manipulations of the raw data have been developed, including averaging a number of responses around the time of stimulation, and full-wave rectification of the EMG signal. In relation to H-reflex studies, Hugon (1973) suggested that measuring the strength of the largest peak or trough (related to the positive or negative components of the response) in the EMG trace at H-reflex latency gave fairly consistent results (Hugon, 1973). Today however, greater accuracy is demanded in the analysis process. Methods such as averaging the non-rectified peak-to-peak (PTP) amplitude of a response (Hugon, 1973; Clancy et al., 2002) or area of the rectified response by using the cumulative sum (CUSUM) are quite common (Brinkworth and Türker, 2003). PTP amplitude analysis simply involves measuring the difference between the maximum and minimum peak of the response from the raw traces. While the CUSUM measurements take into account the duration of the response as well as the peak amplitudes, through more complex calculations (Ellaway, 1978; Brodin et al., 1993; Brinkworth and Türker, 2003). Both methods of analysis are shown on the following page (Figure 3: page 1-52). 1-51

53 H-reflex and M-wave in the Triceps Surae Figure 3: Average trace, average rectified trace and CUSUM of M-wave and H- reflex The top trace is an average of 3 raw surface EMG recordings, showing peak-to-peak (PTP) amplitude of both the M-response and H-reflex. The middle trace shows rectification of the top trace, and the bottom trace shows the CUSUM of the rectified data. Error box limits are shown on the CUSUM (hashed lines) and were set at 2 x the maximum pre-stimulus variation from the mean pre-stimulus value. The same error box size was also used for the H-reflex (not shown) using the mean activity level between the 2 responses as the zero line. The latency of a response is determined by the first turning point in a series of two turning points whereby the vertical distance between the 2 turning points in the CUSUM is larger than the prestimulus error box. The duration of the response is the time between both turning points. 1-52

54 H-reflex and M-wave in the Triceps Surae 1.5. Pain In neurologically intact subjects, the perception of pain is evoked by stimuli that are sufficient or potentially sufficient to produce tissue damage. The perception of pain does vary in patient populations, for example in sufferers of fibromyalgia (Montoya et al., 2005; Perez-Pareja et al., 2005), however this will not be discussed in this overview. The pain sensation is delivered to the central nervous system via free nerve endings of small-diameter myelinated afferents (A-δ) and unmyelinated (C nerve) fibers (Lund et al., 1991; Sessle, 1999). These nerves project to the spinal cord dorsal horn before traveling dorsally to the brain (Sessle and Hu, 1991). The structure of the pain system ensures that a human response to pain will involve both a sensory and motivational response, with the convergence and interaction between cutaneous and deep afferent inputs occurring in the thalamus and somatosensory cerebral cortex (Sessle and Hu, 1991); mid brain and medulla (Fields, 1984), via both complex circuits and direct somatosensory pathways (Di Piero et al., 1994). For a comprehensive review of the pain pathways see (Browm and Lorenz, 1998). There are many primary reasons why ongoing muscle pain will occur in the normal environment. These include muscle fatigue (Miles and Nordstrom, 1995), direct injury to tissue by force, and repeated eccentric contraction (Proske and Morgan, 2001), which will result in a delayed onset of muscle soreness (DOMS) in the days following the contraction (Weerakkody, 2003). Subsequent to the original injury, the secondary inflammatory response can also be related to ongoing pain (Lund et al., 1991). To enable the symptoms associated with muscle pain to be treated appropriately, the mechanisms of muscle pain, and the effect that this has on normal muscle activity 1-53

55 H-reflex and M-wave in the Triceps Surae needs to be fully understood. To date many different methods for experimentally inducing pain have been developed. These include the intramuscular injection of hypertonic saline (normally at 5-8% saline) (Graven-Nielsen et al., 1997b and 1998); where the pain sensation is likely to be due to changes in the intramuscular sodium and potassium levels (Graven-Nielsen et al., 1997b). Eccentric contraction; which is used experimentally to induce micro tears in the test muscle fibres, and which results in a subsequent inflammatory response in the test muscle (MacIntyre et al., 2001; Weerakkody, 2003). Injection of serotonin and bradykinin is also used in the generation of muscle pain and muscular hyperalgesia in humans (Babenko et al., 1999; Babenko et al., 2000), while topical sprays of capsaicin or direct application of hot or cold objects are also used experimentally to specifically activate nociceptors in the skin (McMullan, 2004). Finally, electrical stimulation of the pain afferents are also used to induce experimental pain in some investigations (Bini et al., 1981) The sensation of pain, and reorganisation in the nervous system that can occur following pain depends on the origin of the pain. For example painful stimulation of deep tissues can produce more long-lasting changes in central nervous system excitability compared to cutaneous stimulation, and therefore it may be considered to play a prominent role in chronic pain conditions (Dubner, 1991). The effects of pain on human function are wide and varied as for example high levels of muscle pain are also known to reduce the ability for people to complete very high-level cognitive tasks (Eccleston, 1994). Investigations into the effect of pain on muscle activity have provided varied results (described in full in chapter 5). One of the factors that may contribute to this variability is the lack of a good objective test for the existence of pain, given that the sensation of pain is described in subjective terms. In neuromuscular trials, the level of pain is usually recorded subjectively by a visual analogue scale (VAS) (Salo et al., 1-54

56 H-reflex and M-wave in the Triceps Surae 2003; Lipton, 1991; Kumar, 2002). The VAS system asks subjects to report their pain level from 0/10 (which represents no pain at all), to 10/10 (which represents extreme pain). Although the reproducibility of pain scores using this method is known to be excellent (Goddard et al., 2004), they are obviously dependant on many subjective characteristics of the person describing the pain Painful stimuli and the perception of pain In human studies it has been shown that there is a strong relationship between stimulus intensity, and reported pain intensity (Milne et al., 1988; Milne et al., 1991; Svensson et al., 1997b; Hansen et al., 1999). Many factors have been shown to affect the perception of pain however, including arousal, attention, and emotional stress (Bini et al., 1984; Fields, 1984; Feine et al., 1990). Similarly, either stimulating or anaesthetising sensory receptors (other than the pain receptors) have shown mixed results in the relation to their effect on the perception of pain (Mackenzie et al., 1975; Thunberg et al., 2002). Habituation to a pain response is also known to occur in some situations, where the evoked pain sensation increases, while the electrophysiological response to pain decreases (Milne et al., 1991) The effect of pain on muscle activity The effect of pain on muscle activity has been debated for a long period, and today two main pain models have been developed. These are known as the pain adaptation and vicious cycle models. The pain adaptation model suggests that changes in muscle activity that occur while an individual is in pain, is a normal protective adaptation. This adaptation is suggested to reduce the output of antagonist muscles and increase the output of agonist muscles to reduce the effective movement of the painful area (Lund et al., 1991). Because of the specificity of the 1-55

57 H-reflex and M-wave in the Triceps Surae muscles that respond to pain, it is suggested that these changes occur due to the action of pain on segmental interneurons (Lund et al., 1991). The vicious cycle model, suggests that pain and dysfunction are reciprocally linked (Travell et al., 1942 as reported by Lund et al. (1991). Both models are supported by many studies, and both suggest fundamentally that pain does alter the level of muscle activity in muscles associated with the painful site Pain and background muscle activity While clinical observation suggests that painful muscles are often tense and associated with increased EMG activity as compared to non-painful muscles (Svensson et al., 1997a; Sessle, 1999), some studies show that no EMG hyperactivity occurs during muscle pain at rest or during low levels of contraction (Graven-Nielsen et al., 1997a; Birch et al., 2000). It has also been shown that chronic lower back pain patients have significantly less para-spinal muscle activity than control groups (Collins et al., 1982) and in others, that there is no significant difference between resting EMG activity in painful and non-painful muscles at all (Matre et al., 1999; Schulte et al., 2004). In another study, the level of back ground muscle activity was shown to increase temporarily following the injection of pain, but that this hyperactivity did not last as long as the pain sensation (Svensson et al., 1998b). The variability in results may in part be explained by a difference in the function of the muscle in pain, as the tense muscles that are noticeable in people with pain, are most noticeable when the person is under load or during dynamic postural tasks compared to non-loaded conditions (Ahern et al., 1988; Rossi and Decchi, 1994). Other investigations have looked at the maximal voluntary contraction (MVC) of a muscle during pain; and have shown that it is significantly lower during the painful 1-56

58 H-reflex and M-wave in the Triceps Surae condition (Graven-Nielsen et al., 1997a). In this situation, experimental muscle pain is known to be able to reduce the force output of a muscle while not changing the contractile properties of the muscle, which suggests that the changes are centrally mediated (Graven-Nielsen et al., 2002). Pain has also been shown to reduce endurance duration in static contracting muscles (Graven-Nielsen et al., 1997a); increase EMG activity of antagonist muscles (Graven-Nielsen et al., 1997a; Schulte et al., 2004); and decrease the EMG activity in agonist muscles during dynamic tasks (Graven-Nielsen et al., 1997a) Pain and motor unit firing rate Some authors have found no change to the number, the firing characteristics, or the properties (area or amplitude) of the motor units recruited during pain (Birch et al., 2000). While others have shown significant increases in the firing frequency of motor units during pain, with a direct correlation between the level of pain recorded, and the increase in firing frequency (Farina et al., 2004). No change to the conductance velocity of the motor units has been recorded during experimental pain (Farina et al., 2004) Pain and reflex strength The effect of pain on the H-reflex has not been investigated in full, and much of the current data is quite varied. For example, it is known that when pain is applied by heat to the sole of the foot, the H-reflex in the tibialis anterior is abolished, while it remains unchanged in the soleus (Ellrich and Treede, 1998). Furthermore, it has been shown that pain in either the soleus or tibialis anterior muscles does facilitate the stretch reflex in the soleus muscle, the same pain has no significant effect on the soleus H-reflex strength (Matre et al., 1998, 1999). These results suggest that there 1-57

59 H-reflex and M-wave in the Triceps Surae are nociceptive inputs to the spinal monosynaptic pathways, but that the influence of these inputs may be very specific. It is possible therefore that increased dynamic sensitivity of the muscle spindles during muscle pain is caused by an increased firing rate in the dynamic gamma-motoneurons, while the alpha-motoneurone pool remains unchanged during pain (Matre et al., 1998). All other investigations that consider pain and the H-reflex are limited to facial (Svensson et al., 1998a), and arm (Le Pera et al., 2001) muscles, and are described in full in chapter 5 of this thesis Pain and EMG spectral analysis It is now well understood that the EMG signal undergoes compression in the frequency domain during a sustained muscle contraction, and that this compression of frequency domain is altered during fatigue of muscles (De Luca, 1993). The median frequency of an EMG power density spectrum has been shown to differ between populations with and without lower back pain in both athletes, and nonathletes (Roy et al., 1989, 1990). The power of the mandible peak tremor frequency has also been shown to decrease during local pain, both when the jaw is at rest and during voluntary jaw movements. In contrast, the power of finger peak tremor frequency only increases during voluntary contraction and not while the hand is at rest. The same study showed that there was no change in the peak tremor frequency in either the finger or the jaw during pain (Jaberzadeh et al., 2003) Pain and the recruitment order of muscles Other effects of pain include the finding that the recruitment order of muscles that are required in postural tasks, changes with pain. It has been suggested that people with lower back pain experience changes in the planning of the motor task or to delayed 1-58

60 H-reflex and M-wave in the Triceps Surae transmission of the descending motor command in the nervous system, when making these postural changes (Hodges and Richardson, 1996; Hodges, 2001). Similar changes have been noted in subjects during experimentally induced pain (Hodges et al., 2003), and recently during the anticipation of pain (Moseley et al., 2004) Conclusions and Directions This chapter has introduced the terms required to understand the neuromuscular physiology described in the following chapters. It has also outlined the most commonly used and accepted protocols for eliciting and recording the H-reflex and M-wave in the triceps surae, and reviewed in detail the variety of protocols that will be used in the investigations to follow. A detailed summary of the neuromuscular response to pain has also been included, which clearly shows that results are often conflicting, and that further research in this area is imperative. From this review, it is also clear that future work in the area of the H-reflex in the triceps surae needs to focus initially on obtaining more information about the gastrocnemius. Although much work has been completed on the soleus, given the muscles location, function and innervation by the tibial nerve, they together allow a simultaneous and direct comparison of two functionally different muscles. The following chapters focus on the four main issues that require further research as determined in the review, these are Chapter 2: The lack of comparative information about the H-reflex in the gastrocnemius and the soleus Chapter 3: Fundamental errors that exist with current recording techniques in surface recorded reflex studies. 1-59

61 H-reflex and M-wave in the Triceps Surae Chapter 4: Variability in the maximal M-wave during different experimental conditions. Chapter 5: The effect of pain on agonist and antagonist background muscle activity and the H-reflex in the triceps surae. Chapter 6 then summarises the new information, and suggests possible future directions for research in this area of study. To date 4 of the 5 papers produced as part of this thesis have been published, with the 5 paper currently in review. All published papers are included in the appendix at the end of this thesis. 1-60

62 Signal cancellation in maximal M-wave 2. Muscle spindle feedback differs between the soleus and gastrocnemius in humans This chapter is a slightly edited version (incorporating reference to other sections of the thesis), of the manuscript Muscle spindle feedback differs between the soleus and gastrocnemius in humans by K.J Tucker and K.S. Türker, which has been published in Somatosensory and Motor Research (2004) 21(3/4): Outline of study The Hoffmann (H) reflex and motor (M) response were studied in soleus and gastrocnemius during voluntary contraction in 8 male volunteers. The aims of the study were two fold: To determine if the strength of the muscle spindle input to the soleus and gastrocnemius is the same; and to assess if the strength of the motor evoked M-wave changes during contraction. The results of this study showed that the strength of the maximal M-wave does change during voluntary contraction within subjects. Hence, all H-reflex measurements were normalised to the maximal M-wave at each level of contraction for each subject. The largest H/M max was bigger in soleus than gastrocnemius at every contraction level. The overall largest H/M max for soleus (97%) and gastrocnemius (55%) were achieved at 40% and 100% maximum voluntary contraction (MVC) respectively. Conclusion: Soleus receives greater spindle feedback than the gastrocnemius both at rest and during voluntary contraction. 2-61

63 Signal cancellation in maximal M-wave 2.2. Introduction Although the two muscles are often investigated as a single entity, it is well established that the gastrocnemius and soleus muscles are quite different in both function and composition. The soleus lies immediately deep to the gastrocnemius, they share the same tendon attaching to the ankle and are both involved with plantar flexion; however the soleus does not attach across the knee joint like the gastrocnemius and therefore does not share the same functional role of knee flexion (Hollinshead, 1951). Activities such as walking and running involve mechanical coupling of the ankle and knee joints and can result in simultaneous suppression of the soleus and facilitation of the gastrocnemius (Nichols, 1999). Conversely, the soleus muscle is active constantly and has higher levels of background muscle activity than the gastrocnemius during normal standing (Joseph et al., 1955). The soleus is also most active when the foot is in dorsiflexion (lengthening contraction) and is involved with stabilising the ankle joint, while the gastrocnemius is most active when the ankle is plantarflexed, during strong contraction or rapid development of tension (Herman, 1967). Human histochemical studies have shown that there is an approximately equal distribution of slow (type I) and fast (type II) twitch fibres in the gastrocnemius, while the soleus has about 70-90% slow and 10-30% fast twitch fibres (Johnson et al., 1973; Edgerton et al., 1975; Green et al., 1981). In humans the number and density of muscle spindles is far greater in the soleus (408 spindles; 0.94 muscle spindles / gram) than in the gastrocnemius (156 spindles; 0.4 muscle spindles / gram) (Voss, 1971). The spindle system provides approximately one-third of the total peripheral input to the motoneurons of postural muscles, and is especially important in maintaining tonus in these muscles throughout the day (Macefield et al., 1993). The 2-62

64 Signal cancellation in maximal M-wave characteristics of each muscle therefore support their functional roles, with the soleus having a greater role in postural maintenance (Nardone and Schieppati, 1988b) and the gastrocnemius in ankle movement and flexion of the knee (Hollinshead, 1951). The higher number of soleus muscle spindles also means that a stimulus that excites the tibial nerve (which encapsulates the muscle spindle afferents from both muscles) may result in a greater level of soleus motoneuron recruitment. Furthermore it has been established in animals that the spindle input delivers twice as much current into small motoneurons than larger motoneurons, which has been thought to strengthen the orderly recruitment of motoneurons (Heckman and Binder, 1990; Burke, 1991) and which again favours the soleus given its greater percentage of slow motoneurons. It is well established however, that when the tibial nerve is stimulated in the popliteal fossa, the gastrocnemius and soleus muscles may receive both homonymous and to a lesser degree heteronymous monosynaptic projections from the spindle afferents (Mao et al., 1984). While there is no medial gastrocnemius spindle projection to the soleus in humans the reverse does occur (Mao et al., 1984). The difference in heteronymous facilitation supports the functional roles of the two muscles during the walking cycle (when the gastrocnemius is stretched during knee extension, the soleus must allow the foot to remain in dorsiflexion (Mao et al., 1984); while if the soleus is stretched during postural maintenance, gastrocnemius activity will facilitate stability (Sakamoto and Swie, 2003). Therefore if heteronymous projections are important contributors to the H-reflex, it may be expected that gastrocnemius motoneurons will be more easily recruited given that they will receive both heteronymous and homonymous spindle input from the mixed tibial nerve. The size of the H-reflex that can be elicited at rest varies depending on the muscle being studied. In the soleus, the largest H-reflex obtained at rest is approximately 2-63

65 Signal cancellation in maximal M-wave 50%-60% of the maximum motor (M) response (M max) (Táborìková and Sax, 1968) which suggests that approximately 50-60% of the motoneuronal pool can be recruited by spindle input alone. At higher stimulus intensities, the size of the H-reflex is reduced by the antidromic wave that is co-elicited in the motor nerves along with the M-response, and therefore, the exact value of the spindle input to the motoneuronal pool cannot be determined at rest (Kudina and Pantseva, 1988). The size of the maximum H/M max can increase from less than 10% at rest to 80% in voluntary contracting arm muscles (Verrier, 1985) and from 50-60% at rest to almost 100% in the voluntary contracting soleus muscle (Fumoto et al., 2002). This is because the voluntary drive brings more motoneurons closer to the firing threshold, which allows them to be more easily recruited by the previously subthreshold stimuli (Burke et al., 1989). As the lower intensity stimuli are below the M-response threshold, no antidromic action potential is produced in the motor nerve and no collision can occur. Hence, it is possible to more accurately estimate the extent of the synaptic connections of spindle afferents to a motoneuron pool during high levels of voluntary contraction with low stimulus intensities. Many previous studies examining the H-reflex compare results from different subjects and experimental conditions by normalising results to the maximum evoked M- response. Recent studies however, have shown that the maximum M-response may reduce over the period of an experiment (Crone et al., 1999), be altered by the position of the limb being tested (Fumoto et al., 2002), and be modified by the level of voluntary contraction (Nagata and Christianson, 1995). The current study was designed to accurately record the H-reflex and M-response curves of lower leg muscles throughout the full stimulus range, and at different levels of muscle contraction. The primary aim of the study was to determine if the spindle 2-64

66 Signal cancellation in maximal M-wave input to the soleus and gastrocnemius muscles are different. The second aim was to determine if the M-response changes between these different experimental conditions Materials and Methods Subjects Twelve neurologically intact male subjects between the ages of (mean 23) participated in this study. Four subjects were excluded from the study, as the H-reflex could not be obtained before the M-response in the relaxed gastrocnemius muscle. Hence eight subjects were used where the H-reflex was evoked using stimulus intensities below M threshold. Each subject gave written informed consent to participate in this study. The experiments were approved by the Human Ethics Committee of the University of Adelaide and all procedures used conformed to the Declaration of Helsinki Protocol The subjects lay prone, on a standard physiotherapy table with a face hole that allowed visual feedback of background muscle activity from an oscilloscope positioned under the table. A footrest supported their right foot, which was positioned with an ankle angle of 90 o, and provided the resistance required to produce large levels of voluntary contraction. Given that the subjects were laying prone with a foot support, when the leg was at rest both the gastrocnemius and soleus were in a comfortably stretched position. 2-65

67 Signal cancellation in maximal M-wave Recording Bipolar surface electromyography (SEMG) recording electrodes were placed 2cm apart, level with the motor points (and inferior to these point) on the belly of the right soleus and on the midline between the two heads of the right gastrocnemius muscles. The outputs were amplified, band-pass filtered (5-500Hz), sampled at 2kHz, and recorded on a computer (Lab View ; National Instruments). The electrode impedance was below 10kΩ; and subjects were grounded by a lip clip electrode (Türker et al., 1988) Feedback of voluntary contraction level Muscle activity from the gastrocnemius was used to provide visual feedback of voluntary contraction level to the subject. The position of the subject and recording electrodes were the same as stated above. The SEMG activity was high-pass filtered (5Hz), full wave rectified, and low passed filtered (0.1Hz). The image was then displayed on a feedback oscilloscope. Maximum voluntary contraction (MVC) was obtained over a series of up to 10 trials, where the subject was asked to contract his calf muscles as hard as he could for a 5 second period. Trials were repeated with and without visual feedback, and the strongest contraction over this period was considered to be the MVC. Preset levels of 0, 20, 40, 60, 80, and 100% MVC were randomised then displayed on an oscilloscope and the subjects were required to maintain the displayed level of contraction during the trials Stimulus The H-reflex was elicited via electrical stimulation of the tibial nerve in the popliteal fossa of the right leg. The anode was placed just above the knee, while the cathode was placed in the popliteal fossa at the point where the weakest stimulus intensity was required to elicit an H-reflex (Miles et al., 1989). Square pulse stimuli of 0.5ms 2-66

68 Signal cancellation in maximal M-wave duration were triggered from a computer and delivered by a Grass S44 stimulator. The stimulus artefact was removed using an artefact-suppressing amplifier. The lowest stimulus intensity used was 2V below that needed to elicit an H-reflex with high levels of contraction, while the upper stimulus intensity was 4V higher than that required to obtain a maximal M-response at rest. Up to 20 stimulus intensities between these values were selected, then randomly ordered. Three stimuli at each intensity were given at random intervals, each between one and three seconds apart. Therefore up to 60 stimuli were given at each contraction level. The stimuli were given immediately after the subject reached the required contraction level. Recovery periods of one minute were provided during the high levels of voluntary contraction after approximately nine stimuli, while up to five minutes rest was provided between voluntary contraction levels Analysis The three SEMG traces recorded for each stimulus intensity, at each level of contraction, in each muscle were averaged for each subject (Fig.1). Peak-to-peak (PTP) amplitude was determined by the difference between the maximum positive and negative values of the averaged response. Since the size of the M-response changed with contraction (Figure 6: page 2-72), all results are normalised to the maximum M-response for each contraction level and for each subject. To avoid contamination of the H/M max with the antidromic action potentials and a possible F response, H/M max values were obtained from the H/M max traces at stimulus intensities lower than M threshold (Figure 5: page 2-69). M threshold was set at 10% M max, at which amplitude the M-response could be detected with confidence. 2-67

69 Signal cancellation in maximal M-wave Figure 4: Average of 3 raw data traces of SEMG for one subject throughout the stimulus range Traces shown at each stimulus intensity recorded, while the subject was at rest (0% MVC) and at maximum contraction (100% MVC). Recordings from the gastrocnemius and soleus are illustrated for comparison. The figure clearly shows that the H-reflex is elicited at lower stimulus intensities than the M-response, and that the stimulus intensity required to elicit a M-response, and the maximum H-reflex size increased with voluntary contraction. Note that the size of the M-response changes dramatically when the muscle is contracting in this subject, hence the need to normalise the H- reflex to the maximum M-response for each contraction level. 2-68

70 Signal cancellation in maximal M-wave Figure 5: Normalised H/M max curves Normalised M-wave and H-reflex curves (smoothed lines connect data points). M/M max (top) and H/M max (bottom) values for one subject at rest (light gray), 40% MVC (dark gray), and 100% MVC (black) are shown. The horizontal dotted line in (A,C) indicates 10% M\M max, while the vertical dotted lines (A,B,C,D) correspond to this value for the different contraction levels. The maximum H/M max for each contraction level must occur before 10% M/M max to avoid contamination of the results by the F response. Higher H/M max values are obtained in the soleus than in the gastrocnemius. The higher H/M max values are obtained at 40% MVC in the soleus and 100% MVC in the gastrocnemius. 2-69

71 Signal cancellation in maximal M-wave Statistics One-way ANOVA was used to determine if a significant difference existed between muscles and between contraction levels along with the Bonferroni post hoc test. The Pearson s R 2 correlation statistic was also used to determine if background muscle activity changed in an orderly manner in both muscles. Significance was consistently set at p< Results The size of M max, and the H/M max (before 10% M/M max; Figure 5: page 2-68) were recorded. Examples of the averaged traces (n=3) for one subject at rest (0%) and maximum contraction (100%) in both muscles, over the full range of stimulus levels are shown in Figure 4: page These traces are typical of all subjects used, and show that the H-reflex threshold was below the M-response threshold, and that the H-reflex size steadily increased and then decreased with increasing stimulus intensity, while the M-response continued to increase to M max H-reflex Figure 4 (page 2-68) shows the average of 3 raw traces from one subject at each level of stimulus intensity, at both 0% and 100% contraction levels, while the H/M max curves are shown in Figure 5 (pg 2-69: B and D). Contraction of the lower leg muscles significantly increase the H/M max achieved in both soleus and gastrocnemius muscles M-wave No significant variation in M max existed between different levels of muscle contraction in the pooled data, however variation was evident between contraction 2-70

72 Signal cancellation in maximal M-wave levels in the data from individuals as can be seen in the large standard deviations (SD) in Figure 6 (pg 2-72). To overcome this within subject variation, and to allow data to be pooled, all results were normalised to the M max for each muscle, at each contraction level Gastrocnemius versus soleus The maximum H/M max recorded at rest and at each level of contraction was significantly larger in the soleus than the gastrocnemius (Figure 7, pg 2-73). The overall maximum H/M max value from the pooled data was 55±7% in the gastrocnemius and was achieved at 100% MVC, and 97±5% in the soleus and was achieved at 40% MVC Background Muscle Activity Feedback of background muscle activity was provided from the gastrocnemius. Figure 8: page 2-74, shows a correlation plot between average pre-stimulus (250ms) background muscle activity of the gastrocnemius and soleus, normalised to the maximum pre-stimulus muscle activity for each subject throughout their experimental period. This graph shows that during contractions up to 40% MVC, the level of background muscle activity is similar in both muscles. It also shows that in contractions of greater than 40% MVC, the increases in background muscle activity in gastrocnemius were not matched by increases in the soleus background muscle activity. 2-71

73 Signal cancellation in maximal M-wave Figure 6: Maximal M-wave for each contraction level, normalised to overall maximal M-wave Mean+SD of normalised M maximum are shown for each voluntary contraction level. Considerable variation in M maximum size did exist within subjects at different contraction levels, as can be seem by the large Standard Deviations (SD), however this variation between contraction was not consistent between subjects, which results in no significant difference in the pooled data. All subsequent results were normalised to maximum M-wave for each subject at each contraction level to overcome the individual variation seen here and to allow results to be pooled. 2-72

74 Signal cancellation in maximal M-wave Figure 7: Mean SEM of H/M max, data pooled from 8 subjects Significant differences between voluntary contraction levels (*) (*<.05; **<.01), and between muscles ( ~ ) are shown ( ~ <.05; ~~ <.01). The maximum soleus H/M max is significantly larger than the gastrocnemius maximum H/M max at every level of contraction. 2-73

75 Signal cancellation in maximal M-wave Line of equality Figure 8: Correlation between levels of normalised pre-stimulus background muscle activity in the gastrocnemius and soleus Data points represent normalised background muscle activity given the same subject, stimulus intensity and contraction level (normalised to each subjects maximum pre-stimulus level). Feedback of muscle activity was provided from the gastrocnemius throughout the study. This second order polynomial regression curve (R 2 =.51) demonstrates that up until 40% MVC the level of activity in the soleus increased linearly with the increases in gastrocnemius, but that after this, the muscle activity in the soleus remains pretty constant while the gastrocnemius continues to grow. 2-74

76 Signal cancellation in maximal M-wave 2.5. Discussion There are two main findings in this study. 1. The size of H/M max is larger in the soleus than in the gastrocnemius at all levels of contraction, which suggests that gastrocnemius and soleus motoneurons are recruited differently by spindle input at rest and at all levels of voluntary contraction. 2. The M max size was altered in response to contraction in individual subjects Gastrocnemius and soleus motoneurons are recruited differently The results of the current study suggest that the soleus and gastrocnemius motoneurons are differentially recruited by spindle input. This may mean that the two muscles have different input distributions from muscle spindles and/or that the motoneuronal pools respond differently to spindle input. The soleus and gastrocnemius muscles are ideal for this comparative study as they are both innervated by the same mixed nerve. The tibial nerve encapsulates both spindle afferents and motor nerves for these muscles, and therefore the same stimulus will evoke a response in both muscles. We have demonstrated that the maximum H/M max obtained in these experimental conditions, is greater in soleus than in gastrocnemius. This is reflected in the H/M max at each level of contraction including at rest, and suggests that the soleus motoneurons have greater recruitment tendencies from spindle input than the gastrocnemius in these conditions. The difference in recruitment tendencies may be explained by either the soleus motoneurons having a lower recruitment threshold to the same spindle input than gastrocnemius motoneurons, and/or that more soleus spindle afferents are recruited at a given stimulus intensity, and provide their own 2-75

77 Signal cancellation in maximal M-wave motoneurons with greater input current than do gastrocnemius spindle afferents (Vanden Noven et al., 1986). In humans it has been shown that the soleus muscle has almost three times as many muscle spindles (Voss, 1971), and a far higher proportion of slow twitch fibres (Johnson et al., 1973; Edgerton et al., 1975; Green et al., 1981) than the gastrocnemius. It has also been shown that homogenous spindle projections are stronger than heterogeneous projection to these two muscles (Mao et al., 1984). It is therefore possible that soleus motoneurons receive greater spindle input and hence more input current per motoneuron. Furthermore, the soleus also has larger numbers of smaller motoneurons supplying slow twitch fibres (Burke, 1991). Therefore it is not surprising that the soleus motoneurons have lower recruitment thresholds than the gastrocnemius motoneurons to spindle input. During voluntary contraction it may be argued that other cortical and peripheral factors will influence the size of the H-reflex, and that therefore we cannot conclude that the spindle feedback to these muscles is different simply based on the voluntary contraction data. It is well understood that the output of a motoneuron is a result of the net effect of a complex process of integration (excitatory and inhibitory) of inputs from many other neurons (Henneman et al., 1965; Burke et al., 1970; Calancie and Bawa, 1990; Burke, 1991; Semmler and Türker, 1994). Several studies have shown for example that cutaneous input can alter the recruitment order of motor units (Kanda et al., 1977; Garnett and Stephens, 1981; Kanda and Desmedt, 1983), and it is understood that selective recruitment exists within a motor unit pool, which activates only the muscle units that have appropriate mechanical characteristics to suit the specific demand of a task. It is possible therefore that other peripheral and central modulatory factors are influencing the excitability of the soleus and gastrocnemius motoneuron pool at the same time as the experimentally controlled voluntary contraction level and Ia afferent (H reflex) volley s are being tested. 2-76

78 Signal cancellation in maximal M-wave Although this is true, the current study clearly shows that the size of the H/M max is significantly larger in the soleus than in the gastrocnemius at rest (as well as at every level of contraction). This alone therefore, provides strong evidence for our conclusion Primary afferent depolarisation During voluntary contraction in the leg it is possible that co-activation of antagonist muscles can occur (De Luca and Mambrito, 1987). Therefore it is possible that spindle input from other antagonist muscles pre-synaptically inhibit the test muscle spindle input. Similarly, receptors from within the skin, tendons and joints surrounding the test muscle and the spindle input from the homonymous muscle can generate primary afferent depolarisation (PAD) on the spindle afferent (Jankowska and Lundberg, 1981; Jankowska, 1984; for review see Schieppati, 1987; Zehr, 2002). Furthermore, many central and peripheral inputs Further to this, it is possible that at higher stimulus intensities (around motor threshold and above) afferents from these receptors may be recruited by the electrical stimulation of the mixed nerve (for review see Misiaszek, 2003). In the current investigation the H/M max values were obtained at stimulus intensity at or below the motor threshold, which reduces the chance of electrically induced PAD affecting the results. Furthermore, the effect of stimulating these afferents cannot be considered significant in the soleus given the high levels (close to 100%) of H/M max that were achieved. It cannot be ruled out however, that PAD from antagonist muscles and other receptors may play a part in the smaller H/M max scores seen in the gastrocnemius and in soleus at higher levels of activity. It is also possible that the current protocol for activating these muscles (foot at 90 o ) favours soleus motoneuron recruitment because the soleus motoneurons are more active when the foot is positioned at 90 0 while the gastrocnemius is more active at 2-77

79 Signal cancellation in maximal M-wave higher degrees of plantar flexion (Duysens et al., 1991). This should be overcome, however, by the voluntary controlled contraction, particularly in the gastrocnemius, given that feedback was provided from this muscle Recurrent inhibition It is possible that recurrent inhibition via the Renshaw Cell system could affect the max H-reflex obtained during contraction, particularly in the gastrocnemius. This system is an effective means through which to inhibit small (type I) motor units during an activity in which the larger (type II) motor units are being recruited (Friedmann et al., 1981). Given that the gastrocnemius has a greater percentage of type II motor units than the soleus (Johnson et al., 1973; Edgerton et al., 1975; Green et al., 1981), voluntary contraction of this muscle may lead to greater recurrent inhibition of the smaller type I motor units (via Renshaw cells), and therefore reduce the chance that the whole motor neuron pool will be recruited by the H-reflex at high contraction levels Hyperpolarisation Given the density of low threshold motoneurons in the soleus, it is possible that the majority of the soleus motoneuron pool are active at 40% MVC. If most motoneurons are recruited at 40% MVC, and further voluntary drive is provided, only the frequency of the firing of these neurons can increase in the soleus. With increased firing frequency at very high levels of contraction, more hyperpolarization induced refractory periods will occur in the motoneuron, which could reduce the chance of spindle input recruiting the motoneurons (Gossen et al., 2003). The duration of this effect is variable, and is known to be longer in the soleus muscle than in many others (Matthews, 1996). This may be explained by the overall smaller soleus motoneuron 2-78

80 Signal cancellation in maximal M-wave and axon size. However in the current study, the H/M max size reached almost 100%, and no significant decrease in the H-reflex size was seen at high levels of MVC. This suggests that hyperpolarisation does not have a significant effect on the H-reflex in the soleus. Given that hyperpolarisation is not considered to be an important factor in the soleus, it is not considered to be a contributing factor in the smaller levels of H/M max in the gastrocnemius Changes in M max during contraction studies The M max response represents total activation of all the motor axons to a particular muscle. It is therefore expected to be stable between conditions during an experiment when stimulating and recording conditions remain the same, and fatigue is not an issue. Recently it has been shown that the amplitude of the maximal M- response typically decreases during the course of an experiment (Crone et al., 1999). It has also been shown that the M max size changes in the soleus systematically during plantar flexion, which may have been the result of distortion of the muscle and/or changes in the effectiveness of the electrical stimulation (Simonsen and Dyhre-Poulsen, 1999; Fumoto et al., 2002). In the current study large changes in M max were noted between contraction levels for individual subjects, however when the data were normalised and grouped no significant differences between contraction levels existed. It is most likely that this is due to distortion of the muscle/recording electrode or nerve/stimulating electrode relationship during the experiment, and it is conceivable that the distortion is not consistent between subjects. The lack of significance may also be a result of the limited number (3) of stimuli used at each contraction level, and therefore further investigation is required to determine why the M max can change during experiments. 2-79

81 Signal cancellation in maximal M-wave Further discussion points and limitations Background muscle activity At low levels of contraction, the level of background muscle activity was similar in both the soleus and gastrocnemius. However in contraction levels greater than 40%, the increases of gastrocnemius background muscle activity are not seen reliably in the soleus. This may be expected because feedback is given from the gastrocnemius. However given that the soleus reached its peak of almost 100% H/M max at 40% MVC this is not considered a limitation of the study M-response threshold The M-response threshold was set at 10% in the current investigation. This is higher than that used in previous studies (Funase et al., 1996), however during high levels of voluntary contraction, small peaks and variability exist in the reflex traces (see Figure 4, pg 2-68). This variability can result in small responses being recorded in error during the expected M-response and/or H-reflex periods. If the M-response threshold was set lower than 10% there is greater chance that an error in the recorded max H/M max size (before M-response threshold) would exist. It is also important to not contaminate the max H-reflex size with the F response, as F response can occur particularly during high stimulus levels (Trontelj, 1973b) Voltage stimulator In the current investigation a constant voltage stimulator was used rather than a constant current stimulator. Because the stimuli given to each subject were adjusted according to the stimulus intensity necessary to activate all motor axons (maximum M-response), the use of a constant voltage stimulator is not considered a limitation. 2-80

82 Signal cancellation in maximal M-wave Stimulus rate The stimuli in the current investigation were given randomly between one and three seconds apart. Research suggests that during contraction, stimuli of up to 4Hz can be used without significant loss of reflex amplitude (Burke et al., 1989). The same study suggests that in relaxed conditions, the reduction in H-reflex amplitude may be as great as 85% using high stimulus rates and low stimulus intensities (Burke et al., 1989; for review see Misiaszek, 2003). In the current study, a small reduction (less than 8% at rest and 5% during 100% MVC) in H-reflex size is noted between the first and second stimuli (data not shown), with no further change in the size of the reflex between the second and third stimuli, this indicates that homosynaptic depression is not occurring. The main findings of this study rely on stimulus intensities near motor threshold and on muscles under voluntary contraction conditions; therefore we do not believe that the stimulus rate will have significantly altered the results of this study Significance The present work illustrates that the soleus and gastrocnemius are recruited differently by feedback from muscle spindles. The larger H/M max values suggest that more spindle feedback is provided to the soleus muscle, which is supported by the larger number of spindles in this muscle. The findings are in agreement with the normal recruitment order of motoneurons, as the soleus has a greater density of small motor units, which will be recruited first from spindle input. This work highlights the importance of spindle feedback in postural muscles, and suggests that the soleus has greater spindle input distribution than the gastrocnemius and is therefore potentially better suited for postural work. 2-81

83 Signal cancellation in maximal M-wave 3. A new method to estimate signal cancellation in the human maximal M- wave This chapter is an edited version of the manuscript A new method to estimate signal cancellation in the human maximal M-wave by K J Tucker, and K.S. Türker, which has been published in the Journal of Neuroscience Methods (2005) and is included in the appendix of this thesis Outline of study A new method is introduced that estimates EMG signal cancellation in surface recorded investigations. Its usefulness is demonstrated when determining changes in the maximal motor response (M-wave) magnitude during rest and voluntary contraction. The accuracy of recording and analysis methods and the reliability of the maximal M-wave were assessed in the human gastrocnemius and soleus. The maximal M-wave was recorded by bipolar surface electrodes placed 2 cm, 3 cm and 4 cm apart, and by monopolar (one active and one indifferent reference) surface electrodes. Up to 85% of the maximal M-wave was lost due to signal cancellation during bipolar recording. The maximal M-wave magnitude decreased consistently and significantly during triceps surae contraction compared to rest when recorded by monopolar electrodes, but not when recorded by bipolar electrodes. Area and peakto-peak amplitude analysis methods provided similar results when determining the magnitude of the maximal M-wave. This provides evidence that monopolar recording 3-82

84 Signal cancellation in maximal M-wave is superior to bipolar recording as it removes the signal cancellation error and allows the genuine changes in maximal M-wave magnitude to be observed Introduction The motor response (M-wave) represents the excitation of motor axons while the maximal M-wave represents total activation of the motor neuron pool. The M-wave is evoked by direct electrical stimulation of motor efferents, and is used regularly as a normalisation factor in reflex studies (for review see (Pierrot-Deseilligny and Mazevet, 2000; Zehr, 2002). The stability of M-wave magnitude during muscle contraction, limb movement and rest, illustrates that a constant number of nerve fibres are excited by the test stimuli, (Boorman et al., 1996, Crone et al., 1990), and it is therefore used as an indicator of the nerve to stimulating electrode relationship throughout experimental sessions (for review see Zehr, 2002). A review of the current literature however, suggests that the magnitude of the maximal M-wave may be affected by voluntary contraction or movement of the test muscle (Chapter 2; Nagata and Christianson, 1995; Tucker and Türker, 2004), and decreases over time during the course of long-lasting experimental studies (Crone et al., 1999). Each of these studies describes changes in the magnitude of the maximal M-wave within subject trials, but also report variability between subjects. One of the possible explanations for the variability in maximal M-wave magnitude is inadequate recording techniques given that all of the studies which report this variability during contraction or limb movement use bipolar recording electrodes with an inter electrode distance of less than or equal to 2 cm. The use of bipolar recording in surface EMG studies is favoured because of its ability to remove unwanted noise from the system, via common mode rejection. Common mode rejection aids in the removal of cross talk from muscles in close proximity to the test muscle, and other 3-83

85 Signal cancellation in maximal M-wave electrical artefacts that originate from both inside and outside the body. This process occurs because the two oppositely-poled surface electrodes receive the noise signal at the same time, summates the two signals, and subsequently eliminates the noise from the recorded signal (De Luca and Merletti, 1988; Türker, 1993). One of the potential problems with bipolar recording is that of signal cancellation (also reported as phase cancellation: Moore, 1967; De Luca, 1978; Day and Hulliger, 2001; Keenen et al., 2005). While bipolar recording removes noise from the recorded trace, if the beginning of the maximal M-wave is being recorded by the second electrode before the wave has completely passed by the first electrode, it is plausible that some real signal, originating from the test muscle, may also be cancelled (Moore, 1967; Day and Hulliger, 2001; Keenen et al., 2005). The amount of signal cancellation that will occur is therefore related to the distance between the oppositely poled surface electrodes (Roeleveld et al., 1997), and the duration of the recorded signal (Farina et al., 2002, 2003). The duration of the maximal M-wave is determined by the dispersion and spatial filtering of action potentials that underlie the signal. Dispersion of these action potentials will occur due to recruitment tendencies of the motor nerve, conduction velocity of the motor axons (Rutten et al., 1998) and muscle fibres (Buchthal and Schmalbruch, 1970; for review see Burke, 1981), impairment of neuromuscular propagation (Fuglevand et al., 1993), and summation of the underlying potentials (Day and Hulliger 2001; Roeleveld and Stegeman, 2002; Farina et al., 2003; Beck et al., 2004). The low-pass spatial filtering of the signal occurs as the signal passes through tissue and interstitial fluid from the intramuscular fibres to the surface recording electrodes (Stegeman et al., 2000). 3-84

86 Signal cancellation in maximal M-wave The primary aim of this study was therefore to investigate the possibility of signal cancellation, and its affects on the surface recorded maximal M-wave magnitude; and to illustrate this cancellation clearly using both bipolar recording electrodes placed 2 cm, 3 cm and 4 cm apart on the test muscles belly, and using bipolar recording electrodes where the active electrode was placed over the muscle belly and the indifferent reference electrode place on the ipsilateral heal (reported from here as monopolar: see Roeleveld and Stegeman 2002). This suggested signal cancellation has been demonstrated in simulation studies (Arabadzhiev et al. 2005a,b), but not in true maximal M-wave investigations. It was hypothesised that recording the signal with bipolar electrodes would cancel a proportion of the recordable signal, that the amount of cancellation would be related to the distance between the bipolar electrodes, and that monopolar recording would overcome this cancellation. When significant dispersion of the signal occurs, it is also possible that the analysis methods used could influence the results obtained. This is because analysis methods such as peak-to-peak (PTP) amplitude rely upon simultaneous arrival of action potentials to the recording electrodes. It is therefore possible that area analysis may be more suitable for surface recorded signals because the area calculation allows for dispersion of the response over time (Crone et al., 1999). This has been shown in hand muscles, where during fatiguing contractions, the PTP amplitude of the maximal M-wave decreased, while the duration of the response increased, resulting in no significant change in the maximal M-wave magnitude when analysed by area (Fuglevand et al., 1993). The second aim of this study was therefore to determine if area and PTP amplitude analysis methods affect the reported magnitude of the maximal M-wave. It was hypothesised that area calculations would be a superior analysis tool for determining the magnitude of the surface recorded maximal M-wave. 3-85

87 Signal cancellation in maximal M-wave The medial gastrocnemius and soleus muscles are ideal for such surface recorded reflex studies. Both muscles are innervated by the tibial nerve, allowing them to be activated by the same electrical stimuli (see Figure 2: page 1-42 and Figure 1: page 1-31). Furthermore, human histochemical studies show that there is an approximately equal distribution of slow (type I) and fast (type II) twitch fibres in the gastrocnemius, while the soleus has about 70-90% slow and 10-30% fast twitch fibres (Johnson et al., 1973; Edgerton et al., 1975; Green et al., 1981). Therefore, greater dispersion of the maximal M-wave may be expected in the gastrocnemius where greater variability in muscle fibre types exists. Because of the greater fibre type distribution and potentially greater dispersion of the signal in the gastrocnemius, it was further hypothesised that the signal cancellation would be larger in the gastrocnemius than in the soleus. Finally, to test the accuracy of our signal cancellation hypotheses, and to determine if this process may influence the results of a maximal M-wave study, the maximal M- wave was recorded at rest, and during contraction of the triceps surae. It was hypothesised that the magnitude of the maximal M-wave would be different during contraction and rest trials, and that this difference would only be seen consistently in the monopolar recorded data Materials and Methods Subjects Ten healthy human subjects (6 female, 4 male) between the ages of gave written, informed consent to participate in this study. The experiments were approved by the Human Ethics Committee of the University of Adelaide and all procedures used conformed to the Declaration of Helsinki. 3-86

88 Signal cancellation in maximal M-wave Protocol The subjects lay prone, on a standard physiotherapy table, with a face hole that allowed visual feedback of background muscle activity from a computer monitor positioned under the table. The subject s right ankle lay at rest (R) with no support (R1 at beginning and R2 at end of experiment); and at plantar flexion with 90% maximum voluntary contraction (MVC) of medial gastrocnemius muscle (PF90). A foot bar provided the resistance required to produce the 90% MVC Recording Surface electromyography (SEMG) recording electrodes (all Duo-Trode silver/silver chloride) were placed 2 cm, 3 cm, and 4 cm apart, on the belly of the right medial gastrocnemius and soleus muscles. The superior 3 cm recording electrode was also used as the active electrode for the monopolar recording configuration, with the distal reference electrode placed on the ipsilateral heal (Figure 9: page 3-88). The term monopolar recordings is used only for convenience since they were actually bipolar recordings in which the reference electrode was located at such a distance from the generator that the volume-conducted potentials were comparable to the noise level (Gerilovsky et al., 1985 and 1989; Pierrot- Deseilligny and Mazevet, 2000). A total of 8 SEMG channels were recorded throughout the experiment. The monopolar and bipolar surface EMG recordings were amplified in a custom made EMG amplifier, with built in stimulus artefact suppressor and high pass (20Hz) filter, and were recorded on a computer (LabVIEW : National Instruments) at a sample rate of 5kHz. The skin resistance of all the electrode pairs was below 10kΩ; and subjects were grounded by lip clip electrodes (Türker et al., 1988). 3-87

89 Signal cancellation in maximal M-wave Figure 9: Bipolar and monopolar electrode arrangement The cathode of the stimulator (st) was positioned in the popliteal fossa of the right leg, while the large square anode was placed above the patella on the same leg (not shown). Surface recording electrodes were positioned on the medial gastrocnemius and soleus, 2 cm, 3 cm and 4 cm apart. A monopolar (m) reference electrode was also placed on the heal of the ipsilateral leg. The electrodes were placed 2 cm medial to the midline of the gastrocnemius and soleus muscles, and 3 cm superior (gastrocnemius) and 2 cm inferior (soleus) to the insertion of the gastrocnemius to the Achilles tendon. 3-88

90 Signal cancellation in maximal M-wave Feedback of voluntary contraction level The muscle activity required for feedback during the voluntary contraction trial was recorded from the bipolar electrodes, placed 3 cm apart on the gastrocnemius. MVC was obtained over a series of 3 trials, whereby the subject was asked to contract their calf muscles as hard as they could, for a period of 5 seconds. The SEMG activity was band-pass filtered (20-500Hz), full wave rectified and low passed filtered (1Hz), and the maximum contraction level over that period was defined as 100% MVC. A preset target line of 90% MVC was displayed on the feedback monitor during the contraction condition Stimulus The motor response (M-wave) was elicited via electrical stimulation of the tibial nerve of the right leg. A large (~ 100 cm 2 ) copper anode was placed just above the patella, while the silver ball cathode, with a 1 cm diameter, was placed in the popliteal fossa, at the point where the weakest stimulus intensity was required to elicit a Hoffmann (H) reflex (for details see Miles et al., 1989). Square pulse stimuli of 0.5 ms duration were triggered from a computer and delivered by a Digitimer constant current stimulator (model DS7A). The stimulus intensity employed to obtain the maximal M- wave was determined by increasing the stimulus intensity gradually, and assessing, with online feedback, if the area and peak-to-peak (PTP) amplitude of the M-wave also increased. The maximal M-wave threshold was established as the stimulus intensity which no longer caused an increase in the M-wave magnitude in either the soleus or gastrocnemius. The maximal M-wave threshold was determined for each subject in each of the 8 SEMG channels, at rest and during contraction. The stimulus intensity employed throughout the experiment was 1.5 x the maximal M-wave threshold (for review see Pierrot-Deseilligny and Mazevet, 2000). 3-89

91 Signal cancellation in maximal M-wave Twenty stimuli were given during each of the experimental conditions at random intervals between 3 and 5 seconds apart. During the contraction trial, the stimuli were given immediately after the subject reached 90% MVC. Recovery periods of one minute were provided after each 4 stimuli during the contraction trial. At least 2 minutes rest was provided between each experimental condition Analysis The 20 SEMG traces recorded for each condition were averaged for each subject. PTP amplitude was determined by the difference between the maximum and minimum values of the averaged responses, while area was calculated by the cumulative sum (CUSUM) of the rectified trace (for a full description see Chapter 1; Tucker et al., 2005; also see Ellaway, 1978; Brinkworth and Türker, 2003). Two analysis procedures were conducted on these data. Firstly, we wanted to determine if the inter-electrode distance (including monopolar recording) affected the magnitude of the maximal M-wave. To do this, the magnitude of the maximal M-wave recorded with the different electrode configurations, were normalised to the largest maximal M-wave recorded across these conditions. The second procedure was conducted to determine if the magnitude of the maximal M-wave altered between test muscle conditions. All data were separated into their electrode configuration groups. The maximal M-waves achieved during the different muscle conditions were then normalised to the largest maximal M-wave achieved throughout the muscle conditions Statistics The average (N=20) maximal M-wave magnitude (area and PTP amplitude) was calculated for each subject in each condition. A one-way ANOVA was used on the 3-90

92 Signal cancellation in maximal M-wave pooled data (N=10) to determine if a statistically significant difference existed between the different electrode arrangements, and between rest and voluntary contraction, in each muscle. In each case, the Bonferroni post hoc test was used to determine where any significant differences lay. A paired T-test was used to determine if a statistically significant difference existed between the magnitudes of the maximal M-waves, as recorded by different electrode arrangements in the gastrocnemius compared to the soleus. Pearson s correlation was used to determine if the results obtained by area and PTP amplitude analysis techniques were correlated. In all cases the level of significance was set at p< Results Maximal M-wave magnitude at different electrode placements To determine if the distance between recording electrodes changed the maximal M- wave magnitude, data were normalised to the largest maximal M-wave obtained at different electrode arrangements. In both muscles and using both analysis techniques, the magnitude of the monopolar maximal M-wave magnitude was significantly larger (p<0.01) than the maximum bipolar-recorded response. The magnitude of the maximal M-wave in the bipolar recorded traces was also significantly larger (p<0.01) when the electrodes were placed 4 cm apart, than with the 2 cm and 3 cm electrode placements (data from R1, area analysed trial: gastrocnemius monopolar (0.98±0.05); 4 cm (0.35±0.27); 3 cm (0.17±0.09); 2 cm (0.13±0.08), soleus monopolar (1.0±0.0); 4 cm (0.46±0.25); 3 cm (0.27±0.19); 2 cm (0.25±0.18); Figure 10: page 3-93). The reduction in maximal M-wave magnitude that occurred with bipolar recording was significantly larger in the gastrocnemius than in the soleus at all electrode distances when analysed by PTP amplitude (p<0.05 for 3 cm and 4 cm; p<0.01 for 2 cm; and when analysed by area: p<0.05 for 2cm; data 3-91

93 Signal cancellation in maximal M-wave from R1 trial; Figure 10: page 3-93). Similar results were seen in the R2 and PF90 conditions (data not shown) Maximal M-wave magnitude during rest and contraction The maximal M-wave magnitudes (means ± standard deviations), normalised to the maximal M-wave in each of the trials, are shown in Figure 11: page In the bipolar data, the only significant difference in maximal M-wave magnitude between the rest and contraction exists in the soleus 2 cm traces, when analysed by area. In this case, the maximal M-wave obtained during contraction (0.82±0.2) was significantly smaller (p<0.05) than that obtained in either R1 (0.97±0.05) or R2 (0.96±0.05). When recorded by monopolar electrodes, the maximal M-wave magnitude was significantly smaller (p<0.01) during contraction than during R1 and R2 in both muscles, using either area of PTP amplitude analysis (Figure 11: page 3-94: gastrocnemius area: R1 (0.99±0.02); R2 (0.99±0.02); PF90 (0.61±0.2); soleus area: R1 (0.99±0.02); R2 (0.96±0.04); PF90 (0.74±0.18); gastrocnemius PTP amplitude: R1 (0.94±0.08); R2 (0.95±0.08); PF90 (0.69±0.21); soleus PTP amplitude: R1 (0.98±0.05); R2 (0.95±0.06); PF90 (0.74±0.18)). All records indicate no significant change in the maximal M-wave magnitude existed between the two rest conditions Area vs PTP amplitude analysis methods The correlation of normalised maximal M-wave magnitude between area and PTP amplitude analysis methods is highly significant in both muscles at every electrode arrangement (all data: r 2 > 0.69; p<0.01, except soleus 3cm, R1 and R2 where r 2 =0.42; p<0.05). 3-92

94 Signal cancellation in maximal M-wave Figure 10: Maximal M-wave magnitude at each electrode arrangement Data shown from the R1 condition. The maximal M-wave magnitude at each electrode arrangement is normalised to the largest maximal M-wave magnitude obtained from any of the electrode arrangements. Up to 85% of the maximal M-wave was lost during bipolar recording when compared to the monopolar maximal M-wave magnitude. Significant reductions (**p<0.01; *p<0.05) in maximal M-wave magnitude were also evident between the 4 cm, 3 cm and 2 cm data, particularly when analysed by PTP amplitude. The smaller maximal M-wave magnitude recorded from bipolar electrodes compared to monopolar electrodes, and the increase in size of the maximal M-wave magnitude with increased distance between electrodes, supports the signal cancellation hypothesis. Greater signal cancellation is seen in the gastrocnemius compared to the soleus ( ## p<0.01; # p<0.05) particularly when analysed with PTP amplitude (curved significance lines), which illustrates greater dispersion and signal cancellation of the maximal M-wave in the gastrocnemius compared to the soleus. 3-93

95 Signal cancellation in maximal M-wave Figure 11: Magnitude of the maximal M-wave obtained during all muscle contraction trials The magnitude of the maximal M-wave was normalised to the maximal M-wave obtained during all muscle condition trials. The maximal M-wave (Mean±SD) during R1 (light grey) R2 (dark grey) and PF90 (white), normalised to overall maximal M- wave are shown. Data were recorded by electrodes 2 cm, 3 cm, 4 cm apart, and by the monopolar (Mono) electrodes, in both the gastrocnemius and soleus. Results shown for both area and PTP amplitude analysis techniques. A significant (**p<0.01) reduction in maximal M-wave exists between the rest and contraction conditions when recorded by monopolar electrodes in both muscles when either analysis methods were used. When recorded by 2 cm bipolar electrodes, a significant (*p<0.05) reduction in maximal M-wave magnitude was also seen in the contraction condition compared to rest in the soleus. No significant differences are seen between the two rest trials when recorded by any of the electrode arrangements. 3-94

96 Signal cancellation in maximal M-wave 3.5. Discussion Our results suggest that significant signal cancellation occurs in bipolarly recorded SEMG traces. The magnitude of the maximal M-wave was significantly smaller in all bipolar recordings when compared to monopolar recordings (N.b. monopolar recording consisted of an active electrode placed on the test muscle belly and a reference electrode placed on the ipsilateral heal). Furthermore, the maximal M-wave magnitude also reduced as the distance between the bipolar electrodes decreased. Greater signal cancellation was also noted in the gastrocnemius when compared to the soleus. This was expected because the gastrocnemius has greater fibre type variability which we argue, may lead to increased dispersion of action potentials underlying the maximal M-wave. In the current investigation, both analysis methods produced similar results, and we therefore conclude that either area or peak-to-peak (PTP) amplitude analysis methods may be used in maximal M-wave studies when fatigue is not an issue. No significant change in the maximal M-wave magnitude was seen during rest compared to during contraction in the bipolar-recorded data (with exception of the soleus 2 cm, area analysed data). The maximal M-wave magnitude was significantly different however between the rest and contraction condition when recorded by monopolar electrodes. We suggest that the lack of significant difference noted in the bipolar-recorded data, is due to the large signal cancellation error Monopolar vs bipolar recording and the signal cancellation error Although bipolar SEMG recordings contain less cross talk and noise than monopolar recordings (Türker, 1993; Ohashi, 1995), one of the problems that exists with bipolar 3-95

97 Signal cancellation in maximal M-wave recording is that part of the signal may be cancelled due to the polarity of the electrode discs. Signal cancellation (Figure 12: page 3-98) occurs as a direct result of temporal summation of the same signal, when the beginning of the signal is being recorded by the second (negative) electrode, before the signal has completely passed the first (positive) electrode. The existence of signal cancellation (or phase cancellation) has previously been reported in simulation studies (Moore 1967; Day and Hulliger 2001; Keenen et al., 2005), and is supported by the work of Farina and colleagues (2002 and 2003) and Roeleveld and colleagues (1997) who demonstrated that the PTP amplitude of single fibre potentials, and muscle activity during active postural tasks, increased when the inter electrode distance increases. Our monopolar recordings show that the action potentials that make the maximal M- wave are dispersed by about 10 ms when recorded by surface electrodes (Figure 13: page 3-99 for monopolar recorded actual data). This 10 ms dispersion is a result of: the time that it takes for the original stimulus (0.5 ms duration) to elicit a response in a variety of motor axons; differences in the conduction velocity of action potentials along the length of motor axons; the time taken for action potentials to cross the neuromuscular junction; the differences in the conduction velocity of a variety of different muscle fibre types; and the filtering of the motor unit potential through the interstitial fluid (decreasing its frequency component, and increases the duration of the potential) to the recording electrodes (reviewed in Stegeman et al., 2000). Further to this, a motor unit potential travels at an approximate speed of 4-5 m/s along muscle fibres (Roy et al., 1986). Therefore, theoretically, given a signal with 10 ms dispersion, the electrodes would need to be more than 5 cm apart, parallel to the long length of the muscle fibre and on one side of the motor point, to avoid any cancellation that will occur when the input to oppositely poled electrodes are summated. This suggestion is supported by our results, which show that as the 3-96

98 Signal cancellation in maximal M-wave distance between the recording electrodes increases from 2 cm to 3 cm, then to 4 cm apart, the magnitude of the maximal M-wave increases. It also shows that even at 4 cm apart, the magnitude of the resulting maximal M-wave is up to 60% smaller than the monopolar recorded maximal M-wave (see Figure 10: page 3-93 and Figure 13: page 3-99). We have called the cancellation of this surface recorded signal the signal cancellation error. Figure 13: page 3-99 is an example of both real data and a simulation of the signal cancellation error. In this example, when the same signal is recorded by both positive and negative electrodes (which occurs during bipolar recording) up to 88% of the actual response is lost due to signal cancellation. We suggest that this is the primary reason for inconsistent variability of the maximal M-wave magnitude between subjects, that has been obtained in both our own research (see large SD in Chapter 2, Figure 6, pg 2-72 and Tucker and Türker, 2004) and in other studies (Nagata and Christianson, 1995) when bipolar recording electrodes were used. We further suggest that this variability may prevent the real differences in the maximal M-wave magnitude being observed. Given this finding, we believe that to avoid signal cancellation related problems, and to allow the actual changes in maximal M-wave magnitude (during different test conditions) to be determined, it is important to use monopolar recording electrodes (with a reference electrode placed on a non-active site e.g. the heal) during surface EMG investigations. 3-97

99 Signal cancellation in maximal M-wave Figure 12: Representation of the signal cancellation error Temporal summation of the maximal M-wave from bipolar (positive and negative) electrodes results in a signal cancellation error. This figure shows a signal as recorded by a positive electrode (A); a signal as recorded by a negative electrode (B: the longer latency of B is a result of the electrode being more distal from the motor point); and (C) the trace which results from temporal summation of the signals recorded simultaneously from the positive and negative electrodes (C=A+B). The duration whereby the end of the signal as recorded by the positive electrode (A) and the start of the signal recorded by the negative electrode (B) overlap, is highlighted by the grey box. Given that the bipolar electrodes summate the information from both electrodes and produce one output, the resulting maximal M-wave shape and size can be reduced significantly. This example shows a small temporal summation error (approximately 40% reduction in PTP amplitude), the results of this study, and the example shown in Figure 13 (pg 3-99), demonstrates that this process can cancel up to 85% of the true maximal M-wave. 3-98

100 Signal cancellation in maximal M-wave Figure 13: Simulation and actual data of signal cancellation that occurs during bipolar recording The top four traces are simulations of the cancellation that occurs in the actual data (bottom four traces) from monopolar, 4 cm, 3 cm and 2 cm bipolar electrodes. In the top figures the dotted trace represents the actual motor potential if recorded by a DC amplifier. In each of the figures the bold trace represents the output from an amplifier with 20 Hz high pass filtering. Traces were predicted using a Labview simulator, by creating a positive peak for the monopolar trace and decreasing the distance between the same size negative and positive peaks for the bipolar traces. The maximal M-wave area and PTP amplitude, normalised to the monopolar data, are also shown. Up to 88% of the maximal M-wave (actual gastrocnemius data) is reduced by signal cancellation in example shown. 3-99

101 Signal cancellation in maximal M-wave Signal cancellation and fibre type variability In the current investigation, greater signal cancellation was also noted in the gastrocnemius when compared to the soleus. This finding is supported by the understanding that a strong correlation exists between the cross sectional area and conductance velocity of muscle fibres (Roeleveld and Stegeman, 2002; in humans: Taylor et al., 1997; Sadoyama et al., 1988, and in rats: Kupa et al., 1995). Although there is evidence that the cross sectional area of human muscle fibres do not significantly increase from type I to type II, there are large increases in the mean cross sectional area between these fibre types, and large variability in the cross sectional area of type II muscle fibres over all (Harridge et al., 1996; Miller et al., 1993). Combined, this information does suggest that greater variability in muscle fibre conduction velocity will occur in the gastrocnemius, than in the soleus. Given that the gastrocnemius has greater fibre type variability, and that greater fibre type variability may lead to greater dispersion of the maximal M-wave, a greater amount of signal cancellation was expected to occur in this muscle. It is also possible however, that the greater reduction in recorded signal reported in the gastrocnemius is a result of increased cross talk elimination in this muscle during the bipolar recording. The superior portion of the soleus muscle lies beneath the body of the gastrocnemius with its origins in the superior posterior surface of the tibia and fibula. It is therefore possible that cross talk from this part of the soleus is recorded by the gastrocnemius surface electrodes, while the opposite is less likely given the position of the soleus recording electrodes were inferior to the gastrocnemius insertion into the Achilles tendon. It is well understood that the cross talk signal disperses significantly as it travels via spatial conductance though the 3-100

102 Signal cancellation in maximal M-wave interstitial tissues to be recorded by the surface electrodes placed on other proximal muscles (De Luca and Merletti, 1988; Türker, 1993). We have previously shown that cross talk from the soleus to the gastrocnemius is minimal when recording with 2 cm bipolar electrodes (Chapter 2; Tucker and Türker 2004), which may be largely due to common mode rejection (see below). It is therefore possible that it is cancellation of the dispersed cross talk signal that results in greater signal cancellation in the gastrocnemius than in the soleus when the recording electrodes are placed with smaller inter electrode distances Monopolar vs bipolar recording and changes in the maximal M-wave magnitude In previous investigations we have demonstrated that large but inconsistent variability exists in the magnitude of the maximal M-wave during increased levels of muscle contraction, when 2 cm bipolar recording electrodes are used (Chapter 2; Tucker and Türker, 2004). Similarly, Simonsen and Dyhre-Poulsen, (1999) have shown that the maximal M-wave magnitude (recorded by bipolar electrodes) does change during a walking cycle, however that the variability in the maximal M-wave was not consistent between subjects. In the current investigation, when the maximal M-wave wave was recorded bipolarly, no significant changes existed between the rest and contraction trials due to the large variability between subjects (with exception of the soleus 2 cm electrodes, when area analysis was used). In contrast to this, when the data were recorded by monopolar electrodes (with a reference electrode placed on the ipsilateral heal), the maximal M-wave magnitude was significantly smaller during voluntary contraction than during rest in both muscles (note that the SD s of the monopolar recordings are smaller which illustrates consistent and reliable results)

103 Signal cancellation in maximal M-wave The same, highly significant reduction is seen in both muscles, when either analysis methods were used. The reduction in the maximal M-wave magnitude during contraction may be due to a number of factors. These include: movement of the nerves in relation to the stimulating electrodes during muscle contraction (Capaday, 1997); movement of the muscle bulk under the skin in respect to the recording electrodes (Simonsen and Dyhre-Poulsen, 1999); refractory periods in the motor nerve following an action potential (see section below); or a decrease in the simultaneous arrival of the motor evoked action potentials to the muscle (Crone et al., 1999); caused by either the recruitment tendencies of the motor nerve, or changes in the conduction velocity of motor axons (Rutten et al., 1998; Keenan et al., 2004). Monopolar SEMG recordings are known to be more effective at measuring reflex activity during contraction, as they reduce the effect that changes in the muscle geometry has on the recordings (Gerilovsky et al., 1989). However, we conclude that it is most likely the changes in muscle geometry to recording electrode relationship that may occur during muscle contraction (Capaday, 1997), that is responsible for the changes seen in the maximal M-wave magnitude between test conditions Area vs PTP amplitude analysis: Dispersion of action potentials In the current investigation it was hypothesised that when significant dispersion of the signal occurs, analysis methods that rely upon simultaneous arrival of action potentials to recording electrodes, such as PTP amplitude measurements, may introduce errors. This is based on findings from Fuglevand and colleagues (1993) who demonstrated that during fatiguing contractions, the PTP amplitude of the 3-102

104 Signal cancellation in maximal M-wave maximal M-wave decreases, the duration increases, and subsequently the area remains unchanged (Fuglevand et al., 1993). In the current investigation however, the maximal M-wave magnitudes obtained from the two analysis methods were significantly correlated, and no great differences in the results were seen when using either of the analysis types. This may suggest, that unlike in fatiguing contractions, where impairment of the neuromuscular propagation may occur (Fuglevand et al., 1993), no significant changes in the neuromuscular propagation occurs during the different test muscle conditions used in the current investigation Methodological considerations Plantar flexion The ankle was placed in maximal plantar flexion during the voluntary contraction condition. This was done to ensure maximum geometry changes in the muscle between test conditions. Although it is well understood that this position can significantly alter spinal reflex excitability (Petersen et al., 1998; Morita et al., 2001; Pierrot-Deseilligny 1997), we do not believe that this should significantly effect our investigation since we are studying the maximal M-wave which is induced in the motor nerve, and should therefore not be affected by changes in spinal reflex excitability. The voluntary contraction used in the current investigation may also have affected the possible maximal M-wave magnitude due to the motor nerves absolute refractory period, which occurs directly following the action potential. During this period it is very difficult to elicit a response in the affected nerve. In human nerves, the average absolute refractory period lasts for approximately 4-5 ms (Capaday, 1997). Given that the average firing frequency of units in the soleus is between 7-15 Hz (Capaday, 1997; Kuchinad et al., 2004), when the soleus is firing at its maximum frequency, a 3-103

105 Signal cancellation in maximal M-wave potential exists for an electrical stimulus to be unsuccessful in eliciting a response in the motor nerve for between ms every second (Capaday, 1997) Electrode type and configuration It is well understood that electrode size, and type, will affect the recorded signal. This is because the surface signal is determined by the contributions of many active motor units and their intermingled interference pattern, and because the recording electrode records the average potential under its area (Stegman et al., 2000). For this reason, the same electrodes were used in the bipolar and monopolar recording in the current investigation. Similarly, Beck et al. (2004) suggest that electrodes must run parallel to the underlying fibres, and although this is good practice it is difficult in complex muscles like the gastrocnemius Spatial filtering It is well understood that the duration and amplitude of a motor unit potential recorded from the skin above the test muscle will be a result of the spatial filtering that occurs in both the interstitial fluid and the tissue between the muscle fibres and the recording electrodes (Stegeman et al., 2000). For this reason the depth of the source of the motor unit potential will effect the spatial filtering of that signal (Beck et al., 2004). Although this is true, the effects of spatial filtering do not need special consideration in the current investigation as the recorded response is the maximal muscle response for the test muscle, and the factors effecting spatial filtering are not changed throughout the experimental duration End plate potentials and tendon potentials When recording monopolarly, a motor unit potential can be triphasic, which is a result of the initiation potential, the motor unit potential, and the non propagating potential that exists due to electro-conductance of the motor unit potential when it reaches the 3-104

106 Signal cancellation in maximal M-wave muscle fibre - tendon junction (Roeleveld and Stegeman, 2002; Farina et al. 2003; Beck et al., 2004; Schulte et al., 2005). This obviously complicates our argument, given that summation of many triphasic motor unit potentials, which are slightly out of phase, will also result in some signal cancellation. This type of signal cancellation is inevitable during surface recording, and will result in the amplitude of the maximal M- wave being smaller than the predicted summation of all the individual motor unit potentials (Moore 1967). This becomes particularly important when the motor units are being voluntarily activated, because of their un-synchronised firing nature (Schulte et al., 2005), however we believe that it is not particularly relevant to the current investigation. This is primarily because, in this investigation, the motor units are being activated similarly in each experimental condition due to the supra maximal electrical stimulus, and that therefore the same degree of signal cancellation due to the triphasic shaped potential can be expected in the different experimental conditions. In the current investigation we are reporting the large degree of signal cancellation that occurs due to the bipolar nature of the recording electrodes, rather than that which has been reported previously Common mode rejection and cross talk The current investigation clearly shows that bipolar recording can result in signal cancellation and elimination of up to 85% of the monopolar recorded maximal M- wave, and that this signal cancellation decreases as the distance between electrodes increase. It also shows that greater cancellation occurs in the gastrocnemius than in the soleus, which we argue directly relates to greater fibre type variability, and therefore greater dispersion of the action potentials that underlay the maximal M- wave in the gastrocnemius. It is true however, that monopolar recording does not eliminate noise from the system as well as bipolar recordings, and that this may be a reason for the larger maximal M

107 Signal cancellation in maximal M-wave wave seen in our monopolar data. To help overcome the effect of the noise signals which are created external to the body, a very low impedance, grounding, lip clip electrode was used (Türker et al., 1988). It is also important to recognise that a reference electrode positioned on the ipsilateral heal was used during monopolar recording (approximately cm distal to the active electrodes). This reference electrode will allow some common mode rejection of noise signals, but is not positioned over the test muscle, and therefore, cannot cause the signal cancellation error as discussed in this article. It is also true that the reference electrode used in our monopolar recording will not be as effective as the bipolar recording in removing cross talk from the recorded signal. Cross-talk occurs when electrodes record EMG activity from muscles other than those being studied (De Luca and Merletti, 1988; Türker, 1993). Given that large (1.5 x maximal M-wave threshold) stimuli were used to elicit the maximal M-wave in the current investigation, it is possible that electrical activity from other muscles in the posterior lower leg was elicited, and that therefore some cross talk is likely. Given the size and proximity of the soleus to other lower leg muscles, it may be reasonable to expect greater cross talk in this muscle from those not recorded in this investigation, when compared to the gastrocnemius. As discussed previously however, it is also possible that some cross talk exists particularly in the monopolar recorded traces from the soleus to the gastrocnemius. It is also important to stress however that signals from adjacent muscles are subject to greater spatial filtering, which results in significant loss of signal amplitude, and that the other muscles of the posterior lower limb, that may be simultaneously activated by the test stimulus, are smaller than those we are recording from. Given both these considerations, it is fair to suggest that cross talk from these muscles cannot account for the significant increase in signal magnitude (PTP amplitude and area) seen during monopolar recording

108 Signal cancellation in maximal M-wave Furthermore, we argue that the amount of cross talk originating from proximal muscles following a supra maximal stimuli will be similar in the PF90 and rest conditions in both of the test muscles. So although we cannot eliminate the effect of cross talk on our results, we remain confident that the changes seen in the maximal M-wave magnitude between test conditions is real, and that this real effect is eliminated, particularly due to signal cancellation, in the bipolar recorded data Innervation zones A further methodological consideration when recording muscle activity is the position of the test muscles innervation zones. This zone corresponds with the entrance point of the motor nerves into the muscle. From this region, action potentials propagate symmetrically in both directions, to the end of each muscle fibre within a given motor unit. It is commonly thought that the innervation zone of most skeletal muscles lie perpendicular to the direction of the fibres, in the mid belly of the muscle. Although this is seen in the soleus, recent publications have shown that the gastrocnemius innervation zones are more scattered and distributed in quite complex configurations (Rainoldi et al., 2004; Parratte et al., 2002: Figure 2: page 1-42). This may become particularly problematic when recording with bipolar electrodes, as it is possible that the oppositely poled electrodes are placed on either side of at least some of the muscles motor points, which will result in a reduction in the recorded potentials (Stegman et al., 2000; Beck et al., 2004). To help reduce the effect of this complication, all electrodes were placed inferior to the innervation zones (illustrated by the authors mentioned above) in the soleus. This is more difficult however in the gastrocnemius, where the innervation zones are more diverse. Given that the results obtained from both muscles show the same pattern of significant loss of signal with a smaller inter-electrode distance, this should at least in part reduce the suspicion the gastrocnemius complex innervation zones affected the results

109 Signal cancellation in maximal M-wave Conclusions and significance On average 59-85% of the maximal M-wave magnitude was lost due to signal cancellation in the bipolar recordings. This may be the primary reason for variability of the maximal M-wave magnitude between subjects as demonstrated in previous studies that use bipolar recording electrodes (Chapter 2; Nagata and Christianson, 1995; Crone et al., 1999; Simonsen and Dyhre-Poulsen, 1999; Tucker and Türker, 2004). We believe that it is therefore important to use monopolar electrodes during SEMG recorded reflex or evoked potential investigations. When monopolar recording electrodes were used (with an indifferent reference electrode placed on the ipsilateral heal), significant and consistent differences in maximal M-wave magnitude were noted between the rest and contraction trials. This is most likely due to changes in the relationship between the muscle bulk and recording electrodes, but may also be due to changes in the relationship between the mixed nerve and stimulating electrodes, and / or the effect of the refractory periods in the motor nerve fibres. The variability questions the validity of using M-wave magnitude as an indicator of the nerve to stimulating electrode stability during different experimental conditions, and using a maximal M-wave magnitude as the normalisation factor in reflex studies. This work suggests that when using bipolar recording electrodes signal cancellation can reduce the recorded output of a muscle by as much as 85%, which can significantly reduce the reliability of the data. It shows that the maximal M-wave magnitude can vary considerably during an investigation depending on muscle condition, and highlights the importance of using recording electrode arrangement that do not eliminate data through signal cancellation

110 Stretch and voluntary contraction changes the maximal M-wave 4. Triceps surae stretch and voluntary contraction alters maximal M-wave magnitude This chapter is an edited version of the manuscript Triceps surae stretch and voluntary contraction alters maximal M-wave magnitude by K.J Tucker and K.S. Türker, which has been published in the Journal of Electromyography and Kinesiology (2005) and included in the appendix of this thesis Outline of study Reliability of the motor response (M-wave) is fundamental in many reflex studies; however it has recently been shown to change during some investigations. The aim of this investigation was to determine if triceps surae stretch and voluntary contraction, or recording and analysis techniques, affect the maximal M-wave magnitude. The maximal M-wave was investigated in human gastrocnemius and soleus during different foot positions and during triceps surae contraction. Both bipolar and monopolar recoding methods, and area and peak-to-peak (PTP) amplitude analysis methods were used. Results: Maximal M-wave magnitude changed significantly between test muscle conditions, and is largest during dorsiflexion, probably due to changes in muscle bulk and recording electrode relationship. The maximal M-wave was up to 88% smaller when recorded by bipolar electrodes compared to monopolar electrodes, which is discussed in relation to signal cancellation. Area analysis provided more significant differences in M-wave magnitude between test muscle conditions than did PTP amplitude analysis, and the 4-109

111 Stretch and voluntary contraction changes the maximal M-wave maximal M-wave shape changed significantly between test muscle conditions. This study suggests that maximal M-wave magnitude can vary depending on muscle condition, it highlights the importance of using correct recording and analysis techniques, and questions the reliability of using M-wave magnitude to monitor stimulating electrodes to nerve relationship Introduction The motor response (M-wave) is evoked by direct electrical stimulation of motor efferents. For many years it has been thought that the maximal M-wave, which is obtained by increasing stimulation intensity until no further increases in the M-wave magnitude is observed, is a steady and accurate measurement of total muscle activation. It has therefore been used as the normalisation factor in many reflex studies (for review see Pierrot-Deseilligny and Mazevet, 2000; Zehr, 2002). The stability of a chosen M-wave magnitude is also used as an indicator of the nerve to stimulating electrode relationship during muscle contraction or limb movement investigations (for review see Zehr, 2002). While some studies have shown that the maximal M-wave magnitude reduces over an experimental session (Crone et al., 1999), others do not (Chapter 2; Tucker and Türker, 2005). In individual subjects the maximal M-wave magnitude is also known to be altered during the walking or running cycle (Simonsen and Dyhre-Poulsen, 1999) and to be affected by the level of voluntary contraction in the test muscle (Nagata and Christianson, 1995). However, the variability in maximal M-wave magnitude is not necessarily consistent between subjects (Chapters 2 and 3; Simonsen and Dyhre-Poulsen, 1999; Tucker and Türker, 2004, 2005), which ensures that when data are pooled for analysis purposes, little or no significant differences between conditions are reported

112 Stretch and voluntary contraction changes the maximal M-wave The variability noted between subjects may be due to a number of factors. For example, it is known that surface recording electrodes can pick up electrical activity from muscles other than those being studied (De Luca and Merletti, 1988; Türker, 1993), that the bulk of the test muscle may move in relation to the recording electrodes (Herbert et al., 2002), and that the mixed nerve may move in relation to the stimulating electrodes (Capaday, 1997; Simonsen and Dyhre-Poulsen, 1999; Fumoto et al., 2002). Each of these factors may vary between subjects depending on the placement of the stimulating and recording electrodes, and the size of the muscle being recorded. Other factors such as dispersion of the signal, and signal cancellation may also affect the magnitude and the variability of the recorded M-wave (Chapter 3; Tucker and Türker, 2005). The duration of the maximal M-wave is determined by the dispersion and spatial filtering of action potentials that underlie the signal. Dispersion of these action potentials will occur due to recruitment tendencies of the motor nerve, conduction velocity of motor axons (Rutten et al., 1998) and muscle fibres (Buchthal and Schmalbruch, 1970)for review see (Burke, 1981), impairment of neuromuscular propagation (Fuglevand et al., 1993), and summation of the underlying potentials (Day and Hulliger 2001; Roeleveld and Stegeman, 2002; Farina et al., 2003; Beck et al., 2004). Low-pass spatial filtering of the signal also occurs as the signal passes through tissue and interstitial fluid from the intramuscular fibres to the surface recording electrodes (Stegeman et al., 2000). Dispersion of the action potentials underlying the M-wave will therefore affect both the shape and duration of the response (Gerilovsky et al., 1985; Fuglevand et al., 1993). This becomes particularly important when recording with bipolar electrodes. While bipolar recording removes noise from the recorded trace, if the beginning of the maximal M-wave is being recorded by the second electrode before the wave has 4-111

113 Stretch and voluntary contraction changes the maximal M-wave completely passed by the first electrode, it is plausible that some real signal, originating from the test muscle, may also be cancelled (Chapter 3; Moore, 1967; De Luca 1978; Day and Hulliger, 2001; Keenen et al., 2005, Tucker and Türker, 2005). The amount of signal cancellation that will occur is therefore related to the distance between the oppositely poled surface electrodes (Chapter 3; Roeleveld et al., 1997; Tucker and Türker, 2005), and the duration of the recorded signal (Farina et al., 2003). In the current investigation both bipolar electrodes with an inter-electrode distance of 2 cm, and bipolar recording electrodes where the active electrode was placed over the muscle belly and the indifferent reference electrode place on the ipsilateral heal (reported from here as monopolar: see Chapter 3; Roeleveld and Stegeman 2002; Tucker and Türker, 2005) were used for recording the maximal M-wave. For the purpose of this study, the maximal M-wave was elicited in both the medial gastrocnemius and soleus muscles. These muscles are ideal for such an investigation as a maximal M-wave can be elicited in both muscles simultaneously. Furthermore, although the two muscles are located in close proximity to each other, they are quite different in structure and function, which enables the results of the study to have greater significance and reliability. Both the soleus and gastrocnemius are involved with plantar flexion; are stretched during dorsi flexion; and when the level of voluntary contraction feedback is given from the gastrocnemius, subjects are able to hold equivalent contraction levels in the soleus (up to approximately 50% MVC: Chapter 2; Tucker and Türker, 2004). The first aim of this study was to determine if the magnitude of the maximal M-wave would change in the gastrocnemius and soleus during different levels of triceps surae stretch and voluntary contraction. It was hypothesised that the test muscle condition would alter the magnitude of the maximal M-wave in both muscles

114 Stretch and voluntary contraction changes the maximal M-wave The second aim of this investigation was to determine if bipolar and monopolar recording produce similar maximal M-wave data. It was hypothesised that a large proportion of the maximal M-wave would be eliminated by bipolar recording. It was further hypothesised that therefore the changes in maximal M-wave during different muscle conditions would only be seen in the monopolar-recorded data. The final aim of the study was to determine if the shape of the M-wave changed during different muscle conditions. This was investigated by using both area and PTP amplitude analysis methods. It was hypothesised that that the shape of the maximal M-wave would change between muscle conditions, and that therefore area analysis would provide more reliable results Materials and Methods Eleven healthy subjects (6 female, 5 male) between the ages of gave written, informed consent to participate in this study. The experiments were approved by the Human Ethics Committee of the University of Adelaide and all procedures used, conformed to the Declaration of Helsinki Protocol The subjects lay prone, on a standard physiotherapy table, with a face hole that allowed visual feedback of background muscle activity, from a computer monitor positioned under the table. The subject s right ankle was positioned in 4 different muscle conditions, these being: at rest with no support (R); at plantar flexion with 40% maximum voluntary contraction (MVC) of medial gastrocnemius muscle (PF40); at plantar flexion rest (PFR; ankle in plantar flexion with a foot bar supporting the same angle as produced during PF40); and at maximum dorsiflexion (DF) (Figure 14, pg 4-115). The order of muscle conditions was randomised for each subject. Ankle 4-113

115 Stretch and voluntary contraction changes the maximal M-wave angles were measured from the head of the fibula to the distal point of the fifth metatarsal, via the lateral head of the maleolus (mean ± SD of ankle angles are shown in Fig 1). A foot bar provided the resistance required to produce the voluntary contraction, and was used as a support structure to hold the ankle during PFR Recording Surface electromyography (SEMG) recording electrodes (Duo-Trode silver/silver chloride) were placed 2 cm apart on the belly of the right medial gastrocnemius and soleus muscles, while for monopolar recording, the reference electrode was placed on the Renshaw of the ipsilateral leg. The term monopolar recordings is used only for convenience since they were actually bipolar recordings in which the reference electrode was located approximately 20-30cm distal to the active electrode (Chapter 3; Gerilovsky et al., 1985, 1989; Pierrot-Deseilligny and Mazevet, 2000; Tucker and Türker 2005). The monopolar and bipolar surface EMG recordings were amplified in a custom made EMG amplifier with built in stimulus artefact suppressor and high pass (20Hz) filter, and were recorded on a computer (LabVIEW : National Instruments) at a sample rate of 5 khz. The skin resistance of all the bipolar electrodes was below 10kΩ; and subjects were grounded by lip clip electrodes (Türker et al., 1988)

116 Stretch and voluntary contraction changes the maximal M-wave Figure 14: Ankle positions during test conditions. The subject s right ankle was angled at rest (R; ± 17 0 ), plantar flexion rest (PFR; ± 8 0 ), plantar-flexion with 40% maximum voluntary contraction (MVC) of the medial gastrocnemius (PF40; ± 8 0 ) and at maximum dorsi-flexion (DF; 98 0 ± 8 0 ). Ankle angles were measured from the head of the fibula to the lateral head of the maleolus to the distal point of the fifth metatarsal. A foot bar provided the resistance required to produce voluntary contraction in PF40, and was used as a support structure to rest the foot in PFR

117 Stretch and voluntary contraction changes the maximal M-wave Feedback of voluntary contraction level The muscle activity required for feedback during the voluntary contraction trials was recorded from bipolar electrodes on the gastrocnemius. MVC was obtained over a series of 3 trials, whereby the subject was asked to contract the calf muscles as hard as he/she could, for a period of 5 seconds. The SEMG activity was band-pass filtered (20-500Hz), full wave rectified and low passed filtered (1Hz), and the maximum SEMG level over that period was defined as 100% MVC (Scutter and Türker, 1998). A preset target line of 40% MVC was displayed on the feedback monitor during the contraction condition Stimulus The motor response (M-wave) was elicited via electrical stimulation of the tibial nerve of the ipsilateral leg. A large (~ 100 cm 2 ) copper anode was placed just above the patella, while the silver ball cathode, with a 1 cm diameter, was placed in the popliteal fossa, at the point where the weakest stimulus intensity was required to elicit a Hoffmann (H) reflex (Miles et al., 1989). Square pulse stimuli of 0.5 ms duration were triggered from a computer and delivered by a Digitimer constant current stimulator (model DS7A). The stimulus intensity employed to obtain the maximal M-wave was determined by increasing the stimulus intensity gradually, and assessing with online feedback, if the area and peak-to-peak (PTP) amplitude of the M-wave also increased. The maximal M-wave threshold was determined for each subject in each of the SEMG channels, in both the soleus and gastrocnemius during each of the test muscle conditions. The largest stimulus intensity required to reach the maximal M-wave from each of these channels was recorded as the maximal M- wave threshold. The stimulus intensity employed throughout the experiment was 1.5x 4-116

118 Stretch and voluntary contraction changes the maximal M-wave the maximal M-wave threshold (for review see Chapter 1; Pierrot-Deseilligny and Mazevet, 2000; Tucker et al., 2005). Twenty stimuli were given during each of the five experimental conditions at random intervals between 3 and 5 seconds apart. During the contraction trial, the stimuli were given immediately after the subject reached 40% MVC. Recovery periods of one minute were provided after each 4 stimuli. At least 2 minutes rest was provided between each experimental condition Analysis Data were band passed filtered offline at Hz. The 20 SEMG traces recorded for each condition were then averaged for each subject. PTP amplitude was determined by the difference between the maximum and minimum values of the averaged responses, while area was calculated by the cumulative sum (CUSUM) of the rectified trace (Ellaway, 1978; Brinkworth and Türker, 2003). The results were normalised to the magnitude of the maximal M-wave obtained by each subject during the different muscle conditions (separately for monopolar and bipolar recording) Statistics The average (20) maximal M-wave magnitudes (area and PTP amplitude) were calculated for each subject. Then a one-way ANOVA was used on the pooled data (N=11) to determine if a statistically significant difference existed between the experimental muscle conditions. To determine if the maximal M-wave shape changed between muscle conditions, a one-way ANOVA was also performed on the mean shape ratio (normalised area/normalised PTP amplitude) of the averaged (20) traces. In each case, the Bonferroni post hoc test was used to determine where any significant differences lay. A paired T-test was used to determine if a statistically 4-117

119 Stretch and voluntary contraction changes the maximal M-wave significant difference existed between the magnitude of the maximal M-wave when recorded by bipolar and monopolar electrodes. In all cases the level of significance was set at p< Results Maximal M-wave magnitude during different muscle conditions The maximal M-wave magnitudes (means ± standard deviations) from all 11 subjects recorded from both the gastrocnemius and soleus muscles, are shown in Figure 15, pg In the grouped bipolar data, the only significant difference in maximal M- wave magnitude between ankle conditions existed in the soleus, when analysed by area. In this case, DF (0.94±0.18) was shown to be significantly larger than PFR (0.67±0.19) and PF40 (0.69±0.25; both p<0.05). When recorded by monopolar electrodes and analysed by area, the maximal M-wave was significantly larger (p<0.01) in DF than all other conditions in both the soleus ((DF (1.0±0.0); R (0.63±0.16); PRF (0.58±0.10); PF40 (0.57±0.13)) and gastrocnemius ((DF (1.0±0.02); R (0.74±0.14); PRF (0.57±0.12); PF40 (0.52±0.20)). The gastrocnemius maximal M-wave area was also significantly larger (p<0.01) in R than in PF40. The data are similar when analysed by PTP amplitude (see significance on Figure 15, pg 4-120) Maximal M-wave magnitude: bipolar vs monopolar recording Figure 3 illustrates the averaged (non-normalised) data from the 11 subjects when analysed by both area (CUSUM) and PTP amplitude. In both muscles and both analysis techniques, the magnitude of the monopolar maximal M-wave was significantly larger (p<0.01) than the maximum bipolar recorded response (area; averaged CUSUM (k.ms)): gastrocnemius monopolar (121.5±61.1); bipolar (14.0±7.4); soleus monopolar (84.0±53.6); bipolar (18.3±13.0)). The maximal M-wave 4-118

120 Stretch and voluntary contraction changes the maximal M-wave magnitude was up to 88% smaller when recorded by bipolar electrodes compared to monopolar electrodes Shape of maximal M-wave The shape of the maximal M-wave changed significantly between different muscle conditions. Figure 17 (pg 4-122) shows the averaged data from 20 supra-maximal stimuli in one subject, recorded from both bipolar and monopolar electrodes, for each of the muscle conditions in the gastrocnemius. This figure shows that the R and PFR condition produce similar shaped responses, while the shape of the maximal M-wave in both PF40 and dorsiflexion DF were distinctly different from all other conditions. This qualitative observation is supported by the means ratio (normalised area/normalised PTP amplitude) of the grouped data. When the maximal M-wave was recorded by monopolar electrodes, the means ratio during DF was significantly smaller than in most other conditions (gastrocnemius: (p<0.01); DF (0.99±0.02), PFR (1.30±0.25), PF40 (1.40±0.30), soleus: (p<0.01); DF (1.0±0.0), R (1.36±0.26), PFR (1.41±0.24), PF40 (p<0.05; 1.28±0.25)). In the bipolar recording, no significant differences between the means ratio were seen in the gastrocnemius, however in the soleus DF (0.84±0.19) was again significantly smaller than PFR (p<0.05: 1.27±0.42), and PF40 (p<0.01: 1.32±0.41)

121 Stretch and voluntary contraction changes the maximal M-wave Figure 15: Mean and SD, of normalised maximal M-wave. Data normalised to the overall maximal M-wave in any test muscle condition for each subject. Pooled data from all subjects are shown for both muscles. The monopolarrecorded data did produce significant differences between muscle conditions, particularly when comparing DF to other muscle conditions (*p<0.05; ** p<0.01)

122 Stretch and voluntary contraction changes the maximal M-wave Figure 16: Maximal M-wave magnitude during monopolar vs bipolar recording. In both muscles the maximal M-wave magnitude was significantly larger during monopolar recording compared to bipolar recording when analysed with either area or PTP amplitude (** p<0.01). The smaller maximal M-wave magnitude recorded from bipolar electrodes compared to monopolar electrodes, suggests signal cancellation exists in bipolar recording. The signal cancellation that occurred in the bipolar recording was larger in the gastrocnemius (area: 88% and PTP amplitude: 78%) than the soleus (area: 74% and PTP amplitude: 57%). The greater signal cancellation that occurs in the gastrocnemius supports the suggestion that greater dispersion of the M-wave occurs in this muscle (see also Chapter 3; Tucker and Türker, 2005)

123 Stretch and voluntary contraction changes the maximal M-wave Figure 17: Maximal M-wave: bipolar vs monopolar recording Averaged (n=20) SEMG traces of the maximal M-wave recorded from the gastrocnemius, for one subject during the different test muscle conditions, by both the bipolar and monopolar electrodes. Note that the shape of maximal M-wave is similar in R and PRF. While the DF maximal M-wave shape, and the PF40 shape is quite different compared to all other conditions. This qualitative observation of change in shape is supported by the means ratio (normalised area/normalised PTP amplitude) statistics of the grouped data as described in the text

124 Stretch and voluntary contraction changes the maximal M-wave 4.4. Discussion There were three main findings in the current investigation. Firstly, the maximal M- wave magnitude was significantly different between muscle conditions, which may be a result of changes in the relationship between the muscle bulk and recording electrodes. Secondly, greater changes in maximal M-wave magnitude were noted between muscle conditions when monopolar recording electrodes were used compared to bipolar electrodes. This further supports our previous findings that signal cancellation can affect the results obtained from a maximal M-wave investigation (Chapter 3; Tucker and Türker, 2005). Finally, some variability in the results was noted when comparing the use of area and PTP analysis methods. This variability coincided with shape changes in the maximal M-wave during different conditions, and supports the use of area analysis in future maximal M-wave studies as it takes into account the duration and shape, and not just the peak to peak amplitude of the maximal M-wave The M-wave and muscle condition: monopolar vs bipolar recording Simonsen and Dyhre-Poulsen (1999) showed that maximal M-wave magnitude does change consistently for each subject during a gait cycle, however their study also demonstrated that the variability was not consistent between subjects. We have also previously demonstrated that large and inconsistent variability exists in the magnitude of the maximal M-wave during increased levels of muscle contraction when bipolar recording electrodes are used (Chapter 2; Tucker and Türker, 2004). In the current investigation, and as noted in previous studies, when recorded by bipolar electrodes this variability resulted in limited significant differences in the pooled data

125 Stretch and voluntary contraction changes the maximal M-wave It is possible that the inconsistent variability observed in the maximal M-wave magnitude during bipolar recording may be due to different degrees of signal cancellation in different subjects. Signal cancellation occurs as a direct result of temporal summation of the same signal, when the beginning of the signal is being recorded by the second (negative) electrode, before the signal has completely passed the first (positive) electrode (see full description in Chapter 3; Tucker and Türker, 2005). The amount of signal cancellation in the current investigation is therefore reliant on the dispersion of the maximal M-wave, and the distance between the recording electrodes. Variability in the level of dispersion of this signal can be a result of changes in the time that it takes for the original stimulus (0.5 ms duration) to elicit a response in a variety of motor axons; differences in the conduction velocity of action potentials along the length of motor axons; the time taken for action potentials to cross the neuromuscular junction; the differences in the conduction velocity of a variety of different muscle fibre types (Roeleveld and Stegeman, 2002); and the filtering of the motor unit potential through the interstitial fluid (decreasing its frequency component, and increasing the duration of the potential) to the recording electrodes (reviewed in Chapter 1; Stegeman et al., 2000; Tucker et al., 2005). Given that the reference electrode is not placed on the test muscle for monopolar recording, the variability that may exist between subjects in bipolar recording due to signal cancellation, will be avoided in monopolar recording. It is important to note here that other forms of signal (or phase) cancellation do exist within the maximal M-wave. This is primarily a result of the summation of many single motor unit potentials, that are by nature triphasic (including the initiation potential, the motor unit potential, and the non propagating potential that exists due to electroconductance of the motor unit potential when it reaches the muscle fibre (Roeleveld 4-124

126 Stretch and voluntary contraction changes the maximal M-wave and Stegeman, 2002; Farina et al., 2003; Beck et al., 2004). This form of phase cancellation will exist in both bipolar and monopolar recorded traces, and therefore will not affect the results of the current investigation. The current investigation clearly shows greater significant differences between test muscles conditions when monopolar recording was used compared to bipolar recording. We therefore argue in both this paper and previous work (Chapter 3; Tucker and Türker, 2005) that given up to 88% of the signal is lost due to signal cancellation in bipolar data, the bipolar data cannot be considered as reliable as the monopolar recorded data in the current investigation. In the monopolar recordings, a clear and significant difference in maximal M-wave magnitude existed between the different test muscle conditions (Fig 15, pg 4-120). A similar pattern of difference existed in both muscles when either analysis methods were used. The maximal M-wave magnitude obtained during DF, when the muscles were actively stretched, was significantly larger than during most other conditions, and decreased in the order of R, PFR, PF40. Although it is true that no significant difference in the maximal M-wave magnitude was seen in the current investigation between the PFR and PF40 conditions, previous work in the area has shown significant changes in maximal M-wave magnitude between PFR and PF90 (during a 90% MVC; Chapter 3; Tucker and Türker, 2005). We conclude that the changes that occur in the muscle bulk to recording electrode position will be greater in the 90% MVC contraction compared to the 40% MVC, and that this may result in the lack of significant difference seen in the current investigation between PFR and PF40. We further suggest that it is possible that the absolute refractory periods, which occur directly following an action potential may have influenced the previous PF90 data (Chapter 3; Tucker and Türker 2005). Given fewer absolute refractory periods will occur in the motor nerves of a muscle contracting at 40% MVC compared to 90% 4-125

127 Stretch and voluntary contraction changes the maximal M-wave MVC, there is less chance that this phenomenon will significantly reduce the magnitude of the maximal M-wave in the less contracted muscle. Although monopolar SEMG recordings are known to be more effective at measuring reflex activity during contraction as they reduce the effect that changes in the muscle geometry has on the recordings (Gerilovsky et al., 1989), we conclude that changes in muscle geometry are responsible for the changes seen in the maximal M-wave magnitude in the current investigation during different test muscle conditions. It is also possible that changes in the mixed nerve to stimulating electrode relationship can influence the maximal M-wave magnitude (Capaday, 1997), however given that the stimulus strength was 1.5x maximal M-wave threshold, it is unlikely that a change in this relationship would result in sub maximal stimuli being delivered to the mixed nerve in the current investigation. Given the current findings, and the variability seen in the maximal M-wave magnitude shown in our previous work (Chapters 2 and 3; Tucker and Türker 2004, 2005), we strongly recommend that the maximal M-wave magnitude is recorded with monopolar electrodes, and in all test conditions, if it is to be used as a normalisation factor in the investigation Changes in muscle geometry during muscle stretch and contraction There are many factors that will contribute to changes in the surface recorded EMG signal as the length and contraction strength of a test muscle is altered in an investigation. Firstly, it is well understood that during movement, muscles slide relative to the skin because of variability in muscle length. Subsequently changes in the recording 4-126

128 Stretch and voluntary contraction changes the maximal M-wave electrode to the test muscles active muscle fibres, are expected throughout an investigation (discussed in Schulte et al., 2005). The movement of the electrodes relative to underlying muscle fibres becomes particularly important in the current investigation given the complex innervation zones of the human gastrocnemius and soleus muscles (reviewed in Chapter 1; Tucker et al., 2005), given that large differences in SEMG amplitude are noted when electrodes are placed in different locations relative to the innervation zones of muscles (Rainoldi et al., 2000 and 2004; Saitou et al., 1990, 2000). Furthermore, it is similarly understood that bipolar electrodes should run parallel to the underlying fibres (Beck et al., 2004), however this may again be difficult given the complex structure of the gastrocnemius muscle fibres (reviewed in Chapter 1; Tucker et al., 2005). Secondly, changes in muscle contraction can affect the surface recorded signal given that high levels of contraction can effect the thickness of the subcutaneous layers of tissue between the muscle fibres and recording electrodes. Subcutaneous layers are known to act as a spatial filter to the generated signal (Stegman et al., 2000), and can therefore significantly affect the magnitude of this recorded signal (Farina et al., 2002, discussed in Chapter 3; Tucker and Türker 2005). Thirdly, when a motor unit potential is recorded monopolarly, it is often recorded as a triphasic potential. This is a result of the initiation potential, the motor unit potential, and the non propagating potential that exists due to electro-conductance of the motor unit potential when it reaches the muscle fibre - tendon junction (Roeleveld & Stegeman, 2002; Farina et al., 2003; Beck et al., 2004; Schulte et al., 2005, discussed in Tucker & Türker 2005). It is well understood that the summation of many triphasic motor unit potentials, which are slightly out of phase, will result in some signal cancellation. This type of signal cancellation is inevitable during surface recording, and will result in the amplitude of both background muscle activity and the 4-127

129 Stretch and voluntary contraction changes the maximal M-wave maximal M-wave (as well as any other surface recorded muscle response) being smaller than the predicted summation of all the individual motor unit potentials (Moore 1967). Cancellation of motor unit potentials becomes particularly important when motor units are being voluntarily activated, because of their un-synchronised firing nature (Schulte et al., 2005). Further to this, Gydikov and Kosarov (1972) demonstrated that that the onset of the motor unit potential negative phase and that of the positive end plate potential varies with joint angle and therefore the length of muscle fibres. The degree of muscle stretch and contraction therefore may have a direct effect on the amount of cancellation that occurs. It is conceivable therefore that not only will changes in the muscle length alter the muscle fibres that contribute to the recorded signal, but also that the amount of signal cancellation of motor unit potentials, and the amount of spatial filtering that will occur between the production of the individual motor unit potentials and the recording electrodes, can all effect the size of the recorded response when the length of a test muscle is altered Area vs PTP amplitude analysis: Dispersion of action potentials Fuglevand and colleagues (1993) have previously demonstrated that during fatiguing contractions, the PTP amplitude of the maximal M-wave decreases, the duration increases, and subsequently the area remains unchanged. Although in the current investigation fatigue was not an issue, we similarly show that the shape of the maximal M-wave changed significantly depending on the test muscle condition (Fig 17, pg 4-122). The variability in maximal M-wave shape was seen in both muscles, and in both bipolar and monopolar recordings. Our investigation shows small differences between the results when using area or PTP amplitude analysis methods

130 Stretch and voluntary contraction changes the maximal M-wave The change in shape does promote the use of area analysis as it accommodates for the area under the curve, and therefore any dispersion of the signal, rather than PTP amplitude analysis, which relies on synchronous firing of the motor-units underlying the response Methodological considerations. It is true, that monopolar recording does not eliminate noise from the system as well as bipolar recordings. To help overcome the potential for increased noise created external to the body, a very low impedance, grounding, lip clip electrode was used (Türker et al., 1988). It is also important to recognise that a reference electrode positioned on the ipsilateral heal was used during monopolar recording (approximately cm distal to the active electrodes). This reference electrode will allow some common mode rejection of noise signals, but is not positioned over the test muscle, and therefore, cannot cause the signal cancellation error (discussed in detail in Chapter 3; Tucker and Türker, 2005). Given the reduction in common mode rejection that occurs in monopolar (with reference electrode on heal) compared to bipolar recording, it is possible that increased cross talk from proximal muscles contribute to the maximal M-wave magnitude seen in the monopolar recording (De Luca and Merletti, 1988; Türker, 1993). Given the size and proximity of the soleus to other lower leg muscles, it may be reasonable to expect greater cross talk in this muscle from those not recorded in this investigation when compared to the gastrocnemius. However, it is also possible that some cross talk exists particularly in the monopolar recorded traces from the soleus to the gastrocnemius, as the superior portion of the soleus lies directly deep to the medial gastrocnemius. Signals from adjacent muscles are however subject to greater spatial filtering, which results in significant loss of signal amplitude

131 Stretch and voluntary contraction changes the maximal M-wave Furthermore, the muscles of the posterior lower limb that may be simultaneously activated by the test stimulus are smaller than those we are recording from, and a similar amount of cross talk is expected between experiential conditions as the factors that affect cross talk are not altered between conditions. It is therefore fair to suggest that cross talk from these muscles cannot account for the significant increase in signal magnitude (PTP amplitude and area) seen during monopolar recording compared to the bipolar recording, and that cross talk is not primarily responsible for the changes in maximal M-wave noted between the experimental conditions (as discussed in Chapter 3; Tucker and Türker, 2005) Conclusions and significance. This investigation demonstrates that between 65-88% of the maximal M-wave magnitude is lost due to signal cancellation in bipolar compared to the monopolar recordings. This may be the primary reason for variability of the maximal M-wave magnitude noted between subjects in previous studies, using bipolar recording electrodes (Chapter 2; Nagata and Christianson, 1995; Crone et al., 1999; Simonsen and Dyhre-Poulsen, 1999; Tucker and Türker, 2004). We believe that it is therefore important to use monopolar electrodes during SEMG recorded reflex or evoked potential investigations. When monopolar recording electrodes were used, significant and consistent differences in maximal M-wave shape and magnitude were noted between test conditions. This is most likely due to changes in the relationship between the muscle bulk and recording electrode during different levels of muscle stretch and contraction. The significant change in shape of the maximal M-wave during different ankle conditions does support the suggestion that area analysis methods will provide more accurate results than PTP amplitude. It is therefore recommended that monopolar recording electrodes, and area analysis techniques 4-130

132 Stretch and voluntary contraction changes the maximal M-wave are used in human reflex studies, particularly when the test muscle conditions will change. This work shows that the maximal M-wave magnitude can vary considerably during an investigation depending on muscle condition, and highlights the importance of determining the most accurate means of recording and analysing M-wave data for the muscle being tested

133 Pain and the H-reflex 5. Hypertonic saline induced pain does not affect agonist or antagonist H- reflex strength in the triceps surae This chapter is an edited version of the manuscript Hypertonic saline induced pain does not affect agonist or antagonist H-reflex strength in the triceps surae by K.J Tucker, R.S Brinkworth and K.S. Türker, which has been submitted for publication in Pain Chapter Outline The effect of agonist and antagonist muscle pain on the Hoffmann (H) reflex and background muscle activity was studied in the human soleus and medial gastrocnemius muscles during normal standing. Hypertonic saline was used to induce muscle pain (of approximately 6/10 on a visual analogue scale) in the tibialis anterior and lateral gastrocnemius during separate trials. Control (no injection), and sham (isotonic saline injection) trials were also conducted for comparison to the pain trial. Both bipolar and monopolar (one active and one indifferent reference electrode) recoding methods, as well as area and peak-to-peak (PTP) amplitude analysis methods were used to construct full H-reflex and M-wave curves. Neither background muscle activity in the medial gastrocnemius, soleus or tibialis anterior, nor the H-reflex size in medial gastrocnemius or soleus, were altered by pain induced in either the lateral gastrocnemius or tibialis anterior. These results support similar work completed in active jaw muscles and relaxed leg muscles

134 Pain and the H-reflex 5.2. Introduction The effect of pain on muscle activity has been debated extensively (Lund et al., 1991); however it is not yet entirely clear how musculoskeletal pain affects motor performance. The two main pain theories, known as the pain adaptation and vicious cycle models, both hypothesise reasons for pain affecting normal movement, although neither completely elucidate where or how the changes occur. The pain adaptation model (Lund et al., 1991) suggests that the changes in muscle activity, and the abnormal movement and posture that occur while an individual is in pain, is a normal protective adaptation. This adaptation is suggested to reduce the output of muscles that move the painful site and increase the output of muscles that brace the site, to reduce the effective movement of the painful area, and is the result of the action of pain afferents on segmental interneurons (Lund et al., 1991). In contrasts the vicious cycle model suggests that pain and dysfunction are reciprocally linked (Travell et al., 1942), i.e. pain will cause a change in motor pattern which may lead to more pain and further changes in motor output. While contradictory as to the reason for the alteration in muscle patterns, and whether motor changes are beneficial or detrimental, both models suggest that pain does alter the level of activity in muscles associated with the painful site. While clinical observation suggests that painful muscles are often tense and associated with increased electromyography (EMG) activity as compared to nonpainful muscles (Svensson et al., 1997a; Sessle, 1999), some studies show that no EMG hyperactivity occurs during muscle pain, particularly when there is little or no voluntary contraction (Graven-Nielsen et al., 1997a; Birch et al., 2000). It has also been shown for example, that chronic lower back pain patients have significantly less para-spinal muscle activity than control groups (Collins et al., 1982) while in other 5-133

135 Pain and the H-reflex studies, no significant difference between resting EMG activity in painful and nonpainful muscles have been detected (Matre et al., 1999; Schulte et al., 2004). Other pain studies have shown that the level of background muscle activity increases temporarily following an injection of a painful substance when compared to the injection of a non-painful substance, but that this hyperactivity does not last as long as the sensation of pain (Svensson et al., 1998b). The variability in results may in part be explained by a difference in the general function of the muscle in pain, and the activity level of the muscle at the time of pain. For example, an increase in muscle tension, which is often reported during pain, is most noticeable when the person is under load or during dynamic postural tasks compared to non-loaded conditions (Ahern et al., 1988; Rossi and Decchi, 1994). Given that changes in muscle tension during pain is understood to be related to the task of the painful muscles, it is plausible that this pain, in some way, affects the efficiency of spindle system, which supplies at least one quarter of the drive to the motoneuron pool (Macefield 1993). Increased sensitivity of the fusimotor system during acute muscle pain has been suggested following changes in the stretch reflex in human jaw (Wang et al., 2001, 2002), and cat hindlimb (Thunberg et al., 2002) muscles. This finding is further supported by the work of Matre and colleagues who have shown that when pain is induced in either the human soleus or tibialis anterior muscles, the stretch reflex in the soleus is facilitated (Matre et al., 1998). In the electrical analogue of the stretch reflex, Le Pera and colleagues (2001) have similarly shown that the H-reflex in the human flexor carpi radialis is inhibited during and following a painful stimulus (induced in the same muscle). Ellrich and Treede (1998) have also shown that a noxious heat stimulus applied to the sole of the foot abolished the H-reflex in the tibialis anterior, but had no effect on the soleus H-reflex (Ellrich and Treede, 1998). This study however was completed on just one subject 5-134

136 Pain and the H-reflex and requires further investigation. Other investigators have shown no significant effect of pain on the H-reflex in the human leg (Matre et al., 1999), or jaw muscles (Svensson et al., 1998a). Given such variable results in both human and animal literature, it is quite clear that further investigation is required to determine the extent of the direct influence of pain on the motoneurone pool and/or the Ia α motor neuron synapse. This research is vital in helping to update the current understanding of how pain may affect muscle activity. The goal of the current investigation was to monitor changes in background muscle activity and H-reflex strength in human lower limb muscles while steady muscle pain was induced in either an agonist or antagonist muscles during normal standing. While the study on background muscle activity will provide evidence for a direct effect of pain afferents on motoneurones, the study on the H-reflex (at a given background muscle activity level) will provide evidence for the direct effect on the Ia synapse. The study was divided into four main aims. The first aim of this study was to determine if pain induced in the lateral gastrocnemius would affect the H-reflex in either the soleus or medial gastrocnemius (agonist muscles). The second aim of this study was to determine if pain induced in the tibialis anterior would affect the H-reflex in either the soleus or medial gastrocnemius (antagonist muscles). In both cases it was hypothesised that pain would not alter the H-reflex size. The third aim was to determine if pain in the muscle of one leg would alter the subjects weight distribution between both legs. It was hypothesised that subjects would decrease the weight distributed to the test leg during the pain trials compared to the non-painful trials. The fourth aim of this study was to investigate changes in background muscle activity when pain was induced in the agonist and antagonist muscles. It was hypothesised that pain would not alter background muscle activity in either muscle during standing

137 Pain and the H-reflex 5.3. Materials and Methods Nine healthy human subjects (3 female, 6 male) between the ages of gave written, informed consent to participate in this study. The experiments were approved by the Human Ethics Committee of the University of Adelaide and all procedures used conformed to the Declaration of Helsinki Protocol The subjects stood with their legs a shoulder length apart, and arms comfortably to their side. The subject stood on a force plate beneath their right (test) foot, and a solid, equal height platform beneath their left foot. Change in subject s weight distribution on their right leg was monitored between the test conditions. Experimental muscle pain was induced by a 5% hypertonic saline injection, while 0.9% isotonic saline was used during the sham condition; both were injected via a constant flow device with an average speed of approximately 12.5 ml/h. The level of perceived muscle pain induced in the subject during the trials was monitored with the use of a subject driven visual analogue scale (VAS), where 0 represented no pain and 10 represented extreme pain. Subjects were instructed to represent their pain levels on the VAS throughout the trials. During pain conditions the investigator increased or decreased the saline injection rate in attempt to maintain a muscle pain level of approximately 5/10. During the sham trials, the investigator aimed to maintain a similar injection rate as was used during the pain trials. A total of 9 experimental trials were tested throughout the investigation. The 4 experimental trials were: tibialis anterior sham (TASH); tibialis anterior pain (TAPN); lateral gastrocnemius sham (LGSH); and lateral gastrocnemius pain (LGPN). Each of these test trials was intermitted by a control trial where no injection was given (C1, C2, C3, C4, C5). The 5-136

138 Pain and the H-reflex total investigation took approximately 2.5 to 3 hours to complete as subjects were given time to rest between each of the trials Electrical Stimulus The H-reflex and M-wave were elicited via electrical stimulation of the tibial nerve of the test leg. A large (~ 100 cm 2 ) copper anode was placed just above the patella, while the silver ball cathode, with a 1 cm diameter, was placed in the popliteal fossa, at the point where the weakest stimulus intensity was required to elicit a H-reflex (for details see Chapter 1; Miles et al., 1989; Tucker et al., 2005). The H-reflex was recorded from both the gastrocnemius and soleus. Square pulse stimuli of 0.5 ms duration were triggered from a computer and delivered by a Digitimer constant current stimulator (model DS7A). The smallest stimulus intensity used in this study was 4mA below that needed to elicit a H-reflex in either the medial gastrocnemius or soleus, while the largest stimulus intensity used was 4mA higher than required to obtain a maximal M- response in both muscles. Twenty stimulus intensities were determined between theses values (including both end points), which allowed the construction of full H/M curves for each subjects. Of the twenty stimulus intensities, nineteen were evenly distributed on a logarithmic scale (including the maximal and minimum intensities). The twentieth value was determined as the mid point between the highest two intensities, which ensured sufficient resolution in the upper intensity range. The 20 stimulus intensities were randomised, and were given between 4 and 5 seconds apart. At least 1-minute rest was provided between each trial of 20 stimuli. The fully randomised stimulus range was repeated 6 times in each of the 9 test conditions

139 Pain and the H-reflex Recording Surface electromyography (SEMG) recording electrodes (Duo-Trode silver/silver chloride) were placed 2 cm apart on the belly of the right medial gastrocnemius, soleus and tibialis anterior muscles, while for monopolar recording (medial gastrocnemius and soleus only), the reference electrode was placed on the Renshaw of the ipsilateral leg. The term monopolar recordings is used only for convenience since they were actually bipolar recordings in which the reference electrode was located at such a distance from the generator that the volume-conducted potentials were comparable to the noise level (Gerilovsky et al. 1985, 1989; Pierrot- Deseilligny and Mazevet, 2000). The monopolar and bipolar surface EMG recordings as well as the force output from a transducer positioned beneath the subjects test leg were amplified in a custom made EMG/force amplifier (with built in stimulus artefact suppressor and 20 Hz high pass filter for the EMG channels), and then recorded at a sample rate of 2 khz on a specially designed data acquisition program (National Instruments, LabVIEW : Brinkworth, 2004). The skin resistance of all bipolar recording electrode pairs was kept below 10kΩ; and subjects were grounded by lip clip electrodes (Türker et al. 1988) Analysis The 6 SEMG traces recorded for each stimulus intensity during each condition were averaged for each subject. Peak-to-peak (PTP) amplitude was determined for both the M-wave and H-reflex by the difference between the maximum and minimum values of the averaged responses, while area was calculated via the cumulative sum (CUSUM) of the averaged rectified traces (Brinkworth and Türker 2003; Ellaway 1978). The M-wave and the H-reflex recorded at each stimulus intensity (and measured as PTP and area) were normalised to the magnitude of the maximal M

140 Pain and the H-reflex wave obtained from each subject during each of the conditions. These data were then curve-fitted to minimise fluctuation and permit a reliable estimation of the data between points of stimulation (Figure 18). The M-wave was fitted with a sigmoidal curve (equ. 1) and the H-reflex was fitted with a Gaussian (equ. 2). The maximal H- reflex size before 30% maximal M-wave was calculated from the curve-fitted data for each condition for comparison (Chapters 2 and 3; Tucker and Türker 2004, 2005). a c b + c x x Equation 1: M-wave fit. a is the maximal value, b is the mid-point, c is the slope parameter and x is stimulus intensity. x mid a exp( 0.5( ) SD 2 ) Equation 2: H-reflex fit. a is the maximal value, exp is the power function e, the base of the natural logarithm, x is the stimulus intensity, mid is the mid-point of the curve, the point of maximal amplitude and SD is the standard deviation of the curve Statistics A repeated measures ANOVA was used on the pooled data to determine if a statistically significant difference existed between the experimental muscle conditions. In each case, the Bonferroni post-hoc test was used to determine where any significant differences lay. This was repeated for both recording methods (bipolar and monopolar), and both analysis techniques (PTP amplitude and area). The level of significance was set at p<

141 Pain and the H-reflex Figure 18: Raw and curve fitted H-reflex and M-wave curve Example data from the soleus muscle, recorded with bipolar electrodes in 1 subject, during the 5 th control condition. The top trace shows the curve fitted data highlighting the size of the H-reflex used for comparison in the current investigation; the maximum H-reflex obtained (grey dashed line in top trace) before 30% M-max (black dashed line in the top trace). The bottom trace represents the M-wave and H-reflex size normalised to the maximum M-wave in that trial (n=6)

142 Pain and the H-reflex 5.4. Results All subjects reported pain levels of greater than 4/10 before the stimuli were given during the pain trials, and all but one subject reported appropriate pain levels throughout these trials. A ninth subject, who did not report consistent pain levels was removed from the trial. In this case, the investigator noticed that the subject controlled pain scale had not been altered for a significant period during the lateral gastrocnemius pain trial and asked subject if her pain level was still the same (6/10 VAS). At this time, although the flow rate had remained constant, the subject realised that she felt no pain at all. The subject suggested that she had started thinking about something else, and seemed to ignore the pain. The following results are therefore based on the data from the remaining 8 subjects Pain scores Figure 19 (pg 5-142) illustrates the volume (ml) of both hypertonic and isotonic saline injected into the test muscles in both the painful and sham conditions respectfully (lateral gastrocnemius pain (2.91±1.54) and sham (2.72±0.73); tibialis anterior pain (2.24±1.01) and sham (2.16±0.92)). Subjects reported VAS pain scores (mean±sd/10) of 5.5±1.4 in the lateral gastrocnemius and 6.5±1.3 in the tibialis anterior during the pain trials (Figure 20, pg 5-143). No pain was reported in these muscles during the sham trials with the exception of one subject reported pain that reached 2/10 in the sham lateral gastrocnemius trial. The area of muscle pain reported by each subject is illustrated in Figure 21 (pg 5-144)

143 Pain and the H-reflex Figure 19: Isotonic and hypertonic saline injection Isotonic saline (0.9% NaCl) was injected via a constant flow device during the sham (SH) trials, while hypertonic saline (5% NaCl) was injected during the pain (PN) trials. The saline was injected into either the lateral gastrocnemius (LG) or the tibialis anterior (TA) during the test sessions. Similar volumes (shown as mean±sd) of the isotonic and hypertonic saline was used in each condition. The flow rate was determined as that required to maintain a pain level of approximately 5 (out of 10) in the test muscle during the pain trials (see figure 20, pg 5-144)

144 Pain and the H-reflex Figure 20: Pain scores throughout pain trials Mean±SDs of all VAS scores of all subjects. Subjects were asked to report their pain levels (out of 10) on a visual analogue scale (VAS) throughout both the pain and sham trials. The investigator modified the flow rate of the hypertonic saline (5% NaCl) in attempt to maintain the pain levels of approximately 5/10 throughout the pain trials. The VAS scores (mean±sd) are shown for the pain trial in the tibialis anterior (above) and lateral gastrocnemius (below). No pain was reported in the sham trials, except for 1 subject who reported a maximum pain of 2/10 in the lateral gastrocnemius (data not shown)

145 Pain and the H-reflex Figure 21: Area of pain during hypertonic saline (pain) trials Subjects were asked to represent the area of pain experienced during the trials on a figure similar to that shown. Data are shown for all subjects. Pain experienced during the lateral gastrocnemius pain trial is shown on the left, while pain experienced during the tibialis anterior pain trial is shown on the right. The points of injection are indicated by the black dots

146 Pain and the H-reflex Background muscle activity and H-reflex size Similar results were obtained when recording with either bipolar or monopolar electrode configurations, and when using area or PTP amplitude analysis methods. Due to convention, all results are shown from the bipolar recorded, PTP amplitude analysed data. No significant change in background muscle activity (mv) was seen in any muscle during the different experimental conditions (Table 1). Figure 22 (pg 5-147) illustrates this data normalised to average background muscle activity during the control trials. Similarly, no significant change in the maximal H/Mmax size (% Mmax) was noted in either the soleus or medial gastrocnemius between any of the experimental conditions (Table 1: pg 146). These data are shown as a percentage change from average control H/Mmax in Figure 23 (pg 5-148)

147 Pain and the H-reflex BGMA (mv) H/Mmax (%) Force (N) C1 C2 C3 C4 C5 LGSH LGPN TASH TAPN Sol 9.71 ± ± ± ± ± ± ± ± ±1.09 MG 3.77± ± ± ± ± ± ± ± ±0.37 TA 3.45 ± ± ± ± ± ± ± ± ±0.58 Sol ± ± ± ± ± ± ± ± ± 7.20 MG ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±20.0 Table 1: Background muscle activity (BGMA: mv), maximal H reflex size (%Mmax) and footplate force (N) recorded during each of the test conditions (mean±sem). Both the BGMA and H/Mmax results were obtained from the bipolar recorded, PTP amplitude analysed data. No significant differences in BGMA, H/Mmax size or force data were obtained during the different experimental conditions

148 Pain and the H-reflex Figure 22: Background muscle activity during control, sham and pain trials Percentage difference (mean±sem) of the background muscle activity in the 3 test muscles normalised to the average background muscle activity in the control conditions. SEMG was recorded from the soleus (black), gastrocnemius (mid grey) and tibialis anterior (light grey) of the test limb throughout the study, while subjects stood comfortably. No significant change in the level of background muscle activity occurred in any of the test muscles between the different conditions (LG: lateral gastrocnemius; TA: tibialis anterior; PN: pain trial; SH: sham trial; C1-5: control trials) in any of the muscles tested

149 Pain and the H-reflex Figure 23: H/M max during control, sham and pain trials Percentage difference (mean±sem) of maximal H/Mmax from average control H/Mmax. There was no significant change in the maximal H/Mmax size existed between the conditions (LG: lateral gastrocnemius; TA: tibialis anterior; PN: pain trial; SH: sham trial; C1-5: control trials) in either the soleus (black squares) or gastrocnemius (grey circles)

150 Pain and the H-reflex Foot plate force record No significant difference existed between the weight distributed (force in N) on the test limb during the different experimental conditions (data shown as mean±sem). C1: 325.1±12.96; C2: ±22.61; C3: ±18.30; C4: ±18.92; C5: ±44.18; TAPN: ±27.54; TASH: ±21.33; LGPN ±28 92; LGSH ± These data are shown (in Figure 24, pg 5-150) normalised to average background muscle activity during the control trials. Although a drop in mean force produced by the test leg appears to occur during the pain trials, the change in force is not significant between trials

151 Pain and the H-reflex Figure 24: Weight distribution during control, sham and pain trials Percentage difference (mean±sem) of maximal force (weight in N) recorded from a foot plate placed under the experimental leg, from the average force produced during the control trials. Although there looks to be a reduction of force during the pain trials, no significant change in the force existed between the conditions (LG: lateral gastrocnemius; TA: tibialis anterior; PN: pain trial; SH: sham trial; C1-5: control trials) in the current investigation

152 Pain and the H-reflex 5.5. Discussion The current investigation involved eliciting the full H-reflex and M-wave curve in both the medial gastrocnemius and soleus muscles. These muscles are ideal for such an investigation as the responses can be elicited in both muscles simultaneously. Furthermore, although the two muscles are located in close proximity to each other, they are quite different in structure and function, which enables the results of the study to have greater significance and reliability. There were four main findings in this investigation. Firstly, the controlled injection of hypertonic saline proved to be a good method for maintaining a pain level of approximately 6/10 throughout the investigation in the triceps surae. Secondly, postural adjustment by the subjects in response to the pain stimuli was not great enough to see a significant difference in the force applied to a force plate in the different conditions. Thirdly, background muscle activity recorded from the tibialis anterior, medial gastrocnemius and soleus muscles, was not altered by hypertonic saline induced pain in either the tibialis anterior or lateral gastrocnemius. Finally, hypertonic saline induced pain in either the tibialis anterior (antagonist) or lateral gastrocnemius (agonist) muscles, did not alter the H-reflex size in either the soleus or medial gastrocnemius during normal standing Pain levels throughout the trials The complex mechanisms behind modulation of motor output are of great importance in the clinical setting, and therefore require more investigation. It is currently understood that the sensation of pain is delivered to the central nervous system via free nerve endings of small-diameter mylenated afferents (A-δ) and unmylenated (C 5-151

153 Pain and the H-reflex nerve) fibers (Lund et al., 1991; Sessle, 1999), and that these nerves project to the spinal cord dorsal horn before traveling dorsally to the brain (Sessle and Hu, 1991). The structure of the pain system ensures that a human response to pain will involve both a sensory and motivational response, with the convergence and interaction between cutaneous and deep afferent inputs occurring in the thalamus and somatosensory cerebral cortex (Sessle and Hu, 1991) as well as the mid brain and medulla (Fields, 1984); via both complex circuits and direct somatosensory pathways (Di Piero et al., 1994); for a comprehensive review of the pain pathways see (Bromm and Lorenz, 1998). To date many different methods for experimentally inducing pain have been developed. It is understood that the pain sensation elicited by intramuscular injection of hypertonic saline (normally at 5-8% NaCl; (Graven-Nielsen et al., 1997b; Graven-Nielsen et al., 1998); is likely to be due to changes in the intramuscular sodium and potassium levels of the surrounding tissue (Graven- Nielsen et al., 1997b). One of the advantages of using this method of pain induction is the ability to control the pain levels over a long (20 minute) period. In the current investigation, a consistent level of pain was induced and controlled in each subject throughout the entirety of the painful test conditions. Similar volumes of saline were required to elicit a pain level of approximately 6/10 in both the lateral gastrocnemius and in the tibialis anterior. Furthermore, similar volumes of saline were injected during the isotonic trials (sham) in each of these muscles which ensure reliability of these trials Background muscle activity during symmetrical standing During symmetrical standing, it is well understood that the soleus consistently exhibits continuous, high levels of muscle activity, compared to the gastrocnemius 5-152

154 Pain and the H-reflex muscles that are active only intermediately in most cases (Joseph et al., 1955; Monster et al., 1978). Each of the muscles tested in the current investigation have a postural function, and are therefore required to maintain an upright stance, and are used under load in normal conditions (Nardone and Schieppati, 1988b). In the current investigation a higher level of background muscle activity existed in the soleus compared to the gastrocnemius and tibialis anterior, which supports the basic functions of the muscles during a standing task (Joseph et al., 1955; Monster et al., 1978) Background muscle activity during pain Given that the three muscles recorded during the investigation are required to maintain an upright stance, and each have quite different levels of muscle activity during normal stance, the current protocol provides conclusive evidence to support the suggestion that pain in either agonist, antagonist or test muscles in the lower limb, does not alter the level of background muscle activity during normal stance. Previous data on the changes of background muscle activity in human muscles during pain are quite variable. While some investigations show painful muscles have increased EMG as compared to non-painful muscles (Sessle et al., 1990; Sessle and Hu, 1991; Svensson et al., 1997a; Sessle, 1999), our data clearly supports others that show no EMG increase during muscle pain, particularly at rest or during low levels of contraction (Graven-Nielsen et al., 1997a; Birch et al., 2000) Weight distribution during pain and non pain states During this trial subjects were asked to stand symmetrically with their non experimental leg on a force plate to record alteration weight distribution throughout 5-153

155 Pain and the H-reflex the different conditions. Although force (foot plate) data have been used in aging and rehabilitation studies for many years (Seliktar et al., 1978), no similar information has been reported during limb pain. In the current investigation the average weight distributed in the painful leg decreased during the pain trials, however that this decrease was not statistically significant. It is possible that if more subjects were tested in the current investigation, that the difference may have been significant, and we therefore believe that further investigation into the distribution of body mass during painful episodes is necessary. It would be interesting in further investigations to determine if a possible reduction in body weight distribution on the painful limb coincided with a change in EMG activity. Given the current findings it may be expected that the background EMG of the muscles in the test limb would remain similar during painful and non painful states, while the weight distributed on the test limb decreases H-reflex modulation during pain It is well understood that the maximal H/Mmax size achieved by a subject will change significantly between different postural tasks. For example, the maximal H/Mmax size is significantly different between lying prone and standing (Abbruzzese et al., 1996), is modified during different phases of the walking and running cycle (Crenna and Frigo, 1987; Simonsen et al., 1995; Simonsen and Dyhre-Poulsen, 1999), at different levels of stretch of the soleus (Gerilovsky et al., 1977,1989; Gottlieb and Agarwal, 1978; Burke et al., 1983), and during contraction of the triceps surae (Chapter 2; Tucker and Türker 2005; for review see; Schieppati, 1987; Zehr, 2002; Tucker et al., 2005). In contrast, investigations considering the effect of pain on the H-reflex have produced variable results. While it has been shown consistently that pain does alter 5-154

156 Pain and the H-reflex the stretch reflex in the steadily contracting leg and jaw muscles (Matre et al., 1998; Matre et al., 1999, Svensson et al., 2000; Andersen et al., 2000), simultaneous changes in the H-reflex are sometimes (Le Pera et al., 2001), but not always (Matre et al., 1998; Svensson et al., 1998b; Matre et al., 1999). These findings have led to the suggestion that an increased firing rate in the dynamic γ-motoneurones increases the dynamic sensitivity of muscle spindles during muscle pain, while the α- motoneurone pool remains unchanged (Matre et al., 1998; Svensson et al., 1998b). It is also plausible that the tasks requirement of the test muscle during the experimental session may influence the effect that pain has on the H-reflex magnitude and background muscle activity. This effect, suggested by Rossi (1995) is directly related to the influence of Ib interneurones, which are known to be important in the co-ordination of multi-muscle movements (Rossi and Decchi, 1995). This is supported by the finding that tonic pain facilitates Ib heteronymous pathways in human lower limb motoneurones (Rossi and Decchi, 1995). We therefore argue that it is important in the future to investigate the effect of pain on the H-reflex during movement rather than simple standing or constant level background muscle activity tasks, to determine a more complete understanding of the interaction between pain and the α-motoneuron pool. It is plausible, given the results of the current investigation and previous studies, that a mechanism that helps to control motor output during active movement tasks when a greater chance for further injury/pain may exist, may be more efficient and therefore more likely than a mechanism that modifies motor output directly (via a direct Ia afferent-α-motoneuron pathway). This is because in the latter sinario, the mechanism would be active even in simple tasks where no change in muscle activity is required to reduce movement and therefore the chance of further injury

157 Pain and the H-reflex 5.6. Implications and conclusions The current investigation adds to the previous literature on the effect of pain on α- motoneuron excitability in many ways. Firstly, subjects in the current investigation were standing normally throughout the experimental sessions, which ensure that the results are physiologically significant. This is in contrast to previous leg experiments, where the subjects sat comfortably and plantar flexed (to approximately 10% MVC; Matre et al. 1998); or lay prone (Ellrich et al, 2000). The current investigation also differs in the location of the painful stimuli compared to previous work. Rather than studying the effect of pain on the muscle in pain, the current investigation also considered both antagonist and agonist muscle activity. Furthermore, the protocol used in the current investigation to elicit a full H-reflex and M-wave response function and then curve-fit the data ensures that true sizes can be determined. In previous investigations in the leg the H-reflex size has been measured at a single stimulus intensity which does not allow for movement of the mixed nerve to stimulating electrode relationship that can occur throughout an investigation (Matre et al., 1998; Capaday 1997; Fumoto et al., 2002; Simonsen and Dyhre-Poulsen 1999; reviewed in Tucker at al. 2005). Further to this, the level of muscle pain used in the current investigation ranged from approximately / 10 which is higher to that previously used (VAS: 3-5 / 10: Le Pera et al. 2001; Matre et al. 1998) in limb muscle H-reflex studies. We believe that the combined results from both this investigation the previous work by Matre and colleagues conclusively suggests that pain does not effect the α-motoneuron pool at low levels of static contraction including standing

158 Pain and the H-reflex In the current investigation background muscle activity was not significantly altered when muscle pain was induced in the same muscle (results from tibialis anterior only), an agonist (results from medial gastrocnemius and soleus) or an antagonist muscle (results from tibialis anterior, medial gastrocnemius and soleus). Given that subjects stood throughout the trials, which requires low levels of muscle activity and basic postural control, our investigation supports the suggestion that changes in background muscle activity during pain may relate to the task requirements and movement of the muscle in pain, rather than simply the existence of pain. In the current investigation no change in H-reflex size was seen during the pain or sham condition compared to the control conditions. The different results noted in the previous H-reflex and pain investigations (Ellrich and Treede 1998; Le Pera et al. 2001; Matre et al. 1999; Matre et al. 1998; Svensson et al. 1998a) may be explained by the functional requirements of the different motoneuron pools that were tested, as different results have been found in the jaw and leg muscles compared to the arm muscles. It is plausible that the tasks requirement of the test muscle during the experimental session may influence the effect that pain has on the H-reflex magnitude and background muscle activity given the influence of the Ib interneurons and their direct influence on muscle spindle afferents. All of the investigations to date have considered the effect of pain on steady state contracting or relaxed muscles, rather than on muscles that are actively moving. Given that pain has greater influence on the gamma-motoneuron system in previous investigations (Ellaway et al, 1997; Matre et al. 1999; Matre et al. 1998), it is plausible that pain may have a greater effect on muscles that are actively moving through their dynamic range rather than those that are in a steady state. This 5-157

159 Pain and the H-reflex mechanism would be appropriate because greater protection from further damage/pain may be required in the actively moving muscle compared to a muscle in steady state. Further investigation in this area is completed to determine the effect of pain on the α-motoneuron excitability during movement. We therefore recommend that further research is required in both limb and jaw muscles during active movement tasks. It is possible that pain will modulate the H-reflex during different stages of the active chewing cycle and walking cycle when a greater number of cortical and peripheral inputs are interacting to produce the final movement, and greater chance for tissue damage exists

160 Concluding Remarks 6. Concluding remarks The H-reflex, as described throughout this thesis, is an electrically elicited reflex that has been used as a tool to help answer a number of different neurophysiological questions. The H-reflex is used to help estimate the effect of many different factors on the excitability of the basic reflex pathway (Brooke et al., 1997; Capaday and Stein, 1986); to estimate the threshold of depolarisation of resting motoneurons (Türker and Miles, 1991) and its membrane potential trajectory (Türker, 1995); and to estimate the properties of Renshaw cell circuits (Katz and Pierrot-Deseilligny, 1998). It is also used in conjunction with the stretch reflex to help determine the influence of the muscle spindle and gamma motor system on basic neurological functions (Matre et al., 1998). The investigations conducted as part of this thesis have contributed to our understanding of human motor control with direct relation to changes in the excitability of the H-reflex circuit of two spatially similar, but functionally different muscles in the human triceps surae. The work has also provided a better understanding of the most appropriate stimulating, recording and analysis techniques to be used during H-reflex and M-wave investigations, and has answered questions pertaining to changes in the Ia - α-motoneuron excitability during muscle stretch and contraction. Finally, we sought to contribute to the varied data available concerning the effect of pain on this pathway. In chapter 2 it was demonstrated that the maximum H/M max obtained both at rest and during voluntary contraction was greater in soleus than in gastrocnemius. We argue that this suggests that soleus motoneurons have greater recruitment tendencies from spindle input than the gastrocnemius motoneurons under the test 6-159

161 Concluding Remarks conditions. In humans it has been shown that the soleus muscle has almost three times as many muscle spindles (Voss, 1971), and a far higher proportion of slow twitch fibres (Johnson et al., 1973; Edgerton et al., 1975; Green et al., 1981) than the gastrocnemius. Therefore, the soleus muscle has a much larger number of Ia fibres to be activated by the stimulus. It is also understood that homogenous spindle projections are stronger than heterogeneous projection to these two muscles (Mao et al., 1984), and it is therefore possible that soleus motoneurons receive greater spindle input than their gastrocnemius counterparts. Given the proximity, the dual innervation by the tibial nerve, and the clear differences in recruitment tendencies of these muscles, this study also reinforces the usefulness of using the soleus and gastrocnemius in H-reflex investigations. The ability to record the response to the same stimuli simultaneously from two quite different muscles increases the strength and the reliability of the findings. An example of the increased reliability when using both muscles is shown in the study discussed in chapter 3. This investigation focused on recording methods and variability in the maximal M-wave. The variability in both the H-reflex and M-wave between subjects and between conditions has been previously discussed, however in chapter 3 we introduce signal cancellation as a determining factor in relation to this issue. Signal cancellation results when both bipolar electrodes are recording a portion of the real muscle signal at the same time, and we argue that by recording a large muscle response in this way, a significant proportion of the muscle response will be eliminated. We further argue that this significant signal cancellation error does affect the reliability of the recorded data. Our argument is supported by the larger signal cancellation that was seen in the gastrocnemius when compared to the soleus which we argued is due to the greater dispersion of the signal in this muscle, and 6-160

162 Concluding Remarks which is directly due to greater fibre type variability in the gastrocnemius. From this investigation we strongly argue that monopolar recording electrodes should be used in some reflex studies to ensure that the signal cancellation error does not affect the reliability of the results. Variability in the maximal M-wave was also discussed thoroughly in chapter 4. This investigation supported our previous findings in relation to signal cancellation and the use of monopolar recording electrodes in reflex studies. By focusing on the maximal M-wave and considering different recording and analysis techniques, we were able to clearly show significant changes between conditions. We concluded that the variability may primarily be the result of changes in the stimulating electrode to mixed nerve and / or recording electrode to muscle bulk / motor point relationships. We therefore strongly argue that the maximal M-wave must be recorded in all conditions for normalisation purposes, and believe that this will significantly improve the reliability of H-reflex investigations. Through the work completed as part of this thesis, the collaborative authors of the papers have realised the importance of a greater understanding in the variability that exists in the H-reflex and M-wave curves both within subjects during different experimental sessions, and between subjects during the same experimental sessions. This variability exists because there are a number of factors that can affect the H-reflex size that are not easy to predict or control in advance. These variables include: a) the relative excitability of nerve fibres that lay within a mixed nerve, which also depends upon the distribution of the sensory and motor nerve fibres relative to the stimulating electrodes; b) the excitability of the Ia synapses, which are influenced by PAD from many other inputs via interneurons; c) the excitability of the 6-161

163 Concluding Remarks motoneurons, which are influenced by the varying inputs from thousands of synapses onto the cell; and d) the position of the stimulating / recording electrodes relative to the mixed nerve / muscle bulk and motor points. Although some of these variables have been described in detail; for example the change in excitability of the motoneuron during voluntary contraction (chapter 2) and movement of the stimulating and/or the recording electrodes relative to the mixed nerve position / muscle belly (chapters 3 and 4); and we take care to ensure the investigation are completed with these aspects in mind; we have observed changes in the entire H-reflex curve that we believe cannot be fully explained by these factors. For example, we have observed the full H-reflex curve moving to the left (hence requiring less current to reach the same size) and growing, either relative-to or coinciding-with changes in the M-wave curve during repeated trials in the same day, as well as during different experimental conditions. From this (yet unpublished work) we believe that a fixed magnitude of the H-reflex response relative to the maximal M-wave response (as used in chapters 2 and 5), or a fixed stimulus current that is repeated at different experimental conditions (as used in many previous H-reflex studies) may not fully capture the changes that occur in the reflex circuitry during different experimental sessions. To overcome this variability, we have developed a curve fitting technique that models the M-wave using a hyperbolic function, and the H-reflex using a Gaussian function (Figure 25, pg 6-165; Figure 18, pg 5-140, and described briefly in chapter 5). This modelling allows greater comparison between conditions, as the full H-reflex curves can be compared rather than the H-reflex size at individual stimulus intensities. We believe that this modelling will provide much more information about changes in the 6-162

164 Concluding Remarks excitatory H-reflex circuits than a single value at an arbitrary M-wave size or even the largest H-reflex response, as used by most researchers. With improved techniques for the recording (bipolar and monopolar techniques), and the analysis (area vs amplitude and non curve fitted vs curve fitted data) of reflex data, we foresee that the reliability of results reported in this area will continue to improve. In the final chapter of this thesis we describe the effects, or lack there of, of pain on the H-reflex, background muscle activity and weight distribution of the subjects during control, sham and painful conditions in the triceps surae during normal standing. Although in chapter 3 we argued that large muscle responses should be recorded by monopolar electrodes, the final study is recorded with both bipolar and monopolar electrodes, and is reported using the bipolar data. This is because there are benefits to using both methods of recording (as outlined in the introduction). In the investigation described in the last chapter, both monopolar and bipolar recording provided the same results, and given that conventionally bipolar electrode configurations are used, we chose to report the data recorded in this way. Much of the literature in the area of the effects of pain on muscle activity is quite varied, and although our investigation showed that pain did not affect any of the reported measures, we believe that further work is required in this area. The final investigation discussed in this thesis is unique in that it considers the background muscle activity and H-reflex size in both agonist and antagonist muscles to the area of induced pain; and, in that subjects stood throughout the investigation. We argue that the finding that pain did not affect these measures may be due to the low degree of ongoing motor control needed for the standing task. We therefore believe that 6-163

165 Concluding Remarks future research in this area should focus on the effect of pain on the H-reflex during movement tasks and / or postural variability

166 Concluding Remarks Area (%Mmax) Raw M Raw H Modelled M Modelled H Stimulus Size (m) Figure 25: Modelling the H-reflex and M-wave Modelled (solid lines) and raw (dashed lines and circles) H-reflex (black) and M- response (grey) curves over the full stimulus range. The x-axis has been normalised as described in the methods such that 5m, 50m and 95m correspond to 5%, 50% and 95% of the curve fitted M-response value, while the y-axis has been normalised to the modelled maximal M value 6-165

167 References 7. References Abbruzzese M, Rubino V, Schieppati M (1996) Task-dependent effects evoked by foot muscle afferents on leg muscle activity in humans. Electroencephalography and Clinical Neurophysiology 101: Arabadzhiev TI, Dimitrov GV, Dimitrova NA (2005a) Simulation analysis of the performance of a novel high sensitive spectral index for quantifying M-wave changes during fatigue. Journal of Electromyography and Kinesiology 15: Arabadzhiev TI, Dimitrov GV, Dimitrova NA (2005b) Intracellular action potential generation and extinction strongly affect the sensitivity of M-wave characteristic frequencies to changes in the peripheral parameters with muscle fatigue. Journal of Electromyography and Kinesiology 15: Ahern DK, Follick MJ, Council JR, Laser-Wolston N, Litchman H (1988) Comparison of lumbar paravertebral EMG patterns in chronic low back pain patients and non-patient controls. Pain 34: Ali A, Sabbahi MA (2001) Test-retest reliability of the soleus H-reflex in three different positions. Electromyography and Clinical Neurophysiology 41: Andersen OK, Graven-Nielsen T, Matre D, Arendt-Nielsen L, Schomburg ED (2000) Interaction between cutaneous and muscle afferent activity in polysynaptic reflex pathways: a human experimental study. Pain 84: Babenko V, Graven-Nielsen T, Svensson P, Drewes AM, Jensen TS, Arendt-Nielsen L (1999) Experimental human muscle pain and muscular hyperalgesia induced by combinations of serotonin and bradykinin. Pain 82:1-8. Babenko V, Svensson P, Graven-Nielsen T, Drewes AM, Jensen TS, Arendt-Nielsen L (2000) Duration and distribution of experimental muscle hyperalgesia in humans following combined infusions of serotonin and bradykinin. Brain Research 853:

168 References Basmajian JV (1973) Electrodes and electrode connectors. In: New developments in electromyograpy and clinical neurophysiology (Desmedt JE, ed), pp Basel: Karger. Beck RB, O'Malley M, van Dijk JP, Nolan P, Stegeman DF (2004) The effects of bipolar electrode montage on conduction velocity estimation from the surface electromyogram. Journal of Electromyography and Kinesiology 14: Bigland-Ritchie B, Jones DA, Woods JJ (1979) Excitation frequency and muscle fatigue: electrical responses during human voluntary and stimulated contractions. Experimental Neurology 64: Bini G, Cruccu G, Manfredi M (1981) Acute experimental dental pain: a technique for evaluating pain modulating procedures. Journal of Neuroscience Methods 3: Bini G, Cruccu G, Hagbarth KE, Schady W, Torebjork E (1984) Analgesic effect of vibration and cooling on pain induced by intraneural electrical stimulation. Pain 18: Birch L, Christensen H, Arendt-Nielsen L, Graven-Nielsen T, Sogaard K (2000) The influence of experimental muscle pain on motor unit activity during low-level contraction. European Journal of Applied Physiology 83: Brinkworth RS, Türker KS (2003) A method for quantifying reflex responses from intra-muscular and surface electromyogram. Journal of Neuroscience Methods 122: Brinkworth RSA (2004) Response of the human jaw to mechanical stimulation of teeth. PhD thesis. University of Adelaide, Adelaide, Australia. Brock LG, Coombs JS. Eccles JC (1952) The recording of potentials from motoneurons with an intracellular electrode. Journal of Physiology 117:

169 References Brodin P, Miles TS, Türker KS (1993) Simple reaction-time responses to mechanical and electrical stimuli in human masseter muscle. Archives of Oral Biology 38: Broman H, De Luca CJ, Mambrito B (1985) Motor unit recruitment and firing rates interaction in the control of human muscles. Brain Research 337: Bromm B, Lorenz J (1998) Neurophysiological evaluation of pain. Electroencephalography and Clinical Neurophysiology 107: Brooke MH, Kaiser KK (1970) Muscle fibre types: how many and what kind? Archives of Neurology 23: Brooke JD, McIlroy WE, Miklic M, Staines WR, Misiaszek JE, Peritore G, Angerilli P (1997) Modulation of H reflexes in human tibialis anterior muscle with passive movement. Brain Research 766: Brunia CHM, Burke R, Delwaide PJ, Desmedt JE, Dietrichson P, Gassel MM, Hagbarth KE, Herman R, Hugon M, Isch F, Lance JW, Landau WM, Mark RF, Mayer RF, Peirrot-Deseilligny E, Pinelli P, Rushworth G, Schnecnk E, Struppler A, Wiesendanger M, Young RR (1973) Methodology of the triceps surae propioceptive reflexes. In: New Developments in Electromyography And Clinical Neurophysiology (Desmedt JE, ed), pp Basel: Karger. Buchthal F, Schmalbruch H (1970) Contraction times and fibre types in intact human muscle. Acta Physiologica Scandinavica 79: Buchthal F, Schmalbruch H (1980) Motor unit of mammalian muscle. Physiological Reviews 60: Burke D (1988) Muscle afferent activity and its central projection in man. Australian Paediatrics Journal 24 Suppl 1: Burke D, McKeon B, Skuse NF (1981) Dependence of the Achilles tendon reflex on the excitability of spinal reflex pathways. Annuals in Neurology 10: Burke D, Gandevia SC, McKeon B (1983) The afferent volleys responsible for spinal proprioceptive reflexes in man. Journal of Physiology 339:

170 References Burke D, Gandevia SC, McKeon B (1984) Monosynaptic and oligosynaptic contributions to human ankle jerk and H-reflex. Journal of Neurophysiology 52: Burke D, Adams RW, Skuse NF (1989) The effects of voluntary contraction on the H reflex of human limb muscles. Brain 112: Burke D, McKeon B, Skuse NF, Westerman RA (1980) Anticipation and fusimotor activity in preparation for a voluntary contraction. Journal of Physiology 306: Burke RE (1973) On the central nervous system control of fast and slow twitch motor units. In: New Developments in Electromyography and Clinical Neurophysiology (Desmedt JE, ed). Burke RE, Levine DM, Zajac FE, Tsairis P, Engel WK (1971) Mammalian motor units: physiological-histochemical correlation in three types in cat gastrocnemius. Science 174: Burke RE (1968) Group Ia synaptic input to fast and slow twitch motor units of cat triceps surae. Journal of Physiology 196: Burke RE (1981) Motor units: Anatomy, physiology and functional organisation. In: Handbook of Physiology Sect I. The Nervous System, Vol II. Motor System, 1 Edition (Brooks VB, ed), pp Bethesda MD: American Physiological Society. Burke RE (1991) Selective recruitment of motor units. In: Motor control: concepts and issues (Humphrey DR, Freund HJ, eds), pp 5-21: John Wiley & Sons Ltd. Burke RE, Jankowska E, Bruggencate GT (1970) A comparison of peripheral and rubrospinal synaptic input to slow and fast twitch motor units of triceps surae. Journal of Physiology 207: Burke RE, Levine DM, Tsairis P, Zajac FE (1973) Physiological types and histochemical profiles in motor units of the cat gastrocnemius. Journal of Physiology 234:

171 References Calancie B, Bawa P (1990) Motor unit recruitment in humans. In: The Segmental Motor System (Binder MD, Mendell LM, eds), pp New York: Oxford University Press. Calancie B, Broton JG, Klose KJ, Traad M, Difini J, Ayyar DR (1993) Evidence that alterations in presynaptic inhibition contribute to segmental hypo- and hyperexcitability after spinal cord injury in man. Electroencephalography amd Clinical Neurophysiology 89: Capaday C (1997) Neurophysiological methods for studies of the motor system in freely moving human subjects. Journal of Neuroscience Methods 74: Capaday C, Stein RB (1986) Amplitude modulation of the soleus H-reflex in the human during walking and standing. The Journal of Neuroscience 6: Castaingts V, Martin A, Van Hoecke J, Perot C (2004) Neuromuscular efficiency of the triceps surae in induced and voluntary contractions: morning and evening evaluations. Chronobiology International 21: Clancy EA, Morin EL, Merletti R (2002). Sampling, noise-reduction and amplitude estimation issues in surface electromyography. Journal of Electromyography and Kinesiology 12:1-16. Collins GA, Cohen MJ, Naliboff BD, Schandler SL (1982) Comparative analysis of paraspinal and frontalis EMG, heart rate and skin conductance in chronic low back pain patients and normals to various postures and stress. Scandinavian Journal of Rehabilitation Medicine 14: Crenna P, Frigo C (1987) Excitability of the soleus H-reflex arc during walking and stepping in man. Experimental Brain Research 66: Crone C, Johnson LL, Hultborn H, Orsnes GB (1999) Amplitude of the maximum motor response (M max ) in human muscles typically decreases during the course f an experiment. Experimental Brain Research 124:

172 References Day SJ, M Hulliger (2001) Experimental simulation of cat electromyogram: evidence for algebraic summation of motor-unit action-potential trains. Journal of Neurophysiology 86: De Luca CJ (1993) Use of the surface EMG signal for performance evaluation of back muscles. Muscle Nerve 16: De Luca CJ, Mambrito B (1987) Voluntary control of motor units in human antagonist muscles: coactivation and reciprocal activation. Journal of Neurophysiology 58: De Luca CJ, Merletti R (1988) Surface myoelectric signal cross-talk among muscles of the leg. Electroencephalography and Clinical Neurophysiology 69: Dengler R, Kossev A, Wohlfahrt K, Schubert M, Elek J, Wolf W (1992) F waves and motor unit size. Muscle Nerve 15: Di Piero V, Ferracuti S, Sabatini U, Pantano P, Cruccu G, Lenzi GL (1994) A cerebral blood flow study on tonic pain activation in man. Pain 56: Dubner R (1991) Basic mechanisms of pain associated with deep tissues. Canadian Journal of Physiology and Pharmacology 69: Duysens J, Tax AA, Trippel M, Dietz V (1992) Phase-dependent reversal of reflexly induced movements during human gait. Experimental Brain Research 90: Duysens J, Tax AA, van-der DB, Trippel M, Dietz V (1991) Selective activation of human soleus or gastrocnemius in reflex responses during walking and running. Experimental Brain Research 87: Eccleston C (1994) Chronic pain and attention: a cognitive approach. British Journal of Clinical Psychology 33: Edgerton VR, Smith JL, Simpson DR (1975) Muscle fibre type populations of human leg muscles. Histochemistry Journal 7:

173 References Ellaway PH (1978) Cumulative sum technique and its application to the analysis of peristimulus time histograms. Electromyography and Clinical Neurophysiology 45: Elliott MA, Walter GA, Gulish H, Sadi AS, Lawson DD, Jaffe W, Insko EK, Leigh JS, Vandenborne K (1997) Volumetric measurement of human calf muscle from magnetic resonance imaging. Magma 5: Ellrich J, Treede RD (1998) Convergence of nociceptive and non-nociceptive inputs onto spinal reflex pathways to the tibialis anterior muscle in humans. Acta Physioligica Scandinavica 163: Ellrich J, Steffens H, Schomburg E D (2000) Neither a general flexor nor a withdrawal pattern of nociceptive reflexes evoked from the human foot. Neuroscience Research 37: Eriksson PO, Stalberg E, Antoni L (1984) Flexibility in motor-unit firing pattern in the human temporal and masseter muscles related to type of activation and location. Archives of Oral Biology 29: Faist M, Hoefer C, Hodapp M, Dietz V, Berger W, Duysens J (2006) In humans Ib facilitation depends on locomotion while suppression of Ib inhibition requires loading. Brain Research 1076: Falco FJ, Hennessey WJ, Goldberg G, Braddom RL (1994) H reflex latency in the healthy elderly. Muscle Nerve 17: Farina D, Arendt-Nielsen L, Merletti R, Graven-Nielsen T (2004) Effect of experimental muscle pain on motor unit firing rate and conduction velocity. Journal of Neurophysiology 91: Farina D, Schulte E, Merletti R, Rau G, Disselhorst-Klug C (2003) Single motor unit analysis from spatially filtered surface electromyogram signals. Part I: spatial selectivity. Medical & Biological Engineering and Computing 41:

174 References Farina D, Cescon C, Merletti R (2002) Influence of anatomical, physical, and detection-system parameters on surface EMG. Biological Cybernetics 86: Feine JS, Chapman CE, Lund JP, Duncan GH, Bushnell MC (1990) The perception of painful and nonpainful stimuli during voluntary motor activity in man. Somatosensory and Motor Research 7: Fields HL (1984) Neurophysiology of pain and pain modulation. American Journal of Medicine 77:2-8. Friedman WA, Sypert GW, Munson JB, Fleshman JW (1981) Recurrent inhibition in type-identified motoneurons. Journal of Neurophysiology 46: Fuglevand AJ, Zackowski KM, Huey KA, Enoka RM (1993) Impairment of neuromuscular propagation during human fatiguing contractions at submaximal forces. Journal of Physiology 460: Fujii H (1977) Evoked EMG of masseter and temporal muscles in man. Journal of Oral Rehabilitation 4: Fukunaga T, Roy RR, Shellock FG, Hodgson JA, Day MK, Lee PL, Kwong-Fu H, Edgerton VR (1992) Physiological cross-sectional area of human leg muscles based on magnetic resonance imaging. Journal of Orthopaedics Research 10: Fumoto M, Komiyama T, Nishihira Y (2002) Soleus H-reflex dynamics during fast plantarflexion in humans. Journal of Electromyography and Kinesiology 12: Funase K, Miles TS (1999). Observations on the variability of the H reflex in human soleus. Muscle and Nerve 22: Funase K, Higashi T, Yoshimura T, Imanaka K, Nishihira Y (1996) Evident difference in the excitability of the motoneuron pool between normal subjects and 7-173

175 References patients with spasticity assessed by a new method using H-reflex and M- response. Neuroscience Letters 203: Gandevia SC (1992) Some central and peripheral factors affecting human motoneuronal output in neuromuscular fatigue. Sports Medicine 13: Gandevia SC (2001) Spinal and supraspinal factors in human muscle fatigue. Physiology Review 81: Garland SJ, Gerilovsky L, Enoka RM (1994) Association between muscle architecture and quadriceps femoris H- reflex. Muscle Nerve 17: Garnett R, Stephens JA (1981) Changes in the recruitment threshold of motor units produced by cutaneous stimulation in man. Journal of Physiology 311: Gerilovsky L, Gydikov A, Radicheva N (1977) Changes in the shape of the extraterritorial potentials of tonic motor units, M- and H-responses of triceps surae muscles at different muscle lengths and under conditions of voluntary activation. Experimental Neurology 56: Gerilovsky L, Tsvetinov P, Trenkova G (1985) Is the Mmax/Hmax ration a reliable index of the monosynaptic excitability in man? Recruitment curves for H- and M-responses using monopolar and bipolar recording techniques. Acta Physiologica et Pharmacologica Bulgarica 11: Gerilovsky L, Tsvetinov P, Trenkova G (1989) Peripheral effects on the amplitude of monopolar and bipolar H-reflex potentials from the soleus muscle. Experimental Brain Research 76: Goddard G, Karibe H, McNeill C (2004) Reproducibility of visual analogue scale (VAS) pain scores to mechanical pressure. Cranio 22: Goldberg J, Sullivan SJ, Seaborne DE (1992) The effect of two intensities of massage on H-reflex amplitude. Physical Therapy 72:

176 References Gossen ER, Ivanova TD, Garland SJ (2003) The time course of the motoneuron afterhyperpolarization is related to motor unit twitch speed in human skeletal muscle. Journal of Physiology 552: Gottlieb GL, Agarwal GC (1976) Extinction of the Hoffmann reflex by antidromic conduction. Electroencephalography and Clinical Neurophysiology 41: Gottlieb GL, Agarwal GC (1978) Stretch and Hoffmann reflexes during phasic voluntary contractions of the human soleus muscle. Electroencephalography and Clinical Neurophysiology Supplement 44: Graven-Nielsen T, Svensson P, Arendt-Nielsen L (1997a) Effects of experimental muscle pain on muscle activity and co-ordination during static and dynamic motor function. Electroencephalography and Clinical Neurophysiology 105: Graven-Nielsen T, McArdle A, Phoenix J, Arendt-Nielsen L, Jensen TS, Jackson MJ, Edwards RH (1997b) In vivo model of muscle pain: quantification of intramuscular chemical, electrical, and pressure changes associated with saline-induced muscle pain in humans. Pain 69: Graven-Nielsen T, Fenger-Gron LS, Svensson P, Steengaard-Pedersen K, Arendt- Nielsen L, Staehelin Jensen T (1998) Quantification of deep and superficial sensibility in saline-induced muscle pain--a psychophysical study. Somatosensory and Motor Research 15: Graven-Nielsen T, Lund H, Arendt-Nielsen L, Danneskiold-Samsoe B, Bliddal H (2002) Inhibition of maximal voluntary contraction force by experimental muscle pain: a centrally mediated mechanism. Muscle & Nerve 26: Green HJ, Daub B, Houston ME, Thomson JA, Fraser I, Ranney D (1981) Human vastus lateralis and gastrocnemius muscles. A comparative histochemical and biochemical analysis. Journal of Neurological Science 52: Greenwood LF, Sessle BJ (1976) Pain, brain stem mechanisms and motor functions. In: Mastication and Swallowing (Sessle BJ, Hannam AG, eds), pp Toronto: Toronto University

177 References Gydikov A, Kosarov D (1972) Volume conduction of the potentials from separate motor units in human muscle. Electromyography and Clinical Neurophysiology 12: Hansen PO, Svensson P, Arendt-Nielsen L, Jensen TS (1999) Relation between perceived stimulus intensity and exteroceptive reflex responses in the human masseter muscles. Clinical Neurophysiology 110: Harridge SD, Bottinelli R, Canepari M, Pellegrino MA, Reggiani C, Esbjornsson M, Saltin B (1996) Whole-muscle and single-fibre contractile properties and myosin heavy chain isoforms in humans. Pflugers Archives 432: Heckman CJ, Binder MD (1990) Neural mechanisms underlying the orderly recruitment of motoneurons. In: The Segmental Motor System (Binder MD, Mendell LM, eds), pp New York: Oxford University Press. Heckman CJ (1994) Alterations in synaptic input to motoneurons during partial spinal cord injury. Medical Science in Sports and Exercise 26: Henneman E (1985) The size-principle: a deterministic output emerges from a set of probabilistic connections. Journal Experimental Biology 115: Henneman E, Somjen G, Carpenter DO (1965) Functional significance of cell size in spinal motoneurons. Journal of Neurophysiology 28: Herbert RD, Moseley AM, Butler JE, Gandevia SC (2002) Change in length of relaxed muscle fascicles and tendons with knee and ankle movement in humans. Journal of Physiology 539: Herman R (1967) Function of the gastrocnemius and soleus muscles. A preliminary study in the normal human subject. Physical Therapy 47: Hodges JS, Moseley GL, Gabrielsson A, Gandevia SC (2003) Experimental muscle pain changes feedforward postural responses of the trunk muscles. Experimental Brain Research. Hodges PW (2001) Changes in motor planning of feedforward postural responses of the trunk muscles in low back pain. Experimental Brain Research 141:

178 References Hodges PW, Richardson CA (1996) Inefficient muscular stabilization of the lumbar spine associated with low back pain. A motor control evaluation of transversus abdominis. Spine 21: Hollinshead WH, ed (1951) Functional anatomy of the limbs and back; a text for students of physical therapy and others interested in the locomotor apparatus. Philedelphia: WB Saunders Company. Hugon M (1973) Proprioceptive Reflexes and the H reflex, Methodology of the Hoffmann Relex in Man: IN New Developments in Electromyography And Clinical Neurophysiology (Desmedt JE, ed) pp Basel: Karger. Hutchison DL, Roy RR, Hodgson JA, Edgerton VR (1989) EMG amplitude relationships between the rat soleus and medial gastrocnemius during various motor tasks. Brain Research 502: Hwang IS (2002) Assessment of soleus motoneuronal excitability using the joint angle dependent H reflex in humans. Journal of Electromyography and Kinesiology 12: Isaacs ER, Szumski AJ, Suter C (1968) Central and peripheral influences on the H- reflex in normal man. Neurology 18: Jaberzadeh S, Svensson P, Nordstrom MA, Miles TS (2003) Differential modulation of tremor and pulsatile control of human jaw and finger by experimental muscle pain. Experimental Brain Research 9:9. Jankowska E (1984) On neuronal pathways of presynaptic depolarization of group I muscle afferents. Experimental Brain Research: Jankowska E, Lundberg A (1981) Interneurones in the spinal cord. Trends In Neuroscience September: Johnson MA, Polgar J, Weightman D, Appleton D (1973) Data on the distribution of fibre types in thirty-six human muscles. Journal of the Neurological Sciences 18:

179 References Joseph J, Nightingale A, Williams PL (1955) A detailed study of the electric potentials recorded over some postural muscles while relaxed and standing. Journal of Physiology 127: Kalmar JM, Cafarelli E (1999) Effects of caffeine on neuromuscular function. Journal of Applied Physiology 87: Kanda K, Desmedt JE (1983) Cutaneous facilitation of large motor units and motor control of human fingers in precision grip. In: Motor Control Mechanisms in Health and Disease (Desmedt JE, ed), pp New York: Raven Press. Kanda K, Burke RE, Walmsley B (1977) Differential control of fast and slow twitch motor units in the decerebrate cat. Experimental Brain Research 29: Katz R, Pierrot Deseilligny E (1984) Facilitation of soleus-coupled Renshaw cells during voluntary contraction of pretibial flexor muscles in man. Journal of Physiology 355: Keenan KG, Farina D, Maluf KS, Merletti R, Enoka RM (2005) Influence of amplitude cancellation on the simulated surface electromypgram. Journal of Applied Physiology 98: Knikou M, Rymer WZ (2002) Hip angle induced modulation of H reflex amplitude, latency and duration in spinal cord injured humans. Clinical Neurophysiology 113(11): Knikou M, Rymer WZ (2003) Static and dynamic changes in body orientation modulate spinal reflex excitability in humans. Experimental Brain Research 152(4): Komi PV, Buskirk ER (1970) Reproducibility of electromyographic measurements with inserted wire electrodes and surface electrodes. Electromyography 4: Kuchinad RA, Ivanova TD, Garland SJ (2004) Modulation of motor unit discharge rate and H-reflex amplitude during submaximal fatigue of the human soleus muscle. Experimental Brain Research 158:

180 References Kudina LP, Pantseva RE (1988) Recurrent inhibition of firing motoneurones in man. Electroencephalography and Clinical Neurophysiology 69: Kudina LP, Alexeeva NL (1992) After-potentials and control of repetitive firing in human motoneurones. Electroencephalography and Clinical Neurophysiology 85: Kumar S, Tandon OP, Mathur R (2002) Pain measurement: A formidable task. Indian Journal of Physiology and Pharmacology 46: Kupa EJ, Roy SH, Kandarian SC, De Luca CJ (1995) Effects of muscle fiber type and size on EMG median frequency and conduction velocity. Journal of Applied Physiology 79: Lavigne G, Frysinger R, Lund JP (1983) Human factors in the measurement of the masseteric silent period. Journal of Dental Research 62: Le Pera D, Graven-Nielsen T, Valeriani M, Oliviero A, Di Lazzaro V, Tonali PA, Arendt-Nielsen L (2001) Inhibition of motor system excitability at cortical and spinal level by tonic muscle pain. Clinical Neurophysiology 112: Lipton S (1991) Pain mechanisms and management. British Medical Bulletin 47:1-4 Löscher WN, Cresswell AG, Thorstensson A (1996) Central fatigue during a longlasting submaximal contraction of the triceps surae. Experimental Brain Research 108: Lund JP, Donga R, Widmer CG, Stohler CS (1991) The pain-adaptation model: a discussion of the relationship between chronic musculoskeletal pain and motor activity. Canadian Journal Physiological Pharmacology 69: Macefield VG, Gandevia SC, Bigland Ritchie B, Gorman RB, Burke D (1993) The firing rates of human motoneurones voluntarily activated in the absence of muscle afferent feedback. Journal of Physiology 471: MacIntyre DL, Sorichter S, Mair J, Berg A, McKenzie DC (2001) Markers of inflammation and myofibrillar proteins following eccentric exercise in humans. European Journal of Applied Physiology 84:

181 References Mackenzie RA, Burke D, Skuse NF, Lethlean AK (1975) Fibre function and perception during cutaneous nerve block. Journal Neurology Neurosurgery and Psychiatry 38: Mao CC, Ashby P, Wang M, McCrea D (1984) Synaptic connections from large muscle afferents to the motoneurons of various leg muscles in man. Experimental Brain Research 56: Mark RF, Coquery JM, Paillard J (1968) Autogenetic reflex effects of slow or steady stretch of the calf muscles in man. Experimental Brain Research 6: Matre DA, Sinkjaer T, Svensson P, Arendt NL (1998) Experimental muscle pain increases the human stretch reflex. Pain 75: Matre DA, Sinkjaer T, Knardahl S, Andersen JB, Arendt-Nielsen L (1999) The influence of experimental muscle pain on the human soleus stretch reflex during sitting and walking. Clinical Neurophysiology 110: Matthews PBC (1996) Relationship of firing intervals of human motor units to the trajectory of post-spike after-hyperpolarization and synaptic noise. Journal of Physiology 492: Matthews PB (1999) The effect of firing on the excitability of a model motoneurone and its implications for cortical stimulation. Journal of Physiology 518: McMullan S, Simpson DA, Lumb BM (2004) A reliable method for the preferential activation of C- or A-fibre heat nociceptors. Journal of Neuroscience Methods 138: McPhedran AM, Wuerker RB, Henneman E (1998) Properties of motor units in a homogeneous red muscle (soleus) of the cat. Journal of Neurophysiology 28: Miles TS, Nordstrom MA (1995) Fatigue of jaw muscles and speech mechanisms. Advances in Experimental Medicine and Biology 384: Miles TS, Türker KS, Le TH (1989) Ia reflexes and EPSPs in human soleus motor neurones. Experimental Brain Research 77:

182 References Miller AE, MacDougall JD, Tarnopolsky MA, Sale DG (1993) Gender differences in strength and muscle fiber characteristics. European Journal of Applied Physiology and Occupational Physiology 66: Milne RJ, Kay NE, Irwin RJ (1991) Habituation to repeated painful and non-painful cutaneous stimuli: a quantitative psychophysical study. Experimental Brain Research 87: Milne RJ, Aniss AM, Kay NE, Gandevia SC (1988) Reduction in perceived intensity of cutaneous stimuli during movement: a quantitative study. Experimental Brain Research 70: Misiaszek JE (2003) The H-reflex as a tool in neurophysiology: its limitations and uses in understanding nervous system function. Muscle Nerve 28: Monster AW, Chan H, O'Connor D (1978) Activity patterns of human skeletal muscles: relation to muscle fiber type composition. Science 200: Montoya P, Sitges C, Garcia-Herrera M, Izquierdo R, Truyols M, Blay N, Collado D (2005) Abnormal affective modulation of somatosensory brain processing among patients with fibromyalgia. Psychosomatic Medicine 67: Moore AD (1967) Synthesized EMG waves and their implications. American journal of physical medicine 46: Moseley GL, Nicholas MK, Hodges PW (2004) Does anticipation of back pain predispose to back trouble? Brain 127: Nagata A, Christianson JC (1995) M-wave modulation at relative levels of maximal voluntary contraction. European Journal of Applied Physiology and Occupational Physiology 71: Nardone A, Schieppati M (1988a) Shift of activity from slow to fast muscle during voluntary lengthening contractions of the triceps surae muscles in humans. Journal of Physiology 395:

183 References Nardone A, Schieppati M (1988b) Postural adjustments associated with voluntary contraction of leg muscles in standing man. Experimental Brain Research 69: Nemeth PM, Solanki L, Gordon DA, Hamm TM, Reinking RM, Stuart DG (1986) Uniformity of metabolic enzymes within individual motor units. Journal of Neuroscience 6: Nemeth PM, Pette D, Vrbova G (1981) Comparison of enzyme activities among single muscle fibres within defined motor units. Journal of Physiology 311: Nichols TR (1999) Receptor mechanisms underlying heterogenic reflexes among the triceps surae muscles of the cat. Journal of Neurophysiology 81: Nielsen OB, Clausen T (2000) The Na+/K(+)-pump protects muscle excitability and contractility during exercise. Exercise Sport Science Review 28: Ohashi J (1995) Difference in changes of surface EMG during low-level static contraction between monopolar and bipolar lead. Journal of Physical Anthropology 14: Parratte B, Tatu L, Vuillier F, Diop M, Monnier G (2002) Intramuscular distribution of nerves in the human triceps surae muscle: anatomical bases for treatment of spastic drop foot with botulinum toxin. Surgery Radiology Anatomy 24: Perez-Pareja J, Borras C, Sese A, Palmer A (2005) Pain perception and fibromyalgia. Actas Españolas de Psiquiatría 33: Petersen N, Christensen LOD, Nielsen J (1998) The effect of transcranial magnetic stimulation on the soleus H reflex during human walking. Journal of Physiology 513: Pierrot-Deseilligny E, Pinelli P, Rushworth G, Schnecnk E, Struppler A, Wiesendanger M, Young RR (1973) Methodology of the triceps surae propioceptive reflexes. In J. E. Desmedt (Ed.), New Developments in Electromyography and Clinical Neurophysiology (pp ). Basel: Karger

184 References Pierrot-Deseilligny, E (1997) Assessing changes in presynaptic inhibition of Ia afferents during movement in humans. Journal of Neuroscience Methods 74: Pierrot-Deseilligny E, Mazevet D (2000) The monosynaptic reflex: a tool to investigate motor control in humans. Interest and limits. Neurophysiology Clinical 30: Pierrot-Deseilligny E, Bergego C, Katz R, Morin C (1981) Cutaneous depression of Ib reflex pathways to motoneurones in man. Experimental Brain Research 42: Proske U, Morgan DL (2001) Muscle damage from eccentric exercise: mechanism, mechanical signs, adaptation and clinical applications. Journal of Physiology 537: Rainoldi A, Melchiorri G, Caruso, I (2004) A method for positioning electrodes during surface EMG recordings in lower limb muscles. Journal of Neuroscience Methods 134: Robinson KL, McComas AJ, Belanger AY (1982) Control of soleus motoneuron excitability during muscle stretch in man. Journal of Neurology Neurosurgery and Psychiatry 45: Roeleveld K, Stegeman DF, Vingerhoets HM, Van Oosterom A (1997) Motor unit potential contribution to surface electromyography. Acta Physioligica Scandinavica 160: Roeleveld K, Stegeman DF (2002) What do we learn from motor unit action potentials in surface electromyography? Muscle Nerve Suppl 11:S92-7. Rossi A, Decchi B (1994) Flexibility of lower limb reflex responses to painful cutaneous stimulation in standing humans: evidence of load-dependent modulation. Journal of Physiology 481:

185 References Rossi A, Decchi B (1995) Cutaneous nociceptive facilitation of Ib heteronymous pathways to lower limb motoneurones in humans. Brain Research 700: Rossi-Durand C, Jones KE, Adams S, Bawa P (1999) Comparison of the depression of H-reflexes following previous activation in upper and lower limb muscles in human subjects. Experimental Brain Research 126: Rossignol S, Dubuc R, Gossard JP (2006) Dynamic sensorimotor interactions in locomotion. Physiological Reviews 86: Roy SH, De Luca CJ, Schneider J (1986) Effects of electrode location on myoelectric conduction velocity and median frequency estimates. Journal of Applied Physiology 61: Roy SH, De Luca CJ, Casavant DA (1989) Lumbar muscle fatigue and chronic lower back pain. Spine 14: Roy SH, De Luca CJ, Snyder-Mackler L, Emley MS, CRenshaw RL, Lyons JP (1990) Fatigue, recovery, and low back pain in varsity rowers. Medical Science in Sports Exercise 22: Rutten GJ, Gaasbeek RD, Franssen H (1998) Decrease in nerve temperature: a model for increased temporal dispersion. Electroencephalography and Clinical Neurophysiology 109: Sabbahi M, De Luca CJ (1981) Topical anesthesia: H-reflex recovery changes by desensitization of the skin. Electroencephalography and Clinical Neurophysiology 52: Sadoyama T, Masuda T, Miyata H, Katsuta S (1998) Fibre conduction velocity and fibre composition in human vastus lateralis. European Journal of Applied Physiology and Occupational Physiology 57: Sakamoto K, Swie YW (2003) EMG characteristics of low back and lower limb muscles during forward bending posture. Electromyography and Clinical Neurophysiology 43:

186 References Salo D, Eget D, Lavery RF, Garner L, Bernstein S, Tandon K (2003) Can patients accurately read a visual analogue pain scale? American Journal of Emergency Medicine 21: Scaglioni G, Ferri A, Minetti AE, Martin A, Van Hoecke J, Capodaglio P, Sartorio A, Narici MV (2002) Plantar flexor activation capacity and H reflex in older adults: adaptations to strength training. Journal of Applied Physiology 92: Schieppati M (1987) The Hoffmann reflex: a means of assessing spinal reflex excitability and its descending control in man. Progress in Neurobiology 28: Schmidt RF (1985) Motor systems. In: Fundamentals of Neurophysiology, 3 Edition (Schmidt RF, ed), pp New York: Springer-Verlag. Schulte E, Ciubotariu A, Arendt-Nielsen L, Disselhorst-Klug C, Rau G, Graven- Nielsen T (2004) Experimental muscle pain increases trapezius muscle activity during sustained isometric contractions of arm muscles. Clinical Neurophysiology 115: Schulte E, Dimitrova NA, Dimitrov, GV Rau G, Disselhorst-Klug C (2005) Estimation of the muscle fibre semi-length under varying joint positions on the basis of non-invasively extracted motor unit action potentials. Journal of Electromyography and Kinesiology 15: Schwindt PC, Calvin WH (1972) Membrane-potential trajectories between spikes underlying motoneuron firing rates. Journal of Neurophysiology 35: Scott AJ, Cadden SW (1996) Suppression of an inhibitory jaw reflex by the anticipation of pain in man. Pain 66: Seliktar R, Susak Z, Najenson T, Solzi P (1978) Dynamic features of standing and their correlation with neurological disorders. Scandinavian Journal of Rehabilitation Medicine 10:

187 References Semmler JG, Türker KS (1994) Compound group I excitatory input is differentially distributed to motoneurons of the human tibialis anterior. Neuroscience Letters 178: Sessle BJ (1999) Neural mechanisms and pathways in craniofacial pain. Canadian Journal of Neurological Science 26 Suppl 3:S7-11. Sessle BJ, Hu JW (1991) Mechanisms of pain arising from articular tissues. Canadian Journal of Physiological Pharmacology 69: Sessle BJ, Woodside DG, Bourque P, Gurza S, Powell G, Voudouris J, Metaxas A, Altuna G (1990) Effect of functional appliances on jaw muscle activity. American Journal of Orthodontic Dentofacial Orthopaedics 98: Shall MS, Goldberg SJ (1992) Extraocular motor units: type classification and motoneuron stimulation frequency-muscle unit force relationships. Brain Research 587: Simonsen EB, Dyhre-Poulsen P, Voigt M (1995) Excitability of the soleus H reflex during graded walking in humans. Acta Physiologica Scandinavica 153: Simonsen EB, Dyhre-Poulsen P (1999) Amplitude of the human soleus H reflex during walking and running. Journal of Physiology 515: Stålberg E (1986) Single fibre EMG, Macro EMG, and scanning EMG. New ways of looking at the motor unit. CRC Critical Reviews in Clinical Neurology 2: Stegeman DF, Blok JH, Hermens HJ, Roeleveld K (2000) Surface EMG models: properties and applications. Journal of Electromyography and Kinesiology 10: Stetson DS, Albers JW, Silverstein BA, Wolfe RA (1992) Effects of age, sex, and anthropometric factors on nerve conduction measures. Muscle Nerve 15:

188 References Stuart DG, Enoka RM (1983) Motoneurons, motor units, and the size principle. In: The clinical neurosciences (Rosenberg RN, ed), pp V471-V517. New York: Churchill Livingstone Inc. Svensson P, Houe L, Arendt NL (1997a) Bilateral experimental muscle pain changes electromyographic activity of human jaw-closing muscles during mastication. Experimental Brain Research 116: Svensson P, De Laat A, Graven-Nielsen T, Arendt-Nielsen L, Macaluso GM (1999) Effect of clenching levels on heteronymous H-reflex in human temporalis muscle. Experimental Brain Research 126: Svensson P, De Laat A, Graven-Nielsen T, Arendt-Nielsen L (1998a) Experimental jaw-muscle pain does not change heteronymous H-reflexes in the human temporalis muscle. Experimental Brain Research 121: Svensson P, Graven-Nielsen T, Matre D, Arendt-Nielsen L (1998b) Experimental muscle pain does not cause long-lasting increases in resting electromyographic activity. Muscle Nerve 21: Svensson P, Miles TS, Graven-Nielsen T, Arendt-Nielsen L (2000) Modulation of stretch-evoked reflexes in single motor units in human masseter muscle by experimental pain. Experimental Brain Research 132: Svensson P, Minoshima S, Beydoun A, Morrow TJ, Casey KL (1997b) Cerebral processing of acute skin and muscle pain in humans. Journal of Neurophysiology 78: Táboríková H (1973) Supraspinal Influences on H-Reflexes. In: New Developments in Electromyography and Clinical Neurophysiology (Desmedt JE, ed), pp Basel: Karger. Táborìková H, Sax DS (1968) Motoneurone pool and the H-reflex. Journal of Neurology Neurosurgery and Psychiatry 31: Taylor A (1976) Fibre types in the muscles of mastication. In: Mastication, 1 Edition (Anderson DJ, Matthews B, eds), pp Bristol: John Wright

189 References Taylor A, Hidaka O, Durbaba R, Ellaway PH (1997) Fusimotor patterns in cat jaw muscles inferred from comparision of spindle behaviour in active an dpassive movements. Journal of Physiology 501P:39P-40P. Thunberg J, Ljubisavljevic M, Djupsjobacka M, Johansson H (2002) Effects on the fusimotor-muscle spindle system induced by intramuscular injections of hypertonic saline. Experimental Brain Research 142: Todd G, Taylor JL, Gandevia SC (2003) Measurement of voluntary activation of fresh and fatigued human muscles using transcranial magnetic stimulation. Journal of Physiology 551: Travell J, Rinzler S, Herman M (1942) Pain and disability of the shoulder and arm. Treatment by intramuscular infiltration with procaine hydrochloride. Journal of the American Medical Association 120: Trontelj JV (1973) A study of the H-reflex by single fibre EMG. Journal of Neurology Neurosurgery and Psychiatry 36: Trontelj JV (1973) A study of the F response by single fibre electromyography. New Developments in Electromyography and Clinical Neurophysiology. J. E. Desmedt. Basel, Karger. 3: Tucker KJ, Türker KS (2006) Triceps surae stretch and voluntary contraction alters maximal M-wave magnitude. Journal of Electromyography and Kinesiology In press. Tucker KJ, Tuncer M, Türker KS (2005) A review of the H-reflex and M-wave in the human triceps surae. Human Movement Science 24: Tucker KJ, Türker KS (2005) A new method to estimate signal cancellation in the human maximal M-wave. Journal of Neuroscience Methods 149: Tucker KJ, Türker KS (2004) Muscle spindle feedback differs between the soleus and gastrocnemius in humans. Somatosensory and Motor Research, 21:

190 References Türker KS (1993) Electromyography: some methodological problems and issues. Physical Therapy 73: Türker KS (1995) The shape of the membrane potential trajectory in tonically-active human motoneurons. Journal of Electromyography and Kinesiology 5:3-14. Türker KS, Miles TS (1990) Cross-talk from other muscles can contaminate EMG signals in reflex studies of the human leg. Neuroscience Letters 111: Türker KS, Powers RK (2002) The effects of common input characteristics and discharge rate on synchronization in rat hypoglossal motoneurones. Journal of Physiology 541: Türker KS, Miles TS, Le HT (1988) The lip-clip: a simple, low-impedance ground electrode for use in human electrophysiology. Brain Research Bulliten 21: Türker KS, Scutter SD, Yang J, Jenkins M, Tucker KJ, Abolfathi PP (1999) Reflex control of masticatory muscles in man. In: Neurobiology of Mastication - from Molecular to Systems Approach (Nakamura Y, Sessle BJ, eds), pp Amsterdam: Elsevier Science B.V. Vanden Noven S, Hamm TM, Stuart DG (1986) Partitioning of monosynaptic Ia excitatory postsynaptic potentials in the motor nucleus of the cat lateral gastrocnemius muscle. Journal of Neurophysiology 55: Verrier MC (1985) Alterations in H reflex magnitude by variations in baseline EMG excitability. Electroencephalography and Clinical Neurophysiology 60: Visser SL, Zonneveldt A, De RW, Don JA, Martens EI, Stam D (1983) Normal Hoffmann reflex latencies (H-M interval) in relation to age and body length. Clinical Neurology and Neurosurgery 85: Voss VH (1971) Tabelle der absoluten und relativen Muskelspindelzahlen der menschilichen Skelettmuskulatur. Anat Anz 129:562:

191 References Wang K, Arendt-Nielsen l, Svensson P (2001) Excitatory actions of experimental muscle pain on early and late components of human jaw stretch reflexes. Archives of Oral Biology 46: Wang K, Arendt-Nielsen l, Svensson P (2002) Capsaicin-induced muscle pain alters the excitability of the human jaw-stretch reflex. Journal of Dental Research 81: Weerakkody NS, Percival P, Hickey MW, Morgan DL, Gregory JE, Canny BJ, Proske U (2003) Effects of local pressure and vibration on muscle pain from eccentric exercise and hypertonic saline. Pain 105: Windhorst U (1996) The spinal cord and its brain: representations and models. To what extent do forebrain mechanisms appear at brainstem spinal cord levels? Progress in Neurobiology 49: Wolpaw JR, Carp JS (1990) Memory traces in spinal cord. Trends in Neuroscience 13: Wuerker RB, McPhedran AM, Henneman E (1965) Properties of motor units in a heterogeneous pale muscle (M. Gastrocnemius) of the cat. Journal of Neurophysiology 28: Zehr EP, Stein RB (1999) Interaction of the jendrassik maneuver with segmental presynaptic inhibition. Experimental Brain Research 124: Zehr PE (2002) Considerations for use of the Hoffmann reflex in exercise studies. European Journal of Applied Physiology 86:

192 Appendix 8. Appendix TUCKER, K.J., TUNCER, M., TÜRKER, K.S (2005) A review of the H-reflex and M- wave in the human triceps surae. Human Movement Science 24(5/6) TUCKER,K.J., TÜRKER,K.S. (2005) Muscle spindle feedback differs between the soleus and gastrocnemius in humans. Somatosensory and Motor Research, 21(3/4) TUCKER,K.J., TÜRKER,K.S. (2005) A new method to estimate signal cancellation in the human maximal M-wave. Journal of Neuroscience Methods 149; TUCKER,K.J., TÜRKER,K.S. (2005) Triceps surae stretch and voluntary contraction alters maximal M-wave magnitude. Journal of Electromyography and Kinesiology In press 8-191

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215 Tucker, K.J. & Turker, K.S. (2005) Muscle spindle feedback differs between the soleus and gastrocnemius in humans. Somatosensory and Motor Research v. 21 (3/4) pp NOTE: This publication is included in the print copy of the thesis held in the University of Adelaide Library. It is also available online to authorised users at:

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234