The clinical occurrence of pain due to sprain/strain injuries is frequent, but our understanding of its etiology is rudimentary. There is no clear evidence that a given musculotendinous tissue disruption causes a specific pain response. Furthermore, there is little information that enables one to positively associate the site of tissue trauma to the same area from which the pain is perceived to originate. We are presenting a theoretical rationale to identify and to subsequently guide treatment strategies based on the identification of muscle dysfunction resulting from tissue trauma.
As a first step in developing rehabilitative strategies to enhance recovery from sprain/strain injuries one can carefully characterize changes in muscle function. In the present context, muscle dysfunction is an unusual pattern of muscle recruitment during a prescribed set of movements. These altered neural strategies reflect changes in neural inputs to the motoneuron pools that will be recruited in generating a specific motor task. A practical window into studying these muscle recruitment strategies clinically is surface electromyography (EMG).
We propose that multiple ratios of EMG amplitudes between pairs of different combinations of muscles that perform similar functions can be used to identify altered neural strategies of motoneuron pool recruitment as a result of tissue injury. This assessment strategy differs from one recently proposed by DeLuca (10) to identify muscle dysfunction. DeLuca (10) suggested that by using the frequency spectrum of an EMG signal from an individual motor pool to identify muscle fatigue, one could identify dysfunctional and compensatory muscle recruitment patterns.
We propose that the physiological changes and functional adaptations of the neurological and musculoskeletal systems that occur with muscle injury can also be used for diagnoses. Some factors common to acute and chronic muscle injury contributing to altered neural strategies and abnormal recruitment of motoneuron pools are described. A theoretical basis for expected patterns of both hypoactive and hyperactive EMG activity, given certain types of muscle dysfunction, is presented. Ratios of EMG activity levels between different combinations of compensators at the static end range of carefully defined motor tasks are suggested as useful measures of hypoactive and hyperactive EMG activity and indicators of neuromuscular dysfunction.
Causes of Tissue Injuries
Since muscle dysfunction may, in many instances, be initiated by muscle strain, it is appropriate to briefly examine muscle strains. Muscle strains causing muscle fiber failure may result from either a single contraction or from the cumulative effects of many contractions (2). It has been suggested that muscle strain injuries probably occur most often during eccentric contractions(3,9,15,36), because significantly higher forces can be generated compared to concentric or isometric contractions (15). Furthermore, muscle strains have been reported to occur more readily at the musculotendinous junction(15,16,29,36,39,41) because of the relatively high concentration of stress and, therefore, greater strain at the ends of the muscle fibers (29). This may occur whether the muscle is actively or passively stretched(16). Fast-twitch muscle fibers (Type II) have been reported to be more susceptible to injury than slow-twitch muscle fibers(13), perhaps due to the faster rates of contraction and the higher forces and power that they generate (15), circumstances often associated with muscle injury. The cause of tissue injury may also be related to the fatigability of the muscle, but this relationship is poorly understood as well. Structural or contractile deficiencies caused by fatigue, weakness, and/or scar tissue from previous muscle injury could reduce the ability of the muscle to absorb energy and further increase the susceptibility to injury (9,15,29).
Cellular Aspects of Acute Muscle Tissue Injury
After a muscle injury, the degeneration of muscle fibers may be characterized by the disruption of myofibrils, mitochondria, and sarcoplasmic reticulum (6). There can also be interruptions of the continuity of the sarcolemma (7). Disruption of the sarcolemma or sarcoplasmic reticulum may lead to elevation in intracellular Ca2+ concentrations, which could saturate the capability to pump the Ca2+ from the sarcoplasm, resulting in a loss of Ca2+ homeostasis and uncontrolled contraction of the sarcomeres(2). The sarcomeres could continue to contract in the absence of action potentials as long as the intracellular calcium remains elevated and the ATP supply is adequate (2,42). The mechanical forces in the fiber from these contractions could further damage the structural and contractile components (2). If there is tissue damage of sufficient severity, clinical symptoms, such as pain, swelling, and discoloration, as well as a loss of muscle strength and changes in proprioception, could develop (19).
Inflammatory responses can occur within a few hours and some regenerative or other reparative processes can begin within 4-5 d following injury(2,7). More than a few weeks may be needed for a fiber to regenerate (12). The number of damaged muscle fibers may be related to the severity of the injury (e.g., full or partial tear) and may have an impact on the time necessary for the tissue to heal. If the neuromuscular junction is damaged, the time needed to recover may be delayed further. In many cases the contractile strength of an injured muscle can return to normal (15), but during that healing process the muscle is more likely to be more susceptible to further strain than normal.
Factors that Alter Muscle Recruitment Patterns
We hypothesize that tissue trauma could initiate two kinds of processes that would result in altered recruitment patterns of a group of synergistic and antagonistic muscles and lead to altered EMG ratios(Fig. 1).
Presumably, tissue damage can stimulate pain receptors as a result of strain. Furthermore, nociceptive input to the spinal cord can alter the excitation levels of multiple motoneuron pools. Specific combinations of motoneuron pools affected by nociception are a function of the specific site and intensity of the irritation (37). This neurophysiological characteristic in itself could cause sufficient changes in motoneuron pool net excitability to alter motor unit recruitment and amplitudes of EMG signals. The magnitude of the tissue injury and resulting dysfunction, however, may not always be proportional to the perceived pain and discomfort.
Changes in EMG amplitudes also can result from neural activation patterns that do not directly involve pain sensation, but which accommodate the diminished capacity of an injured muscle to generate force. The nervous system apparently can detect a reduced capacity to generate force from a specific muscle or group of muscles and compensate by recruiting more motoneurons. This compensation can be made by recruiting motor units from an uninjured area of the muscle or from other muscles capable of performing the same tasks that are not directly affected by the injury. These types of compensatory strategies have been demonstrated in an experimental model in which some muscles of a group of synergists are surgically removed (35). Immediately after the removal of the muscles, compensatory events become evident. There is a marked increase in EMG activity and force generated from the muscle compensating for the reduced output of the primary muscle(14,33). Over a period of weeks there is likely to be a continuing process of integrative adaptations. For example, with continued use the compensating muscle hypertrophies and therefore could change the force/EMG ratio.
This sequence of adaptations may not be unlike those occurring with the onset of muscle injury. A ruptured or weakened muscle fiber will cause a reduction in force-generating capacity(2,15,29,30,36,39) and may also result in reduced EMG activity (hypoactivity). Synergistic muscles would be recruited to compensate, leading to an unusually high level of motoneuron activity in these muscles (hyperactivity). Since the recruitment of some motoneuron pools would increase while others decrease, the changes in ratios of the amplitudes of EMG from the affected pools would be readily apparent.
A third way in which altered EMG ratios might occur is via a conscious or subconscious reeducation of the motor system. With a persistent injury, patients learn or acquire new ways to perform a task that enables them to avoid using the musculature that is directly associated with the tissue injury. For example, a patient may favor one arm or rotate the trunk less than normal when performing tasks. Neural adaptations may also occur with a chronically stretched tendon or muscle unit, because faulty feedback on the force-length-velocity relationship could decrease the load the muscle would support (31). Chronic compensation due to excessive force requirements could also affect the spinal reflex pathways(25) causing intrafusal muscle fibers to reset to a higher gain after contraction (24) or the Golgi tendon organ pathways to become desensitized (24,40).
Muscle Spasm and EMG Activity
Palpating muscles at rest to locate muscle spasms is a common clinical technique, although the etiology of muscle spasm is not well understood. Muscle spasms could describe several phenomena; however, in many if not most instances, spasms are presumed to be involuntary sustained muscle contractions. Although several studies have found no evidence of abnormal EMG activity associated with muscle spasms (17,26), we suggest three possible events that could be recognized clinically as muscle spasm, which could result in conflicting findings based on EMG ratios. First, an injured muscle due to tissue trauma may appear rigid during a Ca2+ overload period in the absence of electrical activation and exhibit reduced EMG amplitudes. In a second scenario, a compensating muscle could become overworked, develop localized contractions, and appear rigid upon palpation. Some of these events are likely to reveal predictably hyperactive EMG recordings during motor tasks. As a third example, unusually shortened muscles may also be interpreted as a muscle spasm, since they can be rigid upon palpation (27) even when they are inactive. However, during movement, normal EMG activity would be expected(8). Any of these physiological conditions could result in abnormal EMG ratios.
Use of Ratios of EMG Amplitudes to Identify Normal and Abnormal Muscle Recruitment
Ratios of EMG amplitudes have been used to identify changes in recruitment patterns across motor pools in a variety of motor tasks and species(22,23,34). For example, Hodgson et al.(21) found that the ratios of EMG amplitudes between the human soleus and gastrocnemius muscles changed relative to the rate of development of force during plantarflexion, i.e., there was a bias toward higher activities of the gastrocnemius compared to the soleus at the faster speeds. This approach also was used to identify changes in recruitment of synergists in response to the speed of treadmill locomotion in cats(22) and after adaptation to space flight(20,32). For example, the relative activation of the medial gastrocnemius was elevated compared to the soleus in monkeys performing a specific motor task after compared to before space flight(32). Boucher et al. (5) used ratios of EMG between the vastus medialis and vastus lateralis to diagnose patella femoral syndrome.
To identify muscle dysfunction in the back and neck using muscle ratios, we monitor 14 electrode sites, including most of the major muscle groups that control and stabilize the cervical, thoracic, and lumbar spine. These electrode sites include the cervical paraspinal/superior upper trapezius(cervical paraspinal), upper trapezius, middle trapezius, thoracic paraspinal/lower trapezius (thoracic paraspinal), latissimus dorsi, internal and external obliques (oblique), and lumbar paraspinal muscle groups. We have developed techniques to precisely identify the equivalent anatomical location of muscle loci from subject to subject using three landmarks, i.e., C7, acromioclavicular joint and iliac crest, 13 anthropometric measurements and regression coefficients derived from radiographic images to predict T5, T10, and L1 spinous processes based on gender, height, and distance between C-7 and the iliac crest.
We collected EMG data using 14 channels of a 16- channel system with 1012-ohm input impedance differential amplifiers and a common ground. Analog signals were band-pass filtered, rectified, smoothed using a RC hardware filter with a 145-ms time constant and then underwent A/D conversion. A band-pass from 66 to 215 Hz with 12 dB per octave rolloff at low and high pass limits were selected to reduce interference from ECG artifacts at the low end, and ensured a high signal-to-noise ratio at the higher frequencies. The skin was prepared with rubbing alcohol and an abrasive cream (Omni Prep, Aurora, CO) and a single electrode with two silver-silver chloride 1.0-cm diameter discs spaced 2.5 cm apart (Multi-Biosensors, El Paso, TX) were affixed to each electrode site.
EMG recordings are averaged across four 5-s trials while the subject maintains a “static end of range position” (i.e., isometric phase) for a series of nine carefully defined motor tasks developed to be very inclusive with respect to a wide range of injuries of the back. These motor tasks include: a) the asymmetrical equivalents with one arm raised overhead to 45°, the opposite lower extremity extended posteriorly to 35° with the pelvis maintained in a stable position while balancing on the opposite limb, b) shoulder shrug to chin level, c) bilateral arm abduction to 90°, d) forward trunk flexion to approximately 45° with the head and arms in line with the trunk, e) bilateral anterior arm flexion to 90°, f) reaching overhead with both arms, and g) trunk rotation to the left or right with the arms at the side and the pelvis stabilized and the trunk extended.
For consistency, we carefully train technicians to record and acquire EMG data at defined timepoints. EMG recordings acquired only during the“static end of range position” of the motor tasks are selected for analysis since within and between subject variability, with respect to these anatomical positions and postures, can be more easily monitored. Ratios between muscle pairs across a population of subjects may be more reliable in a“static effort at the end of range position” because: a) the muscles under tension are consistently in a similar position for all joints(1); b) co-activations of agonist and antagonist muscles are matched (4); c) muscle length and force-length relationships are constant (38); and d) the relationships between EMG amplitudes and muscle tension are linear during an isometric contraction (38). Although this approach may underestimate spinal stress as might occur during dynamic activity, it appears to be an acceptable method for quantifying the distribution of activation patterns among muscles and for identifying specific relative contributions of muscles to support the body in a given position against gravity(28).
We minimize variability by using a fixed interelectrode distance of 2.5 cm. Cross talk between overlapping and underlying muscles may occur, given the electrode sites and electrode spacing selected. The high reliability between session-logged EMG amplitudes and logged ratios (r = 0.93 and 0.92, respectively) suggest the EMG data appear to have been obtained consistently and reliably. We also have developed methods similar to those proposed by Hemingway et al. (18) to correct EMG amplitudes for signal attenuation attributable to adipose tissue between the muscle and the surface electrode because of its known filtering effect on EMG. We have found significant gender differences for all nine motor tasks and established separate data bases for men and women.
A Strategy for Distinguishing Normal and Abnormal EMG Patterns in Trunk Muscles
To date we have accumulated a rather extensive data base on normal subjects and patients with sprain/strain injuries and have demonstrated that EMG ratios can be a sensitive discriminator of altered recruitment patterns and muscle dysfunction. Figure 2 simultaneously illustrates normative values of the relative recruitment during one motor task based on 200 men. The normative values for each of 196 muscle ratios (14 × 14 muscle pairs) during bilateral anterior arm flexion to 90° are presented as a transparent surface. Figure 2A and B, and C and D, are mirror images, with the normative range for each muscle pair identified by the data values plotted on the z axis two standard deviations below(Fig. 2A and C) and above (Fig. 2B and D) the mean. Muscle pairs for a male patient 32 d after a serious automobile accident (test, Fig. 2A and B) and 21 d later(retest, Fig. 2C and D) are overlaid onto the normative surface. Aberrant ratios greater than two standard deviations below(Fig. 2A and C) or above (Fig. 2B and D) the mean are considered to be hypoactive (red) and hyperactive (green), respectively. Figure 2A and B identify 31 aberrant ratios. In Figure 2A the bilateral cervical paraspinal, upper trapezius, thoracic paraspinal, and lumbar paraspinal muscle groups exhibit hypoactivity relative to two or more muscle groups. InFigure 2B, the left and right middle trapezius exhibit hyperactivity relative to seven and 10 muscle groups, respectively; the left latissimus dorsi is hyperactive relative to 11 muscle groups; and the right latissimus dorsi is hyperactive relative to two muscle groups.Figure 2C and D identify 15 aberrant ratios for the same patient at retest, i.e., 53 d post-injury. The bilateral cervical paraspinal and left lumbar paraspinal muscle groups exhibit consistent patterns of hypoactivity relative to three or more muscle pairs (Fig. 2C). The bilateral middle trapezius and thoracic paraspinal muscle groups exhibit consistent patterns of hyperactivity relative to three or more muscle groups (Fig. 2D).
We have developed procedures that define hypoactive and hyperactive EMG activity based on patterns of aberrant ratios as illustrated inFigure 2. One fundamental premise to these procedures is the identification of those muscles that are most likely to be recruited to compensate for others in a given motor task based on fundamental kinesiological principles. For example, during bilateral anterior arm flexion to 90°, a hypoactive cervical paraspinal muscle group might result in a compromise in proper elevation, upward rotation, and stabilization of the cervical spine. The ipsilateral upper trapezius muscle group would be the principal muscle recruited to assist with upward rotation and elevation of the shoulder girdle. The contralateral cervical paraspinal would help maintain cervical extension. The ipsilateral middle trapezius muscle group would retract the scapula to help stabilize the shoulder girdle. Recruitment of the ipsilateral thoracic paraspinal muscle group would increase upward rotation of the shoulder girdle and help maintain cervicothoracic extension. The remaining 10 muscle groups share no kinesiological functions with the cervical paraspinal muscle group.
For a muscle group to be categorized as hypoactive or hyperactive in a motor task, the ratio between it and at least two muscle groups must be aberrant, and one of these muscle groups must be categorized as one that could functionally compensate for the muscle group in question, i.e., a compensatory muscle. Levels of severity for hypoactivity or hyperactivity during a motor task are further defined by the number of aberrant compensatory muscle ratios and the total number of ratios that are aberrant.
The aberrant muscle ratios for the same patient depicted inFigure 2 are presented as altered muscle recruitment patterns in Figure 3A and B, respectively.Figure 3A corresponds with Figure 2A and B (test) and Figure 3B corresponds withFigure 2C and D (retest), respectively. Relationships between compensatory muscle pairs with aberrant ratios are identified by connecting lines. For example, in Figure 3A (test) the left thoracic paraspinal is hypoactive (red) whereas the left middle trapezius, bilateral latissimus dorsi, and left oblique muscle groups are hyperactive (green) and thus considered compensatory muscles. InFigure 3B, the left thoracic paraspinal is hyperactive, thus compensatory to the hypoactive left cervical paraspinal and left lumbar paraspinal muscle groups. Note that the bilateral thoracic paraspinal muscle groups were initially hypoactive (test) (Figs. 2A and 3A), but were hyperactive at retest (Figs. 2D and 3B), illustrating that the neural strategies for motoneuron pool recruitment can continue to adapt over time as the patient recovers from injury.
We categorize subjects, as well as individual muscle groups, based on muscle patterns across the nine motor tasks as normal, subclinical, or abnormal. To minimize the probability of misclassification, a muscle group identified as having altered muscle recruitment patterns must exhibit either hypoactivity or hyperactivity in at least three motor tasks. For example, based on a minimum of one aberrant compensatory ratio and a second aberrant ratio, we reported 94% of the normal subjects (N = 400) with at least one muscle group demonstrating hypoactive or hyperactive EMG activity in one motor task, while only 24% of these subjects exhibited consistent patterns of hypoactivity or hyperactivity across three motor tasks. Altered muscle recruitment patterns are further defined as subclinical or abnormal, based on the number of aberrant ratios within and across motor tasks. Normal and subclinical categories are combined for classification purposes since it is likely that no significant dysfunctions exist for a muscle(s) or region(s) indicated to be subclinical, if the clinical evaluation of the subject is asymptomatic.
The procedures described above were used to discriminate individuals who sustained sprain/strain injuries due to automobile or slip-and-fall accidents(N = 61) from a population of normal controls (N = 400). The classification accuracy of this system was 88%, with a specificity of 90%, and a sensitivity of 70%. The between-session reliability for two technicians with a 14-d intertest interval for 40 normal subjects was 95%. The data presented in Figures 2 and 3 are representative of 44 patients with sprain/strain injuries who were retested approximately 2 wk after the initial test. The between-session reliability for the overall classification of these 44 patients was 91%. Seven patients (16%) tested and retested normal. One patient tested normal and retested with abnormal muscle recruitment patterns. Of the remaining 36 patients who initially tested with altered muscle recruitment patterns, 69% showed improvement at retest and 31% retested normal. These results suggest that most of the sources of measurement error using surface electrodes can be minimized sufficiently to reflect altered neural strategies of compensatory and adaptive muscle recruitment for back muscle sprain/strain trauma common to motor vehicle and slip-and-fall accidents. The sensitivity of these procedures to classify patients with different etiologies for back pain (i.e., intervertebral disk disorders with and without radiculopthy, etc.) must still be studied.
We suggest that injuries resulting in indirect or direct loss of muscle force-generating capacity will produce alterations in the neural strategies of recruitment of motoneuron pools when performing prescribed motor tasks. Tissue damage, even in the absence of pain, can cause a loss of force-generating capability and result in hypoactive EMG activity patterns. Neural adaptations and chronic hypoactive EMG patterns also could be attributed to stretched tendons or muscle units or to persistent pain. Muscles that are recruited to compensate for the resulting functional deficits could display increased force generation and hyperactive EMG activity patterns. These compensatory and adapting mechanisms will alter the recruitment patterns among muscles having similar functions that can be readily identified using ratios of EMG activity between pairs of muscles during the static end range of specified motor tasks. These altered recruitment patterns under carefully controlled conditions can be used to objectively and consistently identify dysfunctional muscles. This approach has the potential to provide significant assistance to clinicians in the diagnosis of tissue trauma associated with back injuries, and for the development of rehabilitative strategies to up or down train abnormal reflex responses and reeducate the neural feed-back loop(11,24,43).
1. Andrew, J. G. Biomechanical measures of muscular effort.Med. Sci. Sports Exerc.
2. Armstrong, R. B., G. L. Warren, and J. A. Warren. Mechanisms of exercise-induced muscle fiber injury. Sports Med.
3. Armstrong, R. B. Initial events in exercise-induced muscular injury. Med. Sci. Sports Exerc.
4. Basmajian, J. V. and C. J. DeLuca. Muscles Alive
, 5th Ed. Baltimore: Williams & Wilkins, 1985, pp. 223-229.
5. Boucher, J. P., A. Pepin, and R. Lefebvre. Using the vastus medialis to vastus lateralis IEMG ratio as a neuromuscular imbalance index for the diagnosis of patello-femoral syndrome. Med. Sci. Sport Exerc.
6. Byrd, S. Alterations in the sarcoplasmic reticulum: a possible link to exercise-induced muscle damage. Med. Sci. Sports Exerc.
7. Carlson, B. M. and J. A. Faulkner. The regeneration of skeletal muscle fibers following injury: a review. Med. Sci. Sports Exerc.
8. Cram, J. R. Surface EMG recordings and pain-related disorders: a diagnostic framework. Biofeedback Self-Regul.
9. Cuillo, J. V. and B. Zarins. Biomechanics of the musculotendinous unit: relation to athletic performance and injury.Clin. Sports Med.
10. DeLuca, C. J. Use of the surface EMG signal for performance evaluation of back muscles. Muscle Nerve
11. Evatt, M. L., S. L. Wolf, and R. L. Segal. Modification of human spinal stretch reflexes: preliminary studies. Neurosci. Lett.
12. Fisher, B. D., V. E. Baracos, and T. K. Shnitka. Ultrastructural events following acute muscle trauma. Med. Sci. Sports Exerc.
13. Friden, J. and R. L. Lieber. Structural and mechanical basis of exercise-induced muscle injury. Med. Sci. Sports Exerc.
14. Gardiner, P., R. Michel, C. Bowman, and E. Noble. Increased EMG of rat plantaris during locomotion following surgical removal of synergists. Brain Res.
15. Garrett, W. E. Muscle strain injuries: clinical and basic aspects. Med. Sci. Sports Exerc.
16. Garrett, W. E., M. R. Safran, and A. V. Seaber. Biomechanical comparison of stimulated and non-stimulated skeletal muscle pulled to failure. Am. J. Sports Med.
17. Harell, A., S. M. Mead, and E. Mueller. The problem of spasm in skeletal muscle. J.A.M.A.
18. Hemingway, M. A., H. J. Biedermann, and J. Inglis. Electromyographic recordings of paraspinal muscles: variations related to subcutaneous tissue thickness. Biofeedback Self-Regul.
19. Herring, S. A. Rehabilitation of muscle injuries.Med. Sci. Sports Exerc.
20. Hodgson, J. A., S. C. Bodine-Fowler, R. R. Roy, et al. Changes in recruitment of Rhesus soleus and gastrocnemius muscles following a 14 day spaceflight. Physiologist
21. Hodgson, J. A., F. Ibrahim, and D. L. McLellan. The relationship between soleus and gastrocnemius muscle activity in human movement. J. Physiol.
22. Hodgson, J. A. The relationship between soleus and gastrocnemius muscle activity in conscious cats-a model for motor unit recruitment. J. Physiol.
23. Hutchison, D. L., R. R. Roy, J. A. Hodgson, and V. R. Edgerton. EMG amplitude relationships between rat soleus and medial gastrocnemius during various motor tasks. Brain Res.
24. Hutton, R. S. and S. W. Atwater. Acute and chronic adaptations of muscle proprioceptors in response to increased use.Sports Med.
25. Hutton, R. S., K. Kaiya, S. Suzuki, and S. Watanabe. Post-contraction errors in human force production are reduced by muscle stretch. J. Physiol.
26. Johnson, E. W. The myth of skeletal muscle spasm.Am. J. Phys. Med. Rehabil.
27. Kendell, F. P. and E. K. McCreary. Muscle Testing and Function
, 3rd Ed. Baltimore: Williams & Wilkins, 1983, pp. 9-11.
28. Leskinen, T. P. Comparison of static and dynamic biomechanical models. Ergonomics
29. Lieber, R. L., T. M. Woodburn, and J. Friden. Muscle damage induced by eccentric contractions of 25% strain. J. Appl. Physiol.
30. Nikolaou, P. K., B. L. MacDonald, and R. R. Glisson. Biomechanical and histological evaluation of muscle after controlled strain injury. Am. J. Sports Med.
31. Proske, U. and D. L. Morgan. Tendon stiffness: methods of measurement and significance for the control of movement. J. Biomech.
32. Roy, R. R., J. A. Hodgson, J. Aragon, M. K. Day, I. B. Koslovskaya, and V. R. Edgerton. Recruitment of the Rhesus soleus and medial gastrocnemius before, during, and after spaceflight. J. Gravitational Physiol.
33. Roy, R. R., J. A. Hodgson, G. R. Chalmers, W. Buxton, and V. R. Edgerton. Responsiveness of the cat plantaris to functional overload. In: Medicine and Sports Science: Integration of Medical and Sport Sciences
, Vol. 37, Y. Sato, J. Poortmans, I. Hashimoto and Y. Oshida(Eds.). Basel, Switzerland: Karger Press, 1992, pp. 43-51.
34. Roy, R. R., D. L. Hutchison, D. J. Pierotti, J. A. Hodgson, and V. R. Edgerton. EMG patterns of rat ankle extensors and flexors during treadmill locomotion and swimming. J. Appl. Physiol.
35. Roy, R. R., K. M. Baldwin, and V. R. Edgerton. The plasticity of skeletal muscle: effects of neuromuscular activity.Exerc. Sport Sci. Rev.
36. Russell, B., D. J. Dix, and D. L. Haller. Repair of injured skeletal muscle: a molecular approach. Med. Sci. Sports Exerc.
37. Schouenborg, J., H. R. Weng, and H. Holmberg. Modular organization of spinal nociceptive reflexes: a new hypothesis.N.I.P.S.
38. Soderberg, G. L. and T. M. Cook. Electromyography in biomechanics. Phys. Ther.
39. Taylor, D. C. and J. D. Dalton. Experimental muscle strain injury. Am. J. Sports Med.
40. Thompson, S., J. E. Gregory, and U. Proske. Errors in force estimation can be explained by tendon organ desensitization. Exp. Brain Res.
41. Tidball, J. G. Myotendinous junction injury in relation to junction structure and molecular composition. Exerc. Sports Sci. Rev.
42. Travell, J. G. and D. G. Simons. Myofascial Pain and Dysfunction: The Trigger Point Manual
. Baltimore: Williams & Wilkins, 1983, p. 35.
43. Wolf, S. L., M. Nacht, and J. L. Kelly. EMG feedback training during dynamic movement for low back pain patients. Behav. Ther.