INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

HUMAN LOCOMOTION INVOLVES a complex series of muscular interactions and coordinated movements. Although the kinematics and dynamics of walking and running are well studied, no reliable method exists for measuring force production of individual muscles during locomotion in humans. Information on the forces produced by individual skeletal muscles during locomotion will improve our understanding of muscle physiology, musculoskeletal mechanics, neuromuscular coordination, and motor control. Such information may also aid development of exercise hardware and protocols for physical rehabilitation and training.

In the past, investigators have used mathematical modeling (3, 4, 22, 30) and electromyography (EMG; 16, 18, 24, 27) to estimate the contributions of individual muscles to joint moments during exercise. However, these indirect methods exhibit deficiencies related to the complex nature of human locomotion. With the aid of photography, force platforms, and mathematical models, much is known about the kinematics and dynamics of walking and running. However, because factors such as contraction velocity, muscle length, mode of contraction, muscle architecture, and joint mechanics all affect individual muscle contraction force (19), mathematical models of individual muscles during dynamic activities are extremely complex and often inaccurate (5, 22). Kinematic analyses of locomotion commonly describe actions of muscle groups, but moment contributions of individual muscles are difficult to discern.

Although EMG patterns provide useful information about the phasic electrical activity of muscle, attempts to use EMG magnitude as an index of dynamic muscle contraction force have proved largely unsuccessful. Various disadvantages of this method include nonlinear EMG-force calibration curves, fatigue-related changes, and low reproducibility (24, 26, 27). Although integrated or root mean square EMG is linearly related to individual muscle contraction force during isometric exercise in many muscles (12, 16), this association is unreliable during dynamic activities that involve concentric and/or eccentric movements (1, 23, 27). Because the EMG-force relationship varies with mode and velocity of contraction, EMG is an unreliable index of muscle contraction force during locomotion.

By using a buckle transducer for recording tendon forces, a number of experiments have been performed to measure individual muscle forces in cats during dynamic activities (8, 10, 29). This approach has also been applied to humans, with a buckle transducer surgically implanted around the Achilles tendon (15). Although the buckle transducer is a valuable tool for measuring in vivo tendon tension in animals, inherent surgical risks, subject discomfort, and long recovery periods make it impractical for regular use. Furthermore, a buckle transducer on the Achilles tendon is unable to differentiate between individual contributions of the soleus and gastrocnemius to total tendon tension.

Intramuscular pressure (IMP), or fluid pressure within a muscle, increases linearly with individual muscle contraction force during isometric, concentric, and eccentric activity (1, 17, 23, 25, 26). IMP elevation results directly from increased muscle fiber tension and therefore reflects the mechanical state within the muscle independent of muscle length and muscle activation (1, 9). Thus IMP may be used as a qualitative index of muscle contraction force: the higher the IMP, the higher the force. Furthermore, calibration of IMP values with joint torque may provide quantitative estimates of individual muscle contraction force if the contraction force of that particular muscle can be isolated by a dynamometer. Previous studies of IMP during locomotion, however, were limited by relatively large catheter diameter and by hydrostatic pressure artifacts associated with using saline-filled catheter lines.

Therefore, the purpose of this investigation was to assess IMP as an index of soleus and tibialis anterior function during gait, using a relatively small-diameter, transducer-tipped catheter. These muscles are of particular interest because they are two of the primary muscles of the lower leg involved in locomotion, and a substantial amount of soleus and tibialis anterior EMG and IMP data exist in the literature (1-3, 8, 13, 14, 20, 24, 25, 28, 30). We hypothesized that transducer-tipped catheters provide rapid and reproducible measurements of IMP during walking and running and that IMP parallels muscle force production patterns predicted by kinematic analysis and tendon buckle-transducer measurements.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Subjects. Ten volunteers [age 20-48 (SD) yr; weight 72 ± 13 kg] participated in this investigation, after providing informed, written consent. All subjects were in good health, as determined by comprehensive medical examination. Subjects refrained from caffeine, alcohol, medications, and strenuous exercise for 24 h before study. The protocol was approved by the Human Research Institutional Review Board at the National Aeronautics and Space Adminstration-Ames Research Center.

Instrumentation. IMPs were measured in the soleus and tibialis anterior muscles of the left leg during treadmill walking and running. A 2- or 3-F transducer-tipped catheter (Millar Mikro-Tip, Houston, TX) measured IMP. Before insertion, each catheter was calibrated electronically to maintain sterility. The catheter insertion site for the tibialis anterior was ~3 cm distal and 1 cm lateral to the tibial tuberosity. Soleus catheter insertion was on the posterolateral aspect of the leg, one-third of the distance between the lateral malleolus and the lateral tibial condyle. Each insertion site was first shaved and cleaned with alcohol and Betadine iodine solution. Shaving aids visualization and palpation of muscles to be catheterized. Also, subjects plantar flexed (hip and knee at 90°) and dorsiflexed their feet isometrically to aid visualization of the soleus and tibialis anterior, respectively. In this manner, the tibialis anterior and soleus both lend themselves to identification in nonobese subjects.

Treadmill gait protocol (n = 10). All subjects were familiarized with the protocol and practiced walking and running on the treadmill (Aerobics, Little Falls, NJ) before catheter insertion. Self-selected walking and running speeds for each subject were determined during this familiarization. Self-selected walking speed averaged 1.3 ± 0.3 m/s, whereas self-selected running speed averaged 2.8 ± 0.6 m/s.

Calibration of IMP (n = 2). To convert IMP values into estimated moment contributions of the soleus during walking and running, two of the subjects also performed plantar flexion and dorsiflexion exercises by using a Lido Active isokinetic dynamometer (Loredan Biomedical, Davis, CA) before treadmill exercise. Isometric, concentric, and eccentric contractions were performed. Subjects were positioned and secured with the left knee and hip joints flexed at 90°, thus minimizing contributions of the gastrocnemius to soleus contractions (1, 28). Subjects wore their own athletic shoes, and the left foot was secured to the Lido footplate by two Velcro straps. Ankle neutral position was defined as a 90° angle between foot and tibia. Limits of ankle range of motion were then determined by passive plantar flexion and dorsiflexion of the ankle joint. Isometric contractions were performed at five different joint angles (spaced by ~10°) covering the entire range of motion. Concentric isokinetic contractions were performed at 60, 120, and 240°/s. Eccentric isokinetic contractions were performed at 30, 60, and 120°/s. At each joint angle (in the case of isometric contractions) or velocity (for concentric and eccentric contractions), subjects performed at least four contractions of intensity approximating 100, 75, 50, and 25% of maximal voluntary effort. Subjects rested for ~3 min between each mode of contraction. During each set of contractions, IMP, footplate velocity, and ankle joint torque and angle were continuously recorded at a rate of 100 Hz.

Data analysis. Data were normalized to represent 0 (heel strike) to 100% of each step cycle. A spline interpolation was performed on each data set. Interpolated data were then resampled at 1% intervals to synchronize data points within and across subjects. For each subject, representative traces (showing soleus and tibialis anterior IMP patterns) were produced by calculating means across four step cycles at 1% increments of the cycle. Positions of peak IMP with respect to the normalized gait cycle were recorded, and means (±SE) across subjects were calculated. Paired t-tests identified statistically significant differences between IMP peaks at  = 0.05.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Soleus and tibialis anterior IMP from one representative subject are illustrated in Fig. 1. IMPs within each subject were quite uniform (maximum intrasubject SD equaled 10 mmHg for soleus and 8 mmHg for the tibialis anterior), despite relatively larger variability among subjects in IMP magnitude.

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Soleus (solid lines) and tibialis anterior (heavy dashed lines) intramuscular pressures (IMP) in 1 representative subject during walking (A) and running (B). Each trace represents a mean of 4 step cycles, sampled at 1% intervals. Light dashed lines, SDs over 4 cycles.

In all subjects, soleus IMP closely paralleled GRF_z during the late-stance phase of gait, with single peaks during walking [181 ± 22 (SE) mmHg at 53 ± 1% of gait cycle] and running (269 ± 30 mmHg at 20 ± 1%) (Fig. 2). IMP patterns in the tibialis anterior were somewhat more variable but consistently showed a biphasic response during both walking and running. During walking, the first peak (90 ± 15 mmHg) occurred shortly after heel strike (6 ± 1%), and the second peak was smaller in amplitude (67 ± 11 mmHg) and occurred near toe-off (48 ± 0%). The same pattern was evident during running, with the first tibialis anterior IMP peak averaging 151 ± 25 mmHg (at 3 ± 0% of gait cycle) and the second, smaller peak averaging 109 ± 21 mmHg (at 19 ± 1% of gait cycle). Average peak IMPs during rest and treadmill exercise are given in Table 1. Average self-selected running speed was about twice as fast as self-selected walking speed.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Soleus (solid lines) and tibialis anterior (dashed lines) IMPs during walking (A) and running (B) averaged across all subjects (n = 10). Walking speed averaged 1.3 ± 0.3 m/s; running speed averaged 2.8 ± 0.6 m/s.

In the two subjects who performed dynamometric calibrations before treadmill exercise, linear regression of IMP vs. ankle joint torque produced the following relationships (Fig. 3): IMP = 2.53(torque) + 0.29 [r = 0.97] and IMP = 1.45(torque) + 0.71 [r = 0.97]. Application of these relationships to IMP data during gait yielded estimated peak soleus moment contributions of 0.96-1.40 and 0.95-1.65 N · m · kg body wt¹ for subjects A and B, respectively, during walking and 1.43-1.68 and 1.93-2.70 N · m · kg body wt¹ for subjects A and B, respectively, during running (Fig. 4). In both subjects, peak IMP increased with gait speed.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Dynamometric calibration of soleus IMP with torque during isometric, concentric, and eccentric contractions in 2 subjects [subject A (A); subject B (B)]. Each point represents peak IMP and torque of a single contraction. Linear regression equations were later used to convert IMP values obtained during locomotion into moment contributions of the soleus. con, Concentric; ecc, eccentric.

View larger version (37K): [in this window] [in a new window]   Fig. 4.   Effect of locomotion speed on IMP and moment contributions of soleus [subject A (A); subject B (B)]. Moment values, expressed as N · m · kg body wt1, were derived by using regression equations from Fig. 3. In both subjects, peak IMP increased with speed of walking (solid lines) and running (dashed lines).

None of the subjects reported undue discomfort because of catheter placement or exercise. In two subjects, reliable IMP data from the tibialis anterior were not obtained because of catheter movement or malfunction.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

These results demonstrate that patterns of IMP development in the soleus and tibialis anterior during walking and running are similar to patterns of estimated ankle joint moments (9, 22, 30), Achilles tendon tension measured with a buckle transducer (15), and qualitative patterns of phasic EMG activation (2, 30) (Fig. 5). The soleus exerts a single peak in IMP near push-off, when the ankle joint is undergoing active plantar flexion. Pressure patterns in the tibialis anterior during walking and running are biphasic in nature. The first peak occurs near heel strike as the tibialis anterior is actively contracting to stabilize the ankle joint. The second peak, significantly smaller in amplitude than the first, occurs near the end of the stance phase as the tibialis anterior is eccentrically activated to help stabilize the ankle joint during push-off. Although significant, the difference in magnitude between the two tibialis anterior IMP peaks is not as dramatic as published EMG activation patterns might suggest (Ref. 30, Fig. 5). For example, the tibialis anterior EMG trace in Fig. 5 shows only a small peak at walking push-off (~15% as great as the peak that occurs at heel strike) and a relatively large peak as the foot is dorsiflexed during the swing phase (at ~75% of gait cycle). Because the tibialis anterior is eccentrically coactivated during the push-off phase of the step cycle, and eccentric contractions are known to generate more force per unit EMG (28), it is likely that actual tension in the muscle exceeds tension estimated by EMG.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Comparison of IMP from present study (A; n = 10) with qualitative patterns of EMG activity (by using surface electrodes) (30), ankle joint moment (by using joint kinematics) (30), and Achilles tendon tension (by using a buckle transducer) (15) during walking (B).

Although qualitative patterns of IMP development during locomotion agree generally with phasic EMG activity, the utility of IMP measurement lies in the magnitudes of pressure (1, 25). Because IMP is a physical property related to force development in a muscle, fluid pressure in a muscle increases linearly with increasing tension, apparently regardless of contraction velocity, joint angle, and mode of contraction (Fig. 3), all of which continuously change during dynamic activities. Consider the soleus, for example: at heel strike and through the beginning of the stance phase, the soleus is eccentrically activated (being lengthened during a contraction effort). As the stance phase progresses, the soleus actively contracts and eventually shortens, helping to propel the body forward. Variations in muscle length, contraction mode, and contraction velocity during dynamic activities are major reasons why EMG is unreliable for determining contraction force of individual muscles. Although IMP appears to be directly and linearly related to contraction force, its main disadvantage is the invasiveness of present measurement techniques.

Application of IMP-torque regression equations to the IMP data collected in the present study yields estimated soleus moment contributions of 1.18 and 1.39 N · m · kg body wt¹ (subjects A and B, respectively) during normal (1.25 m/s) walking. Although other plantar flexors, particularly the gastrocnemius, are known to contribute to plantar flexion torque during walking, the soleus is the dominant contributor. Therefore, estimated moment contributions of the soleus presented here agree quite well with those of Winter (30), Groh and Baumann (9), and Cappozzo and co-workers (5), who reported combined plantar flexor torques during walking of 1.5-2.4 N · m · kg body wt¹.

IMP peaks in the soleus and tibialis anterior were higher in all subjects during running than walking (Table 1), indicating increased muscle tension during the stance phase of running. Furthermore, in the two subjects who exercised at multiple speeds of both walking and running, estimated moment contributions of the soleus increased with each increase in treadmill speed. Kirby and co-workers (13) reported similar increases in peak tibialis anterior IMP with increased speed of locomotion.

Although the transducer-tipped catheters used in this investigation generally performed well, negative IMP spikes were sometimes evident during muscle activity immediately after insertion. These spikes usually disappeared after 1-3 min of palpation and muscle contraction, probably as interstitial fluid filled the space above the sensor surface. In a few instances, however, negative spikes persisted during exercise (Fig. 4A). Negative relaxation pressures have been reported previously (6) and may result from slight movement of the catheter during contraction or location of the catheter tip in muscle tissue close to bone or tendon. Alternatively, one might hypothesize that negative or below-baseline relaxation pressures are not due to measurement artifact and instead are physiologically important for muscle perfusion after contraction.

It should be noted that the slope of the IMP-force relationship, while linear, varies both within and between muscles (11, 12, 25, 26). Variations between muscles depend on muscle thickness, fiber curvature, pennation angle, and other factors related to muscle architecture. Within a muscle, pressure increases with depth (21, 26). Repeated catheterizations of the same muscle may show slight differences in baseline pressure and magnitude of pressure response because of variations in positioning of the pressure sensor. When measured in a single location, however, IMP responses to muscle contraction are highly reproducible (1, 26). It is therefore important to ensure that the transducer is in the same position during IMP-torque calibrations as dynamic exercise testing.

During dynamometric calibrations, total torque measured by the dynamometer was probably affected both agonistically and antagonistically by other muscles in the foot and lower leg. Although holding the knee and hip joints at 90° of flexion during dynamometry helped maximize soleus contribution to plantar flexion torques (1, 28), the net effect of surrounding muscles is unknown. Therefore, the linear regression equations of IMP vs. ankle joint torque provide only estimates of soleus moment contributions. Because of the difficulty in isolating forces produced by individual muscles, in vivo calibration of IMP values with torque or force may not be possible in all muscles. Lack of an accurate standard against which to compare IMP-derived moment contributions further illustrates the need for a reliable, reproducible method of monitoring contraction force of specific muscles in vivo.

Our results support the use of IMP measurement to assess function of individual muscles during locomotion in humans. Because IMP magnitude is directly related to muscle force output, measurement of IMP during dynamic exercise provides a valuable index of individual muscle force during locomotion and other dynamic activities.