J. Physiol. (1987), 389, pp. 37-44 With 3 text-figures Printed in Great Britain

37

RELATIVE DISPLACEMENTS IN MUSCLE AND TENDON DURING HUMAN ARM MOVEMENTS

BY A.

AMIS*, A. PROCHAZKAt, D. SHORTt, P. ST. J. TRENDt

AND A. WARDt From the Department of Mechanical Engineering, Imperial College, London SW7 and the Departments of t Physiology and t Rheumatology, St. Thomas's Hospital, London SE1 *

(Received 8 September 1986) SUMMARY

1. X-ray, cine and video recordings were made of the movement of radio-opaque markers injected into the musculo-tendinous junctions of biceps brachii muscle. 2. In strong isometric contractions, the distal tendon of the long head of biceps lengthened by about 2% of its estimated rest length. 3. During voluntary isotonic elbow flexion-extension movements at frequencies up to 5-5 Hz there was no detectable phase shift between intramuscular and joint

displacements.

4. In the fastest alternating movements (5-5-6-7 Hz) small phase advances developed in the muscle. 5. We conclude that human tendons do stretch during muscle contraction, but not enough to cause intramuscular phase reversals in rapid unloaded movements. This in turn means that muscle spindles shorten and lengthen virtually in phase with joint movements under most conditions. INTRODUCTION

Tendons are often treated as inextensible cables, directly transmitting the length changes of muscle fibres to their bony attachments. Yet tendons are sometimes more compliant than the muscles themselves, and can contribute substantially to energy storage and retrieval during cyclical motor tasks (Hill, 1950; Cavagna, Heglund & Taylor, 1977). Indirect estimates have suggested that in an actively contracting mammalian muscle subjected to a small stretch (up to 2 % of muscle fibre rest length), a substantial proportion of the movement is taken up in the tendon (Morgan, Proske & Warren, 1978). The longer the tendon, the more displacement it absorbs. Indeed it was estimated that in wallaby gastrocnemius muscle, where fibre rest length was about one-seventh of in-series tendon length, stretch during maximal contraction produced eight times as much movement in the tendon as in the muscle fibres. Muscles moving inertial loads via compliant tendons would not only 'see' less of the over-all displacement, but beyond the resonant frequency of the tendon-load system, 180 deg phase advances of intramuscular displacement on limb displacement

38

A. AMIS AND OTHERS

could occur. Indeed Fellows & Rack (1986) recently showed that in electrically induced oscillatory movements about the human elbow, such a phase reversal was detectable in the biceps muscle at a frequency as low as 2 Hz. There are two important implications of resonances in tendon-load systems. First, movements at or near the resonant frequency would be sustainable with little modulation in muscle displacement. As such they would tend to arise during, and potentially interfere with, normal controlled movements. Secondly, at the resonant frequency there would be a reversal of phase between intramuscular displacement and origin-to-insertion displacement. The main displacement sensors, the muscle spindles, which respond to intramuscular length variations, would therefore convey out-of-phase information back to the spinal cord and supraspinal structures. Given that limb position or velocity may often be the controlled variable in motor tasks, these unexpected phase relationships between muscle spindle discharge and limb movement would require a reappraisal of current views on how the central nervous system controls movement. In our study, X-ray filming in subjects whose biceps muscles were injected with radio-opaque markers, showed that the tendons lengthened significantly in strong isometric contractions, but there were no phase reversals in voluntary isotonic flexion-extension movements at frequencies up to 7 Hz. We conclude that human tendons do stretch during muscle contraction, but not enough to cause intramuscular phase reversals in rapid unloaded movements. A preliminary report of some of these experiments has been presented to the Physiological Society (Prochazka & Trend, 1986). METHODS

We set out to test these notions in human subjects, and chose the long head of biceps brachii as the test muscle. Biceps, with its long tendons (proximal tendon 110 mm, distal tendon 70 mm, cf. 180 mm muscle belly as measured in a cadaver), and the relatively large mass it moves (moment of inertia of forearm typically 0-08 kg m2), would be expected to have a low resonant frequency. Subjects were comfortably seated beside an X-ray table with their left shoulder abducted and their left forearm semi-pronated. The left elbow was pushed against a firm block, the left wrist was splinted and the distal end of the forearm held a few centimetres clear of the table by a string attached to the ceiling. Voluntary elbow flexion-extension movements up to 5 Hz were easily performed; higher frequency movements up to 7 Hz required some practice and were usually of smaller amplitude. Isometric contractions were made against a stiff spring balance attached to a wristlet. In three subjects (the authors A. P., P.T. and A.W.) we injected 0-05-0-1 ml of radio-dense contrast (Omnipaque 350: Nyegaard, U.K. or Niopam 370: Merck, U.K.) into the musculo-tendinous junctions of biceps brachii and examined the relative movements of muscle, tendon and bone using X-ray cinematography (Siemens Pandoros, 50 frames/s) and video filming (Technicare DR 960, Sony Betamax C7, and Panasonic VHS recorders). All three subjects took part in two or three separate recording sessions spaced over a period of three months. The contrast images remained sharp for 4-5 min after injection, though they tended to change shape slightly during muscle contractions. This detracted from the accuracy of measurements of intramuscular displacement, and so in one subject (A.W.), three small, 0-5 mm diameter vitallium spheres were additionally injected into the biceps muscle proximally and distally (Fig. 1). These were subsequently discernible in video films of isometric contractions, and in cine films of rapid movements. 9 months after the injection of the spheres, they were visualized in X-ray stills, and were found to have migrated distad and towards the surface of the muscle.

MUSCLE DISPLACEMENT DURING MOVEMENT

Forearm

displacement Kmarkers

39

Skin surface markers

iopam streaks

'-c^

o-_-- -B A Distalvitallium balls

..... Proximal-

vitaHlium ball

100 mm

Fig. Tracing from a single frame of an X-ray cine film of voluntary arm movements. The subject's arm had been injected with 0'5 mm vitallium spheres some weeks earlier and two small quantities of a radio-dense marker, 'Niopam', just prior to filming. Elbow displacement was measured either from the outline of the radial tuberosity (point of insertion of biceps brachii), or from markers on an elastic band attached between the forearm and shoulder. Intramuscular displacement was measured between the proximal ball and the distal markers in hundreds of consecutive frames. A: proximal ball to distal Niopam streak. B: proximal to distal ball. C: proximal to deep distal ball. 1.

RESULTS

Isometric contractions In isometric contractions increasing from 0 to 200 N at the wrist (0-60 N m torque about the elbow) the distal biceps tendon in subject A.P. lengthened by about 1'5 + 0-5 mm (ca. 1'8 % ofthe estimated tendon rest length of 85 mm). On the assumption of equal strain in the proximal tendon (estimated rest length 130 mm), the total tendon lengthening was ca. 3-8 mm. This would be equivalent to about 4-6 deg elbow movement, for a biceps moment arm of 48 mm (Amis, Dowson & Wright, 1979).

Voluntary movements In voluntary alternating movements (0-7 Hz) the distal muscle markers, and, when they were visible, the proximal markers, moved in phase with the radius, though the proximal markers nearly always moved less. The net result was that the proximal and distal parts of the muscle came together during flexion and moved apart in extension. We observed this in all three subjects injected with radio-dense contrast. The video recorders used produced 'frozen' frames by interlacing two consecutive images. This resulted in flicker in the mid-range between forearm flexion and extension, precluding the localization of the markers at the resolution required for displacement measurements. At the turning points the images sharpened considerably, but their granularity still made reliable measurements difficult. On the other hand, when sequences of movements were played in slow motion (2-5 frames/ s), changes in the spacing between the markers could be identified with ease. To test

A. AMIS AND OTHERS

40

our own qualitative assessment of these video films, we enlisted the help of four independent observers, none of whom knew the purpose of the investigation. Each observer was shown two separate sequences of movements in slow motion (data from subject P.T.) and was asked to state (a) when the markers were coming together or (b) when they were moving apart. The observers' monitor screens had been masked so Extension

C

|

5 deg|

..

*-

*

FFlexion

'o

c

B a_

c

. '.. ". "." '"'..'

'.-' . . .....'"

.

. .... ·. ........

I

I

1 s

Fig. 2. Frame-by-frame analysis of joint and intramuscular displacements during selfpaced, oscillatory movements of the forearm. A, B and C: intramuscular displacements as shown in Fig. 1. Top trace: forearm displacement derived from external marker band. Note that the intramuscular movements are always broadly in phase with movements of the joint. as to blank out the image of the subject's elbow and forearm. On a separate screen, the experimenters could simultaneously observe the elbow movements, and noted whether the responses 'together' or 'apart' corresponded to elbow flexion or extension. The results of these trials were quite unequivocal. Of the twenty cycles shown to each observer and chosen at random from long sequences of movements, in all but one cycle, 'together' was associated with flexion and 'apart' with extension. To obtain more reliable quantitative data, one subject (A. W.) volunteered to have three 0-5 mm diameter vitallium spheres injected into his biceps muscle proximally and distally (Fig. 1). The spheres could be accurately located in video frames, though flicker and granularity were still a problem during movement. The required resolution was finally obtained with the use of cine filming. Fig. 2 shows the results of a frameby-frame analysis of a cine film of forearm movements ranging in frequency from 2-9 to 6l7 Hz, with the accompanying intramuscular displacements measured between the pairs of proximal and distal markers as identified in Fig. 1. Frames were projected onto a blank white wall so that the image of the arm appeared twice life-size. A strip of white card 250 mm long was hung from threads pinned to the wall, so that the images of the Niopam streaks and the vitallium balls (Fig. 1) fell onto it. The outline of the proximal Niopam streak and the proximal vitallium ball was traced onto the card. At the distal end, a thin strip of millimetre graph paper (50 mm long, 2 mm high) was stuck to the card at the location of the distal balls and marker. As each frame was projected, one experimenter adjusted the

41 MUSCLE DISPLACEMENT DURING MOVEMENT card so that the proximal tracings exactly covered the proximal markers. A sharp pointer anchored the card at the position of the proximal ball. A second experimenter then pivoted the distal end of the card so as to align the millimetre strip with the distal markers, and took readings of displacement along lines A, B and C (Fig. 1). The card was then moved so that the distance between the forearm displacement markers nearest the skin surface markers could be measured. These latter readings were converted into angular displacement by calibrating them against angular changes of the traced outlines of the humerus and radius. Extension E

I

5 deg

u-

V°

Flexion C

E

2I

*

*

c C

A

25

*

advances of the intramuscular movements on joint movements (top). The curves in this Figure were fitted by eye, but quantitative curve fitting showed the phase advances to be: A, 44 deg; B, 12 deg; and C, 15 deg. The durations of the contributing cycles ranged from 150 to 180 ms, corresponding to frequencies of 5-6-6-7 Hz. These were the highest frequencies which this subject could produce voluntarily.

In the range 0-55 Hz, there was no detectable phase shift between the intramuscular and joint displacements. However, in the fastest movements in Fig. 2 (5-6-6-7 Hz), small phase advances developed in the muscle. This is shown more clearly in Fig. 3, in which the first ten cycles of Fig. 2 are averaged. The averaging was performed on the expanded original of Fig. 2 by overlaying a transparent grid which divided a cycle of movement into sixteen bins. For each cycle, the grid was aligned to the onset of extension in the forearm displacement trace, and samples occurring within a particular bin were entered into a cumulative sum. Means were calculated for each bin and plotted in Fig. 3. The intramuscular phase advances were calculated in traces A, B and C by sinusoidal curve-fitting software written for a BBC microcomputer by Miss B. Band. The phase advances were 44 deg (trace A), 12 deg (trace B) and 15 deg (trace C). The differences in phase advance suggest that some portions of a muscle may be activated marginally sooner than others. Though phase reversal did not occur within the frequency range which our subjects were capable of producing voluntarily, this does not rule out that some of the

42

A. AMIS AND OTHERS

displacement between origin and insertion might have been taken up in the tendon. For an elbow flexor moment arm of 41 mm estimated in our subject from Fig. 1 (cf. mean of 48 mm measured in six cadavers: Amis et al. 1979), the 6-9 deg elbow movements in Fig. 3 would have required 4-9 mm of origin-to-insertion displacement. The mean intramuscular displacement in traces A, B and C (4-1 mm peak-to-peak: Fig. 3) accounts for most of this. The intramuscular markers were approximately 100 mm apart, or about 70% of the mean muscle fibre length (Amis et al. 1979). Some further intramuscular displacement might well have occurred beyond the markers, leaving relatively little to be accounted for by the tendons. DISCUSSION

The spring-like properties of tendon might in some instances provide a valuable mechanism for energy storage (Cavagna et al. 1977; Morgan et al. 1978). However, from a control point of view, the non-linear sharing of origin-to-insertion displacement by muscle fibres and tendon would be expected to complicate the feedback signals from the main intramuscular proprioceptors, the muscle spindles (Rack & Westbury, 1984).

Slowly varying forces In maximal isometric contractions we found that the distal biceps tendon lengthened by about 1'8% of its estimated rest length, a displacement which, if added to that presumed to occur in the proximal tendon, was equivalent to some 4-6 deg of elbow displacement. There is substantial evidence from chronic animal recordings that muscle spindle endings provide the c.N.S. with tonic displacement-related signals (for review see Prochazka, 1986). As the muscle spindles generally lie entirely within the muscle fibre component of a muscle-tendon, it follows that the displacement-related afferent signal must to some extent be force dependent. Whether this force dependence should be viewed as a source of error or whether it might provide useful supplementary information is not clear. Furthermore, the importance of tendon compliance probably varies from muscle to muscle, depending among other things on the relative length of the muscle fibres and tendons, and on the maximal physiological stress occurring within the muscle in question. The latter would of course be strongly affected by changes in the sharing of force among synergistic muscles (Fellows & Rack, 1987).

Rapidly varying movements Methodology. Fellows & Rack (1986) obtained a measure of intramuscular displacement during tremor-like movements by inserting needles into human biceps muscles. Displacement was measured between the points at which the needles entered the skin, and this was assumed to indicate intramuscular displacement. The authors felt that as the skin over the muscles was freely mobile, and the needles emerged unbent, the skin had not seriously impeded the needles' movements. In our own Xray recordings, we were struck by the amount of relative movement between skin and muscle. For movements such as those in Fig. 3, the skin would not only be stretched between the needles, but it would also be pulled proximally and distally

43 MUSCLE DISPLACEMENT DURING MOVEMENT with respect to its normal position. Given the small diameter of the needles (0-5 mm), the local pressure developed to achieve the required skin distortion could be quite substantial. Without X-ray verification, we felt that there was some doubt as to whether the tips of the needles remained stationary at single sites in the muscle, and whether phase shifts were not introduced by the viscoelastic impedence represented by the skin. Our technique allows the displacement between discrete points deep within the biceps muscle to be measured, and as such provides data complementary to that of Fellows & Rack (1987). The disadvantages of our approach include the resolution problem with single-frame video images (overcome with cine filming), the limited allowable duration of X-ray screening, and the initial difficulty we encountered in adjusting the X-ray filming parameters to give sharp images of bone, soft tissues and radio-opaque markers simultaneously. Intramuscular displacement. Fellows & Rack (1986) showed that when biceps and triceps brachii were stimulated electrically with alternating impulse doublets of increasing repetition rate, the displacement of the needles inserted into biceps showed a sudden 180 deg phase advance on joint movement at about 2 Hz. This was interpreted as arising from the compliant nature of tendons, which complicates the relationship between joint movement and muscle fibre length. In further experiments, Fellows & Rack (1987) found that the phase reversals did not occur during similar movements produced voluntarily by the subjects at frequencies up to 5 Hz, a result more in line with our own preliminary observations (Prochazka & Trend,

1986).

The authors felt, quite reasonably, that the discrepancy was due to co-activation of muscles in the voluntary movements, which would have kept the tendons under continual, albeit varying, tension, thus avoiding the low-tension range where compliance is greatest. It was suggested that whereas brachialis and brachio-radialis would have contributed to the voluntary movements, electrical stimulation singled out biceps alone (though in our experience it is difficult to obtain such selectivity with monopolar stimulation) and this would have increased the stress variations in this muscle and its tendon. We would also point out that double-shock stimulation per cycle would exaggerate the intramuscular stress variations involved in producing alternating movements of a given amplitude. E.m.g. recordings show that in selfpaced 4-6 Hz movements, biceps is active for about half of each cycle (personal observations) and so a pulsatile activation of muscle is unlikely to occur normally (though it might occur in pathological conditions).

Tendon compliance and tremor The phase reversals at 2 Hz reported by Fellows & Rack (1986, 1987) suggested resonant behaviour of the tendon and its inertial load. As pointed out above, movements at the resonant frequency would be sustainable with little modulation in intramuscular displacement, and a minimal expenditure of energy. As such, they might be expected to occur frequently during the course of controlled movements, unless actively suppressed. The present results indicate that in normal subjects performing voluntary movements, this sort of resonant behaviour is most unlikely. However, the small phase advances we detected during movements at 6-7 Hz, are

44

A. AMIS AND OTHERS

consistent with a possible phase reversal at frequencies beyond the voluntary range. The over-all conclusion of our study is that in the unloaded forearm, intramuscular displacements are in phase with displacements of the elbow for all but the fastest voluntary movements. We thank Dr R. P. Hicks and Mr R. Durado for their help with the X-ray equipment, Mr S. Vincent for technical assistance and Miss B. Band for the curve-fitting software. Dr P. St. J. Trend is supported by the British MRC as a Training Fellow. The St. Thomas's Hospital (Research) Endowments Committee supplied equipment. REFERENCES

AMIS, A. A., DowsoN, D. & WRIGHT, V. (1979). Muscle strengths and musculo-skeletal geometry of the upper limb. Engineering in Medicine 8, 41-48. CAVAGNA, G. A., HEGLUND, N. C. & TAYLOR, C. R. (1977). Mechanical work in terrestial locomotion: two basic mechanisms for minimizing energy expenditure. American Journal of Physiology 233, R243-261. FELLOWS, S. J. & RACK, P. M. H. (1986). Relation of the length of the electrically stimulated human biceps to elbow movement. Journal of Physiology 376, 58P. FELLOWS, S. J. & RACK, P. M. H. (1987). Changes in the length of the human biceps brachii muscle during elbow movements. Journal of Physiology 383, 405-412. HILL, A. V. (1950). The series elastic component of muscle. Proceedings of the Royal Society B 137, 273-280.

MORGAN, D. L., PROSKE, U. & WARREN, D. (1978). Measurements of muscle stiffness and the mechanism of elastic storage of energy in hopping kangaroos. Journal of Physiology 282, 253261.

PROCHAZKA, A. (1986). Proprioception during voluntary movement. Canadian Journal of Physiology and Pharmacology 64, 499-504. PROCHAZKA, A. & TREND, P. (1986). X-ray imaging of human muscle contraction. Journal of Physiology 377, 15P. RACK, P. M. H. & WESTBURY, D. R. (1984). Elastic properties of the cat soleus tendon and their functional importance. Journal of Physiology 347, 479-495.